Organic Coatings
Thả NguyễnRelated papers
Functional Coatings and Microencapsulation: A General Perspective
Functional Coatings, 2006
Coatings Technology Handbook
The world of coatings is very broad. The application techniques are many, and the uses are numerous. Technical people need to be aware of many things. One study said that a coating chemist must be proficient in 27 different disciplines. This book is directed at supplying a broad cross-index of some of the different aspects to help the technical person. It is not meant to be an in-depth treatise on any subject. It is meant to give insight into the various subjects covered. The chapter authors or the editor may be contacted for more information or direction on the subjects. To aid the person involved in coatings, inks, or adhesives, be they chemists, engineers, technicians, researchers, or manufacturers, chapters are given in the areas of fundamentals and testing, coating and processing, techniques and materials, and surface coatings. Each section contains information to expand the awareness and knowledge of someone practicing in the field. The objective is to help people solve problems and increase their level of technology. With time, technology increases, as shown by the chapter on statistical design of experiments, and the chapter on using equipment to determine ultraviolet (UV) resistance. Newer materials such as fluorocarbon resins, polyurethane thickeners, and high-temperature pigments are included as well as older materials such as alkyds, clays, and driers. To accomplish the presentation of technology, this book has been expanded to 118 chapters by adding new material and updating other material. Hopefully, the reader will expand his or her knowledge and further push the envelope of technology. The editor gratefully acknowledges the many contributions of the chapter authors and the publishers who have made this book possible.
Environmentally Friendly Coatings : Past, Present Developments, and Future
Journal of the Japan Society of Colour Material, 1993
Progress in Organic Coatings
2008
Our laboratory recently began work on the use of room temperature ionic liquids ((RTIL’s) to enhance our capabilities for the electrochemical characterization of organic coatings [A.M. Simões, D. Tallman, G.P. Bierwagen, The use of ionic liquids for the electrochemical characterization of water transport in organic coatings, Electrochem. Solid-State Lett. 8 (2005) 60]. The RTIL’s are electrically conductive liquids consisting of large molecules that can be used to investigate the electrochemical properties of coatings in a non-aqueous medium. The enhancement of coating characterization comes from the fact that RTIL’s have sufficient conductivity to be an immersion medium for electrochemical measurements, but they do not directly penetrate and effect organic coatings as do aqueous electrolyte solutions. This allows the separate examination of the effects of water on coatings in immersion or cyclic exposure. Indeed, our initial studies showed that a hydrophilic RTIL could be used to e...
Nanotechnology-A New Prospective in Organic Coating -Review
International Journal of Chemical Engineering and Applications, 2011
New challenges for R&D in coating resins
Progress in Organic Coatings, 2002
In Coatings, Adhesives, and Laminates
Polymer Grafting and Crosslinking, 2008
Organic Coatings
Organic Coatings
Science and Technology
Third Edition
Zeno W. Wicks, Jr.
Frank N. Jones
Eastern Michigan University
S. Peter Pappas
Consultant
Douglas A. Wicks
University of Southern Mississippi
Cover: Scanning electron micrograph of alumina flakes. Courtesy of F. J. Maile, G. Pfaff, and P. Reynders, Merck
KGaA, Pigments R&D. Reprinted from Progress in Organic Coatings, Vol. 54(1), F. J. Maile, G. Pfaff, and
P. Reynders, “Effect pigments—past, present, and future,” pp. 150–163, 2005, with permission from Elsevier.
Copyright # 2007 by John Wiley & Sons, Inc. All rights reserved.
Published by John Wiley & Sons, Inc., Hoboken, New Jersey.
Published simultaneously in Canada.
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by
any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted
under Section 107 & 108 of the 1976 United States Copyright Act, without either the prior written
permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the
Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978)
750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be
addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken,
NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in
preparing this book, they make no representations or warranties with respect to the accuracy or
completeness of the contents of this book and specifically disclaim any implied warranties of merchantability
or fitness for a particular purpose. No warranty may be created or extended by sales representatives or
written sales materials. The advice and strategies contained herein may not be suitable for your situation.
You should consult with a professional where appropriate. Neither the publisher nor author shall be liable
for any loss of profit or any other commercial damages, including but not limited to special, incidental,
consequential, or other damages.
For general information on our other products and services or for technical support, please contact our
Customer Care Department within the United States at (800) 762-2974, outside the United States at
(317) 572-3993 or fax (317) 572-4002.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may
not be available in electronic formats. For more information about Wiley products, visit our web site at
www.wiley.com.
Library of Congress Cataloging-in-Publication Data:
Wicks, Zeno W.
Organic coatings : science and technology / Zeno W. Wicks, Jr., Frank N.
Jones, S. Peter Pappas; consultant, Douglas A. Wicks.
p. cm.
Includes index.
ISBN 978-0-471-69806-7 (cloth)
1. Plastic coatings. I. Jones, Frank N., 1936– II. Pappas, S. Peter
(Socrates Peter), 1936– III. Title.
TP1175.S6W56 2006
6670 .9- -dc22
2006013297
Printed in the United States of America
10 9
8
7
6 5
4
3 2
1
Contents
Preface
xix
Symbols and Units
xxi
1.
What Are Coatings?
1
1.1. Definitions and Scope / 1
1.2. Composition of Coatings / 4
References / 6
2.
Polymerization and Film Formation
7
2.1. Polymers / 7
2.1.1. Molecular Weight / 9
2.1.2. Morphology / 13
2.2. Polymerization / 17
2.2.1. Chain-Growth Polymerization / 17
2.2.2. Step-Growth Polymerization / 23
2.3. Film Formation / 27
2.3.1. Film Formation by Solvent Evaporation from Solutions of
Thermoplastic Binders / 28
2.3.2. Film Formation from Solutions of Thermosetting Resins / 28
2.3.3. Film Formation by Coalescence of Polymer Particles / 34
General References / 38
References / 39
3.
Flow
41
3.1. Shear Flow / 41
3.2. Types of Shear Flow / 42
v
vi
CONTENTS
3.3. Determination of Shear Viscosity / 46
3.3.1. Capillary Viscometers / 47
3.3.2. Rheometers / 48
3.3.3. Rotating Disk Viscometers / 49
3.3.4. Bubble Viscometers / 50
3.3.5. Efflux Cups / 50
3.3.6. Paddle Viscometers / 51
3.4. Shear Viscosity of Resin Solutions / 52
3.4.1. Temperature Dependence of Viscosity / 53
3.4.2. Dilute Polymer Solution Viscosity / 54
3.4.3. Concentrated Polymer Solution Viscosity / 55
3.5. Viscosity of Liquids with Dispersed Phases / 59
3.5.1. Thickeners for Latex Coatings / 62
3.6. Other Modes of Flow / 63
3.6.1. Turbulent Flow / 63
3.6.2. Normal Force Flow / 64
3.6.3. Extensional Flow / 64
General References / 66
References / 66
4.
Mechanical Properties
68
4.1. Introduction: General Comments on Testing of Coating Films / 68
4.2. Basic Mechanical Properties / 70
4.3. Formability and Flexibility / 74
4.3.1. Fracture Mechanics / 82
4.4. Abrasion and Mar Resistance / 83
4.4.1. Abrasion Resistance / 83
4.4.2. Mar Resistance / 85
4.5. Measurement of Mechanical Properties / 86
4.6. Tests of Coatings on Substrates / 89
4.6.1. Field Exposure Tests / 89
4.6.2. Laboratory Simulation Tests / 89
4.6.3. Empirical Tests / 90
General References / 94
References / 95
5.
Exterior Durability
5.1. Photoinitiated Oxidative Degradation / 97
5.2. Photostabilization / 100
5.2.1. UV Absorbers and Excited State Quenchers / 101
5.2.2. Antioxidants / 103
97
CONTENTS
vii
5.2.3. Hindered Amine Light Stabilizers / 104
5.2.4. Pigmentation Effects / 106
5.3. Degradation of Chlorinated Resins / 108
5.4. Hydrolytic Degradation / 109
5.5. Other Modes of Failure on Exterior Exposure / 111
5.6. Testing for Exterior Durability / 112
5.6.1. Accelerated Outdoor Testing / 112
5.6.2. Analysis of Chemical and Mechanical Changes / 114
5.6.3. Accelerated Weathering Devices / 116
General References / 118
References / 118
6.
Adhesion
121
Surface Mechanical Effects on Adhesion / 121
Effects of Internal Stress and Fracture Mechanics / 123
Relationship Between Wetting and Adhesion / 124
Adhesion to Metal Surfaces / 125
6.4.1. Surface Preparation / 126
6.4.2. Coating –Substrate Interaction / 128
6.4.3. Covalent Bonding to Glass and Metal Substrates / 130
6.5. Adhesion to Plastics and to Coatings / 131
6.6. Testing for Adhesion / 133
General References / 135
References / 135
6.1.
6.2.
6.3.
6.4.
7.
Corrosion Protection by Coatings
7.1. Corrosion of Uncoated Steel / 137
7.2. Corrosion Protection of Metals / 140
7.2.1. Passivation: Anodic Protection / 140
7.2.2. Cathodic Protection / 141
7.2.3. Barrier Protection and Inhibition / 142
7.3. Corrosion Protection by Intact Coatings / 142
7.3.1. Critical Factors / 143
7.3.2. Adhesion for Corrosion Protection / 144
7.3.3. Factors Affecting Oxygen and Water Permeability / 146
7.4. Corrosion Protection by Nonintact Films / 148
7.4.1. Minimizing Growth of Imperfections:
Cathodic Delamination / 149
7.4.2. Primers with Passivating Pigments / 149
7.4.3. Cathodic Protection by Zinc-Rich Primers / 151
137
viii
CONTENTS
7.5. Evaluation and Testing / 153
General References / 156
References / 156
8.
Acrylic Resins
159
8.1. Thermoplastic Acrylic Resins / 159
8.2. Thermosetting Acrylic Resins / 160
8.2.1. Hydroxy-Functional Acrylic Resins / 160
8.2.2. Acrylics with Other Functional Groups / 165
8.3. Water-Reducible Thermosetting Acrylic Resins / 166
References / 174
9.
Latexes
176
9.1. Emulsion Polymerization / 177
9.1.1. Raw Materials for Emulsion Polymerization / 178
9.1.2. Emulsion Polymerization Variables / 182
9.1.3. Sequential Polymerization / 186
9.2. Acrylic Latexes / 186
9.3. Vinyl Ester Latexes / 191
9.4. Thermosetting Latexes / 193
General References / 196
References / 196
10.
Polyester Resins
199
10.1. Hydroxy-Terminated Polyester Resins for Conventional
Solids Coatings / 200
10.1.1. Selection of Polyols / 202
10.1.2. Selection of Polyacids / 204
10.2. Polyester Resins for High-Solids Coatings / 205
10.3. Carboxylic Acid – Terminated Polyester Resins / 208
10.4. Carbamate-Functional Polyester Resins / 209
10.5. Water-Reducible Polyester Resins / 209
10.6. Polyester Resins for Powder Coatings / 211
References / 211
11.
Amino Resins
11.1. Synthesis of Melamine –Formaldehyde Resins / 214
11.1.1. Methylolation Reaction / 215
11.1.2. Etherification Reaction / 215
11.1.3. Self-Condensation Reactions / 217
11.2. Types of MF Resins / 218
213
CONTENTS
ix
11.3. MF – Polyol Reactions in Coatings / 220
11.3.1. Catalysis of MF – Polyol Reactions / 222
11.3.2. Kinetics and Mechanism of MF – Polyol
Co-condensation / 224
11.3.3. Package Stability Considerations / 226
11.3.4. MF Resin Reactions with Carboxylic Acids, Urethanes,
Carbamates, and Malonate-Blocked Isocyanates / 227
11.4. Other Amino Resins / 227
11.4.1. Urea – Formaldehyde Resins / 228
11.4.2. Benzoguanamine – Formaldehyde Resins / 228
11.4.3. Glycoluril–Formaldehyde Resins / 228
11.4.4. Poly(meth)acrylamide – Formaldehyde Resins / 229
References / 229
12.
Binders Based on Polyisocyanates: Polyurethanes
231
12.1. Reactions of Isocyanates / 232
12.2. Kinetics of Reactions of Isocyanates with Alcohols / 234
12.2.1. Noncatalyzed Reactions / 235
12.2.2. Catalysts / 237
12.2.3. Interrelationships in Catalysis / 241
12.3. Isocyanates Used in Coatings / 242
12.3.1. Aromatic Isocyanates / 242
12.3.2. Aliphatic Isocyanates / 244
12.4. Two Package Solventborne Urethane Coatings / 246
12.4.1. 2K Urea Coatings / 249
12.5. Blocked Isocyanates / 250
12.5.1. Principles of Blocking and Deblocking / 251
12.5.2. Blocking Groups / 254
12.5.3. Catalysis of Blocked Isocyanate Coatings / 258
12.6. Moisture-Curable Urethane Coatings / 259
12.7. Waterborne Urethane Coatings / 261
12.7.1. Polyurethane Dispersions / 261
12.7.2. Acrylic – Polyurethane Dispersions / 262
12.7.3. 2K Waterborne Urethanes / 264
12.8. Hydroxy-Terminated Polyurethanes / 266
References / 267
13.
Epoxy and Phenolic Resins
13.1. Epoxy Resins / 271
13.1.1. Bisphenol A Epoxy Resins / 271
13.1.2. Other Epoxy Resins / 275
271
x
CONTENTS
13.2. Epoxy – Amine Systems / 278
13.2.1. Pot Life and Cure Time Considerations / 279
13.2.2. Toxicity and Stoichiometric Considerations / 281
13.2.3. Graininess and Blushing / 282
13.2.4. Tg Considerations / 283
13.2.5. Other Formulating Considerations / 283
13.2.6. Waterborne Epoxy –Amine Systems / 285
13.3. Other Cross-Linking Agents for Epoxy Resins / 286
13.3.1. Phenols / 286
13.3.2. Carboxylic Acids and Anhydrides / 286
13.3.3. Hydroxyl Groups / 288
13.3.4. Mercaptans / 288
13.3.5. Homopolymerization / 289
13.4. Water-Reducible Epoxy Acrylic Graft Copolymers / 290
13.5. Epoxy Resin Phosphate Esters / 290
13.6. Phenolic Resins / 290
13.6.1. Resole Phenolic Resins / 291
13.6.2. Novolac Phenolic Resins / 292
13.6.3. Ether Derivatives of Phenolic Resins / 293
General References / 293
References / 293
14.
Drying Oils
295
14.1. Composition of Natural Oils / 295
14.2. Autoxidation and Cross-Linking / 297
14.2.1. Nonconjugated Drying Oils / 297
14.2.2. Catalysis of Autoxidation and Cross-Linking / 300
14.2.3. Conjugated Drying Oils / 301
14.3. Synthetic and Modified Drying Oils / 302
14.3.1. Heat-Bodied Oils, Blown Oils, and Dimer Acids / 302
14.3.2. Varnishes / 302
14.3.3. Synthetic Conjugated Oils / 303
14.3.4. Esters of Higher Functionality Polyols / 303
14.3.5. Maleated Oils / 303
14.3.6. Vinyl-Modified Oils / 304
General References / 305
References / 305
15.
Alkyd Resins
15.1. Oxidizing Alkyds / 307
15.1.1. Monobasic Acid Selection / 308
306
CONTENTS
xi
15.1.2. Polyol Selection / 310
15.1.3. Dibasic Acid Selection / 311
15.2. High-Solids Oxidizing Alkyds / 312
15.3. Water-Reducible Alkyds / 313
15.4. Nonoxidizing Alkyds / 314
15.5. Synthetic Procedures for Alkyd Resins / 315
15.5.1. Synthesis from Oils or Fatty Acids / 315
15.5.2. Process Variations / 316
15.6. Modified Alkyds / 320
15.7. Uralkyds and Other Autoxidizable Urethanes / 320
15.7.1. Uralkyds / 320
15.7.2. Autoxidizable Polyurethane Dispersions / 322
15.8. Epoxy Esters / 322
General Reference / 324
References / 324
16.
Silicon Derivatives
326
16.1. Silicones / 326
16.1.1. Silicone Rubbers and Resins / 327
16.1.2. Silicone-Modified Resins / 329
16.2. Reactive Silanes / 331
16.3. Orthosilicates / 333
16.3.1. Sol-Gel Coatings / 333
References / 333
17.
Other Resins and Cross-Linkers
17.1. Halogenated Polymers / 335
17.1.1. Solution Grade Thermoplastic Vinyl
Chloride Copolymers / 335
17.1.2. Vinyl Chloride Dispersion Copolymers / 336
17.1.3. Chlorinated Rubber, Chlorinated Ethylene Vinyl
Acetate Copolymers, and Chlorinated Polyolefins / 336
17.1.4. Fluorinated Polymers / 337
17.2. Cellulose Derivatives / 338
17.2.1. Nitrocellulose / 338
17.2.2. Cellulose Acetobutyrate / 339
17.3. Unsaturated Polyester Resins / 340
17.4. (Meth)acrylated Oligomers / 342
17.5. 2-Hydroxyalkylamide Cross-Linkers / 342
17.6. Acetoacetate Cross-Linking Systems / 343
17.7. Polyaziridine Cross-Linkers / 345
335
xii
CONTENTS
17.8. Polycarbodiimide Cross-Linkers / 346
17.9. Polycarbonates / 347
References / 347
18.
Solvents
349
18.1. Solvent Composition / 349
18.2. Solubility / 351
18.2.1. Solubility Parameters / 351
18.2.2. Three-Dimensional Solubility Parameters / 354
18.2.3. Other Solubility Theories / 356
18.2.4. Practical Considerations / 356
18.3. Solvent Evaporation Rates / 357
18.3.1. Evaporation of Single Solvents / 357
18.3.2. Relative Evaporation Rates / 358
18.3.3. Evaporation of Mixed Solvents / 361
18.3.4. Evaporation of Solvents from Coating Films / 363
18.3.5. Evaporation of Solvents from High-Solids Coatings / 365
18.3.6. Volatile Loss from Waterborne Coatings / 366
18.4. Viscosity Effects / 367
18.5. Flammability / 369
18.6. Other Physical Properties / 370
18.7. Toxic Hazards / 371
18.8. Atmospheric Photochemical Effects / 372
18.9. Regulation of Solvent Emissions from Coatings / 376
18.9.1. Determination of VOC / 376
18.9.2. Regulations / 377
General References / 379
References / 379
19.
Color and Appearance
19.1. Light / 383
19.2. Light –Object Interactions / 385
19.2.1. Surface Reflection / 385
19.2.2. Absorption Effects / 387
19.2.3. Scattering / 388
19.2.4. Multiple Interaction Effects / 390
19.3. Hiding / 392
19.4. Metallic and Interference Colors / 393
19.5. The Observer / 394
19.6. Interactions of Light Source, Object, and Observer / 396
19.7. Color Systems / 397
382
CONTENTS
xiii
19.8. Color Mixing / 401
19.9. Color Matching / 402
19.9.1. Information Requirements / 402
19.9.2. Color Matching Procedures / 404
19.10. Gloss / 407
19.10.1. Variables in Specular Gloss / 408
19.10.2. Gloss Measurement / 412
General References / 415
References / 415
20.
Pigments
417
20.1. White Pigments / 418
20.1.1. Titanium Dioxide / 418
20.1.2. Other White Pigments / 420
20.2. Color Pigments / 421
20.2.1. Yellow and Orange Pigments / 423
20.2.2. Red Pigments / 425
20.2.3. Blue and Green Pigments / 426
20.2.4. Black Pigments / 427
20.2.5. Metallic, Interference, and Cholesteric Pigments / 428
20.3. Inert Pigments / 430
20.4. Functional Pigments / 431
20.5. Nano-pigments / 432
General References / 433
References / 433
21.
Pigment Dispersion
21.1. Dispersion in Organic Media / 435
21.1.1. Wetting / 435
21.1.2. Separation / 436
21.1.3. Stabilization / 436
21.2. Formulation of Non-aqueous Mill Bases / 442
21.2.1. Daniel Flow Point Method / 443
21.2.2. Oil Absorption Values / 444
21.3. Dispersions in Aqueous Media / 445
21.3.1. Stabilization of Aqueous Dispersions / 446
21.4. Dispersion Equipment and Processes / 449
21.4.1. High-Speed Disk Dispersers / 449
21.4.2. Rotor – Stator Mixers / 451
21.4.3. Ball Mills / 451
21.4.4. Media Mills / 453
435
xiv
CONTENTS
21.4.5. Three Roll and Two Roll Mills / 455
21.4.6. Extruders / 456
21.4.7. Ultrasound Dispersion / 456
21.5. Evaluation of Dispersion / 456
General References / 459
References / 459
22.
Effect of Pigments on Coating Properties
461
22.1. PVC and CPVC / 461
22.1.1. Factors Controlling CPVC / 462
22.1.2. Determination of CPVC / 463
22.1.3. CPVC of Latex Coatings / 464
22.2. Relationships Between Film Properties and PVC / 466
22.2.1. Mechanical Properties / 466
22.2.2. Effects of Porosity / 470
22.2.3. Effects on Curing and Film Formation / 471
References / 471
23.
Application Methods
473
23.1. Brushes, Pads, and Hand Rollers / 473
23.1.1. Brush and Pad Application / 473
23.1.2. Hand Roller Application / 474
23.2. Spray Application / 475
23.2.1. Air Spray Guns / 476
23.2.2. Airless Spray Guns / 478
23.2.3. Electrostatic Spraying / 479
23.2.4. Hot Spray / 481
23.2.5. Supercritical Fluid Spray / 481
23.2.6. Formulation Considerations for Spray-Applied Coatings / 482
23.2.7. Overspray Disposal / 484
23.3. Dip and Flow Coating / 484
23.4. Roll Coating / 485
23.5. Curtain Coating / 487
General References / 488
References / 488
24.
Film Defects
24.1.
24.2.
24.3.
24.4.
Surface Tension / 490
Leveling / 492
Sagging and Drip Marks / 496
Crawling, Cratering, and Related Defects / 499
490
CONTENTS
xv
24.5. Floating and Flooding; Hammer Finishes / 502
24.6. Wrinkling and Wrinkle Finishes / 505
24.7. Bubbling and Popping / 506
24.8. Foaming / 508
24.9. Dirt / 509
General References / 509
References / 509
25.
Solventborne and High-Solids Coatings
511
25.1. Primers / 512
25.1.1. Binders for Primers / 512
25.1.2. Pigmentation of Primers / 513
25.1.3. High-Solids Primers / 514
25.2. Top Coats / 515
25.2.1. Binders for Top Coats / 515
25.2.2. Formulating Solventborne Coatings for Low VOC / 518
General Reference / 522
References / 522
26.
Waterborne Coatings
524
26.1. Water-Reducible Coatings / 525
26.2. Latex-Based Coatings / 529
26.3. Emulsion Coatings / 532
General References / 533
References / 533
27.
Electrodeposition Coatings
535
27.1. Anionic Electrodeposition Coatings / 536
27.2. Cationic Electrodeposition Coatings / 538
27.3. Effect of Variables on Electrodeposition / 541
27.4. Application of Electrodeposition Coatings / 542
27.5. Advantages and Limitations of Electrodeposition / 544
27.6. Autodeposition Coatings / 545
General Reference / 547
References / 547
28.
Powder Coatings
28.1. Binders for Thermosetting Powder Coatings / 548
28.1.1. Epoxy Binders / 550
28.1.2. Hybrid Binders / 550
28.1.3. Polyester Binders / 551
548
xvi
CONTENTS
28.1.4. Acrylic Binders / 553
28.1.5. Silicone Binders / 554
28.1.6. UV Cure Powder Coatings / 554
28.2. Binders for Thermoplastic Powder Coatings / 555
28.3. Formulation of Thermosetting Powder Coatings / 555
28.3.1. Low Gloss Powder Coatings / 559
28.4. Manufacture of Powder Coatings / 560
28.4.1. Production / 560
28.4.2. Quality Control / 563
28.5. Application Methods / 564
28.5.1. Electrostatic Spray Application / 564
28.5.2. Other Application Methods / 567
28.6. Advantages and Limitations of Powder Coatings / 569
General References / 571
References / 571
29.
Radiation Cure Coatings
574
29.1. UV Curing / 574
29.2. Free Radical Initiated UV Cure / 578
29.2.1. Unimolecular Photoinitiators / 578
29.2.2. Bimolecular Photoinitiators / 579
29.2.3. Oxygen Inhibition / 580
29.2.4. Vehicles for Free Radical Initiated UV Cure / 581
29.2.5. Waterborne UV Cure Coatings / 584
29.3. Cationic UV Cure / 585
29.3.1. Vehicles for Cationic UV Cure / 585
29.4. Effects of Pigmentation / 586
29.5. Electron Beam Cure Coatings / 588
29.6. Dual UV/Thermal Cure and Selected Applications / 589
29.7. Advantages and Limitations / 591
General References / 593
References / 593
30.
Product Coatings for Metal Substrates
30.1. OEM Automotive Coatings / 595
30.1.1. Electrodeposition Coatings and Primer – Surfacers / 596
30.1.2. Top Coats / 597
30.1.3. Factory Repair Procedures / 604
30.2. Appliance Coatings / 605
30.3. Container Coatings / 606
30.3.1. Interior Can Linings / 608
595
CONTENTS
xvii
30.3.2. Exterior Can Coatings / 610
30.4. Coil Coating / 611
30.4.1. Advantages and Limitations of Coil Coating / 615
General References / 616
References / 617
31.
Product Coatings for Nonmetallic Substrates
619
31.1. Coatings for Wood / 619
31.1.1. Coatings for Wood Furniture / 619
31.1.2. Panel, Siding, and Flooring Finishes / 626
31.2. Coating of Plastics / 628
31.2.1. In-mold Coating / 628
31.2.2. Post-mold Coating / 629
31.3. Coatings for Glass / 633
General Reference / 634
References / 634
32.
Architectural Coatings
636
32.1. Exterior House Paint / 637
32.2. Flat and Semigloss Paints / 643
32.3. Gloss Enamels / 649
32.3.1. Alkyd Gloss Enamels / 649
32.3.2. Latex Gloss Enamels / 651
General Reference / 655
References / 655
33.
Special Purpose Coatings
658
33.1. Maintenance Paints / 658
33.1.1. Barrier Coating Systems / 659
33.1.2. Systems with Zinc-Rich Primers / 663
33.1.3. Systems with Passivating Pigment Containing Primers / 665
33.2. Marine Coatings / 667
33.3. Automobile Refinish Paints / 670
33.4. Aircraft Coatings / 674
References / 676
34.
Perspectives on Coating Design
679
General References / 687
References / 688
Index
689
Preface
Significant advances have been made since the publication of the second edition of
Organic Coatings: Science and Technology. The third edition has been completely
updated. Our purpose remains the same, which is to provide a reference and textbook
that interrelates coatings technology with current scientific understanding.
Entire books could be written about the subject of each chapter, and many have been.
To be as comprehensive as possible in the limited space available, we have had to limit
coverage of each topic and have selected references for readers seeking more detailed
information. We have striven to enhance the usefulness of this edition both as a classroom
textbook on coatings science, as well as a reference book, by improving the presentation of
each topic. The reader will benefit from having taken college level chemistry courses
through organic chemistry, but no coursework in polymer science is assumed.
Many of the chapters include brief descriptions of coatings compositions and applications, supported by references, which could be omitted or used for outside assignments,
such as term papers, particularly in an advanced course. These descriptions tend to be
placed in the later sections of each chapter and tend to be more prevalent in the application
chapters. These compositions and applications particularly enhance the value of the
volume as a reference book and self-teaching text. We understand that the second
edition was used widely for this purpose. We have also defined the jargon of coatings
to help newcomers to the field understand its specialized language. Although the book
is written specifically about coatings, many of the principles involved apply to the
related fields of printing inks, adhesives, and parts of the plastics industry.
Coatings technology evolved empirically, by trial and error. The last few decades have
seen a marked increase in scientific understanding of the applicable principles, but the
complexities of the field are such that the formulator’s art is still essential in developing
and using coatings. The need to reduce air pollution while maintaining and, preferably,
improving coating performance requires radically new formulations on a short time
scale. Our conviction is that increased understanding of the underlying science can help
xix
xx
PREFACE
formulators work more effectively and that an appreciation of a formulator’s craft is essential for scientists working in the field.
We do not claim to provide a complete literature review on each topic, but believe that
many of the key references are cited. Readers are cautioned that the quality of the literature
in the coatings field is uneven. Many published papers and monographs are excellent, but
some are not; unfortunately, some authors did not fully understand the complexity of the
field.
Many of the chapters in this edition were reviewed by people with extensive working
experience with the particular topic. These reviewers were: David Bittner, Randall Brady,
Adelbert Braig, John Bright, David Cocuzzi, Nico Enthoven, Ray Fernando, Werner
Funke, Loren Hill, George Pilcher, Bradley Richards, Christian Schaller, John
G. Stauffer, and Peter Wolfgang. Special thanks to Clifford Schoff, who reviewed three
chapters and a section of another. We acknowledge the contributions of students and
staff from the Wicks and Thames/Rawlins Research Groups at the University of Southern
Mississippi. Special appreciation is expressed for the assistance of Helen Rassier, Stacy
Trey, and Todd Williams of the Wicks Research Group.
Symbols and Units
A
C
8C
c
E
E0
E00
Ea
F
F̄
f
f̄n
G
Gc
g
g
H
h
i
K
K
Arrhenius preexponential term
concentration, weight per unit volume of solution
degrees Celsius
concentration, moles per liter
modulus; relative evaporation rate
storage modulus (elastic modulus)
loss modulus
thermal coefficient of reaction rate (Arrhenius activation energy)
functionality of a monomer
average functionality of a monomer mixture
functionality of a polymer (resin)
number average functionality of a polymer (resin)
free energy; Small’s molar association constant; force applied in
a tensile test
force to crack in a tensile test
gram
gravitational constant
enthalpy
film thickness
angle of incidence
Kelvin temperature
absorption coefficient
Note that all acronyms are listed in the index.
xxi
xxii
SYMBOLS AND UNITS
KE
k
kg
L
M
c
M
M̄n
M̄w
mL
mP s
N
N
P
P n
P̄w
p
pg
Einstein (shape) constant
rate constant
kilogram
liter
molecular weight
average molecular weight between cross-links
number average molecular weight
weight average molecular weight
milliliter
millipascal second ¼ 1 centipoise
newton
number of moles; refractive index
vapor pressure, degree of polymerization
number average degree of polymerization
weight average degree of polymerization
extent of reaction
extent of reaction at gelation onset
Pa
Pa s
PDI
R
r
S
s
T
Tb
Tg
Tm
tan d
V
Vi
w
X
x
x
g
ġ
d
1
h
he
hr
hs
[h]
[h]w
[h]u
pascal
pascal second ¼ 10 poise
polydispersity index ¼ M̄w/M̄n
gas constant
angle of reflection or angle of refraction
entropy; scattering coefficient
second
temperature; time
brittle– ductile transition temperature
glass transition temperature
melting point
tan delta, loss tangent, E00 /E0
molar volume
volume fraction of internal phase
weight fraction
film thickness
mole fraction
optical path length
surface tension
shear rate
solubility parameter; phase shift in viscoelastic deformation
molar absorbance; strain
absolute shear viscosity
external phase viscosity; extensional viscosity
relative viscosity ¼ h/hs
viscosity of solvent
intrinsic viscosity
weight intrinsic viscosity
intrinsic viscosity under theta conditions
†
†
SYMBOLS AND UNITS
u
l
v
ne
r
s
s0
f
x
contact angle
wavelength
kinematic viscosity; Poisson’s constant
mole of elastically effective network chains per
cubic centimeter
density
shear stress
yield value
packing factor
activity coefficient
xxiii
1
What Are Coatings?
Look around you; coatings are everywhere. If you are indoors, there are coatings on the
walls, refrigerator, cabinets, and furniture; less obviously, coatings are on the wires of
electrical motors, printed circuits, inside television sets, and compact disks. If you are
outdoors, coatings are on your house and car, as well as inside your car, under the
hood, and on components of the automotive stereo and computer systems. Whether you
drink beer or soft drinks, there are coatings on the inside and outside of the cans. The functional and decorative requirements of coatings span a broad spectrum. A diverse science
and technology support the development, production, and use of coatings.
People entering coatings science should realize that although it is an old field, it is
not a mature one; it offers exciting challenges and career opportunities. They will have
opportunities to improve scientific understanding and to contribute to the major thrusts
of coatings development: reducing emissions that cause air pollution, reducing energy
requirements, and protecting metals against corrosion.
1.1. DEFINITIONS AND SCOPE
Coatings may be described by their appearance (e.g., clear, pigmented, metallic, or glossy)
and by their function (e.g., corrosion protective, abrasion protective, skid resistant, decorative, or photosensitive). Coatings may be distinguished as organic or inorganic, although
there is overlap. For example, many coatings consist of inorganic pigment particles
dispersed in an organic matrix (the binder).
A confusing situation results from multiple meanings of the term coating. It is used to
describe the material (usually, a liquid) that is applied to a substrate, the resulting “dry”
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
1
2
WHAT ARE COATINGS?
film, and the process of application. Usually, the intended meaning of the word coating
can be inferred from the context.
We limit our discussion to coatings with organic binders that are applied purposefully
to a substrate. Many types of coatings are not included. Porcelain enamels on kitchen
ranges are coatings, but they do not have organic chemical binders. Electroplated
copper, nickel, and zinc coatings are excluded for the same reason. We further restrict
our discussion of organic coatings to those materials that can be traced historically back
to paints. What is the difference between a coating and a paint? Not much—the terms
are often used interchangeably. However, it is fairly common practice to use coatings
as the broader term and to restrict paints to the familiar architectural and household
coatings and sometimes to maintenance coatings for bridges and tanks. We follow this
practice. Some prefer to call sophisticated materials that are used to coat automobiles
and computer components coatings, perhaps sensing that paint sounds too low brow.
Still another common term that is essentially a synonym for coating and paint is finish.
In limiting the scope of this book to organic coatings that can be related to historic
paints, we exclude many materials that could be called coatings. Printing inks, polymers
applied during production of paper and fabrics, coatings on photographic films, decals
and other laminates, and cosmetics are but a few examples. However, many of the
basic principles that are covered in this book are applicable to such materials. Restrictions
of scope are necessary if the book is to be kept to a reasonable length, but our restrictions
are not entirely arbitrary. The way in which we are defining coatings is based on common
usage of the term in worldwide business. It is close to the definition of organic coatings for
statistical analyses of industrial output used by Bureau of Census of the U.S. Department
of Commerce. The Census Bureau defines four broad categories: (1) architectural coatings, (2) product coatings used by original equipment manufacturers (OEM coatings),
(3) special purpose coatings, and (4) miscellaneous. 2002 U.S. Census data reported
that there were 1139 companies that manufactured paints and coatings [1].
The worldwide coatings market in 2003 was estimated at about $70 billion [2]. The
markets in North America, Europe, and Asia are roughly equal in size, with Asia
growing fastest. The value of coating shipments in the United States in 2002 according
to the Current Industrial Reports—Paint and Coating Manufacturing are shown in
Table 1.1.
Architectural coatings include paints and varnishes (transparent paints) used to
decorate and protect buildings, outside and inside. They also include other paints and
varnishes sold for use in the home and by small businesses for application to such things
as cabinets and household furniture (not those sold to furniture factories). They are often
called trade sales paints. They are sold directly to painting contractors and do-it-yourself
users through paint stores and other retail outlets. In 2003 in the United States, architectural
TABLE 1.1. U.S. Coatings Shipments, 2002
Coatings
Architectural
Product (OEM)
Special purpose
Miscellaneous
Dollars 109
7.211
5.556
3.153
1.181
17.101
Source: Ref. [1].
1.1. DEFINITIONS AND SCOPE
3
coatings accounted for about 58% of the total volume of coatings; however, the unit value
of these coatings was lower than for the other categories, so they made up about 47% of the
total value. This market is the least cyclical of the three categories. Although the annual
amount of new construction drops during recessions, the resulting decrease in
paint requirements tends to be offset by increased repainting of older housing, furniture,
and so on, during at least mild recessions. Latex-based coatings make up about 77% of
architectural coatings [2].
Product coatings, also commonly called industrial coatings or industrial finishes, are
applied in factories on products such as automobiles, appliances, magnet wire, aircraft,
furniture, metal cans, chewing gum wrappers—the list is almost endless. This market is
often called the OEM market, that is, the original equipment manufacturer market. In
2003 in the United States, product coatings were about 29% of the volume and 33% of
the value of all coatings. The U.S. shipments in 2003 were $8.6 billion of the world
market, estimated at $24 billion [2]. The volume of product coatings depends directly
on the level of manufacturing activity. This category of the business is cyclical, varying
with OEM cycles. In most cases, product coatings are custom designed for a particular
customer’s manufacturing conditions and performance requirements. The number of
different types of products in this category is much larger than in the others; research
and development requirements are also higher.
Special purpose coatings include industrial coatings applied outside a factory, along
with a few miscellaneous coatings, such as coatings packed in aerosol containers. It
includes refinish coatings for cars and trucks that are applied outside the OEM factory
(usually in body repair shops), marine coatings for ships (they are too big to fit into a
factory), and striping on highways and parking lots. It also includes maintenance paints
for steel bridges, storage tanks, chemical factories, and so on. In 2003 in the United
States, special purpose coatings constituted about 13% of the total volume and 20% of
the total value of all coatings.
The Census Bureau defines miscellaneous paint and coating products as paint
removers, thinners, pigment dispersions, glazing compounds, and so on.
Coatings are used for one or more of three reasons: (1) for decoration, (2) for protection, and/or (3) for some functional purpose. The low gloss paint on the ceiling of a room
fills a decorative need, but it also has a function: It reflects and diffuses light to help
provide even illumination. The coating on the outside of an automobile adds beauty to
a car and also helps protect it from rusting. The coating on the inside of a beverage can
has little or no decorative value, but it protects the beverage from the can. (Contact
with metal affects flavor.) In some cases, the interior coating protects the can from the
beverage. (Some soft drinks are so acidic that they can dissolve the metal.) Other coatings
reduce the growth of algae and barnacles on ship bottoms, protect optical fibers for
telecommunications against abrasion, retard corrosion of bridges, serve as the recording
medium on compact disks, and so on. Although the public most commonly thinks of
house paint when talking about coatings, all types of coatings are important throughout
the economy, and they make essential contributions to most high-tech fields. For
example, computer technology depends on microlithographic coatings to construct
microprocessors.
Traditionally, coatings have changed relatively slowly in an evolutionary response
to new performance requirements, new raw materials, and competitive pressures. An
important reason for the relatively slow rate of change is the difficulty of predicting
product performance on the basis of laboratory tests. It is less risky to make relatively
4
WHAT ARE COATINGS?
small changes in composition and check actual field performance before making further
changes. Starting in the 1930s, the pace of technical change increased as new synthetic
polymers were introduced. Since 1965 a major driving force for change has been the
need to reduce VOC (volatile organic compound ) emissions because of their detrimental
effect on air quality. Coatings have been second only to the gasoline– automobile complex
as a source of VOC pollutants responsible for excess ozone in the air of many cities
on many days of the year. This situation has resulted in increasingly stringent regulatory
controls on such emissions. The drive to reduce VOC emissions has also been fueled by
the rising cost of organic solvents. Other important factors have also accelerated the rate of
change in coatings. Increasing concern about toxic hazards has led to the need to change
many raw materials that were traditionally used in coatings. Furthermore, manufacturers
often want their coatings modified so that they can be used at faster production rates,
baked at lower temperatures, or changed in color. Product performance requirements
have tended to increase; most notable is the need for increased effectiveness of corrosion
protection by coatings.
1.2. COMPOSITION OF COATINGS
Organic coatings are complex mixtures of chemical substances that can be grouped into
four broad categories: (1) binders, (2) volatile components, (3) pigments, and (4) additives.
Binders are the materials that form the continuous film that adheres to the substrate (the
surface being coated), that bind together the other substances in the coating to form a film,
and that present an adequately hard outer surface. The binders of coatings within the
scope of this book are organic polymers. In some cases, these polymers are prepared
and incorporated into the coating before application; in other cases, final polymerization
takes place after the coating has been applied. Binder polymers and their precursors are
often called resins. The binder governs, to a large extent, the properties of the coating
film. The major resin types used in coatings are given in Table 1.2.
Volatile components are included in a large majority of all coatings. They play a major
role in the process of applying coatings; they are liquids that make the coating fluid
enough for application, and they evaporate during and after application. Until about
TABLE 1.2. U.S. Resin Sales by Type, 2004
Resin Type
Acrylics
Alkyds
Amino resins
Cellulosics
Epoxies
Polyesters
Urethanes
Poly(vinyl aetates)
Poly(vinyl chlorides)
Miscellaneous
Source: Ref. [3].
Pounds 106
Dollars 106
Growth Rate (%)
925
445
335
31
380
275
249
640
240
200
1350
356
94
59
475
261
600
992
156
400
3
0 to 21
1 to 2
21
2
4
3
1 to 2
0
av. 2
3495
4743
av. 2
1.2. COMPOSITION OF COATINGS
5
1945, almost all of the volatile components were low molecular weight organic solvents
that dissolved the binder components. However, the term solvent has become potentially
misleading because since 1945, many coatings have been developed for which the binder
components are not fully soluble in the volatile components. Because of the need to reduce
VOC emissions, a major continuing drive in the coatings field is to reduce the use of solvents by making the coatings more highly concentrated (higher-solids coatings), by using
water as a major part of the volatile components (waterborne coatings), and by eliminating
solvents altogether. Vehicle is a commonly encountered term. It usually means a combination of the binder and the volatile components of a coating. Today, most coatings,
including waterborne coatings, contain at least some volatile organic solvents. Exceptions
are powder coatings, certain solventless liquid coatings, radiation-curable coatings, and a
small but growing segment of architectural coatings.
Pigments are finely divided insoluble solid particles that are dispersed in the vehicle
and remain suspended in the binder after film formation. Generally, the primary
purpose of pigments is to provide color and opacity to the coating film. However, they
also have substantial effects on application characteristics and on film properties.
Although most coatings contain pigments, there are important types of coatings that
contain little or no pigment, commonly called clear coats, or just clears. Clear coats for
automobiles and transparent varnishes are examples.
Additives are materials that are included in small quantities to modify some property of
a coating. Examples are catalysts for polymerization reactions, stabilizers, and flow
modifiers.
Most coatings are complex mixtures. Many contain several substances from each of the
four categories, and each substance is usually a chemical mixture. The number of possible
combinations is limitless. The number of different applications is also limitless.
Formulation of paints started millennia ago as an empirical art or craft. Successive
generations of formulators built on the experience of their predecessors and formulated
coatings with increasingly better performance characteristics. Gradually, formulators
began trying to understand the underlying scientific principles that control the performance of coatings. Most coating systems are so complex that our understanding of them
today is still limited. Real progress has been made, but the formulator’s art is still a critical
element in developing high performance coatings. Because demands on suppliers of
coatings to develop new and better coatings are increasing at an accelerating pace, time
is now too limited to permit traditional trial and error formulation. Understanding the
basic scientific principles can help a formulator design better coatings more quickly. In
the chapters ahead, we present, to as great an extent possible, the current understanding
of the scientific principles involved in coating science. Important considerations when
working on a formulating problem, including experimental design and combinatorial or
high throughput methods, are provided in Chapter 34.
We also identify areas in which our basic understanding remains inadequate and
discuss approaches to more efficient and effective formulation despite inadequate understanding. In some cases, in which no hypotheses have been published to explain certain
phenomena, we offer speculations. Such speculations are based on our understanding of
related phenomena and on our cumulative experience acquired over decades in the
field, but they are dangerous because they may be wrong. We recognize the risk that,
with time, speculation tends to become accepted as fact and may even be cited as evidence
or adopted as an experimentally supported hypothesis. It is our intent, rather, that such
speculations promote the advancement of coatings science and technology by stimulating
6
WHAT ARE COATINGS?
discussion that leads to experimentation designed to disprove or support the speculative
proposal. We believe that the latter purpose outweighs the former risk, and we endeavor
to identify the speculative proposals as such.
Cost is an essential consideration in formulation. Novice formulators are inclined to
think that the best coating is the one that will last the longest time without any change
in properties, but such a coating may be very expensive and unable to compete with a
less expensive coating whose performance is adequate for the particular application.
Furthermore, it is seldom possible to maximize all of the performance characteristics of
a coating in one formulation. Some of the desirable properties are antagonistic with
others; formulators must balance many performance variables while keeping costs as
low as possible.
The formulator’s product is a formula. A coatings company’s formulas are among its
most important assets. They are followed by the manufacturing department to produce
the ultimate products. A formula includes a list of ingredients and amounts by weight
and by volume, with the amounts often totaling 100 or 1000 gallons to facilitate scaling
to different-sized equipment. It also includes specific manufacturing directions, warnings
of safety hazards, information about individual ingredients, quality control tests, specifications, and cautions against potential manufacturing pitfalls, (e.g., “Do not heat above
508C”). In addition, there should be a unique code number, a date, the name of the
formulator, approval by a supervisor, a statement of why the formula was developed (if
it is new) or changed (if it is a modification of an older formula). Typically, responsibility
for a new formula is turned over to the manufacturing department after one to three
successful production batches. The manufacturing department should not change formulas
but should request changes from the technical department if changes are desired. It is
important for companies to manage their formulas well, protecting their confidentiality
while keeping them in good order. It should be anticipated that the formulas will play a
central role in expansions, licensing, acquisitions, and certain legal actions, such as
product liability suits.
REFERENCES
1. U.S. Department of Commerce, Bureau of Census, Current Industrial Reports—Paint and
Coating Manufacturing, http://www.census.gov.
2. Tullo, A. H., Chem. Eng. News, 2004, 82(42), 25.
3. Challener, C., JCT Coat. Tech, 2005, 2(13), 54.
2
Polymerization and
Film Formation
This chapter is designed to introduce the basic concepts of polymer chemistry and film
formation.
2.1. POLYMERS
A polymer is a substance composed of macromolecules. Some authors reserve the term
polymer to describe a substance and use the term macromolecule for the molecules
making up the substance. This use distinguishes between the material and the molecules,
but is not common in the coatings field. We use the term polymer for both meanings;
depending on the context, the term refers to either the molecules or the substance. The
structure of polymers is a multiple repetition of units (mers) derived from molecules of
low relative molecular weight (MW) (monomers). (The more rigorous designation
of MW is molar mass, but we use MW because it is much more commonly used in the
coatings field.)
There is disagreement about how high the MW has to be for a material to qualify as
a polymer. Some people refer to materials with MWs as low as 1000 as polymers;
others insist that only materials with MWs over 10,000 (or even 50,000) qualify. The
term oligomer, meaning “a few mers,” is often used for materials having MWs of a few
hundred to a few thousand. This additional term does not help the definition problem
much because there is no clear-cut boundary between an oligomer and a polymer, but
the term can be useful because it provides a name with which most can agree for materials
containing 2 to about 20 mers.
Polymers occur widely in nature; familiar examples of biopolymers, which are produced by living organisms, are proteins, starch, cellulose, and silk. In the coatings field,
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
7
8
POLYMERIZATION AND FILM FORMATION
we are concerned mainly with synthetic polymers, although some chemically modified
biopolymers are also used.
Synthetic polymers and oligomers are prepared by polymerization, a sequence of
chemical reactions in which small molecules are joined by covalent bonds. A polymer
made from a single monomer is called a homopolymer. If it is made from a combination
of monomers, it is often but not always called a copolymer. An example of a homopolymer
is provided by the polymerization of vinyl chloride:
In this example, the 22(CH2CHCl)22 repeating unit is the mer, and n represents the number
of mers joined together in the molecule. X and Y represent terminal groups on the ends of
the chain of mers.
Polymers can be made in three important classes. When the mers are linked in chains,
the polymers are called linear polymers, a term that is potentially misleading because the
large molecules seldom form a straight line—they twist and coil. In linear copolymers, the
various monomers may be distributed more or less at random throughout the chain
(random copolymers), they may tend to alternate (alternating copolymers), or they may
be grouped together (block copolymers). If there are forks in the chains, the polymers
are called branched polymers. A polymer chain of one type of monomer having
polymer branches of another type of monomer is called a graft copolymer. The third
class, of particular importance in coatings, results from the bonding of chains with
each other at several sites to form cross-linked, or network, polymers. These polymers
are branched polymers where the branches are covalently bound to other molecules,
so the mass of polymer consists mainly of a single interconnected molecule. A
fourth class, dendritic polymers, may become important in coatings applications in
the future.
Reactions that join polymer or oligomer molecules are called cross-linking reactions.
Polymers and oligomers that can undergo such reactions are frequently called thermosetting polymers. Some confusion can result because the term thermosetting is applied not
only to polymers that cross-link when heated, but also to those that can cross-link at
ambient temperature. A polymer that does not undergo cross-linking reactions is called
a thermoplastic polymer, because it becomes plastic (softens) when heated.
Polymers can also be formed from mers that contain multiple H-bonding sites, in which
case the mers are joined together by hydrogen bonds rather than by covalent bonds. Such
polymers are called supramolecular polymers, which are of particular interest when the
hydrogen-bonding sites are four-centered, owing to greatly enhanced strength relative
to three-centered hydrogen-bonds [1]. Coatings compositions in which multicenter hydrogen bonds complement covalent cross-links may be of potential interest for enhancing
coatings properties, owing to thermal reversibility of the hydrogen bonds, as observed
for polyurethanes (Chapter 12). Multicenter hydrogen-bond compositions have been
disclosed for various applications related to coatings, including adhesives and molding
compounds [2].
Another term commonly, but loosely, used in the coatings field is resin. This term
overlaps the meanings of polymer and oligomer. Originally, the term referred to hard,
brittle materials derived from tree exudates, such as rosin, dammar, and elemi.
2.1. POLYMERS
9
A variety of these naturally occurring resins were used since prehistoric times to make
coatings. In the nineteenth and early twentieth centuries, such resins were dissolved in
drying oils to make varnishes (Section 14.3.2). The first entirely synthetic polymers
used in coatings were phenol – formaldehyde polymers (Section 13.6), which replaced
naturally occurring resins in many applications. Hence, it was natural to call them
phenol – formaldehyde resins or phenolic resins. As more synthetic products were
developed to replace naturally occurring resins, these products were also called resins.
When words do not have precise meanings, it is important to understand the context in
which they are placed. Commonly, it is assumed, without much thought, that information
that has been learned about high MW polymers is also applicable to low MW polymers
or oligomers, because all are often called polymers. Many characteristics, however,
depend on MW. Although much of the information available from studies of high-MW
polymers can be useful in the coatings field, it must be used with caution, because the
resins used in making coatings are commonly low MW polymers or oligomers, even
though they are frequently called polymers. In the next set of subsections, we describe
some of the key characteristics of synthetic polymers and oligomers.
2.1.1. Molecular Weight
For most pure organic compounds, the concept of molecular weight is straightforward—
each compound has a MW. For synthetic polymers, however, the situation is more
complex. All methods of synthesis lead to mixtures of molecules with different
numbers of mers and, therefore, with different MWs. Even relatively simple thermoplastic
homopolymers, such as polystyrene or poly(vinyl chloride), contain molecules with hundreds of different, although similar, structures. With copolymers, the number of different
molecules present is much larger. There is a distribution of MWs in a synthetic polymer;
accordingly, MWs can be defined only by a statistical calculation. In the simplest cases,
the distribution of the number of molecules of each MW resembles a skewed Gaussian
distribution, but in other cases, the distribution may be quite complex. Although many
types of average MW can be calculated, the two most widely used are number and
weight average MWs.
1. Number average MW. M̄n is the MW average based on summing the products of
the numbers of molecules and their MWs and dividing by the sum of the number of
molecules in the sample. Mathematically, it is expressed by the following equation,
where M1, M2, and Mi are the MWs of the first, second, and ith species, respectively,
and the N values are the numbers of molecules of each species present:
n¼
M
P
P
N 1 M1 þ N 2 M2 þ
i Ni Mi
P
¼ P
N1 þ N 2 þ
i Ni
P
Ni Pi
P n ¼ Pi
i Ni
A similar equation is often used to represent the number average degree of polymerization P̄n, where P is the number of mers in a molecule and Pi is the number of mers in the
ith polymer. For homopolymers, M̄n ¼ P̄n times the MW of each mer; for copolymers, a
weighted average MW of the mers is used. The differing weights of end groups can be
10
POLYMERIZATION AND FILM FORMATION
neglected in calculating M̄n/M̄w of high polymers, but not of oligomers, for which the
effect can be appreciable.
2. Weight average MW. M̄w is defined by the following equation, in which w1, w2, and
wi are the weights of molecules of species 1, 2, and ith: Since w1 ¼ N1M1, M̄w can also be
calculated from the numbers of molecules of the various species, as shown in the equation.
Weight average degree of polymerization P̄w is defined by analogous equations:
P
P
w 1 M1 þ w 2 M2 þ
Ni Mi2
i wi Mi
P
¼
Mw ¼
¼ Pi
w1 þ w2 þ
i wi
i Ni Mi
Figure 2.1 shows an idealized plot of the weight fraction of molecules of each MW as a
function of degree of polymerization for oligomers made from the same monomer by three
different processes [3]. In relatively simple distributions of MWs, such as those shown in
Figure 2.2, the value of P̄n is at, or near, the peak of the weight fraction ditribution curve.
Because M̄w and P̄w give extra weight to higher MW molecules, they are always larger
than M̄n and P̄n.
The breadth of the MW distribution can have an important effect on the properties of a
polymer and is often critical to achieving satisfactory performance of a coating. The ratio
M̄w/M̄n is widely used as an index of the breadth of distribution. In the case of high MW
polymers, M̄w/Mn ¼ P̄w/P̄n, but in the case of oligomers, differences in end groups can be
significant and affect the equality of the ratios. These ratios are called polydispersity (PD),
or sometimes, polydispersity index (PDI). We use the symbols M̄w/M̄n and P̄w/P̄n.
The ratios provide a convenient way to compare the MW distributions of different
polymers. However, one must be cautious in the use of a single value to describe a possibly
Figure 2.1. Degree of polymerization distribution plots calculated for three types of chain-growth
polymers. P̄w/Pn ¼ 1.07 is for an ideal anionic polymerization, P̄w/Pn is 1.5 for an ideal free radical polymerization with termination by combination, and P̄w/Pn is 3.0 for a typical free radical polymerization.
P̄n is 12 for all plots, and P̄w is 12.84, 18, and 36, respectively. (From Ref. [4], with permission.)
2.1. POLYMERS
11
Figure 2.2. (a) Molecular weight distribution of a typical polyester resin (from Ref. [5], with permission);
(b) molecular weight distributions of three alkyd resins, as measured by GPC with a UV detector
(from Ref. [6], with permission).
complex distribution. As shown in Figures 2.1 and 2.2, synthetic polymers commonly
have broad distributions of MWs. As M̄w/M̄n increases, the fractions of polymer at the
extremes above and below the number average MW increase. Even the oligomer with a
number average of 12 mers and with M̄w/M̄n ¼ 1.07 has substantial numbers of molecules
12
POLYMERIZATION AND FILM FORMATION
containing 7 to 18 mers, and a polymer with a more typical M̄w/M̄n ¼ 3 has molecules
spanning several orders of magnitude of MW.
M̄n is the MW of most importance for relating stoichiometric ratios of reactants and
for comparing certain physical properties. M̄w often proves more useful than M̄n when
considering the relationship between MW and many physical properties of polymers,
including some of the properties that are crucial to coating performance.
MW measurement is difficult and beyond the scope of this book to discuss in detail.
(See Ref. [4] for a discussion of the various ways of determining M̄w and M̄n.) In practice,
most scientists in the coatings field use gel permeation chromatography (GPC), more
properly called size exclusion chromatography (SEC), to measure MWs. In this convenient method, a dilute solution of an oligomer or a polymer is pumped at high pressure
through a series of columns containing porous gels. The molecules are “sorted” by sizes,
with the largest ones coming out first and the smaller ones, which are slowed by entering
and leaving more of gel pores, coming out later. The concentration of polymer in the
solvent is analyzed as it leaves the column and is plotted as a function of time. A computer
program compares the plot to plots of standard polymers of known MWs and calculates
M̄n, M̄w, and several other quantities that characterize the polymer. The results appear
precise, but they are not accurate; errors of +10% can be expected, and much larger
errors are possible. Errors can result because the MW is not measured directly—
instead, the size of the polymer molecules in solution is measured—and from differences
in detector response to different compositions. Despite its inaccuracy, GPC is useful,
especially for comparing polymers of similar structure.
The M̄n of oligomers can be measured accurately using colligative methods such as
freezing point depression and vapor pressure osmometry. However, the accuracy decreases as MW increases, and colligative methods are of little use above M̄n ¼ 50,000. Mass
spectroscopic methods are available that can accurately measure the MWs of individual
molecules in oligomers and even in fairly high polymers. (In Section 10.2 we give
examples.)
Some polymers and oligomers have MW distributions approaching the idealized distributions shown in Figure 2.1, as illustrated by the GPC trace of a polyester oligomer in
Figure 2.2(a). However, many polymers used in coatings have complex distribution
patterns, as exemplified by the alkyds in Figure 2.2(b). The M̄w and M̄n can be calculated
for an entire trace or for portions of complex traces. But such polydispersity numbers must
be used with caution for complex traces.
The MW of resins is an important factor affecting the viscosity of coatings made with
solutions of the resins: Generally, the higher the MW, the higher the viscosity. The MW
of oligomers used in higher-solids coatings is especially critical. It is often desirable to
make oligomers with as narrow a range of MW as possible, since this minimizes the
amount of very low and very high MW resin. The low MW fraction is generally undesirable from the standpoint of film properties, whereas the high MW fraction increases
resin solution viscosity disproportionately. However, alkyd resins having broad,
complex MW distributions may perform better than alkyds with similar compositions
having narrow distributions [6].
MW is often a critical factor controlling the strength of films that are not cross-linked.
In general, the higher the MW, the higher the tensile strength of such films, at least up to a
point. The acrylic copolymer in automotive acrylic lacquers must have an M̄w greater
than about 75,000 for acceptable film properties but less than 100,000 for acceptable
application properties. For other lacquers the MWs required depend on polymer
2.1. POLYMERS
13
composition and application methods. Film property considerations argue in favor of using
high MW polymers in formulating coatings, but viscosity considerations favor low MWs.
As is often the case in coatings, compromises are needed.
An important advantage of many waterborne coatings is that MW of the polymer
generally does not affect viscosity directly, since the polymers are dispersions of
polymer particles rather than solutions.
2.1.2. Morphology
Morphology is the study of the physical forms of materials. Like MW, morphology is more
complex with polymers than with small-molecule substances. Pure small molecules
generally solidify to crystals if the temperature is sufficiently low. In contrast, few synthetic polymers crystallize completely, and many do not crystallize at all. Non-crystalline
materials that appear to be solids are called amorphous solids. There are at least two
reasons that synthetic polymers are at least partly amorphous. In general, synthetic
polymers are not pure compounds, so it is difficult to achieve the completely regular
structure characteristic of a crystalline material. In addition, the molecules are so large
that the probability of complete crystallization is low. Part of a molecule can associate
with part of a different molecule or with another part of the same molecule, reducing
the odds of pure crystal formation. However, small crystalline domains are common in
synthetic polymers; polymers with fairly regular structures, usually homopolymers, are
most likely to crystallize partially. In these crystalline domains, fairly long segments of
molecules associate with each other in a regular way. The remaining parts of the same
molecules are unable to fit together regularly and remain amorphous. Whereas polymers
used in fibers and films (e.g., polyethylene and nylon) are often partly crystalline, polymers used in coating applications are, with few exceptions, completely amorphous.
Amorphous materials behave quite differently from crystalline materials. An important
difference is shown in Figure 2.3(a) and (b), which compare schematically changes in
specific volume of crystalline and amorphous materials with temperature. In the case of
a pure crystalline material, as temperature increases, initially there is a slow increase in
specific volume, owing to increasing vibrations of the atoms and molecules. Then, at a
Figure 2.3. Specific volume as a function of temperature for (a) a crystalline material and (b) an amorphous material; (c) free volume within an amorphous material as a function of temperature. Units of
specific volume are volume per mass (usually, cm3 g1). (Adapted from Ref. [3], with permission.)
14
POLYMERIZATION AND FILM FORMATION
specific temperature, the substance melts. The melting point Tm is the lowest temperature
at which the vibrational forces pushing molecules apart exceed the attractive forces
holding them together in crystals. With almost all substances (water is a notable exception), the molten compound occupies more volume at the same temperature than do the
crystals; because the molecules are freer to move in a molten compound, they “bounce”
their neighbors out of the way, leading to an abrupt increase in specific volume at Tm.
Above Tm, the specific volume of a liquid increases slowly with further increase in
temperature.
Amorphous materials behave differently, as shown in Figure 2.3(b). Starting from a
low temperature, there is a slow increase in specific volume as temperature increases,
but there is no temperature at which an abrupt change in volume occurs, there is no
melting point. Rather, there is a temperature at which there is a change in the rate of
increase of specific volume with temperature. Above that temperature, the thermal
expansion coefficient is larger than below it. This change of slope is not a phase
change; it is a second-order transition; that is, there is a discontinuity in a plot of the
derivative of volume change as a function of temperature. The temperature at which it
occurs is called the glass transition temperature, Tg. Tg is properly defined as the temperature at which there is an increase in the thermal expansion coefficient. By comparison, Tm
is a first-order transition; that is, there is a discontinuity in change of volume as a function
of temperature corresponding to the solid – liquid phase change. Unfortunately, Tg is often
improperly defined as the temperature below which a material is brittle and above which it
is flexible. Although there are many cases when this is true, there are cases when this
definition is misleading (Section 4.2). Perhaps part of the reason for the misunderstanding
is the connotation of the word glass, which we associate with a brittle material. Like all
amorphous materials, glasses undergo a second-order transition. In fact, the phenomenon
was first observed in the study of glasses—hence the name glass transition temperature.
The idea has proliferated that Tg is a phenomenon associated only with polymers. This
assumption is not true; many small molecules can be supercooled without crystallization
to form amorphous glasses that have a Tg. For example, the Tg of m-xylene is 125 K [7].
The Tg is always lower than Tm. Partially crystalline polymers show both a Tm and a Tg.
What is happening physically at Tg? As an amorphous material is heated, atoms in
the molecules vibrate with increasing energy, colliding with neighbors and shoving molecules apart for brief periods of time. At Tg, a few of the short-lived “holes” between the
molecules become large enough that an adjacent molecule or a segment of a polymer
molecule can fit between two molecules. Thus, Tg can be considered the lowest temperature at which segments of polymer molecules can move with some facility relative to
neighboring segments. The increase in coefficient of thermal expansion above Tg results
from the greater degree of freedom available to the molecule segments. The larger
volume between molecules gives more degrees of freedom, so the same increase in temperature gives a greater increase in volume. As the temperature rises, the specific volume
increases, but there is no more material—just the same material occupying more space.
What is in this “extra” volume? Nothing. It is called free volume, represented by the
hatched area in Figure 2.3(c). The molecular motions involved can be detected by spectroscopic techniques such as solid state nuclear magnetic resonance (NMR), and change as
polymers are heated through Tg [8,9].
Although it is difficult to overemphasize the importance of the concepts of Tg and free
volume in coatings science, our understanding of these parameters and our ability to
measure them are limited. Tg values of a material that are determined by different
2.1. POLYMERS
15
methods may not agree with each other; one must be careful in comparing Tg values of
different materials to be sure that they are based on consistent test methods. The classical
method of measuring Tg is dilatometry (measurement of specific volume as a function of
temperature). More commonly, it is measured by differential scanning calorimetry (DSC)
or by dynamic mechanical analysis (DMA) (Section 4.5). The measured Tg depends on the
measurement method and the conditions under which the measurement was made. Heating
rate is an important variable. The faster the rate of heating during the determination, the
higher the apparent Tg. When free volume is small, the rate of movement of molecules or
segments is slow. If the rate of heating is slow, there is more time for movement, and,
hence expansion, and the measured Tg is lower. (See Ref. [10] for further discussion of
methods of measuring Tg.)
Some scientists argue that Tg is not a real thermodynamic parameter. They point out
that if the determination of specific volume were done at a slow enough heating rate,
no transition would be observed and that rather than two straight lines, as shown in
Figure 2.3(c), there would be a smooth curve. Despite this controversy, Tg is a very
useful concept and is well understood in qualitative terms. Relationships between
polymer structure and Tg are understood well enough that it is often possible to make
reasonable predictions of Tg from knowledge of composition and M̄n. Important factors
affecting the Tg of thermoplastic polymers include:
1. Number average MW. Tg increases with increasing M̄n, approaching a constant
value at M̄n in the range 25,000 to 75,000, depending on the polymer structure. It is
logical that Tg is related to M̄n because decreasing M̄n results in an increasing fraction
of chain ends to chain middles, since chain ends have more freedom of movement than
chain middles. The relationship of Tg to M̄n is approximated by the equation
Tg ¼ Tg1
A
Mn
where Tg1 is the Tg at infinite MW and A is a constant (Tg is in Kelvin).
2. Polymer backbone flexibility. Tg is affected by the ease of rotation about bonds
in the polymer backbone. For example, the siloxane bond, Si22O22Si, rotates easily;
the Tg of poly(dimethylsiloxane) is 146 K (21278C) [11]. Aliphatic polyethers such as
poly(ethylene oxide), (22CH222CH222O22)n, also have low Tg, generally in the range
158 to 233 K, because there is considerable ease of rotation around the ether bond. The
Tg of polyethylene varies because although we commonly think of polyethylene as
being just chains of methylene groups, the backbone is actually substituted to varying
degrees with alkyl side chains such as ethyl groups. Also, most grades of polyethylene
are partially crystalline; only the amorphous areas show a Tg. However, all would agree
that the Tg of a long linear aliphatic chain is low, perhaps less than 200 K. The presence
of rigid aromatic rings in the polymer backbone increases Tg substantially.
3. Side chains. Pendant aromatic rings also lead to high Tg, for example, 373 K for
polystyrene, since ease of rotation is decreased. Similarly, pendant methyl groups and carboxymethyl groups increase Tg: for example, 281 K for poly(methyl acrylate) and 378 K
for poly(methyl methacrylate), which has both groups on alternate carbons of the chain. If
the side chains are several atoms long and flexible, they reduce Tg: for example, 219 K
for poly(n-butyl acrylate). However, if the side chain is short, bulky, and inflexible, it
has less effect and in some cases, raises Tg: for example, 314 K for poly(t-butyl acrylate).
16
POLYMERIZATION AND FILM FORMATION
TABLE 2.1. Glass Transition Temperatures (88 C) for Homopolymers of Various Monomers
Monomer
Free acid
Methyl
Ethyl
Isopropyl
n-Butyl
Isobutyl
t-Butyl
n-Hexyl
2-Ethylhexyl
iso-Decyl
a
Methacrylate
185
105
65
81
20
53
114
25
210
Acrylate
106
9
223
28
254
240
74
257
250
230
Monomer
Methacrylate
a
n-Tridecyl
iso-Tridecyl
2-Hydroxyethyl
2-Hydroxypropyl
Other monomers
Styrene
Vinyl acetate
Vinyl chloride
Vinylidene chloride
Acrylate
246
239
55
73
100
29
81
218
Mixture of C12 – 14.
Care must be taken when comparing Tg values to be sure that the determinations have
been done under consistent conditions and that the MWs are high enough to eliminate MW
effects. Table 2.1 gives the Tg of high MW homopolymers of a group of acrylic and
methacrylic esters and of other monomers commonly used as comonomers in polymers
for coatings [11 –13].
Copolymers have Tg values intermediate between those of the homopolymers. It is
common to use the Fox equation to estimate the Tg of copolymers: where w1, w2, w3,
etc. are the weight fractions of the various monomers in the copolymer and Tg1, Tg2,
Tg3, etc. are the Tg (Kelvin) of their high MW homopolymers.
1
w1 w2 w3
þ
þ
þ
¼
Tg(copolymer) Tg1 Tg2 Tg3
Somewhat better approximations can be calculated using a different mixing equation,
also devised by Fox, in which v1, v2, v3, etc. are the volume fractions from the various
monomers in the copolymer:
Tg(copolymer) ¼ v1 Tg1 þ v2 Tg2 þ v3 Tg3 þ
This equation is not widely used because some of the homopolymer densities needed to
calculate v1, v2, v3, etc. are not readily available.
Gupta [14] reports an extensive study of the estimation of Tg of acrylic copolymers.
He recommends use of the van Krevelen equation for estimation of Tg:
Tg ¼
Yg
M
where M is the MW of the repeat unit and Yg is a molar glass transition factor. Gupta’s
values for the Tg of n-butyl methacrylate (108C) and of 2-ethylhexyl acrylate (2638C)
vary considerably from the values given in Table 2.1, illustrating that different values
are often found in the literature.
The Tg of cross-linked polymers is controlled by four factors and their interactions: Tg
of the segments of polymer between the cross-links, the cross-link density (XLD), the presence of dangling ends, and the presence of cyclic segments [15]. Although generalized
2.2. POLYMERIZATION
17
equations showing the relationships of these factors with Tg have been developed, the
complex relationships are not fully understood. The Tg of the polymer segments
between cross-links is governed by the chemical structures of the resin and the crosslinking agent, by the ratio of these components, and by the extent of the cross-linking reaction. The factors discussed in connection with thermoplastic polymers apply in terms of
their effects on the Tg of the segments of cross-linked polymer chains. Since cross-links
restrict segmental mobility, Tg increases as cross-link density increases. On the other
hand, Tg decreases with an increasing proportion of dangling ends—that is, chain segments with unreacted cross-link sites.
Solutions of polymer in solvent and of solvent in polymer have Tg values intermediate
between the Tg of the polymer and that of the solvent. The Tg of solutions increases with
increasing polymer concentration. When the weight fraction of solvent ws is less than
about 0.2, a simple mixing equation gives reasonable correlation between experimental
and predicted results [16]: Over a wider range of concentrations, this simple equation
gives poor correlations.
Tg(solution) ¼ Tg(solution) k ws
For solutions of oligomeric n-butyl methacrylate in m-xylene [7], Eq. 2.1 gives a good
fit between observed and predicted data over the entire range from pure solvent to
solvent-free oligomer. Here, ws and wo are weight fractions and Tgs and Tgo are the Tg
of the solvent and the oligomer, respectively. Although Eq. 2.1 accurately describes a
limited number of oligomer and polymer solutions, its generality is not fully established.
1
Tg(solution)
¼
ws
wo
þ
þ Kws wo
Tgs Tgo
(2:1)
An excellent general review of glass transition is available in Ref. [17]. A review of free
volume considerations in coatings may be found in Ref. [18].
2.2. POLYMERIZATION
There are two major classes of polymerization reactions: chain-growth and step-growth.
The common denominator of chain-growth polymerization is that reactions are chain
reactions. Frequently, chain-growth polymerization is called addition polymerization,
but this terminology is inadequate. While all chain-growth polymerizations involve
addition reactions, not all addition polymerizations involve chain-growth reactions—
some are step-growth reactions.
2.2.1. Chain-Growth Polymerization
Chain-growth polymerization, initiated by free radicals, is the most commonly used chaingrowth polymerization for making polymers for coatings. Its mechanisms and kinetics
have been studied extensively. Reference [19] gives an extensive review of the topic,
especially the kinetics of the reactions. Free radical chain-growth polymerizations of
most interest to coating applications are solution polymerization (Chapter 8) and emulsion
polymerization (Chapter 9). A related process of importance in coatings is the autoxidation
18
POLYMERIZATION AND FILM FORMATION
involved in cross-linking drying oils and drying oil derivatives (Chapter 14). The discussion in this section applies to solution polymerization, although many of the principles are
also true for emulsion polymerization.
Three types of chemical reactions—initiation, propagation, and termination—are
always involved in chain-growth polymerization; and a fourth, chain transfer, often
plays a significant role. Initiation occurs when an initiator (I) reacts to form an initiating
free radical (I) (Eq. 2.2), which, in turn, adds rapidly to a monomer molecule to form a
second free radical (Eq. 2.3):
ð2:2Þ
ð2:3Þ
The polymer chain grows by the propagation reaction, in which the monomer free
radical adds to a second monomer molecule to extend the chain while forming a new
free radical:
ð2:4Þ
Propagation reactions are very fast, so fast that a chain with hundreds of mers can
grow in a fraction of a second. At any moment, the concentrations of monomer and
polymer greatly exceed the concentration of growing polymer molecules (ca. 1026 ML21).
(Exceptions are controlled free radical polymerizations, Section 2.2.1.1.)
The final stage is termination of the growing chain. Two common types of termination
reactions are combination (Eq. 2.5) and disproportionation (Eq. 2.6). In most free radical –
initiated polymerizations, the rate of propagation (Eq. 2.3) is faster than the rate of
initiation, which is limited by the rate of Eq. 2.2.
ð2:5Þ
ð2:6Þ
where P represents an attached polymer chain.
Side reactions also occur; among the most important are chain transfer reactions, in
which the free radical on the end of the propagating polymer chain abstracts a hydrogen
atom from some substance X22H present in the polymerization reaction mixture:
ð2:7Þ
The net effect of chain transfer is to terminate the growing chain while generating a free
2.2. POLYMERIZATION
19
radical, which may start a second chain growing. X22H may be a solvent, a monomer,
a molecule of polymer, or a chain transfer agent, a reactant that is added to the polymerization reactants to cause chain transfer. When chain transfer is to a solvent or a
chain transfer agent, MW is reduced. When chain transfer is to a polymer molecule,
growth of one chain stops but a branch grows on the polymer molecule; the result is a
higher M̄w/M̄n.
Note that the structures of the propagating polymer chains show substitution on
alternate carbon atoms. This structure results from the favored addition of free radicals
to the CH2 end of the monomer molecules: head-to-tail addition. With almost all monomers, head-to-tail addition predominates, but a small fraction of head-to-head addition
occurs. The result is a polymer with most of the substitution on alternating carbons in
the chain, but with a few chain segments having substitution on adjacent carbons. The
effect of a small fraction of head-to-head structure is generally negligible but sometimes
has significant consequences for exterior durability and thermal stability.
Initiators, incorrectly called catalysts, are used in low concentration [usually in the
range 1 to 4 wt% (weight percent), but sometimes higher when low MW is desired]. A
variety of free radical sources has been used. Two classes of initiators are used most
often: azo compounds such as azobisisobutyronitrile (AIBN) and peroxides such as
benzoyl peroxide (BPO) or t-amyl peracetate. AIBN is fairly stable at 08C but decomposes
relatively rapidly when heated to 70 to 1008C to generate free radicals. A substantial
fraction of the resulting radicals initiate polymerization, although some combine to
form a coupling product. The half-life of AIBN is about 5 hours at 708C and about
7 minutes at 1008C.
BPO decomposes at similar temperatures—its half-life is about 20 minutes at 1008C.
The reactive benzoyloxy free radical generated can initiate polymerization; also, it can
dissociate (rapidly at higher temperatures, such as 1308C) to yield a very highly reactive
phenyl free radical and CO2:
A range of monomers is capable of propagating a free radical – initiated chain reaction.
Most are alkenes having an electron-withdrawing group; methyl acrylate (MA) and methyl
methacrylate (MMA) are important examples:
Copolymers containing a preponderance of acrylic and methacrylic ester monomers are
20
POLYMERIZATION AND FILM FORMATION
called acrylic polymers, or often, just acrylics. They are used extensively in coatings.
Control of MW and MW distribution is critical in preparing polymers for coatings.
There are four major factors that affect MW with the same monomer, initiator, and
solvent:
1. Initiator concentration. The higher the initiator concentration, the lower the MW.
When the initiator concentration is higher, more initiating free radicals are generated to
react with the same total amount of monomer. More chains are initiated and terminated;
the M̄n and M̄w of the polymer produced are lower.
2. Temperature. At higher temperatures, more initiator is converted into initiating
free radicals in a given time, increasing the concentration of growing chains and the
probability of termination. As with increasing initiator concentration, the result is lower
M̄n and M̄w.
3. Monomer concentration. The higher the monomer concentration, the higher the M̄n
and M̄w. The highest MW is obtained in a solvent-free reaction mixture. With the same
concentration of growing free radical ends, a higher monomer concentration increases
the probability of chain growth relative to termination.
4. Solvent concentration. The higher the solvent concentration, the lower the M̄n and
M̄w. Higher solvent concentration leads to lower monomer concentration, reducing MW
as described previously. Furthermore, most solvents undergo chain-transfer reactions to
varying extents, which also reduce MW.
To the extent that any of these factors change during a polymerization process, M̄n
and M̄w of the polymer molecules also change. The usual result is broader MW distribution. Changes in monomers also produce changes in MW distribution. Consider the
difference between MA and MMA. Since the free radicals at the ends of growing
chains of poly(methyl methacrylate) (PMMA) are sterically hindered, termination by
combination is impeded, and termination by disproportionation predominates. On the
other hand, with MA, a major fraction of the termination reactions occur by combination. Theoretical calculations show that for high MW polymers, the lowest M̄w/M̄n
attainable with termination by combination is 1.5, while the minimum with termination
by disproportionation is 2.0. In actual polymerization processes, M̄w/M̄n is usually
higher, although with very high initiator concentrations, polydispersities tend to be
lower. No basic studies have been reported to account for the low polydispersities
with high initiator concentrations. It is speculated that lower polydispersity with
high initiator concentration reflects the resulting higher rate of initiation relative to
propagation, which is expected to promote “living polymerization,” as discussed in
Section 2.2.1.1.
Chain transfer to polymer must also be considered. This reaction occurs to a degree
in the polymerization of MMA but is more important in the polymerization of MA.
The hydrogen on the carbon to which the carboxymethyl group is attached is more
susceptible to abstraction by free radicals than is any other hydrogen in PMA or
PMMA. The new free radical on the PMA chain can now add to a monomer molecule,
initiating growth of a branch on the original polymer molecule. The result is a polymer
containing branched molecules and having a larger M̄w/M̄n than predicted for ideal
linear polymerization. In extreme cases, chain transfer to polymer results in very
2.2. POLYMERIZATION
21
broad MW distributions and, ultimately, to formation of gel particles through
cross-linking.
Branching can also result from the abstraction of hydrogen atoms from a polymer chain
by initiating free radicals. Phenyl free radicals from high temperature decomposition
of BPO are so reactive that they will abstract almost any aliphatic hydrogen, leading to
substantial branching. BPO at high (e.g., 1308C) temperature is a good choice when
branching is desired, but in most cases it is desirable to minimize branching. In these
cases, azo initiators such as AIBN, or aliphatic peroxy initiators, are preferred over
BPO. Initiator residues remain attached to the polymer chain ends. For high MW polymers, they have a negligible effect on most properties; for oligomers, they may have an
appreciable effect, particularly on exterior durability (Section 8.2.1).
MW and MW distribution also depend on solvent structure. For example, substituting
xylene for toluene and keeping other variables constant leads to a decrease in MW. Since
each xylene molecule has six abstractable hydrogen atoms, whereas toluene has only
three, the probability of chain transfer increases and average MW decreases. To prepare
a low MW polymer or oligomer, one can add a compound that undergoes facile hydrogen
abstraction as a chain-transfer agent. If the hydrogen atoms are readily abstracted, the
addition of even relatively low concentrations of a chain transfer agent can lead to a substantial reduction in MW. Mercaptans (RSH) are widely used as chain-transfer agents.
Other variables affecting MW and MW distribution are the decomposition rate of the
initiator and the reactivity of the resulting free radicals. To achieve a low M̄w/M̄n, concentrations of reactants must be kept as constant as possible through the polymerization. It is
undesirable to simply charge all of the monomers, solvents, and initiators into a reactor
and heat the mass to start the reaction. This procedure is sometimes used in small scale
laboratory reactions, but never in production. At best, it yields a high M̄w/M̄n; at worst,
the reaction may run violently out of control, because free radical polymerizations are
highly exothermic. Instead, one charges some of the solvent into the reactor, heats to
reaction temperature, and then adds monomer, solvent, and solutions of initiator to the
reactor at rates such that the monomer and initiator concentrations are kept as constant
as possible. Adding monomer at a rate that maintains a constant temperature leads to a
fairly constant monomer concentration. The appropriate rate of addition of the initiator
solution can be calculated from the rate of its decomposition at the temperature being
used. Keeping solvent concentration constant is more complex since as the polymerization
proceeds, polymer is accumulating; in a sense, the polymer becomes a part of the “solvent”
for the polymerization. Solvent is added at a decreasing rate so that the other concentrations stay as constant as possible. Perfect control is not possible, but careful
attention to detail makes an important difference in the M̄w/M̄n of the polymer produced.
Copolymerization of mixtures of unsaturated monomers further complicates the
situation. The rates of reaction involved in the various addition reactions depend on the
structures of the monomers. If the rate constants for all of the possible reactions were
the same, the monomers would react randomly and the average composition of molecules
of substantial length would all be the same. However, the rate constants are not equal.
If polymerization is carried out by putting all of the reactants in a flask and heating, the
first molecules formed would contain more than proportional amounts of the most reactive
monomer, and the last molecules formed would have an excess of the least reactive
monomer. This situation is usually undesirable. Such effects have been studied extensively, and equations have been developed to predict the results with different monomer
22
POLYMERIZATION AND FILM FORMATION
combinations. (See Ref. [19] for a detailed discussion of copolymerization.) In actual
practice, the problem is somewhat less complex. Reactions are not run in bulk, but
rather, as mentioned above, monomers, solvent, and initiator solution are added gradually
to the reaction mixture. If the additions are carefully controlled so that the rate of addition
equals the rate of polymerization, copolymers having reasonably uniform composition that
corresponds to the feed ratio are obtained with most monomers. This procedure, called
monomer-starved conditions, results in polymerization under conditions in which
the concentration of monomers is low and fairly constant. Further process refinements
are possible by adding individual reactants or mixtures of reactants in two or three
streams at different rates. Computer modeling of the processes can help achieve the
desired results.
2.2.1.1. Controlled Free Radical Polymerizations For several decades, considerable effort has been directed to the preparation of acrylic polymers with narrow MW distributions. This can only be accomplished by finding chain-growth polymerization processes
in which the rate of initiation is much faster than the rate or propagation and the rates of termination reactions are very slow. Under these circumstances, all polymer chains start growing
early in the process and grow at about the same rate under about the same conditions. In these
processes the polymer chain ends often remain reactive after all monomer has been consumed,
in which case they are called living radical polymerizations. Early efforts focused on anionic
polymerizations and group transfer polymerizations. More recently, attention has shifted to
controlled free radical polymerizations (CFRP). There has been controversy about how to
define living radical polymerizations and about nomenclature in general. A sensible proposal
is that if after all monomer has been consumed, polymerization continues when additional
monomer is supplied, the polymerization is living [20].
All CFRP methods mediate the rate of propagation by including some substance that
bonds reversibly to the radical at the growing end of the polymer chain. It is possible to
slow the propagation rate by many orders of magnitude, satisfying the kinetic requirements for living polymerization. Since these polymerizations are free radical processes,
they are relatively insensitive to impurities with active hydrogens and can be used to
copolymerize monomers with active hydrogens. These methods enable preparation of
block, alternating block, and gradient copolymers by sequential addition of monomers;
and they can be adapted to synthesize a limitless variety of linear, graft, and star polymers,
copolymers, and macromonomers (polymerizable oligomers or polymers) from most
of the common acrylic and styrenic monomers. The extensive literature about CFRP
processes lacks consistency about how the processes are named. One way to classify
them is as follows:
1. Stable free radical polymerization (SFRP), also called nitroxide-mediated polymerization (NEM), in which polymerization is mediated by nitroxide compounds [21].
2. Reversible addition-fragmentation chain transfer polymerization (RAFT), which
involves mediation by certain dithioesters or xanthates (trithiocarbonates) [22].
3. Transition metal – mediated living radical polymerizations (TMMLRP), two
methods can be distinguished, although some authors lump them together:
a. Atom transfer radical polymerization (ATRP), in which the mediator is a metal
salt, often of copper, with a carefully chosen organic ligand and an organic halide that
can undergo a redox reaction with the metal to trigger the polymerization [23 –25].
2.2. POLYMERIZATION
23
b. Catalytic chain transfer polymerization (CCTP), which uses conventional initiator,
usually an azo type, in the presence of a chelated cobalt (or other) metal salt. CCTP is
especially useful for preparing macromonomers, and it yields products with relatively
little color [26,27].
TMMLRP methods have thus far been the primary techniques used in work aimed
at coating applications. Much information on other techniques has been published, and
there is a book covering a symposium on controlled free radical polymerizations [28].
Another source is Ref. [29]. The various processes, with an emphasis on ATRP, are
discussed in a review paper. It emphasizes the importance of interrelationships among
monomer, halogen compound, and concentrations on control of the polymerization.
Possible application to the preparation of oligomers for UV curing is also discussed [30].
Recent patents and journals are rich with examples of CFRP synthesis of acrylic
polymers; a few examples are mentioned here. ATRP is used for preparation of an
acrylic resin for use in powder coatings [31]. A reactor is loaded with toluene, CuBr2,
Cu powder, 2,20 -bipyridyl, diethyl 2-bromo-2-methylmalonate, MMA, BMA, and
t-BMA, heated for 4 hours at 858C, cooled, and filtered. After solvent removal by
distillation, the residue is dissolved in dioxane, 1 M aqueous HCl is added and the
solution is refluxed for 4 hours to hydrolyze the malonate ester groups. The solution is
cooled, neutralized with Na2CO3, filtered, and the solvent and water are distilled off to
give a solid resin that is pulverized. The resin had a P̄w/P̄n of 1.25 compared with 1.95
for a similar resin made by conventional free radical polymerization. The melt viscosity
at 1808C of the ATRP polymer was 11.2 Pa. s, compared with 57.2 Pa. s for the
conventional resin. A powder coating using bis(N,N-di-2-hydroxyethyl) adipamide as a
cross-linker was compared with a similar powder coating made with the conventional
free radical copolymer. The ATRP powder coating gave cured films with better leveling
than the conventional polymer and had superior storage stability. The lower melt viscosity
due to the narrow MW distribution resulted in better leveling. The reduced amount of very
low MW molecules reduced the fusion of powder particles during storage.
Block acrylic copolymers prepared by ATRP are useful aqueous pigment dispersants
[23], as are graft copolymers made with CCTP [32]. See Section 21.3.1 for further
examples and discussion of the use of the copolymers in pigment dispersion.
An acrylic resin prepared by CFRP is being used in automotive clear coats. The solids of
the SuperSolidsTM clear coat are reported to be 65%, with the possibility of future coatings
having up to nearly 90%. At the same time, the scratch and mar resistance are substantially
higher. Dupont was awarded the EPA clean air achievement award for the work [33].
High-solids acrylic block copolymers, including ones in which HEA is a comonomer,
have been synthesized by CFRP using a mediating nitroxy. The oligomers obtained have
narrow MW distributions and hence low viscosities for their MW [34].
2.2.2. Step-Growth Polymerization
A second class of polymerization that is important in the coatings field is step-growth
polymerization. As the name indicates, the polymer is built up a step at a time. The
term condensation polymerization has been used for this process because early examples
involved condensation reactions, reactions in which a small molecule by-product such as
water is eliminated. Although both terms are still used, step-growth polymerization is more
appropriate because many step-growth polymerizations are not condensation reactions.
24
POLYMERIZATION AND FILM FORMATION
Step-growth polymerization reactions are used in coatings in two ways. One is in the
preparation of resins for use as vehicles, and the other is for cross-linking after a coating
has been applied to a substrate. In this introductory section, polyester formation is used to
illustrate the principles involved; polyesters are discussed more broadly in Chapter 10.
Of the many reactions that form esters, three are commonly used in making step-growth
polymers and oligomers for coatings: direct esterification of an acid with an alcohol,
transesterification of an ester with an alcohol, and reaction of an anhydride with an
alcohol. The first two of these reactions proceed rapidly only at elevated temperature,
and process temperatures of 2008C and higher are common.
When one of the reactants is monofunctional, polymer cannot form. If, however, both reactants have two or more functional groups, a polymer can be made. When all monomers are
difunctional, linear polymers form. Linear step-growth polymers are commonly used in fibers,
films, and plastics, but their use in coatings is restricted to low-MW oligomers. Most polyester
resins used in coatings have relatively low MWs and are branched resins made using at least one
monomer with three or more functional groups. After application of the coating, the terminal
groups on the branch ends are reacted with a cross-linker to form the cured coating.
When a difunctional acid (AA) reacts with a difunctional alcohol (BB) in a direct
esterification reaction, the MW builds up gradually. Under ideal conditions, polymer
chains averaging hundreds of mers per molecule can be made, but this can occur only
if (1) the reactants AA and BB contain no monofunctional impurities, (2) exactly equimolar amounts of AA and BB are used, (3) the reaction is driven virtually to completion,
and (4) side reactions are negligible. If one reactant is present in excess, terminal groups of
the excess monomer predominate. The MW of the completely reacted system is progressively lower as the difference from equal equivalents is increased. For example, if 7 mol of
dibasic acid are completely reacted with 8 mol of a dihydroxy compound (a diol), the
average molecule will have terminal hydroxyl groups, as shown in the following equation
(here, for convenience, AA and BB represent both the reactants and the mers in the polymer):
The symbol F is used for the functionality of monomers, which is the number of reactive
groups per molecule, for example:
2.2. POLYMERIZATION
25
Note that the anhydride groups in phthalic anhydride and trimellitic anhydride count as two
functional groups since they can form two ester groups during polymerization.
The average functionality, represented byF̄, of a mixture of monomers containing equal
equivalents of hydroxyl and carboxyl groups is calculated as follows:
total equivalents
F ¼
total moles
Most coating polyester resins are hydroxy-functional and are made using monomer
mixtures with excess hydroxyl groups. Some of the hydroxyl groups have no carboxyl
groups to react with, so the equation must be modified to reflect only the total number
of equivalents that can react. In a resin with excess hydroxyl groups prepared from
dicarboxylic acids, the total equivalents that can react correspond to twice the number
of equivalents of carboxylic acid groups:
total equivalents that can react 2(equiv of COOH)
¼
F ¼
total moles
total moles
A simple formulation for a polyester oligomer is given in Table 2.2.
A further type of functionality important in designing resins is the functionality of the
resin. To distinguish this functionality from that of the monomers and the monomer
mixture, the symbol f is used. Since almost all coating resins are made using some triol
or tetrol, one generally uses a number average functionality f̄n:
number of functional groups in a sample
f n ¼
number of molecules in the sample
The value of f̄n can be calculated from M̄n and the number of functional groups per
sample weight obtained by analysis.
Esterification of a carboxylic acid with an alcohol is acid catalyzed. In the absence of a
strong acid catalyst, the rate r is approximately third order in reactants: with one carboxyl
group reacting with the alcohol and a second catalyzing the reaction. Since water is
generally removed rapidly from the reaction mixture, it is reasonable to use Eq. 2.8,
which disregards the reverse reaction.
r ¼ k½RCOOH2 ½R0 OH
(2:8)
Because of the second order dependence on acid concentration, the rate decreases
precipitously as reaction proceeds. For example, polyesterification of equal moles
TABLE 2.2. Polyester Formulation
Component
Adipic acid
Phthalic anhydride
Neopentyl glycol
Glycerol
F ¼ 8.6/3.8 ¼ 1.89
Moles
Equivalents
0.9
0.9
1.0
1.0
3.8
1.8
1.8
2.0
3.0
8.6
26
POLYMERIZATION AND FILM FORMATION
of diethylene glycol with adipic acid at 1608C without a catalyst is 60% complete in 1 hour
but requires 27 hours to reach 94.5% conversion and would require years to reach 99.8%
conversion [35]. The reaction is catalyzed by strong acids. In many cases, conventional
strong acids cause side reactions and discoloration, so the most appropriate catalysts are
organotin compounds such as monobutyltin dioxide, or titanate esters. It has been
shown that both the organotin compound and carboxylic acid act as catalysts [36].
The kinetics of ideal step-growth polyesterification for difunctional reactants can be
analyzed in terms of p, the fractional extent of reaction; np, the number fraction of differing
degrees of polymerization; Pn, the degree of polymerization; and wp, the weight fraction of
molecules [37]. As p increases, the degree of polymerization builds up slowly at first—at
Figure 2.4. (a) Weight fraction distribution wP of molecules in a linear step-growth polymer for several
extents of reaction p; (b) number, or mole fraction, distribution nP. (From Ref. [37], with permission.)
2.3. FILM FORMATION
27
p ¼ 0.5 (corresponding to 50% conversion), P̄n is only 2, P̄n is only 10 at p ¼ 0.9, and
p ¼ 0.998 is required to reach a P̄n of 500. Thus, with difunctional monomers, high MW
can only be attained when the COOH/OH mole ratio is 1.00 and when esterification is
driven beyond p ¼ 0.99. This is difficult because of the decreasing reaction rate at
high values of p. Note that as shown in Figure 2.4(a), the number of unreacted monomer
molecules remains higher than that of any other single species in the reaction mixture, no
matter how high p becomes. As shown in Figure 2.4(b), P̄n, the peak of the P distribution
curve, reaches substantial values only at high p values. In the case of high MW linear
polymers, under ideal conditions, the M̄w/M̄n obtained in step-growth polymerizations is 2.
2.3. FILM FORMATION
Most coatings are liquids with a viscosity appropriate for the application method, generally in the range 0.05 to 1 Pa. s at high shear rates. After application, the liquid is converted
to a dry (i.e., solid) film. In powder coatings, the powder is liquefied after application and
then converted to a solid film. The chemical and physical changes that occur in this process
are called film formation, a process that is critical to the ultimate appearance and performance of the coating.
If the applied coating were crystalline, there would be no difficulty in defining a solid
film. The film would be solid if the temperature were below its freezing point; however,
binders of coatings are almost always amorphous, with no melting point and no sharp
demarcation between a liquid and a solid. A useful definition of a solid film is that it
does not flow significantly under the pressures to which it is subjected during use.
Thus, one can define whether a coating is a solid under a set of conditions by stating
the minimum viscosity required such that flow does not exceed a specified extent under
specified pressure and time. For example, it is reported that a film is dry-to-touch if its
viscosity is greater than about 103 Pa . s [38]. However, if the definition of a solid is that
the film resists blocking—that is, sticking together when two coated surfaces are put
against each other for 2 seconds under a pressure of 1.4 kg cm22 (20 psi)—the
viscosity has to be greater than about 107 Pa . s.
For thermoplastic binders, we can use this information to predict polymer structures
that could meet such tests. Using a simplified form (Eq. 2.9; in the equation, T is in
Kelvin) of the Williams – Landel – Ferry (WLF) equation (Section 3.4), using “universal
constants” and assuming that the viscosity at Tg is 1012 Pa . s, one can estimate the Tg of
a binder required so that a film does not flow under some set of circumstances:
ln h ¼ 27:6
40:2(T Tg )
51:6 þ (T Tg )
(2:9)
Using Eq. 2.9, we can estimate the appropriate (T 2 Tg) value required for a film to be
dry-to-touch, that is, to have a viscosity of 103 Pa . s. The T 2 Tg value calculated is 548C,
which corresponds to a Tg of 2298C for a film to be dry-to-touch at a temperature T
of 258C. The Tg calculated for block resistance (at 1.4 kg cm22 for 2 s at 258C, i.e., for
a viscosity of 107 Pa . s) is 48C. Because there is considerable variation in the WLF
universal constants, these Tg values are not exact, but they can serve as a formulation
guide. Since we have a reasonable idea of the relationships between structure and Tg
(Section 2.1.2), we can approximate the requirements to make a binder with the viscosity
28
POLYMERIZATION AND FILM FORMATION
necessary to pass a particular test. If the coating has to pass a test at a temperature higher
than 258C, the Tg of the binder must be higher, since the free volume dependence is on
T 2 Tg. If the pressure to which the film is to be subjected is higher or the time under
pressure is to be longer, the Tg must be higher.
2.3.1. Film Formation by Solvent Evaporation from
Solutions of Thermoplastic Binders
Films can be formed in a variety of ways. One of the simplest is to dissolve a polymer in
solvent(s) at a concentration needed for application requirements, apply the coating, and
allow the solvent to evaporate. Let us illustrate with a copolymer of vinyl chloride, vinyl
acetate, and a hydroxyl-functional vinyl monomer with an M̄n of 23,000 that is reported to
give coatings with good mechanical properties without cross-linking [39]. The Tg of the
copolymer is reported to be 798C. A solution in methyl ethyl ketone (MEK) with a
viscosity of 0.1 Pa . s required for spray application would have about 19 NVW (nonvolatile weight, i.e., wt% solids) and about 12 NVV [nonvolatile volume, i.e., volume
percent (vol%) solids]. MEK has a high vapor pressure at room temperature and evaporates rapidly from a thin layer. In fact, a sizable fraction of the MEK evaporates from the
atomized spray droplets between the time they leave the spray gun and the time they
reach the substrate. As solvent evaporates from a film, viscosity increases, and the film
will be dry-to-touch soon after application. Also, in a short time, the coating will not
block under the conditions mentioned previously. Nevertheless, if the film is formed at
258C, the “dry” film contains several percent of retained solvent. Why?
In the first stage of solvent evaporation from a film, the rate of evaporation is essentially
independent of the presence of polymer. The evaporation rate depends on the vapor
pressure at the temperature, the surface area/volume ratio, and the rate of air flow over
the surface. However, as solvent evaporates, viscosity increases, Tg increases, free
volume decreases, and the rate of loss of solvent becomes dependent on how rapidly
solvent molecules can diffuse to the surface of a film so that they can evaporate. The
solvent molecules must jump from free volume hole to free volume hole to reach the
surface. As solvent loss continues, Tg increases, free volume decreases further, and
solvent loss slows. If the film is formed at 258C from a solution of a polymer that,
when solvent free, has a Tg greater than 258C (in this example, it is 798C), the film
retains considerable solvent even though it is a hard “dry” film. Solvent slowly leaves
such a film, but it has been shown experimentally that 2 to 3% of solvent remains after
several years at ambient temperature. To assure nearly complete removal of solvent in
a reasonable period of time requires baking at a temperature significantly above the Tg
of the solvent free polymer. Solvent loss from films is discussed in more detail in
Section 18.3.4.
2.3.2. Film Formation from Solutions of Thermosetting Resins
A drawback of solution thermoplastic polymer –based coatings is that the MWs required
for film properties require high solvent levels (on the order of 80 to 90 vol%) to achieve the
viscosity for application. Less solvent is needed for a coating based on solutions of
lower-MW thermosetting resins. After application, the solvent evaporates, and chemical
reactions cause polymerization and cross-linking to impart good film properties. Many
combinations of chemical reactions are used in thermosetting coatings, as discussed in
2.3. FILM FORMATION
29
Chapters 8 to 17. A critical aspect of the design of a coating is the selection of components
that give required mechanical properties (Chapter 4). In this section we cover only the
general principles of cross-linking reactions.
A problem with all thermosetting systems is the relationship between coating stability
during storage and the time and temperature required to cure a film after application.
Generally, it is desirable to be able to store a coating for many months, or even several
years, without a significant increase in viscosity resulting from reaction during the
storage period. On the other hand, after application, one would like to have the crosslinking reaction proceed rapidly at the lowest possible temperature.
As formulations are shifted to higher solids to reduce VOC emissions, there are
higher concentrations of functional groups and there is greater difficulty in formulating
storage stable coatings. The problem results not only from the presence of less solvent,
but also from the lower MWs and lower equivalent weights needed to achieve a
cross-linked film of acceptable cross-link density. Both factors increase the concentration of functional groups in a stored coating. Concentration of reactants in a film
increases after application and evaporation of solvent, which increases reaction rates,
but since the MW is lower, more reactions must occur to achieve the desired cross-linked
film properties.
What controls the rate of a reaction? We can consider this question broadly as a reaction
between two groups, represented by the symbols A and B, that react to form a cross-link
A – B:
In the simplest cases, one can express the rate of reaction r of A and B as where k is the
rate constant for the reaction between A and B by Eq. 2.10 at a specified temperature, and
[A] and [B] represent the concentration of the functional groups in terms of equivalents per
liter. The rate constant is the reaction rate when [A] [B] equals 1 equiv.2 L22.
r ¼ k½A½B
(2:10)
To minimize the temperature required for curing while maintaining adequate storage
stability, it is desirable to select cross-linking reactions for which the rate depends strongly
on temperature. This dependence is reflected in the rate equation by the dependence of k on
temperature. It is commonly taught in introductory organic chemistry classes that rate constants double with each 108C rise in temperature. That generalization is true for only a
limited number of reactions within a narrow temperature range near room temperature.
A better estimate, but still an estimate, of the temperature dependence of k is given by
the empirical Arrhenius equation where A is the preexponential term, Ea is the thermal
coefficient of reactivity (commonly labeled activation energy), R is the gas constant,
and T is temperature (in Kelvin).
ln k ¼ ln A
Ea
RT
(2:11)
Reaction rate data that fit the equation give straight lines when ln k is plotted against
1/T, as illustrated in Figure 2.5. As seen in plot (a) of competing reactions (1) and (2),
where A(1) ¼ A(2) and Ea(1) . Ea(2), the temperature dependence of rate constants
30
POLYMERIZATION AND FILM FORMATION
Figure 2.5. Arrhenius plots for competing reactions: (a) A(1) ¼ A(2), Ea(1) . Ea(2); (b) A(3) . A(1),
Ea(1) ¼ Ea(3); (c) A(4) . A(1), Ea(4) . Ea(1). (From Ref. [40], with permission.)
increases with increasing values of Ea. However, the rate of reaction (2) is slower than
that of reaction (1) at all temperatures. This effect can be counteracted by selecting a
reaction with a higher A value, as shown in plot (b), where A(3) . A(1) and Ea for
the two reactions is equal. If both A and Ea are sufficiently greater for one reaction
than for another, the rate constant at storage temperature could be smaller while the
rate constant at a higher temperature could be larger, as shown schematically in plot
(c) [40].
Term A is controlled predominantly by entropic factors, or more specifically, by
changes in randomness or order, as the reaction proceeds to the activated complex in
the transition state. Three important factors are: (1) unimolecular reactions tend to
exhibit larger A values than those with a higher molecular order; (2) ring opening reactions
tend to have high A values; and (3) reactions in which reactants become less polar exhibit
larger A values. The importance of these factors (in particular, factor 3) depends on the
2.3. FILM FORMATION
31
reaction medium; accordingly, solvent selection can have a significant effect on storage
stability.
Although unimolecular reactions are desirable for high A values, cross-linking reactions are necessarily bimolecular. A way around this problem is to use a blocked reactant
BX that thermally releases a reactant B by a unimolecular reaction—most desirably, with
ring opening and decreasing polarity—followed by cross-linking between A and B:
Another approach is to use a blocked catalyst CX, where C catalyzes the cross-linking
of A and B:
An important consideration is that the cross-linking reaction, which follows unblocking, should be faster than the reverse reaction, which regenerates the blocked reactant
or catalyst. Although one often encounters “threshold” or “unblocking” temperatures
for reactions in the literature, such minimum reaction temperatures do not exist in the
kinetics of reactions. Reactions proceed at some rate at any temperature. Threshold or
unblocking temperatures are actually the temperatures at which some observable extent
of reaction has occurred within a specified time interval.
These concepts are helpful for understanding the differences we see between storage
stabilities, but another reason for understanding these kinetic considerations is that they
can be used to predict whether any chemical reaction will ever be found to meet some
combination of stability and cure schedule that might be desired for a coating. Pappas
and Hill have carried out calculations to permit such predictions [25,41]. They made a
set of reasonable assumptions about reactive group concentrations during storage and in
the applied film, permissible degrees of reaction during storage, and required extents of
reaction during curing. Using these assumptions, they calculated the ratio of rate constants
needed to permit any time of storage with any time of curing. In turn, this allowed calculation of Ea and A values as a function of any combination of storage and curing temperatures. Kinetic parameters calculated for unimolecular blocked reactant systems that
proceed 5% over a six-month period at 308C (storage temperature) and 90% in 10
minutes at various curing temperatures are listed in Table 2.3.
TABLE 2.3. Kinetic Parameters as a Function of Cure
Temperature
T (8C)
175
150
125
100
Source: Ref. [40].
a
1 kJ ¼ 0.239 kcal.
A (s21)
Ea (kJ mol21)a
1010
1012
1017
1024
109
121
146
188
32
POLYMERIZATION AND FILM FORMATION
Rate constants and kinetic parameters are independent of concentration for unimolecular reactions but not for bimolecular reactions. Kinetic parameters calculated for bimolecular (i.e., second order) reactions using concentrations corresponding to high-solids
coatings were of similar magnitude to those for unimolecular reactions. Although
the values in Table 2.3 represent order of magnitude calculation, they provide useful
guidelines for avoiding development projects aimed at impossible goals and provide
insights for design of thermosetting coatings. Kinetic parameters are known for many
chemical reactions. As a point of reference, a reasonable upper limit of A values for
unimolecular reactions is 1016 s21, which corresponds to an upper value for the frequency
of simple vibrations. For bimolecular reactions, A values tend to be less than
1011 L mol21 s21, an upper limit for the rate constant of diffusion, which must precede
reaction. However, as shown in Table 2.3, a coating stable at 308C would require an A
value of 1024 s21 if it were to cure within 10 minutes at 1008C. No such reaction is
known or even conceivable. Users would like to have package stable coatings that cure
in a short time at 808C, a convenient temperature for low pressure steam heat, but it is
pointless to pursue this goal via kinetic control. That is not to say that it is impossible
to make such a coating: It means that the problem must be solved by an approach other
than kinetic control.
Storage life can be extended by refrigeration, but users are seldom willing to incur the
expense. More reactive combinations can be used in two package coatings, in which one
package contains a resin with one of the reactive groups and the second package contains
the component with the other reactive group. Alternatively, the second package could
contain a catalyst for the reaction. The packages are mixed shortly before use. Two
package coatings are often called 2K coatings, and single package coatings are sometimes called 1K coatings. The K stands for the German word for component. Two
package coatings are used on a large scale commercially, but they pose substantial
extra problems for the user; mixing and cleaning equipment take extra time, material
is usually wasted, they are generally more expensive, and there is a chance of error in
mixing. Even 2K coatings have the analogous problem of pot life: that is, the length
of time after the two packages are mixed that the viscosity stays low enough for application. Pappas and Hill have made similar calculations of A and Ea values for the shorter
times involved [41].
There are several approaches to increasing package stability while permitting cure at
ambient or moderately elevated temperature. One or more of these approaches are
included in the discussions of many of the cross-linking reactions that are covered in
later chapters. The following list gives some of the approaches to solving this “impossible”
problem:
1. Use of a radiation-activated cross-linking reaction instead of a thermally activated
reaction.
2. Use of a cross-linking reaction requiring an atmospheric component as a catalyst or
reactant; reactions involving oxygen or water vapor in the air are examples. The same principle is involved in passing a coated article through a chamber containing a catalyst vapor.
3. Use of a volatile inhibitor that works when the coating is stored in a closed container
but volatilizes after application as a thin film, permitting the reaction to proceed. Examples
include use of a volatile antioxidant in a coating that dries by oxidation, and use of oxygen
as an inhibitor in a composition that cures anaerobically.
2.3. FILM FORMATION
33
4. Use of a cross-linking reaction that is a reversible condensation reaction
involving loss of a volatile reaction product; some of the monofunctional volatile reactant
can be used as a solvent in the coating. The equilibrium of the reaction favors the
non-cross-linked side during storage but shifts to the cross-linked side after application,
when the solvent evaporates. A similar approach can be used with blocked reactants
and catalysts in which the blocking group is volatile.
5. Use of an encapsulated reactant or catalyst, whereby the capsules are ruptured
during application. Encapsulation has been useful in adhesives but has not been very
useful in coatings because residual capsule shells interfere with appearance and/or
performance.
6. Use of a reactant that undergoes a phase change. Although there are no threshold
temperatures for kinetic reactions, phase changes can occur over narrow temperature
ranges. A crystalline blocked reactant or catalyst, insoluble in the vehicle, could give an
indefinitely stable coating; heating above the melting point permits the unblocking reaction
to occur, releasing a soluble reactant or catalyst. Over a somewhat wider temperature
range, the same idea can be used for an amorphous blocked reactant, having a Tg
about 508C above the storage temperature and perhaps 308C below the desired cure
temperature.
Another consideration in selecting components for thermosetting coatings is the potential effect of the availability of free volume on reaction rates and reaction completion.
For reactions to occur, the reacting groups must diffuse into a reaction volume to form
an activated complex that can then form a stable bond. If the diffusion rate is greater
than the reaction rate, the reaction will be kinetically controlled. If the diffusion rate is
slow compared to the kinetic reaction rate, the rate of the reaction will be mobility controlled. The major factor controlling the diffusion rate is the availability of free volume.
If the reaction is occurring at a temperature well in excess of Tg, the free volume is
large and the rate of reaction is controlled by concentrations and kinetic parameters.
If, however, the temperature is well below Tg, the free volume is so limited that the
polymer chain motions needed to bring unreacted groups close together are very slow,
and reaction virtually ceases. At intermediate temperatures, the reaction can proceed,
but the reaction rate is controlled by the rate of diffusion—that is, by the mobility of
the reactants.
Since cross-linking starts with low MW components, Tg increases as the reaction
proceeds. If the reaction temperature is well above the Tg of the fully reacted polymer,
there will be no mobility effect on the reaction rate. However, if, as is often the case in
ambient cure coatings, the initial Tg is below the ambient temperature and the Tg of the
fully reacted polymer is above the ambient temperature, the reaction rate will become
mobility controlled as cross-linking proceeds. As further reaction occurs, cross-linking
may essentially stop before reaching completion. As Tg increases toward the temperature
at which the reaction is occurring, reactions become slower. When Tg equals T, reactions
become very slow, and vitrification (glass formation) is said to occur. Unless the experiment is continued for a relatively long time, the reactions have been said to cease [42].
However, with extended reaction times, it can be seen that reactions continue slowly.
One paper [43] reports that reaction rate constants drop by about three orders of magnitude
when Tg equals T, but that the reaction continues at a slow rate until Tg increases to
34
POLYMERIZATION AND FILM FORMATION
T þ 508C [43]. It is interesting to consider the similarity of this value to the universal B
constant of 51.6 in the WLF equation (Eq. 2.9). Tg 2 B is the temperature at which
viscosity goes to infinity and free volume theoretically approaches zero, corresponding
to a Tg of T þ B. The effect of variables on mobility control of reaction rates has been
studied by Dusek and Havlicek [44]. In the reaction of bisphenol A diglycidyl ether and
1,3-propanediamine, they determined the effects of temperature, polymer – solvent interaction, and solvent volatility on reaction rates and extent of reaction. They also reviewed
the theories involved.
It seems reasonable for a formulator to assume that cross-linking reactions begin to
slow as Tg increases to about 108C below the curing temperature and get progressively
slower until Tg is about 508C above the curing temperature, where reaction essentially
ceases. The slow rates of reaction mean that properties of many ambient cure coatings
can change substantially for several weeks or even months. Caution is required, since
Tg values are dependent on the method of determination and the rate of heating
used. For this purpose, Tg values obtained at very slow rates of heating and at low
rates of application of stress are most appropriate. It has been pointed out that if
reactions are very fast, equilibrium Tg may not be reached [44]. Another factor that
may affect the development of mobility control is the size of the diffusing reactants.
Small molecules may diffuse more readily to reaction sites than functional groups on a
polymer chain. Water plasticizes coatings such as polyurethanes and epoxy – amines,
lowering their Tg.
If the initial reaction temperature is well below the Tg of the solvent-free coating, it is
possible that little or no reaction can occur after solvent evaporation and that a “dry” film
forms merely due to solvent evaporation, without much cross-linking. The result is a weak,
brittle film. One must be careful when defining what is meant by a dry film, especially
when dealing with ambient temperature cure coatings. One consideration is whether the
film is dry-to-handle. This stage could be reached with little cross-linking if the Tg of
the solvent free binder is high enough. Another consideration is whether some required
extent of cross-linking has occurred. This must be tested by some method other than
hardness, most easily by determining resistance to dissolution or by the extent of
solvent swelling (Section 4.2).
Mobility control is less likely to be encountered in baking coatings because in most
cases, the final Tg of the film is below the baking temperature. Furthermore, the Tg is
usually well above ambient temperatures, so even if there are some unreacted groups,
reactions do not continue after cooling to ambient temperature. In moderate temperature
cure powder coatings, mobility control of reaction can be a limitation, since the initial Tg
of the reactants has to be above 508C so that the powder will not sinter during storage
(Section 28.3). To achieve a high extent of reaction, the baking temperature must be
above the Tg of the fully reacted coating. Gilham and co-workers have studied factors
influencing reaction rates in high Tg epoxy – amine systems; Gilham’s work is summarized
in Ref. [45], with a focus on its applicability to powder coatings.
2.3.3. Film Formation by Coalescence of Polymer Particles
In contrast to the processes of film formation from solutions of thermoplastic or thermosetting polymers, dispersions of insoluble polymer particles form films by coalescence
(fusion) of particles. After application and loss of volatile components, the particles
2.3. FILM FORMATION
35
form a continuous film. The largest volume of coatings having such binders are latexes,
which are a dispersion of high MW polymer particles in water (Chapter 9).
For a given latex, the lowest temperature at which coalescence occurs sufficiently
to form a continuous film is called its minimum film formation temperature (MFFT);
some authors call it the minimum filming temperature (MFT). MFFT is measured by
placing samples on a heated metal bar with a temperature gradient. A major factor controlling MFFT is the Tg of the polymer in the particles. The Tg of poly(methyl methacrylate)
(PMMA) is about 1058C, and one cannot form a useful film from a PMMA latex at room
temperature; instead, one gets a layer of material that powders easily. Many latexes are
designed to have layers of material with different Tg values within each particle (Sections
9.1.3 and 9.2), making it difficult to relate MFFT directly to Tg.
The mechanism of film formation from latexes has been studied extensively, but is
still not fully understood. Film formation is divided into three overlapping steps:
1. Evaporation of water and water-soluble solvents that leads to a close packed layer
of latex particles
2. Deformation of the particles from their spherical shape that leads to a more or
less continuous but weak film
3. Coalescence a relatively slow process in which the polymer molecules interdiffuse
across the particle boundaries and entangle, strengthening the film
As water evaporates, the particles come closer together, becoming close packed.
The internal phase volume of a latex film is highly dependent on particle size distribution:
The broader the distribution, the higher the internal phase volume.
As the particles become close packed, the polymer particles are deformed. It has been
proposed that as the particles approach each other, the spaces between them act like the
equivalent of a capillary, and capillary forces apply pressure to the close packed particles,
leading to deformation. It is estimated that such forces may generate as much as 3.5 MPa
(5000 psi) where latex particles touch [46].
Tg of the latex particles is an important factor controlling the rates of particle deformation. Surfactant stabilizers can increase water absorption and also act as plasticizers
for the polymer [47,48]. In general, higher surfactant content further reduces MFFT.
Structure of the surfactant also affects MFFT; for example, nonyl phenol ethoxylates
with fewer than 9 ethoxylate units reduced MFFT further than those with 20 or 40
units. Lower Tg particles have a lower modulus—that is, they are softer (Section 4.2)
and hence are more easily deformed. For latex coatings, which are baked, the Tg is commonly above ambient temperature, and coalescence does not fully occur prior to baking.
When water is evaporated, the latex particles become a layer of close packed spheres.
Sulfate groups from the polymerization initiator can form H-bonds between spheres,
thereby raising the Tg. The MFFT of latex particles can be affected by water, which can
act as a plasticizer [48]. Reduction of Tg is largest with hydrophilic polymers. It has
been shown that MFFT can be reduced by as much as 58C by forming the film in a
humid atmosphere [49].
Deformation of the close packed particles increases the area of contact between particles and promotes coalescence. For coalescence to occur it is necessary to overcome
the stabilizing repulsion. Although capillary forces may be a factor, Croll points out
36
POLYMERIZATION AND FILM FORMATION
that the time span during which the forces are high is very short [46]. His data on drying
rates support a proposal of Kendall and Padget that a major driving force for coalescence is
surface free energy reduction [47]. The surface area of a coalesced film is only a small
fraction of the surface area of the particles, so the driving force resulting from reduction
in surface area must be large. Other workers strongly adhere to the position that capillary
pressure is the dominant driving stress for film formation [50]. Probably both forces are
involved, with capillary pressure an important factor in deformation and surface free
energy reduction for coalescence [51,52].
Consistent with the differences in surface area, small particle size latexes form films at
somewhat lower temperatures than larger particle size latexes [51]. However, it has been
shown that this is not always the case [53]. With other latexes, larger size promotes
coalescence. Still other work indicates that there is no effect of particle size on MFFT.
Since broader distribution of particle size increases close packed volume fraction, it
leads to more rapid film formation. Possibly, these different conclusions are caused not
just by particle size, but also by particle size distribution and differences in the compositions of the latexes used. It is speculated that the larger proportion of surface active
material in small particle size latexes may also play a role.
With the large number of variables involved in coalescence, one would expect interactions among the different variables. The rate of interdiffusion is driven primarily by
T 2 Tg. Adequate interdiffusion will occur only when the Tg of the particles is lower
than the temperature at which the film is to be formed. Theoretical and experimental
studies of coalescence suggest that it is only necessary for the molecules to interdiffuse
a distance comparable to the radius of gyration of one molecule to develop maximum
film strength [54]. This distance is considerably less than the diameter of a typical latex
particle. The rate of interdiffusion is directly related to Tg and is therefore controlled by
free-volume availability [54]. The major factor affecting free volume is the difference
between the temperature of film formation and the Tg of the particles. Presumably, the
Tg of the material near the surface of the original particle is most important. As a
general rule, coalescence will not occur unless the temperature is at least slightly higher
than Tg. Latexes for paints are generally copolymers of acrylic and vinyl esters that
have a Tg well below room temperature so that they can coalesce readily within the
usual range of application temperatures.
Most studies of the mechanism of film formation use carefully prepared uniform
films that are dried in a laboratory under conditions of intermediate humidity and little
air flow over the surface. Real-life conditions are quite different; the latex has been
formulated into a paint with pigments and many additives, and dried under a wide
variety of conditions, In many cases, film thickness is not uniform. A further complication
is that water evaporation is seldom uniform across the entire area of a film. Instead, a
drying front starts, perhaps at a thin spot, and spreads laterally across the panel [55].
Both water and polymer particles move sideways near the drying front, and smaller particles appear less likely to give uniform films. The practical implications of this finding are
only partly understood. If exterior house paint is applied on a hot windy day with low
humidity, film formation is very poor.
Plasticizers, which dissolve in the polymer, can be added to lower the Tg and MFFT
of the formulation. Since nonvolatile plasticizers permanently reduce Tg, most latex
paints contain volatile plasticizers, which are called coalescing solvents, although they
accelerate deformation as well as coalescence. A coalescing solvent must be soluble in
the polymer and have a low, but appreciable evaporation rate. It acts as a plasticizer to
2.3. FILM FORMATION
37
lower the MFFT, but after the film has formed, it diffuses to the surface of the film and
evaporates. Since free volume in the film is relatively small, the rate of loss of the last
of the coalescing solvent is very slow. Although the films feel dry in a short time, they
will still block for days, or even weeks, after application. Effects of coalescing solvents
on film formation have been studied quantitatively using fluorescence decay measurements to follow the extent of polymer diffusion in films as they coalesce [56]. The efficiency of coalescing solvents varies over a wide range; an example of a relatively
efficient solvent is the acetate ester of propylene glycol monobutyl ether (PnBA) [57].
The rate of evaporation from films also varies; for example, dipropylene glycol dimethyl
ether leaves a film more rapidly than does PnBA, but somewhat more is required for film
formation [57]. A widely used coalescing solvent is Texanol, the isobutyric ester of
2,2,4-trimethylpentane-1,3-diol.
Although films form rapidly from latexes when the temperature is above MFFT,
complete coalescence is a relatively slow process; in many cases, the film probably
never equilibrates to become completely uniform. The rate of equilibration is affected
by T 2 Tg. From the standpoint of rapid coalescence, it is desirable to have a latex with
a Tg well below the temperature at which the film is to be formed. There are review
papers that discuss factors affecting development of the cohesive strength of films from
latex particles [54,58]. The extent of coalescence has been studied using small-angle
neutron scattering, direct energy transfer of particles labeled with fluorescent dyes, and
scanning probe microscopy [59,60].
Usually, architectural paints are formulated so that film formation occurs at temperatures as low as 28C. Film formation at such a low temperature requires a low-Tg latex.
However, as discussed in the introductory paragraphs of Section 2.3, T 2 Tg also affects
whether the film obtained will be solid. It was estimated that for a film to withstand the
relatively mild blocking test described, T 2 Tg would have to be on the order of 218C.
If the film is to be exposed to the blocking test when the temperature is 508C (not an
unreasonable expectation during direct exposure to summer sun), the Tg should be
about 298C or higher. Thus, the paint formulator is faced with a difficult
challenge to design a system that can form a film when applied at 28C, yet resist blocking
at 508C.
Coalescing solvents help solve this problem, but environmental regulations are limiting
permissible emissions of volatile organic compounds (VOC). A second approach is to
design latex particles so that there is a gradient of Tg from a relatively high Tg in the
center of the particles to a relatively low Tg at the outer periphery [61] (Sections 9.1.3
and 9.2). The low Tg of the outer shell permits film formation at low temperature. Over
time, the Tg of the coalesced film approaches the average Tg of the total polymer. This
higher average Tg reduces the probability of blocking. Relatively small amounts
of coalescing solvents can be sufficient with such a latex. A third approach is to use
blends of high and low Tg latexes, which can reduce MFFT without the presence of
coalescing solvents [62]. It has been proposed that the films contain particles of high Tg
polymer dispersed in a matrix of lower Tg polymer, reinforcing the matrix, increasing
its modulus, and hence, decreasing blocking. There must be an adequate ratio of the
soft polymer to the hard polymer so that the soft polymer forms a continuous film enclosing the hard particles. These studies were done without pigment; it would be interesting to
compare the results with films made with a low Tg polymer with an equal volume content
of pigment, which can also reinforce a film. A fourth approach involves use of core – shell
latexes [63] (Section 9.1.3).
38
POLYMERIZATION AND FILM FORMATION
Polyurethane dispersions (PUDs) have low MFFT values in relation to dry film Tg
owing to the plasticizing effect of hydrogen bonding with water. Accordingly, coalescing
solvents, which are generally required for film formation of latexes at low temperatures,
are not needed with PUDs. When PUDs are used with high Tg acrylics, MFFT is
reduced without reducing hardness (Section 12.7.2).
Although the majority of latexes are thermoplastic polymer latexes, there are applications for thermosetting latexes as well. An excellent and extensive review has been published [64]. In many cases, two package coatings are required, but they are useful only for
industrial applications; a few stable systems have been developed that can be used for
architectural paints. (See Section 9.4 for further discussion and examples.) Cross-linked
films have better mechanical properties and increased resistance to solvents. In general,
a lower Tg latex can be used in coatings, which permits coalescence without addition of
a coalescing solvent and/or at lower film formation temperatures. Film properties
depend on the relative rate of interdiffusion between polymer particles and the rate of
cross-linking. If the rate of cross-linking is fast compared to the rate of diffusion, crosslinking within particles will occur and interfere with good film formation. If the rate of
cross-linking is slow, achievement of desirable film properties may take excessive time.
A balance between the two is needed. In some cases the functional groups on a latex
particle are on the particle surface, and an external cross-linker serves to cross-link the
particles together. In other cases the functional groups are on the same polymer and interdiffusion is required to bind the particles together.
Winnik [65] has reviewed the literature on thermosetting latexes and experimental
results related to the importance of the rate of interdiffusion and cross-linking in films
of thermosetting latexes. To develop good properties, significant interdiffusion must
occur before extensive cross-linking. The rate of interdiffusion is controlled by T 2 Tg
and the chain length of the latex. Thus, thermosetting latexes are designed with lower
MW, which both reduces Tg and shortens chain length. Straight chains diffuse more
rapidly than branched chains.
Most studies of the mechanism of film formation have been done with latexes, not
with fully formulated paints. It is to be expected that other components of paints, such
as pigments, pigment dispersing agents, and water soluble polymers used as thickening
agents, among others, affect MFFT, rate of film formation, and blocking resistance.
(See Chapter 31 for further discussion of latex paints.)
Other types of coatings that involve coalescence of particles are discussed in later
chapters, including aqueous polyurethane dispersions, organosols, water-reducible
resins, and powders.
GENERAL REFERENCES
Mark, J. E., Ed., Physical Properties of Polymers Handbook, American Institute of Physics,
Woodbury, NY, 1996.
Odian, G. W., Principles of Polymerization, 4th ed., Wiley-Interscience, New York, 2004.
Provder, T.; Urban, M. W., Film Formation in Coatings: Mechanisms, Properties, and Morphology,
American Chemical Society, Washington, DC, 2001.
REFERENCES
39
REFERENCES
1. Sijbesma, R. P.; et al., U.S. patent 6,320,018 (2001); Brunsveld, L.; et al., J. Polym. Sci. A
Polym. Chem., 1999, 37, 3657.
2. Eling, B.; Lindsay, C. I., U.S. patent appl. 20040087755.
3. Hill, L. W.; Wicks, Z. W., Jr., Prog. Org. Coat., 1982, 10, 55.
4. Elias, H. G., in Macromolecules, Plenum Press, New York, 1984, pp. 301 – 371.
5. Sullivan, C. J.; et al., J. Coat. Technol., 1990, 62(791), 37.
6. Kumanotani, J.; et al., Org. Coat. Sci. Technol., 1984, 6, 35.
7. Wicks, Z. W., Jr.; et al., J. Coat. Technol., 1986, 57(725), 51.
8. Dickinson, L.; et al., Macromolecules, 1988, 21, 338.
9. Mathias, L. J.; Colletti, R. F., Polym. Prepr., 1989, 30(1), 304.
10. Roe, R. J., Encyclopedia of Polymer Science and Technology, 2nd ed., Wiley, New York, 1987,
Vol. 7, pp. 531 – 544.
11. Andrews, R. J.; Grulke, E. A., in Polymer Handbook, 4th ed., Brandrup, J.; et al., Eds., Wiley,
New York, 1999, VI/193.
12. Lesko, P. M.; Sperry, P. R., in Emulsion Polymerization and Emulsion Polymers, Lowell, P. A.;
El-Aasser, M. S., Eds., Wiley, New York, 1997, pp. 622 –623.
13. Neumann, C.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans LA,
2004, Paper 32.
14. Gupta, M. K., J. Coat. Technol., 1995, 67(846), 53.
15. Stutz, H.; et al., J. Polym Sci. B Polym. Phys., 1990, 28, 1483.
16. Ferry, J. D., Viscoelastic Properties of Polymers, 3rd ed., Wiley, New York, 1980, p. 487.
17. Eisenberg, A., in Physical Properties of Polymers, Mark, J. E.; et al., Eds., American Chemical
Society, Washington, DC, 1984, pp. 55– 95.
18. Wicks, Z. W., Jr., J. Coat. Technol., 1986, 58(743), 23.
19. Odian, G. W., Principles of Polymerization, 4th ed., Wiley-Interscience, New York, 2004,
pp. 39– 197.
20. Darling, T. R.; et al., J. Polym. Sci. A Polym. Chem., 2000, 38, 1706.
21. Auschra, C.; et al., Prog. Org. Coat., 2002, 45, 83.
22. Perrier, S.; et al., Macromolecules, 2004, 37, 2709.
23. White, D.; et al., U.S. patent 6,462,125 (2002).
24. Auschra, C.; et al., Prog. Org. Coat., 2002, 45, 83.
25. Gaynor, S. C.; Matyjaszewski, K., in Controlled Radical Polymerization, Matyjaszewski, K.,
Ed., ACS Symp. Ser. 685, American Chemical Society, Washington, DC, 1998, pp. 396 – 417.
26. Janowicz, A. H., U.S. patent 4,694,054 (1987).
27. Chiefari, J.; et al., Macromolecules, 2005, 38, 9037.
28. Matyjaszewski, K., Ed., Controlled Radical Polymerization, ACS Symp. Ser. 685, American
Chemical Society, Washington, DC, 1998.
29. Odian, G. W., Principles of Polymerization, 4th ed., Wiley-Interscience, New York, 2004,
pp. 313 – 330.
30. De Clercq, B.; et al., Prog. Org. Coat., 2005, 53, 195.
31. Barkac, K.A.; et al., U.S. patent 6,670,043 (2003).
32. Visscher, K. B.; McIntyre, P. F., U.S. patent 6,599,973 (2003).
33. htpp://automotive.dupont.com/en/newsEvents/article20030305b.html.
40
POLYMERIZATION AND FILM FORMATION
34. Callais, P.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 197 –210.
35. Flory, P. J., J. Am. Chem. Soc., 1939, 61, 3334.
36. Chang, W. L.; Karalis, W. L., J. Polym. Sci. A Polym. Chem., 1993, 31, 493.
37. Odian, G. W., Principles of Polymerization, 4th ed., Wiley-Interscience, New York, 2004,
pp. 40–62.
38. Burrell, H., Off. Dig., 1962, 34(445), 131.
39. Mayer, W. P.; Kaufman, L. G., XVII FATIPEC Congress Book, 1984, Vol. I, p. 110.
40. Pappas, S. P.; Feng, H.-B., Int. Conf. Org. Coat. Sci. Technol., Athens, Greece, 1984,
pp. 216 –228.
41. Pappas, S. P.; Hill, L. W., J. Coat. Technol., 1981, 53(675), 43.
42. Aronhime, M. T.; Gilham, J. K., J. Coat. Technol., 1984, 56(718), 35.
43. Blair, H. E., Polym. Prepr., 1985, 26(1), 10.
44. Dusek, K.; Havlicek, I., Prog. Org. Coat., 1993, 22, 145.
45. Simon, S. L.; Gilham, J. K., J. Coat. Technol., 1993, 65(823), 57.
46. Croll, S. G., J. Coat. Technol., 1987, 58(734), 41.
47. Kendall, K.; Padget, J. C., Int. J. Adhesion Adhesives, 1982, 2(3), 149.
48. Vandezande, G. A.; Rudin, A., J. Coat. Technol., 1996, 68(860), 63.
49. Sheetz, D. P., J. Appl. Polym. Sci., 1965, 9, 3759.
50. Lin, F.; Meier, D. J., Prog. Org. Coat., 1996, 29, 139.
51. Eckersley, S. T.; Rudin, A., J. Coat. Technol., 1990, 62(780), 89.
52. Routh, A. F.; Russel, W. B., Ind. Eng. Chem. Res., 2001, 40, 4302.
53. Kan, C. S., J. Coat. Technol., 1999, 71(896), 89.
54. Winnik, M. A., in Emulsion Polymerization and Emulsion Polymers, Lovell, P. A.; El-Aasser,
M. S., Eds., Wiley, New York, 1997, pp. 467 – 518.
55. Salamanca, J. M.; et al., Langmuir, 2001, 17, 3202.
56. Winnik, M. A.; et al., J. Coat. Technol., 1992, 64(811), 51.
57. Geel, C., J. Oil Colour Chem. Assoc., 1993, 76, 76.
58. Daniels, E. S.; Klein, A., Prog. Org. Coat., 1991, 19, 359.
59. Rynders, R. M.; et al., J. Coat. Technol., 1995, 67(845), 59.
60. Butt, H.-J.; Kuropka, R., J. Coat. Technol., 1995, 67(848), 101.
61. Hoy, K. L., J. Coat. Technol., 1979, 51(651), 27. Bassett, D. R.; Hoy K. L., in Bassett, D. R.;
Hamielec, A. E., Eds., ACS Symp. Ser. 165, American Chemical Society, Washington, DC,
1981, p. 371.
62. Winnik, M. A.; Feng, J., J. Coat. Technol., 1996, 68(852), 39.
63. Juhue, D.; Lang, J., Macromolecules, 1995, 28, 1306.
64. Taylor, J. W.; Winnik, M. A., JCT Res., 2004, 1, 3.
65. Winnik, M. A., J. Coat. Technol., 2002, 74(925), 49.
3
Flow
Rheology is the science of flow and deformation. In this chapter we deal only with the
flow of relatively liquid materials; deformation of relatively solid materials are discussed
in Chapter 4. Flow properties of coatings are critical for the proper application and appearance of films. For example, in brush application of a paint, the flow properties govern
settling of pigment during storage, how much paint is picked up on the brush, film thickness applied, leveling of the applied film, and control of sagging of the film. Depending on
how stress is applied to a fluid, there are several types of flow. Of major importance in
coatings is flow under shear stress. We consider shear flow first and then, more briefly,
other types of flow.
3.1. SHEAR FLOW
To understand and define shear flow, consider the model shown in Figure 3.1. The lower
plate is stationary, and the upper parallel plate is movable. The plates are separated by a
layer of liquid of thickness x. Force F is applied to the top movable plate of area A, so the
plate slides sidewise with velocity v. The model assumes that there is no slip at the interfaces and that there is no fluid inertia. When the plate moves, the liquid near the top moves
with a velocity approaching that of the movable plate, and the velocity of the liquid near
the bottom approaches zero. The velocity gradient dv/dx for any section of the liquid is
constant and, therefore, equals v/x. This ratio is defined as shear rate ġ. The units of
shear rate are reciprocal seconds, s21.
g_ ¼
dv v
¼ ;
dx x
units: cm s1 =cm ¼ s1
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
41
42
FLOW
Figure 3.1. Model of shear flow of an ideal liquid. (In current usage, the symbol for shear stress is s and
its units are Pa, the symbol for shear rate is ġ and the units of h are Pa.s) (From Ref. [1], with
permission.)
Force F acting on top plate of area A results in shear stress, s. The units of shear stress
are pascal (Pa).
s¼
F
;
A
units: m kg s2 m2 ¼ N m2 ¼ Pa
A liquid exerts a resistance to flow called viscosity, h, defined as the ratio of shear stress
to shear rate. This type of viscosity is correctly called simple shear viscosity, but since it is
the most widely encountered type of viscosity, it is usually just called viscosity. The separation of molecules in flow dissipates energy, primarily as heat. Thus, viscosity is the
energy per unit volume dissipated to attain a unit velocity gradient. The units are pascal
seconds (Pa.s). The older and still commonly used unit is the poise (P). One Pa.s equals
10 P, and 1 mPa.s equals 1 cP.
h¼
s
;
ġ
units: Pa=s1 ¼ Pa s
˙
When a liquid flows through a hole or a capillary, part of the energy is diverted into
kinetic energy; then the resistance to shear flow is called kinematic viscosity, n, with
units of m2 s21, formerly stokes, where 1 m2 s21 ¼ 104 stokes. When the acceleration
results from gravity, kinematic viscosity equals simple shear viscosity divided by the
density, r, of the liquid:
h
n¼ ;
r
units: m2 s1
3.2. TYPES OF SHEAR FLOW
When the ratio of shear stress to shear rate is constant, liquids are Newtonian and viscosity
is independent of shear rate (or shear stress). A plot of shear rate as a function of shear
stress is linear [Figure 3.2(a)], and the slope equals the inverse of viscosity. Such plots
3.2. TYPES OF SHEAR FLOW
43
Figure 3.2. Plots of flow of various types of liquids.
also appear in the literature with axes opposite to those shown; then the slope is the viscosity. Newtonian flow is exhibited by liquids composed of miscible small molecules.
Many solutions of oligomeric resins also approximate Newtonian flow.
Many liquids are non-Newtonian; that is, the ratio of shear stress to shear rate is not
constant. One class of non-Newtonian liquids exhibits decreasing viscosity as shear
rate (or shear stress) increases; these liquids are shear thinning. As stress is applied, the
molecules line up parallel to the direction of flow, reducing the energy required to displace
them further. In the case of liquids with a disperse phase, the particles line up as bead
chains again reducing the energy required to displace them further. Shear thinning has
been called shear-induced order. When the molecules or particles are all aligned in the
direction of flow the viscosity is again independent of shear rate and the curve becomes
linear [Figure 3.2(b)].
In some cases, no detectable flow occurs unless a minimum shear stress is exceeded.
Such materials exhibit plastic flow and are sometimes called Bingham bodies. The
minimum shear stress required is called the yield value, or yield stress, and is designated
by the symbol s0.
s s0 ¼ hp g_
A schematic flow diagram is shown in Figure 3.2(c). The yield value depends
strongly on the rate at which stress is increased. The faster the rate of stress increase
the higher the measured value that will be obtained. It is fairly common to extrapolate
the linear part of the curve to the intercept with the shear stress axis and call the intercept a yield value.
Another class of liquids exhibits increasing viscosity as shear rate (or shear stress)
increases. A schematic plot of such behavior is given in Figure 3.2(d ). Such liquids
are shear thickening. Shear thickening has been called shear-induced disorder. If
shear thickening liquids also increase in volume under shear, they are called dilatant
fluids. Dilatant fluids have dispersed phases that become less ordered and hence
occupy more volume. Examples are pigment and resin dispersions in which the
dispersed phase is sufficiently concentrated so that the particles approach being
randomly close packed. If sufficient stress is applied for flow, microscopic voids are
created, increasing the volume and increasing the energy necessary to induce flow.
Thus, the viscosity increases. Another example is quicksand. See Ref. [2] for a discussion
of factors affecting dilatant flow.
44
FLOW
Figure 3.3. Casson plot of viscosity as a function of shear rate, showing the dependence t0 with
constant h and h1. (From Ref. [1], with permission.)
The Casson equation linearizes the viscosity/shear rate data of shear thinning or
thickening fluids:
The slope of the line is the yield stress, and extrapolation gives the viscosity at infinite
shear rate h1.
hn ¼ hn1 þ
s0a
g_
(3:1)
In many cases, the value of n is 0.5, and commonly, the Casson equation is shown with
just the half-power relationship. In plots of log viscosity against shear rate, the degree of
curvature is related to the value of s0. In Figure 3.3 the values of h and h1 are held
constant to show the effect of changes in t0 on flow response [3]. For a Newtonian
fluid, t0 equals zero and the plot is a straight line parallel to the shear rate axis.
Some fluids show time or shear history dependence of viscosity, as illustrated in
Figure 3.4(a). The curves in the figure result from shear stress readings taken at successively higher shear rates to some upper limit (right-hand curve), followed immediately
by shear stress readings taken at successively lower shear rates (left-hand curve). At
any shear rate on the initial curve, the stress would decrease with time to an equilibrium
value between the two curves; that is, the viscosity would decrease. On the other hand, if
such a system had been exposed to a high rate of shear and then the shear rate decreased,
the shear stress would increase to an equilibrium value as the measurement was continued;
that is, the viscosity would increase with time. This behavior is called thixotropic flow.
Thixotropic fluids are shear thinning fluids; their viscosity is also dependent on time
and prior shear history. All thixotropic fluids are shear thinning fluids, but not all shear
thinning systems are thixotropic. Unfortunately, the term thixotropy is often improperly
used as a synonym for shear thinning.
Thixotropy usually results from reversible formation of a structure within a fluid; an
example is association of dispersed particles held together by weak forces. Thixotropic structure is broken down by applying shear for a sufficient time, and the structure reforms over
3.2. TYPES OF SHEAR FLOW
45
Figure 3.4. Plots of systems exhibiting thixotropic flow. (a) The curve to the right is based on readings
taken as the shear rate was being increased, and the curve to the left is based on readings taken as the
shear rate was then being decreased. (b) The viscosity drops as shear continues, then increases as the
shear rate is decreased.
time when shear is stopped. Some thixotropic fluids undergo viscosity reduction to equilibrium values in short time periods and recover their viscosity rapidly when shearing is
stopped; others change more slowly with time. In early work, areas within hysteresis
loops were compared, as shown in Figure 3.4(a). However, the areas of such loops
depend on the shear history, the peak shear, and the time intervals between successive
measurements. Another way to represent the effect is to plot the viscosity at a series of
shear rates as a function of time, as illustrated in Figure 3.4(b). The time for recovery
after shifting from a high shear rate to a lower shear rate is a useful method for comparing
different paints. In many coatings, thixotropic flow is desirable since it controls sagging,
leveling, and pigment settling. It is desirable to formulate a coating so that the rate of recovery is rapid enough to avoid sagging but slow enough to permit leveling.
Additives to impart thixotropy to coatings have been developed. Clays treated with
quaternary ammonium compounds, which form salts with the clay platelets and make
them organophilic, are widely used in solventborne coatings. Attapulgite clay is used in
both solventborne and waterborne coatings. The needles of this clay associate with each
other, providing an increase in viscosity that reverses with agitation. In waterborne coatings, the clay also acts by absorbing water to make swollen particles that distort in shape
with stress. Fine particle size silica has been used for many years. Castor oil derivatives
and powdered polyethylene are effective thixotropic agents in baking coatings.
Polyamides such as alkyds made with aminoalcohols are used as thixotropes especially
in Europe. Basic calcium sulfonate derivatives are liquid thickeners that do not reduce
gloss and retain their effectiveness when a coating is heated. They are particularly
useful in moisture-curable urethane coatings since their basicity neutralizes the CO2 produced by reaction with water hence reducing the bubbling problem as well as imparting
thixotropy. In latex-paints, hydroxyethyl cellulose is widely used. Reference [4] provides
a review of thixotropes and additives for modifying paint flow.
Rheologists frequently discuss the properties of such systems in terms of viscoelasticity; that is, their flow is a combination of viscous flow and elastic deformation. (Viscoelasticity is discussed in Section 4.2.) Such interpretations are valid and useful, but they
have seldom been applied in the coatings industry. Time dependency can best be measured
in terms of viscoelasticity (see Refs. [3] and [5] for such analyses). Increasing use of viscoelastic flow analysis can be expected in the future.
Another way to show the effect of shear on a thixotropic fluid is by a different type of
Casson plot, as shown in Figure 3.5. The square root of the viscosity is plotted against the
square root of the reciprocal of the shear rate; the steeper the slope, the greater the degree
46
FLOW
Figure 3.5. Casson plots of a sheared and an unsheared thixotropic coating. The degree of divergence
gives an estimate of the degree of thixotropy. (From Ref. [6], with permission.)
of shear thinning. If the sample had been sheared until all of the thixotropic structure was
broken down, and if the measurements could be made before any structure buildup
occurred, the plot would be linear and parallel to the x axis. Although comparisons
of the differences between the slopes of such lines give a qualitative expression of
the extent of thixotropy, the slopes of the curves are dependent on prior shear history,
the rate of acceleration of shear, and the length of time that the sample was exposed
to the highest shear rate.
3.3. DETERMINATION OF SHEAR VISCOSITY
A variety of instruments is available to determine viscosity. They vary in cost, time
required for measurements, operator skill required, susceptibility to abuse, precision,
accuracy, and ability to measure shear rate variability or time dependency effects. Data
obtained on the same sample with different instruments and by different operators with
the same instruments can vary substantially, especially for shear thinning liquids at low
rates of shear [7]. Some of the variation can result from lack of attention to details,
especially temperature control and possible solvent loss; major errors can result from
comparing samples that have had different shear histories.
Temperature must be carefully controlled. Since viscosity depends strongly on temperature, it is critical that the sample has reached a constant, known temperature before
measurement. When high viscosity fluids are sheared at high shear rates, heat is
evolved and the temperature of the sample increases unless the heat exchange
efficiency of the viscometer is adequate. If viscosities are determined as both shear rate
and temperature are increasing, it may be impossible to tell whether or not a fluid is
shear thinning.
Viscometers can be divided into three broad classes: (1) those that permit quite accurate
viscosity determinations, (2) those that permit determination of reasonable approximations of viscosity, and (3) those that provide flow data marginally related to viscosity.
Space limitations restrict our discussion to major examples of each class. The general
references at the end of this chapter are good sources of further information.
3.3. DETERMINATION OF SHEAR VISCOSITY
47
3.3.1. Capillary Viscometers
Figure 3.6 shows an example of a capillary viscometer [8]. The time required for a known
amount of liquid to flow through a capillary tube is measured. Although viscosities can be
calculated based on the diameter of the capillary, usually each instrument is standardized
with liquids of known viscosity; then, calculation is based simply on instrument constants
and time.
Flow is usually driven by gravity, so kinematic viscosity is measured. Simple
shear viscosity can be calculated by correcting for density. Capillary viscometers with a
range of diameters permit the determination of viscosities from 1027 to 1021 m2 s21.
For liquids with a density of 1, these values correspond to a range of 1 mPa.s to 1000 Pa.s.
Capillary viscometers are the instruments of choice for research work, since the accuracy attained is high. They are applicable only to Newtonian fluids and are not appropriate
for routine work, because determinations are relatively time consuming especially when
temperature dependence of viscosity data is desired. Temperature equilibration is slow
because of the relatively large sample sizes and the low rate of heat transfer by glass.
Capillary viscometers are particularly appropriate for use in determining the viscosity
of volatile liquids or solutions containing volatile solvents, since they are essentially
closed systems.
Various modified capillary viscometers are available that overcome limitations of the
standard types. Vacuum viscometers are available in which the sample is sucked up into
the tube. The time to pass from the lower marker to the upper one is measured. Since the
flow rate is not affected by gravity, density does not affect the time, and viscosity is
Figure 3.6. Ostwald capillary viscometer. (From Ref. [8], with permission.)
48
FLOW
measured in Pa.s, avoiding the need to determine density. Similar viscometers can be used
for opaque liquids since the passage of liquid from the lower mark to the upper mark can
easily be seen.
3.3.2. Rheometers
For non-Newtonian liquids, including pigmented liquids, the highest accuracy over a wide
range of shear rates is obtained with rotational rheometers, exemplified by cone and plate
rheometers. A schematic diagram is shown in Figure 3.7. The sample is placed on the
plate, which is then raised to a level with a small clearance from the cone. The cone
can be rotated at any desired number of revolutions per minute (rpm) and the torque is
measured. The angle of the cone is designed to make the shear rate constant throughout
the sample in the gap. Shear rate is proportional to the rpm, and shear stress is related
to the torque. Temperature is controlled by passing temperature-controlled water
through the plate; temperature control problems are minimized by the small sample size.
A variety of rheometers are available, varying in the range of shear rates that can be
used and the time required to increase or decrease shear rate. (See the general references
and Refs. [6] and [8] for further discussion.) The least expensive instruments are
sufficiently rugged, simple to use, and fast enough for quality control applications. The
most versatile ones are sensitive scientific instruments requiring skill in use and are
most appropriate for research applications. In some instruments, the edge of the liquid
sample is exposed to the atmosphere, and volatiles can evaporate. When rheometers are
used with solutions containing volatile solvents, the cone and plate unit should be
shrouded in an atmosphere saturated with solvent vapor. Two types are available:
Figure 3.7. Cone and plate viscometer geometry. (From Ref. [1], with permission.)
3.3. DETERMINATION OF SHEAR VISCOSITY
49
controlled strain and controlled stress instruments. The latter type offers advantages for
coatings, as it is generally superior for measurements at very low shear rates.
The viscosity of highly viscous materials can be determined at high rates of shear by the
use of mixing rheometers, which are small, heavy-duty mixers. The test sample is confined
to a relatively small space and subjected to intense mixing by dual rotors in the form of
sigma-shaped blades. A dynamometer measures the work input through a reaction
torque that is converted to a strip chart readout. The speed is set by a tachometer. The
newer instruments are computerized. These instruments were originally designed for
studying the molding of plastics, but are also used in studying the effect of pigments on
viscosity. Heat buildup can be substantial with high viscosity fluids. A micro twin
screw rheometer is available in which the rpm of the screws can be varied over a wide
range, permitting measurement of shear stress as a function of rpm.
3.3.3. Rotating Disk Viscometers
Rotating disk viscometers, such as a Brookfield Viscometer, have a motor that rotates a
disk immersed in a liquid over a range of rpm, and the resulting torque is measured. A
schematic diagram is shown in Figure 3.8. The instruments must be calibrated with
standards. Measurements should be made in a container of the same dimensions as that
in which the standardization is carried out, since the distance of the disk below the
surface of the liquid, above the bottom of the container, and from the side walls can
affect the response. In reporting results, the rpm setting should always be specified. The
instruments are sturdy and relatively inexpensive, although they must be recalibrated
periodically. When used properly, they provide relatively accurate viscosity measurements for Newtonian liquids. For non-Newtonian liquids, the viscosity reading represents
an average response corresponding to the viscosities resulting from a span of shear
stresses.
Rotating disk viscometers can detect whether a liquid is shear thinning or shear
thickening by measurements carried out at different rpm settings. They can detect thixotropy by a change in response over time at the same rpm setting. Useful comparisons of the
Figure 3.8. Disk viscometer. (From Ref. [1], with permission.)
50
FLOW
Figure 3.9. Determination of viscosity with a bubble tube. (From Ref. [1], with permission.)
thixotropic properties can be obtained by applying a high shear rate, then suddenly
reducing the rate of shear to a lower value and measuring the time required for the
viscosity to reach equilibrium at that lower rate.
3.3.4. Bubble Viscometers
Bubble viscometers are widely used in quality control to approximate the viscosity of
solutions of resins (Section 15.6.2). The measurement is based on the rate of rise of an
air bubble in a tube of liquid; the higher the viscosity, the slower the bubble rises. A
glass tube is filled with a liquid to a graduation mark and stoppered so that a definite
amount of air is enclosed at the top, as shown in Figure 3.9. The tube is placed in a thermostatic bath and kept there long enough for the temperature to equilibrate. Equilibration is
slow, but is essential if meaningful measurements are to be made. The tube is then
inverted, and the time required for the air bubble to travel between two calibration
marks on the tube is measured. Provided that the length of the bubble is greater than its
diameter, the rate of rise is independent of the bubble size. The density of the liquid
affects the rate of rise of the bubble, so kinematic viscosity is measured. Sets of standard
tubes are designated as A, B, C, and so on; after Z, tubes are designated as Z1, Z2, and so
on. Kinematic viscosities range from about 1025 to 0.1 m2 s21. Bubble viscometers are
appropriate only for Newtonian, transparent fluids. They are low in cost and simple to
use. If the bubble has a tail, the resin is approaching gelation.
3.3.5. Efflux Cups
The most widely used control device for measuring flow of industrial coatings, especially
for spray application, is the efflux cup. A variety of efflux cups are used. (Refs. [6] and [8]
provide a comparison of about two dozen that are used commercially.) A schematic
diagram of one of the most common efflux cups, the Ford No. 4 cup, is shown in
3.3. DETERMINATION OF SHEAR VISCOSITY
51
Figure 3.10. Ford No. 4 efflux cup. (From Ref. [1], with permission.)
Figure 3.10. One holds a thumb over the hole in the bottom of the cup, fills the cup with
coating, removes the thumb, and then times when the stream of coating flowing through
the hole breaks. The result is expressed in seconds. The data should not be converted
into kinematic viscosity numbers, since a significant amount of the force is converted
into kinetic energy, especially with low viscosity coatings. The method is not appropriate
for non-Newtonian liquids, although efflux cups are frequently used for coatings that
exhibit a small degree of shear thinning. Despite their limitations, efflux cups are useful
quality control devices. They are low in cost, rugged in construction, and easily
cleaned. Results are simply and quickly obtained, but reproducibility is poor, reportedly
only in the range 18 to 20% [8].
The proper way to use an efflux cup to control viscosity for spraying, for example, is to
adjust the viscosity of the coating by solvent addition until the coating sprays properly, and
then measure the time it takes to flow through the efflux cup. This time can then be used as
the standard for spraying that particular coating through that spray gun at that distance
from the object being sprayed. Proper efflux cup flow times for spraying vary with
different coatings for a given application system and with different application systems
for a given coating.
3.3.6. Paddle Viscometers
In the United States, the most widely used viscometer in architectural paint formulation is
the Stormer Viscometer. The instrument paddle is immersed in the paint and rotated at 200
rpm. The force required to maintain this rotation rate is measured by adding weights to a
platform at the end of a cord over a pulley connected to the paddle by a geartrain.
A schematic diagram is shown in Figure 3.11. The weight loaded is converted into
Krebs units (KU) by an arbitrary conversion scale. Originally, a KU value of 100 was
supposed to correspond to a good brushing consistency. In current practice, paints are
usually formulated with somewhat lower KU values. The instrument is of little utility
with Newtonian fluids, and the readings have no real meaning for non-Newtonian
52
FLOW
Figure 3.11. Paddle viscometer. (From Ref. [1], with permission.)
fluids, typical of most architectural paints. Even when used only for quality control, a
paddle viscometer is not very satisfactory.
The proper way to establish the flow properties for a trade sales paint is to apply it
with a brush or roller and adjust it until the best compromise of such properties as ease
of brushing, leveling, sagging, settling, and so on, has been achieved. Having done this
work, a quality control test can be set up. Appropriate quality control instruments are
cone and plate viscometers (some of which are low cost) or, lacking that, rotating disk
viscometers. Stormer Viscometers are widely used, but that does not make them satisfactory. The director of research of a large manufacturer of trade sales paints in the
United States said some years ago that the Stormer Viscometer was responsible for
setting back the formulation of one-coat hiding latex paints by 20 years.
3.4. SHEAR VISCOSITY OF RESIN SOLUTIONS
The viscosity of liquids depends on free volume availability. To put it somewhat simplistically, there are free volume holes opening and closing rapidly in a liquid; molecules
move randomly through these free volume holes. When a stress is applied, movements
in the direction that relieves the stress are favored, and the liquid flows. Therefore,
factors that control viscosity of resin solutions are those that control the availability of
free volume. Many coatings are based on solutions of polymers or oligomers. The variables affecting the flow behavior of these concentrated solutions are not fully understood.
The variables that govern the flow of very dilute polymer solutions have been studied
extensively and are better understood. Factors affecting the flow of dilute solutions are
discussed in Section 3.4.2, and factors affecting the flow of more concentrated solutions
are discussed in Section 3.4.3.
3.4. SHEAR VISCOSITY OF RESIN SOLUTIONS
53
3.4.1. Temperature Dependence of Viscosity
Temperature dependence of viscosity for a range of low molecular weight (MW) resins
and their solutions has been shown to fit a Williams– Landel – Ferry (WLF) equation
[9– 12] (Section 2.3): where Tr, the reference temperature, is the lowest temperature for
which experimental data are available and hr is the viscosity at the reference temperature,
Tg. Except for very dilute solutions, data fit Eq. 3.2 when the reference temperature is Tg
and the viscosity at Tg is assumed to be 1012 Pa.s [9].
ln h ¼ ln hr
c1 ðT Tr Þ
AðT Tg Þ
ffi 27:6
c2 þ ðT Tr Þ
B þ ðT Tg Þ
(3:2)
For higher MW polymers at temperatures above Tg þ 1008C, the temperature dependence of viscosity approximately fits an Arrhenius equation, shown in Eq. 3.3, where
Ev is the activation energy for viscous flow. (Note that the Arrhenius A value is not the
same as the constant A in Eq. 3.2.)
ln h ¼ K þ
B
Ev
¼ ln A þ
RT
T
(3:3)
Using the data available for low MW resins and their solutions, Arrhenius plots of ln h
as a function of 1/T have been found to be curved and not linear [9,11]. On the other hand,
the data do fit a WLF equation. From a practical viewpoint, the differences in the models
are small if the temperature range is small. However, over a wide range of temperatures,
the differences are relatively large. Figure 3.12 shows plots of the temperature dependence
of viscosity of commercial standard liquid bisphenol A epoxy resin (Section 11.1.1), calculated from both Arrhenius and WLF equations, together with experimental data points;
the data fit the WLF equation.
A major factor controlling the viscosity of resin solutions is T 2 Tg, but it is not the only
factor. When the differences between Tgs are small, differences in WLF constants A and B
may overshadow the small difference in T 2 Tg. Constant A depends on the difference in
thermal expansion coefficients above and below Tg, but no studies have been reported on
the structural factors that control these coefficients. Constant B is the value of Tg 2 T at
which viscosity goes to infinity. The universal value of this constant is 51.68C, but the
“constant” varies considerably with composition. No studies have been reported on the
relationship between structure and the value of constant B.
Generally, in designing resins, it is reasonable to predict that a lower Tg will lead to a
lower viscosity of the resins and their solutions. (See Section 2.1.2 for a discussion of the
factors controlling the Tg of polymers.) Linear poly(dimethylsiloxanes) have low Tgs
and low viscosities. Linear polyethylene glycols have almost as low Tgs and viscosities.
Poly(methyl methacrylate) resin solutions have higher Tg values and viscosities than
those of comparable poly(methyl acrylate) resin solutions. BPA epoxy resins have
higher Tg values and viscosities than those of corresponding hydrogenated derivatives.
Exceptions to this generalization on the effect of Tg have been reported for some
high-solids acrylic resins (Section 8.2.1). Acrylic resins made with a comonomer that
has a bulky group, such as 3,3,5-trimethylcyclohexyl methacrylate [13] or isobornyl
methacrylate [14], are reported to have low viscosities at high solids even though their
homopolymers have high Tg values; no explanation of this effect has been advanced.
54
FLOW
Figure 3.12. Viscosity dependence of standard liquid BPA epoxy resin on temperature. (From Ref. [9],
with permission.)
3.4.2. Dilute Polymer Solution Viscosity
Measurement of viscosities of very dilute solutions of polymers played an important
part in the early development of polymer science, since they were the earliest method
of estimating MWs. If the concentration of a solution is low enough that individual
polymer molecules and their associated solvent molecules are isolated from each other,
the relative viscosity (hr) of the solution can be expressed by the empirical Huggins
equation where [h] is intrinsic viscosity (sometimes called the limiting viscosity
number) and C is the concentration of the polymer solution, expressed as weight of
polymer per unit volume of polymer solution. Relative viscosity is the ratio (it is unitless)
of solution viscosity to viscosity of the solvent. In recent literature, the units of C most
commonly used are grams of polymer per milliliter of solution, but in older literature, g
dL21 is common, so one must be careful to check units. When C is expressed in g
mL21, the units of intrinsic viscosity are mL g21. Intrinsic viscosity is obtained by extrapolating a plot of ln hr/C as a function of C to zero concentration. It is related to the hydrodynamic volume of the sphere swept out by an isolated polymer molecule and its
associated solvent as it moves through a dilute solution. Further information on the
Huggins equation and a table with data covering [h] for a large number of polymers
with various solvents and temperatures may be found in Ref. [15].
ln½hr ¼ ½hC þ ½h2 C2
(3:4)
3.4. SHEAR VISCOSITY OF RESIN SOLUTIONS
55
Intrinsic viscosity depends on temperature. As temperature increases, the coil size of
the polymer molecule usually increases, more solvent is entrapped, and intrinsic viscosity
increases. Intrinsic viscosity is also affected by solvent –polymer interactions. The greater
the extent of solvent association with a polymer molecule, the more the coil expands, and
therefore, the higher the intrinsic viscosity. Intrinsic viscosity depends on MW: The higher
the MW, the larger the intrinsic viscosity. Another factor is the rigidity of the chain.
Everything else being equal, polymers with flexible, randomly kinked chains have
lower intrinsic viscosities than those with rigid, rod-like structures. The relationship
between intrinsic viscosity and molecular weight M is expressed by the Kuhn – Mark –
Houwink – Sakurada equation:
½h ¼ KM a
(3:5)
in which K and a are constants.
If the solvent is too poor or the temperature is too low, the polymer molecules
precipitate rather than staying in solution. The combination of minimum temperature
and poorest solvent that just maintains solubility is called the theta condition (u). Under
theta conditions, intrinsic viscosity [h]u is at a minimum. If the chains are flexible, as
for example with acrylic polymers, [h]u is proportional to the half-power of MW, as
shown in Eq. 3.6. Note that Eqs. 3.5 and 3.6 are based on polymers that have narrow
MW distributions.
½hu ¼ Ku M 1=2
(3:6)
In better solvents the isolated polymer coils expand, intrinsic viscosity increases, and
the exponent a in Eq. 3.5 increases to as high as 0.78 for flexible polymers, and even
higher for rigid polymers. In THF, branched poly(methyl methacrylate) (PMMA) has
been shown to have an intrinsic viscosity of 0.40 compared to 0.688 for linear PMMA
[16]. The lower value of a is attributed to the lower hydrodynamic volume of the branched
polymer solvent swollen molecules.
3.4.3. Concentrated Polymer Solution Viscosity
Concentrated solutions of polymers behave quite differently than dilute solutions. There
have been relatively few fundamental studies of the factors controlling the viscosity of
more concentrated solutions of polymers and resins, such as those used in the coatings
field. Several empirical relationships have been proposed, such as the concentration
dependence of relative viscosity: In which wr is the weight fraction resin and the k
values are constants.
ln hr ¼
wr
k1 k2 wr þ k3 w2r
(3:7)
Nonlinear regression analysis of the limited number of sets of data that were available
in the literature in 1985 fits Eq. 3.7 over a wide range of concentrations [9]. Even with this
many constants, there is some systematic deviation from the model at very low concentrations. Constant k1 is the reciprocal of weight intrinsic viscosity, [h]w which although
formally unitless, is based on the number of grams of solution containing a gram of
56
FLOW
resin. Weight intrinsic viscosity can be converted to the more familiar volume intrinsic
viscosity [h] by dividing by the density of the solution at the concentration wr ¼ k1. No
physical significance of the other two constants, k2 and k3, has been elucidated; they are
presumably related to further solvent – resin interactions and to free volume availability.
Over narrower ranges of concentration, the experimental data reasonably fit the simpler
Eq. 3.8. The even simpler Eq. 3.9 has been used extensively to calculate approximate
relative viscosities over a narrow range of concentrations with viscosities from around
0.01 to 10 Pa.s.
wr
k1 k2 wr
wr
¼ ½hw wr
ln hr ¼
k1
ln hr ¼
(3:8)
(3:9)
As explained in Section 3.4.2, relative viscosity of dilute solutions of polymers
increases as the solvent gets “better.” However, in concentrated solutions, relative viscosity is higher in poor solvents than it is in good solvents. In the few cases reported in
the literature, log of relative viscosity increases with the square root of MW of resins
dissolved in good solvents at these higher concentrations [9,17]. It has been postulated
that intrinsic viscosities of oligomers exhibit theta condition response; that is, the relative
viscosity of oligomer solutions is proportional to the square root of MW [18]: as shown in
Eq. 3.10. This relationship appears to be true for solutions in good solvents of resins
having relatively narrow MW distributions and viscosities between about 0.01 and 10
Pa.s; but further research is needed.
ln hr ¼ Kwr M 1=2
or
ln h ¼ ln hs þ Kwr M 1=2
(3:10)
As shown in Eqs. 3.7 to 3.10, the viscosity of the solvent is a factor affecting the
viscosity of resin solutions. At first glance it might appear that a small difference in the
viscosity of the solvent would have a trivial effect on the much higher viscosity of
the resin solution. However, there are examples in which solvents differ in viscosity by
as little as 0.2 mPa.s, whereas 50 wt% resin solutions in those same solvents differ in
viscosity by as much as 2 Pa.s.
Also important are the effects of resin –solvent interactions. In good solvents, there
are stronger interactions between solvent molecules and resin molecules than in poor solvents. In very dilute solutions, this means that the chains become more extended and
sweep out larger hydrodynamic volumes in good solvents than in poor solvents.
However, in more concentrated solutions, resin molecules are constrained by the hydrodynamic volumes swept out by neighboring resin molecules. If interaction between
solvent and resin is stronger than resin –resin interactions, the molecules can flow easily
through the hydrodynamic volumes swept out by other molecules (provided that the
free volume is adequate). On the other hand, if resin – resin interactions are stronger,
more or less transient clusters of resin molecules form, and the viscosity is higher. In solutions containing good solvents, flow is generally Newtonian. In many cases, flow of more
concentrated resin “solutions” in poor solvents behave somewhat like dispersed systems;
they are non-Newtonian because shear can break up or distort resin clusters.
3.4. SHEAR VISCOSITY OF RESIN SOLUTIONS
57
Although the difference in the viscosity of resin solutions in good and poor solvents
is reasonably well understood, there is little definitive work in the literature on
comparisons between solutions in various good solvents in which some of the solvents
are “more good” than others. Erickson studied relative viscosities of solutions of several
low MW resins in a range of solvents [19]. He concluded that relative viscosities decrease
as one changes from a very good solvent to a good solvent, pass through a minimum, and
then increase rapidly in very poor solvents. As can be seen in Eqs. 3.7 to 3.9, which relate
relative viscosity to concentration, the hydrodynamic volume of the isolated resin molecule
and its associated solvent molecules is a factor in determining the viscosity not just of very
dilute solutions, but also of more concentrated ones. In changing from a very good to a good
solvent, the equations predict that intrinsic viscosity, and hence relative viscosity, should
decrease; this prediction conforms with Erickson’s hypothesis. The range of error in Erickson’s work is not small enough to establish his conclusions beyond doubt. He may well be
right, but there is need for further research.
Solvent effects on hydrogen bonding between resin molecules can be substantial.
Figure 3.13 shows the viscosities of solutions of an acrylated epoxidized linseed oil in
three solvents chosen because of their similar viscosities but very different hydrogenbonding characteristics [20]. The resin molecules have multiple hydroxyl groups. Note
that the viscosities of the solutions in xylene are highest. Xylene is a poor hydrogen-bond
Figure 3.13. Viscosity reduction of a hydroxy-functional UV-curable oligomer with xylene, MEK, and
methyl alcohol compared to the viscosity predicted if the viscosity reduction were a log-linear additive
relationship by weight. (From Ref. [20], with permission.)
58
FLOW
acceptor and hence promotes intermolecular hydrogen bonding between the resin molecules.
Methyl ethyl ketone (MEK) is a good hydrogen-bond acceptor and reduces the viscosity
more effectively than xylene by reducing intermolecular hydrogen bonding. Although
methyl alcohol is a much stronger hydrogen-bonding solvent than MEK, it is only marginally
better at reducing viscosity. Methyl alcohol is both a hydrogen-bond donor and acceptor.
Possibly, methyl alcohol can bridge resin molecules by functioning as a hydrogen-bond
donor with one resin molecule and a hydrogen-bond acceptor with the other; such bridging
would counteract the effectiveness of viscosity reduction.
Intermolecular hydrogen bonding between carboxylic acid groups is particularly
strong. Correspondingly, solvent effects on the viscosity of carboxylic acid– substituted
resins are particularly large [21]. It has been known for many years that simple carboxylic
acids, such as acetic acid, exist as dimers in poor hydrogen-bond acceptor solvents such as
benzene, whereas the dimers are dissociated in good hydrogen-bond acceptor solvents
such as acetone. The effect in resin solutions was demonstrated in a study of a primarily
monocarboxylic acid– substituted acrylic oligomer [22]. Relative viscosity in xylene was
higher than in methyl isobutyl ketone (MIBK). MW, as determined by vapor pressure
depression, was lower in acetone than in benzene. In xylene, hydrogen bonding between
carboxylic acid molecules was promoted; in MIBK, the predominant hydrogen bonding
is between ketone and carboxylic acid groups rather than between carboxylic acids.
Another solvent effect on viscosity is the solvent Tg. The Tg of resin solutions depends
on the concentration and the Tg of both the resin and the solvent. This effect has been
recognized in the addition of plasticizers to polymers, but has not been widely studied
in resin solutions having concentrations and viscosities in the range of interest for coatings. In one study it was found that the data fit Eq. 3.11, where Tgs is the Tg of the
solvent and Tgr is the Tg of the solvent-free resin [9]. In this study, Eq. 3.11 fits the data
over the entire range of concentrations from pure solvent to pure resin. Equation 3.11
needs to be tested with other systems.
1
ws
wr
þ
þ kwr ws
¼
Tg Tgs Tgr
(3:11)
The relationships are further complicated because solvent –resin interactions can be
further affected by resin MW and the number of polar groups per molecule. For
example, consider the data in Table 3.1 on the viscosity of solutions of a pair of
styrene/allyl alcohol (SAA) copolymers in methyl ethyl ketone (MEK) and toluene
TABLE 3.1. Effects of Molecular Weight and
Functional Group Content on Viscosity
Characteristic
SAA-I
SAA-II
M̄n
M̄w/M̄n
OH content (wt%)
Viscosities (mPa.s)
80 wt% in MEK
70 wt% in MEK
60 wt% in MEK
50 wt% in MEK
50 wt% in toluene
1600
1.5
5.7
1150
1.5
7.7
10,000
300
80
34
760
6500
230
65
30
3840
3.5. VISCOSITY OF LIQUIDS WITH DISPERSED PHASES
59
[20]. SAA-I has a higher MW, but a lower functional group content than SAA-II. In MEK,
an effective hydrogen-bond acceptor, the effect on viscosity of the OH content is diminished so that the higher MW of SAA-I results in a somewhat higher viscosity then that of
SAA-2. In toluene, which does not hydrogen bond effectively, the difference in OH
content dominates over the difference in molecular weight, so that the SAA-II solutions
have the higher viscosity. Comparison of the 50 wt% solutions in MEK and toluene
shows that the hydrogen-bonding solvent is more effective for viscosity reduction of
both SAAs. See Table 18.6 in Section 18.4 for examples of effects of solvents on the viscosity of solutions of a high-solids acrylic resin.
The viscosity of mixtures of hydrogen-bond acceptor solvents such as ketones and
esters are reported to be nearly ideal in their effects on viscosity; that is, the viscosity
of a mixture can be estimated quite well by calculating the weighted average viscosity
from those of the components [23]. However, in the case of mixing alcohols with other
solvents, the viscosities of mixtures varied substantially from ideal behavior. The deviation can be attributed to reduction of intermolecular hydrogen bonding of alcohols by
the other solvents, and was particularly pronounced when water was one of the solvents
in a mixture. Although these results with solvent mixtures add to our understanding of
the effects of intermolecular interactions on viscosity, they are not directly applicable to
the problem of mixed solvent effects on the viscosity of resin solutions. Since most
resins have multiple hydrogen-bond donor and acceptor sites, the interactions with
solvent are greater and more complex than in solvent blends. Little work has been published on this important question. In one paper, the authors suggest that the “best”
solvent in the mixture dominates in determining the effect on the intrinsic viscosity of
the resin solution [24]. The rationale is that the “best” solvent interacts most strongly
with resin molecules, hence controls the degree to which the resin molecules are extended.
It is hoped that further research will be reported on the effects of mixed solvents on resin
solution viscosity.
In reducing viscosity to spray viscosity by adding solvent to a coating, it is desirable to
have an equation that combines the effects of temperature and concentration on viscosity.
For the limited ranges of temperature and concentration involved, Eqs. 3.3 and 3.9 can be
combined to give such a relationship [25]. (In Eq. 3.12, log of solvent viscosity is
combined into constant K.)
ln h ¼ K þ
B wr
þ
T k1
(3:12)
Flow within coatings during film formation has a major influence on final film appearance and uniformity. It has proven to be a very difficult area to study because of the many
variables involved, including nonuniformities within the forming film. Progress is being
made, however. For example, Eley and Schwartz used mathematical modeling and
numerical simulation to develop a theory that correlated with the behavior of two
architectural paints [26].
3.5. VISCOSITY OF LIQUIDS WITH DISPERSED PHASES
Since many coatings contain dispersed pigment and/or resin particles, it is important to
consider the effect of dispersed phases on the viscosity of liquids. When a small
60
FLOW
amount of a dispersed phase is present, there is only a small effect on viscosity (unless the
dispersed phase is flocculated); however, as the volume of dispersed phase increases, there
is a sharply increasing effect. More energy is used to rotate particles, and the presence of
the particles interferes increasingly with the ability of other particles to move. When the
system becomes closely packed with particles, the viscosity approaches infinity.
Several equations have been developed that model the effect of a dispersed phase on
viscosity; the earliest was proposed by Einstein. Here we use a treatment by Mooney
[27]. Equation 3.13 is a useful form of the Mooney equation for understanding the
effects of variables on viscosity: where he is the viscosity of the continuous or external
phase, KE is a shape constant, Vi is the volume fraction of internal phase, and f is the
packing factor.
ln h ¼ ln he þ
KE Vi
1 Vi =f
(3:13)
The packing factor is the maximum volume fraction of internal phase that can be fit into
the system when the particles are randomly close packed and the external phase just fills
all the interstices between the particles. When Vi equals f, the viscosity of the system
approaches infinity. Figure 3.14 shows an example of a plot of the relationship between
log viscosity and Vi for a dispersed-phase system. Two major assumptions are involved
in the Mooney equation: (1) the particles are rigid, and (2) there are no particle –particle
interactions other than physical collisions.
The shape constant KE for spheres is 2.5. Some of the particles in coatings are spheres
or are reasonably close to being spheres, but many pigments are irregularly shaped. In the
case of uniform diameter spheres (i.e., monodisperse systems), the value of f is 0.637.
This value is the packing factor that has been calculated for a random mixture of
cubical and hexagonal close-packed spheres and has been confirmed experimentally. To
the surprise of many people first considering the question, the packing factor of monodisperse spheres is independent of particle size. The packing factor for basketballs is 0.637;
the packing factor for marbles is 0.637; the packing factor for monodisperse latex particles
is 0.637. However, marbles fit into the interstices between close-packed basketballs, and
latex particles fit into the interstices between close-packed marbles. In other words, the
packing factor depends strongly on particle size distribution: The broader the particle
size distribution, the higher the packing factor.
The viscosity of dispersions of nonrigid particles does not follow the Mooney equation.
When a shear stress is applied to such a dispersion (e.g., an emulsion), the particles can
distort. When the particles are distorted, the shape constant changes to a lower value
and the packing factor increases [28]; both changes lead to a decrease in viscosity. Commonly, such systems are thixotropic. This is logical since depending on the difference
between the viscosities of the internal and external phases, there would be time dependency of the distortion of the particles and hence a decrease in viscosity as a function
of time at a given shear rate. There are modifications of the Mooney equation that
account for viscosities of the two phases, but not time dependency [29]. Time dependency
can be studied using viscoelastic deformation analysis [5,6].
Fluids with readily distorted internal phases include emulsions, water-reducible acrylic
and polyester resins, urethane dispersions, some latexes in which the outer layer of the
latex particles and layers adsorbed on them are highly swollen by water, and some
pigment dispersions with comparatively thick adsorbed layers of polymer swollen with
3.5. VISCOSITY OF LIQUIDS WITH DISPERSED PHASES
61
Figure 3.14. Effect of increasing the volume fraction of noninteracting spherical particles on the viscosity of a dispersion. (From Ref. [20], with permission.)
solvent. (See Ref. [2] for a discussion of the effect of adsorbed layers on particles on
rheology.)
Although other factors may be involved, many thixotropic agents act by creating a
swollen dispersed phase that can be distorted. For example, very small particle size
SiO2 adsorbs a layer of polymer swollen by solvent that is thick compared to the
pigment and is distortable in a shear field. Up to a point, the degree of distortion increases
as the shear stress increases and/or as the time of shearing increases. When shearing is
stopped or decreased, the polymer layers recover their equilibrium shape and viscosity
increases. In other cases, lightly cross-linked polymer gel particles are used; the particles
swell with solvent, giving a distortable dispersed phase. The shear thinning behavior
depends on the particle size, concentration, and internal viscosity of the dispersed
phase. Smaller particles lead to higher shear rates for shear thinning. Shear thinning
decreases with decreasing concentration and increasing internal viscosity of particles.
The viscosity of dispersions is also affected by particle –particle interactions. If clusters
of particles form when stirring of a dispersion is stopped, the viscosity of the dispersion
62
FLOW
Figure 3.15. Effect of cluster formation on viscosity. (From Ref. [20], with permission.)
increases; if these clusters separate again when shear is exerted, the viscosity drops.
Examples of such shear thinning systems are flocculated pigment dispersions and flocculated latexes. Another is the gelation (really, flocculation) induced by water in coatings
containing treated clay dispersions [30]. When clusters of particles form, continuous
phase is trapped in the clusters; as a result, at low shear rates, Vi is high. At high shear
rates, when the clusters break up, the value of Vi is reduced to just that of individual
particles of the dispersion without a trapped continuous phase. As Vi increases, viscosity
increases, and vice versa.
One can also consider the dispersion from just the point of view of the Vi of the primary
particles only, adjusting the KE and f in Eq. 3.13 to account for aggregation rather than
adjusting the Vi. An example is shown in Figure 3.15. The vertical axis is the ratio of
the viscosity of the dispersion to the viscosity of the external phase (h/he). The viscosity
of the dispersion increases rapidly as the number of particles n in aggregates increases.
Polymer solutions containing dispersed phases are complex physical systems whose
flow properties are still the subject of continuing research. In this discussion we have
used the Mooney equation; alternative treatments, such as those of Krieger and Dougherty
[32] and Russel and Sperry [33], are also useful. A detailed analysis of effects of particle
interactions on dispersion rheology is included in Ref. [34].
3.5.1. Thickeners for Latex Coatings
Latex paints require a thickener to adjust viscosity. Viscosity of the paint as a function of
shear rate particularly affects ease of brushing, film thickness, leveling, sagging,
3.6. OTHER MODES OF FLOW
63
and settling. The viscosity of the external phase controls its rate of penetration into a
porous substrate, such as wood. If penetration is rapid, the viscosity of the paint above
the porous surface increases rapidly, resulting in poorer leveling. The water soluble
polymer hydroxyethylcellulose (HEC) is a widely used thickener. Since high MW HEC
leads to increased extensional viscosity, which in turn leads to spattering, larger
amounts of medium to low MW HEC have been used (Section 3.6.3). However, latex
paints thickened with HEC show a high degree of shear thinning, which may result in
the application of thin films of paint.
Progress in minimizing the shear thinning problem has been made with associative thickeners. Many kinds are available; all are moderately low MW, hydrophilic polymers with two
or more long chain nonpolar hydrocarbon groups spaced along the backbone. Examples
include hydrophobically-modified ethoxylated polyurethanes (HEUR), styrene–maleic
anhydride terpolymers (SMAT), and hydrophobically-modified alkali-swellable emulsions
(HASE). An example of a HASE thickener is a copolymer of MAA/EA/TMI on which
the NCO group has been reacted with an ethoxylated stearyl alcohol [35]. Another group
is hydrophobically-modified ethoxylated aminoplast (HEAT) polymers.
Use of associative thickeners permits formulation of latex paints that exhibit less shear
thinning, so that viscosity at high shear rates can be higher; as a result, thicker wet films
can usually be applied [36]. Associative thickeners interact by shared ionic bonds and by
absorption of the long aliphatic chains on the surface of latex particles. (The thicker
wet film in itself helps promote leveling too, since the rate of leveling depends on wet
film thickness, as discussed in Section 24.2.) Low shear viscosity can also be maintained
at an appropriate level, as with HEC, for controlling sagging, settling, and penetration into
porous surfaces. Combinations of associative thickeners and HEC are sometimes used.
Reynolds has reviewed possible mechanisms of thickening by associative thickeners
and factors involved in their use in formulating latex paints [37]. He emphasizes
that greater care and skill is required in formulating with associative thickeners than
with conventional water-soluble thickeners. The results obtained can be very sensitive
to the combination of the particular latex and thickener and to the amounts and types of
surfactant present in the formulation [37,38]. It has been shown that formulations with
associative thickeners not only level better, but also give somewhat higher gloss [39]. A
study of interactions between ethylated octylphenols with various ethoxylation levels
and a HEUR and a HEAT associative thickener has been published [40]. The associative
thickeners and surfactant with only 4.5 ethoxyl groups were incompatible. Compatibility
is particularly important in selecting the surfactant to be used in tint colors for use with
paints thickened with associative thickeners.
3.6. OTHER MODES OF FLOW
Although flow from application of shear stress is the most common type of flow encountered in making and using coatings, other modes of flow are sometimes involved.
3.6.1. Turbulent Flow
Turbulent flow occurs at very high rates of shear or in irregularly shaped containers and
pipes. At low shear rates, flow occurs in a laminar fashion, as illustrated in Figure 3.1.
However, as shear rate increases, a critical point is reached where flow suddenly
64
FLOW
Figure 3.16. (a) Conventional compared to (b) normal-direction flow of liquids on stirring.
becomes chaotic. Laminar flow is disrupted; swirling eddies and vortices occur, and flow
changes to turbulent flow. Even with Newtonian fluids, viscosity increases more than proportionally with shear rates above this critical level.
3.6.2. Normal Force Flow
When Newtonian fluids are stirred with a rotary stirrer, the liquid level becomes low in
the center near the shaft of the stirrer and high on the walls of the vessel as a result of
centrifugal force. This is shown schematically in Figure 3.16(a). However, some liquids
climb the shaft of the stirrer, as shown in Figure 3.16(b), rather than the sides of the
vessel. Such flow is normal (perpendicular) to the plane of force. This normal force
flow behavior is typical of viscoelastic materials and systems that are starting to gel. In
the early stages of cooking resins, the flow pattern is as shown in Figure 3.16(a), but if
cross-linking starts and gelation begins, the flow pattern can change abruptly to that
shown in Figure 3.16(b). If this occurs, it is time to stop heating and to dump the reaction
mixture out as quickly as safely possible before the gel becomes intractable.
Normal force flow effects have been observed in the handling and application of some
coatings whose flow is viscoelastic. Under stress, the flow of these coatings has both
elastic and viscous components. It seems logical to assume that a significant extent of
elasticity could affect atomization in spraying, film splitting in roller coating, and leveling;
however, there have been few studies of correlation between normal force flows and
coating application [8]. At least part of the reason for the lack of studies is that expensive
specialized, instrumentation such as an oscillatory plate rheometer is required for
measurements. Unfortunately, even many of those viscometers measure normal force
with such poor precision that correlations are not possible.
3.6.3. Extensional Flow
Another mode of flow encountered in some methods of coating application is extensional
flow, sometimes called elongational flow. Extensional flow occurs when fluid deformation
is the result of stretching. Various types of stretching are possible. In spin coating, extension occurs in two dimensions. The extensional flow of greatest importance in most other
coating processes is uniaxial (i.e., in one direction). In uniaxial flow, the viscosity is
3.6. OTHER MODES OF FLOW
65
properly called uniaxial extensional viscosity; we simply use the term extensional viscosity, but it should be remembered that there are several types of extensional viscosity.
The difference between extensional flow and shear flow was first observed in fiber
drawing. When the fiber material passes through the spinneret, the mode is shear flow.
However, as the fiber is pulled after leaving the spinneret, there is no further shearing
action; rather, the fiber is extended. The flow is extensional flow, and the resistance to
flow is extensional viscosity. The symbol used for extensional viscosity is he. In the
case of Newtonian fluids, he/h ¼ 3.
Extensional flow is encountered when applying coatings by direct roll coating
(Section 22.4). The material to be coated is passed through the nip between two rollers,
one of which is covered with a layer of coating. Prior to entering the nip, the coating is
subject to ambient pressure. As the coating goes through the nip, pressure rises and, as it
exits the nip, pressure begins to drop and can drop below ambient pressure. The pressure
drop can cause volatilization of solvents and release of dissolved air within the coating,
thus causing cavitation. The cavities lead to the formation of ligaments of coating that
get elongated under the extensional deformation in the nip exit region. If the extensional
viscosity is relatively low, the film splits quickly, leaving a ridged film. However, with
higher extensional viscosity, fibers grow; longer fibers tend to split in two places, resulting
in formation of droplets, which are thrown out into the air. This is called misting or
spattering. Figure 3.17 shows an extreme case of fiber development in roll coating of
paint with an impractically high extensional viscosity [41]. Soules et al. [42] discuss
the relationship of variables and extensional viscosity effects in roll coating.
A large fraction of latex flat wall paints is applied by roller. During roller application,
latex paints spatter, some to a major degree. Paints with high extensional viscosity are
likely to spatter severely [43]. Extensional viscosity increases when high-MW watersoluble polymers with very flexible backbones are used as thickeners in a latex paint
[41]. As a ligament of the paint is extended, the molecules line up in the direction of extension, leading to increased resistance to the extension. Spattering can be minimized by
using low MW water-soluble thickeners with rigid segments in the polymer backbone,
such as low MW HEC. Since the selling price of HEC is the same for various MWs,
and more low MW polymer is needed to reach the same shear viscosity, paint cost
increases when low MW HEC is used. It is common to compromise by using an intermediate MW or a mixture of high and low MW grades of HEC.
Extensional flow can also be encountered in spray application. If, for example, a solution of a thermoplastic acrylic resin with M̄W above about 100,000 is sprayed, instead
Figure 3.17. Fiber development in roll coating a high-extensional-viscosity paint. (From Ref. [34], with
permission.)
66
FLOW
of droplets coming out of a spray gun orifice, fibers emerge. As the fiber extends, the mode
of flow is extensional. The behavior is called cobwebbing. Although cobwebbing is undesirable when painting a car, it can give a desirable effect for applying decoration to a
Christmas tree. Soules et al. [44] discuss other possible extensional viscosity phenomena
in spray application. High extensional viscosity of latex paints that are airless sprayed
results in large particle size droplets and hence poor film appearance [45]. Fernando
et al. [45] also provide a review of various aspects of extensional viscosity effects in
coatings.
Measuring extensional (elongational) viscosity of liquids is difficult. One way is by
forming a fiber, wrapping it around a drum, and measuring the rate of extension and the
force required for further extension. In another adaptation of this method, the drum is
replaced with a vacuum nozzle that is vertically aligned with the nozzle that pumps the
fluid. Opposing jet viscometers are a second type of instrument. A third type is the
screen viscometer, a simple device that involves flow of a polymer solution through a
stack of screen [46]. ASTM is said to be developing a method that employs a screen viscometer. Recently, commercial instruments for measuring extensional viscosity have
become available.
GENERAL REFERENCES
Eley, R. R., Rheology and viscometry, in Paint and Coating Testing Manual: Gardner– Sward
Handbook, 14th ed., Koleske, J., Ed., ASTM, Philadelphia, PA, 1995, pp. 333 – 368.
Patton, T. C., Paint Flow and Pigment Dispersion, 2nd ed., Wiley-Interscience, New York, 1979.
Reynolds, P. A., The rheology of coatings, in The Chemistry and Physics of Coatings, Marrion, A.,
Ed., Royal Society of Chemistry, London, 1994.
Schoff, C. K., Rheology, Federation of Societies for Coatings Technology, Blue Bell, PA, 1991.
REFERENCES
1. Patton, T. C., Paint Flow and Pigment Dispersion, 2nd ed., Wiley-Interscience, New York,
1979.
2. Mewis, J.; Vermant, J., Prog. Org. Coat., 2000, 40, 111.
3. Hester, R. D.; Squire, D. R., Jr., J. Coat. Technol., 1997, 69(864), 109.
4. Hare, C. H., J. Prot. Coat. Linings, 2001, April, 79.
5. Boggs, L. J.; et al., J. Coat Technol., 1996, 68(855), 63.
6. Schoff, C. K., in Encyclopedia of Polymer Science and Engineering, 2nd ed., Wiley, New York,
1988, Vol. 14, pp. 454 – 540.
7. Anwari, F.; et al., J. Coat. Technol., 1989, 61(774), 41.
8. Schoff, C. K., Rheology, Federation of Societies for Coatings Technology, Blue Bell, PA, 1991.
9. Wicks, Z. W., Jr.; et al., J. Coat. Technol., 1985, 57(725), 51.
10. Toussaint, A.; Szigetvari, I., J. Coat. Technol., 1987, 59(750), 49.
11. Jones, F. N., J. Coat. Technol., 1996, 68(852), 25.
12. Haseebuddin, S.; et al., 1997, Prog. Org. Coat., 30, 25.
13. Kruithof, K. J. H.; van den Haak, H. J. W., J. Coat. Technol., 1990, 62(790), 47.
14. Wright, A. J., Eur. Coat. J., 1996, 32, 696.
REFERENCES
67
15. Schoff, C. K., in Polymer Handbook, 4th ed., Branderup, J.; et al., Eds., Wiley, New York, 1999,
p. VII/265.
16. Simon, P. F. W.; et al., Macromolecules, 2001, 34, 1677.
17. Sperry, P. R.; Mercurio, A., ACS Coat. Plast. Chem. Prepr., 1978, 43, 427.
18. Lee, W. A.; Rutherford, R. A., in Polymer Handbook, 4th ed., Branderup, J.; et al., Eds., Wiley,
New York, 1999, p. VI/193.
19. Erickson J. R., J. Coat. Technol., 1976 48(620), 58.
20. Hill, L. W.; Wicks, Z. W., Jr., Prog. Org. Coat., 1982, 10, 55.
21. Sherwin, M. A.; et al., J. Coat. Technol., 1981, 53(683) 35.
22. Wicks, Z. W., Jr.; Fitzgerald, G., J. Coat. Technol., 1985, 57(730), 45.
23. Rocklin, A. L.; Edwards, G. D., J. Coat. Technol., 1976, 48(620), 68.
24. Erickson, J. R.; Garner, A. W., ACS Org. Coat. Plast. Chem. Prepr., 1977, 37(1), 447.
25. Eiseman, M. J., J. Coat. Technol., 1995, 67(840), 47.
26. Eley, R. R.; Schwartz, L. W., J. Coat. Technol., 2002, 74(932), 43.
27. Mooney, M., J. Colloid Sci., 1951, 6, 162.
28. Jones, D. A. R.; et al., J. Colloid Interface Sci., 1992, 150(1), 84.
29. Nielsen, L. E., Polymer Rheology, Marcel Dekker, New York, 1977, pp. 56– 61.
30. Kemnetz, S. J.; et al., J. Coat. Technol., 1989, 61(776), 47.
31. Lewis, T. B.; Nielsen, L. E., Trans. Soc. Rheol., 1968, 12, 421.
32. Krieger, I. M.; Dougherty, T. J., Trans. Soc. Rheol., 1959, III, 137. Choi, G. M.; Krieger, I. M.,
J. Colloid Interface Sci., 1986, 113, 94, 101.
33. Russel, W. B.; Sperry, P. R., Prog. Org. Coat., 1994, 23, 305.
34. Goodwin, J. W.; Hughes, R. W., in Technology for Waterborne Coatings, Glass, J. E., Ed., ACS
Symp. Ser. 663, American Chemical Society, Washington, DC, 1997, pp. 94– 125.
35. Wu, W.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002, pp.
343 – 355.
36. Fernando, R. H.; et al., J. Oil Colour Chem. Assoc., 1986, 69, 263.
37. Reynolds, P. A., Prog. Org. Coat., 1992, 20, 393.
38. Chen, M.; et al., J. Coat. Technol., 1997, 69(867), 73.
39. Hall, J. E.; et al., J. Coat. Technol., 1986, 58(738), 65.
40. Garcia, R.; et al., FSCT Meeting, 2004.
41. Glass, J. E., J. Coat. Technol., 1978, 50(641), 56.
42. Soules, D. A.; et al., J. Rheol., 1988, 32, 181.
43. Massouda, D. B., J. Coat. Technol., 1985, 57(722), 27.
44. Soules, D. A.; et al., in Polymers as Rheology Modifiers, Glass, J. E., Ed., American Chemical
Society, Washington, DC, 1991, pp. 322 – 332.
45. Fernando, R. H.; et al., Prog. Org. Coat., 2000, 40, 35.
46. Prud’homme, R. K.; et al., Polym. Mater. Sci. Eng., 2005, 92, 241.
4
Mechanical Properties
4.1. INTRODUCTION: GENERAL COMMENTS ON TESTING
OF COATING FILMS
The critical properties of most coating films relate to their ability to withstand use without
damage. The range of potential mechanical damage is large. The coating on the outside of
an automobile should withstand being hit by a piece of flying gravel without film rupture.
The coating on the outside of a beer can must be able to withstand abrasion when cans rub
against each other during shipment in a railroad car. The coating on wood furniture should
not crack when the wood expands and contracts as a result of changing temperatures
during winter shipment or due to swelling and shrinkage resulting from changes in the
moisture content of the wood. The coating on aluminum siding must be flexible enough
to withstand fabrication of the siding and must resist scratching during installation on a
house. In addition, many coatings must also withstand the effects of weather (Chapter 5),
retain adhesion (Chapter 6), and protect metals from corrosion (Chapter 7).
In the introductory part of this chapter we discuss all aspects of durability, including the
complexities of developing, evaluating, and testing coatings. Development of coatings
with adequate durability is made complex by the wide range of conditions to which coatings are exposed. It is safe to generalize that the only way to know how a coating will
perform in actual use is to apply the coating to the final product, use the product over
its lifetime, and see whether the coating performs satisfactorily. But in many cases, the
lifetime of the product can be very long. The coating on the outside of an automobile
should maintain its integrity and appearance for well over five years. The coating on furniture should perform satisfactorily for 20 or more years. No laboratory tests are available
that permit satisfactory product performance predictions for many applications; however,
the formulator must have some way of judging the merits of a new formulation.
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
68
4.1. INTRODUCTION: GENERAL COMMENTS ON TESTING OF COATING FILMS
69
The most powerful tool is a data bank of field use performance of previous formulations. Formulators have made judgments about the effects of formulation changes on
durability based on their years of experience; experienced formulators try to pass on
their accumulated experience to novice formulators. Historically, new formulations
were relatively small modifications of formulations with known field performance. If
the change was significant, initial field use was usually limited. In the automobile field,
for example, after a promising new formulation had been developed, it was common to
coat just a few cars with it. Then, the next year, if no problems were encountered, the
new formulation might be adopted for one color on one model of automobile. The following year, the use might be extended to three or four colors on two or three models. Finally,
if all of this history was satisfactory, the new formulation might be widely adopted. This
gradual approach to formulating and testing worked quite successfully. However, in recent
years, there has been pressure to accelerate the process. This has resulted from increasing
performance requirements, pressures to reduce costs, and the need to meet regulation
requirements. Reduction of VOC emissions has been a major driving force; but
other factors, such as increasing recognition of possible toxic hazards, especially from
long term exposure to relatively low levels of some chemicals, have required changes
in relatively short time spans.
Increasingly, databases created from actual field use are being accumulated. For
example, for many years, teams of representatives of automobile manufacturers and suppliers have surveyed cars in parking lots in various parts of the country. The serial number,
which can be seen through the windshield, permits identification of the coatings on that
car. In other fields, performance of pipelines, exterior siding, and offshore oil rigs are monitored. Abrasion resistance of exterior beer can coatings can be related to shipment variables. Computers make possible analysis of masses of data correlating actual performance
with coating composition and application variables. The computer makes available to all
formulators in a company all records, rather than just the memory of one formulator. Data
on coatings that fail are as important as data on coatings that are satisfactory. More use of
this approach in the future is critical to future progress in formulating superior coatings. A
word of caution: Some technical people think only in terms of maximizing performance,
but economic factors are important as well. It is foolish to apply an expensive coating that
will last 20 years on a product that will last only five years. A five-year life coating can be
fully satisfactory and usually costs less than a 20-year life coating.
Accumulating a database takes time. Meanwhile, formulators need tools to work with.
The formulator has three kinds of needs when considering the mechanical properties of
films. To select the most appropriate components of the coating, the formulator needs
to understand the relationships between composition and properties; research instruments
can help develop this understanding. Laboratory tests are needed to follow the effects of
changes in formulation. Quality control tests are needed to compare production batches to
a standard. The accuracy and the appropriate uses of laboratory and quality control tests
are often quite limited. Unfortunately, many people working with coatings do not appreciate the limitations of the tests they use. It is common for formulators and users of coatings
to assume that an empirical quality control test or an accelerated laboratory test that has
been correlated with use performance for one type of coating can predict performance
for coatings of a different type. That is almost never the case.
Dickie has proposed a methodology for systematic consideration of the factors involved
in service life prediction [1]. He suggests that predictive models can provide the framework for assessing the importance and relevance of available information and may
70
MECHANICAL PROPERTIES
give insight as to what information may be missing from the evaluation of a given
material or application. A monograph by Martin et al. [2] provides a discussion of the problems of predicting service lives and proposes reliability theory methodology for database
collection and analysis.
4.2. BASIC MECHANICAL PROPERTIES
Understanding relationships between composition and basic mechanical properties of
films can provide a basis for more intelligent formulation. Most coating formulators
were educated as chemists, not as engineers, and few have had any education on mechanical properties. Terms such as loss modulus, storage modulus, and tan delta may have little
meaning to chemists, but such concepts have been used for many years in the plastics,
rubber, and fiber industries in developing products with superior performance. Because
of the diversity of coatings, the relatively small volume of most types of coatings, and
this lack of understanding of the physical behavior of films, the coatings industry has
lagged behind in applying these concepts. In 1977, Hill published a review paper discussing stress analysis as a tool for understanding coatings performance [3]. He did an excellent job of presenting an introduction to stress analysis in terms that a coatings formulator
could understand. Much of the information presented had to be based on examples from
plastics, rubber, and fiber work; there were few such papers that dealt with coatings at the
time. Thirty years later, application of stress analysis to coatings has mushroomed.
Basic to understanding the study of mechanical properties of coatings is a recognition
that coating films are viscoelastic materials. The mode of deformation of any solid can be
elastic and/or viscous. In ideal elastic deformation, or Hookean deformation, a material
elongates under tensile stress in direct proportion to the stress applied in conformance
with Hooke’s law, as exemplified by a steel spring. When the stress is released, the
material returns to its original dimensions essentially instantaneously. An ideal viscous
material, a Newtonian fluid, also elongates when a stress is applied in direct proportion
to the stress; but in contrast to an ideal elastic material, it does not return to, or even
toward, its original dimensions when the stress is released; the deformation is permanent.
Almost all coating films are viscoelastic—that is, they exhibit intermediate behavior.
Thermoplastic films frequently do not recover their original shape completely after deformation; the viscous flow part of the deformation is permanent. In cross-linked films, if
there is no yield point, the recovery of the original dimensions may be complete, even
though there was viscous flow. The stress on the cross-links supplies the force to
reverse the viscous flow. If there is a yield point, there is partial but incomplete recovery
of the original dimensions.
Figure 4.1 shows a schematic plot of the results of a stress –strain test, in which a
coating film is elongated (strain) at a constant rate and the resulting stress is recorded
[4]. (Methods for measurement of mechanical properties are discussed in Section 4.5.)
By convention, the stress (force per unit of cross-sectional area) is based on the original
dimensions. Strain is expressed in terms of percent elongation. The slope A of the
initial, essentially straight line, portion of the graph is the modulus: that is, the ratio of
stress to strain. One must be careful to know how the term modulus is used in a specific
case. In the initial part of this plot, modulus is independent of strain. However, as strain
increases, the ratio is no longer constant, and the modulus depends on the strain.
The end of the curve signifies that the sample has broken. This point is defined in
4.2. BASIC MECHANICAL PROPERTIES
71
Figure 4.1. Stress– strain plot. (From Ref. [4], with permission.)
two ways: (1) elongation-at-break is a measure of how much strain (E) is withstood before
breaking; and (2) the tensile strength, or tensile-at-break, is a measure of the stress (B)
when the sample breaks. The area under the curve represents the work-to-break (energy
vol21). Quite commonly, as shown in Figure 4.1, at an intermediate strain, the stress
required for further elongation decreases. The maximum stress (C) at that point is
called the yield point. Yield point can also be designated in two ways: elongation-at-yield
(D) and yield strength (C ).
An ideal elastic material deforms virtually instantaneously when a stress is applied and
recovers its original shape virtually instantaneously when the stress is released. Elastic
deformation is, over a wide range, almost independent of temperature. In contrast,
viscous flow is time dependent; the flow continues as long as a stress is applied. The
rate of deformation depends on the viscosity of the material as well as on the temperature.
As a result, viscoelastic deformation is dependent on the temperature and the rate at which
a stress is applied. If the rate of application of stress is rapid, the response can be primarily
elastic; if the rate of application of stress is low, the viscous component of the response
increases and the elastic response is correspondingly lower. Similarly, if the temperature
is low, the response can be primarily elastic, whereas at a higher temperature, the viscous
response is correspondingly greater.
These differences are illustrated in Figure 4.2(a) and (b). Plot (a) shows, schematically,
the results of elongating a film at two different rates at the same temperature. Curve A is
the same curve shown in Figure 4.1; curve B results from a more rapid application of
stress. In curve A there is time for the sample to undergo some viscous flow along with
the elastic deformation. In curve B, the stress was applied at such a rapid rate that there
was little time for viscous flow, and the elastic response dominated. Note that as commonly occurs in real samples, elongation-at-break is less, and tensile-at-break is greater,
when the rate of application of stress is higher. Plot (b) shows the results of pulling the
films at the same rate: equal to the slower rate in plot (a), but at two different temperatures.
Curve A is at the same temperature as in plot (a); curve B is at a lower temperature. At
the lower temperature, the viscosity was higher, so that even at the slower rate of
72
MECHANICAL PROPERTIES
Figure 4.2. Effects of (a) rate of application of stress and (b) temperature on stress– strain responses.
extension, there was essentially no viscous flow, and elastic deformation dominated. At
the higher temperature, the viscosity was low enough to permit substantial viscous flow
during stretching. As commonly occurs in real samples, elongation-at-break is less and
tensile-at-break is higher when the temperature is lower.
Plots (a) and (b) are identical. The rates of application of stress and temperatures were
chosen so that the change in the viscous response would be the same. In viscoelastic
materials, the effects of higher rates of application of stress and lower temperatures are
in the same directions. It is possible to do time – temperature superpositioning of curves
mathematically. If one’s instrument cannot operate at as high a rate of application of
stress as one would want in order to evaluate stress –strain behavior, one can operate at
a lower temperature and then calculate data points at higher stress rates, as discussed in
Ref. [4].
Stress – strain determinations are also run in creep or relaxation modes rather than in a
tensile mode. In creep experiments, a constant stress is applied and the resulting strain is
determined. An ideal elastic material undergoes an instantaneous strain, with stress application at time t1, that stays constant until the strain is removed at time t2, as shown in
Figure 4.3(a). A Newtonian fluid subjected to a creep test shows linearly increasing
strain with time, as shown in Figure 4.3(b). For a viscoelastic sample, strain is observed
to increase over time in a nonlinear manner. Typical responses for cross-linked and thermoplastic coatings (viscoelastic) are shown in Figure 4.3(c) and (d ), respectively. In a
relaxation test, one applies an instantaneous strain, elongating the sample, then follows
the change in stress with time. The stress stays constant with time for an ideal elastic
material, and a Newtonian liquid exerts no stress. For viscoelastic samples, the stress is
initially high and drops to lower values (relaxes) with time.
Stress – strain analysis can also be done dynamically by using instruments that apply an
oscillating strain at a specific frequency. The stress and strain vary according to sine
waves, due to the alternation from the oscillations. The stress and the phase angle difference between applied strain and resultant measured stress are determined. For an ideal
elastic material, the maximums and minimums occur at the same angles since there is
an instantaneous stress response to an applied strain; the phase shift is 08. For a Newtonian
fluid, there would be a phase shift of 908. Viscoelastic materials, on the other hand, show
an intermediate response, as illustrated in Figure 4.4. If the elastic component is high, the
phase shift d is small; if the elastic component is low compared to the viscous component,
4.2. BASIC MECHANICAL PROPERTIES
73
Figure 4.3. Plots of creep test responses. A constant stress is applied at t1 and removed at t2 for (a) an
ideal elastic solid, (b) a Newtonian fluid, (c) a viscoelastic sample with cross-links that result in complete
recovery, and (d) a viscoelastic sample that shows incomplete recovery. (Adapted from Ref. [4], with
permission.)
the phase shift is large. The phase shift, along with the maximum applied strain 10 and the
maximum measured stress s0, are used to calculate the dynamic properties.
The storage modulus E0 , also called the elastic modulus, a measure of elastic response,
equals (s0 cos d)/10. Its magnitude and physical significance are similar to modulus
Figure 4.4. Dynamic mechanical analysis plot as a sinusoidal strain is applied and a sinusoidal stress is
determined. (From Ref. [4], with permission.)
74
MECHANICAL PROPERTIES
obtained from the initial straight-line slope of a stress – strain curve, such as shown in
Figure 4.1. The term storage modulus reflects the fact that E0 measures the recoverable
portion of the energy imparted by the applied strain. The loss modulus E00 , a measure of
the viscous response, equals (s0 sin d)/10. The term loss modulus reflects the fact that
viscous flow leads to dissipation (as heat) of part of the energy imparted by the applied
strain. The square of the total modulus equals the sum of the squares of the storage and
loss moduli. The ratio E00 /E0 is called the loss tangent, since all of the terms cancel
except the ratio sin d/cos d, corresponding to the tangent of an angle, tan d, commonly
called tan delta.
Dynamic mechanical analysis (DMA) has the advantage over static stress –strain
studies, such as that shown in Figure 4.1, that the elastic and viscous components of a
modulus can be separated. The frequency of oscillation is a variable related to the rate
of application of strain. The higher the frequency, the greater the elastic response: that
is, the smaller the phase angle. The lower the frequency, the greater the viscous
response: that is, the larger the phase angle. The higher the frequency, the less time
there is for viscous flow; hence, elastic response dominates. Similarly, lowering the
temperature reduces viscous flow, decreasing the phase angle, and raising the temperature
gives greater viscous flow, increasing the phase angle. Generally, it is possible to run
experiments over a range of frequencies in dynamic tests that is wider than the range of
rates of application of stress possible in linear stress – strain experiments. Furthermore,
dynamic testing can be done over a wider range of temperatures and rates of heating.
In dynamic tests it is not possible to determine tensile-at-break (tensile strength),
elongation-at-break, or work-to-break, since the sample must remain unbroken in order
to run the test.
Stress – strain analysis discussed above is based on elongation of samples by application
of tensile stress. Oscillating (dynamic) and linear (static) stress – strain analysis can also be
carried out by application of shear or compressive forces. In shear tests, the stress is
applied sideways—analogous to shear viscosity tests. The ratio of shear stress to shear
strain, called shear modulus, is represented by the symbol G. For many coatings, the
tensile modulus E0 equals three times the shear modulus G00 . Recall from Section 3.6.3
that the analogous extensional and shear viscosity have the same relationship.
4.3. FORMABILITY AND FLEXIBILITY
In many cases, a coated metal object is subjected to mechanical forces either to make a
product, as in forming bottle caps or metal siding, or in use, as when a piece of gravel
strikes the surface of a car with sufficient force to deform the steel substrate. To avoid
film cracking during such distensions, the elongation-at-break must be greater than the
extension of the film under the conditions of fabrication or distortion.
To illustrate some of the variables involved, let us consider the simpler case of a plastic,
poly(methyl methacrylate) (PMMA), for which data on stress –strain relationships are
available. Tensile stress – strain curves of PMMA (Tg ¼ 378 K) as a function of temperature are shown in Figure 4.5. At low temperatures, there is no yield point, the modulus is
high, and the elongation-at-break is low. Breakage at low temperature is called brittle
failure. The terminology confuses some people; at low temperature, PMMA approaches
being an ideal elastic material, but its failure is classified as brittle because the
4.3. FORMABILITY AND FLEXIBILITY
75
Figure 4.5. Stress–strain plots of PMMA as a function of temperature. (From Ref. [3], with permission.)
elongation-at-break is low. At higher temperatures, but still below Tg, greater elongations
without breaking are possible, and there is a yield point.
Wu studied the modulus, elongation-at-yield, and elongation-at-break of PMMA and
other plastics as a function of temperature [6]. Figure 4.6 shows plots of these three properties versus temperature. The abrupt drop in modulus at Tg is typical of amorphous high
molecular weight (MW) thermoplastic homopolymers. Similar data are in the literature for
many such polymers. Fewer examples of the plots of elongation-at-break appear. Note that
Figure 4.6. Elongation-at-break and elongation-at-yield as functions of temperature, superimposed on
a storage modulus–temperature plot. (From Ref. [5], with permission.)
76
MECHANICAL PROPERTIES
elongation-at-break starts to increase rapidly at a temperature below the Tg of PMMA. Wu
defines the temperature at the intercept of the elongation-at-break and elongation-at-yield
plots as the brittle– ductile transition temperature, Tb. (Tb can also be defined as the intercept of plots of tensile strength and yield strength versus temperature.) Below Tb, the
polymer is brittle; between Tb and Tg, it is hard and ductile; above Tg, the polymer
becomes increasingly soft. Plastics can be deep-drawn above Tb; it is not necessary
to be above Tg, which is contrary to the common but erroneous definition of Tg as the
temperature below which an amorphous material is brittle.
Copolymers of MMA with methacrylates of longer chain alcohols, such as n-butyl
methacrylate, have lower Tg and Tb values. There is substantial variation in the difference
between Tg and Tb of various thermoplastic polymers, as illustrated by the data in
Table 4.1. Since these polymers are viscoelastic, the deformation is dependent on the
rate of application of stress as well as on temperature. Tg and Tb are both dependent on
the rate of application of stress: The higher the rate, the higher the Tg and Tb values.
Brittle – ductile transition temperature has been studied less with thermosetting
systems. The limited data available indicate that polymers with a low degree of crosslinking show a differential between Tb and Tg, but the differential decreases as
the cross-link density (XLD) increases. Probably, most thermoset coatings show little,
if any, differential between Tg and Tb, but few data have been published. The Tg of crosslinked polymers depends on several factors: structure of the segments between cross-links,
XLD, the amount of dangling chain ends, and the extent, if any, of cyclization of the
backbone [7].
It has been known for many years that there is a proportional relationship between XLD
and the modulus of low XLD elastomers above Tg, but it had been uncertain whether the
relationship could be extended to relatively higher XLD coatings. Hill demonstrated that
this relationship also holds for melamine – formaldehyde (MF) cross-linked acrylic
and polyester coating films [8,9]. The relationship is surprisingly simple, as shown in
Eq. 4.1, in which ve is XLD expressed as the number of moles of elastically effective
network chains per cubic centimeter of film. An elastically effective network chain is
one that is connected at both ends to the network at different junction points: Short cyclical
chains and dangling ends are not elastically effective. The terms E and G are tensile and
shear storage moduli, respectively. Since E00 and G00 are low at temperatures well above Tg,
E E0 and G G0 . It has been suggested that the simple form of Eq. 4.1 is favored when
network chains are short and when E0 is obtained with very small deformations, as used in
DMA [10]:
E ¼ 3ve RT(or G ¼ ve RT)
(when T .. Tg )
(4:1)
Thus, at least for the classes of films studied, it is possible to calculate XLD (ve) from
modulus against temperature plots [9]. Note that ve divided by the film density provides the
TABLE 4.1. Transition Temperatures of Homopolymers
Polymer
Polystyrene
Poly(methyl methacrylate)
Poly(vinyl chloride)
Bisphenol A polycarbonate
Tg (8C)
TB (8C)
Tg 2 Tb
100
105
80
150
90
45
10
2200
10
60
70
350
4.3. FORMABILITY AND FLEXIBILITY
77
moles of network chains per gram. The inverse, grams per mole of network chains, corresponds to the average MW of network chains, frequently called the average MW between
cross-links, M̄c. Commonly, M̄c is erroneously defined and used as the MW per branch
point; one must be careful when reading any paper dealing with MW of network chains
and XLD. Cross-link density can also be calculated and correlated with XLD measurements by swelling. Correlation is good, although it is necessary to assume an interaction
parameter from the extent of swelling of a film by solvent [8]. Although cross-linked films
do not dissolve in solvent, solvent does dissolve in a cross-linked film. As cross-links get
closer together (i.e., as XLD increases), the extent of swelling decreases.
XLD can also be determined by nuclear magnetic resonance imaging (NMR). The technique has the advantage that one can determine XLD as a function of depth through a film
[11]. For example, as expected, the top surface of air dry films of alkyds with only cobalt
driers could be seen to have cross-linked selectively. Also, the poor cure of the surface of
UV curing films could be seen to be air inhibited.
Equation (4.1) can also be used to predict the storage modulus above Tg from the XLD.
In a system with stoichiometric amounts of two reactants whose functional groups react
completely, one can estimate the XLD from the equivalent weights and the average functionality. If the reactant mixture contains molecules of several different functionalities,
calculation becomes more difficult. A more general approach is provided by the
Scanlan equation, where n is moles of elastically effect network chains per cm3 of
polymer film [8]:
3
4
5
n ¼ C3 þ C4 þ C5 þ
2
2
2
(4:2)
The C values are the concentrations of reactants with functionality of 3 to 5 (or more),
expressed in units of moles per cubic centimeter of final cured film. The volume of the
final film depends on the density of the cured film and loss of volatile by-products of
the reaction. Equation (4.2) does not include a term for difunctional reactants because
these reactants do not create junction points in a network; they only extend chains.
Recent modifications permit consideration of incomplete conversion [10]. Although the
Scanlan equation is convenient for stoichiometric reactions, it does not apply to other
cases. For non-stoichiometric mixtures and/or incomplete conversion, Miller –Macosko
equations are useful general equations. Bauer selected the Miller – Macosko equations
as most useful for coatings, gave examples of their applications, and provided a computer
program [12].
Properties are affected by the extent to which cross-linking has been carried to completion. Incomplete reaction leads to lower XLD and, hence, lower storage modulus
above Tg. The extent of reaction can be followed by determining storage modulus as
a function of time [13]. As cross-linking continues, storage modulus increases until a
terminal value is reached.
Thus, one can, at least in theory, design a cross-linked network to have a desired storage
modulus above Tg by selecting an appropriate ratio of reactants of appropriate functionality. By proper selection of the structures between cross-links and cross-link density, one
can design the Tg of the cross-linked network.
Effects related to cross-linking can be achieved by intermolecular hydrogen bonding.
For example, the strong hydrogen bonds between urethane groups leads to high
abrasion resistance in urethane coatings. When subjected to stress the hydrogen bonds
78
MECHANICAL PROPERTIES
rupture without breaking chemical bonds, relieving the stress. When the stress is released,
hydrogen bonds re-form, restoring the original properties.
A related phenomenon results from the presence of mesogenic groups on resins, which
when the film is cured, form liquid crystal domains. When stress is applied, the crystal
structure can be disrupted, and when the stress is relieved, crystal domains can re-form.
The initial examples of such a coating system were prepared by cross-linking a mesogenic
diol with hexamethoxymethylmelamine (HMMM) [14]. A variety of resin classes were
investigated, including epoxy –amine, polyester, alkyd, and acrylic resins with side
chains from p-hydroxybenzoic acid. There were non-cross-linked acrylic lacquers,
Class I melamine – formaldehyde cross-linked thermosetting coatings, and waterborne
coatings. Appropriate levels of mesogenic groups provided both high impact resistance
and film hardness, which are generally difficult to achieve in combination [15]. MF
resins in which HMMM is reacted with a mesogen such as the hydroxytriphenyl ether
of 1,6-hexanediol were used to replace part of the methoxy groups of the HMMM. The
resins were tested as coil coating vehicles and showed liquid crystal domains; they provided both high flexibility and hardness [16]. A UV cure coating formulated with an
acrylic ester of a mesogenic alcohol also provided both hardness and flexibility [17].
High cost has limited the commercialization of such coatings. Specialty electronic
coatings are reported to be in use.
Determination of dynamic mechanical properties has proved to be a valuable tool for
studying cross-linking of hydroxy-functional resins with MF resins [8,9]. As the stoichiometric ratio of methoxymethyl groups from the MF resin to hydroxyl groups was raised
from values less than 1, the storage modulus above Tg of the fully cured films increased
up to the point that the ratio became 1. As discussed in Section 11.3.2, these results
show that all of the functional groups on the MF resin can react with hydroxyl groups
and the reaction is not limited by steric hindrance, as thought previously. When the
amount of MF resin was increased so that excess methoxymethyl groups were present,
storage modulus above Tg increased at higher temperatures during dynamic mechanical
testing. This behavior is explained on the basis that the excess methoxymethyl groups
can undergo self-condensation reactions during testing. The self-condensation reaction
is relatively slow and was incomplete during the baking cycle used in preparing the
film; hence, the reaction continued at the higher temperatures used in dynamic mechanical
analysis, leading to the higher storage modulus. Self-condensation during baking of coatings also occurs when excess MF resin is used; then, the extent of self-condensation
increases as baking time and temperature increase.
As done in Figure 4.7, it is common to assign the peak of the tan delta curve as the Tg.
Some authors prefer to assign the peak of the loss modulus plot as the Tg. As can be seen,
there is a substantial difference between the two. The peak of the loss modulus plot is usually
nearer to the Tg determined using DSC. As noted before, Tg is dependent on the rate of
heating in DSC (Section 2.1.2) and on the rate of application of stress—the frequency of
oscillation—as well as the rate of heating in a dynamic test. This dependence on experimental technique might lead one to wonder if Tg means anything. It does, but one must always be
careful to compare Tg values determined in the same way. With regard to fabrication or
deformation of coatings, the most relevant Tg is the one determined at a rate of stress
application close to the rates of stress encountered in deformation of the coating.
An additional factor that can affect the mechanical properties of polymeric materials is
the breadth of the Tg transition region [18]. In some materials, the slope of the transition
region is steep, as shown in Figure 4.6. In others, this slope is shallower, as shown in
4.3. FORMABILITY AND FLEXIBILITY
79
Figure 4.7. Dynamic properties of a highly cross-linked acrylic coating film (solid lines). Dashed lines
indicate values E 0 (min), tan d(max), and Tg that are measures of the extent of cure. The values of
E 0 (min) and Tg tend to increase, whereas tan d(max) tends to decrease as the extent of cure increases.
(From Ref. [5], with permission.)
Figure 4.7. The same effect can be seen in tan delta plots, which exhibit various breadths.
The breadth of the Tg transition region can also be estimated by differential scanning
calorimetry (DSC), but dynamic mechanical analysis gives a clearer picture.
Factors controlling the breadth of the Tg transition have been only partly elucidated, but
broad tan d peaks are frequently associated with heterogeneous polymeric materials.
Blends of different thermoplastic resins often display two distinct Tg’s presumably
because of phase separation. Other blends of thermoplastics have a single, often broad,
Tg, presumably when phase separation is indistinct or when the phases are very small.
For thermosetting polymers, the Tg transition region is generally broader than for thermoplastics, and the breadth may vary considerably. Breadth of the distribution of chain
lengths between cross-links is a factor, and blends of thermosetting resins such as acrylics
and polyesters often display a single broad Tg transition [9]. As a rule—to which there
are probably exceptions—materials with broad and/or multiple Tgs have better impact
resistance than comparable polymers with a sharp, single Tg.
Some homogeneous polymers exhibit one or more, usually small, tan delta peaks at
temperatures below Tg. These peaks are called low temperature loss peaks, or b- and
g-transitions. They are thought to result from the onset of some specific small-scale
motions of parts of the polymer molecules as temperature is increased. These transitions
have been detected in a number of polymers and, in some cases, related to specific motions
of molecule segments [19]. In plastics it is well established that tough, impact-resistant
materials generally have low temperature loss peaks; polymers made from bisphenol A
(epoxies and polycarbonates) are common examples. It is reasonable to speculate that
coatings with low temperature loss peaks may have good impact resistance (if adhesion
is good), but the relationship is not well documented in the literature.
Mechanical properties of coatings are generally more complex than those of most plastics. One reason is that coatings are used as thin films on a substrate. Interaction with the
substrate affects the mechanical properties of thin films. The substrate can limit the extent
of deformation that occurs. The substrate can act as an energy sink to dissipate the energy
80
MECHANICAL PROPERTIES
so that there is less effect on the coating film. Adhesion can have a profound affect on the
ability to withstand fabrication. If adhesion is good, the fabrication and impact resistance
of the films are almost always superior. Stress within the film can act to pull the film off
the substrate. Coatings have been known to pop off postformed metal bottle caps when a
jar is sitting on a supermarket shelf. Stress within films can also arise during the last stages
of solvent loss and/or cross-linking of films [20]. Solvent loss and cross-linking result in
shrinkage. If the shrinkage occurs when the temperature is near the Tg of the film, the
resulting internal stresses may persist indefinitely. (See Section 6.2 for a discussion of
the effects of internal stresses on adhesion.)
Film thickness is also a factor in the ability of a coating film to withstand fabrication
without cracking. Thin films can be used for deeper draws than can thick films. In
making coated exterior siding, the hardness of the film can be increased without encountering cracking by limiting the film thickness. Of course, thinner films of pigmented coatings give poorer hiding; a common compromise in this case is at 20- to 25-mm film
thickness. Two-piece fish cans are coated as flat sheets with a relatively highly crosslinked phenolic coating that is quite brittle, to minimize swelling with fish oil. Such
cans can be formed successfully without cracking only if the film thickness is 5 mm or less.
Pigmentation affects the mechanical properties of films. In many cases, as the pigment
volume concentration of films is increased up to the critical pigment volume concentration, the tensile strength of the films increases (Section 21.1). A DMA study of aluminum pigmented base coats showed that the E0 increased tenfold as the pigmentation
increased in five steps from 0 to 13 PVC [21]. It is also possible that imperfections resulting from some types of pigmentation may lead to crack propagation. In some cases, the
storage modulus above Tg is increased by the presence of pigment. For pigmented coatings, one cannot expect a direct proportional relationship between storage modulus (E0 )
above Tg and XLD [8]; however, if the pigment content is constant in a series of films
varying in the polymeric portion, the relative values of E0 should still indicate relative
XLDs for the pigmented films. A study of the effects of pigmentation on mechanical properties of coatings is available. Tan delta is lower, broader, and shifted to higher temperature by pigmentation. E0 is increased, but above CPVC it decreases. Tensile strength
generally increases with PVC and then drops sharply above CPVC. However, if there is
poor bonding between the pigment particles and the resin, as is the case with CaCO3 in
an acrylic binder, the tensile strength drops steadily as the PVC increases [22]. (See
Section 22.1.1 for further discussion.)
Nano-pigments can have substantial effects on mechanical properties. Nano-pigments
are those with small particle size; some authors include those with average particle sizes of
less than 100 nm, but more commonly the particle size is below 25 nm. The term is
new but nano-pigments have been used in coatings for many years. For example, high
strength channel blacks have particle sizes in the range 5 to 15 nm. A review paper on
nano-material technology applications in coatings is available [23].
Compared to the effects of conventional pigments, nano-pigments, especially nanofillers, improve scratch, abrasion, heat, radiation, and swelling resistance, decrease
water permeability, and increase hardness, weatherability, modulus, and strain-to-failure
while maintaining toughness. These properties reflect the high surface area of nanopigments, resulting in an increase in adsorption of resin on the pigment surfaces and a
larger fraction of immobilized resin. In some cases, the films have two Tgs. Nano-pigments
have been shown to increase E0 and, in some cases, the Tg of clear-coat films [21]. When
the particle size of the pigments is smaller than the wavelength of light, coatings
4.3. FORMABILITY AND FLEXIBILITY
81
pigmented with nano-pigments are transparent [24]. Ball mills are used to disperse
nano(aluminum silicate) in an acrylic resin with a commercial dispersing agent. The
dispersion is used in making mar resistant clear coats [25].
The effect of heterogeneity in coating films has been recognized increasingly as an
important factor with regard to mechanical properties [26]. Most obviously, as pigment
volume concentration (PVC) of a film increases, the elastic modulus increases due to
the reinforcing effect of pigments. If the pigment is increased above the critical
pigment volume concentration (CPVC), the tensile strength drops precipitously
(Chapter 22). The increase in elastic modulus is affected by the aspect ratio of the pigments: fibrous . laminar . spherical. One can obtain significant effects by incorporating
elastomeric particles in a coating. The use of a combination of low Tg and high Tg
components gives films with higher blocking resistance and longer elongation.
Another important variable can be the timing of fabrication or flexing after curing a
coating. It is common for coatings to become less flexible as time goes on. Particularly
in air dry coatings, some solvent may be retained in films. Since most coatings have Tg
values near or a little above room temperature, solvent loss may be very slow (Section
18.3.4). Solvents generally act as plasticizers, so as solvent is lost, Tg and storage
modulus increase and coatings tend to become less flexible. In cross-linkable coatings,
if the cross-linking reaction was not complete, the reaction may continue slowly, increasing XLD and, hence, storage modulus and decreasing flexibility. Continued cross-linking
is particularly likely to occur in air dry coatings, since reaction rates are likely to become
mobility rate controlled, and hence, the last part of the reaction is slow, as discussed
in Section 2.1.3. Reactions during the use life of the coating, especially during exterior
exposure (Section 5.1), can result in embrittlement. Dynamic mechanical analyses of
a variety of clear coats after Florida exposure and QUV testing show the changes in
mechanical properties resulting from exposure [27].
Hardening of baked cross-linked coatings over time is also commonly observed.
Although in some cases, further volatile loss or continued cross-linking may be responsible, another possible factor is densification. If a polymer is heated above its Tg and
then cooled rapidly (quenched), the density is commonly found to be lower than if the
sample had been cooled slowly. During rapid cooling, more and/or larger free volume
holes are frozen into the matrix than with slower cooling, which provides greater
chances for molecular motion. On storage, the molecules in quenched films slowly
move even though the temperature is somewhat below Tg and free volume decreases,
causing densification. Since this process results in changes in properties with ageing
with no chemical change, it is also called physical ageing. With the decrease in free
volume and the increase in density, cracking during fabrication is more likely. This
phenomenon has been widely observed in plastics and is increasingly being recognized
in coatings. It may occur when coatings are baked on metal at high temperatures and
then cooled rapidly after coming out of the oven. It has been suggested that densification
may be a common cause of embrittlement during ageing of baked coating films. Greidanus
has studied physical ageing at 308C of polyester/MF films that had been baked at 1808C
and then quenched to 308C [28]. There was a small but reproducible increase in modulus
with time at 308C. The aging rate (i.e., the rate of increase of modulus) decreased with
time. If the sample was heated again at 1808C and again quenched to 308C, the
modulus returned to its lower value and underwent physical ageing again. Perara has
shown that relative humidity can also affect physical ageing; samples stored at high
humidity age more quickly that those stored at 0% RH because of the plasticizing
82
MECHANICAL PROPERTIES
effect of water relaxing the polymer so that it can rearrange to its “normal state” [29].
Perara [29] also reviews the field of physical ageing extensively. Further work is
needed, but it is evident that physical ageing can be an important phenomenon.
When coatings are baked industrially, there are further complications involved. Oven
temperatures can vary not only the air temperature in the oven as a whole, but also within
the oven. The rate of heating of a coating can depend on the substrate thickness. For
example, the temperature of the coating on the sheet metal roof of a car increases faster
than the temperature of the coating on joints, where the metal thickness is greater. To
achieve the desired properties, some minimum time at a temperature is required, but overbaking can lead to excessive cross-linking. There is a cure window for any baked coating;
within this range of time and temperature, satisfactory properties are obtained. As discussed
in Section 11.3, high-solids acrylic melamine coatings have narrower cure windows than
conventional solids coatings had. The effects of some of the variables in thermal history
have been modeled [30].
4.3.1. Fracture Mechanics
When exposed to stress, some coatings crack, leading to drastic film failure. Weathered
clear coats on automobiles, fabrication of precoated metal, grain cracking of some wood
furniture finishes, and cracking of exterior house paint are examples. All of these are
examples of failure by fracture mechanics. For a variety of reasons, stresses can accumulate in films; if the stresses exceed the tensile strength of the films, failures will result. The
main sources of stress in coatings are thermal expansion mismatches, humidity stresses
due to mismatches in swelling between the coating and the substrate, increased crosslinking such as by exterior exposure, and densification due to physical ageing.
The failure of clear coats on automobiles has been studied. Such failures are encountered especially after weathering. When coatings are exposed to UV radiation and high
humidity, cross-link density increases, which leads to shrinkage of the film, increasing
stress. Fracture mechanics of weathered automobile coatings has been discussed in an
excellent paper [31]. With UV absorbers, the protection is least at the surface, leading
to higher cross-linking at film surfaces and in turn to surface cracks. To study the stresses
required to grow cracks in weathered automobile coatings, cracks of known size are
made in the base coat and clear coat on steel. Then they are extended in a tensile
tester and the strain is determined at which cracks in the clear coat are observed. The
driving force, G, can be calculated with the equation where Ef is the modulus divided
by (1 2 nf)3 (n is Poisson’s ratio, typically taken to be near 0.35), Z is a geometry
constant (3.95 for surface cracking), s is the stress applied, and h is the thickness of
the clear coat.
G¼
Zs2 h
Ef
When the driving force exceeds the fracture energy (G Gc), crack propagation ensues.
Note that fracture energy depends on film thickness and on the stress squared.
In the same study coatings on strips cut from weathered panels were pulled in tension
in a mechanical testing machine at 20 mm min21 and the strain at which cracks appeared
was noted. Uncertainty in results was large; testing 8 to 10 samples reduced uncertainty
4.4. ABRASION AND MAR RESISTANCE
83
to +25%. The following equation was used to calculate Gc:
Gc ¼
12 hE f pgða; bÞ
2
where 1 is the strain at cracking and g(a, b) is a constant relating to the mismatch between
the coating and the underlying layer. For a coating on steel g(a, b) ¼ 0.78, and for a clear
coat on a base coat it equals 1.26. In coatings that cracked on exposure, failures almost
always occurred in the clear coat layer. If adhesion between the clear coat and base
coat was poor, failure resulted in delamination.
Further understanding of changes in Gc was gained by study of films of the separate
base and clear coatings [32]. The effects of temperature change and exposure to different
levels of relative humidity were reported. Overbaking was shown to reduce Gc. Three
MF –acrylic clear coats showed that the Gc varied considerably with formulation; an
epoxy –acid clear coat showed the lowest Gc. Cracks can also propagate to the surface
between the top coat and the lower interface, which may then lead to delamination.
(See Section 6.2 for a discussion of fracture mechanics and adhesion.)
4.4. ABRASION AND MAR RESISTANCE
Abrasion is the wearing away of a surface, whereas marring is a disturbance of a surface
that alters its appearance. Both phenomena are included in the field of tribology, the
science of surfaces in sliding contact. Terminology is not standardized, and terms such
as scratching, buffing, gouging, and wearing are used with meanings that sometimes
overlap abrasion and marring.
4.4.1. Abrasion Resistance
One might suppose that hard materials are less likely than soft materials to fail by abrasion.
In some cases, this assumption is true, but in many other cases, softer materials are more
abrasion resistant; for example, rubber tires resist abrasion far better than steel tires. Evans
studied the mechanical properties of a series of floor coatings with known actual wear life
[33]. He determined tensile-at-break, elongation-at-break, and work-to-break. His data are
given in Table 4.2; the coatings are listed in order of increasing wear life. One might
suppose that higher tensile strength would give higher abrasion resistance; the data
show the reverse. (It should not be assumed from these limited data that abrasion resistance is always inversely related to tensile strength.) Elongation-at-break values gave
TABLE 4.2. Mechanical Properties of Floor Coatings
Floor
Coating
Hard epoxy
Medium epoxy
Soft epoxy
Urethane elastomer
a
Tensile
Strength (psi)
Elongationat-Break (%)
Work-toBreak (in.-lb/in3)
Tabera (rev/mil)
9000
4700
1100
280
8
19
95
480
380
600
800
2000
48 103
33 103
23 103
36 103
See Section 4.6.3.4 for a discussion of these results.
84
MECHANICAL PROPERTIES
the proper rank order, but Evans concluded that work-to-break values best represented the
relative wear lives. Intuitively, it seems reasonable that abrasion resistance would be
related to work-to-break. Work-to-break values vary with the rate of application of
stress and should be determined at a rate comparable to that encountered in use.
In studies on another series of coatings, Evans and Fogel determined that work-to-break
did not always correlate with abrasion resistance determined by loss of gloss in a ball mill
abrasion tester when the stress –strain tests were carried out at ambient temperatures [34].
They reasoned that the strain rate of their instrument was too low relative to the stress
application in the test. Using a time – temperature superposition relationship, they calculated that the tests at an accessible strain rate should be carried out at 2108C in order
to compensate for the instrumentally inaccessible high rate of stress application at
ambient temperature. The resulting work-to-break values did correlate with abrasion
resistance for urethane films with a Tg equal to or greater than 2108C.
In studies of erosive wear of clear coats for automobiles, it has been shown that wear
resistance increases as energy-to-break of films increases [35]. Erosion rate is also affected
by the substrate; for example, clear coats applied directly to steel showed significantly less
durability than when applied over primer and base coats. Wear tends to increase as the
angle of application of stress decreases.
Urethane coatings generally exhibit superior abrasion resistance combined with solvent
resistance. This combination of properties may result from the presence of intersegment
hydrogen bonds in addition to the covalent bonds. At low levels of stress, hydrogen
bonds act like cross-links, reducing swelling on exposure to solvent. At higher levels of
stress, the hydrogen bonds can dissociate, permitting the molecules to extend without rupturing covalent bonds. When the stress is released, the molecules relax and new hydrogen
bonds form. Urethanes are used as wear layers for flooring, as well as top coats in aerospace
applications, where this combination of properties is desirable.
Factors in addition to work-to-break are involved in abrasion resistance. The coefficient
of friction of the coating can be an important variable. For example, abrasion of the
coating on the exterior of beer cans during shipment can be minimized by incorporating
a small amount of incompatible wax or fluorosurfactant in the coating. When the two
coated surfaces rub against each other, the incompatible additive reduces surface
tension and the coefficient of friction, so that transmission of shear force from one
surface to the other is minimized and abrasion is reduced.
Another variable is surface contact area. Incorporation of a small amount of a small
particle size SiO2 pigment in a thin silicone coating applied to plastic lens eyeglasses
reduces abrasion, adding to the effect of the low surface tension of the silicone surface.
The pigment particles reduce contact area, permitting the glasses to slide more easily
over a surface. Another example of the same principle is the incorporation of a small
amount of coarse SiO2 inert pigment in wall paints to reduce burnishing. If a wall paint
without such a pigment is frequently rubbed, as around a light switch, it abrades to a
smoother glossier surface: that is, it burnishes. The coarse inert pigment reduces burnishing
by reducing contact area.
An approach that has been used for many years in resin-bonded pigment print colors on
textiles is to incorporate rubber latex in the print paste. The latex particles are not soluble
in the resin and end up as individual particles in the resin, along with the pigment particles.
The abrasion resistance is markedly improved by the latex addition. Similar work is now
being done to improve the abrasion resistance of continuous coatings. Presumably, the
relatively soft rubber particles act to dissipate stresses on the film, minimizing the
4.4. ABRASION AND MAR RESISTANCE
85
chance of a stress concentration leading to film rupture. Glass microspheres have been
shown to increase abrasion resistance of epoxy coatings by damping the energy released
by impacts on the surface [36]. Lee reviewed abrasion resistance as one type of wear in a
broad approach to fracture and surface energetics of polymer wear [37].
4.4.2. Mar Resistance
Mar resistance is related to abrasion resistance, but there is an important difference.
Abrasion may go deeply into the coating, whereas marring is usually a near-surface
phenomenon; mars less than 0.5 mm deep can degrade appearance. Marring is a major
problem with automobile coatings, particularly those with a clear coat as the final top
coat (Section 30.1.2). In going through automatic car washes, the surfaces of some
clear coats are visibly marred and may lose gloss due to marring [38,39]. Mar resistance
is also a critical requirement in coatings for floors and for transparent plastics: for example,
on polycarbonate glazing or eyeglasses. Physically, marring is a complex phenomenon;
while mars may be very shallow, the physical properties deep in the coating film can influence mar resistance.
The physics of marring is complex. Different authors use different terms to describe the
phenomena involved. Various models have been proposed to describe what happens to a
viscoelastic material when a hard object is drawn over its surface. One such model classifies the response of the material as elastic, plastic, or fracture. Since the elastic response
recovers essentially instantaneously, only plastic deformation and fracture lead to marring.
Qualitatively, mars caused by plastic deformation have shoulders, whereas those caused
by fracture do not. Although simplistic, this model has the advantage that the three
responses can be measured quantitatively by scanning probe microscopy [40] or with a
nano-indenter [41]. Most coatings exhibit a mixture of responses. Different coatings
vary widely in their responses, and the same coating may respond quite differently as
the force or rate of marring stress changes. A further complication is that mars in some
coatings can heal slowly by reflow (creep).
Efforts are underway to relate mar resistance to the chemical structure of coatings, but
relatively few systematic studies have been published. In general, MF cross-linked acrylic
clear coats are more resistant than isocyanate cross-linked coatings (urethanes) to marring,
but MF cross-linked coatings have poorer environmental etch resistance. MF cross-linked
polycarbamates are an exception, combining etch and mar resistance (Section 11.3.4).
Since urethanes generally have superior abrasion resistance, it is surprising that they
have inferior mar resistance; possibly, this might be explained by differences between
surface and bulk properties. A study of marring of clear coats by a scanning probe microscope indicated that an acrylic polyurethane had a thin layer of deformable plastic material
on its surface, whereas an acrylic –MF clear coat had a layer of elastic material [40]. Reference [42] reviews mar test methods and compares them by subjective estimations of performance. Methods include two simulated car wash tests, a crockmeter test originally
developed for evaluating abrasion of colored textiles, microindentation hardness tests,
scrub resistance tests, a Rota-Hub test, a collate sanding machine, microscratch resistance,
a nano-scratch tester, and dynamic mechanical analysis. Reference [43] compares a variety
of scratch tests on cross-linked trimethoxypropylsilane-functional clear coats for plastics.
New instruments, nanomechanical analyzers (nano-indenters), that can characterize
near-surface mechanical properties are now available [44,45].
86
MECHANICAL PROPERTIES
Two strategies are available for the design of coatings with exceptional mar resistance.
They can be made hard enough that the marring object does not penetrate far into the
surface, or they can be made elastic enough to recover after the marring stress is
removed. If the hardness strategy is chosen, the coating must have a minimum hardness.
If the Tb (Section 4.2) is above the testing temperature, groove formation is minimized;
however, such coatings may fail by fracture. Film flexibility is an important factor influencing fracture resistance. Use of 4-hydroxybutyl acrylate instead of 2-hydroxyethyl acrylate
in an acrylic resin cross-linked with MF resin gave improved results, as did use of a polyolmodified hexamethylene diisocyanate isocyanurate instead of isophorone diisocyanate
isocyanurate in cross-linking urethane coatings [46]. Further improvement was obtained
using silicone-modified acrylic resins. Courter proposes that maximum mar resistance
will be obtained with coatings having as high a yield stress as possible without being
brittle [47]. In this way, high yield stress minimizes plastic flow, and avoidance of brittleness thereby minimizes fracture. Courter’s paper provides a good review of attempts to
relate bulk mechanical properties of coatings to their mar resistance, but these studies
have not led to a broadly applicable theory of marring. This is understandable, since the
mechanical properties near a coating’s surface are likely to be quite different from the
mechanical properties of the bulk material.
A study of the scratch resistance of 49 different clear top coats subjected to a
simulated car wash test has been reported [48]. High correlation between measurements
and visual evaluation was obtained. Determination of dynamic mechanical proper ties
showed that best performance was associated with high storage modulus at a relaxation
time of 1 s and high minimum strain that would cause plastic deformation of the films.
A further problem related to mar resistance is metal marking. When a metal edge is
rubbed across a coating, a black line is sometimes left on the coating where metal has
rubbed onto the surface of the coating. A common test for metal marking resistance is to
draw a coin across the surface of a coating to see if a dark streak is left. Coating lore says
that a Canadian nickel is the best kind of coin to use. Metal marking usually occurs with relatively hard coatings. The problem can be reduced or eliminated by reducing the surface
tension of the coating, so the coefficient of friction is low; then the metal slips over the
surface. Additives can be incorporated to increase slip. Modified polysiloxanes have been
reported to be particularly effective [49]. Care must be exercised in selecting the particular
grade of silicone additive and the amount of the additive used so as to minimize marring,
scratching, and metal marking without causing other defects, such as crawling.
4.5. MEASUREMENT OF MECHANICAL PROPERTIES
Most instruments require free films for measurement of mechanical properties. Two major
disadvantages to using free films are that (1) the interaction of the film with the substrate
can have major effects on some film properties, and (2) free films are sometimes difficult to
prepare and handle. Test results are generally more reproducible with thick films than with
thin films; however, the results with thick films may not be applicable to thin films. Preparation of thin unsupported films can be difficult. In some cases, it is possible to make a
film by drawing down a coating on a release paper with a wire wound bar. Release
papers are coated with low surface tension materials to minimize adhesion; but if the
surface tension of the release paper is lower than the surface tension of the coating
4.5. MEASUREMENT OF MECHANICAL PROPERTIES
87
being applied, there is the possibility of crawling; that is, the coating tries to minimize
surface free energy by drawing up into a ball. (See Section 23.4 for a discussion of crawling.) One tries to find a release paper with a low enough surface tension so that adhesion is
poor, but high enough so that crawling does not occur. A generally more effective method
is to apply the coatings to tin-plated steel panels. After curing, one end of the panel is
placed in a shallow pool of mercury. Mercury creeps under the coating, forming an
amalgam with the tin, and the film comes free of the panel. Mercury vapor is toxic, and
care must be taken to minimize the hazard. The safety regulations of some laboratories
forbid such use of mercury. After the film is freed from the substrate, a specimen is cut
from it. Cutting free films may result in nicks or cracks along the edge of the film.
When subjected to stress, cracked films commonly tear easily starting at the imperfection,
leading to meaningless results. Handling films with a Tg above room temperature is
especially difficult; they tend to be brittle and easily broken.
One must be careful about changes that may occur during storage before testing, such
as loss of residual solvent, chemical changes, or physical ageing. Test results will correspondingly change with time. Storage conditions can be critical. Most films absorb some
water from the atmosphere. If the Tg is near room temperature, and especially if the film
has groups such as urethanes that hydrogen bond strongly with water, the Tg and film properties can be strongly affected by the humidity conditions in storage, since water acts as a
plasticizer. Comparisons should be done with samples that have been stored at the same
temperature and humidity. In actual use, films encounter a variety of humidity conditions
and hence show a variation in properties.
Several types of instruments are available for determining mechanical properties. The
Instron Tester is used for tensile (nondynamic) experiments. The free sample is mounted
between two jaws of the tester; care must be taken to ensure that the film is in line with the
direction of pull. The instrument can be run with a range of rates of jaw separation, but
even the highest rates are slow compared to the rates of stress application found in
many real situations. This problem can be partially overcome by running the tests at
low temperatures. This method has the advantage that stress can be increased until the
film fails, making possible determination of tensile strength, tensile modulus,
elongation-at-break, and work-to-break. However, one cannot separate the viscous and
elastic components of the mechanical properties.
A thermal mechanical analyzer (TMA) is a penetrometer that measures indentation
versus time and temperature. An advantage over most tensile instruments is that a TMA
includes a furnace and temperature programmer, so that heating, cooling, and isothermal
operations are possible. TMAs can be used with films on a substrate. An example of TMA
is measuring the softening point, which is related to the extent of cure of cross-linking
films. Figure 4.8 shows a plot of probe penetration as a function of temperature for an
undercured and a well-cured 25 mm acrylic coil coating. The softening points for the
two samples are marked on the graph. The softening point is related to, but not identical
to, Tg; it is frequently used as an index of flexibility [50].
Various dynamic mechanical analyzers (DMA) are also available [51]. The most versatile are those in which the free sample is subjected to an oscillating strain by attachment
under tension to a fixed clamp on one end and a vibrating clamp on the other. Oscillating
stresses are imparted to the sample. A range of frequencies can be used, and properties can
be determined over a wide range of temperatures. The most sophisticated instruments are
set up in line with a computer that analyzes the data and provides storage and loss modulus
88
MECHANICAL PROPERTIES
Figure 4.8. TMA plot of probe displacement against temperature for an undercured and a well-cured
acrylic coil coating. (From Ref. [5], with permission.)
and tan delta figures and plots as functions of temperature. A schematic diagram is shown
in Figure 4.9.
Two types of DMA rheometers are available: controlled stress (CSS) types and controlled shear rate (CSR) types. In CSS rheometers one presets a stress and measures the
resulting deformation or shear rate. In CSR rheometers one presets the shear rate and
measures the stress. Rheometers that can operate in either mode are also available.
Another type of instrument used for dynamic mechanical analysis is a torsional pendulum. In its simplest form, one end of a film is fastened in jaws, and the lower end is
attached to a disk to which weights can be added. The lower weight is twisted, setting
up a pendulum motion whose decay can be analyzed to give the dynamic properties.
Torsional pendulums have been most widely used not with film but with a fiber braid
that is saturated with liquid cross-linking polymers. The instrument has the advantage
that it can be used to follow changes in dynamic properties, starting with liquid coatings
as reactions occur on the braid. It has the disadvantages that the sample is not a film and
that there are large surface areas of fiber – polymer interface that may affect properties.
A torsional pendulum apparatus has been modified so that shrinkage can be measured
simultaneously with mechanical properties [52]. Since the volume change is largely
Figure 4.9. Dynamic mechanical analyzer.
4.6. TESTS OF COATINGS ON SUBSTRATES
89
proportional to the extent of cure, it can be used as a measure of the progress of cure, while
simultaneously measuring changes in mechanical properties.
The range of tensile modulus that can be measured with a torsional pendulum without
changing the weight on the pendulum is limited. Furthermore, since the resonance frequency
is used as a measure of viscoelastic behavior, modulus results and the measuring frequency are
not independent. These difficulties can be overcome by using a DMA instrument with glass or
carbon fiber braid support [53].
4.6. TESTS OF COATINGS ON SUBSTRATES
A variety of test methods has been established to characterize properties of coatings. In
general, these methods do not permit calculation of the basic mechanical properties, but
rather, test some combination of properties of the coating. There are two categories of
such tests: One type can be appropriate for prediction of actual performance; the
second type is only suitable for quality control. There are very real needs for both types
of tests. But too commonly, paint tests, which may be appropriate for quality control,
are used to predict performance, even though they are not capable of providing results
that permit performance predictions. Three broad classes of coating tests are field exposure
tests, laboratory simulation tests, and empirical tests.
4.6.1. Field Exposure Tests
As stated in the introduction to this chapter, the only reliable way of knowing how a
coating will perform is to use it and see how it performs. The next best approach is to
use the coating in field applications on a smaller scale and under especially stringent conditions that may accelerate possible failure. The more limited the test and the greater the
degree of acceleration, the less reliable are the predictions, but carefully designed and
analyzed tests of this type can be very useful. There are many examples; we cite a few
to illustrate the principles.
Highway marking paints can be tested by painting stripes across the lanes of traffic
instead of parallel to the traffic flow. In this way, exposure of the paint to wear is
greater, and many paints can be tested and compared on a short length of highway,
where they will receive the same amount of traffic wear. Controls with known performance should be tested alongside new coatings. Tests should be done at different times
of the year because the effects of hot sun, snow plows, salt application, and so on, must
be considered. Tests should be set up on different types of highway materials, such as
concrete and asphalt. Floor paints can be evaluated similarly in busy hallways.
Test automobiles painted with new coatings are driven on torture tracks with stretches
of gravel, through water, under different climate conditions, and so on. Sample packs of
canned goods are made; the linings are examined for failure and the contents evaluated
for flavor after various lengths of storage.
4.6.2. Laboratory Simulation Tests
Many tests have been developed to simulate use conditions in the laboratory. The value of
these tests depends on how well use conditions are simulated and how thorough a validation procedure has been used. An important key to the value of any test for performance
90
MECHANICAL PROPERTIES
prediction is the simultaneous evaluation of standards with known performances that
cover the range from poor to excellent performance. It is not enough to use only the
extremes of standards; although such information may be a first step for checking the possible applicability of a test, performance prediction requires more than two standard data
points. Enough replica tests must be run to decide how many repeat tests must be run
to give results within desired confidence limits. Chemists commonly think of standard
deviations, but these are only 67% confidence limits; the odds are 33% that the proper
value is outside the standard deviation range. (See Ref. [2] for a comprehensive discussion
of problems in predicting service life.)
An example of a well-validated test is the laboratory shaker test devised to simulate the
abrasion of can coatings when six-packs of beer are shipped by railcar [54]. Six-packs are
loaded onto a shaker designed to simulate the pressures, rate of shaking, range of motion,
and so on, actually encountered in rail shipments. The test was validated using cans with
known field performance. The times to reach different degrees of abrasion failure were
established. In unpublished work, in laboratories of several coatings suppliers and
can manufacturers, the results were compared with known performance and with the
results of standard abrasion tests. It was found that none of the standard laboratory
abrasion tests gave satisfactory predictions, but this test gave results that could be used
for performance prediction.
The automobile industry uses gravelometers to evaluate resistance of coatings to chipping when struck by flying gravel. Pieces of standard gravel or shot are propelled at the
coated surface by compressed air under standard conditions, usually at low temperature.
The tests have been standardized by comparison to a range of actual results and have
been found to give reasonably good predictions of actual performance. A more sophisticated instrument, a precision paint collider, which permits variations in angle and velocity
of impact and temperature has been described [55]. It was used in studies of impact
failure of various coatings on automotive bumpers made with various plastic polymers.
In some cases, cohesive failure occurred in the upper layers of the plastic; in other
cases, the coating delaminated from the plastic.
Several laboratory devices are available that approximately reproduce stamping or
other forming operations to test the ability of coated metal to withstand fabrication. Individual companies design such tests to match the conditions of their factory’s forming operations as closely as possible. Commonly, such simulation tests check only one or a few of
the performance requirements, so they must be used along with other tests to predict
overall performance. For example, the shaker test for beer cans obviously can give no
information on the important ability of the coating to withstand the pasteurization procedure, but separate simulation tests are available for testing pasteurization performance.
In most cases, simulation tests are designed for use in performance prediction rather than
quality control. Generally, the sample preparation and testing time are too long for checking
whether production batches are equal to the standard.
4.6.3. Empirical Tests
A range of empirical tests is used for coatings. In some cases, the results can be used as
part of the data to predict performance, particularly when comparing formulations that
are very similar to standards with known performance. In most cases, empirical tests
are more appropriate for quality control. They are frequently required in product specifications, an appropriate practice when empirical tests are used for quality control. But
4.6. TESTS OF COATINGS ON SUBSTRATES
91
quality control tests established for one coating should be revalidated if they are to be
applied to a new type of coating. Commonly, there is a considerable range of error, and
many replicates should be run.
We can only mention a few of the most widely used tests. ASTM annually publishes
books describing tests; most of the tests of importance to the coatings field are in
Volumes 06.01, 06.02, and 06.03: Paint—Tests for Formulated Products and Applied
Coatings. The books are published annually, but most of the methods are not changed,
although they are supposed to be reviewed every four years. Each method has a
number, such as ASTM D-2794-93 (Reapproved 1999). The “93” indicates that the test
was approved in 1993. In this particular case, the test was reevaluated and reapproved
in 1999. If one finds a reference to a test as D-1876-71 and then refers to a 1997
ASTM book and finds it listed as D-1876-88, it means that the test was last reviewed,
and possibly revised, in 1988. In general, one should use the new test method. Sometimes,
a method will be designated as D-459a. The “a” indicates that there was a minor rewriting
that did not change the basic method. A too common assumption is that because a test
method is given an ASTM designation, it can be used not just for quality control, but
also to predict performance. This is frequently not the case. However, following ASTM
procedures does provide the best chance of obtaining comparable test results. Some
ASTM methods include precision statements, usually based on repeatability and reproducibility studies involving different laboratories. The precision statements should not be
ignored; many people believe that the tests are more precise than they proved to be in
ASTM round robin tests [56]. In Europe, tests established by the German Institute for
Standards (Deutsches Institut für Normung, DIN) are widely used.
A useful reference book is the Paint and Coating Testing Manual: Gardner–Sward Handbook [57]. It provides descriptions of a wide range of test methods and summaries of each
major class of properties, as well as background information and comparisons of the utility
of various tests. Hill provides a brief informative, discussion of the more important tests in
his monograph on mechanical properties [51].
4.6.3.1. Hardness Measuring the hardness of a viscoelastic material is not as straightforward as it might appear. Units of hardness, force per unit area (Nm22), are the same as
those of modulus. When interpreting hardness and modulus data, it is important to know
how the force was applied (tensile, shear, bending, or compressive?), the rate of application of stress, and the temperature. Since the area changes during deformation, it
is also important to know what the area was before, during, and after deformation.
As reviewed by Guevin, three types of empirical hardness tests are used for coatings:
indentation, scratch, and pendulum tests [58].
One indentation test is run with a Tukon Indentation Tester (ASTM D-1474-98). An
indenter with a diamond-shaped tip is pressed into the film with a fixed weight for a
fixed time. The indenter is raised, and the indentation left in the film is measured with a
calibrated microscope. The results are expressed in Knoop Hardness Numbers (KHN),
which are related to the weight divided by the area of the indentation. Results are affected
by film thickness; thin films on hard substrates give higher values than do thicker films of
the same coating. Meaningful results are obtained only with high Tg films. Intermediate-Tg
materials may have partial recovery of the indentation in the time needed to move the
sample under the microscope and make the measurement. Low Tg films give considerable
response variation, and the indenter may leave no indentation at all on a rubbery material.
A careless tester may conclude that a rubbery material is very hard, even though it is
92
MECHANICAL PROPERTIES
obviously soft. The Tukon method is most appropriate for baked coatings, since they are
more likely to have a Tg above the testing temperature.
A widely used scratch test is the pencil hardness test (ASTM D-3363-00). Pencils with
hardnesses varying from 6B (softest) to 9H (hardest) are available. The “lead”—actually
graphite and clay—in the pencil is not sharpened as for writing, but is squared off by
rubbing perpendicularly on abrasive paper. For the test, the pencil is held at a 458 angle
to the panel and pushed forward with a pressure just below that which will break the
lead. Hardness is reported as the grade of pencil that does not cause any marring of
the surface. Experienced testers can reproduce results to +1 hardness grade. Probably,
the test reflects some combination of modulus, tensile strength, and adhesion.
A pendulum test, particularly used in the United States, is the Sward rocker (ASTM
2134-93). The rocker is a circular device made up of two rings joined with a glass level
gauge and is weighted off center. The circumference of the rings rests on the panel.
The rocker is rolled to a marked angle and released. The number of rocks (swings)
required to dampen the motion down to a smaller fixed angle is determined. The rocker
is calibrated to give 50 rocks (a hardness reading of 100) on polished plate glass. Hard
coatings give higher readings (but less than 100) than soft coatings. Dampening is
caused by rolling friction as well as by mechanical loss. The results depend on film thickness and surface smoothness. The Sward rocker is probably most useful for following the
increase of hardness of a coated panel during drying of an ambient cure coating. It has only
limited validity for comparing the hardness of different coatings.
In Europe, Koenig and Persoz pendulums are used. In these tests, a pendulum makes
contact with a coated panel through two steel balls. As the pendulum swings back and
forth through a small angle, movement of the balls requires some deformation of the
coating near the surface. Test results are reported as the time (in seconds) for the swing
to be dampened from a higher to a lower angle (from the perpendicular)—from 6 to 38
in the case of the Koenig pendulum and from 12 to 48 in the Persoz test. Usually,
harder coatings give longer times. However, soft, rubbery coatings may also give
longer times. Based on the reasonable assumption that the main contribution to dampening
the pendulum is absorption of mechanical energy by the film, these apparently conflicting
results can be explained in terms of the loss modulus. As shown in Figure 4.7, loss
modulus values are low in both the regions below and well above Tg and are highest in
the transition region. Low loss modulus could account for longer dampening times for
both soft, rubbery films, with Tg values below ambient temperature, and hard films,
with Tg values well above ambient temperature. This hypothesis predicts that dampening
times for coatings in the transition region at ambient temperature may be very sensitive to
temperature, since loss modulus goes through a maximum in this region.
4.6.3.2. Formability and Flexibility One flexibility test is a mandrel bend test, in
which a coated panel is bent around a rod or cone (the mandrel). The panel is bent with
the coated side away from the mandrel. Any crack in the coating on the bend is reported
as a failure. In the cylindrical mandrel test, a series of different mandrels is used, and the
smallest diameter that permits a bend without failure is reported as the mandrel flexibility.
In a conical mandrel test, one end of the panel is clamped, and a lever is used to bend the
panel over a cone. The distance from the small end of the cone to the end of a crack is
reported. This distance, which is proportional to the radius of curvature of the mandrel
at that point, can be used to estimate the elongation-at-break [57]. Thicker films crack
more easily than thinner films because the elongation at the same distance along the
4.6. TESTS OF COATINGS ON SUBSTRATES
93
mandrel is greater. The bent edge should be inspected with a magnifying glass to see
hairline cracks, and the panels should also be inspected the next day because cracks sometimes appear later. If adhesion is poor, the film comes loose from the panel. The severity of
the test can be increased by testing at low temperature (by putting the panel and tester in a
freezer before testing). The severity of the test is also affected by the rate of bending.
Another formability test, widely used for testing coil coatings, is the T-bend test. The
coated panel is bent back on itself with the coating side out. If there is no crack at the edge,
the result is reported as 0T. The zero means there is no additional layer of metal inside the
bend. If the coating cracks, the panel is bent back on itself again. Repeated bends back
over the original bend are made until the coating does not crack. The radius of curvature
gets greater as the number of bends increases. The results are reported as 0T, 1T, 2T, 3T,
and so on, counting the number of thicknesses of metal inside the bend. The severity of the
test is affected by temperature and rate of bending, and the panels should be reexamined
after some time has elapsed.
4.6.3.3. Impact Resistance Impact tests evaluate the ability of a coating to withstand extension without cracking when the deformation is applied rapidly [ASTM
D-2794-93 (reapproved 1999)]. A weight is dropped down a guide tube onto a hemispherical indenter that rests on a coated panel. An opening opposite the indenter in the base
support on which the panel rests permits deformation of the panel. If the coated side is
up (i.e., the coating is hit directly by the indenter), the test is called a direct impact test.
If the back of the panel is up, the test is called a reverse impact test. The weight is
dropped from greater and greater heights until the coating on the panel cracks. In the
United States, the results are reported in in.-lbs (i.e., the number of inches the weight
falls times its weight). In the most common apparatus, the maximum is 160 in.-lb. Generally, reverse impact tests, in which the coating is extended, are more severe than direct
impact tests in which the coating is compressed. The thickness, mechanical properties,
and surface of the substrate affect the results substantially. If the substrate is thick
enough, it is not distorted by the impact, and almost any coating will pass. Comparison
of results on different substrates is meaningless. Different lots of test panels of the same
type can have subtle differences in their surfaces that can affect test results. The test
panels should be aged for at least one day before testing and should be inspected immediately after the test and again the next day.
4.6.3.4. Abrasion Resistance In the Taber Abraser test, abrasive wheels roll on a
panel, creating a circular wear path. The test is continued until the coating is worn
through. Results are reported as the number of rotations required to wear through 1 mil
(25 mm) of coating. The results often do not correlate with field use tests. For example,
in the Taber abrasion results for the four floor coatings described in Table 4.2, the hard
epoxy coating had the poorest abrasion resistance in actual field use, but it exhibited
the highest rating in the Taber test. Another example in which correlation was poor is
the beer can abrasion problem mentioned previously. On the other hand, some authors
report that Taber tests correlate with visual observation of abrasion of clear coats on
cars by automatic car washes [39], although others disagree. Generally, softer coatings
tend to give poorer Taber Abraser results, probably because the abrasion disks rotate at
a constant speed and therefore impart more energy to softer coatings. However, very
soft coatings may clog the abrasive on the wheel and give spurious results.
94
MECHANICAL PROPERTIES
Another abrasion test is the falling sand test, in which sand falls out of a hopper through
a tube onto a coated panel held at a 458 angle to the stream of sand. The result is reported in
liters of sand required to wear through a unit thickness of coating. A more sophisticated
version of the falling sand test has been reported [35]. A gas blast erosion apparatus has
been shown to give reproducible results. When the stream of particles is perpendicular to
the surface, the erosion scar produced is circular; the radius of the scar provides an accurate
measure of abrasion. The velocity of the particles and the angle of impact can be varied.
Yet another method of determining abrasion uses a ball rotating in a slurry of small
abrasive particles to measure abrasive wear of a small area of coating in the upper
30 mm of a film with good reproducibility [35].
4.6.3.5. Solvent Resistance Solvent resistance is not a mechanical property, but it is
included here because it is one of the properties that must be balanced with mechanical
properties for many applications. It is also appropriate to mention here because resistance
to swelling of cross-linked films is related to XLD, which affects many mechanical
properties.
The most common test of solvent resistance is the methyl ethyl ketone (MEK) double
rub test (ASTM D-4752-98). This test can be done by rubbing a tissue soaked with MEK
on the panel, but it is more conveniently done with a felt tip marker pen filled with solvent.
The test can be mechanized so that there is one stroke back and forth (one double rub) on
the film each second, and a timer can be used as a counter. A soft thermoplastic coating
will rub off with very few rubs. In the case of thermosetting coatings, the number of
rubs that the coating withstands increases as the degree of reaction increases. The test
is sensitive to the development of low cross-link density but insensitive to changes as
the cross-link density gets higher. Usually, the test is stopped after 200 double rubs,
and therefore, a series of highly cross-linked coatings may all be reported to give 200þ
double rub resistance, even though there may be differences in the extent of cross-linking.
At higher cross-link densities, DMA or solvent swelling can be used to determine the
extent of cross-linking, as discussed in Section 4.2.
Another solvent resistance test is to expose the coating to the solvent for a certain period
of time (say, 15 minutes) and then carry out a pencil hardness test on the exposed area.
For example, a test for aircraft top coats specifies that the coating shall not lose more
than two pencil hardness units after exposure to hydraulic fluid for 15 minutes.
GENERAL REFERENCES
Hill, L. W., Dynamic mechanical and tensile properties, in Paint and Coating Testing Manual:
Gardner– Sward Handbook, 14th ed., Koleske, J. V., Ed., ASTM, Philadelphia, PA, 1995,
pp. 534 – 546.
Koleske, J. V., Ed., Paint and Coating Testing Manual: Gardner– Sward Handbook, 14th ed.,
ASTM, Philadelphia, PA, 1995.
Mark, J. E., Ed., Physical Properties of Polymers Handbook, American Institute of Physics,
Woodbury, NY, 1996 (particularly, Brostow, W.; Kubat, J.; Kubat, M. M., Mechanical properties, pp. 313 – 334).
Rubenstein, M.; Colby, R. H., Polymer Physics, Oxford University Press, New York, 2004.
Sperling, L. H., Introduction to Physical Polymer Science, 3rd ed., Wiley-Interscience, New York,
2001.
REFERENCES
95
REFERENCES
1. Dickie, R. A., J. Coat. Technol., 1992, 64(809), 61.
2. Martin, J. W.; et al., Methodologies for Predicting Service Lives of Coating Systems, Federation
of Societies for Coatings Technology, Blue Bell, PA, 1996.
3. Hill, L. W., Prog. Org. Coat., 1977, 5, 277.
4. Hill, L. W., Mechanical Properties of Coatings, 2nd ed., Federation of Societies for Coatings
Technology, Blue Bell, PA, 1996.
5. Hill, L. W., Mechanical Properties of Coatings, Federation of Societies for Coatings Technology, Blue Bell, PA, 1987.
6. Wu, S. J., J. Appl. Polym. Sci., 1976, 20, 327.
7. Stutz, H.; et al., J. Polym. Sci. B Polym. Phys., 1990, 28, 1483.
8. Hill, L. W., J. Coat. Technol., 1992, 64(808), 29.
9. Hill, L. W., J. Coat. Technol., 1987, 59(751), 63.
10. Hill, L. W., Prog. Org. Coat., 1997, 31, 235.
11. Hellgren, A. C.; et al., Prog. Org. Coat., 2001, 43, 85.
12. Bauer, D. R., J. Coat. Technol., 1988, 60(758), 53.
13. Skrovanek, D. J., Prog. Org. Coat., 1990, 18, 89.
14. Dimian, A. F.; Jones, F. N., Polym. Mater. Sci. Eng., 1988, 56, 640.
15. Jones, F. N.; et al., U.S. patent 6,218,045 (1993).
16. Yoshida, K.; et al., Prog. Org. Coat., 2005, 52, 227.
17. Yoshida, K.; Kakuchi, T., Prog. Org. Coat., 2005, 52, 165.
18. Roller, M. B., J. Coat. Technol., 1982, 54(691), 33.
19. Fried, J. R., in Physical Properties of Polymers Handbook, Mark, J. E., Ed., American Institute
of Physics, Woodbury, NY, 1996, pp. 161 – 175.
20. Perera, D. Y.; Schutyser, P., FATIPEC Congress Book, 1994, Vol. I, p. 25.
21. Schlesing, W.; et al., Prog. Org. Coat., 2004, 49, 197.
22. Perera, D. Y., Prog. Org. Coat., 2004, 50, 247.
23. Fernando, R., JCT Coat. Tech, 2004, 1(5), 32.
24. Perera, D. Y., Prog. Org. Coat., 2004, 50, 247.
25. Vanier, N. R.; et al., U.S. patent 6,916,368 (2005).
26. Kivilevich, A., JCT Coat. Tech, 2004, 1, 38.
27. Hill, L. W.; et al., Prog. Org. Coat., 1994, 24, 147.
28. Greidanus, P. J., FATIPEC Congress Book, 1988, Vol. I, p. 485.
29. Perara, D. Y., Prog. Org. Coat., 2003, 47, 61.
30. Dickie, R. A.; et al., Prog. Org. Coat., 1997, 31, 209.
31. Nichols, M. E., J. Coat. Technol., 2002, 74(924), 39.
32. Nichols, M. E.; et al., Prog. Org. Coat., 1999, 35, 153.
33. Evans, R. M., in Treatise on Coatings, Myers, R. R.; Long, J. S., Eds., Marcel Dekker,
New York, 1969, Vol. 2, Part I, pp. 13– 190.
34. Evans, R. M.; Fogel, J., J. Coat. Technol., 1977, 47(639), 50.
35. Rutherford, K. L.; et al., Wear, 1997, 203 – 204, 325.
36. Kotnarowska, D., Prog. Org. Coat., 1997, 31, 325.
37. Lee, L. H., Polym. Mater. Sci. Eng., 1984, 50, 65.
38. Hamada, T.; et al., Prog. Org. Coat., 1997, 30, 271.
96
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
MECHANICAL PROPERTIES
Betz, P.; Bartelt, A., Prog. Org. Coat., 1993, 22, 27.
Jones, F. N.; et al., Prog. Org. Coat., 1998, 34, 119.
Shen, W.; et al., JCT Res., 2004, 1(2), 117.
Osterhold, M.; Wagner, G., Prog. Org. Coat., 2002, 45, 365.
Yoneff, P. S., et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2001, pp. 109 – 125.
Alers, P.; Consiglio, R., Prog. Org. Coat., 2000, 40, 287.
Lin, L.; et al., Prog. Org. Coat., 2000, 40, 85.
Gregorovich, B. V.; Hazan, I., Prog. Org. Coat., 1994, 24, 131.
Courter, J. L., J. Coat. Technol., 1997, 69(866), 57.
Hara, K.; et al., Prog. Org. Coat., 2000, 40, 39.
Fink, F.; et al., J. Coat. Technol., 1990, 62(791), 47.
Skrovanek, D. J.; Schoff, C. K., Prog. Org. Coat., 1988, 16, 135.
Hill, L. W., in Paint and Coating Testing Manual, 14th ed., Koleske, J. V., Ed., ASTM,
Philadelphia, PA, 1995, p. 534.
Lange, J.; et al., Surf. Coat. Int., 1996, 79, 486.
Frey, Th.; et al., Prog. Org. Coat., 1996, 27, 59.
Vandermeersche, G. A., Closed Loop, 1981, April, 3.
Ryntz, R. A.; et al., J. Coat. Technol., 1995, 67(842), 23.
Athey, R. D., Jr., Am. Paint Coat. J., 1992, December, 7.
Koleske, J. V., Paint and Coating Testing Manual, 15th ed., ASTM, West Conshohocken, PA,
2006.
Guevin, P. R., Jr., J. Coat. Technol., 1995, 67(840), 61.
5
Exterior Durability
Exterior durability of coatings refers to their resistance to change during outdoor
exposure; such changes include changes of modulus, loss of strength, embrittlement, discoloration, loss of adhesion, chalking, loss of gloss, and environmental etching. Thus, both
aesthetic and functional properties are involved. The terms outdoor durability and weatherability are also used. Corrosion protection by coatings is discussed in Chapter 7.
The most common chemical processes leading to degradation of coatings are photoinitiated oxidation and hydrolysis resulting from exposure to sunlight, air, and water.
These processes are interrelated, including enhanced photoxidative degradation in high
humidity and enhanced hydrolytic degradation during photoexposure. Furthermore, both
processes are accelerated by higher temperatures. Hydrolytic degradation may be
enhanced by exposure to acid, as from acid rain. Other atmospheric degradants include
ozone and oxides of nitrogen and sulfur. Changes in temperature and humidity may
result in cracking, which arises from the expansion and contraction of coatings or substrates. Rates at which these processes occur vary, depending on exposure site(s), time
of year, coating composition, and substrate.
5.1. PHOTOINITIATED OXIDATIVE DEGRADATION
Coatings formulated for exterior durability should exclude or minimize resin components
that are readily oxidized and that absorb UV radiation at wavelengths longer than 290
nm. Photoinitiated oxidation of polymers by a chain reaction is outlined in Scheme
5.1. Absorption of UV by a polymer (P) or other coating component produces highly
energetic photoexcited states (P ) that undergo bond cleavage to yield free radicals
(P.). Free radicals undergo a chain reaction with O2 (autoxidation), leading to polymer
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
97
98
EXTERIOR DURABILITY
degradation. Hydroperoxides (POOH) and peroxides (POOP) are unstable products of
photoinitiated oxidation; they dissociate with sunlight and moderate heat to yield
alkoxy (PO.) and hydroxy (HO.) radicals, so the degradation reactions are autocatalytic.
These radicals are highly reactive toward hydrogen abstraction and yield polymer radicals (P.), which enter into the propagation stage of polymer degradation. Tertiary
alkoxy radicals dissociate into ketones and a lower-molecular-weight polymer radical
(P0 .), resulting in scission of the polymer. As shown in Scheme 5.1, chain propagation
leading to oxidative degradation proceeds by hydrogen abstraction from the polymer
autocatalytically.
To achieve good exterior durability, functional groups in a coating that are vulnerable
to hydrogen abstraction should be avoided or minimized. A general ordering of common
CH groups having decreasing vulnerability to free radical abstraction is the following
(the groups at the bottom of the list are least susceptible to oxidative degradation):
Scheme 5.1
5.1. PHOTOINITIATED OXIDATIVE DEGRADATION
99
The order is based on available bond dissociation energies and well-established activating effects of heteroatoms and aryl/vinyl groups; neighboring groups and steric hindrance as
well as statistical effects should also be considered. For example, poly(vinylidene fluoride),
)2,
22
( CH222CF22
n is known to weather substantially better than other polymers having CH2
groups, which probably reflects the neighboring F groups as well as the statistical reduction
of CH2 groups. Linolenic esters having 1,4,7-triene groups are more readily oxidized than
linoleic esters having 1,4-diene groups, which may be attributed to a neighboring group
effect of the third ene group in linolenates as well as to the statistical effect of two highly
activated CH2 groups between double bonds in linolenates compared to one in linoleates.
In general, hydrogen abstraction from tertiary alkyl groups is disfavored statistically and
is expected to be particularly sensitive to steric hindrance effects.
Enhancement of hydrogen abstraction by electron-donating N and O groups and retardation by strongly electron-withdrawing F groups can be attributed to polar contributions
to the transition state for hydrogen abstraction. Since most hydrogen abstracting agents are
electronegative (e.g., peroxy or alkoxy radicals), the transition state will be stabilized by
partial negative charge on the abstracting agent and, correspondingly, partial positive
charge on the C from which the hydrogen is abstracted, as shown, where RX represents
the hydrogen abstracting agent. Accordingly, the energy of the transition state will be
lowered by electron-donating groups, which stabilize the partial positive charge and
raised by electron-withdrawing groups, which destabilize the partial positive charge.
None of these functional groups absorbs sunlight directly. In the absence of absorbing
aromatic groups, absorption of sunlight occurs primarily by inadvertently present peroxide
and ketone groups, both of which absorb ultraviolet (UV) above 290 nm. Essentially, all
resins used in coatings contain some hydroperoxides, as do most organic substances. Peroxides, ketones, and aldehydes are also formed in photoinitiated oxidation. (Scheme 5.1)
Photolysis of peroxides and ketones is shown in Schemes 5.1 and 5.2, respectively. As
shown in Scheme 5.2, photoxidation of aldehydes and ketones yields peracids, which
are strong organic oxidants. Peracids may play a significant role in oxidative degradation.
Methyl-substituted silicones and silicone-modified resins exhibit high photoxidation
stability, generally in proportion to the silicone content. The excellent exterior durability
100
EXTERIOR DURABILITY
Scheme 5.2 Oxidation of aldehydes and ketones (peracid formation).
of fluorinated resins may be attributed, at least in part, to the absence (or reduced level) of
C22H groups.
Aromatic groups with directly attached heteroatoms, as found in aromatic urethanes
(Ar22NH22CO22OR) and bisphenol A (BPA) epoxies (Ar22O22R), absorb UV above 290
nm and undergo direct photocleavage to yield free radicals that can participate in oxidative
degradation. BPA epoxies also have readily abstractable hydrogens positioned on carbon
directly attached to ether and alcohol oxygen. Accordingly, BPA epoxies are expected to
be susceptible to oxidative degradation, as evidenced by studies on model compounds [1].
Coatings made using aromatic isocyanates yellow badly after only short exposures to
UV radiation. Coatings based on BPA epoxies generally chalk rapidly on exposure outdoors. Since ketones absorb UV, they should be avoided. When acrylic resins are polymerized in ketone solvents such as methyl amyl ketone, ketone groups can be incorporated in
the resin by chain transfer [2]; to avoid introduction of ketone groups, it is preferable to use
polymerization solvents such as esters or toluene.
Highly chlorinated resins such as vinyl chloride copolymers, vinylidene chloride
copolymers, and chlorinated rubber degrade by autocatalytic dehydrochlorination on
exposure to either heat or UV. They must be formulated with stabilizers, as discussed
in Section 5.3.
5.2. PHOTOSTABILIZATION
A major step toward formulation of a coating with exterior durability is to start with
binders that resist photoxidation and hydrolysis. Substantial further improvement can
often be achieved by use of photostabilizers. Defenses against photoinitiated oxidative
degradation include UV absorbers to reduce UV absorption by the polymer (Scheme 5.1,
Eq. 5.1), excited state quenchers to compete with bond cleavage of P (Scheme 5.1,
Eq. 5.2), and antioxidants to reduce oxidative degradation (Scheme 5.1, Eqs. 5.3 and
5.4). Review articles and books on photostabilization and thermal stabilization of coatings,
including degradative pathways, are available [3].
5.2. PHOTOSTABILIZATION
101
Scheme 5.3
5.2.1. UV Absorbers and Excited State Quenchers
Important characteristics of UV stabilizers—both absorbers (A) and quenchers (Q)—are
photostability and chemical and physical permanence. Photostability requires that the photoexcited stabilizer (A or Q ) can return to the ground state by converting the UV energy into
thermal energy, as shown in Scheme 5.3. This process generally occurs by reversible intramolecular hydrogen transfer or E–Z (cis–trans) isomerization of double bonds.
An important feature of excited state quenchers is their effective quenching volume,
within which quenching of photoexcited polymers occurs efficiently. Effective quenching
volume depends on the mechanism of energy transfer. These considerations have been discussed and related to the prospects for stabilizing aromatic polymers [4]. Stabilization by
UV absorbers requires strong absorption in the wavelength region in which the polymer
and/or trace impurities also absorb. Stabilization by quenchers require strong absorption
in the wavelength region in which the excited state polymer and/or impurities emit radiation. Stabilizers, which function as UV absorbers in some coatings, may perform as
excited state quenchers in other coatings. The most effective stabilizers probably function
by both roles.
One cannot eliminate UV absorption by the resin by adding a UV absorber, which can
only reduce absorption by the binder to slow the rate of photodegradation reactions. Since
absorption increases as the path length increases and incident radiation is attenuated in the
lower layers by absorption in the upper layers, UV absorbers, as well as quenchers, are
most effective in protecting the lower parts of a film or substrate (e.g., a base coat,
wood, or plastic under a clear top coat containing an absorber) and least effective in protecting the layer at the air interface. It follows that the thickness of a clear top coat, which
contains a UV absorber/quencher, can be an important variable affecting the protection of
base coats or plastics under them, since thicker films transmit less radiation. A critical consideration in the design and selection of UV absorbers/quenchers is their absorption
spectra. In general terms, one would like to have very high absorption of UV radiation,
from 290 to 380 nm. To avoid color effects by an absorber or quencher, ideally, there
would be no absorption above 380 nm.
102
EXTERIOR DURABILITY
Substituted 2-hydroxybenzophenones, 2-(2-hydroxyphenyl)-2H-benztriazoles, 2-(2hydroxyphenyl)-4,6-phenyl-1,3,5-triazines, benzylidenemalonates, and oxalanilides are
classes of UV stabilizers. Specific members of these classes may act as UV absorbers,
as excited state quenchers, or as both, but they are frequently called UV absorbers
(UVA) in the coatings literature.
These UV stabilizers convert UV energy into heat by intramolecular hydrogen transfer
or cis – trans isomerization. For example, UV absorption or excited state quenching by a
2-hydroxyphenyl-substituted stabilizer yields a photoexcited state that converts excess
electronic energy into chemical energy by undergoing intramolecular hydrogen transfer
to yield an unstable intermediate. The unstable intermediate spontaneously undergoes
reverse hydrogen transfer to regenerate the UV stabilizer with conversion of the chemical
energy into heat. The process is illustrated with 2-hydroxybenzophenone.
The UV stabilizer must be soluble in the coating film. (Pigments that serve as UV
absorbers, discussed in Section 5.2.4, are insoluble.) Several grades of the various stabilizers are available with different substituents on the aromatic rings that provide for solubility in different polymer systems. Commonly, the stabilizer is added to the top coat of
a multicoat system. However, especially in baking systems, migration may result in the
stabilizer being distributed through the whole coating, reducing the concentration in the
top coat. This effect has been demonstrated by analyzing sections through a film of
clear coat –base coat automotive finishes [5]. In one combination of coatings, in which
UV stabilizer was added only to the clear coat, the content through the whole film, both
clear coat and base coat, was essentially uniform. In a second case, with a different
5.2. PHOTOSTABILIZATION
103
type of base coat, a major fraction of the stabilizer stayed in the clear coat. In coatings for
plastics, it has been shown that stabilizers can migrate into the plastic [6].
A critical requirement of a UV stabilizer is permanence. Loss of stabilizer can result
by two mechanisms. There can be physical loss by vaporization, leaching, or migration
and/or there can be chemical loss by deterioration of absorbance by photochemical reactions
of the stabilizer. If a UV stabilizer has even a small vapor pressure, it slowly volatilizes from
the surface over the long term period for which durability is desired. Analysis of one UV
stabilizer as a function of depth into the film initially and after one year exterior exposure
[on a Florida black box (Section 5.6.1)] showed significant losses, particularly near the
surface of the film [5]. Most hydroxyphenyltriazines have very low vapor pressures and
show the best photo-permanence, followed by benzotriazoles, benzophenones, and oxalanilides. Longer term physical permanence may be achieved by using oligomeric photostabilizers. Polymer-bound stabilizers can also be used assuring physical permanence.
UV stabilizers can degrade chemically and lose effectiveness. The degradation presumably proceeds via free radical intermediates, and therefore degradation proceeds most
rapidly in binders that are most readily photodegraded by a free radical mechanism.
(The complex factors involved in permanence are discussed extensively in Ref. [7].)
5.2.2. Antioxidants
Antioxidants may be classified into two groups: preventive and chain-breaking antioxidants. Preventive antioxidants include peroxide decomposers, which reduce hydroperoxides to alcohols and become oxidized into harmless products. Examples of peroxide
decomposers are sulfides and phosphites that are initially oxidized to sulfoxides and phosphates, respectively, as shown in Eqs. 5.5 and 5.6 for dilauryl thiodipropionate (LTDP)
and triphenylphosphite. Further reactions may occur.
ð5:5Þ
ð5:6Þ
Metal complexing agents are a second type of preventive antioxidant. They tie up transition metal ions that may be present as contaminants and that catalyze the undesired conversion of hydroperoxides into peroxy and alkoxy radicals by redox reactions, as exemplified in
Scheme 5.4. The resulting peroxy and alkoxy radicals are undesirable because they promote
oxidative degradation, as shown in Scheme 5.1. For oxidative cure of drying oils and alkyds
(Section 14.2.2), transition metal ions are deliberately added as catalysts (driers) to accelerate
autoxidation. Thus, the reactions in Scheme 5.4 are involved in both oxidative cross-linking
and oxidative degradation. Accordingly, drier concentrations should be minimized to reduce
subsequent degradation during exterior exposure.
Scheme 5.4
104
EXTERIOR DURABILITY
The tetrafunctional bidentate imine derived from o-hydroxybenzaldehyde (salicylaldehyde) and tetraaminomethylmethane is reported to complex a large number of transition
metal ions effectively, including Co, Cu, Fe, Mn, and Ni (represented by Mnþ in
complex 1) [8].
Chain-breaking antioxidants function by interfering directly with chain propagation
steps of autoxidation, shown in Scheme 5.1. An example is a hindered phenol,
2,20 -methylenebis(4-methyl-6-tert-butylphenol), which reacts with peroxy radicals in competition with hydrogen abstraction from a polymer (PH) to yield a resonance-stabilized, less
reactive phenoxy radical:
ð5:7Þ
Note that hydroperoxides are generated in Eq. 5.7; this is a basis for synergistic stabilization of polymers by a combination of peroxide decomposers and chain-breaking phenolic
antioxidants. Synergistic stabilization means that the combination of stabilizers is more
effective than the additive effect of each stabilizer by itself.
5.2.3. Hindered Amine Light Stabilizers
Hindered amine light stabilizers (HALS) are amines with two methyl groups on each of
two alpha carbons; most are derivatives of 2,2,6,6-tetramethylpiperidine, as shown in
general formula 2. They are reported to function both as chain-breaking antioxidants
[9] and transition metal complexing agents [10]. The former function appears to be
most important. Note that the 2,2,6,6-methyl groups prevent oxidation of the ring
carbons attached to the nitrogen.
5.2. PHOTOSTABILIZATION
105
Scheme 5.5
HALS derivatives undergo photoxidative conversion into nitroxyl radicals (R2NO.)
that react with carbon-centered radicals by disproportionation and combination to yield
corresponding hydroxylamines and ethers, respectively, as shown in Eq. 5.8 in Scheme
5.5. The hydroxylamines and ethers, in turn, react with peroxy radicals to regenerate
nitroxyl radicals, shown in Eq. 5.9 in Scheme 5.5. In this manner, HALS derivatives interfere with propagation steps involving both carbon-centered and peroxy radicals in autoxidation (Scheme 5.1). In contrast to nitroxyl radicals, hindered phenols do not react
with carbon-centered radicals. The chain breaking antioxidant activity of hindered
phenols is limited to reaction with peroxy radicals, and oxygen is required as a costabilizer.
To be effective, HALS derivatives must undergo rapid photoxidation to form nitroxyl
free radicals on exterior exposure of the coating. In a photoexposed film, only a small fraction (about 1%) is present as the nitroxyl radical, the major storage components being the
corresponding hydroxylamine (R2NOH) and ethers (R2NOP). Continued stabilization
requires the presence of the nitroxyl radical, and its disappearance is followed shortly
by rapid polymer degradation. Probably, the ultimate demise of HALS occurs, at least
in part, by oxidation of the nitroxyl radical, accompanied by opening of the piperidine
ring. Transition metal ions and peracids are potential oxidants for this process.
A variety of HALS compounds is available. The R in general formula 2 is often a
diester group that joins two piperidine rings; this increases the molecular weight, decreasing volatility. The first commercial HALS compounds, still used to a degree, had R0 ¼ H.
Later versions with R0 ¼ alkyl exhibit better long-term stability. Both of these types are
basic and interfere with acid catalyzed cross-linking reactions, such as those involving
melamine – formaldehyde resins. Alkanoyl HALS compounds with R0 ¼ C(55O)R0 are
not basic, but their initial reaction to form nitroxyl free radicals is slower (Section
5.6.2) [3(c)]. More recently, hydroxylamine ethers (R0 ¼ OR00 ) have gained wide acceptance. The octyl ether provides a HALS compound with low basicity that converts
rapidly to nitroxyl free radicals [11]. HALS compounds, especially with R0 ¼ H, can
accelerate degradation of polycarbonate plastics, possibly by base catalyzed hydrolysis.
Combinations of UV absorbers and HALS compounds can act synergistically [3(c),
12]. The UV absorber reduces the rate of generation of radicals, whereas the HALS compound reduces the rate of oxidative degradation by the radicals. A further factor is that UV
absorbers are inefficient at protecting the outer surface of a film; in contrast, HALS compounds can effectively scavenge free radicals at the surface. Analysis of films after exterior
exposure shows that significant amounts of HALS derivatives remain after two years black
box Florida exterior exposure [5] (Section 5.6.1).
For automobile use, a base coat and a clear coat are applied over a primed TPO
(polypropylene–EPDM rubber blend). For example, an acrylic–MF base coat and an
acrylic polyurethane clear coat could be applied. A UV absorber and HALS compound are
included in the clear coat to assure exterior durability. However, in contrast to the same
coating applied over steel, it has been found that the UVA and HALS can migrate through
the coatings and into a TPO plastic resulting in poorer durability than over steel. To minimize
106
EXTERIOR DURABILITY
this problem, the use of UVAs and HALS with hydroxyl groups that will react with the isocyanate in the clear coat to bind the stabilizers to the polymer chains has been studied. With
the stabilizers bonded, retention of stabilizers in the clear coat is much higher and durability is
improved [13]. An extensive study of the use of similar reactive UVAs and HALSs using an
acrylic MF base coat and a 2K urethane clear coat also showed marked improvement in performance as compared to unreactive UVAs and HALSs [14].
Wood discolors rapidly on outdoor exposure to UV, as well as visible light, and more
slowly indoors. A major reason for the effect is the reactions of lignin in the wood
on exposure to the radiation. A two coat system has been recommended to protect
wood against discoloration. The first coat contained the nitroxyl derivative of
4-hydroxy-2,2,6,6-tetramethylpiperidine to trap radicals formed by the radiation. Although
as noted above, in most cases it is desirable to put the HALS in the top coat, in this case it
was most important to put it in a coat in direct contact with the lignin-containing wood. In
the second coat a mixed bis, tris, and tetra-para-substituted-tris-resorcinol-triazine was
used as the UV absorber to screen the radiation [15].
The N-2,2,6,6-pentamethylhydroxypiperidinyl diester of 4-methoxyphenylmethylenemalonic acid combines the UV absorber function of a benzylidene malonic ester and
the antioxidant properties of HALS derivatives [16]. Furthermore, on exposure of films,
this HALS derivative grafts onto the polymer chains, eliminating volatility and migration
losses. It reacts with MF cross-linked acrylic and 2K urethane acrylic coatings when
exposed to radiation, hence preventing loss by volatilization or leaching. It also absorbs
UV radiation and is used with UV stabilizer, o-hydroxyphenyl-s-triazine, to give excellent
stabilization over long periods of time [17]. The use of three types of UV stabilizers with
this HALS compound has shown that photostability diminishes more rapidly in the order
o-hydroxyphenyltriazenes . o-hydroxybenzotriazoles . oxanilides [17].
5.2.4. Pigmentation Effects
Many pigments absorb UV radiation. The strongest UV absorber known is fine particlesize carbon black. Many carbon blacks have structures with multiple aromatic rings and
in some cases, phenol groups on the pigment surface. Such black pigments are both UV
absorbers and antioxidants. Enhanced exterior durability is obtained with carbon black
pigmented coatings. When Henry Ford said, “The customer can have any color car he
wants, as long as it is black,” he was not just being crotchety. Black was by far the
most durable paint available at the time.
5.2. PHOTOSTABILIZATION
107
Other pigments absorb UV radiation to varying degrees. For example, 50-mm coatings pigmented with fine particle size, transparent iron oxide pigments absorb virtually all radiation
below about 420 nm [18]. This strong absorption is particularly useful in stains for use over
wood, since the pigmented transparent coating protects the wood from photodegradation.
Nano-zinc zirconium silicate has been patented for use as a UV absorber in automotive
clear coats. Because of the small particle size, it does not affect the transparency of the
clear coat. It has the advantage that it in insoluble and nonvolatile [19].
Rutile TiO2 white pigment absorbs UV strongly. Absorption is a function not only of
wavelength and concentration, but also of the particle size of the pigment [20]. The
optimum particle size for absorption of UV by rutile TiO2 increases from 0.05 mm for
300 nm radiation to 0.12 mm for 400 nm radiation. This size is smaller than the optimum particle size of 0.19 mm for hiding (Section 19.2.3). Rutile TiO2 with an average particle size of
0.23 mm still absorbs UV strongly. Anatase TiO2 also absorbs UV strongly, although not as
strongly as rutile TiO2 in the near UV. Thus, TiO2, especially rutile TiO2, functions as a UV
absorber in coatings. However, TiO2 can accelerate photodegradation of films on exterior
exposure, causing chalking of coatings: that is, degradation of the organic binder and
exposure of unbound pigment particles on the film surface that rub off easily, like chalk
on a blackboard. Degradation of the binder is enhanced by interaction of photoexcited
TiO2 with oxygen and water to yield oxidants, as shown in Scheme 5.6 [21]. Photoexcitation
of TiO2 results in promoting a low energy valence band electron into the higher energy conduction band, creating a separated electron (e)/hole (p) pair, signified by TiO2 (e/p) in
Scheme 5.6. Electron capture by O2 (reduction) and hole capture by H2O (oxidation)
result in regeneration of ground-state TiO2 and lead to hydroperoxy and hydroxy radicals
that can participate in oxidative degradation, as shown in Scheme 5.1. Anatase TiO2 is
more active in promoting oxidative degradation than rutile TiO2, so much so that anatase
is generally not used in outdoor coatings except in rare cases where chalking is desired.
The photoactivity of TiO2 pigments is reduced by coating the pigment particles with a
thin layer of silica and/or alumina (Section 20.1.1) to form a barrier layer against the
redox reactions. Treated rutile pigments are available that accelerate chalking to very
minor extents, and these grades are used in exterior coatings. A laboratory test has been
developed to compare the photoreactivity of various grades of TiO2 [22]. Various stabilizing additives, including HALS, have also been reported [23].
Scheme 5.6
108
EXTERIOR DURABILITY
Chalking of exterior house paints eventually leads to complete erosion of the film.
Chalking also leads to reduction in gloss, since the film becomes rougher. However, loss
of gloss does not necessarily correlate with ease of chalking [24]. It was shown that
initial gloss loss in some TiO2 pigmented coatings resulted from film shrinkage, which
in some cases was greater with more resistant grades of TiO2. In paints containing both
TiO2 and color pigments, chalking results in color changes as a result of the gloss loss;
the higher surface reflectance of the low gloss films gives an appearance of weaker colors.
Formulating coatings for exterior use can be complicated by pigment –binder interactions. In some cases, a pigment shows excellent color retention after exterior exposure
when formulated with one class of resins, but poor durability when formulated with
another class. For example, thioindigo maroon had excellent color retention when used
in nitrocellulose lacquers, but poor color stability when used in acrylic lacquers. Although
one can use experience in other systems as an initial basis for selecting pigments for use
with a new class of resins, field tests are required to assure that each combination of
pigments and resins is suitable.
5.3. DEGRADATION OF CHLORINATED RESINS
Highly chlorinated resins undergo dehydrochlorination on exposure to either heat or UV.
The reactions are autocatalytic; with poly(vinyl chloride) the ultimate products are
Scheme 5.7
5.4. HYDROLYTIC DEGRADATION
109
conjugated polyenes. As the number of conjugated double bonds in series increases, the
polymer progressively discolors, finally becoming black. The resulting highly unsaturated
polymer undergoes autoxidation, resulting in cross-linking and embrittlement.
The mechanism of degradation of poly(vinyl chloride) has been studied extensively.
Dehydrochlorination is promoted when chlorine is situated on a tertiary carbon or on an
allylic carbon. It has been proposed that at least one major weak point results from addition
of a vinyl chloride monomer in a head-to-head fashion, as in Eq. 5.10 in Scheme 5.7, to the
growing polymer chain, followed by chain transfer of Cl to monomer as in Eq. 5.11 [25].
The resulting allylic chloride is highly susceptible to dehydrochlorination, which generates a new allylic chloride with two conjugated double bonds (Eq. 5.12). Progressive dehydrochlorination is favored because the growing number of conjugated double bonds
increases the lability of allylic chlorides down the chain.
A variety of stabilizing agents are used. Since dehydrochlorination is catalyzed by
hydrogen chloride, HCl traps such as epoxy compounds and basic pigments are useful.
Diels –Alder dienophiles can act as stabilizers; the Diels– Alder addition breaks up the
chain of conjugated double bonds. Dibutyltin diesters are effective stabilizers. It is proposed that the activated chlorine atoms are interchanged with ester groups of the tin compounds to form the more stable ester-substituted polymer molecules. Dibutyltin maleate is
a particularly effective stabilizer, since it acts as both an ester interchange compound and a
dienophile. Barium, cadmium, and strontium soaps act as stabilizers. Choice of stabilizer
combinations can be system specific, especially depending on whether stabilization is
needed against heat, UV, or both. In the case of UV stabilization, UV absorbers can
further enhance stability.
5.4. HYDROLYTIC DEGRADATION
A general ordering of functional groups subject to hydrolysis is esters . carbonates .
ureas . urethanes ethers, with esters being the most vulnerable and ethers the least.
However, activated ethers, such as in etherified melamine – formaldehyde (MF) crosslinkers (Chapter 11), and etherified resole phenolics (Section 13.6), as well as their crosslinked products with hydroxy-functional resins, are more reactive than ureas and
urethanes. The tendency of each type of group to hydrolyze can be reduced by steric hindrance: for example, by placement of alkyl groups in the vicinity of the susceptible groups
such as esters (Section 10.1.1). The alkyl groups may also reduce hydrolysis by decreasing
the solubility of water, hence reducing rates of hydrolysis. Studies have shown that the
lower the water solubility of the diacid or diol used to make a polyester, the greater the
resistance to hydrolysis [26]. Rates of hydrolysis are also influenced by neighboring
groups. For example, phthalate half esters, in which the groups are ortho, are more
readily hydrolyzed under acidic conditions (often encountered during exterior exposure)
than are isophthalate half esters, in which the groups are meta (Section 10.1.2). Hydrolysis
of polyesters results in backbone degradation. On the other hand, the backbones of
(meth)acrylic resins are completely resistant to hydrolysis, since the linkages are
carbon – carbon bonds. Acrylate, and particularly methacrylate, ester side groups are
very resistant to hydrolysis, owing to the steric effects of the acrylic backbone. Carbonate
esters are more stable to hydrolysis than are carboxylic esters.
Hydrolysis of MF cross-linked hydroxy-functional resins is enhanced by residual acid
catalysts (generally, sulfonic acids) used to catalyze cross-linking (Section 11.3.1). Curing
temperatures can be reduced by increasing the concentration of sulfonic acid in the
110
EXTERIOR DURABILITY
coating, but the sulfonic acid remaining in the cured film enhances susceptibility to
hydrolysis. An apparently ideal solution would be to use transient, or fugitive, acid catalysts, which either leave the film or become neutralized after cure. Degradation of acrylic –
MF coatings at 508C and different humidity levels was studied with infrared spectroscopy
(FTIR) and atomic force microscopy (AFM). Formation of 22NH2 and 22COOH groups
was observed, and material was lost. It is interesting that the degradation was not
uniform; pits in the surface form, deepen, and grow during exposure [27].
Base coat –clear coat finishes for automobiles are subject to environmental etching. In
warm, moist climates with acidic rain or dew, such as Jacksonville, Florida, small,
unsightly spots appear in the clear coat surface, sometimes within days. The spots are
uneven shallow depressions in the clear coat surface. Presumably, they result from hydrolytic erosion of resin in the area of a droplet of water containing a significant acid concentration. While testing by exposure of panels for 14 weeks in Jacksonville has been widely
used to evaluate coatings, it has been shown that the variations in rain, dew, sunshine, and
acidity there cause considerable variability in results year to year [28]. An artificial weathering test is recommended. The laboratory test was continuous exposure to UV with
repeated cycles of application of spraying panels with a dilute solution of a mixture of sulfuric, nitric, and hydrochloric acids to simulate acid dew, a dry period of 9 hours at 358C
and 30% RH, another dry period at 608C and 5% RH, and a 4-hour rain period spraying
with demineralized water.
Several factors are involved in differences in resistance to environmental etching
[29,30]. Since urethane linkages are more resistant to acid hydrolysis than are the activated
ether cross-links obtained with MF cross-linked hydroxy-functional resins, generally,
urethane – polyol clear coats are less susceptible to environmental etching than many
MF– polyol clear coats. When MF resins are used to cross-link carbamate-functional
resins (Section 8.2.2), good resistance to environmental etching can be combined with
mar resistance. Temperature and Tg are also important, as is surface tension of the clear
coat. A variety of approaches have been undertaken to minimize the problem (see
Section 30.1.2 for further discussion).
Acrylic – urethane coatings are reported to be stabilized more effectively by HALS than
are acrylic –melamine coatings [31]. This may reflect, at least in part, the tendency of
acrylic –melamines to degrade by hydrolysis; HALS derivatives do not stabilize against
hydrolysis. Both urethane and melamine coatings undergo oxidative degradation. CH
groups adjacent to urethane nitrogen groups are expected to be activated toward
H-abstraction by free radicals, as are CH groups adjacent to nitrogen and ether oxygen
groups in melamines. However, evidence has been presented that hydroperoxide levels
are significantly lower in melamine than in urethane cross-linked coatings, indicative of
lower free radical formation rates. The results have been attributed to the ability of melamines to decompose hydroperoxides [32] but may also be explained by lower susceptibility to oxidation of CH groups on melamine.
Hydrolytic degradation of acrylic – urethane [31] as well as acrylic – melamine [33]
coatings are reported to be accelerated by UV exposure. This may result from the increase
of hydrophilic groups, such as hydroperoxides, alcohols, ketones, and carboxylic acids,
resulting from photoxidation, which increases the solubility of water in the coating. Photoxidation may also occur at specific sites to generate groups that are more susceptible to
hydrolysis. Photodegradation of acrylic –melamine coatings is also reported to be accelerated in high humidity [33]. This result has been attributed to conversion of formaldehyde
(from hydrolysis) into performic acid, a strong oxidant.
5.5. OTHER MODES OF FAILURE ON EXTERIOR EXPOSURE
111
Silicone coatings, which are highly resistant to photodegradation, are subject to
hydrolysis at cross-linked sites, where silicon is attached to three oxygens [34,35]. Apparently, the electronegative oxygens facilitate nucleophilic attack at Si by water (Section
16.1.2). The reaction is reversible, so the cross-links can hydrolyze and re-form. If a
silicone-modified acrylic or polyester coating is exposed to water over long periods or
is used in a climate with very high humidity, the coating can get softer. It is common to
include some MF resin as a supplemental cross-linker in the formulation. Apparently,
the MF cross-links with the acrylic or polyester are more hydrolytically stable than the
bonds between the silicone resin and the polyester or acrylic.
5.5. OTHER MODES OF FAILURE ON EXTERIOR EXPOSURE
Although exposure to UV and the hydrolytic effects of rainfall and humidity are major
causes of exterior failure of coatings, many other phenomena can occur.
Automotive finishes can undergo microcracking when there is a rapid change of temperature: for example, when a car is driven out of a heated garage on a very cold winter day. The
stress built up by the differential of coefficients of expansion of coating and steel as the temperature drops rapidly can lead to shrinkage in excess of the elongation-at-break of the coating.
When paint is applied to wood, it must be able to withstand the elongation that results
from the uneven expansion of wood grain when it absorbs moisture; otherwise, grain
cracking, cracking parallel to the grain, occurs. This failure mode can occur with interior
coatings; however, it is more likely to happen with exterior coatings, particularly with
alkyds or drying oils that embrittle with exterior exposure. As a result of their greater
exterior durability and extensibility, acrylic latex paint films seldom fail this way.
Another problem of exterior, oil-based house paints on wood siding is blistering. Blistering results from the accumulation of water in the wood beneath the paint layer. The vapor
pressure of the water increases with heating by the sun, and blisters form to relieve the
pressure. Since latex paints have higher moisture vapor permeability than that of oil-based
paints, the water vapor can pass through a latex paint film, relieving the pressure before blisters grow. However, the high moisture vapor permeability of latex paint films can lead to
failures of other types. For example, if calcium carbonate fillers are used in an exterior latex
paint, frosting can occur. Water and carbon dioxide permeate into the film, dissolving
calcium carbonate by forming soluble calcium bicarbonate, a solution of which can
diffuse out of the film. At the surface of the film, the equilibrium changes direction, and
the calcium bicarbonate is converted back to a deposit of calcium carbonate.
Dirt retention can be a difficult problem with latex house paints. Latex paints must be
designed to coalesce at relatively low application temperatures. At warmer temperatures,
soot and dirt particles that land on the paint surface may stick tenaciously and not be
washed off by rain. This problem is minimized with low gloss paints because the high
pigment content increases the viscosity of the paint film surface. The problem is particularly
severe in areas where soft coal is burned. As would be expected, dirt pickup is less for paints
formulated with higher Tg polymers [36]. It was also shown that the more hydrophobic
styrene–acrylic latex paints picked up less dirt than do acrylic latex paints having the same
Tg. Surface tension is also a factor; for example, 2-ethylhexyl acrylate (EHA) as a co-monomer
reduces dirt retention [37]. Cross-linking of functional latexes also reduces dirt pickup.
Growth of fungi (mold and mildew) can lead to blotchy, dark deposits on the surface of
coatings. Retardation of mildew growth is discussed in Section 32.1.
112
EXTERIOR DURABILITY
Often, one needs to be a detective to determine the origin of a paint failure. For
example, there were suddenly numerous reports of ugly dirt spots on houses in Bismarck,
North Dakota. It turned out that the flight pattern of airplanes taking off from the Bismarck
airport had been changed. Oily droplets from plane emissions were landing on the paint,
softening its surface to make it a good adhesive for dirt. The wide range of things that can
happen to coatings outdoors makes the problem of predicting performance from the results
of any simplified test conditions difficult.
5.6. TESTING FOR EXTERIOR DURABILITY
Unfortunately, no test is available that reliably predicts the exterior durability of all coatings. The best way to determine whether a coating will be durable for some length of time
in a particular environment is to apply the coating to the product and use the coated
product in that environment to determine its lifetime. Even this is difficult, due to the
wide variety of environments and variability in coatings and application conditions.
The limitations of accelerated tests, the need for data based on actual field experience,
and methods of building a database are described in the introduction to Chapter 4 and
more extensively in Ref. [38]. Bauer [39] recommends use of reliability theory using statistical distribution functions of material, process, and exposure parameters for predicting
exterior durability of automotive coatings. Valuable data can be obtained by analysis of
customer complaints; they are especially useful for defining performance standards. By
analyzing field experience and customer complaint data over many years, coatings and
automobile manufacturers have accumulated an increasingly useful data bank. The information can be used to test hypotheses relating structure to performance. Established
structure – performance relationships are useful starting points for designing new coatings.
The data can also be used to evaluate laboratory tests. Similar databases are accumulated
by some manufacturers and users of maintenance paints, marine coatings, and coil coatings for exterior siding.
5.6.1. Accelerated Outdoor Testing
Accelerated test methods are desirable to permit prediction of performance in shorter
times than possible by actual field use. Reference [40] provides an analysis of various
test methods. The most reliable accelerated tests are outdoor fence exposures of coated
panels, especially if they are carried out in several locations with quite different environments. Various means of exposure and testing of exposed panels are reviewed in Ref. [41].
Even at one location, there are sometimes substantial variations in conditions that lead to
variability of test results [38,39,42]. The most commonly used exposure for architectural
paints and industrial coatings is in southern Florida with panels facing south at a 458 angle;
automotive coatings are commonly tested at 58 from horizontal. For coil coatings, a major
supplier exposes at 908 for sidewall panels and 458 for roofing panels [43]. It is important
to maintain records of the weather throughout the exposure period. Modern practice is to
use photocells to record the cumulative amount of radiant energy striking a test specimen.
Since different colors absorb different amounts of infrared radiation, the maximum temperatures reached by coatings during exposure depends on their color; this difference can
affect exposure test results.
5.6. TESTING FOR EXTERIOR DURABILITY
113
Southern Florida has a subtropical climate with high humidity, temperature, and sunshine levels. Arizona has more hours of sunshine per year and a higher average daily
high temperature but lower humidity. Martin [44] presents a review of the differences
between exposure conditions in Florida and Arizona. Cities in the northeastern states
and in northern Florida have higher levels of acid in the atmosphere; the Denver area
has a higher UV intensity because of the high altitude and a greater temperature change
from day to night. Bauer estimates that for automotive coatings, a lower humidity in
Arizona approximately offsets the higher UV intensity and temperature, and that one
year of exposure in Florida has a comparable effect to two years in Michigan, where
the UV intensity and temperature are lower but the humidity is higher [39]. The most
severe conditions are found in humid equatorial locations (e.g., Venezuela).
Test specimens are examined periodically to compare their appearance before and after
exposure. Usually, at least part of the coating surface is cleaned for the comparison. Ease
of cleaning, change in gloss, change in color, degree of chalking, and any gross film failures are reported. Changes in color are particularly difficult to assess, since change in gloss
or chalking changes the color even if there has been no change in the color of the components of the coating. The effects of changes in gloss and chalking on the color of the
components can be minimized by cleaning the panels and then coating part of each
panel with a thin layer of clear gloss lacquer. The lacquer layer minimizes the effect of
differences in surface reflection on the color. Test fence exposures can eliminate some
formulations as inadequate after a few months of exposure; however, two to five years
may be required to permit one to conclude that resistance to exposure at that location
under those conditions is adequate.
Film degradation can be accelerated by using black box exposure. Panels are mounted
at 58 to the horizontal on black boxes rather than with the backs of the panels open to the
air. This method is widely used for testing automotive top coats. A substantial increase in
the temperature of the coating when the weather is sunny accelerates degradation. The
temperature increase, and therefore the extent of acceleration, can vary substantially
from coating to coating, especially with color. A variable that can affect the extent of
change due to black box exposure is the relationship between the temperatures and the
Tg of the film. Coatings that will seldom be exposed to direct sunlight (indoor architectural
paints, auto interior paints) are tested outdoors under glass that absorbs UV similarly to
absorbtion by window or auto glass.
Loss of gloss was formerly the primary mode of failure of automotive top coats, but
since the adoption of clear coat – base coat finishes, a major mode of failure has become
clear coat cracking in combination with delamination between top coats and base coat,
primer, or plastic substrate [39]. Application of sufficient film thickness and proper
choice of UV stabilizers and HALS compounds can be critical to avoiding delamination.
Results can be obtained in substantially shorter times by using Fresnel reflectors to concentrate sunlight on test panels. High intensity is achieved by reflecting sunlight from
moving mirrors that follow the sun to maintain a position perpendicular to the sun’s
direct beam radiation. Fresnel reflector devices with trade names such as EMMAQUA
(Equatorial Mount with Mirrors for Acceleration plus water) FRECKLE, and Sun-10
are operated in the Arizona desert [44]. They enhance the intensity of sunlight on the
panel surfaces by a factor of 8 over direct exposure; it is said to accelerate degradation
rates 4 to 16 times the rate for nonaccelerated exposure [41]. Panels are cooled by
blowing air both over and under the samples. Since rainfall and humidity in the desert
are low, the test facilities permit periodic spraying of water on the panel surfaces. In
114
EXTERIOR DURABILITY
any highly accelerated test, it must be remembered that the rate of volatilization of UV
stabilizers depends more on temperature than on radiant energy intensity; stabilizer
losses may be smaller during accelerated testing than during long term exposure.
Fresnel reflector weathering devices record actual exposure energies per unit area,
expressed in MJ m22. For comparison, the results can be approximately translated into
Arizona or Florida exposure times based on relative solar intensity (i.e., energy per unit
time). Comparisons based on intensity of exposure to a 10 nm band of UV around
313 nm give results that correlate most closely with field exposures. In addition, spraying
with water at night has been found to give results that more closely correspond to field
exposure than spraying only during the daylight hours, as was done in earlier procedures.
5.6.2. Analysis of Chemical and Mechanical Changes
Analysis of chemical changes occurring in coatings that have been subjected to exterior
exposure is a powerful tool. Chemical changes become detectable long before the physical
changes (gloss loss, cracking, delamination, etc.) that characterize failure become evident.
Analysis can permit early detection of impending failure after relatively brief exposure,
shortening evaluation time for new coatings. However, to have predictive value, chemical
changes must be correlated with actual field experience for each type of coating, and only a
few correlations have been reported. In addition to predicting failure, studies of chemical
changes can help determine the mechanisms of failure, providing a basis for formulating
more resistant coatings. Various approaches are reviewed in Refs. [11] and [40]. Early
mechanical changes, measured by scanning electron microscopy (SEM) or by dynamic
mechanical analysis (DMA), can also be correlated with chemical changes and with
field experience [45].
Coatings, in general, embrittle on exposure, with an increase in Tg making them more
subject to cracking; but it has been pointed out that the degree to which stress builds up in
films is also an important factor in determining whether a film will crack [46]. If no stress
were to build up, even a brittle film would not fail. But other factors are also involved. Flat
surfaces will have lower stress buildup than will curved surfaces. Thin clear coats build up
less stress than thick coats but offer less protection of base coats, owing to reduced UV
absorption. Humidity and thermal changes can result in stress buildup, owing to expansion
and contraction as a consequence of temperature and/or humidity increase or decrease,
respectively. Repeated cycles of temperature and humidity changes can lead to fatigue.
Overbaking can increase stress. (See Section 4.3.1 for a discussion of the effects of fracture
mechanics on cracking of weathered clear coats.)
Electron spin resonance spectrometry (ESR) can monitor changes in free radical concentrations within a coating. The rate of disappearance of stable nitroxyl radicals has been
correlated with loss of gloss in long-term Florida exposure [12]. In this technique, nitroxyl
radical precursors are incorporated into TiO2 pigmented acrylic – melamine coatings subjected to ambient or accelerated short-term UV exposure. Similarly, the rate of buildup of
phenoxy radicals from bisphenol A epoxy resins incorporated into acrylic – melamine
coatings exposed to UV radiation in an ESR spectrometer cavity has been correlated
with cracking of coating films exposed in a QUV weathering device [47]. The latter technique was used to evaluate UV stabilizers.
Use of ESR spectrometry to monitor the rate of disappearance of nitroxyl radicals in
acrylic –melamine coatings allows calculation of photoinitiation rates (PR) of free
5.6. TESTING FOR EXTERIOR DURABILITY
115
radical formation, which were found to correlate with rates of gloss loss (GLR):
GLR / (PR)1/2. This proportional relationship of GLR with (PR)1/2 is consistent with
a free radical process and results from termination by second-order radical – radical reactions [12]. Photoinitiation rates, determined by this method, have also been used to evaluate experimental conditions for the synthesis of acrylic polyols by free radical
polymerization, including the effect of initiator, temperature, and solvent on the projected
exterior durability of the resulting acrylic –melamine coating [48].
The nitroxyl early detection method has also been used to investigate the synergistic
stabilizing effect of a benzotriazole UV stabilizer and a HALS derivative in acrylic –
urethane coatings. The UV stabilizer reduces the photoinitiation rate (PR) of free
radical formation, whereas the HALS derivative reduces the propagation rate by lowering
the concentration of free radicals.
A somewhat more direct application of ESR to studying photostability of coating films
is the determination of free radical concentration after UV irradiation of films at a temperature of 140 K, well below Tg [49]. Under these conditions, free radicals are stable. The
method is appropriate for evaluation of stabilizers such as HALS compounds by comparing the radical concentrations with and without stabilizer. Useful comparisons can be
made in 3 hours [12].
ESR instruments are expensive; an alternative approach is to grind exposed films cryogenically to a fine powder and measure the buildup of hydroperoxides by titration with
periodate. A linear relationship between photoxidation rates (PR) and hydroperoxide concentrations in acrylic – urethane and acrylic –MF coatings was reported [48,50]. The
relationship GLR / (PR)1/2, established by ESR as shown previously, was assumed to
apply to data obtained by titration.
Fourier transform infrared spectroscopy (FTIR) is another powerful tool for following chemical changes on a surface [42]. Photoacoustic FTIR has also been used and
has the advantage that the sample does not have to be removed from the substrate.
Furthermore, not only the surface, but also different depths within a film can be
analyzed [51].
A study of a variety of exposure tests and analytical tools to give early predictions of
gloss retention of seven latex paints with known exterior durability has been reported.
The most useful results were obtained with x-ray photoelectron spectroscopy (XPS).
XPS was used to analyze the surface of films for titanium, aluminum, and silicon in comparison with polymeric carbon before and after exterior exposure. (The aluminum and
silicon are from the surface treatment of the TiO2 in the paint.) Exposure leads to the
degradation of polymer from the film surface, increasing the ratio of the metals to the
polymeric carbon. Initially, surfactant bloom to the surface interfered with the analysis;
but this disappeared after 2 weeks exposure. It was found that results obtained after 12
weeks Florida exposure (after the surfactant effect was gone) gave good predictions of
gloss loss after 96 weeks of Florida exposure [52]. Atomic force microscopy (AFM)
has been recommended for studying mechanical and chemical changes in the surface
of films over time [53].
A study of changes in clear coats after five years of exterior exposure has led to the
suggestion that four nontraditional tests should be added to the usual tests for performance:
UVA concentration retention, HALS retention at the film surface and in the film bulk,
FTIR analysis of the ratio of (22OH, 22NH)/22CH peak areas, and clear coat fracture
energy [54].
116
EXTERIOR DURABILITY
A variation on accelerating degradation of a binder in TiO2 pigmented films is to substitute the surface-treated rutile TiO2 that would be used in a commercial coating with
untreated, photoactive anatase TiO2. Similarly, one can study delamination failures of
automotive coatings by exposing panels with no UV stabilizer in the clear coat over
various substrates [38]. The results could permit one to decide whether a major effort
should be put into improving the resistance of the substrate to photoxidation or aid in predicting the required minimum film thickness of a clear coat containing a UV stabilizer to
protect the substrate.
5.6.3. Accelerated Weathering Devices
Many laboratory devices for accelerating degradation are available. See Ref. [55] for
descriptions of various devices and their advantages and disadvantages. The various
devices expose panels to UV sources with different wavelength distributions. Also, the
panels are subjected to cycles of water spray (or high humidity). Although these tests
are widely used, results frequently do not correlate with actual exposure results. A
general problem with accelerated weathering methods is the difficulty in accelerating
the effects of radiation, heat, and moisture uniformly, not to mention the effects of
other atmospheric degradants. The predictive value of accelerated weathering with artificial light sources is particularly questionable when a light source includes wavelengths
below 290 nm. Variability of performance of the test instruments can also be a
problem, especially when comparing results from laboratory to laboratory [42]. Typically,
a number of specimens are tested at the same time, and the device geometry must expose
them all to the same level of radiation.
An evaluation of accelerated weathering devices available in 1985 for a polyester –
urethane coating using photoacoustic –FTIR spectroscopy concluded that none of the conventional devices were suitable, including a carbon arc device with Corex D filters, a
xenon arc device with borosilicate inner and outer filters, and a UV device with FS-40
fluorescent bulbs [56]. All artificial light sources resulted in the loss of isophthalate
groups in the coating, which was not observed during accelerated weathering in Florida
(58 south) or in Arizona (EMMAQUA). These results correlate with the general wisdom
that polyesters, presumably with phthalate or isophthalate groups, perform worse than
acrylics in accelerated weathering devices relative to their performance in Florida or
Arizona exposure. The unnatural weathering that resulted in loss of isophthalate groups
was attributed to excessive amounts of short wavelength light (lower than 290 nm)
from the artificial light sources.
A comparison of spectra of sunlight (Miami average optimum) and artificial weathering
devices that illustrates this disparity is provided in Figure 5.1. As shown, the FS-40 fluorescent bulbs utilized in Atlas UVCON and Q-Panel QUV devices exhibit strong unnatural
emission below 300 nm.
Many examples of reversals of results comparing coatings with known exterior durability with laboratory tests have been found. For example, poor results in early evaluations of HALS compounds by QUV testing might have led to their abandonment.
Fortunately, remarkable outdoor durability results were obtained, which prompted continued development. Another example is that some automotive base coat– clear coats
containing alkanoyl HALS (Section 5.2.3) performed very well in QUV tests but had
poor resistance to delamination in the field. It is suspected that this type of HALS compound is quickly activated by short wavelength UV but that sunlight turns it on too
5.6. TESTING FOR EXTERIOR DURABILITY
117
Figure 5.1. Wavelength distribution of radiation from various sources. (Courtesy of Atlas Electric
Devices, Chicago.)
slowly. In this case, reliance on the QUV test to predict performance had severe economic
consequences. Not only are there examples of poorer coatings showing better laboratory
test results, and vice versa, but there are also disparities and reversals when comparing
various laboratory test methods. Thus, one may optimize a coating formulation for
118
EXTERIOR DURABILITY
performance in a particular weathering device only to find that it performs poorly in a
different weathering device or outdoors. More serious, as might have occurred in the original work on HALS compounds, is the possibility of rejecting a superior material on the
basis of accelerated laboratory testing. The argument that laboratory tests are at least
useful for eliminating coatings that exhibit clearly inadequate performance is specious,
since accumulated knowledge of failure mechanisms is adequate to eliminate such formulations without testing.
More recently, device manufacturers claim considerable improvement over older
devices. Filtered xenon-lamp weather test devices that closely match the short UV
content of sunlight have become available. In addition, UV-A-340 fluorescent bulbs are
available that closely approximate sunlight between 290 and 340 nm. Automotive companies favor xenon-lamp testers.
Although accelerated weathering devices have been improved by closer correspondence of UV emissions with sunlight, it appears unlikely that a completely satisfactory
device will be developed because the variations in exterior environments to which coatings are exposed are so great. As with any other highly accelerated test, stabilizer loss
by volatilization may well be insignificant during the test period but very important
over the long time periods of actual use. This situation serves to emphasize the need to
accumulate data banks of real field use results and to correlate these results with structure
to develop a basic understanding of the mechanisms of failure.
GENERAL REFERENCES
Bauer, D. R.; Martin, J. W., Service Life Prediction of Organic Coatings: Methodologies and
Metrologies, American Chemical Society, Washington, DC, 2001.
Valet, A., Light Stabilizers for Paints, translated by Welling, M. S., Vincentz, Hannover, Germany,
1997.
Wypych, G., Handbook of Material Weathering, 3rd ed., ChemTech, Northbrook, IL, 2003.
REFERENCES
1. Timpe, H.-J.; et al., Polym. Photochem., 1985, 6, 41.
2. Gerlock, J. L.; et al., Macromolecules, 1988, 21, 1604.
3. (a) Hawkins, W. L., in Encyclopedia of Polymer Science and Technology, 2nd. ed., Wiley,
New York, 1989, Vol. 15, pp. 539 –583. (b) Dexter, M., in Kirk – Othmer Encyclopedia of
Chemical Technology, 4th ed., Wiley, New York, 1992, Vol. 3, pp. 424 – 447. (c) Valet,
A., Light Stabilizers for Paints, translated by Welling, M. S., Vincentz, Hannover,
Germany, 1997.
4. Pappas, S. P.; et al., in Photodegradation and Photostabilization of Coatings, Pappas, S. P.;
Winslow, F. H., Eds., ACS Symp. Ser. 151, American Chemical Society, Washington, DC,
1981, pp. 109 – 116.
5. Bohnke, H.; et al., J. Coat. Technol., 1991, 63(799), 53.
6. Haacke, G.; et al., J. Coat. Technol., 1996, 68(855), 57.
7. Valet, A., Light Stabilizers for Paints, translated by Welling, M. S., Vincentz, Hannover,
Germany, 1997, pp. 64– 70, 116 – 124.
REFERENCES
119
8. Shelton, J. D., in Polymer Stabilization, Hawkins, W. L., Ed., Wiley-Interscience, New York,
1971, pp. 80– 84.
9. Carlsson, D. J.; et al.; Makromol. Chem. Suppl., 1984, 8, 79.
10. Fairgrieve, S. P.; MacCallum, J. R., Polym. Degrad. Stabil., 1984, 8, 107.
11. Bauer, D. R., J. Coat. Technol., 1994, 66(835), 57.
12. Gerlock, J. L.; et al., J. Coat. Technol., 1985, 57(722), 37.
13. Yaneff, P. V.; et al., JCT Res., 2004, 1, 201.
14. Cliff, N.; et al., JCT Res., 2005, 2, 371.
15. Hayoz, P.; et al., Prog. Org. Coat., 2003, 48, 297.
16. Sanduvor PR-31, tech. bull., Clairiant Corp., Charlotte, NC, 1996.
17. Avar, L.; Bechtold, K., Prog. Org. Coat., 1999, 35, 11.
18. Sharrock, R. F., J. Coat. Technol., 1990, 62(789), 125.
19. Vanier, N. R.; Munro, C. H., U.S. patent appl. 20030158316.
20. Stamatakis, P.; et al., J. Coat. Technol., 1990, 62(789), 95.
21. Voelz, H. G.; et al., in Photodegradation and Photostabilization of Coatings, Pappas, S. P.;
Winslow, F. H., Eds., ACS Symp. Ser. 151, American Chemical Society, Washington, DC,
1981, pp. 163 – 182.
22. Irick, G., Jr.; et al., in Photodegradation and Photostabilization of Coatings, Pappas, S. P.;
Winslow, F. H., Eds., ACS Symp. Ser. 151, American Chemical Society, Washington, DC,
1981, pp. 147 – 162.
23. Braun, J. H., J. Coat. Technol., 1990, 62(785), 37.
24. Braun, J. H.; Cobranchi, D. P., J. Coat. Technol., 1995, 67(851), 55.
25. Starnes, W. H., Jr., Pure Appl. Chem., 1985, 57, 1001. Georgiev, G.; et al., J. Macromol. Sci.
Chem., 1990, A27, 987.
26. Jones, T. E.; McCarthy, J. M., J. Coat. Technol., 1995, 67(844), 57.
27. Nguyen, T.; et al., J. Coat. Technol., 2003, 75(941), 37.
28. Schulz, U.; et al., Prog. Org. Coat., 2000, 40, 151.
29. Betz, P.; Bartelt, A., Prog. Org. Coat., 1993, 22, 27.
30. Gregorovich, B. V.; Hazan, I., Prog. Org. Coat., 1994, 24, 131.
31. Bauer, D. R.; et al., Ind. Eng. Chem. Res., 1988, 27, 65.
32. Mielewski, D. F.; et al., Polym. Degrad. Stabil., 1991, 33, 93.
33. Bauer, D. R., Prog. Org. Coat., 1986, 14, 193.
34. Brown, L. H., in Treatise on Coatings, Myers, R. R.; Long, J. S., Eds., Marcel Dekker,
New York, 1972, Vol. I, Part III, pp. 536 – 563.
35. Hsaio, Y.-C.; et al., J. Appl. Polym. Sci., 1975, 19, 2817. See also Pappas, S. P.; Just, R. L.,
J. Polym. Sci. Polym. Chem. Ed., 1980, 18, 527.
36. Smith, A.; Wagner, O., J. Coat. Technol., 1996, 68(862), 37.
37. Baumstark, R.; et al., Surf. Coat. Int. B, 2004, 87(5), 210.
38. Martin, J. W.; et al., Methodologies of Predicting Service Lives of Coatings Systems, Federation
of Societies for Coatings Technology, Blue Bell, PA, 1996.
39. Bauer, D. R., J. Coat. Technol., 1997, 69(864), 85.
40. Johnson, B. W.; McIntyre, R., Prog. Org. Coat., 1996, 27, 95.
41. Hicks, L. S.; Crewdson, M. J., in Paint and Coating Testing Manual, 14th ed., Koleske, J. V.,
Ed., ASTM, Philadelphia, PA, 1995, p. 619.
42. Fischer, R. M.; et al., Prog. Org. Coat., 1991, 19, 151.
120
EXTERIOR DURABILITY
43. Pilcher, G. R., J. Coat. Technol., 2001, 72(921), 135.
44. Martin, J. L., Proc. Adv. Coat. Technol. Conf. Eng. Soc., Society of Automotive Engineeers,
Detroit, MI, 1991, p. 219.
45. Hill, L. W.; et al., Prog. Org. Coat., 1994, 24, 147.
46. Nichols, M. E.; et al., Prog. Org. Coat., 1999, 35, 153.
47. Okamoto, S.; et al., in Proc. FATIPEC Congr., 1986, pp. 239 – 255.
48. Gerlock, J. L.; et al., Prog. Org. Coat., 1987, 15, 197.
49. Sommer, A.; et al., Prog. Org. Coat., 1991, 19, 79.
50. Bauer, D. R., Prog. Org. Coat., 1993, 23, 105.
51. Bauer, D. R.; et al., Polym. Degrad. Stabil., 1992, 8, 507.
52. Gebhard, M. S.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2005, pp. 69– 83.
53. Gu, X., JCT Res., 2004, 1(3), 191.
54. Gerlock, J. L.; et al., J. Coat. Technol., 2001, 73(918), 45.
55. Sherbondy, V. D., in Paint and Coating Testing Manual, 14th ed., Koleske, J. V., Ed., ASTM,
Philadelphia, PA, 1995, p. 643.
56. Bauer, D. R.; et al., J. Coat. Technol., 1987, 59(755), 103.
6
Adhesion
Adhesion is an essential characteristic of most coatings. Unfortunately, there is inadequate
basic scientific understanding of the variables affecting adhesion. As a result, some of the
comments in this discussion are based only on reasonable deductions that fit in with accumulated experience.
One difficulty in dealing with adhesion is defining what adhesion means. In most cases,
a coatings formulator thinks of adhesion in terms of the question: How hard is it to remove
the coating? But a physical chemist would think in terms of the work required to separate
two interfaces that are adhering. These perspectives can be different considerations; the
latter is only one aspect of the former. Removal of a coating requires breaking or
cutting through the coating and pushing the coating out of the way, as well as separating
the coating from the substrate.
6.1. SURFACE MECHANICAL EFFECTS ON ADHESION
Resistance to separation of coating and substrate can be affected by mechanical interlocking. Consider the schematic representations in Figure 6.1. With a very smooth interface
between coating and substrate, as shown in sketch A, the only forces holding the substrate
and coating together are the interfacial attractive forces per unit of geometric area. With
a rough surface on a microscopic scale, as represented in sketch B, two other factors are
important. In some places, there are undercuts in the substrate; to pull the coating off the
substrate, one would either have to break the substrate or break the coating to separate
them. The situation is analogous to using a dovetail joint to hold two pieces of wood
together. Another factor is that the actual contact area between the coating and the
rough substrate is substantially larger than the geometric area.
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
121
122
ADHESION
Figure 6.1. Geometries of surface interactions between a coating and a substrate. A, smooth interface
between coating and substrate; B, rough surface on a microscopic scale; C, rough surface with
incomplete penetration of coating.
Better adhesion can generally be obtained if the surface of a substrate is roughened
before coating; however, as can be seen in sketch C, surface roughness can be a disadvantage. If the coating does not completely penetrate into the microscopic pores and
crevices in the surface, dovetail effects are not realized, and the actual interfacial
contact area can be smaller than the geometric area. Furthermore, when water permeates through the film to the substrate, there will be areas of contact of water with
uncovered substrate, which can be a major problem, especially if the coating is to
protect steel against corrosion.
The importance of surface roughness is widely recognized, but its effect on coatings has
not been subjected to many scientific studies. The scale of the roughness of surfaces can
vary from macroscopic to microscopic to submicroscopic. It is important to consider the
situation on a microscopic and submicroscopic scale. What factors control the rate of penetration of a liquid into such pores and crevices? The situation is analogous to penetration
of a liquid into a capillary on the micro scale, although possibly not on the nano scale.
Equation 6.1 shows the variables affecting distance of penetration L (cm) into a capillary
of radius r (cm) in time t (s), where g is surface tension (mN m21), u is contact angle, and h
is viscosity (Pa.s):
g
L ¼ 2:24 (r cos u)t
h
1=2
(6:1)
The rate of penetration, L/t, increases with increasing surface tension of the coating.
However, there is an upper limit to this surface tension effect because the rate is strongly
affected by contact angle. The rate is fastest when the cosine of the contact angle is 1: that
is, when the contact angle is zero. The cosine can be 1 only if the surface tension of the
liquid is less than that of the solid substrate. The radius of the capillary is a variable of
the substrate, not of the coating.
Accordingly, the variable over which the formulator has greatest control is viscosity
(h). The critical viscosity is that of the continuous (external) phase of the coating, not
the bulk viscosity of the coating, including the internal phase of pigment and polymer particles, which are too large to participate, as effectively, in penetrating the microscopic and
submicroscopic surface irregularities. The lower the viscosity of the external phase, the
more rapid the penetration. Since in most cases, the viscosity of the vehicle increases
after application; it is important to keep the viscosity low for a long enough time for penetration to approach completion. Since the viscosity of resin solutions increases with molecular weight (MW), one would expect that lower MW resins would provide superior
adhesion after cross-linking, everything else being equal. This hypothesis has been confirmed in the case of epoxy resin coatings on steel [1]. Another possible advantage of
6.2. EFFECTS OF INTERNAL STRESS AND FRACTURE MECHANICS
123
low MW resins is that their molecules may be better able than high MW molecules to
penetrate smaller crevices. Coatings with low viscosity external phases, slow evaporating
solvents, and relatively slow cross-linking rates have been found, in general, to give better
adhesion. In general, baked coatings give better adhesion than do air dry coatings. (The
term air dry is used widely but is potentially confusing. It generally means that the film
is formed at ambient temperature, but does not necessarily mean that oxygen is required
for cross-linking.) When a coated article goes into an oven, the temperature increases, the
viscosity of the external phase decreases, and penetration into surface irregularities
becomes easier. This is only one of several possible explanations for the advantages of
baked coatings when adhesion is critical.
6.2. EFFECTS OF INTERNAL STRESS AND FRACTURE MECHANICS
Internal stresses in coatings amount to forces that counteract adhesion; less external
force is required to disrupt the adhesive bond. The common statement that internal
stresses result from shrinkage is misleading; rather, internal stresses result from the
inability of coatings to shrink as they form films on rigid substrates. When solvent evaporates from a thermoplastic coating (lacquer), in the early stages, the polymers can
accommodate the resulting voids by relaxation; thus, shrinkage occurs. However, as
film formation proceeds, Tg rises and free volume is reduced; it becomes more difficult
for the polymer to accommodate the voids from solvent evaporation, so it becomes fixed
in unstable conformations, and internal energy (stress) increases. This phenomenon is
particularly likely to occur in coatings in which the Tg approaches the film-forming
temperature [2]. In some cases, the stresses build up sufficiently that spontaneous delamination occurs [3]. Stresses can result not only from volume contractions, but also
from volume expansions, such as swelling of films by exposure to high humidity [3]
or water immersion [4].
In thermosetting coatings, cross-linking reactions lead to the formation of covalent
bonds that are shorter than the distance between two molecules before they react.
When such reactions occur at temperatures near the Tg of the film, stresses result
from the inability of the coating to undergo shrinkage. As the rate of cross-linking
increases, stresses also tend to increase, since less time is available for polymer relaxation to occur. An extreme example is UV curing of acrylated resins by free radical
polymerization that occurs in a fraction of a second at ambient temperatures (Section
29.2). Shrinkage, measured by thermomechanical analysis (TMA), has been shown to
lag significantly behind polymerization [5]. The high rates of polymerization, together
with the relatively large shrinkage that accompanies polymerization of double bonds,
contribute to the generally observed poor adhesion of UV cure acrylated resins to
smooth metal surfaces. Heating after UV curing relaxes the cross-link network and
often improves adhesion.
Nonuniform curing, particularly in the later stages, as well as film defects or imperfections in the film, can lead to localized stresses that exert adverse effects on adhesion [6].
Localization of stresses at imperfections falls within the discipline of fracture mechanics.
The effects of fracture on abrasion resistance are considered in Section 4.3.1. The phenomena have been studied most extensively in adhesive bonding of substrates [7], but must
also affect adhesion of coatings. If there is a local imperfection in a film, any stress
applied to that part of the film could be concentrated at the imperfection. Localized
124
ADHESION
imperfections result in greater stress (force per unit area) and greater probability of
forming a crack. Once a crack starts, the stress concentrates at the point of the crack,
leading to crack propagation. If the crack propagates to the coating – substrate interface,
stress concentration can cause the film to delaminate. (Fractures and stress fractures
within metal substrates are important considerations in adhesion and corrosion, but they
are beyond the scope of this book.)
Although the principle is widely recognized, identification of causes of imperfections is
more difficult. Pigment particles with sharp crystal corners and air bubbles are examples of
potential sites for concentration of stresses. On the other hand, incorporation of particles of
rubber may lead to dissipation of stresses (Section 4.3.1). Presumably, such stress dissipation would reduce the probability of fracture mechanical adhesive failure.
6.3. RELATIONSHIP BETWEEN WETTING AND ADHESION
Wetting is a major, and perhaps limiting, factor in adhesion. If a coating does not spread
spontaneously over a substrate surface so that there is intermolecular contact between the
substrate surface and the coating, there cannot be interactions and hence no contribution to
adhesion. The relationships between wetting and adhesion were studied extensively by
Zisman [8]. A liquid spreads spontaneously on a substrate if the surface tension of the
liquid is lower than the surface free energy of the solid. [Surface free energies of solids
have the same dimensions as surface tension (Section 24.1).] If the surface tension of a
liquid is too high, a drop of the liquid stays as a drop on the surface of the solid—it has
a contact angle of 1808. If a liquid has a sufficiently low surface tension, it spreads spontaneously on the substrate and has a contact angle of 08. At intermediate surface tensions,
there are intermediate contact angles. A schematic drawing of a drop of liquid with an
intermediate surface tension is shown in Figure 6.2.
The relationship between contact angle u and the surface free energy of the substrate gS,
the surface tension of the liquid gL, and the interfacial tension between the solid and the
liquid gSL for a planar surface is given as
cos u ¼
gS gSL
gL
(6:2)
Maximum adhesion requires a contact angle of 08. Experimental determination of contact
angles of complex systems such as coatings, especially on rough substrates with heterogeneous compositions, is difficult [9]. In general, it is sufficient to conceptualize the
Figure 6.2. Contact angle.
6.4. ADHESION TO METAL SURFACES
125
relationships by saying that adhesion requires that the liquid have a lower surface tension
than the surface free energy of the substrate to be coated. From a practical standpoint, it is
useful to do cruder but easier experiments. One can apply a drop of coating on a substrate,
put the sample in an atmosphere saturated with the solvents in the coating, and watch the
spreading. If the droplet of coating stays as a small ball, spreading is poor and adhesion
problems should be expected. If the drop spreads out to a thin, wide circle, the coating
meets at least one criterion for good adhesion. It is also useful to carry out a second
type of experiment, in which a coating is spread mechanically on a substrate under a
solvent saturated atmosphere and then observed. Generally, a liquid that spreads spontaneously in the first experiment will remain spread out in the second experiment. But
sometimes, a liquid that spreads on the surface will draw up into droplets, or at least
pull away in sections from the substrate when standing, without solvent evaporation.
Consider, for example, the application of n-octyl alcohol to a clean steel surface. The
surface tension of n-octyl alcohol is lower than the surface free energy of steel, and it
spreads on steel spontaneously. However, if one spreads out a film of n-octyl alcohol
on steel, the film draws up to form droplets on the surface of the steel. The low surface
tension of the n-octyl alcohol results from the linear hydrocarbon chain; however, after
spreading on the polar surface of the steel, the hydroxyl groups of the n-octyl alcohol molecules interact with the surface so that a monolayer of oriented n-octyl alcohol molecules
forms on the surface. This makes a new surface, aliphatic hydrocarbon, which has a low
surface tension, lower than the surface tension of n-octyl alcohol. The n-octyl alcohol
above a monolayer dewets. The behavior of n-octyl alcohol illustrates a principle important in formulating coatings: One must be careful about using additives with single polar
groups and long hydrocarbon chains in coatings to be used directly on metals. An
example is the poor adhesion to steel that can result from use of dodecylbenzenesulfonic
acid as a catalyst (Section 11.3.1.1). Adhesion of latex films can be affected by a layer of
surfactant forming at the interface between the coating and the substrate [10].
6.4. ADHESION TO METAL SURFACES
The metal and the surface characteristics of the metal can have major effects on adhesion;
Ref. [11] provides a review of metal surface characteristics, cleaning, and treatments. The
surface tension of a clean metal surface (usually, metal oxide) is higher than that of any
potential coating. However, metal surfaces are frequently contaminated with oil, and
such surfaces can have very low surface tensions. Whenever possible, it is desirable to
clean the surface of the metal before applying a coating. Sometimes, the metal is wiped
with rags wet with solvent. A more effective method is vapor degreasing, in which the
object is hung from a conveyor that carries it into a tank above a boiling chlorinated
solvent. The cold steel surface acts as a condenser, condensing solvent on the surface
of the steel, where it dissolves oils. The solution drips off, removing the oil. The
solvent is purified by distillation for reuse. Surfactant solutions are also used to clean
oil from metals [11]. Care must be exercised in selecting surfactants and in rinsing the
surface after cleaning. It is possible for some surfactants to adsorb on the surface, creating
a hydrocarbon layer on the surface of the metal.
Steel can be cleaned by abrasive particle blasting such as sandblasting. (See Ref. [11] for a
discussion.) The surface of the steel, including rust, is removed, leaving a rough surface.
Sandblasting is widely used for steel structures such as bridges and tanks, but it leaves the
126
ADHESION
steel surface too rough for products such as automobiles and appliances. Sandblasting is
effective, but there are health hazards to workers from inhalation of silica dust and old
paint debris. A variety of alternative blasting techniques are being used or evaluated [12].
These include use of other dry, abrasive materials, such as steel grit or water-soluble abrasives, including sodium bicarbonate and salt, to replace sand; cryogenic cleaning by blasting
with dry ice pellets; vacuum blasting using aluminum oxide; and for cleaning softer metals
such as aluminum, plastic pellet blasting. Ultrahigh pressure hydroblasting at pressures above
175 mPa (25,000 psi) is very effective in removing oil and surface contaminants such as salt.
Formulations are tested on laboratory panels, but the surfaces of these panels are not the
same as those of the product to which the coating will be applied. Furthermore, commercial test panels vary: for example, the side of the panel next to the wrapping of a package of
panels was shown to have a different surface analysis than that of inner panels from the
same package [13]. Washing the panels with warm water and rinsing with acetone
before coating generally improved adhesion. When possible, laboratory tests should
also be carried out on the ultimate substrate, or at least on sample pieces of metal to be
used in production, simulating, as closely as possible, factory cleaning and treating
procedures.
6.4.1. Surface Preparation
When good adhesion, corrosion protection, and a relatively smooth surface are required, it
is common to treat the surface of the metal chemically. The treatments for steel are called
conversion coatings or chemical pretreatments. A variety of phosphate-based conversion
coatings are used. One such coating involves spray or immersion application of a phosphoric acid– based “iron phosphate” solution. This method provides a mild etch of the steel
surface and precipitation of a ferrous –ferric phosphate monolayer. Adhesion of coatings
is markedly increased; however, corrosion protection is enhanced only slightly. More
enhancement is often possible by using zinc phosphate conversion coating. The steel
object is immersed in a bath of zinc acid phosphate solution and a coprecipitate of zinc
and ferric phosphates is formed on the steel surface, as shown in Scheme 6.1. Both
adhesion and corrosion protection are enhanced.
The coprecipitate forms a mesh of crystals that adheres tightly to the surface, increasing
the surface area on a microscopic scale. Depending on the zinc concentration in the treatment bath, different crystals can be deposited. At relatively high zinc concentrations, the
crystals are predominantly hydrated zinc phosphate, Zn3(PO4)2 . 4H2O, called hopeite.
Under zinc-starved conditions, the crystals have been identified as phosphophyllite, Zn2Fe(PO4)2 . 4H2O [14]. The performance of the conversion coating is affected by the
Scheme 6.1
6.4. ADHESION TO METAL SURFACES
127
uniformity and degree of surface treatment. Zinc phosphate coatings are generally applied
in the range 1.5 to 4.5 g m22. A variety of other phosphate coatings are also used.
(See Ref. [11] and Sections 7.3.1.1, 27.1, 27.2, and 30.1.1 for further discussions of
conversion coating.)
The reactions shown in Scheme 6.1 are straightforward. Achieving the high rates of
reaction required to permit minimum dwell times for treatment is more difficult. Proprietary formulations reduce times to the order of minutes or seconds. Treatment quality is
dependent on time, temperature, and pH; these and other variables must be closely controlled to assure that the desired type and dimensions of crystals are formed. The
treated surface must be rinsed thoroughly to remove any soluble salts, since these salts
could lead to blister formation when water vapor permeates through a coating film
applied over the soluble salts; rinsing also removes loosely adhering crystals. It has
been common for the last rinse to contain a low concentration of chromic acid to
protect against corrosion. Due to the toxic risks of hexavalent chromium, replacements
for the chromate rinse are being sought. Over zinc phosphate, a rinse of 0.5% methyltrimethoxysilane with sufficient H2ZrF6 to give a pH of about 4 is reported to give
better performance than a chromic acid rinse [15]. However, over iron phosphate, the
silane – H2ZrF6 rinse gave a poorer performance then did a chromic acid rinse.
The mechanism of action of a phosphate crystal layer is not fully understood. One
factor is that the coating penetrates into the crystal mesh, providing a mechanical
attachment to the crystals attached to the surface. The interfacial area for interaction
is greater than for a relatively smooth steel surface. It is also possible that hydrogenbond interactions between these crystals and the resin molecules are stronger (i.e.,
less readily displaced by water) than those between the steel surface and the resin
molecules.
The surface of aluminum is a thin, dense, coherent layer of aluminum oxide. For many
applications, no treatment other than cleaning is required. However, for applications in
which there might be exposure to salt, surface treatment is necessary. Most treatments
for aluminum have been chromate treatments. An example of one such treatment is an
acid bath containing chromate, fluorides, and a ferricyanide salt as an accelerator. The
resulting coating is said to have the following composition: 6Cr(OH)3 . H2CrO4 . 4Al2O3 .
8H2O. (See Ref. [11] for further discussion.) In the past several years, many chromeand cyanide-free proprietary aluminum conversion coatings have been developed that
demonstrate a performance equivalent to that of chrome coatings.
To provide greater protection against corrosion, steel coated with zinc is widely used in
construction and automobiles. Several types of zinc-coated steel are used; the best known
is galvanized steel. There can be large variations in adhesion, depending on the condition
of the zinc layer of the galvanized steel. If zinc-coated steel has been exposed to rain or
high humidity before coating, there may have been some degree of surface oxidation,
leading to formation of a combination of ZnO, Zn(OH)2, and ZnCO3; all of these are
basic and somewhat soluble in water. Therefore, it is important to use saponification resistant resins in primers for galvanized steel. Resins such as alkyds, which tend to saponify,
are likely to give poor adhesion in service. Zinc surfaces on automobile bodies are treated
with zinc, manganese, or nickel phosphate conversion coatings before cationic electrodeposition of primer [16]. The effectiveness of phosphate coatings is affected by the presence
of small amounts of other metals in the zinc coating [17]. The coating of galvanized steel is
reviewed in Ref. [17].
Stainless steel, because of its smoothness and lack of oxide and hydroxide groups on the
surface, is difficult to adhere to. In some cases, roughening the surface provides a basis for
128
ADHESION
anchoring; in others, pretreatment with an electrochemically produced flash of chromium/
chromium oxide is required [18].
Plasma treatment is a new approach being investigated for cleaning and treating the
surface of cold-rolled steel [19]. Initially, the steel is cleaned by plasma discharge, and
then trimethylsilane is introduced into the plasma chamber, leading to polymerization
of a thin layer of polymer bonded to the surface. Laboratory tests indicated superior
adhesion of electrodeposition coatings (Chapter 27) to the surface compared to a conventionally pretreated galvanized steel substrate. If the process can be scaled up successfully
for production, substantial benefits can be foreseen. Cold-rolled steel is less expensive and
more readily recycled than electrogalvanized steel, and the waste disposal problems
associated with electrogalvanizing and phosphating would be eliminated.
Another approach being investigated is use of bis(trialkoxysilyl)alkanes to react with
the surface of steel (Section 6.4.3).
6.4.2. Coating – Substrate Interaction
The surface of clean steel is not iron; rather, hydrated iron oxides are present as a monolayer on the iron. The layer is not a layer of rust particles but a monolayer of hydrated
oxide [20]. The surface is rarely, if ever, uniform on the molecular level. Adhesion to
this surface is promoted by developing hydrogen bonds between groups on the resin molecules and the oxide and hydroxide groups on the surface of the steel. (In the case of phosphate conversion – coated steel, there is the further possible hydrogen bond interaction
with the phosphate groups.) Some authors prefer to interpret the interactions in terms of
association between soft acids and soft bases [21].
It follows that adhesion is promoted by using resins having multiple hydrogen-bond
donating and accepting groups, such as carboxylic acid (strongly hydrogen-donating),
amine (strongly hydrogen-accepting), hydroxyl, urethane, amide (the latter three being
both hydrogen-donating and hydrogen-accepting), and phosphate (hydrogen-donating
and hydrogen-accepting). One might assume that a large number of such substituents
on a molecule would be desirable. However, it is known from adsorption studies that if
there are large numbers of polar groups, at equilibrium the adsorbed layer can be very
thin. The principle can be illustrated by considering a polymer molecule with an aliphatic
backbone chain with polar groups on every other carbon atom. At equilibrium, adsorption
of adjacent polar groups is favored sterically, resulting in a thin adsorbed layer with the
polar groups on the steel surface and only hydrocarbon groups exposed to the rest of
the coating. The interactions between the rest of the coating and the hydrocarbon
groups would be expected to be weak, resulting in a weak boundary layer and poor cohesion. If a smaller number of hydrogen-bond-donating groups are scattered along a resin
chain, adsorption of resin molecules may occur with loops and tails sticking up from
the surface so that some of the polar groups are adsorbed on the surface and some are
on the loops and tails, where they can interact with the rest of the coating. On those
parts of the resin molecule in the loops and tails, there can be groups to hydrogen-bond
with molecules in the coating, as well as functional groups, which can react with a crosslinker in the coating.
BPA epoxy resins (Sections 13.1.1 and 15.8) and their derivatives commonly provide
excellent adhesion to steel. These resins have hydroxyl and ether groups along the chain,
which can provide interactions with both the steel surface and other molecules in the
coating. The cross-linkers employed with these resins often introduce additional
6.4. ADHESION TO METAL SURFACES
129
hydrogen-bonding groups, such as amines. It may also be important that the backbone consists of alternating flexible 1,3-glyceryl ether and rigid bisphenol A groups. It seems
logical that such a combination could provide the flexibility necessary to permit multiple
adsorption of hydroxyl groups on the surface of the steel, along with the rigidity to prevent
adsorption of all of the hydroxyl groups. The remaining hydroxyl groups can participate in
cross-linking reactions or can hydrogen bond with the rest of the coating. The effects of
variations in epoxy resin composition on adhesion are discussed in Refs. [1] and [22].
In general, the results of many studies of adsorption of polymer molecules on metal
surfaces are consistent with the picture given above. However, many of these studies
involve adsorption from dilute solutions. Observations are made over relatively long
time intervals, permitting equilibrium conditions to develop. With polydisperse MW
adsorbents, low MW species are adsorbed first, but at equilibrium, they are displaced
by higher MW molecules with larger numbers of polar groups. What really happens
when a coating is applied? The resin is in a relatively concentrated solution, and the
solvent evaporates in a relatively short time. There may not be time for equilibrium to
be established. Depending on the coating, those groups that happened to be near the
surface when the film was applied might remain there and could lead to poor adhesion,
even if the same resins could provide good adhesion given the opportunity for appropriate
orientation and equilibration to occur. Such a scenario is compatible with improved
adhesion using slow evaporating solvents, which can also permit more complete penetration into surface crevices. Perhaps another reason that baking coatings commonly
leads to improved adhesion is greater opportunity for orientation of molecules at the substrate –coating interface at the higher temperature. There is need for further research on
interactions and orientation at the interface of coatings and substrates.
Surface analysis can be useful in understanding factors affecting adhesion. The surfaces
of uncoated steels have been studied by Auger analysis; organic compounds have been
detected on the surface of some cold-rolled steels. These organic compounds apparently
become embedded in the surface of the steel during coil annealing. If this happens, it
becomes difficult to obtain high quality phosphate conversion treatments on the steel
[23]. Such steels more commonly lead to adhesion failures and to inferior corrosion protection by coatings.
X-ray photoelectron spectroscopy (XPS) can be used to study the surface of steel from
which a coating has been removed as well as the underside of the coating. This technique
can be particularly powerful for showing where failure occurred: that is, whether failure
was between the steel and the coating or between the main body of the coating and a monolayer (or a very thin layer) of material on the surface of the steel. Other valuable analytical
procedures for thin surface layers are attenuated total reflectance (ATR) and FTIR. Such
techniques have been most useful for diagnosing problems; further use of these and other
analytical techniques for understanding the mechanisms of adhesion can be anticipated.
Although strong interaction between coating and steel is critical for achieving good
adhesion, it is also important to develop interactions that cannot be displaced easily by
water. A reason that the presence of multiple groups on resin molecules, such as hydroxyl
groups on epoxy resins, may be desirable is that some may remain bonded to the steel
while others are reversibly displaced by water. This phenomenon has been termed cooperative adhesion [24]. It has been found empirically that amine groups on the cross-linked
resin molecules promote corrosion protection. Explanations for the effect are controversial; one hypothesis is that the amine groups interact strongly with the steel surface and
are not as easily displaced by water from the surface as are hydroxyl groups. Phosphate
130
ADHESION
groups are another substituent group that has been found to impart improved adhesion in
the presence of water (i.e., wet adhesion). For example, the use of epoxy phosphates in
epoxy coatings has been shown to improve both adhesion and wet adhesion [22]
(Section 13.5). Phosphate functional (meth)acrylate esters as comonomers have been
shown to improve wet adhesion of acrylic latex and solventborne acrylic coatings [25].
6.4.3. Covalent Bonding to Glass and Metal Substrates
Stronger interactions with the substrate surface should be possible by forming covalent
bonds compared to the more readily displaced hydrogen-bonds. One such approach is
the use of reactive silanes, which are very effective in enhancing adhesion of coatings
to glass [26] (Section 16.2). A variety of reactive silanes is available; they all have a trialkoxysilyl group attached to a hydrocarbon chain, the other end of which has a functional
group such as amine, mercaptan, epoxy, or vinyl. The alkoxysilyl group can react with
hydroxyl groups on the surface of glass and with other alkoxysilyl groups after hydrolysis,
so the surface of the glass becomes covalently bonded to a series of hydrocarbon tails substituted with reactive groups that can cross-link with the coating being applied.
In formulating an epoxy –amine coating for glass, one could add 3-aminopropyltrimethoxysilane to the amine package of the two package coating. After application, the trimethoxysilyl group can react with silanol groups on the surface of the
glass to generate siloxane bonds, as shown in the first step of Scheme 6.2. The trimethoxysilyl groups can also react with water to produce silanol groups, which can in turn react
with remaining silyl methoxy groups to generate polysiloxane groups at the glass
surface; see the second step in Scheme 6.2. The terminal amine groups can react with
epoxy groups in the resin so that the coating is multiply bonded to the surface of the
glass, as shown in the third step in Scheme 6.2.
When water vapor penetrates through the coating to the glass interface, hydrolysis of
some of the interfacial Si22O bonds occurs. However, with multiple interfacial bonds,
Scheme 6.2
6.5. ADHESION TO PLASTICS AND TO COATINGS
131
some of the bonds remain intact and prevent the coating from delaminating. Furthermore,
hydrolysis is reversible, so the hydrolyzed bonds can re-form. Before the advent of reactive silanes, it was difficult to formulate coatings that maintained adhesion to glass after
exposure to a humid atmosphere.
Reactive silanes have also been added to coatings with the objective of improving
adhesion to steel and other metal surfaces [26]. There is a growing body of indirect evidence of improvement. In theory, the trialkoxysilyl group could react with hydroxyl
groups attached to iron, although direct evidence for this theory is in short supply. In
the past, reactive silanes were not widely adopted as additives to improve the adhesion
of coatings to steel, although their use is said to be increasing. Perhaps the lack of widespread use results from the lower stability of the bonds formed with iron than of those
formed with glass and other metals, such as aluminum [27]. In the case of amino
silanes, the amine group may be preferentially adsorbed on iron. An extensive review
of the journal and patent literature is available [28].
Many investigations of the use of bis(trialkoxysilyl)alkanes for the treatment of metals
have been reported [29]. For example [29], clean steel is rinsed with water and then the wet
steel is dipped into an aqueous solution of bis(trimethoxysilyl)ethane (BTSE). BTSE
hydrolyzes in the water and then adsorbs on the surface of the steel, reacting with hydroxyl
groups and cross-linking with other BTSE molecules. The multiple covalent attachments
form a coating that is stable when immersed in water. The surface can then be treated with
a reactive silane that reacts with silanol groups and also provides a reactive group to react
with a coating binder. Many variables affect the treatment; they are currently being investigated in the laboratory, and commercial use is projected. Laboratory tests indicate excellent adhesion in the presence of water and good corrosion protection.
Another approach to achieving chemical bonding to steel is the use of resins containing
groups that can form coordination complexes with ferric compounds. For example, one
can make resins with acetoacetic ester substituents (Section 17.6). Such esters are
highly enolized and can coordinate with metal ions, including ferric salts. Preliminary
reports indicate improvement in adhesion and corrosion protection [30]. Because of the
potential hydrolysis of acetoacetic esters, evaluation over relatively long time intervals
will be required to assess their commercial utility.
6.5. ADHESION TO PLASTICS AND TO COATINGS
In contrast to clean steel and other metals, commonly, there is a problem wetting the
surface of plastic substrates with a coating. Wetting of many plastics is difficult, owing
to their low surface free energies, and can be further reduced by the presence of mold
release agents on molded plastic parts. Mold release agents should be avoided if at all
possible. If essential, release agents should be selected that are relatively easily
removed from the molded part, and care should be exercised to remove all traces. Even
after cleaning, the surface free energies of some plastics are lower than the surface tensions of many coatings, resulting in a contact angle .08. The contact angle between
coating and substrate should be 08 to permit spreading. Determination of contact angle
is experimentally difficult, especially due to surface roughness and inhomogeneity [9].
Attainment of satisfactory adhesion to polyolefins, which generally have low surface
free energy, usually requires treatment of the surface to increase its surface free energy.
This can be done by oxidation of the surface to generate polar groups such as hydroxyl,
132
ADHESION
carboxylic acid, and ketone groups. The presence of these groups not only increases
surface free energy so that wetting is possible with a wider range of coating materials,
but also provides hydrogen-bond acceptor and donor groups for interaction with complementary groups on coating resin molecules. A variety of processes can be used to
treat the surface [31,32]. The surfaces of films, flat sheets, and cylindrical objects can
be oxidized by flame treatment with gas burners using air/gas ratios such that the
flames are oxidizing. Oxidation can also be accomplished by subjecting the surface to a
corona discharge atmosphere; the ions and free radicals generated in the air by the electron
emission serve to oxidize the surface of the plastic.
Adhesion to untreated polyolefins can be assisted by applying a thin tie coat of a
low solids solution of a chlorinated polyolefin or chlorinated rubber. Ryntz has
reviewed the various approaches and provided the results of various types of surface
analysis [31]. (See Section 31.2.2 and Ref. [33] for further discussion of surface
treatments.)
Both theoretical and experimental studies show that the molecules at the surface of a
polymer are more mobile than those in the bulk material [34]. One theoretical study indicated that the layer of mobile molecule segments is about 2 nm in thickness [35]. Because
polymer surfaces are dynamic, they adjust to the environment; polar groups will, with
time, move away from a polymer – air interface and toward a polymer – water interface.
Thus, the polar groups placed on the surface by oxidation treatments can be expected to
move to the interior within hours or days after treatment [34].
Although adhesion between coatings and plastic substrates can be enhanced by
hydrogen-bond interactions, still further enhancement can be obtained if the temperature
is above the Tg of the plastic substrate. At temperatures above Tg, there is adequate free
volume to permit resin molecules from the coating to move into the surface of the
plastic, and vice versa. Presumably, penetration can be deeper than the thin layer of
mobile molecule segments at the surface. When the solvent evaporates, the intermingled
molecules increase adhesion. This interaction can be enhanced by having the structure of
the coating resin be sufficiently similar to the structure of the plastic that the resin molecules are somewhat soluble in the plastic substrate. In some cases, promotion of adhesion
by heating the plastic substrate above its Tg is not feasible because the plastic substrate
may undergo heat distortion.
Solvents in the coating that are soluble in the plastic can enhance adhesion. The solvent
swells the plastic, lowering its Tg, and facilitating penetration of coating resin molecules
into the surface of the plastic. The solvents should evaporate slowly to permit time for penetration to occur. Fast evaporating solvents such as acetone can cause crazing of the
surface of high Tg thermoplastics such as polystyrene and poly(methyl methacrylate).
Crazing is the development of large numbers of minute surface cracks. (See Section
31.2.2 for a further discussion of crazing and coatings for plastics.)
Adhesion to other coatings, commonly called intercoat adhesion, is another example of
adhesion to plastics. The same principles apply. The surface tension of the coating being
applied must be lower than the surface free energy of the substrate coating to permit
wetting. The presence of polar groups in both coatings permits hydrogen bonding; in
the case of thermosetting coatings, covalent bonding enhances intercoat adhesion. It has
been found empirically that the presence of relatively small amounts of amine groups
on resins commonly gives coatings with superior intercoat adhesion. Such comonomers
as 2-(N,N-dimethylamino)ethyl methacrylate and 2-aziridinylethyl methacrylate have
been used to make acrylic resins with enhanced intercoat adhesion.
6.6. TESTING FOR ADHESION
133
Curing temperatures above Tg increase the probability of satisfactory adhesion. Compatible resins in the substrate coating and top coat also increase the probability of satisfactory adhesion. Using solvents in the coating that can swell the substrate coating is a
commonly used technique for enhancing intercoat adhesion. Coatings with lower crosslink density are more swollen by solvents and, in general, are easier to adhere to than
are coatings with high cross-link density. Sometimes, one can undercure the primer,
thus having a lower cross-link density when the top coat is applied. Cure of the primer
is completed when the top coat is cured.
Adhesion to high gloss coatings is difficult to achieve because of their surface smoothness. Gloss coatings that have undergone excessive cross-linking on ageing are particularly difficult surfaces on which to apply an adherent coating. Sanding to increase
surface roughness may be necessary to achieve intercoat adhesion. One reason for formulating primers with low gloss is that their rougher surfaces enhance adhesion. When possible, increasing the pigment loading of a primer above critical pigment volume
concentration (CPVC) facilitates adhesion of a top coat (Chapter 23). Above CPVC, the
dry film contains pores. When a top coat is applied, vehicle from the top coat can penetrate
the pores in the primer, providing a mechanical anchor to promote intercoat adhesion.
Care must be exercised not to have PVC too much higher than CPVC, or so much
vehicle will be drained away from the top coat that the PVC of the top coat will increase,
leading to a loss of gloss.
An essential requirement of many industrial coatings is recoat adhesion: that is, the
ability of a coating to adhere to itself well enough that flawed or damaged objects can
be repainted without extensive preparation. This requirement can be difficult to satisfy,
especially with highly cross-linked gloss enamels. Additives to overcome film defects
during application (Section 24.4) may interfere with recoat adhesion.
6.6. TESTING FOR ADHESION
In view of the complexity of adhesion phenomena, it is not surprising that there is difficulty in devising suitable tests for adhesion. As is so often the case in coatings, the only
really conclusive way of telling whether adhesion of a coating is satisfactory is to use
the product and see whether the coating adheres over its useful life.
A common method formulators use to evaluate adhesion is to see how easily a penknife
can scrape a coating from a substrate. By comparing the resistance of a new coating –
substrate combination to combinations with known field performance, the formulator
has some basis for performance prediction. While a penknife in the hand of an experienced
person can be a valuable tool, it has major disadvantages as a test method. The experience
is not easily transferred from one person to another; even the technique for the test is not
easily transferred. Also, there is no good way to assign numerical values to the results.
Thus, it does not provide a basis for following small changes in adhesion as a result of
changes in composition to aid in developing hypotheses to relate composition and
adhesion.
Relatively satisfactory test methods for evaluating adhesives have been developed, but
few of these methods are applicable to coatings. Many investigators have worked on a
variety of methods in attempts to devise meaningful tests for evaluating the adhesion of
coatings [36,37]. None of these tests are very satisfactory. For research purposes, the
most useful technique is a direct pull test. A rod is fastened perpendicular to the upper
134
ADHESION
surface of the coated sample with an adhesive. The panel is fastened to a support with a
perpendicular rod on its back so that the two perpendicular rods are lined up exactly opposite each other. The assembly is put into the jaws of an Instron Tester (Section 4.5), and the
tensile force required to pull the coating off the substrate is recorded. Since the procedure
is subject to considerable experimental error, multiple determinations must be made.
Experienced operators can achieve precisions of +15%. The adhesive must bond the
rod to the coating surface more strongly than the coating is bonded to the substrate. It
is also essential that the adhesive not penetrate into the coating to perturb the coating –
substrate interface. Cyanoacrylate adhesives are generally satisfactory. The rods must
be aligned with each other exactly and be perpendicular to the coating. If the rod is at
even a slight angle to the surface, stress is concentrated on only part of the substrate –
coating interface, and less force is required to break the bond. Sometimes, the weakest
component is the substrate—this may be nice for advertising purposes, but it does not
provide a measure of the adhesive strength.
Another potential complication is cohesive failure of the coating; again, no information
on adhesion is obtained. In reading the literature, one sometimes finds data from tests in
which cohesive, adhesive, and mixed cohesive – adhesive failures have occurred. The
authors may then discuss the improvement in adhesion from some change that resulted
in a greater force to get adhesive failure than that of another sample that failed cohesively.
Clearly, such comparisons are invalid. When there is cohesive failure, all that is known is
that the adhesive strength is above the measured value (within experimental error).
One must use caution in interpreting the results even when the sample appears to have
failed adhesively at the substrate –coating interface. Sometimes, when no coating can be
seen on the substrate surface after the test, there is a monolayer (or thin layer) of material
from the coating left on the substrate surface. In this event, failure was not at the substrate
surface but between the material adsorbed on the surface and the rest of the coating.
Surface analysis is useful in determining the locale of failure and the identity of the
adsorbed material. Fairly often, there is a combination of adhesive and cohesive failure.
A possible explanation of such failures is that there was a fractural failure starting at
some imperfection within the film, and the initial crack propagated down to the interface.
Tensile values from samples that fail in this way cannot be compared to the tensile values
of samples that failed adhesively.
The direct pull test does not evaluate the potentially important differences between the
difficulty of breaking through a coating film and of shoving it out of the way, as mentioned
at the beginning of the chapter. Despite all the difficulties, direct pull tests are the most
useful available. Instruments have been devised for direct pull tests under field conditions.
The method is quite widely used for quality control in high performance maintenance and
marine coatings. Serious disadvantages for use on actual products are that the test is
destructive and the area tested must be repainted. See Ref. [36] for a more detailed discussion of the effect of variables on the test results.
Adhesion can be affected by the angle of application of stress. An instrument called
STATRAM II has been devised to combine a normal load and lateral traction to
measure friction induced damage [38]. Optical measurements are combined with measurements of total energy consumed during the scraping process. The test has been used to
study delamination of coatings when plastic automobile bumpers rub together or scrape
against solid objects. In many cases, cohesive failure of the plastic occurred near the
surface of thermoplastic olefin (TPO) rather than adhesive failure between the coating
and the substrate. The composition of coatings, especially solvents, can affect the structure
REFERENCES
135
of the upper layer of the plastic. (See Section 31.2.2 for further discussion of adhesion of
coatings to plastics.)
Acoustic testing is of increasing interest. When a film cracks, it emits acoustical energy.
A coating is applied to a film, placed in a tensile tester, and a transducer is attached to the
surface; then as the jaws of the tensile tester are separated, the vibrations of noise are
recorded. This permits evaluation of the stress required to lead to film cracking. The
devices are particularly useful in comparing dry films to those previously exposed to
moisture testing.
Probably, the most widely used specification test is the cross hatch adhesion test. Using
a device with 6 or 11 sharp blades, a scratch mark pattern is made across the sample, followed by a second set cut perpendicular to the first. A strip of pressure-sensitive adhesive
tape is pressed over the pattern of squares and pulled off. Adhesion is assessed qualitatively
on a 5 to 0 scale by comparing to a set of photographs, ranging from trace removal along the
incisions to removal of most of the area. The test is subject to many sources of error, one
being the rate at which the cuts are made. If the cuts are made slowly, they are likely to be
even. However, if the cuts are made rapidly, it is possible that there will be cracks proceeding out from the sides of the cuts, due to more brittle behavior of coatings at higher rates of
application of stress. Other important variables are the adhesive tape; the pressure with
which it is applied; the angle and rate at which the tape is pulled off the surface;
bending, if any, of the substrate during the test; and the surface of the coating to which
the tape is applied. Some additives that appear to improve adhesion may actually only
improve the test results by decreasing adhesion of the tape to the coating. The test may
be useful for distinguishing between samples having poor adhesion and those having
fairly good adhesion, but it is not very useful in distinguishing among higher levels of
adhesion. See Ref. [36] for a more detailed discussion of the variables affecting the test.
Wet adhesion is a critical property for interior linings on tin cans. A paper provides evidence that electrochemical impedance spectroscopy (EIS) is a useful way to study wet
adhesion of the coatings [39]. (See Sections 7.3.2 and 7.5, which deal with corrosion,
for further discussion of factors affecting adhesion and testing for adhesion in the presence
of water.)
GENERAL REFERENCES
Baghdachi, J. A., Adhesion Aspects of Polymeric Coatings, Federation of Societies for
Coatings Technology, Blue Bell, PA, 1996.
Hartshorn, S. R., Structural Adhesives: Chemistry and Technology, Plenum Press, New York, 1986.
Mittal, K. L., Adhesion, Aspects of Polymeric Coatings, Plenum Press, New York, 1983.
Nelson, G. L., Adhesion, in Paint and Coating Testing Manual, 14th ed., Koleske, J. V., Ed., ASTM,
Philadelphia, PA, 1995, pp. 513 – 524.
REFERENCES
1. Sheih, P. S.; Massingill, J. L., J. Coat. Technol., 1990, 62(781), 25.
2. Perera, D. Y.; Van den Eynden, D., J. Coat. Technol., 1987, 59(748), 55.
3. Perera, D. Y., Prog. Org. Coat., 1996, 28, 21.
136
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
ADHESION
Negele, O.; Funke, W., Prog. Org. Coat., 1995, 28, 285.
Kloosterboer, J. G., Adv. Polym. Sci., 1988, 84, 1.
Basin, V. E., Prog. Org. Coat., 1984, 12, 213.
Pocius, A. V., in Structural Adhesives: Chemistry and Technology, Hartshorn, S. R., Ed.,
Plenum Press, New York, 1986, pp. 23– 68. Portelli, G. B., in Structural Adhesives: Chemistry
and Technology, Hartshorn, S. R., Ed., Plenum Press, New York, 1986, pp. 407 – 449.
Zisman, W. A., J. Coat. Technol., 1972, 44(564), 42.
Yekta-Fard, M.; Ponter, A. B., J. Adhesion Sci. Technol., 1992, 6, 253.
Charmeau, J. Y.; et al., Prog. Org. Coat., 1996, 27, 87.
Perfetti, B. M., Metal Surface Characteristics Affecting Organic Coatings, Federation of
Societies for Coatings Technology, Blue Bell, PA, 1994.
Rex, J., J. Prot. Coat. Linings, 1990, October, 50.
Skerry, B. S.; et al., J. Coat. Technol., 1991, 62(788), 55.
Dyett, M. J., J. Oil Colour Chem. Assoc., 1989, 72, 132.
Tang, N.; et al., Prog. Org. Coat., 1997, 30, 255.
Schoff, C. K., J. Coat. Technol., 1990, 62(789), 115.
Maeda, S., Prog. Org. Coat., 1996, 28, 227.
Lori, L.; et al., Prog. Org. Coat., 1996, 27, 17.
Lin, T. J.; et al., Prog. Org. Chem., 1997, 31, 351.
Reinhard, G., Prog. Org. Coat., 1987, 15, 125.
Jacobasch, H. J., Prog. Org. Coat., 1989, 17, 115.
Massingill, J. L.; et al., J. Coat. Technol., 1990, 62(781), 31.
Maeda, S., J. Coat. Technol., 1983 55(707), 43.
Funke, W., J. Coat. Technol., 1983, 55(705), 31.
Yang, H. S.; et al., JCT Coat. Tech, 2005, 2(13), 44.
Pluddemann, E. P., Prog. Org. Coat., 1983, 11, 297.
Witucki, G. L., J. Coat. Technol., 1993, 65(822), 57.
Sathyanaarayana, M. N.; Yaseen, M., Prog. Org. Coat., 1995, 26, 275.
van Ooij, W. J.; Child, W. J., Chemtech, 1998, February, 26.
Rector, F. D.; et al., J. Coat. Technol., 1989, 61(771), 31.
Ryntz, R. A., Polym. Mater. Sci. Eng., 1992, 67, 119.
Lane, J. M.; Hourston, D. J., Prog. Org. Coat., 1993, 21, 269.
Ryntz, R. A., Painting of Plastics, Federation of Societies for Coatings Technology, Blue Bell,
PA, 1994.
Garbassi, F.; et al., Polymer Surfaces from Physics to Technology, revised and updated ed.,
Wiley, New York, 1998, pp. 49– 68.
Mansfield, K. F.; Theodorou, D. N., Macromolecules, 1991, 24, 6283.
Nelson, G. L., in Paint and Coating Testing Manual, 14th ed., Koleske, J. V., Ed., ASTM,
Philadelphia, PA, 1995, pp. 513 – 524.
Bullett, T. R.; Prosser, J. L., Prog. Org. Coat., 1972, 1, 45.
Ramamurthy, A. C.; et al., Wear, 1997, 203 – 204, 350. Mihora, D. J.; Ramamurthy, A. C., Wear,
1997, 203 – 204, 362.
Barilli, F.; et al., Prog. Org. Coat., 2003, 46, 91.
7
Corrosion Protection
by Coatings
Corrosion is a process by which materials, especially metals, are worn away by
electrochemical and chemical actions. The estimated annual economic cost of corrosion
in the United States was $296 billion in 1995, so the incentive to improve corrosion protection technology is enormous [1]. In this chapter we discuss the principles of corrosion and
the protective role of organic coatings; specific types of coatings for corrosion control are
covered in Chapters 27, 30, and 33.
7.1. CORROSION OF UNCOATED STEEL
The major economic losses are from the electrochemical corrosion of steel, which is the
main focus of this chapter. Aluminum is also important, especially in aircraft and military
vehicles. Protection of various other metals has been reviewed [2]. Before we consider the
role of coatings, it is important to understand the corrosion of uncoated steel.
In the corrosion of metals, electrochemical and chemical reactions are involved. An
electrochemical (galvanic) element is formed when two pieces of different metals are connected with a conductive wire and partly immersed in an electrolyte, generally water containing some dissolved salts. An electrochemical reaction begins spontaneously with the
oxidation of metal atoms and dissolution of ions at the anodic parts of the corrosion
elements. Metals may be arranged in an electromotive series such that the most noble
(least easily oxidized) metal forms the cathode. However, because polarization processes
may change the character of a metal (e.g., by formation of thin layers of corrosion products
or by the adsorption of gases such as hydrogen), the practical electromotive series differs
somewhat from the theoretical one. The metals considered herein, in decreasing order of
ease of oxidation, theoretically, are magnesium, aluminum, zinc, iron, tin, and copper.
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
137
138
CORROSION PROTECTION BY COATINGS
There are many kinds of steel, which corrode at different rates, depending on their composition and on the presence of mechanical stresses; all are alloys of iron and carbon with
other metals. The composition of the surface varies from location to location; as a result,
some areas are anodic relative to other areas that are cathodic. Stresses and the morphological structure of the metal surface can also be factors for setting up anode – cathode
pairs. Cold-rolled steel has more internal stresses than hot-rolled steel and is generally
more susceptible to corrosion, but cold-rolled steel is widely used because it is stronger.
Internal stresses can also be created during fabrication or by the impact of a piece of
gravel on an auto body, for example.
Steel adsorbs on its surface a thin layer of water, which dissolves traces of soluble salt,
thus providing the electrolyte for electrochemical reactions. In the absence of oxygen,
ferrous ions are the primary anodic corrosion products of steel. At cathodes, hydrogen
is formed, which stops further dissolution of iron by cathodic polarization, except when
the pH of the electrolyte is low.
However, in the presence of oxygen, depolarization of the cathode takes place, hydroxyl
anions are formed at the cathodes, and dissolution of iron continues:
The rate of corrosion of steel depends on the concentration of oxygen dissolved in the
water at the steel surface, as shown in Figure 7.1. At low concentrations, the rate increases
with increasing dissolved oxygen concentration. At high concentrations, the rate declines
because of passivation (Section 7.2.1). The equilibrium concentration of oxygen in water
exposed to the atmosphere at 258C is about 6 mL L21.
Figure 7.1. Effect of oxygen concentration on corrosion of mild steel in slowly moving distilled water,
48-hour test, 258C. (From Ref. [3], with permission.)
7.1. CORROSION OF UNCOATED STEEL
139
Figure 7.2. Effect of sodium chloride on corrosion of iron in aerated solutions at room temperature
(composite of data from several investigators). (From Ref. [3], with permission.)
Corrosion can occur at a significant rate only if there is a complete electrical circuit.
The rate of corrosion depends on the conductivity of the water at the steel surface. Dissolved salts increase conductivity, which is one reason that the presence of salts increases
the rate of corrosion of steel. Effects of salts on corrosion rates are complex; the reader is
referred to Ref. [1], or other general texts on corrosion, for detailed discussions. The
relationship between NaCl concentration and corrosion rate is shown in Figure 7.2.
The dashed vertical line in the figure indicates the salt concentration in seawater. At
higher salt contents, the rate of corrosion decreases since the solubility of oxygen
decreases as the NaCl concentration increases.
The rate of corrosion also depends on pH, as shown in Figure 7.3. Since iron dissolves
in strong acid even without electrochemical action, it is not surprising that corrosion is
most rapid at a low pH. Corrosion rate is nearly independent of pH between about 4
and 10. In this pH region, the initial corrosion causes a layer of ferrous hydroxide to precipitate near the anode. Subsequently, the rate is controlled by the rate of oxygen diffusion
through the layer. Underneath, the surface of the iron is in contact with an alkaline solution
Figure 7.3. Effect of pH on corrosion of iron in aerated soft water at room temperature. (From Ref. [3],
with permission.)
140
CORROSION PROTECTION BY COATINGS
Figure 7.4. Effect of temperature on corrosion of iron in water containing dissolved oxygen. (From
Ref. [3], with permission.)
with a pH of about 9.5. When the environmental pH is above 10, the increasing alkalinity
raises the pH at the iron surface. The corrosion rate then decreases because of passivation
(Section 7.2.1).
Corrosion rate also depends on temperature, as shown in Figure 7.4. The reactions
proceed more rapidly at higher temperatures, as indicated by the increase in corrosion
rate in a closed system. However, the solubility of oxygen in water decreases as temperature increases, so in an open system, where the oxygen can escape, the rate of corrosion
goes through a maximum at some intermediate temperature. The temperature at which
the corrosion rate maximizes is system dependent.
7.2. CORROSION PROTECTION OF METALS
In this section we deal with strategies for controlling electrochemical corrosion without
the use of organic coatings, which serve to introduce principles for corrosion protection
using organic coatings (Sections 7.3 and 7.4). One can suppress the anodic reaction, as
discussed in Section 7.2.1; one can suppress the cathodic reaction, as discussed in
Section 7.2.2; and one can prevent water, oxygen, and corrosion stimulants from contacting the surface, as discussed in Section 7.2.3.
7.2.1. Passivation: Anodic Protection
As shown in Figure 7.1, increasing oxygen concentration up to about 12 mL L21 increases
the corrosion rate because it acts to depolarize the cathode. At higher concentrations, more
oxygen reaches the surface than is reduced by the cathodic reaction; beyond that concentration, corrosion is retarded. The mechanism of retardation has not been fully elucidated. According to one theory, if the oxygen concentration near the anode is high
enough, ferrous ions are oxidized to ferric ions soon after they are formed at the anodic
surfaces. Since ferric hydroxide is less soluble in water than ferrous hydroxide, a
barrier of hydrated ferric oxide forms over the anodic areas. Suppression of corrosion
by retarding the anodic reaction is called passivation. The iron is said to be passivated.
The critical oxygen concentration for passivation depends on conditions. It increases
with dissolved salt concentration and with temperature, and it decreases with increases
7.2. CORROSION PROTECTION OF METALS
141
in pH and velocity of water flow over the surface. At about pH 10, the critical oxygen concentration reaches the value for air-saturated water (6 mL L21) and is still lower at higher
pHs. As a result, iron is passivated against corrosion by the oxygen in air at sufficiently
high pH values. It is impractical to control corrosion by oxygen passivation below
about pH 10, since the concentrations needed are in excess of those dissolved in water
in equilibrium with air. However, a variety of oxidizing agents can act as passivators.
Chromate, nitrite, molybate, plumbate, and tungstate salts are examples. As with
oxygen, a critical concentration of these oxidizing agents is needed to achieve passivation,
and lower concentrations may promote corrosion by cathodic depolarization. The reactions with chromate salts have been studied most extensively. Partially hydrated mixed
ferric and chromic oxides are deposited on the surface, where they presumably act as a
barrier to halt the anodic reaction.
Certain nonoxidizing salts, such as alkali metal salts of boric, carbonic, phosphoric, and
benzoic acids, also act as passivating agents. Their passivating action may result from their
basicity. By increasing pH, they may reduce the critical oxygen concentration for passivation below the level reached in equilibrium with air. Alternatively, it has been suggested
that the anions of these salts may combine with ferrous or ferric ions to precipitate
complex salts of low solubility to form a barrier coating at the anode. Possibly, both mechanisms operate to some extent.
A fairly new approach to passivation is applying a film of electrically conductive
polymer to a steel surface to protect it from corrosion. Polyphenyleneamine, commonly
called polyaniline, is available commercially under the trade names Zypan [4] (E. I.
DuPont de Nemours & Co.), Versicon (Allied Signal Co.), and Panda (Monsanto Co.).
It is said to be effective by leading to the formation of a tight, very thin, metal oxide
passivating layer on the surface of the metal. The polymer powders are insoluble in all
solvents, non-fusible, and difficult to disperse because of high surface tension. Dispersions
are available in a variety of vehicles. Reviews of the effects of conducting polymer
coatings on metals are available [5,6]. Electropolymerization and deposition of
polyaniline from a phosphate buffer solution onto stainless steel increases the corrosion
resistance of the steel [7]. Steel coated with polythiophene by electropolymerization
also provides corrosion protection [8]. Another review of conducting polymer coatings
emphasizes protection of aluminum with poly[2,5-bis(N-methyl-N-propylamino)phenylenevinylene] [9].
Aluminum is higher than iron in the electromotive series and is more easily oxidized.
Yet aluminum generally corrodes more slowly than steel. A freshly exposed surface of
aluminum oxidizes quickly to form a dense, coherent layer of aluminum oxide. In other
words, aluminum is passivated by oxygen at concentrations in equilibrium with air. On
the other hand, aluminum corrodes more rapidly than iron under either highly acidic or
highly basic conditions. Also, salt affects the corrosion of aluminum even more than it
affects the corrosion of iron; aluminum corrodes rapidly in the presence of sea water.
7.2.2. Cathodic Protection
If steel is connected to the positive pole of a battery or a direct current source while the
negative pole is connected to a carbon electrode, and both electrodes are immersed in
salt water, the steel does not corrode. The impressed electrical potential makes the
entire steel surface cathodic relative to the carbon anode. The result is electrolysis of
water rather than corrosion of steel. This is an example of cathodic protection.
142
CORROSION PROTECTION BY COATINGS
A related method is to connect the steel electrically to a piece of metal higher in the electromotive series than iron: for example, magnesium, aluminum, or zinc. When a block of
one of these metals is connected to steel and immersed in an electrolyte, the metal is the
only anode in the circuit, and all corrosion takes place at that anode. The metal is called
a sacrificial anode. This method is often used to protect pipelines and steel hulls of ships.
The sacrificial anode is gradually used up and must be replaced periodically. Zinc and
magnesium are generally the preferred sacrificial metals. Aluminum is often ineffective
because a barrier layer of aluminum oxide forms on its surface (Section 7.2.1). However,
aluminum is appropriate for marine applications, since it corrodes readily in salt water.
Another method of cathodic protection is coating steel with zinc to make galvanized
steel. The steel is protected in two ways: Zinc functions as a sacrificial anode and also
acts as a barrier preventing water and oxygen from reaching the steel surface. Since
zinc is easily oxidized, it is passivated by oxygen at concentrations below 6 mL L21. If
the surface of a galvanized sheet is damaged and bare steel and zinc are exposed, the
zinc corrodes but not the steel. After exposure to the atmosphere, the surface of the
zinc becomes coated with a mixture of zinc hydroxide and zinc carbonate. Both are
somewhat soluble in water and strongly basic.
7.2.3. Barrier Protection and Inhibition
For steel to corrode, oxygen and water must be in direct molecular contact with the surface
of the steel. Barriers that can prevent oxygen and water from reaching the surface prevent
corrosion. The zinc layer on galvanized steel acts as a barrier. It may even be considered
that a layer of irreversibly adsorbed small molecules can act as a barrier. It is often incorrectly assumed that the tin coating on steel in tin cans acts electrochemically, similarly to
zinc in galvanized steel. Tin is lower than iron in the electromotive series, so the iron is the
anode and the tin is the cathode. Before a can is opened, the tin coating is intact and acts as
a barrier so that no water or oxygen reaches the steel. After a can has been opened, the cut
bare edges expose both steel and tin to water and oxygen, and the steel corrodes relatively
rapidly.
Many organic compounds are corrosion inhibitors for steel. Most are polar substances
that tend to adsorb on high energy surfaces [10]. Amines are particularly widely used.
Clean steel wrapped in paper impregnated with a volatile amine or the amine salt of a
weak acid is protected against corrosion. Amines are also used in boiler water to minimize
corrosion. The reason for their effectiveness is not clear. They may act as inhibitors
because they are bases and neutralize acids. It may be that amines are strongly adsorbed
on the surface of the steel by hydrogen bonding or salt formation, with acidic sites on the
surface of the steel. This adsorbed layer then may act as a barrier to prevent oxygen
and water from reaching the surface of the steel. However, this mechanism works only
in the presence of an aqueous medium containing the inhibitor.
7.3. CORROSION PROTECTION BY INTACT COATINGS
Organic coatings can be effective barriers to protect steel when it is anticipated that the
coating can be applied to cover essentially all of the substrate surface and when the
layer remains intact in service. However, when it is anticipated that there will not be complete coverage of the substrate or that the layer will be ruptured in service, alternative
7.3. CORROSION PROTECTION BY INTACT COATINGS
143
strategies using coatings that can suppress electrochemical reactions involved in corrosion
may be preferable; they are discussed in Section 7.4. It is seldom effective to try to use
both strategies in the same coating layers—one must choose one or the other.
7.3.1. Critical Factors
Until about 1950, coatings were generally believed to protect steel by acting as a barrier
to keep water and oxygen away from the steel surface. Then it was found by Mayne [11]
that the permeability of paint films was so high that the concentration of water and
oxygen coming through the films would be higher than the rate of consumption
of water and oxygen in the corrosion of uncoated steel. Mayne concluded that barrier
action could not explain the effectiveness of coatings and proposed that the electrical conductivity of coating layers is the variable that controls the degree of corrosion protection.
Presumably, coatings with high conductivity would give poorer protection than would
coatings of lower conductivity. It was confirmed experimentally that coatings having
very high conductivity afforded poor corrosion protection. However, in comparisons of
layers with relatively low conductivity, little correlation between conductivity and protection has been found. It may be that high conductivity layers fail because they also have
high water permeability; some investigators believe that the conductivity of all coatings
is at least a factor in corrosion protection [12,13].
Current understanding of protection of steel against corrosion by intact films is based to
a significant degree on the work of Funke [14 –17]. He found that an important factor not
given sufficient emphasis in earlier work was adhesion of a coating to steel in the presence
of water. Funke proposed that water permeating through an intact film could displace areas
of the coating from steel. In such cases, the film shows poor wet adhesion. Water and
oxygen dissolved in the water would then be in direct contact with the steel surface;
hence, corrosion would start. As corrosion proceeds, ferrous and hydroxide ions are generated, leading to formation of an osmotic cell under the coating film. Osmotic pressure
can provide a force to remove more coating from the substrate. Osmotic pressure can
be expected to range between 2500 and 3000 kPa, whereas the resistance of organic coatings to deformational forces is lower, ranging from 6 to 40 kPa [14]. Thus, blisters form
and expand, exposing more unprotected steel surface. It has also been proposed that
blisters can grow by a nonosmotic mechanism [18]. The suggestion has been made that
water absorbed by a coating induces in-plane compressive stress within the coating and
extends the interfacial bonds elastically between the coating and the steel substrate
[19]. At a point of weak adhesion between the coating and the substrate, the stress can
lead to disbondment. It has been demonstrated that the rate of growth of blisters is
dependent on the modulus of the coating layer; at a sufficiently high modulus, blister
growth is minimized [20].
In either osmotic or nonosmotic mechanisms, the key to maintaining corrosion protection by a coating is sufficient adhesion to resist displacement forces. Both mechanisms
predict that if the coating covers the entire surface of the steel on a microscopic as well
as a macroscopic scale, and if perfect wet adhesion could be achieved at all areas of the
interface, the coating would protect steel against corrosion indefinitely. It is difficult to
achieve both of these requirements in applying coatings, so a high level of wet adhesion
is important but is not the only factor affecting corrosion protection by coatings. For
example, owing to nano- or micro-roughness of the metal surface, not all of its cavities
may be filled by binder molecules, for steric reasons. These unprotected cavities, although
144
CORROSION PROTECTION BY COATINGS
small, are large enough to allow a local accumulation of water molecules that provides a
medium for the action of corrosion cells [21]. Accordingly, Funke found that in addition to
wet adhesion, low water and oxygen permeability help increase corrosion protection [16].
In any case, if wet adhesion is poor, corrosion protection is also poor. However, if the
adhesion is fairly good, a low rate of water and oxygen permeation may delay loss of
adhesion long enough so that there is adequate corrosion protection for many practical
conditions.
7.3.2. Adhesion for Corrosion Protection
Chapter 6 deals broadly with adhesion, but wet adhesion, which is especially critical to
corrosion protection, is reviewed here. Good dry adhesion must be taken as a given for
achieving corrosion protection. If there is no coating left on the substrate, it cannot
protect the steel. It has not been so obvious, however, that good wet adhesion is required.
Good wet adhesion means that the adsorbed layer of the coating will not desorb when
water permeates through the film.
The first step to obtaining good wet adhesion is to clean the steel surface before painting, especially to remove any oils and salts. Application of phosphate conversion coatings
gives further advantages (Section 6.4.1). Various types of steel and coated steel may
require different cleaning and treatment methods [22]. In most cases, after the phosphate
coating is applied, the treated metal is rinsed with chromic acid solution as a “sealing
coat.” Waste disposal problems are severe, especially considering the chromic acid
wash step. Soluble chromium(VI) compounds have been shown to be carcinogenic and
are being replaced as rapidly as possible. Toxic hazards are also associated with soluble
nickel compounds.
It is reported that addition of polyethyleneimine to a Ca – Zn phosphating treatment
bath gives a satisfactory conversion treatment without a chromate rinse [23]. It has
been shown that a dilute solution of trimethoxymethylsilane, with enough H2ZrF6 to
bring the pH down to about 4, gave even better performance than a chromate rinse as a
sealing treatment for phosphated steel [24]. Treatment with a zinc –manganese phosphate
solution containing silicotungstates, followed by a hexafluorozirconic acid rinse, has
also been disclosed [25]. A nickel and chromium(VI) free treatment system for steel
and galvanized steel is carried out by washing with an alkaline cleaner, rinsing, treating
with hexafluorozirconic acid, rinsing, and then sealing with a solution of epoxy
phosphate [26].
For many applications, aluminum does not have to be treated for corrosion control,
due to the coherent aluminum oxide surface of the aluminum. But if exposure to salt
is to be expected, the surface must be treated before applying a coating. Chromate
surface treatments have been the standard in the industry, but with the concern about
carcinogenicity of chromium(VI), alternative chromium(VI) free treatments have been
developed, including chromium(III) sulfate plus potassium hexafluorozirconate [27]; an
aqueous solution of potassium manganate, potassium fluoride, potassium hydroxide,
sodium hydrosulfite, and ortho-phosphoric acid [28]; vanadium tetrafluoroborate
(among similar compounds) [29], and chromium(III) sulfate plus potassium hexafluorozirconate [30]. Treatment of aluminum alloys with a waterborne solution of a mixture of
bis(trimethoxypropylsilyl)amine and triacetoxysilane is another approach. The treated
surface is said to provide corrosion protection equal to that of chromate coatings when
coated with powder coatings [31].
7.3. CORROSION PROTECTION BY INTACT COATINGS
145
DC cathodic polymerization of trimethylsilane in a closed system is used to plasma coat
the surface of aluminum-coated aluminum alloys. The coated surface provides corrosion
protection and good adhesion for epoxy primers [32]. Surface treatment of aluminum with
combinations of tetraethoxysilane and 3-glycidyloxypropyltrimethoxysilane showed
that treatments with a high ratio of 3-glycidyloxypropyltrimethoxysilane and a low
ratio of water to alkoxy groups gave superior corrosion protection [33]. A patent
application for the treatment of cadmium and zinc-nickel coated steel with basic chromium(III) sulfate and potassium hexafluorozirconate and potassium hexafluorosilicate has
been disclosed [34].
Bis(trialkoxysilyl)alkanes are being investigated to treat the surface of steel. Clean steel
is treated with an aqueous solution of bis(triethoxysilyl)ethane (BTSE). BTSE is hypothesized to react with water and hydroxyl groups on the steel to give a water-resistant anchor
on the steel. This hypothesis is supported by a large body of indirect evidence, although
direct evidence is in short supply. After drying, the treated metal can be coated and
baked. For some types of coatings it is desirable to react the BTSE treated steel with a reactive silane that reacts with silanol groups of the BTSE and provides a functional group to
react with a coating binder [35]. A mixture of bis(trimethoxysilylpropyl)amine and vinyltriacetoxysilane has been reported to provide corrosion protection equal to chromate
primers on aluminum, galvanized steel, stainless steel, steel, and cobalt/chromium
alloy [36,37]. A combination of BTSE and a ureidotriethoxysilane to treat steel, galvanized steel, zinc, and aluminum has also been disclosed [38]. It is reported that galvanized
steel treated with zirconium nitrate followed by aminoethylaminopropyltrimethoxysilane
and a polyester primer gives a performance equal to that of commercial chromate primed
metal [39]. A review of potential uses of silane treatments for steel, galvanized steel, and
aluminum is available [40].
An investigation of the effect of treatment of the surface of galvanized steel with rare
earth nitrates has been published [41]. Lanthanum nitrate was particularly effective. The
results are based on electrochemical testing of samples without organic coatings. Tests
with organic coatings will be required before the utility of the treatment can be
determined.
After cleaning and treating, a surface should not be touched and should be coated as
soon as possible. Finger prints leave oil and salt on the surface. After exposure to high
humidity, fine blisters can form, disclosing the identity of the miscreant by the finger
prints. A rusty handprint was once observed on a ship after only one ocean and lake
passage [10]. It is critical to avoid having any salt on the metal surface when the
coating is applied near the ocean.
It is also critical to achieve as complete penetration as possible into the micropores and
irregularities in the surface of the steel to prevent pockets that could fill with water and
oxygen (Section 7.3.1). An important factor for achieving penetration is that the viscosity
of the liquid coatings external phase be as low as possible and remain low long enough to
permit complete penetration (Section 6.1). It is desirable to use slow evaporating solvents,
slow cross-linking coatings, and when possible, baking primers. Macromolecules may be
large compared to the size of small crevices, so lower molecular weight components may
give better protection.
Wet adhesion requires that the coating be adsorbed strongly on the surface of the steel
and not be desorbed by water that permeates through the coating. Empirically, it is found
that wet adhesion is enhanced by having several adsorbing groups scattered along the resin
chain, with parts of the resin backbone being flexible enough to permit relatively easy
146
CORROSION PROTECTION BY COATINGS
orientation and other parts rigid enough to assure that there are loops and tails sticking up
from the surface for interaction with the rest of the coating. Thus, wet adhesion will be
better if the coating has a Tg lower then the temperature at which corrosion protection
is needed, since desorption will be minimized. However, as discussed in Section 7.3.3,
a higher Tg is preferable for reducing oxygen and water permeability. Another reason
that baking primers commonly provide superior corrosion protection is that at the
higher temperature, there may be greater opportunity for orientation of resin molecules
at the steel interface. Amine groups are particularly effective polar substituents for promoting wet adhesion. Perhaps water is less likely to displace amines than to displace
other groups from the surface. Phosphate groups also promote wet adhesion. For
example, epoxy phosphates have been used to enhance the adhesion of epoxy coatings
on steel [42]. Phosphate-functional (meth)acrylic esters as comonomers increase the
wet adhesion of solventborne and latex coatings [43].
Saponification resistance is another important factor in wet adhesion [44,45]. Corrosion
generates hydroxide ions at the cathode, raising the pH levels as high as 14. Ester groups in
the backbone of a binder can be saponified, degrading the polymer near the interface and
reducing wet adhesion. Epoxy – phenolic primers are an example of high bake primers that
are completely resistant to hydrolysis. In some epoxy – amine primers, there are no hydrolyzable groups. Amine-terminated polyamides, which are used widely in air dry primers to
react with epoxy resins, have amide groups in the backbone that can hydrolyze. However,
amides are more resistant than esters to base-catalyzed hydrolysis. Alkyd resins are used
when only moderate corrosion protection is required and low cost is important. The
superior corrosion protection of epoxy ester primers is often attributed to greater resistance
to saponification of epoxy esters.
Water-soluble components that may stay in barrier primer layers should be avoided
because they can lead to blister formation. For example, zinc oxide is generally an undesirable pigment to use in primers. Its surface interacts with water and carbon dioxide to form
zinc hydroxide and zinc carbonate, which are somewhat soluble in water and can lead to
osmotic blistering. Insoluble pigments with water-soluble contaminants should also be
avoided. Passivating pigments, discussed in Section 7.4.2, cannot function unless they
are somewhat soluble in water; their presence in coating layers therefore leads to blistering. Funke showed that hydrophilic solvents, which become immiscible in the drying film
as other solvents evaporate, can be retained as a separate phase and lead to blister formation [14].
7.3.3. Factors Affecting Oxygen and Water Permeability
Many factors affect the permeability of coating layers to water and oxygen [46]. Water and
oxygen can permeate, to some extent at least, through any amorphous polymer film, even
though the film has no imperfections such as cracks or pores. Small molecules travel
through the film by jumping from free volume hole to free volume hole. The free
volume increases as the temperature increases above Tg. Therefore, normally, one
wants to design coatings with a Tg above the temperature at which corrosion protection
is desired. Since cross-linking reactions become slow as the increasing Tg of the crosslinking polymer approaches the temperature at which the reaction occurs and become
very slow at T , Tg, air dry coatings cannot have Tg values much above ambient temperatures. Furthermore, since wet adhesion is improved by having a lower Tg (Section 7.3.2),
compromises may be necessary. Permeability is also affected by the solubility of oxygen
7.3. CORROSION PROTECTION BY INTACT COATINGS
147
and water in a film. The variation in water solubility can be large. The permeability of
oxygen and water through coating layers is reviewed in Ref. [47].
Oxygen dissolves to only a small degree in any of the resins used in coatings, but there
are many other factors that affect permeability of films to oxygen, related primarily to free
volume. Hydrophilic polymers show less permeability to O2 than do hydrophobic polymers because hydrogen bonding reduces free volume availability. Lateral side chains
increase free volume and increase permeability. The higher Tg values that can be
reached with baked coatings may be another factor in their generally superior corrosion
protection. In general, higher cross-link density leads to lower permeability. Both Tg
and cross-link density affect other coating properties, so that some compromise
between Tg and cross-link density and performance must be accepted.
Permeability of coatings to water is affected by the same factors that affect permeability
to oxygen, but there are additional factors as well. Water is more soluble than oxygen in
many coatings, and a higher solubility increases permeability. Water can act as a plasticizer for coatings such as epoxy – amines and polyurethanes; the swelling caused by water
increases internal stress that can lead to delamination [48]. Internal stress increases
when a coating is cycled through wet and dry stages.
Salt groups on a polymer increase the solubility of water in coatings. This makes it difficult to formulate high performance air dry, water-reducible coatings that are solubilized
in water by amine salts of carboxylic acids. Although to a lesser degree than salts, resins
made with polyethylene oxide backbones are likely to give higher water permeabilities.
On the other hand, water has low solubility in halogenated polymers; hence vinyl chloride
and vinylidene chloride copolymers and chlorinated rubber are commonly used in formulating top coats for corrosion resistance. Fluorinated polymers have low permeabilities and
good wetting properties; hydroxy-functional poly(vinylidene fluoride) cross-linked with
polyisocyanates is reported to give good corrosion protection even as a single coat [49].
Although silicones give low surface tension and are used as water repellants, films of silicone resins are highly permeable to water.
Pigmentation can have significant effects on water and oxygen permeability. Oxygen and
water molecules cannot pass through pigment particles; therefore, permeability decreases as
pigment volume concentration (PVC) increases. However, if the PVC exceeds the critical
pigment volume concentration (CPVC), there are voids in the film and the passage of
water and oxygen through the film is facilitated. (PVC and CPVC are discussed in
Chapter 22.) Some pigments have high-polarity surfaces that adsorb water, and in cases
in which water can displace polymer adsorbed on such surfaces, water permeability can
be expected to increase with increasing pigment content. As explained in Section 7.3.2, pigments should be insoluble and as free as possible of water-soluble impurities. In theory, the
use of hydrophilic pigment dispersants should be avoided, or at least minimized.
Pigments with platelet shaped particles can reduce permeability rates as much as fivefold when they are aligned parallel to the coating surface [15,50]. Since oxygen and water
vapor cannot pass through the pigment particles, the presence of aligned platelets can
reduce the rate of vapor permeation through a coating. The alignment is critical to the
action of the platelets; if they are not aligned, permeability may be increased, especially
if the film thickness is small relative to the size of the platelets. A factor favoring alignment is shrinkage during solvent evaporation. (See Section 30.1.2 for a discussion of
additional factors.) Mica, talc, micaceous iron oxide, glass flakes, and metal flakes are
examples of such pigments. Aluminum flake is widely used; stainless steel and nickel
platelets, although more expensive, have greater resistance to extremes of pH. When
148
CORROSION PROTECTION BY COATINGS
appearance permits, use of leafing aluminum pigment in the top coat is particularly effective (Section 20.2.5). Leafing aluminum is surface treated, so its surface free energy is very
low. As a result, the platelets come to the surface during film formation, creating an almost
continuous barrier. In formulating coatings with leafing aluminum, it is necessary to avoid
resins and solvents that displace the surface treatment from the flakes.
A Monte Carlo simulation model of the effect of several variables on diffusion through
pigmented coatings has been devised [51]. The model indicates, as would be expected, that
finely dispersed, lamellar pigment particles at a concentration near, but below, CPVC give
the best barrier performance.
There are advantages to applying multiple layers of coatings. The primer can be
designed for optimal wet adhesion and penetration into the substrate surface without particular concern about other properties. Intermediate and top coat(s) can provide for
minimum permeability and other required properties. The primer film does not need to
be thick as long as the top coat is providing barrier properties; the lower limit is probably
controlled by the need to assure coverage of the entire surface. Funke has reported good
results with a 0.2-mm primer thickness [17] or a 10-nm layer of a wet adhesion promoting
polymer. Another advantage of applying multiple coats is the decrease in probability that
any area of the substrate will escape being coated.
Film thickness affects the time necessary for permeation through coatings. Thicker films
are expected to delay the arrival of water and oxygen at the interface but are not expected to
affect the equilibrium condition. Were this the only factor, the corrosion protection afforded
by intact films would be expected to be essentially independent of film thickness. However,
film thickness affects the mechanical performance of films. For example, erosion losses
would take longer to expose bare metal as film thickness increases. On the other hand,
the probability of cracking on bending increases as film thickness increases, so there may
be some optimum film thickness for the maintenance of an intact film. However, in
certain air dry, heavy duty maintenance coatings, there is a film thickness, dependent on
the coating, that provides a more than proportional increase in corrosion protection relative
to thinner films. Commonly, this film thickness is as much as 400 mm or more. Funke
suggests that below certain coating thicknesses, there may be microscopic defects extending
down through the film to the substrate [52]. The film may look intact, but there may be
microscopic defects that are large compared to the free volume holes through which permeation in fully intact films occurs. A potential source of such defects is cracks resulting
from shrinkage of films as the last solvent is lost from a coating, particularly when Tg of
the solvent-free system is around ambient temperature. Funke suggests that if the layer is
thick enough, such defects may not reach the substrate, hence substantially reducing
passage of water and oxygen. This hypothesis is consistent with the general observation
that greater protection is achieved by applying more coats to reach the same film thickness.
In line with this proposal, the use of barrier platelet pigments permits a reduction in the
required film thickness without loss of protection. The platelets may minimize the probability of defects propagating through the film to the substrate. Such defects are less
likely to occur in baked films, and this may be another factor in the generally superior corrosion protection afforded by baked films, even though thinner film thicknesses are used.
7.4. CORROSION PROTECTION BY NONINTACT FILMS
There are situations in which it is not possible to assure full coverage of the entire metal
surface as required for barrier coatings. Furthermore, coatings that were intact initially
7.4. CORROSION PROTECTION BY NONINTACT FILMS
149
may be damaged during their service lives, even those designed to minimize the probability of mechanical failure. In such cases it is generally desirable to design coatings
to suppress electrochemical reactions rather than primarily for their barrier properties.
7.4.1. Minimizing Growth of Imperfections: Cathodic Delamination
If there are gouges through the film down to bare metal, water and oxygen reach the metal
and corrosion starts. If the wet adhesion of the primer to the metal is not adequate, water
creeps under the coating, and the coating comes loose from the metal over a wider and
wider area. Poor hydrolytic stability can be expected to exacerbate the situation. This
mode of failure is called cathodic delamination. Control of cathodic delamination requires
wet adhesion and saponification resistance. It has also been shown that blisters are likely to
develop under a film near the location of a gouge [10,15].
When wet adhesion varies on a local scale, filiform corrosion can occur [53]. It is
characterized by development of thin threads of corrosion wandering randomly under
the film, but never crossing another track. Formation of these threads often starts from
the edge of a scratch. At the head of the thread, oxygen permeates through the film, and
cathodic delamination occurs. The head grows following the path of poorest wet adhesion.
Behind the head, oxygen is consumed by oxidation of ferrous ions and ferric hydroxide
precipitates, passivating the area, explaining why threads never cross. Since the ion concentration decreases, osmotic pressure drops, and the thread collapses, but it leaves a discernible rust track. Filiform corrosion can be difficult to see through pigmented films.
7.4.2. Primers with Passivating Pigments
Passivating pigments promote formation of a barrier layer over anodic areas, passivating
the surface (Section 7.2.1). To be effective, such pigments must have minimum solubility.
However, if the solubility is too high, the pigment would leach out of the coating film too
rapidly, limiting the time that it is available to inhibit corrosion. For the pigment to be
effective, the binder must permit diffusion of water to dissolve the pigment. Therefore,
the use of passivating pigments may lead to blistering after exposure to humid conditions.
Such pigments are most useful in applications in which the need to protect the steel substrate after film rupture has occurred outweighs the desirability of minimizing the probability of blistering. They are also useful when it is not possible to remove all surface
contamination (blistering will probably occur anyway) or when it is not possible to
achieve complete coverage of the steel by the coating.
Red lead pigment, Pb3O4 containing 2 to 15% PbO, has been used in oil primers as a
passivating pigment since the mid-nineteenth century. Red lead in oil primers were used
for air dry application over rusty, oily steel. The mechanisms of action are not fully understood. They presumably include oxidation of ferrous ions to ferric ions followed by
co-precipitation of mixed iron –lead salts or oxides. The somewhat soluble PbO raises
the pH and neutralizes any fatty acids formed over time by hydrolysis of the drying oil.
Toxic hazards of red lead have resulted in widespread prohibition of its use.
The utility of chromate pigments for passivation of steel and aluminum is well established. Various mechanisms have been proposed to explain their effectiveness [10]. All the
proposed mechanisms require that the chromate ions be in aqueous solution. Like all passivators, chromate ions accelerate corrosion at low concentrations. The critical minimum
21
. The critical
concentration for passivation at 258C is approximately 1023 mol CrO22
4 L
concentration increases with increasing temperature and increasing NaCl concentration.
150
CORROSION PROTECTION BY COATINGS
Sodium dichromate is an effective passivating agent but would be a poor passivating
21
). It would be rapidly
pigment, owing to its high solubility in water (3.3 mol CrO22
4 L
leached out of a film and would probably cause massive blistering. At the other
21
) that
extreme, lead chromate (chrome yellow) is so insoluble (5 1027 mol CrO22
4 L
it has no electrochemical action.
“Zinc chromates” have been widely used as passivating pigments. The terminology is
poor, since zinc chromate itself is too insoluble and could promote corrosion, rather than
passivate. Zinc yellow pigment is [K2CrO4 . 3ZnCrO4 . Zn(OH)2 . 2H2O]. [Alternative ways
of giving the same composition are Zn4K2Cr4O20H6 and (4ZnO . K2O . 4CrO3 . 3H2O).] It
21
at 258C) and has been widely
has an appropriate solubility (1.1 1022 mol CrO22
4 L
used in primers. Zinc tetroxychromate [ZnCrO4 . 4Zn(OH)2, also written as
L21), but is
Zn5CrO12H8] has a solubility lower than desirable (2 1024 mol CrO22
4
used in wash primers. Phosphoric acid is added to wash primers before application; it
may be that this changes the solubility so that the chromate ion concentration is raised
to an appropriate level. Strontium chromate (SrCrO4) has an appropriate solubility in
L21) and is often used in primers, especially latex paint
water (5 1023 mol CrO22
4
primers, in which the more soluble zinc yellow can cause problems of package stability.
It has been established that zinc chromates, and presumably other soluble chromates,
are carcinogenic to humans. They must be handled with appropriate caution. In some
countries, their use has been prohibited, and prohibition worldwide is probable in the
future. Enormous efforts have been undertaken to develop less hazardous passivating pigments [54]. However, it is difficult to conclude from the available literature and supplier
technical bulletins how these pigments compare with each other and with zinc yellow. In
some cases, a formulation that has been optimized for one pigment is compared to a formulation containing another pigment that may not be the optimum formulation for that
pigment. (A common example is the substitution of one pigment for another on an
equal weight basis rather than formulating to the same ratio of PVC to CPVC; the
results could be very misleading, since primer performance is quite sensitive to the
PVC/CPVC ratio.) Much of the published data is based on comparing corrosion resistance
in salt fog chamber tests (or other laboratory tests) rather than on actual field experience. A
problem with accelerated tests is that an important factor in the performance of a pigment
is the rate at which it is leached from a film. As discussed in Section 7.5, there is no
laboratory test available that provides reliable predictions of field performance.
Basic zinc and zinc – calcium molybdates are said to act as passivating agents in the presence of oxygen, apparently leading to precipitation of a ferric molybdic oxide barrier
layer on the anodic areas. Barium metaborate is the salt of a strong base and a weak
acid. It may act by increasing the pH, thus lowering the critical concentration of
oxygen required for passivation. To reduce its solubility in water, the pigment grade is
coated with silicon dioxide. Even then, the solubility may be too high for use in long-term
exposure conditions. Zinc phosphate, Zn3(PO4)2 . 2H2O, has been used in corrosion
protective primers and may act by forming barrier precipitates on the anodic areas.
There is considerable difference of opinion as to its effectiveness. Calcium and barium
phosphosilicates and borosilicates are being used increasingly; they may act by increasing
pH. Calcium tripolyphosphate has also been recommended [55].
A series of passivating pigments were evaluated in a waterborne epoxy coating [56].
Zinc iron phosphate, zinc aluminum phosphate, zinc molybdenum phosphate, and basic
zinc phosphate all performed well in laboratory tests; zinc phosphate was slightly
poorer. Field performance remains to be demonstrated. It has been reported that such
7.4. CORROSION PROTECTION BY NONINTACT FILMS
151
dual metal phosphate pigments provide passivation superior to that of zinc chromate under
acidic conditions but inferior to that of zinc chromate under alkaline conditions [57]. Zinc
aluminum phosphate has been shown to be a good corrosion inhibitor for waterborne
epoxy –amine coatings [58]. A study of corrosion protection of steel by zinc phosphate,
calcium zinc phosphate, zinc ferrite, and calcium-exchanged silica each in a styrene –
acrylic latex and a fatty acid modified polyurethane dispersion has been reported. The
corrosion resistance was evaluated by a series of tests, including EIS and Prohesion.
The polyurethane vehicle gave better resistance than the latex, and the best combination
was zinc ferrite with polyurethane; next best was calcium zinc phosphate [59].
These are all inorganic pigments; a wider range of potential oxidizing agents and bases
would seem to be available if organic pigments were used. An example of a commercially
available organic pigment is the zinc salt of 5-nitroisophthalic acid. It is reported to be as
effective as zinc yellow at lower pigment levels. However, since over a long time,
any passivating pigment will be lost by leaching, it seems doubtful that an equal lifetime could be achieved at a substantially lower pigment content. The zinc salt of
2-benzothiazoylthiosuccinic acid has been recommended as a passivating agent. A recurrent problem with organic passivators is that many perform extremely well in the early
parts of tests but lose effectiveness later.
Sinko reviewed the subject of passivating pigments and concluded that no other practical inorganic can match or surpass the performance of strontium chromate, but that
hybrid inorganic– organic pigments can equal or excel its performance [60]. The same
author provides manufacturing procedures for inorganic– organic anticorrosion pigments
such as Zn(NCN)2/Zn(2-mercaptobenzothiazole) [61] and discloses that the zinc salt of
2,5-dimercapto-1,3,4-trithiazole outperforms strontium chromate in protecting aluminum,
and the triisocyanuric acid salt of dicyclohexylamine with strontium doped amorphous
silica outperforms chromate on galvanized steel [62].
The morpholine salt of toluylpropionic acid, the zirconium complex of toluylpropionic
acid, metal salts of benzothiazoylthiosuccinic acid, and the tridecylammonium salt of benzothiazoylthiosuccinic acid are reported to impart corrosion inhibition and, further, that
hindered amine light stabilizers (HALS) and UV absorbers improve corrosion protection
[63]. Benzothiazoylthiosuccinic acid and calcium-exchanged silica enhance corrosion
resistance of powder coatings; the morpholine salt of toluylpropionic acid improves the
corrosion resistance of 2K epoxy –amine primers [63]. Copper can be protected by electropolymerizing 2-mercaptobenzimidazole on the surface [64].
7.4.3. Cathodic Protection by Zinc-Rich Primers
Zinc-rich primers are another approach to protecting steel with nonintact coatings. They
were designed initially to provide the protection given by galvanized steel, but have
been found to provide excellent protection to steel structures after fabrication [65].
The primers contain high levels of powdered zinc—over 84 wt% is usual. On a volume
basis, the zinc content exceeds CPVC, to assure good electrical contact between the
zinc particles and with the steel. Furthermore, when the PVC is greater than the CPVC,
the film is porous, permitting water to enter, completing the electrical circuit. The
CPVC of zinc powder varies, depending primarily on particle shape and particle size distribution; values on the order of 67% have been reported [66]. The zinc serves as a sacrificial anode, and zinc hydroxide is generated in the pores. The structure of the zinc
particles can affect performance. Smaller particle size spherical zinc gives better results
152
CORROSION PROTECTION BY COATINGS
than larger particle size zinc. Lamellar zinc particles give films with better mechanical
properties and satisfactory corrosion protection with a lower weight fraction than with
spherical zinc particles [67].
Vehicles for zinc-rich primers must be saponification resistant; alkyds are not appropriate resins for this application. Both organic and inorganic binders are widely used. Among
organic binders, epoxies are the most commonly used, and polyurethanes are gaining
favor. The principal inorganic vehicles are oligomers derived from tetraethyl orthosilicate
by controlled partial hydrolysis with a small amount of water (Section 16.3). Ethyl or isopropyl alcohol is used as the principal solvent, since an alcohol helps maintain package
stability. After application, the alcohol evaporates, and water from the air completes the
hydrolysis of the oligomer, to yield a film of polysilicic acid partially converted to zinc
salts. Cross-linking is affected by relative humidity (RH); properties can be affected
adversely if the RH is low at the time of application.
Properly formulated and applied, zinc-rich primers are very effective in protecting
steel against corrosion. Their useful lifetime is not completely limited by the amount
of zinc present, as one might first assume. Initially, the amount of free zinc decreases
from the electrochemical reaction; later, loss of zinc metal becomes slow, but the
primer continues to protect the steel for some time if the coating remains intact. Possibly,
the partially hydrated zinc oxide formed in the initial stages of corrosion of the zinc fills
the pores and, together with the remaining zinc, acts as a barrier coating [66,68]. It is also
possible that zinc hydroxide raises the pH to the level at which oxygen can passivate
the steel.
Zinc is expensive, especially on a volume basis. Early attempts to replace even 10% of
the zinc with low-cost inert pigment caused a serious decrease in performance, presumably
due to a decrease in metal to metal contact, even though the PVC was above CPVC. A
relatively conductive inert pigment, iron phosphide (Fe2P), has shown promise. In ethyl
silicate –based coatings, up to 25% of the zinc can be replaced with Fe2P [69];
however, with epoxy –polyamide coatings, replacement of part of the Zn with Fe2P
leads to a reduction in protection [70].
Zinc-rich primers are frequently top coated to minimize corrosion of the zinc, to protect
against physical damage, and to improve appearance. Formulation and application of top
coats require care. If the vehicle of the top coat penetrates the pores in the primer film,
conductivity of the primer may be substantially reduced, rendering it ineffective. (See
Section 33.1.2 for further discussion.) Zinc-rich primers can be formulated so that the
painted objects can be welded.
In response to the need to reduce VOCs, waterborne inorganic zinc-rich primers have
been developed. The chemistry of silicate binders is reviewed in Ref. [71]. The binder is a
combination of potassium, sodium, and/or lithium silicates with a dispersion of colloidal
silica [72]. The ratio of silicate to SiO2 is critical; if there is insufficient base, phase separation can occur. Commonly, a ratio of 1 to 4 or 5 is appropriate. The ratio can be changed
by adding amorphous silica. High humidity impairs drying. High temperature accelerates
drying. Carbon dioxide decreases pH and brings the silicate solution nearer the instability
region. Anodic inhibition of the metal surface with amines before application reduces problems encountered in drying at high humidity. Hydroxy-functional silicones promote
water compatibility, as does addition of methyltrimethoxysilane. It is critical to use zinc
dust that has not reacted with CO2 and water to give a surface of zinc carbonate and
zinc hydroxide [73]. Excellent performance on oil and gas production facilities in
7.5. EVALUATION AND TESTING
153
marine environments has been reported [74]. Organic zinc– rich primers based on
water borne epoxy vehicles are also available.
7.5. EVALUATION AND TESTING
There have been many attempts to develop laboratory tests to predict corrosion protection
by coatings, and these efforts continue. However, no laboratory test is available that can be
used to predict corrosion protection performance of a new coating system. This unfortunate situation is an enormous obstacle to research and development of new coatings,
but it must be recognized and accommodated.
Use testing is still the only reliable approach: that is, to apply a coating system and then
observe its condition over years of actual use. The major suppliers and end users of coatings
for such applications as bridges, ships, chemical plants, and automobiles have collected data
correlating the performance of various systems over many years. These data provide a basis
for selection of current coatings systems for particular applications. They also provide insight
into how new coatings could be formulated to improve chances of success.
Simulated tests are the next most reliable tests for predicting performance. One
common approach is to expose laboratory prepared panels on test fences in inland south
Florida or on beaches in south Florida or North Carolina. The difficulties in developing
tests to simulate corrosion in marine environments are discussed in Ref. [75]. Test conditions must simulate actual use conditions as closely as possible. For example, exposure
at higher temperatures may accelerate corrosion reactions; however, oxygen and water
permeability can be affected by T 2 Tg. If use temperatures are below Tg but tests are
run above Tg, no correlation should be expected.
Variables in the preparation of test panels are frequently underestimated. The steel used
is a critical variable [76], as are how the steel is prepared for coating and how the coating is
applied. Film thickness, evenness of application, flash off time, baking time and temperature, and many other variables affect performance. Results obtained with carefully prepared and standardized laboratory panels can be quite different than results with actual
production products. In view of these problems, it is desirable, when possible, to paint
test sections on ships, bridges, chemical storage tanks, and so on, and to observe their condition over the years. The long times required for evaluation are undesirable, but the
results can be expected to correlate reasonably with actual use.
Since wet adhesion is so critical to corrosion protection, techniques for studying wet
adhesion can be very useful. Electrochemical impedance spectroscopy (EIS) is widely
used to study coatings on steel. A series of papers discuss the theory, application, interpretation, and limitations of EIS [77]. Impedance is the apparent opposition to flow of an alternating electrical current and is the inverse of apparent capacitance. When a coating layer
begins to delaminate, there is an increase in apparent capacitance. The rate of increase of
capacitance is proportional to the area delaminated by wet adhesion loss. Highperformance systems show slow rates of increase of capacitance, so tests must be continued for long time periods. EIS is very sensitive for detecting defects but cannot determine
whether the defect is characteristic of the coating system or a consequence of poor application. EIS can be very useful in comparing various coating formulations, but because of
the wide variation in actual exposure conditions, cannot reliably predict field performance.
A review of evaluation of coatings by electrochemical, physicochemical, and physical
measurements has been published [78].
154
CORROSION PROTECTION BY COATINGS
The most widely used test method for corrosion resistance is the salt spray (fog) test
(ASTM Method B-117-95). Coated steel panels are scribed (cut) through the coating in
a standardized fashion, exposing bare steel and hung in a chamber where they are
exposed to a mist of 5% salt solution at 100% relative humidity at 358C. Periodically,
the nonscribed areas are examined for blistering, and the scribe is examined to see how
far from the scribe mark the coating has been undercut or has lost adhesion. It has been
shown repeatedly that there is little, if any, correlation between results from salt spray
tests and actual performance of coatings in use [79 – 82].
Many factors are probably involved in the unreliability of the salt spray test. One major
flaw is that the test conditions are constant; no attempt is made to replicate the changes in
conditions (warm/cool, wet/dry) that occur during most field exposures. Furthermore,
outdoor exposure can have a significant effect on coating properties, and environmental
factors such as acid rain vary substantially from location to location. The application of
the scribe mark can be an important variable; narrow cuts generally affect corrosion
less than broader ones. Also, if the scribe mark is cut rapidly, there may be chattering
of cracks out from the main cut, whereas slow cutting may lead to a smooth cut. A passivating pigment with high solubility might be very effective in a laboratory test, but may
provide protection for only a limited time under field conditions, owing to the loss of
passivating pigment by leaching.
Since with intact films, it is common for the first failure to be blister formation, humidity resistance tests are also widely used (ASTM Method D-2247-94). The unscribed face
of a panel is exposed to 100% relative humidity at 388C while the back of the panel is
exposed to room temperature. Thus, water condenses continuously on the coating
surface. This humidity test is a more severe test for blistering than the salt fog test
because pure water on the film generates higher osmotic pressures with osmotic cells
under the film than the salt solution used in the salt fog test. It is common to run the
test at 608C “because it is a more severe test.” The pitfalls of this approach are obvious
in view of the previous discussion of the importance of T 2 Tg. Humidity tests do not
provide a prediction of the life of corrosion protection, but may provide useful comparisons of wet adhesion. Funke recommends testing for wet adhesion by scribing panels after
various exposure times in a humidity chamber, followed immediately by applying
pressure sensitive tape across the scribe mark and then pulling the tape off the panel
[52]. A peel adhesion test for wet adhesion has been described [83]. Wet adhesion can
also be checked after storing panels in water [84].
It is often observed that alternating high and low humidity causes faster blistering than
does continuous exposure to high humidity. A possible explanation is that intermediate
corrosion products form colloidal membranes, causing polarization and temporary inhibition of corrosion. The membranes are not stable enough to survive drying out and
aging. Another factor may be the increase in internal stress that has been reported by
cycling through wet and dry periods [48]. A large number of humidity cycling tests
have been described, commonly involving repeated immersion in warm water and
removal for several hours. In some industries, such tests have become accepted
methods of screening coatings, although their predictive value is questionable. Simply
correlating them with salt fog tests proves nothing.
A testing regimen called Prohesion (a trademark of BP Chemicals) has been reported to
correlate better with actual performance than the standard salt spray test [85]. The procedure combines care in selection of substrates that will reflect real products, use of
thin films, emphasis on adhesion checks, and a modified salt mist exposure procedure.
7.5. EVALUATION AND TESTING
155
Instead of 5% NaCl solution, a solution of 0.4% ammonium sulfate and 0.05% NaCl is
used. Scribed panels are sprayed with the mixed salt solution cycling over 24 hours, six
3-hour periods alternating with six 1-hour drying periods using ambient air. During
these cycles, water can penetrate through the film to a greater extent than in salt fog
chamber testing, in which the humidity is always at 100%; the 5% salt solution minimizes
water penetration because of reverse osmotic pressure. In some laboratories, a QUV
exposure cycle is included in the cycling regime [86].
Automotive companies have developed a variety of cycling tests, and comparisons
have been made with auto company proving ground tests and two-year on-vehicle
field results. A study of three tests, the GM Cyclic Corrosion Test 9540-P, Hoogovens’
Cyclic Test, the Chrysler Chipping Test, has demonstrated good correlation with
on-vehicle performance [87]. The focus of this study was to compare different coated galvanized and cold-rolled steels, not different coatings. The study also demonstrated the
lack of correlation with salt spray tests. Other tests that merit consideration are the
Society of Automotive Engineers test SAE J-2334 and ASTM D-5894-96, a cyclic
corrosion test [88].
Neither salt fog nor humidity tests have good reproducibility. It is common for differences between duplicate panels to be larger than differences between panels with different
coatings. Precision can be improved by testing five to eight replicate panels of each
coating system. (Commonly, decisions are based on the results from testing two or
three panels.) A further problem of evaluating panels for corrosion protection is the difficulty of detecting small blisters and rust areas underneath a pigmented coating film
without removing the film. Infrared thermography has been recommended as a nondestructive testing procedure [89]. A great deal of effort has been expended on electrical
conductivity tests of paint films and electrochemical tests of coated panels. (See Ref. [90]
for an extensive review.)
A variety of cathodic disbonding tests specifically for testing of pipeline coatings
has been established by ASTM G-8-96, G-42-96, and G-80-88 (reapproved 1992). In
these tests, a hole is made through the coating and the pipe is made the cathode of a
cell in water with dissolved salts at a basic pH. Disbonding (loss of adhesion) as a
function of time is followed. Although there is considerable variability inherent in
such tests and their utility for predicting field performance is doubtful, useful guidance
in following progress in modifying wet adhesion may be obtained not only for pipeline
coatings, but also more broadly. (For a discussion of research on cathodic delamination,
including investigation of the migration of cations through or under coating films, see
Ref. [91].)
Appleman has reported the results of an extensive survey of accelerated test methods
for anticorrosive coating performance [92]. The need for everyone in the industry to
become aware of the current testing situation and to work cooperatively to develop
more meaningful methods of testing is emphasized. The lack of laboratory test methods
that reliably predict performance puts a premium on collection of databases, permitting
analysis of interactions between actual performance and application and formulation variables. It is especially critical to incorporate data on premature failures in the database.
Availability of such a database can be a powerful tool for a formulator and may be
especially useful in testing the validity of theories about factors controlling corrosion.
In time, it may be possible to predict performance better from a knowledge of the underlying theories than from laboratory tests. Many workers feel that this is already true in
comparison with salt fog chamber tests.
156
CORROSION PROTECTION BY COATINGS
GENERAL REFERENCES
Dickie, R. A.; Floyd, F. L., Eds., Polymeric Materials for Corrosion Control, ACS Symp. Ser. 322,
American Chemical Society, Washington, DC, 1986.
Munger, C. C., Corrosion Prevention by Protective Coatings, National Association of Corrosion
Engineers, Houston, TX, 1997.
Revie, R. W., Ed., Uhlig’s Corrosion Handbook, 2nd ed., Wiley, New York, 2000.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
Revie, R. W., Ed., Uhlig’s Corrosion Handbook, 2nd ed., Wiley, New York, 2000.
Hare, C. H., J. Prot. Coat. Linings, 1998, 15(12), 17.
Uhlig, H. H., Corrosion and Corrosion Control, 2nd ed., Wiley, New York 1971.
Zypan, product bull., E. I. DuPont de Nemours & Co., Wilmington, DE, 1999.
Spinks, G. M.; et al., J. Solid State Electrochem., 2002, 6(2), 85– 100.
Sitram, S. P.; et al., J. Coat. Technol., 1997, 69(866), 65.
Moraes, S. R.; et al., Prog. Org. Coat., 2003, 48, 28.
Kousik, G.; et al., Prog. Org. Coat., 2001, 43, 286.
Zares, P.; et al., Semiconducting Polymers, ACS Symp. Seri. 735, American Chemical Society,
Washington, DC, 1999, pp. 280 – 292.
Leidheiser, H., Jr., J. Coat. Technol., 1981, 53(678), 29.
Mayne, J. E. O., Off. Dig., 1952, 24(325), 127.
Mayne, J. E. O., in Corrosion, Shreir, L. L., Ed., Butterworth, Boston, 1976, Vol. 2,
pp. 15:24– 15:37.
Leidheiser, H., Jr., Prog. Org. Coat., 1979, 7, 79.
Funke, W., J. Oil Colour Chem. Assoc., 1979, 62, 63.
Funke, W., J. Coat. Technol., 1983, 55(705), 31.
Funke, W., J. Oil Colour Chem. Assoc., 1985, 68, 229.
Funke, W., Farbe Lack, 1987, 93, 721.
Martin, J. W.; et al., J. Coat. Technol., 1990, 62(790), 25.
Perera, D. Y., Prog. Org. Coat., 1996, 28, 21.
Chuang, T.-J.; et al., J. Coat. Technol., 1999, 71(895), 75.
Funke, W., Prog. Org. Coat., 1996, 28, 3.
Perfetti, B. M., Metal Surface Characteristics Affecting Organic Coatings, Federation of
Societies for Coatings Technology, Blue Bell, PA, 1994.
Nair, U. B.; Subbaiyan, M., J. Coat. Technol., 1993, 65(819), 59.
Tang, N.; et al., Prog. Org. Coat., 1997, 30, 255.
Gottwald, K. A.; et al., U.S. patent 6,395,105 (2002).
Hauser, B. T.; et al., U.S. patent 6,312,812 (2001).
Green, J.; et al., U.S. patent 6,375,726 (2002).
Krishnaswamy, S., U.S. patent appl., 20020033208.
Dolan, S. E.; et al., PCT Int. appli., WO/0228550 (2002).
Matzdorf, C.; et al., U.S. patent 6,375,726 (2002).
Zhu, D.; van Ooij, W. J., Prog. Org. Coat., 2004, 49, 42.
REFERENCES
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
157
Yu, Q.; et al., Prog. Org. Coat., 2002, 44, 37.
Metroke, T. L.; et al., Prog. Org. Coat., 2002, 44, 295.
Matzdorf, C.; et al., U.S. patent 6,527,841 (2003).
van Ooij, W. J.; Child, T. F., Chemtech, 1998, February 26.
Zhu, D.; van Ooij, W. J., Prog. Org. Coat., 2004, 49, 42.
van Ooij, W. J.; et al., U.S. patent 6,261,638 (2001).
Song, J.; et al., U.S. patent 6,361,592 (2002).
Montemorm, M. E.; et al., Prog. Org. Coat., 2000, 38, 17.
Child, T. F.; van Ooij, W. J., Trans. Inst. Metal Finish., 1999, 77(2), 64– 70.
Montemor, M. F.; et al., Prog. Org. Coat., 2002, 44, 111.
Massingill, J. L.; et al., J. Coat. Technol., 1990, 62(781), 31.
Yang, H. S.; et al., J. Coat. Technol., 2005, 2(13), 44.
Holubka, J. W.; et al., J. Coat. Technol., 1980, 52(670), 63.
Holubka, J. W.; Dickie, R. A., J. Coat. Technol., 1984, 56(714), 43.
Thomas, N. L., Prog. Org. Coat., 1991, 19, 101.
Sangaj, N. S.; Malshe, V. C., Prog. Org. Coat., 2004, 50, 28.
Negele, O.; Funke, W., Prog. Org. Coat., 1995, 28, 285.
Barbucci, A.; et al., Prog. Org. Coat., 1996, 29, 7.
Bieganska, B.; et al., Prog. Org. Coat., 1988, 16, 219.
Bentz, D. P.; Nguyen, T., J. Coat. Technol., 1990, 62(783), 57.
Funke, W., private communication, 2005.
Bautista, A., Prog. Org. Coat., 1996, 28, 49.
Smith, A., Inorganic Primer Pigments, Federation of Societies for Coatings Technology,
Blue Bell, PA, 1988.
Vetere, V. F.; et al., J. Coat. Technol., 2001, 73(917), 57.
del Amo, B.; et al., Prog. Org. Coat., 2002, 45, 389.
Bethancourt, M.; et al., Prog. Org. Coat., 2003, 46, 280.
Galliano, F.; Landolt, D., Prog. Org. Coat., 2002, 44, 217.
Zubielewicz, M.; Gnot, W., Prog. Org. Coat., 2004, 49, 358.
Sinko, J., Prog. Org. Coat., 2001, 42, 267.
Sinko, J., U.S. patent 6,139,610 (2000).
Sinko, J., U.S. patent appl. 20020197468.
Braig, A.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2000,
pp. 360 – 372.
Trachli, B.; et al., Prog. Org. Coat., 2002, 44, 17.
Mayne, J. E. O.; Evans, U. R., Chem. Ind., 1944, 63, 109.
Leclercq, M., Mater. Technol., 1990, March, 57.
Kalendova, A., Prog. Org. Coat., 2003, 46, 124.
Feliu, S.; et al., J. Coat. Technol., 1989, 61(775), 63, 71.
Fawcett, N. C.; et al., J. Coat. Technol., 1984, 56(714), 49.
Feliu, S., Jr.; et al., J. Coat. Technol., 1991, 63(793), 31.
Parashar, G.; et al., Prog. Org. Coat., 2001, 42, 1.
Montes, E., J. Coat. Technol., 1993, 65(821), 79.
Szokolik, A., J. Prot. Coat. Linings, 1995, 12(5), 56.
158
CORROSION PROTECTION BY COATINGS
74. Montes, E., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2001,
pp. 211 –216.
75. Lee, T. S.; Money, K. L., Mater. Perform., 1984, 23, 28.
76. Groseclose, R. G.; et al., J. Coat. Technol., 1984, 56(714), 31.
77. Loveday, D.; et al., JCT Coat. Tech, 2004, 1(8), 46; 2004, 1(10), 88; 2005, 2(13), 20.
78. Sekine, I.; et al., Prog. Org. Coat., 2002, 45, 1.
79. Athey, R.; et al., J. Coat. Technol., 1985, 57(726), 71.
80. Mazia, J., Met. Finish., 1977, 75(5), 77.
81. Wyvill, R. D., Met. Finish., 1982, 80(1), 21.
82. Funke, W., in Corrosion Control by Coatings, Leidheiser, H., Jr., Ed., Science Press, Princeton,
NJ, 1979, pp. 35– 45.
83. McKnight, M. E.; et al., J. Prot. Coat. Linings, 1995, 12, 82.
84. Hemmelrath, M.; Funke, W., FATIPEC Congress Book, 1988, Vol. IV, p. 137.
85. Timmins, F. D., J. Oil Colour Chem. Assoc., 1979, 62, 131.
86. Skerry, B. S.; Simpson, C. H., NACE Corrosion ’97, Cincinnati, OH, 1991.
87. Petschel, M., Jr., Proc. 5th Automot. Corros. Prevent. Conf., SAE International, Dearborn,
MI, 1991.
88. Appleman, B. R., J. Prot. Coat. Linings, 1989, November, 71.
89. McKnight, M. E.; Martin, J. W., J. Coat. Technol., 1989, 61(775), 57.
90. Murray, J. N., Prog. Org. Coat., 1997, 30, 225; 1997, 31, 255; 1997; 31, 375.
91. Parks, J.; Leidheiser, H., Jr., FATIPEC Congress Book, 1984, Vol. II, p. 317.
92. Appleman, B. R., J. Coat. Technol., 1990, 62(787), 57.
8
Acrylic Resins
Acrylic resins are used as the primary binder in a wide variety of industrial coatings. Their
major advantages are photostability and resistance to hydrolysis.
8.1. THERMOPLASTIC ACRYLIC RESINS
Thermoplastic acrylic (TPA) polymers have many excellent properties, especially exterior
durability, but their use has declined because they require large amounts of solvent to
reduce viscosity low enough for application. Lacquers based on TPA polymers were
used for automotive top coats from the 1950s to the 1970s. Their high solvent levels
make possible brilliant metallic colors since they permit orientation of aluminum flake
pigment in the films parallel to the surface (Section 30.1.2). Acrylic lacquers are still
used to a degree in refinishing automobiles (Section 33.3) and for other specialized coatings, but as volatile organic compound (VOC) regulations become more restrictive, their
use is declining.
Solution thermoplastic acrylic polymers are prepared by chain-growth polymerization
(Section 2.2.1). They are copolymers of various (meth)acrylic monomers and, commonly,
styrene. For automotive top coats, Tg must be over 708C. Monomers are selected on the
basis of cost and effect on properties: notably, outdoor durability and Tg. Methyl methacrylate (MMA), styrene (S), and n-butyl acrylate (BA) are often used. The acrylic copolymers are used with plasticizers. Control of molecular weight (MW) in the synthesis is
critical. Film strength increases with MW, although above M̄w of about 90,000, the
change in properties with increasing MW is small. The upper end of MW is limited
because solutions of acrylic polymers with M̄w greater than about 100,000 exhibit cobwebbing on spraying; rather than atomizing to small droplets from the orifice of a spray gun,
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
159
160
ACRYLIC RESINS
the lacquer comes out as threads. As the MW increases, the viscosity of solutions increases
and the solids that can be spray-applied decreases. Viscosity is particularly affected by the
very high MW fractions; therefore, it is critical to minimize this fraction by controlling
MW distribution (M̄w/M̄n) to a narrow range (Section 2.2.1). Commercial TPAs have
M̄w of 80,000 to 90,000 and M̄w/M̄n of 2.1 to 2.3. The solids of TPA lacquers at application viscosity are in the range 11 to 13 NVV (nonvolatile volume).
8.2. THERMOSETTING ACRYLIC RESINS
Solids can be increased by using thermosetting acrylic resins (TSA). Since they are
designed to react chemically after application, they can have lower MWs. The postreaction
leads to a polymer network, which is ideally one interconnected molecule with a very high
MW and hence potentially good film properties. The cross-linked film is not soluble in any
solvent. The term thermosetting implies that the resins cross-link by themselves and
require heat to cross-link; however, most TSAs are resins bearing functional groups that
are reacted with a different functional polymer or cross-linker, sometimes at ambient
temperature.
8.2.1. Hydroxy-Functional Acrylic Resins
Hydroxy-functional acrylic resins are copolymers of non-functional monomers, often
MMA, S, and BA, with a hydroxy-functional monomer. They are cross-linked with
either melamine –formaldehyde (MF) resins (Chapter 11) or with polyisocyanates
(Chapter 12). Solvent-borne hydroxy-functional TSAs for conventional-solids coatings
were developed in the 1950s. For conventional (low) solids formulations, their M̄ns are
usually 10,000 to 20,000 with M̄w/M̄n of 2.3 to 3.3. Thermosetting acrylics are prepared
by free radical – initiated chain-growth polymerization under monomer-starved conditions,
as described in Section 2.2.1. Polymerization of conventional TSAs is easier to control
than that of TPAs, since the effect of MW on application and film properties is not as critical. Due to their lower MW, one can use less expensive aromatic hydrocarbon solvents for
the polymerization medium. However, the lower MW of TSA polymers results in more
end groups per unit weight. Therefore, the structure of the end groups can have greater
effects on performance than in the case of TPA polymers. (Reactions in free radical polymerizations that can lead to various end groups are reviewed in Ref. [1].) Azo initiators are
used predominantly, since they cause relatively few side reactions and form end groups
with minimal photochemical reactivity. Certain peroxy initiators also meet these requirements, but others leave end groups that reduce exterior durability. Chain transfer to solvent
can also result in end groups that reduce exterior durability. For example, terminal groups
resulting from chain transfer to a ketone solvent, such as methyl n-amyl ketone (MAK),
can have a deleterious effect [2].
Hydroxyl groups are often introduced by using a hydroxy-functional comonomer such
as 2-hydroxyethyl methacrylate (HEMA). If a lower Tg reactive monomer is desired,
2-hydroxyethyl acrylate (HEA) can be used, but its greater toxic hazard increases handling
costs; it is often preferable to reduce Tg by increasing the proportion of low Tg nonreactive
monomers such as n-butyl acrylate. Commercial grades of HEMA (and HEA) contain
appreciable amounts of diester, ethylene glycol dimethacrylate (EGDMA) in the case of
HEMA. Separation of the mono- and diesters is not easy since their boiling points are
8.2. THERMOSETTING ACRYLIC RESINS
161
similar. When a relatively small fraction of hydroxy-functional monomer is used, small
amounts of diester cause some branching, which does not seem to be harmful.
However, larger amounts cause higher molecular weight, broader MW distribution, and
even gelation.
2-Hydroxypropyl methacrylate (HPMA) can also be used. It costs less than HEMA and
generally has a lower diester content. HPMA is a mixture of isomers in which the secondary alcohol predominates. Since secondary alcohols are less reactive than primary
alcohols, TSAs with HPMA substituted for HEMA require higher baking temperatures,
longer baking times, or more catalyst to cure to a given cross-link density with MF
resins. For example, baking temperatures may have to be increased by 10 to 208C. Isocyanate cross-linkers also react slower with HPMA-containing TSAs, but there is the
accompanying advantage of longer pot life. Use of HPMA in place of HEMA may
reduce resistance to photoxidation as a result of the hydrogens on tertiary carbons,
which are more easily abstractable. An alternative method for introducing hydroxy functionality is to react a carboxylic acid –functional copolymer made using methacrylic (or
acrylic) acid (MAA or AA) with an oxirane such as propylene oxide. Raw material cost
is lower with propylene oxide than with HPMA, but exacting process control is required.
Another approach to making hydroxy-functional acrylics is to react an AA or MAA copolymer with a glycidyl ester such as glycidyl versatate [3]. Rapid cross-linking hydroxyfunctional acrylic resins for use in refinish coatings have also been prepared by reacting
glycidyl versatate with methacrylic acid and then copolymerizing this monomer with
BA, styrene, and MMA. Rapid cure depends on reacting the methacrylic acid and glycidyl
versatate under conditions that maximize formation of primary hydroxy groups [4].
Acrylic resins for high solids coatings can be made by polymerizing BA, BMA,
styrene, acrylic acid, and HEMA in the presence of glycidyl versatate. During the
polymerization the glycidyl versatate reacts with the carboxylic acid groups, resulting
in the formation of another hydroxy group and a long alkyl side chain. The Tg of the
resins is lower than conventional acrylics at similar MW and hence permit higher solids
coatings. To further increase the solids, the incorporation of the reaction product of
TMP with glycidyl versatate as a reactive diluent is recommended. Polyisocyanates are
used as cross-linkers [5]. Use of t-butyl methacrylate as a comonomer has been recommended. Such resins are reported to give films with better appearance. Refinish clear
coats with a VOC content of 3.5 lb/gal have been formulated [6].
An example of a hydroxy-functional TSA is an MMA/S/BA/HEMA/MAA copolymer with a weight ratio of 50 : 15 : 20 : 14 : 1, corresponding to a mole ratio of
54.3 : 15.6 : 16.9 : 11.7 : 1.5. Typical MWs are M̄w of about 35,000 and M̄n of about
15,000 (M̄w/M̄n ¼ 2.3). With the given ratio of comonomers, P̄w is about 320 and the
P̄n is about 140. For stoichiometric comparisons, one uses the number average, M̄n.
This resin has a hydroxyl equivalent weight of slightly over 900, with a number
average functionality f̄n of about 16 hydroxyl groups per polymer molecule. The small
amount of carboxylic acid –functional monomer (MAA) is introduced to reduce the
probability of pigment flocculation in a liquid coating.
The non-functional monomers (MMA, S, and BA, in this case) are selected on the basis
of their effect on Tg, exterior durability, and cost. This monomer combination provides
good exterior durability with a relatively high Tg at a moderate cost, as is appropriate
for an automotive top coat. This resin also has a relatively high level of hydroxy-functional
monomer, appropriate when relatively high cross-link density (XLD) films are desired.
For applications requiring more flexible films, such as coil coatings or exterior can
162
ACRYLIC RESINS
coatings, one would use a lower HEMA content. By adjusting the Tg and the functionality,
TSAs can be designed for a wide range of end uses. Countless variations can be prepared.
Thermosetting resins and cross-linkers should be selected or designed for use as a
system. The appropriate composition of hydroxy-functional TSAs depends on the crosslinker to be used. A TSA resin optimized for a TSA –MF coating would probably not
yield optimum properties with a polyisocyanate cross-linker. Depending on the polyisocyanate used with a TSA, the Tg of the final cross-linked films could be higher or lower
than would be obtained by cross-linking the same TSA with an MF resin. TSA resins
for polyisocyanate cross-linking tend to have lower Tg and lower f̄n than TSAs for crosslinking with MF resins, in order to compensate for the effect of the intermolecular hydrogen bonding between urethane groups in the cross-linked films (Section 12.4).
During the 1970s, major efforts were launched to make high-solids solution acrylic
resins to meet lower VOC emission requirements. It may seem that this problem is
simple; why not just reduce M̄n from 15,000 to 1500? Actually, the problem is
complex; as the MW goes down, greater care must be exercised in every aspect of
polymerization, coating formulation, and application. In polymerization, the amount of
non- or monofunctional resin must be kept to a very low fraction. Due to the larger
number of chain ends, the effect of chain transfer to solvents such as ketones, which
can introduce photoreactive end groups, is potentially more serious. More care is
needed in establishing the stoichiometric ratio of cross-linker, and as noted in Section
25.2.2, the cure window becomes narrower. Controlling film defects during application
becomes more difficult as the solids content becomes higher (Chapter 24). It is particularly
difficult to avoid sagging of spray applied baking coatings.
An example of a TSA resin has the following composition: S/MMA/BA/HEA; the
weight ratio 15 : 15 : 40 : 30; M̄w, 5200; P̄w, 54; M̄n, 2300; P̄n, 20; P̄w/P̄n ¼ 2.7; equivalent
weight, 400; f̄n ¼ 5.7 [7]. It was made for studying MF cross-linking of fairly high solids
coatings. The resin was prepared at 65 NVW in methyl amyl ketone (MAK). In comparing
this TSA with the conventional TSA described earlier in this chapter, note that M̄n has
been reduced by a factor of 6.5, whereas f̄n is reduced by only a factor of 2.8. This difference results from the higher content of hydroxy-functional monomer required in the highsolids resin. To attain about the same XLD in the final film, the low MW resin must
undergo more reactions. Although not reported, application solids of coatings formulated
with this resin would be around 45 NVV. This solids content corresponds to that used in
high-solids metallic automotive top coats. However, the solids are still too low to meet
EPA regulations for some current end uses, and even stricter regulations are anticipated.
As explained in Section 2.2.1, conventional free radical polymerization inherently
limits the extent to which MW of a TSA can be reduced. Satisfactory performance requires
that a high fraction of the molecules have at least two hydroxyl groups, but this becomes
statistically less probable as the MW is reduced. The problem can be illustrated by comparing the conventional TSA (M̄n ¼ 15,000, P̄n ¼ 140) described in Section 8.2 with a
hypothetical high-solids TSA having the same monomer ratio with M̄n ¼ 1070
(P̄n ¼ 10), which is about the level needed to formulate with VOCs around 300 gL21.
The conventional resin has an f̄n of 16 hydroxyl groups. While individual molecules
have more or less than 16 hydroxyl groups, statistically the number of molecules with
fewer than two hydroxyl groups is very low, and virtually all the molecules are capable
of cross-linking. In contrast, the f̄n of this high-solids TSA would be only about 1.2,
and a very large fraction of the molecules could not be cross-linked. Molecules with no
hydroxyl groups would either volatilize or remain in the film as plasticizers, to the
8.2. THERMOSETTING ACRYLIC RESINS
163
detriment of film properties. Molecules with one hydroxyl group would terminate crosslinking reactions, leaving loose ends in the coating. The theory of elasticity predicts
that loose ends seriously diminish the mechanical properties of a network. Experimental
experience confirms that loose ends have a substantial effect on initial film properties
[8,9]. For example, each percent by weight of monohydroxy oligomer in a TSA may
reduce the Tg of the TSA – MF enamel by about 18C [8].
The problem results from both MW and functional group distribution. Figure 2.2
shows the distribution of molecules of differing degrees of polymerization of a series
of resins with P̄n ¼ 12 and P̄w/P̄n ¼ 1.07, 1.5, and 3.0. The theoretical minimum polydispersity that can be obtained by anionic polymerization is 1.07, and the theoretical
minimum achievable by free radical –initiated polymerization is 1.5. The difference is
explained by different kinetic characteristics; with anionic polymerization, the rate of
propagation can be slower than the rate of initiation, while with free radical polymerization, the rate of propagation is generally faster. It is difficult to approach the theoretical
minima in practice; P̄w/P̄n ¼ 2.5 is typical of reasonably well controlled free radical
polymerizations. The other factor involved is sequence length distribution, which is
the distribution of sequences of different lengths of non-functional monomers separating
functional monomers. If these sequence lengths are short compared to the chain length of
the low MW parts of the resin, multiple functional groups will be incorporated in the
molecules. But if the sequence lengths of non-functional monomers are long in comparison to chain lengths, molecules with one or no functional monomer units can be
expected.
Statistical methods have been used to calculate the proportions of non-functional molecules that would be formed during random copolymerization of monomer mixtures with
differing monomer ratios for different MWs and MW distributions [10,11]. Due to the
assumptions involved, the calculations are approximate. An example of such results for
S/BA/HEA (30 : 50 : 20 wt%) copolymers of varying P̄n is given in Table 8.1. For
P̄n ¼ 9.5 (M̄n ¼ 1125), about 36% of all molecules, corresponding to 13% of the
weight, is non-functional. (The mol% and wt% differ because of the MW distribution;
low MW molecules are more likely to be nonfunctional than higher MW molecules.)
The weight fraction of monofunctional oligomer is more difficult to calculate but would
be expected to exceed that of non-functional oligomer. At the higher HEA content
(30%) of the example given previously with P̄n ¼ 20, the proportion of monofunctional
oligomer molecules would be minimized. But if one wished to have volume solids of
about 70 NVV, one would need to have a P̄n of about 10 with a narrow distribution of
MW. The calculations show that the functional monomer content would have to be
very high; so high that the XLD of the cross-linked film would be too high for acceptable
film properties.
TABLE 8.1. Percentage of Non-functional Molecules Statistically
Predicted for S/BA/HEA (30 : 50 : 20 by wt%) Copolymers
P̄n
M̄n
Mol%
Wt% Nonfunctional
36.8
19.2
9.5
4357
2273
1125
15
24
36
1.8
5.8
12.8
Source: Ref. [12].
164
ACRYLIC RESINS
The viscosity of acrylic resins can be reduced by changing the proportions of comonomers to reduce Tg. However, this approach also reduces the Tg of the cross-linked film,
other factors being equal. The effect of reduced Tg on film properties can be offset to
an extent by increasing cross-link density, but if carried too far, this approach leads to
an unsatisfactory combination of properties. Methacrylate esters of relatively rigid alcohols, such as 3,3,5-trimethylcyclohexyl [12] and isobornyl [13] methacrylates, with a
compact shape can combine relatively low viscosity and high Tg. Accordingly, such
esters can be used as partial replacements for MMA and S.
MW is reduced by using chain-transfer agents, but some chain-transfer agents can have
an adverse effect on exterior durability [9]. Use of a chain-transfer agent with a functional
group is an interesting approach. For example, when 2-mercaptoethanol is used as a chaintransfer agent, initiating free radicals bearing a hydroxyl group are formed. These lead to a
high fraction of molecules with a hydroxyl group on one end, reducing the fractions of
non- and monofunctional molecules in the resin [14]. Improved film properties are
obtained, but the unpleasant odor of mercaptoethanol causes handling problems and
may leave residual odor in the resin.
High levels of initiator with close control of variation of temperature and concentrations during polymerization can minimize polydispersity. Although no basic studies
have been published, it appears that with high initiator concentration, polydispersity; is
narrower than theory predicts. Proprietary acrylic resins for general-purpose white coatings with moderately demanding durability requirements have been reported with sprayable viscosities at about 54 to 56 NVV (70 to 72 NVW) [15]. Another commercial resin
with M̄n of 1300 and M̄w/M̄n ¼ 1.7 can be formulated with class I MF resins in white coatings at about 77 NVW [16]. These resins provide hard, chemical resistant films with many
good properties, although the films tend to be brittle.
Initiator choice can be critical. Benzoyl peroxide gives wide distributions due to chain
transfer to polymer resulting from hydrogen abstraction. Azo initiators such as azobisisobutyronitrile (AIBN) give less branching; AIBN is also generally preferred over benzoyl
peroxide because exterior durability of films is superior. Initiators based on t-amyl peroxides, such as ethyl 3,3-di(t-amylperoxy)butyrate, are reported to give narrow distributions [17]. Other workers recommend initiators derived from t-butyl peroxide, such
as t-butyl peracetate [18]. A comparison of the use of azobis(2-methylbutyronitrile
(ABMBN) with t-butyl peroctanoate as initiators showed that the azo initiator gave narrower MW distributions [19]. It was also shown that with ABMBM, hydroxy-functional
acrylic resins could be polymerized in solutions of MF resin. In the presence of mercaptan
chain-transfer agent, ABMBN gave narrower MW distribution than did t-butyl peracetate.
Copolymers of hydroxy-functional allylic monomers and acrylate esters are reported to
give copolymers with more uniform distribution of hydroxy groups than copolymers with
hydroxy-functional acrylates. The allylic monomers are less reactive than the acrylate
monomers; hence, allylic monomers rarely react with each other, leading to more
uniform incorporation of hydroxy functionality. Also, they are chain-transfer agents,
which increases the amount of terminal hydroxy functionality. Three allylic monomers
were used in the study: allyl alcohol, allyl monoethoxylate, and allyl monopropoxylate.
Resins were designed for use in 2K urethane coatings [20].
Controlled free radical polymerization (CFRP), a living free radical polymerization,
permits the synthesis of block copolymers with narrow MW distribution and controlled
introduction of functional groups (Section 2.2.1.1). An acrylic resin prepared by CFRP
is being used in automotive clear coats. The solids of the SuperSolids clear coat are
8.2. THERMOSETTING ACRYLIC RESINS
165
reported to be 65% with the possibility of future coatings having up to nearly 90% solids.
At the same time the scratch and mar resistance are substantially higher. Dupont has been
awarded the EPA clean air achievement award for the work [21].
Another example of the use of CFRP is the preparation of acrylic resins for use in
powder coatings [22]. The resin had a M̄w/M̄n of 1.25, compared with 1.95 for a
similar resin made by conventional free radical polymerization. The melt viscosity at
1808C of the ATRP polymer was 11.2 Pa.s, compared with 57.2 for the conventional
resin. A powder coating using bis(N,N-di-2-hydroxyethyl)adipamide as a cross-linker
was compared with a similar powder coating made with the conventional free radical
copolymer. The ATRP powder coating gave cured films with better leveling than the conventional polymer and had superior storage stability. The lower melt viscosity due to the
narrow MW distribution resulted in better leveling. The reduced amount of very low MW
molecules reduced the fusion of the powder particles during storage.
Another approach to increasing solids is to blend acrylic polyols with other low viscosity polyols, such as polyesters. Some reduction in VOC can be achieved in this way.
The film properties, such as exterior durability and chemical resistance, with such
blends are often not quite as good as those of straight TSA – MF coatings, but they may
be adequate for many uses, and some highly weather resistant coatings have been
reported [23].
In general, acrylic resin coatings do not show good adhesion to metals and are used over
primers. However, phosphate-functional (meth)acrylate esters are now available. Resins
prepared with 1 to 4% of such a monomer are reported to give acrylic coatings with
excellent adhesion to a broad range of metals [24].
8.2.2. Acrylics with Other Functional Groups
Carbamate-functional acrylics can be cross-linked with Class I MF resins to give films
with better environmental etch resistance than MF cross-linked hydroxy-functional
acrylics while retaining the advantage of mar resistance [25] (Section 11.3.4). Carbamatefunctional acrylic resins can be prepared by reacting isocyanato-functional acrylic resins
with hydroxypropyl carbamate [26,27], by reacting acrylic resins with urea [27] and by
ester interchange between a hydroxyl group on a hydroxy-functional acrylic and the carbamate from propylene glycol monomethyl ether [28]. Use of a combination of a
carbamate-functional acrylic resin, a carbamate-functional urethane, and MF resins has
been disclosed [29]. Carbamate-functional acrylics can be blended with hydroxylfunctional resins to reduce costs while retaining satisfactory levels of mar resistance.
Trialkoxysilylalkyl-functional acrylics can be prepared using a trialkoxysilylalkyl
methacrylate as a comonomer [30]. They provide high-performance clear coats by
moisture-cure in the air (Section 16.2). Trialkoxysilyl-functional acrylics have also
been synthesized by copolymerization of vinyltrimethoxysilane and butyl acrylate.
Studies of the structure indicate that some Si22O22Si bonds form during the polymerization. Films of the resin cure rapidly with an acid catalyst by reaction with atmospheric
water [31].
Carboxylic acid– functional acrylic resins are made using acrylic or methacrylic acid as
comonomers; they are cross-linked with epoxy resins (Section 11.3.4). Note that close
quality control of acrylic acid monomer is required because it can dimerize during storage.
Acrylics with amide groups can be used in several ways. Acrylic copolymers containing acrylamide can be cross-linked with MF resins by reaction with the amide group; the
166
ACRYLIC RESINS
curing temperature required is higher than for hydroxy-functional resins. Alkoxymethyl
derivatives of acrylamide copolymers can be made by reacting the amide groups with formaldehyde, followed by etherification, analogous to the preparation of amino resins, as
described in Section 11.4.4. These resins can be used as cross-linkers for hydroxyfunctional TSAs. Alternatively, one can make alkoxymethyl derivatives of monomeric
acrylamides for use as comonomers. For example, N-(isobutoxymethyl)methacrylamide
has been used as a comonomer along with hydroxy-functional comonomers to make copolymers that are “self cross-linking” [32]. Acrylic copolymers containing acrylamide and
HEMA can also be reacted with formaldehyde and then with an alcohol to make a self
cross-linking resin.
Epoxy-functional acrylics are made using glycidyl methacrylate (GMA) as a comonomer (Section 13.3.2). Such resins are used in powder clear coats for automobiles [33 –35]
(Section 28.1.4). They can be cross-linked with dicarboxylic acids such as dodecanoic acid
[33,34] or with carboxylic acid-functional acrylic resins [35]. GMA acrylic copolymers,
which contain 4-hydroxybutyl acrylate (HBA), have also been combined with an
acrylic resin using maleic anhydride as a comonomer that has been reacted with methanol
to give a half ester. During curing the half ester cyclizes to re-form anhydride groups that
react with the hydroxy groups from the HBA; epoxy groups react with the COOH groups
generated after the anhydride reacts with the OH group. The cured films are reported to
exhibit excellent mar and acid resistance as automotive clear coats [36]. Three alternative
systems based on GMA acrylic copolymers have been proposed. A GMA copolymer can
be cross-linked with carboxy-functional polyesters. Copolymers using both GMA and
acrylic acid as comonomers are self cross-linking resins. Also, an acrylic using GMA
and t-butyl acrylate are self cross-linking; with heat the t-butyl group leaves as isobutylene, leaving a COOH to cross-link with the epoxy groups. The coatings are reported to
exhibit excellent mar and acid resistance and to be useful not only in high-solids solventborne systems but also in powder coatings and, with a nonionic emulsifier, in waterborne
coatings [37].
Isocyanato-functional acrylics can be prepared by copolymerizing isopropenyl
dimethylbenzylisocyanate (TMI) with acrylates (Section 12.3.2) [38]. They can be crosslinked with polyols and hydroxy-functional acrylic resins.
New under body plastisol coatings prepared from core – shell acrylic copolymers have
been reported. For example, an emulsion copolymer with a core of isobutyl
methacrylate with 2% 1,6-hexanediol diacrylate and a shell of MMA with 6%
1,6-hexanediol diacrylate is converted to a dry powder and dispersed in dioctyl phthalate
or tricresylphosphate [39].
8.3. WATER-REDUCIBLE THERMOSETTING ACRYLIC RESINS
Another way to reduce VOC emissions from TSA coatings is to make TSA resins that can
be diluted (reduced) with water. Such resins are sometimes called “water-soluble resins,”
but this terminology is misleading since the resins are not soluble in water. Rather, solutions of amine salts of these resins in organic solvents can be diluted with water to
form reasonably stable dispersions of polymer aggregates swollen by solvent and water.
To minimize confusion, we have chosen to use the following terminology: waterborne
is used broadly for all coatings with aqueous media, including latex coatings; waterreducible coatings is used only for waterborne coatings based on resins having hydrophilic
8.3. WATER-REDUCIBLE THERMOSETTING ACRYLIC RESINS
167
groups in most or all molecules, which excludes latexes; and water-soluble is limited
specifically to substances that are soluble in water.
A typical water-reducible acrylic resin is a copolymer of MMA/BA/HEMA/AA, in a
weight ratio of 60 : 22.2 : 10 : 7.8, prepared by free radical – initiated polymerization using
an azo initiator [40]. The polymerization is carried out at high solids (generally, 70 NVW
or higher) in a water-miscible solvent. Glycol ethers, such as 1-(n-propoxy)-2-propanol or
2-butoxyethanol, and butyl alcohols are the most widely used solvents. Except for the
solvent and the higher proportion of acrylic acid, such resins are similar to conventional
solids TSAs, having M̄w and M̄n of about 35,000 and 15,000, respectively, as described
in Section 8.3.
Water-reducible acrylic resins typically have acid numbers of 40 to 60. (The acid
number is determined by titration and is defined as milligrams of KOH required to neutralize 1 g of resin solids. Equivalent weight equals 56,100/acid number.) Residual monomer
is removed by distilling off a small fraction of the solvent, and the resin is stored as a concentrated solution. The first step in preparing a coating is to add an amine, such as
2-(dimethylamino)ethanol (DMAE). As discussed later, less than the theoretical amount
of amine required to neutralize all of the carboxylic acid groups is commonly used. The
ratio used is called the extent (of) neutralization (EN). For example, if 75% of the theoretical amount were used, the EN would be 75. Other coating components (pigments, MF
resin, sulfonic acid catalyst) are dispersed or dissolved in this solution and the coating is
diluted with water prior to application.
The change in viscosity with dilution of water-reducible resins is abnormal. An
example is given in Figure 8.1 showing the log viscosity as a function of concentration
for a model resin [41]. The height of the peak in the dilution curve is dependent on the
particular resin and formulation, and the systems are highly shear thinning in the peak
region. Another abnormality is that when water-reducible resins are neutralized and
reduced with water, their pH is over 7 (commonly 8.5 to 9.5), even though less than the
theoretical amount of amine necessary to neutralize the carboxylic acid is used. For comparison purposes, a dilution curve for the same resin using an organic solvent (t-butyl
alcohol in this case) is included in the figure, as is a typical dilution curve of a latex.
The morphology of water-reducible TSAs has been studied fairly extensively [41–43].
In Figure 8.1, the water-reducible resin is a 54% solution of a 90 : 10 butyl methacrylate
(BMA)/AA copolymer in t-butyl alcohol at 75 EN with DMAE. The solvent dilution
curve of log viscosity versus concentration is fairly linear, which is typical for most resin solutions in good solvents. The water dilution curve shows the abnormal response that is typical
for water-reducible resins. During dilution with water, two kinds of changes are occurring
simultaneously: The concentration is reduced, as is the ratio of solvent to water.
In the first stages of dilution, viscosity drops more rapidly with water than with solvent. It
is hypothesized that the viscosity is high before dilution because of association of ion pairs
on different molecules. Water associates strongly with ion pairs, separating most of the
intermolecular ion pairs and causing rapid reduction in viscosity. However, as dilution
with water continues, the viscosity levels off and then increases, passing through a
maximum. Further dilution causes a very steep drop of viscosity. At an application viscosity
of about 0.1 Pa.s, such systems typically have solids in the range 20 to 30 NVW. Thus, the
application solids of coatings made with such resins is low; however, since a major part of
the volatile components is water, VOC emissions are still low.
The viscosity changes on dilution with water are explained as follows. The amine salt of
the resin is soluble in the t-butyl alcohol and in t-butyl alcohol–water solutions in which the
168
ACRYLIC RESINS
Figure 8.1. Viscosity dependence on concentration for a 10 mol% acrylic acid copolymer, 75 EN with
DMAE, dissolved in t-butyl alcohol and then diluted with water. Also shown are curves for dilution of
the same resin with t-butyl alcohol and a typical viscosity–solids latex dilution with water. (From
Ref. [41], with permission.)
ratio of solvent to water is high. But as additional water is added, the ratio of solvent to water
decreases to the point where some of the molecules are no longer soluble in the mixed
solvent. These molecules do not precipitate in a separate macrophase, but the nonpolar segments of various molecules associate with each other to form aggregates. The predominantly
nonpolar parts of the molecules are in the interior of the aggregates, and the highly polar
carboxylic acid salt groups are on their periphery. As dilution continues, more and more
molecules join in aggregates. Since the solvent is soluble in the resin, some dissolves in
the aggregates, swelling their volume. Also, water associates with the salt groups, and
some dissolves in the solvent inside the aggregates, further swelling them. As the aggregates
form, the system changes from a solution to a dispersion of aggregates in a continuous phase.
As the number and volume of aggregates grow, the volume fraction of internal phase
increases and the aggregates get more and more closely packed, leading to an increase in
viscosity. (See Section 3.5 for discussion of the factors controlling the viscosity of liquids
containing dispersed phases.) At the maximum viscosity, the material is predominantly a
dispersion of highly swollen aggregates in water containing some solvent. The high viscosity can lead to difficulty in stirring. On still further dilution, the viscosity drops
rapidly. This drop results from two factors. There is a dilution effect, that is, the decrease
8.3. WATER-REDUCIBLE THERMOSETTING ACRYLIC RESINS
169
in volume fraction internal phase. However, the drop is even steeper than would result from
this effect alone. The balance of the decrease is due to a decrease in the swelling of the
aggregates. The partitioning of solvent and water between the aggregates and the continuous
phase changes throughout the dilution process, and as more water is added, more solvent
moves to the continuous phase, shrinking the swollen aggregates.
The viscosity of the solutions in the organic solvent is Newtonian, as it is when only
small amounts of water have been added. However, in the midrange around the peak of
the water dilution curve, the dispersions show a high degree of shear thinning [42]. The
dilute systems show little or no shear thinning. This behavior can be explained on the
basis that in the shear thinning stage, the swollen aggregate particles (internal phase)
are not rigid. When shear is applied, the particles distort, decreasing the shape factor,
increasing the packing factor, and as a result, decreasing the viscosity as a function of
increasing shear rate. With further dilution, some of the organic solvent is extracted
from the aggregates into the continuous phase, and the aggregates become smaller and
less easily distorted. In the more dilute stage, the viscosity is lower, so the shear stress
exerted on the aggregates at a given shear rate is less, reducing the probability of distortion. Thus, the flow properties are Newtonian or only slightly shear thinning at application
viscosities.
The abnormal pH effect noted earlier can be explained in a way consistent with this
picture of the morphology of the system [44]. When one neutralizes a simple carboxylic
acid, such as acetic acid, with 75% of the theoretical equivalent amount of an amine such
as DMAE, the pH is about 5.5. However, the situation with the carboxylic acid groups on
the polymer chain is different. The acrylic acid groups are relatively randomly spaced
along the polymer chain. In some cases, there are carboxylic acid groups near each
other; in other cases, there are single acrylic acid groups separated from others on
both sides by several hydrophobic ester monomer units. As aggregation occurs during
dilution, many of the carboxylic acid groups are near the surface, where all, or almost
all, are neutralized by the water-soluble amine. However, geometric factors require
that many other carboxylic acid groups reside in the interior of the aggregates; those
most widely separated from other carboxylic acid groups by hydrophobic monomer residues are presumably most likely to be “hidden” in the interior of aggregates. DMAE partitions among the continuous (water–solvent) phase, the surface region of the aggregates,
and the interior of the aggregates. Since DMAE and DMAE salts are highly soluble in
water, it is expected that DMAE concentrates in the first two regions, leaving a fraction
of carboxylic acid groups unneutralized in the interior of aggregates. Even when there is
only 75% of the amine necessary to neutralize all of the carboxylic acid groups, there is
more than enough to neutralize those carboxylic acid groups at or near the surface of the
aggregates. Part of the amine is in the continuous phase, resulting in a basic pH reading.
Since the interaction is between a weak base and a weak acid, the change of pH with
amine addition is slow. Due to this insensitivity, pH is not appropriate as a quality
control specification.
This description of the morphology of a water-reducible system accounts for its behavior but does not provide direct evidence for the presence of aggregates. Evidence was
found by studying the effects of changing solvent to water ratios at a constant resin/
amine concentration. A BMA –AA copolymer (mole ratio 84 : 16) at 75 EN with
DMAE was “dissolved” in a series of solvent/water solutions at the same concentration
of 21.2% of resin plus DMAE [42]. The viscosity of this series of dispersions goes
through a peak at 30 : 70 solvent to water. The samples were then subjected to ultrafiltration. No resin was retained on the filter in the case of samples with a ratio of 80 : 20
170
ACRYLIC RESINS
solvent to water up to pure solvent. As the solvent ratio decreased from 80 : 20 to 30 : 70,
the amount of resin retained increased until at 30 : 70, essentially all of the resin was
retained by the ultrafilter. All of the resin in samples with still less solvent was
also retained. The samples were also examined by phase contrast microscopy, which
revealed the presence of particles in all those samples where resin was retained by the
ultrafilter.
Many variables affect the morphology of these systems. The shape of the viscosity –
concentration (dilution) curve varies. In some cases, there is a very high viscosity
peak—higher than the viscosity of the original undiluted material; in other cases, there
may be only a shoulder in the dilution curve. Dependence of the viscosity on the MW
depends on solvent structure and the ratio of solvent to water, in addition to the concentration [42]. The effect of MW on viscosity is different at different stages of dilution. The
log viscosity of the all solvent solution varies with approximately the square root of the
MW. The viscosity in the peak area of the dilution curve is very dependent on MW. As
the MW increases, the viscosity of the interior of the aggregate particles is higher, so
they are more difficult to distort; hence, the viscosity of the entire system does not decrease
as much at any given shear rate. Very high peak viscosity leads to difficulty in dilution. It is
necessary to limit the MW so that the coating can be thoroughly agitated throughout the
dilution cycle with the available mixing equipment. The viscosity of systems diluted with
water to application viscosity is independent of MW. This advantage permits application
of coatings using resins having a MW in the same range as those used in conventional solventborne TSA coatings but with VOC contents substantially lower and comparable to that
of relatively high solids coatings.
The dilution behavior of TSAs depends on their carboxylic acid content. The effect of
varying the mole percent of acrylic acid in a series of BA/AA copolymers from 10 to
50% is shown in Figure 8.2. With 50% acrylic acid, the salts exhibit a viscosity dependence on dilution behavior approaching that of solution systems. As the carboxylic acid
content is reduced, the abnormal rheological properties become more pronounced. Notice
that the concentration at viscosities near those required for application are highest for the
lowest acid content systems. This is an important reason that resins of this type are
designed with the lowest carboxylic acid content (acid numbers 40 to 60) that still provides a stable dispersion at application viscosity. Low acid numbers favor formation of
relatively unswollen aggregates, keeping viscosity down, but if the acid numbers are too
low, the system separates into macrophases instead of forming a stable microphase
dispersion.
The required acid content is lower for resins with increasing hydroxyl group content.
Although salts of carboxylic acid groups are much more hydrophilic than hydroxyl
groups, hydroxyl groups are sufficiently hydrophilic to promote the solubility in water –
solvent blends, as indicated by the water solubility of the homopolymer of 2-hydroxyethyl
acrylate. Therefore, the minimum required acid content decreases as the content of
hydroxyl groups increases.
Polycarbonate-modified water-reducible acrylic resins, having pendant carbonate
groups, are synthesized by copolymerizing MMA and styrene with a monomer obtained
by reacting 5,5-dimethyl-1,3-dioxane-2-one with HEMA. Aqueous dispersions of the
resins exhibited excellent hydrolytic stability; and formulations with MF resin cured
more rapidly than did conventional water-reducible acrylic resins [45].
Various types and amounts of amine can be used [41,46,47]. Generally, less than the
stoichiometric amount of amine is used. The lower the amine content, the lower the
8.3. WATER-REDUCIBLE THERMOSETTING ACRYLIC RESINS
171
Figure 8.2. Viscosity variations during water dilutions of 42 NVW copolymers having 50, 30, 20, and 10
mol% AA, each at 75 EN with DMAE. (From Ref. [41], with permission.)
viscosity of the fully diluted systems, corresponding to higher solids at a fixed application
viscosity. Figure 8.3 shows viscosity response to dilution as a function of extent neutralization of a 90 : 10 BMA – AA resin with DMAE. For any resin – amine combination, there
is a minimum amount of amine required to give a stable dispersion at application viscosity,
that is, to prevent macrophase separation. In the example shown in Figure 8.3, the 50 EN
sample could not be fully diluted without macrophase separation. In a similar experiment
using an 80 : 20 BMA – AA resin, 50 EN with DMAE gave a stable dispersion even when
diluted until the viscosity was below 0.1 Pa.s. Viscosity at application viscosity changes
rapidly with concentration, and it is possible to overshoot a desired reduction. The viscosity is also sensitive to amine content, so if a coating has been reduced with too
much water, resulting in too low a viscosity, the viscosity frequently can be brought
back up by addition of small amounts of amine.
Another variable is the structure of the neutralizing amine. Although there may be some
effect of base strength (less amine may be required with increasing base strength), the principal variable seems to be the water solubility of the amine. The amount of amine required
for stable dispersions increases in the order DMAE , triethylamine (TEA) , tripropylamine. Hydroxyl-substituted amines are the most widely used, but morpholine derivatives
172
ACRYLIC RESINS
Figure 8.3. Viscosity as a function of weight percent resin when neutralized with varying levels of
DMAE. Resin has 10 mol% AA; dilution started at 54 wt% solids. (From Ref. [41], with permission.)
such as N-ethylmorpholine (NEM) are also effective. Potential toxic hazards of the amines
should be assessed carefully.
8.3. WATER-REDUCIBLE THERMOSETTING ACRYLIC RESINS
173
In selecting amines it is important to consider not only the effect on dispersion stability,
but also the effect on package stability and curing of the coating [44]. The amine improves
the package stability by minimizing reactions of MF resins during storage. If Class II MF
resins are to be used, the amine must be tertiary. If, as is more common, Class I MF resins
are used, primary or secondary amines can also be used.
DMAE, although widely used, has a disadvantage. It has been shown that transesterification occurs between ester groups on the acrylic resin and the hydroxyl of the amino
alcohol, thereby covalently bonding part of the amine to the resin [46]. Since bound
amine cannot volatilize during application and baking, it inhibits the acid-catalyzed crosslinking reactions. NEM, which has essentially the same boiling point as DMAE, permits
faster curing, since it cannot undergo transesterification. 2-Amino-2-methyl-1-propanol
(AMP) can be used with Class I MF resins. It provides stable dispersions and despite
its higher boiling point, gives coatings that cure more rapidly than even those neutralized
with NEM. AMP has a hydroxyl group that can transesterify, but also a primary amine that
can undergo aminolysis reactions with ester groups on the resin [46]. By reacting at both
sites, AMP can act as a cross-linking agent, supplementing cross-linking with the MF
resin. Furthermore, AMP has been shown to cyclize to an oxazolidine under curing conditions, reducing basicity and increasing the cure rate [47]. The difference in curing
response with DMAE and AMP has also been demonstrated by dynamic mechanical
analysis [48].
Another factor involved in the choice of amine is the effect on wrinkling [46]. When
using TEA as the neutralizing amine, it is common to get low gloss coatings. The low
gloss results from the development of a fine wrinkle pattern on the surface during
curing of the film, which is caused by faster curing of the surface layer than the interior
of the film. When the interior of the film does cure, it shrinks, causing the immobilized
surface of the film to wrinkle. (See Section 24.6 for further discussion of wrinkling.)
This phenomenon occurs when amine is essentially absent near the surface of the film
but still present in the lower layers of the film; the MF resin readily cross-links the
acrylic at the surface, whereas the reaction is still inhibited in the lower layers.
The probability of wrinkling increases as film thickness increases because there is more
likely to be a differential in amine content as film thickness increases. The probability of
wrinkling is also affected by amine structure. A series of amines gave decreasing probability of wrinkling in the order TEA . DMAE . NEM [46]. TEA is a relatively
strong base (pKa of conjugate acid ¼ 10.9) and hence diffuses slowly through a film containing COOH groups; it also has a relatively low boiling point (908C), so on reaching the
surface, it volatilizes rapidly. This combination of factors leads to a differential amine
content even with relatively thin films. NEM, which is a weaker base (pKa of conjugate
acid ¼ 7.8) and has a higher boiling point (1398C), can diffuse relatively more rapidly
through the film, but volatilizes more slowly from the surface. The result is a more nearly
uniform amine gradient and more uniform timing of curing of surface and lower layers of
a film. DMAE has an intermediate base strength (pKa ¼ 9.3) and a relatively high boiling
point (1348C) and exhibits intermediate behavior between that of TEA and NEM.
Melamine –formaldehyde resins are most commonly used as cross-linkers for the coatings. Class I and Class II monomeric methylolated MF resins are miscible in these
systems. They cross-link the TSA by reacting with hydroxyl and carboxyl groups. The
reaction with OH groups is faster, and the resulting ether bonds are more stable to hydrolysis than are ester bonds resulting from cross-linking with COOH groups. Thus, while the
174
ACRYLIC RESINS
cross-link density of the final film depends on the total functionality of COOH and OH
groups, it is generally desirable to attain as much of the cross-linking as possible
through OH groups and to adjust cross-link density by varying the f̄n of OH. A further
advantage of using the lowest possible level of COOH is that less amine is required.
Amines are relatively expensive and must be included as part of the VOC emissions.
2K waterborne urethane coatings use polyisocyanates as cross-linkers for water-reducible
acrylics (Section 12.7.3).
REFERENCES
1. Chung, R. P.-T; Solomon, D. H., Prog. Org. Coat., 1992, 21, 227.
2. Gerlock, J. L.; et al., Macromolecules, 1988, 21, 1604.
3. Ryan, R. W.; Walt, M. B., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 1994, p. 786.
4. Petit, H.; et al., Prog. Org. Coat., 2001 43, 41.
5. Slinckx, M.; et al., Prog. Org., Coat, 2000 38, 163.
6. Ball, P. A.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 119 –124.
7. Hill, L. W.; Kozlowski, K., J. Coat. Technol., 1987, 59(751), 63.
8. Nakamichi, T.; Shibato, K., J. Jpn. Soc. Colour Mater., 1986, 59, 592.
9. Spinelli, H. J., Org. Coat. Appl. Polym. Sci., 1982, 47, 529.
10. Hill, L.W.; Wicks, Z. W., Jr., Prog. Org. Coat., 1982, 10, 55.
11. O’Driscoll, K., J. Coat. Technol., 1983, 57(705), 57.
12. Kruithof, K. J. H.; van den Haak, H. J. W., J. Coat. Technol., 1990, 62(790), 47.
13. Wright, A. J., Eur. Coat. J., 1996, 696.
14. Gray, R. A., J. Coat. Technol., 1985, 57(728), 83.
15. Nylund, J. E.; Pruskowski, S., Resin Rev., 1989, 39, 17.
16. Joncryl 500, tech. bull. PSN 0-104-12/83, S.C. Johnson & Son, Inc., Racine, WI, 1983.
17. Kamath, V. R.; Sargent, J. D., Jr, J. Coat. Technol., 1987, 59(746), 51.
18. Myers, G. G., J. Coat. Technol., 1995, 67(841), 31.
19. Ziemer, P. D.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2002, pp. 99– 107.
20. Guo, S.-H.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 211 –223.
21. htpp://automotive.dupont.com/en/newsEvents/article20030305b.html.
22. Barkac, K. A.; et al., U.S. patent 6,670,043 (2003).
23. Okada, K., Proc. Int. Conf. Org. Coat. Sci. Technol., Athens, Greece 1997, Vol. 23, p. 337.
24. Yang, H. S.; et al., J. Coat. Technol., 2005, 2(13), 44.
25. Green, M. L., J. Coat. Technol., 2001, 73(918), 55.
26. Rehfuss, J. W.; St. Aubin, D. L., U.S. patent 5,356,669 (1994).
27. Rehfuss, J. W.; St. Aubin, D. L., U.S. patent 5,605,965 (1997).
28. Barancyk, S. V.; et al., Int. patent appl. WO 95/29947 (1995).
29. Ohrbohm, W. H.; et al., U.S. patent 6,165,618 (2000).
30. Chen, M. J.; et al., J. Coat. Technol., 1997, 69(870), 43.
31. Douskey, M. C.; et al., Prog. Org. Coat., 2002, 45, 145.
REFERENCES
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
175
Reitberg, J.; et al., Proc. Int. Conf. Org. Coat. Sci. Technol., Athens, Greece 1989, p. 209.
Ooka, M.; Ozawa, H., Prog. Org. Coat., 1994, 23, 325.
Kenny, J. C.; et al., J. Coat. Technol., 1996, 68(855), 35.
Agawa, T.; Dumain, E. D., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 1997, p. 342.
Fushimi, H.; et al., Prog. Org. Coat., 2001, 42, 159.
Flosbach, C.; Schubert, W., Prog. Org. Coat., 2001, 43, 123.
Vazken, A.; et al., U.S. patent 5,254,651 (1993).
Wang, G. J.; et al., Prog. Org. Coat., 2004, 50, 55.
Watson, B. C.; Wicks, Z. W., Jr., J. Coat. Technol., 1983, 55(698), 59.
Hill, L. W.; Richards, B. M., J. Coat. Technol., 1979, 51(654), 59.
Wicks, Z. W., Jr.; et al., J. Coat. Technol., 1982, 54(688), 57.
Armat, R.; et al., J. Appl. Polym. Sci., 1996, 60, 1927.
L. W. Hill; Wicks, Z. W., Jr., Prog. Org. Coat., 8, 161 (1980).
Nakano, S., Prog. Org. Coat, 1999, 35, 141.
Wicks, Z. W., Jr.; Chen, G. F., J. Coat. Technol., 1978, 50(638), 39.
Ferrell, P. E.; et al., J. Coat. Technol., 1995, 67(851), 63.
Hill, L. W.; et al., in Film Formation in Waterborne Coatings, Provder, T.; et al., Eds., American
Chemical Society, Washington, DC, 1996, p. 235.
9
Latexes
A latex is a dispersion of polymer particles in water. Most synthetic latexes are made by a
free radical – initiated chain-growth polymerization process in which the monomers are
emulsified in water at the start; hence, the process is called emulsion polymerization. It
is not surprising that latex paints are sometimes misnamed emulsion paints, but this terminology is best avoided, since it can cause confusion with systems that are true emulsions.
Aqueous dispersions of polymers are also prepared by other methods; for example, polyurethane dispersions in water are prepared by step-growth polymerization. Usually, these
materials are called aqueous dispersions rather than latexes; they are discussed in Section
12.7.1. To further confuse the terminology, some authors call latexes colloidal dispersions
or polymer colloids.
Molecular weights (MW) of polymers prepared by emulsion polymerization are generally high; an M̄w of 1,000,000 or higher is common. However, unlike solution polymers,
the MW of the polymer in latex particles does not affect the viscosity of the latex. Instead,
latex viscosity is governed by the viscosity of the medium in which the polymer particles
are dispersed (the continuous phase), by the volume fraction of particles, and by their
packing factor. (See the discussion of the Mooney equation in Section 3.5.) This lack of
dependence of viscosity on MW makes it possible to formulate latex coatings at higher
solids than those of solutions of polymers having high MW. Latexes are used as the principal vehicle in a large majority of architectural coatings in the United States. A growing
part of the OEM (original equipment manufacture) product and special purpose coatings
markets is latex based. Most latex paints form films by coalescence at ambient temperatures (Section 2.3.3). A limitation of most latex-based coatings is the inability to
achieve very high gloss. On the other hand, the durability of the films is superior to that
of films formed from drying oil and alkyd paints; and VOC emissions are often lower.
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
176
9.1. EMULSION POLYMERIZATION
177
9.1. EMULSION POLYMERIZATION
Emulsion polymerization is carried out in water using monomer(s), surfactant(s), and a
water-soluble initiator. Many of the same monomers are used as in solution polymerization, and the reactions are broadly similar at the molecular level. However, the physical
circumstances of polymerization are different, affecting the polymerization chemistry.
The process is easily varied, and the properties of the latex are affected by changes in
physical conditions of the polymerization.
There have been many studies of the mechanisms of emulsion polymerization;
however, there are so many variables that no general theory has been developed that
can predict the results obtained in all emulsion polymerizations. Monomer structure, solubility, and concentration; surfactant(s) structure and concentration; initiator concentration
and rate of radical generation; presence and concentration of added electrolytes; and temperature are some of the critical variables. The rate of agitation and reactor design can affect
the results as well. A qualitative discussion of proposed mechanisms is provided in Section
9.1.2. The reader is referred to the general references provided at the end of the chapter for
extensive discussion and proposed equations for predicting the number of particles and
particle sizes of latexes.
Many early laboratory studies were carried out by a small-scale batch process in which
all of the ingredients were put into a sealed bottle, which was then shaken in a temperature
controlled water bath. This process is often called the pop-bottle process. Many papers
describing the mechanism of emulsion polymerization are based on such studies.
However, batch processes cannot be used commercially because the heat evolved by
the exothermic polymerization process would be uncontrollable in a large vessel.
Instead, commercial latexes are produced by a semicontinuous batch process, using apparatus such as that shown schematically in Figure 9.1. Monomers and initiators are added in
proportions and rates such that rapid polymerization occurs. In this way, the monomer
Figure 9.1. Semicontinuous batch process production unit. (Adapted from Ref. [1], with permission.)
178
LATEXES
concentration at any time is low; the polymerization is said to be done under monomerstarved conditions. This facilitates temperature control. Furthermore, the composition
of copolymers formed under monomer-starved conditions approximately equals the
composition of the monomer feed, regardless of the relative reactivity of the various
monomers. This procedure also permits changing monomer composition during the
course of a polymerization, as discussed in Section 9.1.3. It is common to start the
polymerization in the presence of a seed latex (Section 9.1.2). A laboratory scale
process involving a seed latex is described in detail in Ref. [2].
The products obtained from batch and semicontinuous emulsion polymerizations of the
same monomers are often quite different. The differences are particularly large when
copolymers are prepared by the two methods (Sections 9.1.2 and 9.3). Care must be exercised in reading the literature since conclusions derived from studies on the batch process
are often misleading when applied to the semicontinuous process, and vice versa. Since
many variables affect the composition and properties of the latex produced, it is vital to
plan laboratory procedures to simulate eventual production operations as closely as
possible to minimize problems in scaling up to production equipment.
9.1.1. Raw Materials for Emulsion Polymerization
9.1.1.1. Monomers Many monomers have been converted to latexes by emulsion
polymerization. The main requirements are that the monomers can undergo free radical
chain polymerization and that they do not react with water. Monomers with limited solubility in water are most useful, although water-miscible comonomers can be used in small
amounts. Monomers with very high or very low solubility in water pose some difficulties
but can be accommodated in copolymerization. The two major classes of latexes used in
coatings are based on acrylic and methacrylic esters (Section 9.2) and on vinyl esters
(Section 9.3). Vinylidene chloride/acrylic ester copolymer latexes give films with exceptionally low water permeability (Section 9.2). The earliest latexes used in coatings were
copolymers of styrene with butadiene; they are seldom used today in architectural coatings, but are used in coatings for paper. Various functional monomers are also used.
Methacrylic acid (MAA) and acrylic acid (AA) improve colloidal stability, affect flow
properties, improve adhesion, and give carboxylic acid groups to serve as cross-linking
sites. Hydroxyethyl acrylate (HEA) and hydroxyethyl methacrylate (HEMA) provide
hydroxyl groups for cross-linking. Monomers have been developed that promote wet
adhesion to painted surfaces; methacrylamidoethylethyleneurea is used widely for this
purpose. Phosphate-functional (meth)acrylate esters are available that can be used as
comonomers to promote adhesion to metals [3].
Methacrylamidoethylethyleneurea
9.1.1.2. Initiators The principal initiators used in emulsion polymerization are soluble
in water. The most common initiators are persulfate salts, especially ammonium persulfate
9.1. EMULSION POLYMERIZATION
179
(more correctly, ammonium peroxydisulfate). Persulfates cleave thermally in the water
phase to sulfate anion radicals (Eq. 9.1) that initiate polymerization, as shown with
ethyl acrylate in Eq. 9.2. Note that the terminal group is a sulfate half ester anion.
ð9:1Þ
ð9:2Þ
The sulfate anion radical can also abstract hydrogen from water, leading to the formation of bisulfate and a hydroxy free radical (HO.). Although this reaction is not
favored thermodynamically, it occurs to some extent under conditions in which
monomer concentrations are low. Initiation by HO. leaves a terminal hydroxyl group.
Bisulfate lowers the pH, often requiring the addition of a buffering agent.
The half-life of persulfate is such that a 0.01 M solution of (NH4)2S2O8 at pH 10 produces
8.4 1012 radicals per mL s21 at 508C and 2.5 1015 radicals per mL s21 at 908C. For
rapid polymerization at lower temperatures, radical production can be accelerated by reducing agents. For example, mixtures of ferrous, thiosulfate (S2O22
3 ), and persulfate salts react
faster than persulfate alone as a result of a sequence of oxidation and reduction reactions.
Processes using this type of initiation are called redox emulsion polymerization.
Using redox systems, polymerization can be initiated at room temperature; the exothermic
polymerization reaction warms the reaction mixture to the desired temperature (often, 50 to
808C). Cooling is required to prevent overheating. Sometimes, redox initiation is used in the
first stage of polymerization, and the latter stages are continued using thermal initiation. A
variety of reducing agents have been used. Sodium formaldehyde sulfoxylate has been
widely used. Since it generates formaldehyde, it has been largely replaced by ascorbic
acid, isoascorbic acid, and sodium erythorbate. However, all three tend to give yellowing.
A mixed reducing agent consisting of 2-hydroxy-2-sulfinatoacetic acid disodium salt,
sodium sulfite, and 2-hydroxy-2-sulfanatoacetic acid disodium salt have been introduced.
It is reported to be very effective as a reducing agent and not to give yellowing [4].
It is common to add a second, more lipophylic “chaser” initiator in the late stages of the
semicontinuous process to facilitate conversion of monomer to high levels, preferably
greater than 99%. An initiator such as t-butyl hydroperoxide, which is more soluble in
polymer particles than in water, is more effective than ammonium persulfate in the late
stages, when most of the unreacted monomer is dissolved in the polymer particles.
Even when an auxiliary initiator is used, latexes still contain some unreacted monomer.
t-Butyl hydroperoxide has also been recommended as the sole initiator together with
sodium formaldehyde sulfoxylate instead of potassium persulfate; higher conversion of
monomers is obtained and the problem of residual persulfate ion is eliminated [5].
9.1.1.3. Surfactants Surfactants are an important component in emulsion polymerizations. A variety of both anionic and nonionic surfactants is used. An example of an
anionic surfactant is sodium lauryl sulfate (often incorrectly called sodium lauryl
180
LATEXES
sulfonate):
Among nonionic surfactants, nonylphenylethoxylates have been popular. However, they
now require eco-topological labeling in Europe. Reference [6] reviews the biodegradation of
nonylphenylethoxylates to toxic products. Ethoxylates of C12 – 14 mixed primary and secondary aliphatic alcohols are effective surfactants, are biodegradable and do not require labeling.
They are available with different lengths of the ethoxyl chain, since the most suitable chain
length can vary depending on the monomers to be polymerized. For example, acrylic
latexes can be started with a seed latex, made using sodium lauryl sulfate as the surfactant,
followed by adding the nonionic surfactant with 40 ethoxyl groups [6]:
Surfactants generally have limited solubility in water. Above that solubility limit, the critical micelle concentration (CMC), nonpolar ends of surfactant molecules associate with each
other in clusters called micelles. Micelles are submicroscopic aggregates typically containing
30 to 100 surfactant molecules, with the hydrophobic parts of each molecule oriented toward
the center and the hydrophilic parts oriented outward in contact with water. The CMCs of
different surfactants vary over a wide range from about 1027 to 1023 gL21. (See Ref. [7]
for a review of surfactants.) Surfactants play a critical role in stabilizing the dispersion of
polymer particles and preventing coagulation of the latex on standing. If the particles
approach each other closely, van der Waals forces tend to hold them together and the particles
flocculate. As noted in Section 3.5, flocculation of a dispersed phase system leads to changes
in flow properties; the viscosity increases and the flow becomes shear thinning.
There are two general mechanisms by which stabilization of dispersions is affected; one
is by charge repulsion in which the surfaces of the particles have an excess of one electrostatic charge, most commonly negative. For example, a latex can be stabilized by an
anionic surfactant adsorbed on the surface of the polymer particles. The surfactant
molecules orient with the long hydrophobic hydrocarbon tails in the polymer and the
hydrophilic salt groups on the periphery associated with the water. As a result, the
surface of the particles is covered with anions; each anion has an associated cation.
This surface layer of cations, called a Stern layer, is somewhat rigid and behaves as if
it were physically part of the particle. Its presence induces a second diffuse layer of
anions to surround the particle, giving it a negatively charged double layer. When two
particles approach each other, their diffuse, negatively charged layers repulse each
other electrostatically. The stability of such dispersions can be affected by the addition
of salts. The stability is particularly sensitive to the presence of multivalent ions with
the opposite charge of the stabilizing charge; hence, deionized water is generally used.
The second general mechanism for stabilization is by repulsion resulting from the outer
layers of the particle. Three terms are widely used in the literature to characterize this type
of stabilization: steric, entropic, and osmotic repulsion, which may be described as
follows. If the outer surface of the particles is hydrophilic, water is absorbed and swells
the surface. If this swollen layer is thick enough, the particles are not able to approach
each other closely enough to flocculate: hence the term steric repulsion.
9.1. EMULSION POLYMERIZATION
181
If a nonionic surfactant, for example, a surfactant with a hydrophobic nonpolar segment
and a relatively long chain of repeating ethylene ether groups as a hydrophilic segment, is
adsorbed on the surface of a latex particle, the hydrophilic ether groups will be on the outer
surface. The ether groups hydrogen bond with water, and more water is absorbed in the
layer. The absorbed layer can exist in a very large number of conformations, and water
molecules can move in and out of the layer. When two particles approach each other,
the layers are compressed, leading to a reduction in the number of conformations that
molecules in the layers can assume. As a result, the system becomes less random,
corresponding to a reduction in entropy. Resistance to the reduction in entropy leads to
repulsion: hence the term entropic repulsion.
Some authors prefer to focus on the reduction in the amount of water in the layer if compression occurs. There is a tendency for the water to return to equilibrium concentration in
the layer; some consider this to be analogous to osmosis and call the result osmotic repulsion. Other authors prefer to avoid the controversy over terms by simply calling the repulsion steric repulsion, which includes both entropic and enthalpic factors. We favor entropic
repulsion, because the principal factor is considered to be entropy. In industrial practice,
stabilization of the dispersion of polymer particles in a latex is found to be most effective
when charge repulsion and entropic repulsion are combined.
Surfactants are a major factor controlling the particle size and particle size distribution
of latexes (Section 9.1.2). In general, anionic surfactants are used at levels of 0.5 to 2 wt%
based on polymer; they cost less than nonionic surfactants, and they reduce particle size
more efficiently on a weight basis. On the other hand, nonionic surfactants, used at
levels of 2 to 6 wt%, are more effective in stabilizing the latex against coagulation
during freeze – thaw cycling, make it less sensitive to coagulation by salts (especially,
salts of polyvalent cations), less sensitive to changes in pH, and sometimes less likely
to stabilize foam. In general terms, anionic surfactants stabilize primarily by charge repulsion, and nonionic surfactants stabilize primarily by entropic repulsion. The two surfactant
types impart different rheological characteristics to the latex. Anionic surfactants lead to
formation of essentially rigid particles. Such latexes can have low viscosity at relatively
high solids. On the other hand, with nonionic surfactants, the thicker, swollen layers on
the surface of entropically stabilized particles lead to lower polymer solids or higher
viscosity at a given solids. The surface layer of an entropically stabilized latex particle
is not rigid, so can be distorted when shear stress is applied, imparting shear thinning
characteristics. Commonly, both anionic and nonionic surfactants are used in an emulsion
polymerization recipe. More explicit generalizations about the use of surfactants in emulsion polymerization are imprudent because there are exceptions to almost any statement
that can be made.
All surfactants impart some water sensitivity to films from latex paints. This sensitivity
is reflected, for example, in the tendency of house paints to waterspot if rained on soon
after application and in the somewhat limited corrosion resistance of most latex coatings
on steel. Selection of surfactant combinations that minimize such shortcomings is an
ongoing subject of research. So-called “soap-free” latexes have been developed in
which emulsion polymerization is effected without the use of conventional surfactants.
Such approaches may reduce the water sensitivity associated with surfactants. (Emulsifierfree latexes have been reviewed extensively in Ref. [8].) The most widely used approach is
to incorporate a hydrophilic comonomer such as (meth)acrylic acid.
Another approach to the problem of water sensitivity is to use polymerizable surfactants [9]. For example, it has been shown that a surfactant composed of allyl alcohol,
182
LATEXES
butylene oxide, ethylene oxide, and a sulfonate salt terminal group can be used to make
vinyl acetate/butyl acrylate latexes [10]. The surfactant is reported to be completely incorporated into the latex during the polymerization. Presumably, a hydrogen is abstracted
from the surfactant, yielding a free radical that initiates growth of another polymer
chain, covalently bonding the surfactant to a polymer molecule. Alkenyl-functional surfactants in latex preparation have been reported to copolymerize with the monomers
during polymerization, giving latexes having superior shear stability and freeze – thaw
resistance. Paints prepared with such latexes overcome the problems (e.g., surfactant
blooming) associated with the use of conventional surfactants [11].
Additional entropic stabilization can be achieved by including in the polymerization
recipe a small amount of water-soluble polymer (sometimes called a protective colloid )
that can form graft copolymers readily. A polymer commonly used for this purpose is
poly(vinyl alcohol) (PVA), which has many abstractable hydrogens. During polymerization, initiator radical ions abstract a hydrogen from PVA to form a free radical on a
PVA chain. A graft of the monomers being polymerized grows on the PVA chain;
more than one such graft can grow on a single PVA molecule. As the graft chain
becomes longer, it becomes hydrophobic and associates with other polymer molecules
in particles, carrying the PVA part of the molecule to the particle surface. The PVA
part associates with water to form an entropic stabilizing layer analogous to that formed
by nonionic surfactants. Such layers can be thicker than the layer from adsorbed surfactant
molecules and are effective in promoting stability. The water sensitivity of films may be
reduced because less water-soluble surfactant is needed. Hydroxyethyl cellulose (HEC)
can also be used. It has been shown that grafting occurs and that the resulting latex exhibits
thixotropy and is more shear thinning than a latex prepared with no water-soluble polymer
[12]. Although monomer – HEC grafting improves latex stability, it favors large particle
size, broad particle size distribution, and low film clarity [13].
9.1.1.4. Water and Other Additives Variations in water quality can cause a variety
of problems, especially when anionic surfactants are used. Therefore, deionized water is
used in emulsion polymerization and is generally used in latex paint production. Water
is somewhat soluble in some polymers and acts as a plasticizer, reducing film formation
temperature. Other ingredients sometimes used in emulsion polymerization processes
include buffers, which protect sensitive monomers from hydrolysis and sensitive surfactants from deactivation, and thickeners, which control viscosity.
9.1.2. Emulsion Polymerization Variables
Many quantitative theories for the mechanisms of emulsion polymerization have been
proposed, but none has proven to be applicable to all systems, because of the wide
range of variables involved. Various theories and mathematical models that have been
proposed are reviewed concisely in Ref. [14], an example of current theoretical efforts
is provided in Ref. [15], and controversial issues as of 2004 are reviewed in Ref. [16].
Since none of the theoretical models have proven to be generally applicable, the discussion
here is limited to qualitative discussion of important considerations.
The monomers are emulsified in water with surfactant molecules oriented at the surface
of the emulsion droplets, stabilizing the emulsion. There is also surfactant dissolved in the
water and excess surfactant present in micelles. In most systems of interest in coatings, at
least one of the monomers is somewhat soluble in water, so there are monomer molecules
9.1. EMULSION POLYMERIZATION
183
in solution as well as in the emulsion droplets. For example, the solubility of methyl
methacrylate (MMA) and ethyl acrylate (EA) is approximately 1% in water. The initiator
is also dissolved in the water.
There are three potential sites for initiation and initial propagation of polymerization:
(1) in the monomer swollen micelles, (2) in the aqueous phase (called homogeneous
nucleation), and (3) in the emulsion droplets. It is suggested that all three of these
modes may occur in any given polymerization, but that the ratio of the three modes is
system dependent [14]; monomer solubility in water, surfactant composition, and emulsion droplet size may be particularly important variables. Usually, initiation in the emulsion droplets is minimal. The number of particles predicted and the particle size are
different in each of the three cases (or combinations of them). However, qualitatively,
one can say that particle size decreases and number of particles increases as surfactant
concentration increases.
When at least one of the monomers is somewhat soluble in water, it is commonly
assumed that initiation occurs predominantly in the water phase. The initiating radical
reacts with a monomer molecule, forming a new free radical that is soluble in water;
these free radicals can react with additional dissolved monomer molecules to form oligomers with terminal free radicals. As the MW of the growing oligomer chain increases, its
solubility in water decreases, and one of three things can happen: (1) the growing chain
may enter a micelle, where propagation continues, as monomer molecules also enter the
micelle; (2) surfactant from the solution may be adsorbed on the surface of the growing
oligomer molecule; or (3) the growing chain can enter a monomer droplet. In any case, a
surfactant-stabilized polymer particle containing a single free radical is formed. Such
particles rapidly imbibe monomer, and propagation continues within the particles.
When all of the monomers are low in solubility, initiation in micelles may predominate.
Again, the monomer in the micelle polymerizes rapidly to form a surfactant-stabilized
polymer particle containing a single free radical.
Termination (by combination or disproportionation, as described in Section 2.2.1)
occurs only after a second free radical enters a particle. After termination, the
surfactant-stabilized polymer particle is temporarily inactive because it contains no free
radicals. More monomer molecules then migrate into the inactive polymer particle. A
third free radical (IMMMMM.) enters the monomer-swollen particle and initiates polymerization. Chain growth continues until a fourth free radical enters the particle and causes
termination. Dissolution of monomer from the monomer droplets replenishes the
monomer in solution. The process is repeated over and over, leading to formation of
additional polymer molecules in the particle. (This model does not apply to cases in
which the Tg of the polymer is higher than the polymerization temperature, but this is not
usually the case.)
As the polymer particles grow, surface expansion requires additional surfactant. This
need is satisfied by adsorption of more surfactant molecules from solution, which, in
turn, is replenished by dissolution of micelles in which polymerization has not occurred.
The result is the growth of particles that acquired a polymer molecule early on and disappearance of micelles that did not. Polymer particle growth continues until unreacted
monomer or initiator is used up; enough initiator is used to assure high conversion of
monomer. After the early stages of the process, much of the polymerization occurs
within a fixed number of polymer particles. The number of particles per unit volume
depends on the concentration of micelles in the early stages. If the concentration of
micelles is high, the number of particles is high, and hence the particle size at the end
184
LATEXES
of the polymerization is small. At the same surfactant concentration, a surfactant with a
low CMC gives a smaller particle size latex than a surfactant with a higher CMC.
In the homogeneous nucleation mechanism, surfactant is absorbed on the surface of the
growing oligomer chains. As the number of particles and their size increase, the amount of
surfactant becomes too small to stabilize them, and particles coalesce to form fewer and
larger particles until the surfactant concentration is just high enough to stabilize the particles.
The third mechanism, initiation in the droplets of monomer emulsion, is usually unfavorable. In most cases, the size of the droplets of monomer emulsion is relatively large and
their surface area is small relative to the surface area of the more numerous micelles, so
that the probability of an initiating radical or a growing oligomer radical entering an emulsion droplet is small. However, if the rate of agitation is very high and/or a high surfactant
concentration is used, emulsion droplet size becomes smaller, and the probability of
propagation in emulsion droplets increases. If a water-insoluble initiator were used in
the early stages of polymerization, most of the polymerization would occur in the emulsion droplets, resulting in large, unstable particles. This method is the basis for a different
polymerization method, called suspension polymerization.
The number of particles, and hence their particle size, are also affected by the rate of
initiation (a function of the type of initiator, its concentration, and the temperature).
The faster the rate of formation of initiating radicals, the larger the number and the
smaller the size of the particles, everything else being equal.
Latexes are frequently made starting with a seed latex, a small particle size latex [17].
Monomer and initiator are added slowly, and the polymerization occurs primarily in the
seed particles, resulting in a constant number of particles as polymerization proceeds.
The seed latex can be prepared as the first step in a polymerization so that it contains
10 to 20% solids. Alternatively, it is common practice to produce a large batch of seed
latex and subdivide it for subsequent use in many batches. The seed latex is diluted to
3 to 10% solids. Increasing the number of seed particles, with other factors being equal,
yields a latex of smaller average particle size. This practice can improve batch-to-batch
reproducibility and is particularly useful in sequential polymerizations.
The M̄w obtained by both batch and semicontinuous processes is generally very high; an
M̄w over 1,000,000 is common. In many cases, there is a broad distribution of MWs. A
high M̄w results from several factors. First, during polymerization, the fraction of
polymer particles containing more than one free radical at a given instant is low. As a
result, the rate of termination reactions within the particle is reduced, and a single
growing chain may consume most of the monomer present in the particle before it terminates. Second, the viscosity inside the particles is very high, reducing the mobility of the
polymer chains. It is then easier for more mobile monomer molecules to diffuse and to add
to the growing chain and increase the MW than for two less mobile growing chains to react
with each other and terminate growth. Third, hydrogen atoms on polymer chains can be
abstracted, leading to chain transfer to polymer and formation of branched-chain
molecules. The environment within the polymer particle favors chain transfer to
polymer because the concentration of free radicals is low (often, one per particle) while
the concentration of polymer approaches 100%. Chain transfer to polymer increases M̄w
and M̄w/ M̄n because it is statistically probable that when the growth of short chains is
ended by chain transfer, the new chain will grow as a branch on a polymer molecule
that already has a high MW. Sometimes a chain transfer agent is added to the monomer
mixture deliberately to reduce the MW, and sometimes a very small amount of difunctional monomer is added to increase it.
9.1. EMULSION POLYMERIZATION
185
In a study comparing batch and semicontinuous homopolymerization and copolymerization of vinyl acetate and butyl acrylate, M̄w/M̄n for batch homopolymers and
copolymers was broad, ranging from 15 to 21; and the semicontinuous process gave
even broader distributions, ranging from 9 to 175 [18]. The semicontinuous process
commonly afforded bimodal distributions with considerable amounts of relatively low
MW polymer. Semicontinuous poly(vinyl acetate) homopolymer had the broadest distribution; GPC indicated substantial fractions of material with MW below 660 and above
1,700,000.
Particle size distribution is an important variable in the formulation of latex paints.
Particle size distribution is commonly measured by light scattering and can also be
measured by hydrodynamic chromatography, ultracentrifuge, or transmission electron
microscopy. It may be expressed as number average diameter (D̄n) and volume average
diameter (D̄v). The ratio D̄v/D̄n is a convenient index of particle size distribution. A
broad or bipolar distribution is often advantageous because the packing factor is affected
by particle size distribution. The volume fraction of latex particles that can be present at a
standard viscosity is highest with broad particle size distribution.
The particle size of a latex is determined by the number of particles relative to the
amount of monomer. The number of particles is governed primarily by the amount
and type of surfactant(s) and initiator(s). Anionic surfactants tend to produce more particles than do nonionic surfactants when used on an equal weight basis and hence, lead to
smaller particles. The specific anionic surfactant used affects the particle size; for
þ
example, use of NP(OCH2CH2)20OSO2
3 NH4 (NP stands for the nonpolar part of the
molecule) in making an acrylic latex gave a particle diameter of 190 nm, whereas use
þ
of the less hydrophilic NP(OCH2CH2)9OSO2
3 NH4 gave a diameter of 100 nm [19].
In general, particle size distribution is broader with anionic surfactants and more
nearly approaches a monodisperse particle size with nonionic surfactants. The effect
of changing the ratio of nonionic surfactant (nonylphenol ethoxylate with 40 mol of
ethylene oxide) to anionic surfactant (sodium dodecylbenzenesulfonate) on particle
size and particle size distribution with a constant amount of total surfactant in the
polymerization of a vinyl acetate/dibutyl maleate copolymer has been studied [18].
The lowest ratio gave the smallest particle size and the broadest distribution of
particle size.
Particle size is also affected by agitation intensity and turbulence during polymerization. High turbulence leads to larger particle size, thus influencing rheological properties.
(See Ref. [20] for a study of the relationship of agitation during polymerization of
styrene – acrylic latexes and particle size.) The effect increases the problem of scaling
up of laboratory synthesis to production requiring designing lab equipment to approximate
the agitation to be encountered in plant reactors.
Emulsion polymerization usually produces latexes with essentially spherical particles.
During the 1980s, latexes were commercialized in which groups of a few particles were
fused together into nonspherical lobed particles [21]. At equal concentration, the viscosity
of these latexes was higher than the viscosity of comparable latexes with spherical
particles, since the packing factor is smaller than for a spherical particle latex. Paints formulated with a lobed latex are shear thinning and give a higher high shear viscosity at the
same low shear viscosity than do conventional latexes. The higher viscosity at shear rates
experienced during application leads to thicker films with low levels of thickener and
hence, lower cost. Adhesion to chalky surfaces is said to be improved over that of
paints formulated with conventional latexes.
186
LATEXES
9.1.3. Sequential Polymerization
When latexes are prepared by semicontinuous polymerization under monomer-starved
conditions, the composition of the monomer feed can be changed during the course of
the polymerization. Depending on a number of variables, the resulting latex particles can
have a variety of morphologies. In some cases, latexes having core –shell morphology
can be made by shifting from the initial monomer feed composition to a different feed
partway through the process. Many such core – shell latexes display two Tgs after coalescence. In some cases the composition of the polymer in the center of each particle (the core)
reflects the initial monomer feed composition, and the composition near the surface (the
shell ) reflects the second monomer feed composition. In other cases, an inverse core –
shell morphology can be obtained, with the composition of the core polymer being that
of the second monomer feed and the shell polymer derived from the initial monomer
feed. In yet other cases, a homogeneous composition is obtained despite a change in
monomer feed composition. In still other cases, it is possible to make a gradient morphology
by changing the monomer feed composition continually during the polymerization. There is
an extensive literature on sequential polymerization. (The literature applicable to latexes of
interest in coatings and discussion of the variables involved is reviewed briefly in Ref. [22].)
Several factors affect the morphology of particles obtained in sequential polymerization. Formation of outer layers having the lowest interfacial tension with the water
phase is favored; thus, when different monomer compositions are fed, the one containing
the most polar monomers is likely to end up as the shell, regardless of its place in the
sequence. Morphology can be affected by the amount of free monomer present during
the polymerization, since free monomer acts as a plasticizer; hence, formation of core –
shell morphology tends to be favored under highly monomer-starved conditions. The
presence of di- or trifunctional cross-linking monomers in at least one of the monomer
combinations increases the probability of phase separated morphology. In latexes
designed for coatings, generally only limited amounts of polyfunctional monomers can
be used because of the need for coalescence for good film formation. But it is a
common practice to use a low level of cross-linking monomer in the first composition
to increase the likelihood that it will remain in the core. Examples of applications of
sequential polymerization are given in Section 9.2.
9.2. ACRYLIC LATEXES
Acrylic latexes are widely used for exterior paints because of their resistance to photodegradation. They are also more resistant to hydrolysis and saponification than are
vinyl acetate latexes. These properties are critical for exterior paints; they also make
acrylic latex paints useful for alkaline substrates such as masonry and galvanized metal
and for applications in which there is exposure to high humidity. Acrylic and styrene –
acrylic latexes are being used increasingly for industrial maintenance coatings. Acrylic
latexes are finding increasing interest for kitchen cabinet finishes and for OEM automotive
applications. For example, base coats based partially on latexes are being developed for
OEM automotive waterborne base coats. Hybrid acrylate – urethane dispersions have
become increasingly important (Section 12.7.2).
A critical decision in designing a latex is monomer selection. An important consideration is to select a monomer combination that produces a copolymer with the appropriate
Tg. Monomer cost is also important; the paint business is highly price competitive. For
9.2. ACRYLIC LATEXES
187
lower priced paints, adequate film properties must be attained using low cost monomers.
The Tg must be low enough to permit coalescence of the latex at the lowest anticipated
application temperature, yet high enough to assure adequate film hardness and toughness.
Tg in the vicinity of 5 to 158C is common for exterior house paints, which might be applied
at temperatures as low as 28C. The minimum film forming temperature (MFFT) (Section
2.3.3) of latexes is related to Tg but is influenced by other factors, such as particle size,
phase separation within the latex particles, and the plasticizing effects of water and surfactants. The MFFT tends to be somewhat lower than Tg. A Tg and MFFT in the desired range
are usually attained by copolymerizing monomers whose homopolymers Tg values are
much higher and much lower than the target values.
As a high Tg comonomer, MMA imparts excellent exterior durability and hydrolytic
stability at moderate cost. Styrene is often substituted partly or completely for MMA
because it costs less and provides a similar Tg effect. Such products are sometimes
called styrene –acrylic latexes but usually, simply acrylic latexes. Styrene imparts excellent hydrolytic stability, but its effect on overall exterior durability is not clear cut. Styrene
homopolymer degrades relatively rapidly outdoors. However, in a copolymer, some MMA
can be replaced with styrene without decreasing exterior durability substantially. The
amount of styrene that can be substituted apparently varies, depending on the other comonomers and probably on other variables, such as process conditions. For example, processes that would tend to produce blocks of homopolystyrene are suspect. It is safest to
test each composition thoroughly outdoors before use. It is interesting that styrene acrylics
are said to be used more in northern Europe than in southern Europe, where weather
exposure is more severe. Other properties may also be affected. For example, it cannot
be assumed that the mechanical properties of the copolymer will be equal when styrene
is substituted for MMA, even though its Tg is about the same. Polystyrene has a significantly higher brittle – ductile transition temperature than poly(MMA), even though it
has a slightly lower Tg (Section 4.2).
Acrylic esters are generally used as the low Tg monomer(s), the choice of which is
affected by cost and durability considerations. If the cost of the common acrylic esters
is intermediate between that of MMA and of styrene, it is more economical to use ethyl
acrylate (EA; homopolymer Tg ¼ 2248C) than butyl acrylate (BA; homopolymer
Tg ¼ 2548C) in styrene-free latexes in order to get the highest possible proportion of
EA in the composition. However, with styrene – acrylic latexes, BA may be more economical when styrene is less expensive than EA or BA. Presumably, the different acrylic
esters impart different film properties at a given Tg, but published data are inadequate
to evaluate the effects. Recent decades have seen an industry trend toward use of BA in
preference to EA [23]. While published data are scarce, one can speculate that BA has
given better durability in proprietary tests. One can also speculate that it imparts better
resistance to hydrolysis.
An example of a laboratory semicontinuous batch process for preparation of a 40 : 59 : 1
MMA/EA/MAA copolymer latex is provided in Table 9.1. Many more examples are
provided in Ref. [24], and very detailed directions for laboratory-scale seeded emulsion
polymerization with a more contemporary composition are given in Ref. [2]. Note from
Table 9.1 that the measured MFFT of the latex at 98C is somewhat lower than the Tg of
178C estimated by the Fox equation for a 40 : 60 MMA/EA copolymer. Also, note that
changing to a 50 : 50 MMA/EA ratio increases MFFT to 218C, while the Fox equation
predicts a Tg value of 288C. These figures give a general picture for how Tg and MFFT
relate to composition.
188
LATEXES
TABLE 9.1. Laboratory Procedure for Preparation of a MMA/EA/MAA
Copolymer Latex
Materials
1L
96 g
320 g
480 g
8g
1.6 g
Deionized water
Triton X-200 (sodium salt of an alkylaryl polyether sulfate ester) (Union Carbide Co.)
Methyl methacrylate (10 ppm MEHQ stabilizer) (MEHQ is the monomethylether
of hydroquinone)
Ethyl acrylate (15 ppm MEHQ stabilizer)
Methacrylic acid (100 ppm MEHQ stabilizer)
Ammonium persulfate
Procedure
Prepare an emulsion of all reactants in 800 mL of water. Place 200 mL of the emulsion and 200 mL of
water in a 3-L flask fitted with an inert gas inlet tube, a thermometer, a stirrer, an addition funnel, and a
reflux condenser. Heat in a 928C water bath while stirring until the internal temperature reaches 828C.
The mixture will begin to reflux, and its temperature will rise to about 908C within a few minutes,
indicating rapid polymerization. When refluxing subsides, add the remaining emulsion continuously
over a period of 1.5 hours. Heat to maintain refluxing; the internal temperature will be 88 to 948C. After
monomer addition is complete, heat to 978C to complete conversion of the monomer. Cool and strain.
Properties
NVW 42.9 (calculated 43.1); pH 2.7; viscosity (Brookfield), 11.5 mPa.s; MFFT 98C; film hardness
(Tukon), 1.2 KHN. A similarly prepared 50 : 50 MMA/EA copolymer latex (no MAA), had similar
properties, except that MFFT was 218C and film hardness was 6.2 KHN.
Source: Ref. [43].
Commercial suppliers of latexes provide little information as to the composition of the
latexes they offer. Although a simple acrylic latex made as described in Table 9.1 could be
used to make a house paint, it would not be competitive. Commercial production involves
many process nuances that are held confidential by the latex producing companies. It is
difficult to replicate commercial latexes, even though the monomer composition can be
analyzed precisely.
Acrylic acid (AA), or the somewhat less water-soluble methacrylic acid (MAA), is
commonly used in preparation of acrylic latexes, usually on the order of 1 to 2 wt% of
the monomer charge. The carboxyl functionality incorporated with these monomers
enhances the mechanical stability of wet latex paint, reducing the amount of surfactant
needed. The effects of acid monomers on stability and viscosity are maximized when
they are incorporated in the last part of the monomer feed and the polymerization
medium is acidic [1,25]. The viscosity of a latex made in this way depends on pH. As
ammonia is added, there is little change in viscosity until the pH reaches about
7. Above pH 7, viscosity increases steeply until pH reaches 9 or 10; at still higher pH
levels, viscosity decreases. The pH of acrylic latex paints is usually adjusted to about 9,
where there is a substantial viscosity effect: Not only is the viscosity high, the paint exhibits shear thinning, which is sometimes a desirable application characteristic (Chapter 31).
Hoy describes a latex that shows this effect [1]; it is a MMA/EA/BA/AA (40 : 52 : 6 : 2
by weight) copolymer, in which the acrylic acid was added late in the process. The Tg of
this copolymer was 158C. Neutralization with ammonia to pH 9 caused the surface layer of
the particles to expand, so the diameter of the particles was about 1.8 times that at pH 7.
The corresponding increase in volume was almost sixfold. The expansion resulted from
9.2. ACRYLIC LATEXES
189
the association of water with the highly polar salt groups in the polymer near the particle
surface. The presence of the expanded layer at pH 9 increased the viscosity at low shear
rates. However, because this layer could be distorted by shear forces, the viscosity
decrease at high shear rates was less marked.
In the same paper, Hoy describes the effects of using sequential polymerization during
semicontinuous batch polymerization [1]. In one experiment, an MMA/BA ratio of 40 : 60
was used, with 2 wt% of AA added late in the process. When the 40 : 60 ratio was maintained throughout monomer addition, the product, called a uniform feed latex, had a Tg of
208C. A second staged feed latex of the same overall composition was made by a process
designed to yield core – shell particles [1] (Section 9.1.3). In the first half of the monomer
addition, the MMA/BA ratio was 70 : 30, and in the second half, it was 10 : 90. The resulting latex showed two Tgs, one at 608C and the other at 2108C. Because the shell of this
latex has a low Tg, its MFFT is substantially lower than that of the uniform feed latex.
However, coalesced films are cloudy. Apparently the high Tg cores remain uncoalesced;
the composition of the cores and the shells is sufficiently different that the polymers are
not miscible with one another. Because the refractive index of the two polymers is different, light passing through the film is scattered, making the film cloudy. Since the Tg of the
continuous phase is low, the blocking resistance of a film of the staged feed latex is inferior
to a film of the uniform feed latex.
Hoy describes a third latex with the same monomer composition made by a process called
linear power feed addition, using the apparatus sketched in Figure 9.2. In this process there are
two stirred tanks for monomer mixtures. The near tank was loaded with half of the monomer
charge at an MMA/BA ratio of 70 : 30, and the far tank was loaded with the other half at a
10 : 90 ratio. During polymerization, monomer was pumped from the near tank into the
reactor in the usual way. At the same time, monomer was being pumped at the same rate
from the far tank to the near tank. Again, 2 wt% of AA was added in the late stage of the
process. In this procedure, the monomer feed starts at 70 : 30 MMA/BA and ends at
10 : 90, but changes continuously through the process. Each latex particle presumably contains high Tg polymer at its center and low Tg polymer (with COOH groups) near its
surface. It differs from the usual core–shell latex in that there is a linear gradient of
Figure 9.2. Apparatus for semicontinuous latex polymerization using a linear power feed addition.
(Adapted from Ref. [1], with permission.)
190
LATEXES
composition and Tg from the center to the surface. Films cast from this latex have equal clarity
and the same Tg, about 208C, as those cast from the uniform feed latex. However, the glass
transition is much broader for the linear power feed latex than for the uniform feed latex.
The linear power feed latex has a lower MFFT than that of the uniform feed latex and also
has a higher blocking resistance temperature.
Exterior house paints for large surfaces are usually formulated at low gloss in the
United States; the formulations are heavily pigmented. Block resistance requirements
are modest, and the pigment helps improve this property; Tg values of about 5 to 158C
are adequate. Latex paint films are more permeable to water than films from oil or
alkyd paints, an advantage because it reduces blistering of paints applied to wood surfaces,
but it is a disadvantage for coatings applied over metal.
Applications such as gloss trim and door paints, as well as interior trim paints (for
kitchen cabinets, windowsills, etc.) have different requirements; to be glossy, they must
be formulated at lower pigment levels, yet block resistance is required. This situation
presents a challenge to designers of latexes. The problems have been at least partly overcome, as described by Mercurio [26]. A latex Tg of about 558C is reported to be required to
achieve adequate block resistance. Coalescence of acrylic latexes having this Tg can be
attained by using a small particle size latex with substantial amounts of carefully selected
coalescing solvent(s). The coalescing solvent increases VOC to near the current regulatory
limit of 250 g of VOCs per liter of paint as applied, excluding water. Commercial products
based on these principles are available; details of how they are made are proprietary.
Although the VOC of latex paints is usually lower than that of solventborne paints,
there is increasing pressure to reduce VOC further. Reducing the amount of coalescing
solvent required is the most evident approach to this end. Sequencing the composition
of the monomer feed during polymerization so that the last part of the monomer feed
has a higher fraction of monomers, which impart a low Tg, as in the power feed approach
discussed previously, permits reduction of coalescing solvent (Section 9.1.3). Addition of
acrylic acid in the later part of the comonomer feed may permit reduction of coalescing
solvent. Water associates with the neutralized carboxylic acid salt, plasticizing the
surface of the particles, promoting coalescence, and reducing the need for coalescing
solvent. Another method is to use cross-linkable latexes having a lower Tg; cross-linking
after film formation increases block resistance to offset the effect of the lower Tg (Section
9.4). It may also be possible to design a coalescing solvent that would cross-link after film
formation. Another approach for eliminating the need for coalescing solvent is to use
blends of high and low Tg latexes [27,28]. When a film of such a blend dries, the high
Tg latex does not coalesce, but is dispersed in the continuous phase from the low Tg
polymer. The hard particles act to reinforce the low Tg polymer film increasing the
modulus of the film so that its block resistance is superior to that of a film using only
low Tg latex. It is critical that the ratio of the two latexes be such that there is enough
low Tg latex so that there is enough to enclose the high Tg latex particles completely. It
would be interesting to compare the results with those of a film from the low Tg latex
with a volume content of pigment equal to that of the high Tg latex.
Acrylated castor oil has been used as a comonomer, resulting in latexes that are
formulated into semigloss paints with better block resistance while still permitting
low-temperature film formation. The latexes have the further advantage that coalescing
solvents are not needed [29].
Another approach to eliminating VOC is to make nano polymer/nano clay composite
latexes. The latexes are made with relatively low Tg acrylic monomers polymerized with
9.3. VINYL ESTER LATEXES
191
clay dispersed in the water in which the emulsion polymerization is carried out. Films
prepared from the latex have unusual toughness without compromising film formation.
They also exhibit low tack and dirt pickup in zero VOC paints [30]. Polymerization of
BA, MMA, and MAA, with sodium montmorillonite dispersed in the water phase has
been disclosed [31]. The latex is reported to give coatings with improved block, print,
and dirt pickup resistance, as well as enhanced barrier properties.
For latexes to be used over metal, it is desirable to have low moisture vapor
permeability. Acrylic latexes give films with quite high permeability. One approach to
reducing water permeability is to use vinylidene chloride as a comonomer. Reference
[32] provides details of procedures for the copolymerization and characterizations of
several vinylidene chloride/acrylate latexes. Since the reactivities of vinylidene chloride
and acrylic ester monomers are quite different, monomer-starved polymerization conditions are used to achieve reasonably uniform compositions. Although the moisture
vapor permeability is greatly reduced, stabilizing the chlorinated copolymer against
photodegradation may be difficult.
A different use for an acrylic emulsion polymerization is for the preparation of microgel
particles. For example, acrylic latexes containing some cross-links by using a small fraction of divinylbenzene as a comonomer were prepared by emulsion polymerization using a
water-soluble azo initiator, 4,40 -azobis(cyanovaleric acid) [33]. The latex polymer was
isolated as a powder by precipitating with an equal volume of isopropyl alcohol,
washing with a 4 : 1 methyl alcohol/water ratio, and drying under vacuum at 508C. The
resultant powder readily dispersed in aliphatic hydrocarbon to yield a dispersion of an
acrylic microgel. Among other applications, acrylic microgels are used as additives to
control sagging of high-solids coatings (Section 24.3).
9.3. VINYL ESTER LATEXES
Vinyl acetate (VAc) is less expensive than (meth)acrylic ester monomers. However, VAc
latexes are inferior to acrylic latexes in both photochemical stability and resistance to
hydrolysis. When an acetate group on a PVAc chain hydrolyzes, the liberated acetic acid catalyzes hydrolysis of more acetate groups, and the hydroxyl group may exert a neighboring
group (anchimeric) effect that promotes hydrolysis of adjacent acetate groups. Hence,
PVAc latexes find their principal use in interior coatings that do not have to withstand
high humidity exposure or frequent wetting. One such use, flat wall paints, is the largest
volume type of paint sold in the U.S. market. The Tg of PVAc homopolymer is 298C, too
high for film formation under ambient conditions. Thus, one must either formulate PVAc
with a plasticizer or copolymerize VAc with a monomer that reduces Tg. A widely used
comonomer is n-butyl acrylate (BA), although other comonomers, such as 2-ethylhexyl acrylate and di-n-butyl maleate, have been used. Longer chain vinyl esters are also being used as
comonomers; the resulting latexes have superior hydrolytic stability and exterior durability.
Use of diesters of 3-butene-1,2-diol to reduce the Tg of vinyl acetate copolymers has been
reported. Best hydrolytic resistance was obtained using the di(2-ethyl)hexanoate ester [34].
Copolymerization of VAc with BA poses potential problems because of reactivity
ratios; specifically, the rate of reaction of a free radical on a terminal BA group with
another molecule of BA is greater than its rate of reaction with VAc; furthermore, the
rate of reaction of a free radical on a terminal VAc group is much greater with BA than
with VAc. Thus, a 50 : 50 mixture of BA and VAc will produce a copolymer very rich
192
LATEXES
in BA at the outset of batch polymerization and very rich in VAc in the late stages. This
problem is overcome by a semicontinuous batch process, wherein the monomer mixture is
added at a rate equal to the polymerization rate. A steady state is reached in which the concentration of unreacted VAc is low but the concentration of BA is even lower. Under these
monomer-starved conditions, a relatively uniform copolymer forms with a composition
similar to the ratio of monomers being fed. Such copolymers are made commercially
on a large scale. An extensive study of the copolymerization of VAc with BA by batch
and semicontinuous batch processes, as well as characterization of the copolymers, has
been reported [18,35] and is discussed in part in Section 9.1. Since the reactions rates
are relatively low, the rate of monomer addition must be quite slow. If the initial
charge is high in the least reactive monomer, the addition can be made faster while still
producing reasonably uniform copolymerization. Mathematical simulations can predict
the effects of different addition modes [36].
Vinyl acetate is more soluble in water than acrylic esters, as are the initial oligomeric
free radicals formed from VAc than those from oligomeric radicals of acrylic esters at
equal chain length. This factor increases the probability of termination of VAc-rich oligomers, while still in aqueous solution, and leads to low MW fractions. On the other hand,
chain transfer to polymer is more facile with polyvinyl acetate, leading to high MW and
broad MW distribution. The high water solubility of VAc also increases the need for effective initiator systems to remove residual monomer. Vinyl acetate latexes, including the
effect of different initiator combinations, are reviewed in Ref. [37].
An essential component of the recipe for polymerization of VAc is a buffer, such as
sodium bicarbonate, to maintain the pH near 7. A buffer is essential because VAc
monomer hydrolyzes at appreciable rates under either acidic or basic conditions. The
hydrolysis is irreversible because the products are acetaldehyde and acetic acid. At pH
less than 7, hydrolysis is autocatalytic because acetic acid catalyzes further reaction.
Furthermore, acetaldehyde is oxidized by persulfate, which consumes initiator and
generates still more acetic acid to catalyze hydrolysis.
Copolymer latexes of VAc and (meth)acrylic monomers have been designed for dualpurpose use in both exterior and interior flat paints. The objective is to save cost in two
ways: lower raw material cost for exterior paints and lower inventory and storage costs
by basing all production on a single resin. This approach has been successful commercially, although the latexes undoubtedly represent a compromise between the best resins
for exterior and interior use. The amount of vinyl acetate in an acrylic latex can be
increased by sequential polymerization to form inverse core – shell latexes, lowering the
cost with minimal loss of properties [38]. For example, a seed latex of MMA/BA/
MAA made with persulfate initiator was further reacted with vinyl acetate using azobisisobutyrylnitrile as initiator. Since the initial polymer has polar acrylate salt groups, the
acrylate polymer becomes the shell, with a polyvinyl acetate core.
Copolymer latexes of the vinyl ester of a C10 branched acid (vinyl neo-decanoate, Vinyl
Versatate) with VAc are used in both interior and exterior paints. These copolymers have
better hydrolytic stability than conventional vinyl acetate copolymers [39]. A combination
of ethoxylated undecyl alcohol, a cellulose ether, and sodium vinyl sulfonate can be used
as a surfactant/protective colloid in preparing the latexes. A study has been published
9.4. THERMOSETTING LATEXES
193
covering the effect of changes in chain length of the ethylene oxide units on the properties
of latexes and paints. Increasing the degree of ethoxylation gave smaller-particle size latex
particles and higher viscosity. In flat paints, optimum properties are obtained with 2 to 3%
emulsifier concentration and 17 to 28 degrees of ethoxylation. In gloss and semigloss
paints, the gloss improves with increased emulsifier content up to about 4%, where it
levels off, and as the degree of ethoxylation increases to 17. Above 4% concentration
and 17 ethoxyl units, blocking increases drastically. Similar results were obtained with
vinyl acetate –ethylene latexes [40].
ðR1 ; R2 ; and R3 are alkyl groups totaling C8 H19 Þ
The use of a variety of vinyl esters, including vinyl pivalate and vinyl 2-ethylhexanoate,
in latexes is discussed in Ref. [41]. Such monomers yield polymers that are more hydrophobic than vinyl acetate homopolymers and that have superior hydrolytic stability and
scrub resistance. The advantages of using vinyl neo-decanoate in both vinyl acetate and
acrylic copolymers is reported in Ref. [42].
Emulsion copolymerization of VAc with vinyl neo-decanoate can be problematic
because of the very low water solubility of the latter. Satisfactory processes have been
developed, and the mechanism of copolymerization is reasonably well understood [43].
Copolymers of ethylene and VAc are made by emulsion polymerization under high
pressure. The copolymers are often called VAE latexes. Their use, especially in interior
wall paints, has been increasing rapidly; consumption of the latexes (on a solids basis)
in 2003 was 27,000 metric tons in the United States and 20,000 metric tons in Europe.
The number average MW of VAc/ethylene is lower than that of most emulsion polymers.
In one series of varying MW, a latex having M̄n value of 89,700 showed improved scrub
resistance. Scrub resistance increased further with still higher M̄n, but films cracked with
touch-up at 408F. Water permeability and scrub resistance are superior to VAc/BA
latexes. Lower-VOC paints can be prepared and the odor is much lower than with
VAc/BA latexes [44]. Ethylene is not only a more effective monomer than butyl acrylate
for reducing the Tg of VAc polymers, and it appears to provide more effective plasticization. Paints made with VAc/ethylene latexes require less coalescing solvent and are said
to show better touch-up properties at low and room temperatures, as well as higher scrub
resistance [45]. Copolymer latexes with a hydrophobic comonomer give paints with
enhanced water, block, and scrub resistance [46]. An example of such a terpolymer
latex is vinyl acetate/ethylene/vinyl neo-decanoate [47].
9.4. THERMOSETTING LATEXES
Although the majority of latexes are thermoplastic polymer latexes, there are applications
for thermosetting latexes as well, as reviewed in Refs. [48] and [49]. The reviews emphasize experimental results related to the importance of the rate of interdiffusion and
194
LATEXES
cross-linking in films of thermosetting latexes. To develop good properties, significant
interdiffusion must occur before extensive cross-linking. Film properties depend on the
relative rate of interdiffusion between polymer particles and the rate of cross-linking. If
the rate of cross-linking is fast in comparison to the rate of diffusion, cross-linking
within the particles will occur and interfere with good film formation. If the rate of crosslinking is too slow, achievement of desirable film properties may take excessive time. A
balance between the two is needed. In some cases, the functional groups on a latex particle
are on the particle surface and an external cross-linker serves to cross-link the particles
together. In other cases, the functional groups are on the same polymer, and interdiffusion
is required to bind the particles together.
The rate of interdiffusion is controlled by T 2 Tg and the chain length of the latex. Thus,
thermosetting latexes are designed with lower MW, which both reduces Tg and shortens
chain length. Straight chains diffuse more rapidly than branched chains. A copolymer
of isobutoxymethylacrylamide (IBMAA) with MMA and BA having 2% IBMAA and
M̄w ¼ 200,000 was studied. The MW was controlled with dodecyl mercaptan chain transfer agent. IBMAA groups react with each other to form 22CO22NH22CH222O22CH222
NH22CO22 cross-links, a reaction catalyzed with acid. It was found that interdiffusion
increases more rapidly with temperature than does cross-linking. At 808C with 0.5%
pTSA, the cross-linking occurred in 40 min. Similarly, copolymers of BA/IBMAA/
HEA/MAA showed good film formation. In addition to self cross-linking, IBMAA
reacts with the OH groups of HEA; the COOH group of MAA serves as a catalyst [48].
In many cases, two package coatings are required, but they are useful only for industrial
applications; a few stable systems have been developed that can be used for architectural
paints. Cross-linked films have better mechanical properties and increased resistance to
solvents. In general, a lower Tg latex can be used in coatings, which permits coalescence
without addition of a coalescing solvent and/or at lower film formation temperatures.
For two package coatings, a variety of cross-linkers for hydroxy-functional and
carboxylic acid –functional polymers can be used. Hydroxy-functional polymers are
readily prepared by using hydroxyethyl (meth)acrylate as a comonomer. The resulting
latexes can be formulated with urea –formaldehyde (UF) or melamine – formaldehyde
(MF) resins as cross-linkers. UF resins tend to evolve high levels of formaldehyde from
hydrolysis; MF resins evolve less formaldehyde. MF resins are soluble in BA/HEMA/
MAA latex polymers, so that they act as plasticizers to promote interdiffusion [48]. As
temperature increases, interdiffusion rate increase faster than the rate of cross-linking
until many of the potential reactions have occurred, but at some point the cross-link
density (XLD) reaches a point that interdiffusion ceases. The rate of cross-linking can
be increased by the addition of a catalyst such as pTSA. Although this permits shorter
cure times, the pot life of the coating decreases and the residual acid in the final film
can lead to reduced stability of the films to hydrolysis since the catalyst also catalyzes
hydrolysis.
An approach to assuring distribution of the MF resin through the polymer particles is to
dissolve the MF resin in the monomer mixture before polymerization [50]. Premature crosslinking is minimized by controlling pH at above 5. With addition of catalyst, films cross-link;
the pot life of the resin after adding catalyst is 1 to 2 days.
The following examples in this section illustrate specific applications and/or specific
cross-linkable systems.
Traffic paints must dry very rapidly so that vehicles can be driven over them soon after
application. Three-package traffic marking paints have been patented. The first component
9.4. THERMOSETTING LATEXES
195
is a white (or yellow) pigmented carboxylic acid functional latex paint, the second paint is
pigmented with CaCO3, the third component is a polyamine. The polyamine is a copolymer of dimethylaminoethyl methacrylate and MAA, dissolved in aqueous methanol and
neutralized with NH4OH. The packages are mixed and applied; as the water and
ammonia evaporate, the COOH functional latex and the polyamine rapidly form a crosslinked salt [51].
Acetoacetate-functional (Section 17.6) latexes can be cross-linked with polyamines,
but pot life is short before coalescence is adversely affected [52]. A BA/MMA/MAA/
acetoacetoxyethyl methacrylate latex made using dodecyl mercaptan for MW control is
used in traffic paint with aminoethylaminopropyltrimethoxysilane and a dimethylaminoethyl methacrylate/AA latex, neutralized with ammonia. Pot life is said to be
adequate [53].
Carbodiimides (Section 17.8) can be used as cross-linkers [54]. A water-emulsifiable
polycarbodiimide that is terminated with a mixture of the methyl ether of diethylene
glycol and the methyl ether of a polyethylene glycol is used as a cross-linker for COOHfunctional latexes. The design of the latex markedly affected properties. The best
properties were obtained with a BA/MMA/MAA latex that was prepared with the core
containing 2% MAA and the shell 1% MAA [55].
Another system studied used two latexes with different functional groups. One latex
was a copolymer of t-butylcarbodiimidoethyl methacrylate and 2-ethylhexyl methacrylate
(EHMA), the other was a copolymer of MA, EHMA, and MAA. The carbodiimide group
and COOH group react to form N-acylurea groups. This graft copolymer is said to increase
the solubility of one latex in the other. The reaction occurs at ambient temperature, but
improved properties were obtained at 608C, because interdiffusion increases more
rapidly with temperature than the cross-linking reaction [48].
Polyfunctional aziridines (Section 17.7) such as the propylene imine addition product
with pentaerythritol triacrylate serve as cross-linkers for carboxylic acid-functional latexes
[56]. Pot lives of 48 to 72 hours are reported.
Carboxylic acid– functional latexes can also be cross-linked with epoxysilanes, such as
b-(3,4-epoxycyclohexyl)ethyltriethoxysilane [57]. Addition of an emulsion of this crosslinker to the latex increased solvent resistance and hardness, especially when baked
10 minutes at 1168C. Latexes with a high COOH content and equivalent amounts of
epoxysilane showed the largest improvements. Package stability is reported to be at
least one year. Cross-linking can be catalyzed: for example, using 1-(2-trimethylsilyl)
propyl)-1H-imidazole. m-Isopropenyl-a,a-dimethylbenzyl isocyanate (TMI) (Section
12.3.2) reacts slowly with water and can be used to make thermosetting latexes [58].
Acrylic latexes with surface methacrylate groups are prepared by treating a carboxylic
acid –functional acrylic latex with an emulsion of cyclohexylcarbodiimidoethyl methacrylate. The carbodiimide group reacts with COOH groups on the surfaces of latex particles to
give surface methacrylate ester groups. Films prepared with the latex (to which t-butyl
peroxybenzoate was added) were cured at 1608C for 30 minutes, or by UV after adding
a photoinitiator [59].
There have been fewer reports of thermosetting latexes that cross-link at room temperature and are storage stable for the long times required for architectural paints. One
approach is to use a cross-linking mechanism based on autoxidation. A latex with
allylic substitution cross-links on exposure to air after application [60,61]. Hybrid
alkyd – acrylic latexes have been prepared by dissolving an oxidizing alkyd (Section
15.1) in the monomers used in emulsion polymerization, yielding a latex with alkyd
196
LATEXES
grafted on the acrylic backbone [62 – 64]. Stable thermosetting latexes can be prepared
using triisobutoxysilylpropyl methacrylate as a comonomer [57]. In contrast to the
ethoxy derivative, the isobutoxy derivative has sufficient hydrolytic stability to permit
emulsion polymerization at 80 to 908C without extensive reaction with water. The resulting latex has a package stability of over a year, yet cross-links after application in a week
using organotin catalysts. This unusual combination of storage stability and reactivity may
signify that the trialkoxysilyl groups hydrolyze during storage, but in the presence of large
amounts of water, do not cross-link until the water evaporates after the coating is applied.
Acrylic latexes prepared using vinyltriethoxylsilane as a comonomer are reported to be
package stable if buffered with NaCO3. Compared to a similar latex without the silane
comonomer, films show increased solvent resistance and exterior durability and improved
heat flow resistance [65].
Core – shell latexes, prepared using N-(3-keto-2,2-dimethylbutyl)acrylamide (diacetoneacrylamide) as a comonomer, resulting in ketone groups in the shell for cross-linking
with adipic dihydrazide, have been reported [66], as have corresponding latexes having
a continuous gradient composition rather than a core – shell structure. Diffusion of the
cross-linker through the latter particles before cross-linking is said to afford better film
properties than reaction with keto groups at the surface of the core – shell particles [67].
GENERAL REFERENCES
El-Aasser, M. S.; Lovell, P. A., Eds., Emulsion Polymerization and Emulsion Polymers, Wiley,
New York, 1997.
Gilbert, R. O., Emulsion Polymerization: A Mechanistic Approach, Academic Press, San Diego, CA,
1995.
Vanderhoff, J. W., J. Polym. Sci. Polym. Symp., 1985, 72, 161 – 198.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Hoy, K. L., J. Coat. Technol., 1979, 51(651), 27.
Jones, F. N.; et al., Prog. Org. Coat., 2005, 52, 9.
Yang, H. S.; et al., J. Coat. Technol., 2005, 2(13), 44.
Rothmann, H.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2001, pp. 465 – 479.
Anderson, L. L.; Brouwer, W. M., J. Coat. Technol., 1996, 58(855), 75.
Fernandez, A. M.; et al., Prog. Org. Coat., 2005, 53, 246.
Lynn, J. L., Jr.; Bory, B. H., Kirk– Othmer Encyclopedia of Chemical Technology, 4th ed.,
Wiley, New York, 1997, Vol. 23, pp. 478 – 541.
Aslamazova, R., Prog. Org. Coat., 1995, 25, 109.
El-Aasser, M.; et al., J. Coat. Technol., 2001, 73(920), 51.
Holmberg, K., Prog. Org. Coat., 1992, 20, 325.
Grade, J.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 145 –157.
Nguyen, B. D.; Rudin, A., J. Coat. Technol., 1986, 58(736), 53.
Craig, D. R., J. Coat. Technol., 1989, 61(779), 49.
REFERENCES
197
14. Hansen, F. K., in Polymer Latexes, Daniels, E. S.; et al., Eds., American Chemical Society,
Washington, DC, 1992, pp. 12– 27.
15. van Berkel, K. Y.; et al., Macromolecules, 2003, 36, 3921.
16. Herrera-Ordonez, J.; et al., J. Macromol. Sci. C Polym. Rev., 2004, C44(3), 207.
17. Gardon, J. L., in Technology for Waterborne Coatings, Glass, J. E., Ed., ACS Symp. Seri 663,
American Chemical Society, Washington, DC, 1997, p. 27.
18. El-Aasser, M. S.; et al., J. Polym. Sci. Polym. Chem. Ed., 1983, 21, 2363.
19. Alahapperuma, K.; Glass, J. E., Prog. Org. Coat., 1992, 21, 53.
20. Oprea, S.; Dodita, T., Prog. Org. Coat., 2001, 42, 194.
21. Chou, C.-S.; et al., J. Coat. Technol., 1987, 59(755), 93. Rhoplex Multilobe 200, Techn. bull.,
Rohm & Haas, Philadelphia, PA, 1992.
22. Padget, J. C., J. Coat. Technol., 1994, 66(839), 89.
23. Learner, T., Stud. Conserv., 2001, 46, 225 – 241.
24. Emulsion Polymerization of Acrylic Monomers, tech. bull. CM-104 A/cf, Rohm & Haas,
Philadelphia, PA.
25. Emelie, B.; et al., Makromol. Chem., 1988, 189, 1879.
26. Mercurio, A.; et al., J. Oil Colour Chem. Assoc., 1982, 65, 227.
27. Winnik, M. A.; Feng, J. J. Coat. Technol., 1996, 68(852), 39.
28. Eckersley, S. A.; Helmer, B. J., J. Coat. Technol., 1997, 69(864), 97.
29. Zhu, P.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 139 – 144.
30. Lorah, D. P.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2004, Paper 19.
31. Chou C.-S.; et al., U.S. patent 6,838,507 (2005).
32. Friedli, H. R.; Keillor, C. M., J. Coat. Technol., 1987, 59(748), 65.
33. Nair, M., Prog. Org. Coat., 1992, 20, 53.
34. Webster, D. C.; Crain, A. L., Prog. Org. Coat., 2002, 45, 43.
35. Misra, S. C.; et al., J. Polym. Sci. Polym. Chem. Ed., 1983, 21, 2383.
36. Arzamendi, G.; Asua, J. M., J. Appl. Polym Sci., 1989, 38, 2019.
37. Vandezande, G. A.; et al., in Emulsion Polymerization and Emulsion Polymers, Lovell, P. A.;
El-Aasser, M. S., Eds., Wiley, New York, 1997, pp. 563 –587.
38. Vandezande, G. A.; Rudin, A., J. Coat. Technol., 1994, 66(828), 99.
39. Bassett, D. R., J. Coat. Technol., 2001, 73(912), 43.
40. Heldmann, C.; et al., Prog. Org. Coat., 1999, 35, 69.
41. Prior, R. A.; et al., Prog. Org. Coat., 1996, 29, 209.
42. Decocq, F.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1997,
p. 169.
43. De Bruyn, H.; et al., Macromolecules, 2002, 35, 8371.
44. Caldwell, R.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
1999, pp. 60– 75.
45. Gilicinski, A. G.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2000, pp. 516 – 526.
46. Avramidis, K. S., FSCT Annual Meeting, 2004.
47. Avramidis, K. S.; et al., U.S. patent 6,339,447 (2001).
48. Taylor, J. W.; Winnik, M. A., JCT Res., 2004, 1(3), 163.
198
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
LATEXES
Winnik, M. A., J. Coat. Technol., 2002, 74(925), 49.
Huang, Y.; Jones, F. N., Prog. Org. Coat., 1996, 28, 133.
Friel, J. M.; et al., U.S. patent 6,689,824 (2004).
Geurink, P. J. A.; et al., Prog. Org. Coat., 1996, 27, 73.
Hermes, A. H.; et al., U.S. patent appl. 20030224184.
Taylor, J. W.; Bassett, D. W., in Technology for Waterborne Coatings, Glass, J. E., Ed.,
American Chemical Society, Washington, DC, 1997, p. 137.
Porzio, R. S.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2003, pp. 129 – 143.
Pollano, G., Polym. Mater. Sci. Eng., 1996, 77, 73.
Chen, M. J.; et al., J. Coat. Technol., 1997, 69(875), 49.
Inaba, Y.; et al., J. Coat. Technol., 1994, 66(833), 63.
Taylor, J. W.; et al., Prog. Org. Coat., 1999, 35, 215.
Monaghan, G., Polym. Mater. Sci. Eng., 1997, 76, 178.
Collins, M. J.; et al., Polym. Mater. Sci. Eng., 1997, 76, 172.
Nabuurs, T.; et al., Prog. Org. Coat., 1996, 27, 163.
Gooch, J. W.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
1997, p. 366.
Van Hamersfeld, E. M. S.; et al., Prog. Org. Coat., 1999, 35, 235.
Guyer, K. L.; Gai, W. J., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2006, pp. 85– 96.
Lee, S.-B.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2003,
pp. 523 –534.
Mestach, D.; Ahmed, M., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 2001, pp. 195 –210.
10
Polyester Resins
In fibers and plastics, polyester refers to high molecular weight (MW), partially
crystalline, linear, thermoplastic polymeric esters of a short-chain diol and terephthalic
acid. In coatings, however, it is used to describe a different sort of material. Most coating
polyesters have relatively low MWs and are amorphous and branched, and they must be
cross-linked to form useful coating films. Furthermore, the term is applied only to certain
polyesters, those prepared from polyols and polybasic acids. Alkyd resins are also polyesters in the chemical sense of the word, but they are not called polyesters in coatings.
As discussed in Chapter 15, alkyds are prepared from polyols and dibasic acids, and in
addition, monobasic acids, usually derived from vegetable oils. Coating polyesters are
sometimes called oil-free polyesters or oil-free alkyds to distinguish them from alkyds.
Polyesters are made by step-growth polymerization; the general principles are discussed in Section 2.2.2. The relatively low MW polyesters used in coatings are usually
made from mixtures of diols, triols, and dibasic acids. Most commonly, excess polyol is
used; hence, the polyesters are hydroxy-terminated polyesters. They are most commonly
cross-linked with melamine– formaldehyde (MF) resins or polyisocyanates. Resins with
excess carboxylic acid are also made; these carboxylic acid – terminated polyesters are
cross-linked with epoxy resins, MF resins, or 2-hydroxyalkylamides. Water-reducible
polyesters are synthesized with both terminal hydroxyl and carboxylic acid groups; they
are usually cross-linked with MF resins.
Branched polyesters are made from monomer mixtures that include one or more monomers having functionality F . 2. Kinetic analysis of such polymerizations is complex, but
a few generalizations can be pointed out. As the proportion of monomer of F . 2
increases, M̄n, M̄w/M̄n, and the number average functionality f̄n all increase. The
average F̄ must be controlled to avoid gelation at high extents of reaction.
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
199
200
POLYESTER RESINS
In general terms, thermosetting polyesters give coatings with better adhesion to metal
substrates and better impact resistance than thermosetting acrylics (TSAs) (Section 8.2).
On the other hand, TSAs give coatings with superior water resistance and exterior durability. The differences can be attributed to the presence of the ester linkages in the backbone of polyesters: linkages that impart flexibility but are vulnerable to hydrolysis. Thus,
polyesters tend to be used for one-coat coatings on metal, and TSAs tend to be used in
applications for which exterior durability and moisture resistance are particularly important, often with primers that assure excellent adhesion of the coating system to the metal.
The superior exterior durability of TSAs results partly from the greater effectiveness of
UV screeners in acrylics than in polyesters. Many exceptions to these generalizations
can be found (Section 5.2.1).
10.1. HYDROXY-TERMINATED POLYESTER RESINS FOR
CONVENTIONAL SOLIDS COATINGS
Most hydroxy-terminated polyesters are made by coesterifying four types of monomers:
two polyols (a diol and a triol) and two diacids (an aliphatic dibasic acid or its anhydride
and an aromatic dicarboxylic acid or its anhydride). Tribasic and cycloaliphatic acids are
used to some extent. The ratio of moles of dibasic acid to polyol must be less than 1 to
give terminal hydroxyl groups and avoid gelation. MW is controlled by this ratio; the
smaller the ratio, the lower the MW. The MW distribution, M̄n, and f̄n are all controlled
by the diol/triol ratio; f̄n is critical because it affects the potential cross-link density after
cure. With the same mole ratio of dibasic acid to total polyol, an increase in triol
component increases the average number of hydroxyl groups per molecule, decreases
hydroxy equivalent weight, broadens MW distribution, and increases cross-link
density of a fully cross-linked film. The ratio of aromatic to aliphatic dibasic acids is
the principal factor controlling the Tg of the resin, but differences in polyol structure
also affect Tg. Mixtures of polyols and diacids also reduce the tendency of the resins
to crystallize.
In commercial processes, mixtures of the monomers are esterified at 220 to 2408C
while water is removed. Organotin compounds, titanium orthoesters, or zinc acetate
are frequently used as catalysts. Strong protonic acids also catalyze esterification but
can cause side reactions and discoloration. The reaction is continued until a high
conversion has been reached. Acid numbers of resins to be cross-linked with MF
resins are generally 5 to 10 mg KOH per grams of resin solids; those for use with polyisocyanates are commonly esterified until the acid number is less than 2. Polyesters
made from polyfunctional reactants can cross-link during polymerization, causing
gelation. Since the probability of gelation increases with conversion, and high conversions are generally desirable, an equation relating formula composition, extent of reaction, and probability of gelation would be useful. Progress has been made in predicting
gelation in simple systems in which the reactivity of all hydroxyl groups and carboxylic
acid groups are equal, but even with these “simple” systems, the equations give only fair
predictions because they do not take into consideration the formation of cyclic esters
(see Ref. [1] for a discussion of such equations). In the case of polyesters used in coatings, the situation is far from simple; the reactivity of the various functional groups commonly varies substantially, and side reactions occur. Therefore, the textbook gelation
equations are of little value for polyesters used in coatings. If the mole ratio of
10.1. HYDROXY-TERMINATED POLYESTER RESINS FOR CONVENTIONAL SOLIDS COATINGS
201
dibasic acid to polyol exceeds, or even closely approaches 1, gelation may occur before
esterification is essentially complete. This is seldom a problem in making hydroxyfunctional polyesters for coatings, but it can become critical in synthesis of alkyds
(Section 15.1).
Misev [2] has published equations that permit calculation of ratios of polyols and
polybasic acids to use to make polyesters with a desired M̄n and f̄n at a desired extent
of reaction of the acid groups. The equations are complex and require iterative computer
solution. The equations are simplified if the choice of raw materials is limited, as it generally is, to diols, triols, and dibasic acids, and if essentially all of the acid groups are to
be esterified. In commercial practice, polyesters are cooked until about 98 to 99.5% of
the carboxylic acid groups are esterified. Therefore, an assumption of complete esterification is reasonable. In these simplified equations Np2 represents the number of moles of
diol; Np3, the number of moles of triol; and Na, the number of moles of dibasic acid.
Mp2 represents the MW of the diol; Mp3, the triol; and Ma, the dibasic acid. If more
than one dibasic acid (or other component) is used, additional terms would be added
to the equation for each additional component. Using these equations, one can calculate
how much diol, triol, and dibasic acid should be used in an initial laboratory cook for a
desired M̄n and f̄n.
In making hydroxy-terminated polyesters, the number of moles of polyol is one more
than that of dibasic acid:
Np2 þ Np3 ¼ Na þ 1
(10:1)
Na ¼ Np2 þ Np3 1
(10:2)
or
The relationship between the average functionality and the moles of reactants is
expressed as
f n ¼ 2Np2 þ 3Np3 2Na
(10:3)
Substituting the value of Na from Eq. 10.2 into Eq. 10.3 gives Eqs. 10.4 and 10.5
f n ¼ Np3 þ 2
(10:4)
Np3 ¼ fn 2
(10:5)
or
When the esterification is carried out 2 mol of water (total MW ¼ 36) is split out for
each mole of dibasic acid, resulting in Eq. 10.6 for M̄n:
Mn ¼ Np2 Mp2 þ Np3 Mp3 þ Na (Ma 36)
(10:6)
If an anhydride is used, only 1 mol of water (M ¼ 18) is split out for each mole of
anhydride, and the appropriate adjustment in Eq. 10.6 is needed. Substituting in
202
POLYESTER RESINS
TABLE 10.1. Starting Formulation for a Conventional Polyester Resin
Raw Material
NPG
TMP
IPA
AA
Weight (g)
Moles
Equivalents
2050
270
1740
1530
19
3
10.5
10.5
38
9
21
21
Eq. 10.6 for Na using Eq. 10.2 and for Np3 using Eq. 10.5 gives Eq. 10.7, which relates
M̄n to f̄n:
n ¼ Np2 Mp2 þ ( f n 2)Mp3 þ (Np2 þ f n 3)(Ma 36)
M
(10:7)
A wide range of polyesters has been used commercially. In conventional polyester – MF
coatings, polyesters with M̄n on the order of 2000 to 6000, M̄w/M̄n on the order of 2.5 to 4,
and with f̄n of 4 to 10 hydroxyl groups per molecule are usual. Table 10.1 gives an example
of an initial laboratory formulation calculated for a polyester with an M̄n of 5000 and an f̄n
of 5 using Eqs. 10.1 to 10.7. Neopentyl glycol is NPG, trimethylolpropane is TMP, isophthalic acid is IPA, and adipic acid is AA. Structural formulas for these compounds
are given in Sections 10.1.1 and 10.1.2.
A formulation such as that in Table 10.1 does not take into consideration all of the
complexities of polyesterification; thus, it can only be expected to provide a starting
point for experimentation. A common complication is partial loss of monomer from
the reactor by volatilization. The diol is usually the most volatile component, so it is generally necessary to use excess diol. The amount of excess diol depends on the particular
reactor and conditions, the efficiency of separating water and diol, the rate of flow of
inert gas, the temperature of reaction, and so on. A formula established in the laboratory
has to be adjusted when production is scaled up in a pilot plant, again when it is scaled
up in a production reactor, and yet again when production is shifted from one reactor
to another.
Esterification is a reversible reaction, the formation of high yields of ester requires
removal of the water. A few percent of a reflux solvent, such as xylene, may be added
to accelerate removal of water and to assist in return of volatilized diol to the reactor.
Alternatively, an inert gas purge can be used in the later stages of the process to help
remove the last of the water. Polyols can undergo self-condensation to form polyethers,
which also produces water. Ester groups that form during the process may hydrolyze or
transesterify and re-form many times, leading to a mixture of kinetic and thermodynamic
control in formation of the final products.
10.1.1. Selection of Polyols
Polyols are selected on the basis of cost, rate of esterification, stability during high temperature processing (minimal decomposition and discoloration), ease of separation from
water during processing, viscosity of esters at equal MW and functionality, effect on
Tg, rate of cross-linking with MF resins (or other cross-linkers), and hydrolytic stability
of their esters. Obviously, compromises are necessary. The following polyols are
10.1. HYDROXY-TERMINATED POLYESTER RESINS FOR CONVENTIONAL SOLIDS COATINGS
203
representative of the many available:
The most widely used diol is probably NPG, and the most widely used triol is TMP.
CHDM esterifies significantly more rapidly than NPG, and trimethyl-1,3-pentanediol
(TMPD) more slowly [3]. Generally, but not always, the slower the rate of esterification,
the greater the resistance of the ester to hydrolysis. The hydrolytic stability of esters of
NPG and TMP is better than those from less hindered glycols such as ethylene glycol
or propylene glycol. Neopentyl glycol (bp 2138C) is volatile at esterification temperatures,
so well designed fractionating and condensing equipment are needed to permit the
removal of water with a minimum loss of NPG during processing at 220 to 2408C.
Most diols start decomposing in the presence of strong acids at temperatures above
about 2008C, so strong acid esterification catalysts should be avoided. Organotin compounds and titanium orthoesters are examples of appropriate catalysts.
Although the kinetics and mechanisms of esterification and hydrolysis reactions have
been extensively studied from a fundamental point of view, the literature contains few
basic data dealing with polyols of most interest in coatings resins. Turpin [4] has proposed
an empirical steric factor based on Newman’s rule of six that may be useful in predicting
the relative hydrolytic stability of esters. The steric factor is based on the number of
substituents in the 6 and 7 positions of an ester, starting with the carbonyl oxygen as 1:
steric factor ¼ 4(No. 6 position atoms) þ (No. 7 position atoms)
For example, for an NPG ester,
The value of 21 for NPG esters is the same for TMP esters, but for ethylene glycol and
204
POLYESTER RESINS
1,6-hexanediol esters, the steric factors are only 13 and 15, respectively. Turpin’s steric
factor seems to correlate reasonably well with experience and with the limited published
data for esters of acyclic alcohols. Turpin also pointed out that an anchimeric, or neighboring
group, effect may be involved. For example, the location of the hydroxyl group in a half ester
of ethylene glycol may promote the hydrolysis of its neighboring ester through a cyclic transition state and contribute to the poorer hydrolytic stability of ethylene glycol esters as compared to those of 1,6-hexanediol [4]. TMPD is also used frequently. Despite its relatively low
steric factor, some polyesters made from it have good exterior durability.
Esters of cycloaliphatic diols, such as cyclohexanedimethanol (CHDM), seem more
resistant to hydrolysis than predicted by Turpin’s steric factor. A comparison of the
results of testing films of coatings made with a series of polyesters from several polyols
has been published [5]. Coatings based on CHDM polyesters cross-linked with MF resins
showed the best test results for properties related to hydrolytic stability. In contrast to
general behavior, CHDM also esterifies more rapidly than NPG [3]. It has also been
shown that CHDM polyesters give cured films with MF resins more rapidly than NPG-based
polyesters [6]. A 1 : 1 mixture of (cis,trans) 1,3-cyclohexanedimethanol and (cis,trans)
1,4-cyclohexanedimethanol is a liquid that is reported to give polyesters for cross-linking
with either MF resins or polyisocyanates that exhibit an excellent balance of hardness
and flexibility with excellent acid etch resistance and durability [7]. The enhanced hydrolytic resistance of CHDM may arise from hydrophobicity of the cyclohexane ring.
10.1.2. Selection of Polyacids
Most polyesters are made using a mixture of aromatic and aliphatic diacids. The ratio of
aromatic to aliphatic acid is a major factor that controls the Tg of the resin.
Aromatic acid esters hydrolyze more slowly than aliphatic esters, unless there is an
anchimeric effect. Isophthalic acid (IPA) is the predominant aromatic acid used even
though phthalic anhydride (PA) can be processed at lower cost. The preference for IPA
10.2. POLYESTER RESINS FOR HIGH-SOLIDS COATINGS
205
is based on superior exterior durability of coatings made with IPA polyesters; this has been
attributed to greater hydrolytic stability of isophthalic esters. Half esters of (ortho)phthalic
acid are more easily hydrolyzed than half esters of isophthalic acid in the pH range 4 to 8
[8]. This difference results from anchimeric assisted hydrolysis by the ortho-carboxylic
acid group. During exterior exposure, resistance to hydrolysis in contact with rainwater,
which has a pH in the range 4 to 6, is most important. Under these conditions, isophthalic
acid polyesters are more resistant than phthalic acid polyesters. On the other hand, monoand diesters of isophthalic acid hydrolyze more rapidly than those of phthalic acid under
alkaline conditions [9,10].
A processing cost advantage of phthalic anhydride results from its lower melting point
(1318C); it can be handled as a molten liquid, it is readily soluble in the reaction mixture,
and its reactive anhydride structure rapidly forms monoesters at about 1608C. Isophthalic
acid (mp . 3008C) is more difficult to process because it dissolves more slowly in the reaction medium, and lower reactivity slows the processing. Due to a longer time at high temperatures, volatilization losses and acid catalyzed side reactions (e.g., etherification or
dehydration of polyols) are more serious problems with IPA than with PA (Section 15.5.2).
Adipic acid is probably the most widely used aliphatic dibasic acid. Succinic and
glutaric acids give esters with inferior hydrolytic stability, presumably due to the anchimeric effect [4]. Longer chain acids such as azelaic and sebacic acids may give somewhat
better hydrolytic stability and greater reduction of Tg (on a molar basis) than those given
by adipic acid, but are more expensive. Dimer acids, derived from dimerization of drying
oil fatty acids and hence predominantly C36 acids, are widely used. Fatty acid dimerization
is discussed in Section 14.3.1. Dimer acids are relatively inexpensive and are available in
various grades. The high quality grades of dimer acids normally used in polyesters have
little monobasic or tribasic acid contaminants and have been hydrogenated to saturate
the double bonds in order to reduce oxidative degradation. Dimer acid –isophthalic polyesters impart an excellent balance of properties to MF cross-linked coil coatings, providing
high resistance to marring combined with excellent resistance to cracking during fabrication of coated metal.
Cycloaliphatic dibasic acids or anhydrides can be used. Hexahydrophthalic anhydride
(HHPA) is said to give exterior durability comparable to IPA. Aside from being aliphatic, HHPA half esters may be less susceptible to anchimeric assisted hydrolysis
than are phthalic acid half esters possibly owing to isomerizing to the E (trans)
isomers during processing. 1,4-Cyclohexanedicarboxylic acid (CHDA) provides a
good balance of hardness, flexibility, and stain resistance. CHDA esters are much
more stable than IPA esters to hydrolysis at pH 8 to 9, but somewhat less stable to
hydrolysis at pH 4 to 5 [9,10].
A statistically designed evaluation of the factors affecting the exterior durability
of IPA/adipic acid polyester/MF coatings showed that the major factor was the
IPA/adipic acid ratio. The polyols were NPG and TMP. Higher IPA content polyesters
gave coatings with improved durability [11].
10.2. POLYESTER RESINS FOR HIGH-SOLIDS COATINGS
Low viscosity polyesters are required for reduction of VOC emissions; viscosity of concentrated polyester solutions depends on several variables. MW and MW distribution
are two important factors. The number of functional groups per molecule also affects
206
POLYESTER RESINS
the viscosity; increasing the number of hydroxyl groups (and to an even greater extent, the
number of carboxylic acid groups) increases the probability of intermolecular hydrogen
bonding, in turn increasing viscosity. The hydrogen-bond effect can be minimized, but
not eliminated, by using hydrogen-bond acceptor solvents such as ketones. The Tg of
the resin also affects viscosity at a given concentration. For this reason, reducing the
ratio of aromatic to aliphatic dibasic acids and using acyclic polyols rather than cyclic
polyols, such as CHDM [5,6], give lower viscosities, but there is a lower limit of Tg
below which desired film properties cannot be attained.
Polyesters have significant advantages over thermosetting acrylics (TSAs) in highsolids coatings. In contrast to the difficulty of making low MW TSAs with at least two
hydroxyl groups on essentially all molecules, the synthesis of polyesters with two or
more hydroxyl groups on essentially all of the molecules is relatively easy. With the
low mole ratios of dibasic acids to polyols (2 : 3 is typical) required to make low MW
resins, the probability of cyclization reactions to form nonfunctional materials is small.
Virtually all of the final reactions result in terminal hydroxyl groups. While the lowest viscosity is obtained with difunctional resins, such resins also require the most careful formulation and curing to achieve good film properties. If a hydroxyl is left unreacted on a
conventional polyester with an average of five or so hydroxyl groups per molecule, the
effect on film properties would probably be relatively small. However, failing to react
all of the hydroxyl groups on a resin with only two hydroxyl groups per molecule
would magnify the effect. A compromise is required; commonly, an f̄n of between two
and three hydroxyl groups per molecule is used.
The major factor controlling the MW (M̄n) of a polyester is the mole ratio of dibasic
acid to polyol. A ratio on the order of 2 : 3 is typical. Loss of polyol during production
of a polyester results in a higher MW than predicted from the starting ratio. Condensers
are designed to remove water rapidly while returning a high fraction of NPG (or other
polyol), but none is completely efficient. It is necessary to add some extra glycol to compensate for losses. Since production equipment and conditions vary, the final decision on
the mole ratio to load must be based on experience with the particular reactor and with the
particular formula. The amount of glycol being lost can be estimated by checking the
refractive index of the water being removed from the reaction.
Some low MW components of a resin may volatilize when a thin film is baked. Such
loss has to be counted as part of the VOC emissions, together with solvent loss. Belote and
Blount studied this volatility both in the absence of cross-linking agent and in its presence
[12]. They made a series of model resins from NPG and a 1 : 1 mole mixture of AA and
IPA; the polyacid/polyol ratio was varied from 1 : 2 to 1 : 1.15. The presence of the crosslinker (class I MF resin) reduced volatile loss. They concluded that an M̄n of about 800 to
1000 was optimum for achieving the lowest VOC with resins having normal MW distributions. At M̄n around 600, so much low MW fraction volatilized that the total VOC was
higher than at an M̄n of 800, even though the amount of solvent required was lower. A
broad MW distribution not only increases volatile losses because of the low MW material,
but also gives a higher viscosity because of the disproportionate effect on viscosity of the
high MW fraction. Lowest viscosity results from making as narrow a MW distribution as
possible.
The use of TMPD in combination with a two-stage addition of the triol TMP during
synthesis of a polyester gives lower viscosity and improved film properties [13]. The
use of TMPD, an unsymmetrical diol, presumably gives narrower MW distribution
because of the differential reactivity of the two hydroxyl groups. A drawback is the
need for somewhat higher baking temperatures for coatings, presumably a result of the
10.2. POLYESTER RESINS FOR HIGH-SOLIDS COATINGS
207
TABLE 10.2. High-Solids Polyester Formulations
(Equivalents)
Raw Material
TMPD
TMP
IPA
AA
Resin 1
Resin 2
Resin 3
11.96
1.72
4.56
4.56
11.96
0.86
4.56
4.56
12.24
0.88
4.38
4.38
0.86
1075
76.5
0.88
1000
77.6
Second Stage
TMP
Resin M̄n
Coating NVW
—
1500
75.6
low reactivity of the hindered terminal hydroxyl groups derived from TMPD. Three of the
resin formulations used in this study are given in Table 10.2. Addition of the TMP in two
stages resulted in a reduction in M̄n from 1500 to 1100 (compare resins 1 and 2) with a
corresponding reduction in viscosity. Further reduction was achieved by adjusting the
monomer ratios (resin 3).
A polyester diol with a narrow MW distribution and a reported M̄n of 425 is commercially available [14]. This resin is presumably manufactured by preparing a polyester with
excess CHDM and then distilling off the low MW fraction with a vacuum thin-film evaporator. Low viscosity polyester diols and triols prepared by reacting caprolactone with a
diol or a combination of diol and triol are also available [15]. These types of low-M̄n
resins seldom provide satisfactory coatings when used by themselves, but they are
useful in blends to increase the solids content of somewhat higher MW polyester or
acrylic-based coatings. They are often called reactive diluents.
Linear polyesters made by transesterification of 1,4-butanediol with a mixture of
glutarate, adipate, and azaleate methyl esters give low viscosity hydroxy-terminated
polyesters having an M̄n of 680 [16]. They gave solvent-free two package coatings with
hexamethylene diisocyanate trimer. However, with Class I MF resins, baked films were
soft due to low Tg.
A relatively new analytical method, matrix-assisted laser desorption ionization
time-of-flight mass spectrometry (MALDI-ToF-MS), has made it possible to study the
complex mixtures of molecules in coatings polymers in unprecedented detail. For
example, Willemse et al. [17] studied polyesters designed for very high solids and solventless liquid formulations. One such polyester was made from 1,4-butanediol, adipic acid
(AA), and isophthalic acid (IPA) in a 1 : 0.375 : 0.375 mole ratio with an organotin catalyst. Figure 10.1(a) shows an overall MALDI spectrum of the product. Each cluster of
peaks represents molecules with a specific degree of polymerization. Figure 10.1(b) is
an expanded spectrum focused on molecules composed of seven diol and six diacid monomers. The numbers designate the numbers of AA and IPA residues in the molecules; for
example, the peak designated 4-2 with a mass of 1369.57 Da contains four AA and two
IPA residues. Here the peaks are clusters because of the presence of molecules containing
isotopes (e.g., 2H, 18O). Further expansion makes it possible to identify small peaks from
cyclic structures and from molecules with one acid and one hydroxyl end group.
A significant finding of this study was that the compositions of molecules within this
polyester are close to random as calculated by Bernouillan statistics. Sequences of monomers within molecules cannot be determined by MALDI-ToF-MS alone.
208
POLYESTER RESINS
Figure 10.1. (a) MALDI-ToF-MS spectrum of a polyester resin made from 1,4-butanediol, adipic acid
(AA), and isophthalic acid (IPA) in a 1 : 0.375 : 0.375 mole ratio; ( b) expanded spectrum focused on molecules composed of seven diol and six diacid monomers. (Reprinted with permission from Ref. [17]; copyright # 2005 American Chemical Society.)
In another study, Arnould et al. [18] used MALDI-ToF-MS, tandem mass spectroscopy
(MS/MS), and gel permeation chromatography (GPC) to study a higher-MW polyester. It
was made from NPG, TMP, TPA, and AA in a 50 : 1 : 45 : 5 ratio, a composition similar to
powder coating resins (Section 10.6). The polymer was fractionated by GPC to facilitate
study of higher MW fractions. Only low levels of cyclic products were detected. The MS/
MS spectra suggested that within individual molecules, certain specific sequences of
monomers are favored.
10.3. CARBOXYLIC ACID – TERMINATED POLYESTER RESINS
Carboxylic acid –terminated polyesters are copolymers of a diol and a molar excess of
dibasic acid. A modest amount of a tri- or polyfunctional monomer, most commonly
trimellitic anhydride (TMA), is almost always included to increase f̄n. Carboxylic
10.5. WATER-REDUCIBLE POLYESTER RESINS
209
acid –functional polyesters can be cross-linked with MF resins but in most cases are crosslinked with epoxy resins or 2-hydroxyalkylamides, as discussed in Sections 13.3.2 and
17.5, respectively. With high ratios of high melting point diacids such as IPA, it is
usually preferable to use a two-step process. The slow dissolution and high melting
point of IPA prevent a clean reaction when the IPA is used in excess. First, one prepares
a hydroxy-functional polyester with the IPA and then reacts the terminal hydroxyl groups
with a lower melting acid or anhydride. Careful control is needed to assure reproducible
products, since varying degrees of transesterification occur in the second stage. Final
product structure is governed both by kinetic and thermodynamic factors; thus, resin
properties are affected by reaction time and other process variables.
10.4. CARBAMATE-FUNCTIONAL POLYESTER RESINS
Carbamate-functional resins cross-linked with MF resins have been shown to give clear
coats with good mar and acid resistance, exterior durability, and hydrolytic resistance.
An example is an oligomer prepared by reacting citric acid with glycidyl neodecanoate
and then transesterified with methyl carbamate [19].
10.5. WATER-REDUCIBLE POLYESTER RESINS
Polyesters are also made for use in waterborne coatings with a reduced VOC. Most such
polyesters have both hydroxyls and carboxylic acids as terminal groups. As with other
water-reducible resins, acid numbers in the range 35 to 60 are required to give amine
salt solutions in solvent that can be diluted with water to give reasonably stable dispersions
of aggregates of resin molecules swollen with water and solvent. These resins show abnormal viscosity dilution curves similar to those described in Section 8.3 for water-reducible
acrylic resins.
In theory, one could prepare such water-reducible polyesters by simply stopping the
esterification of a combination of diol, triol, AA, and IPA at the desired acid number;
in practice, it is almost impossible to do so reproducibly. Rather, it is necessary to find
combinations of polybasic acids that have sufficiently different reactivity to permit
control of the ratio of unreacted hydroxyl and carboxyl groups. The most widely used
method has been to prepare a hydroxy-functional resin and then add enough TMA to esterify a fraction of the hydroxyl groups, generating two carboxyl groups at each site. Advantage is taken of the higher reactivity of the anhydride groups of the TMA at 1808C relative
to the carboxylic acid groups formed in the reaction. Table 10.3 gives an example of a formulation for preparing a water-reducible polyester [20].
This approach has been used on a large scale, but it has disadvantages. Most seriously, the
ester linkages in the resin are vulnerable to hydrolysis during storage of the formulated
coating. The ester group of partially esterified TMA is particularly sensitive to hydrolysis,
owing to anchimeric assistance by the adjacent carboxylic acid group. Hydrolysis of the
TMA partial ester results in removal of the solubilizing carboxyl groups and thereby destabilizes the resin dispersion. The film properties may also be affected adversely. In addition,
the use of primary alcohols as a solvent, as recommended in Table 10.3, has been found to
be undesirable. Primary alcohols have been found to esterify carboxylic acid groups and
transesterify ester groups during the thinning at 1608C and at a slow but appreciable
rate during storage of the resin. This problem can be minimized by using a secondary
210
POLYESTER RESINS
TABLE 10.3. Water-Reducible Polyester Formulation
Material
NPG
Adipic acid
TMA
IPA (85 : 15 IPA to TPA)
Weight (g)
Mole Ratio
Equivalent Ratio
685
192
84
655
1.0
0.2
0.067
0.6
1.0
0.2
0.1
0.6
React at 2358C until the acid number is 16–18; cool to 1808C, then add:
TMA
84
0.067
0.1
React at 1808C until the acid number is 40–45; cool to 1608C, and thin to 80% solids with diethylene
glycol monobutyl ether.
alcohol as solvent, along with careful control of process temperatures. 1,4-Cyclohexanedicarboxylic acid esters are more stable than IPA to hydrolysis under basic conditions
[10]. Water-reducible polyesters are used in applications for which good storage stability
and hydrolytic stability are not important, such as industrial coatings with a fast
turnover.
An alternative approach that provides somewhat better, but still limited hydrolytic stability is to use 2,2-dimethylolpropionic acid as one of the diol components. The carboxylic
acid group of this monomer is highly hindered by being located on a tertiary carbon. The
resulting differential reactivity makes it possible to esterify the hydroxyl groups while
leaving many of the acid groups unreacted. Although they are too hindered to esterify
readily, the acid groups still form salts readily.
Hydrolytic stability is also affected by the choice of polyol. In addition to the steric
effect discussed earlier, it has been shown that polyols with low water solubility give polyesters that are more stable than those with higher water solubility to hydrolysis under basic
conditions, presumably because the polymers are more hydrophobic [10]. For example,
polyesters made with 2-butyl-2-ethyl-1,3-propanediol have greater stability to hydrolysis
than do NPG esters.
The problem of hydrolysis can be minimized by making powdered solid polyesters.
An example of such a solid polyester is made from purified isophthalic acid, adipic
acid, neopentyl glycol, cyclohexanedimethanol, hydrogenated bisphenol A, and trimellitic anhydride [21]. The resin is powdered and stored until a coating is to be made; then it
is stirred into a hot aqueous solution of dimethylethanolamine to form a dispersion.
Another possible problem with water-reducible polyesters results from intramolecular reaction of terminal hydroxyl and carboxylic acid groups to form some low MW
nonfunctional cyclic polyesters. When coatings are baked, small amounts of cyclic
esters can volatilize out of the coating and gradually accumulate in cool spots in the
oven. Eventually, sufficient resin can accumulate to drip down on products going
through the ovens, marring their finish. Since the amounts are small, dripping may not
start until after weeks or months of operation of the coating line.
Water-thinnable polyester coatings have been formulated with low MW oligomeric
hydroxy-terminated polyesters [16]. Up to about 20% of water dissolves in a polyester –
Class I MF resin binder, reducing the viscosity to about half. This permits making solvent
free coatings. Further work is needed, but the idea is appealing not only because of the low
VOC from the lack of need of cosolvent, but also because no amine is needed. Absence of
amine should reduce the problem of hydrolysis.
REFERENCES
211
10.6. POLYESTER RESINS FOR POWDER COATINGS
Polyester resins for powder coating are brittle solids with a relatively high Tg (50 to 608C),
so the powder coating does not sinter (partially fuse) during storage. These requirements
are met with terephthalic acid (TPA) and NPG as the principal monomers. Resins in which
much of the TPA is replaced with isophthalic acid (IPA) are said to give coatings with
superior outdoor weatherability [22]. Smaller amounts of other monomers are added to
increase f̄n and to reduce Tg to the desired level. The relatively high Tg makes it possible
to prepare hard, tough films with relatively low cross-link density. Widely used commercial products are amorphous, not crystalline. Both hydroxy- and carboxy-terminated polyesters are used. The former are most commonly cross-linked with blocked isocyanates
(Section 12.5) and the latter with epoxy resins (Chapter 13). Other cross-linkers include
2-hydroxyalkylamides (Section 17.5) and tetramethoxymethylglycoluril (Section
11.4.3). (See Chapter 28 for a discussion of powder coatings.)
1,4-Cyclohexanedicarboxylic acid (CHDA) has been suggested as a replacement for
isophthalic acid [23]. CHDA polyesters have a lower Tg and lower melt viscosities; if
the Tg is too low, NPG can be fully or partially replaced with hydrogenated bisphenol
A to provide for storage stability. Processes involving direct esterification of TPA are
used, but since TPA has a very high melting point, it is common to make the hydroxyterminated polyester by transesterification of dimethyl terephthalate. An appropriate transesterification catalyst is tetraisopropyl titanate. If desired, the hydroxy-terminated TPA
polyester can then be reacted with other polyacids to form a carboxylic acid– terminated
product in a second stage.
Recycled polyethylene terephthalate (PET) has been evaluated as a raw material for
making polyesters for powder coatings. Recycled PET was dissolved in a mixture of
NPG and TMP with a catalyst at 2008C. After cooling to 1808C, IPA was added and
the reaction mixture heated to 2408C until the required acid number was reached. Properties of powder coated films were identical to those obtained with a polyester prepared from
IPA, ethylene glycol, NPG, and TMP [24].
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Odian, G. W., Principles of Polymerization, 4th ed., Wiley, New York, 2004, pp. 40– 62.
Misev, T. A., J. Coat. Technol., 1989, 61(772), 49.
Van Sickle, D. E.; et al., Polym. Mater. Sci. Eng., 1997, 76, 288.
Turpin, E. T., J. Paint Technol., 1975, 47(602), 40.
Golob, D. J.; et al., Polym. Mater. Sci. Eng., 1990, 63, 826.
Chu, G.; Jones, F. N., J. Coat. Technol., 1993, 65(819), 43.
Argyropoulos, J.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2003, pp. 107 – 113.
Bender, M. L.; et al., J. Am. Chem. Soc., 1958, 80, 5384.
Cambon, A.; Jullien, R., Bull. Soc. Chim. Fr., 1973, 2003.
Jones, T. E.; McCarthy, J. M., J. Coat. Technol., 1995, 67(844), 57.
Heidt, P. C.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2000,
pp. 295 – 307.
Belote, S. N.; Blount, W. W., J. Coat. Technol., 1981, 53(681), 33.
212
POLYESTER RESINS
13. Hood, J. D.; et al., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA, 1986, p. 14.
14. Calbo, L. J., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA, 1986, p. 356.
15. TONE Polyols, tech. bull., union Carbide Corp., Specialty Polymers and Composites,
Charleston, WV, 1986.
16. Jones, F. N., J. Coat. Technol., 1996, 68(852), 25.
17. Willemse, R. X. E.; et al., Macromolecules, 2005, 38, 6877.
18. Arnould, M. A.; et al., Prog. Org. Coat., 2002, 45, 305.
19. Green, M. L., J. Coat. Technol., 2001, 73(918), 55.
20. Tech. Bull. TMA109e, Amoco Chemicals Corp., 1984.
21. R. Engelhardt, Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA, 1996, p. 408.
22. Merck, Y., Surf. Coat. Int. B Coat. Trans., 2001, 84(B3), 231.
23. Johnson, L. K.; Sade, W. T., J. Coat. Technol., 1993, 65(826), 19.
24. Kawamura, C.; et al., Prog. Org. Coat., 2002, 45, 185.
11
Amino Resins
Amino resins, also called aminoplast resins, are major cross-linking agents for baked
thermosetting coatings; the amino resins most commonly used are derived from melamine:
that is, 2,4,6-triamino-1,3,5-triazine. Urea, benzoguanamine, glycoluril, and copolymers
of (meth)acrylamide are also used.
Amino resins for coatings are made by reacting one of these compounds with formaldehyde (H2C55O) and subsequently, with alcohols (ROH) to yield activated ethers with
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
213
214
AMINO RESINS
the general structure .NCH2OR. Ethers of amino resins are activated toward nucleophilic
substitution by the neighboring N and are much more reactive than aliphatic ethers. When
the nucleophile is the alcohol of a polyol (POH), transetherification can occur, as shown in
Eq. 11.1, resulting in formation of a cross-linked polymer. The reaction is catalyzed by
acid. Carboxylic acids, urethanes, and phenols with an unsubstituted ortho position also
react, as shown in Eqs. 11.2, 11.3, and 11.4.
ð11:1Þ
ð11:2Þ
ð11:3Þ
ð11:4Þ
11.1. SYNTHESIS OF MELAMINE – FORMALDEHYDE RESINS
The first step in the synthesis of melamine – formaldehyde (MF) resins is methylolation,
the reaction of melamine with formaldehyde under basic conditions. With excess formaldehyde, the reaction can be driven to form a mixture that contains substantial amounts of
hexamethylolmelamine (1, where R is H). With less than the stoichiometric 6 mol of formaldehyde per mole of melamine, a mixture of partially methylolated derivatives, including species such as symmetrical trimethylolmelamine (2, where R is H) is formed.
Methylolmelamine resins of this type are often used to fabricate laminates and plastics
but are rarely used in coatings.
For coatings, the second step is acid-catalyzed etherification of methylolmelamines
with an alcohol, such as methyl or butyl alcohol. Complete etherification of methylolmelamines 1 and 2, where R is H, yields alkoxymethyl derivatives 1 and 2, where R is alkyl.
Many commercial MF resins are only partially etherified. In addition to monomeric
species, commercial MF resins contain oligomeric species in which triazine rings are
linked by methylene (.NCH2N,) and dimethylene ether bridges (.NCH2OCH2N,).
11.1. SYNTHESIS OF MELAMINE –FORMALDEHYDE RESINS
215
11.1.1. Methylolation Reaction
A probable mechanism for base catalyzed reaction of melamine with formaldehyde is outlined in Scheme 11.1, where 22NH2 represents the melamine amino groups. The first step
involves nucleophilic attack by the amino group on the electrophilic C of formaldehyde,
which is facilitated by removal of a proton from N by the base (B2 ). This is followed by
proton transfer from the resulting B22H to the negative oxygen atom, which yields the
methylolated product and regenerates the base catalyst; both steps are reversible.
Early studies of the kinetics of the reaction indicate that the presence of one methylol
group on N deactivates the group for a second reaction by a factor of 0.5 to 0.6. On the
other hand, substitution on one amino group has less effect on the reactivity of the
other amino groups. These kinetic factors favor formation of the symmetrical trimethylolmelamine (TMMM) (2), where R is H. However, the preference is not strong enough to
overcome the kinetic and thermodynamic tendencies to produce mixtures of products.
Thus, at equilibrium, the reaction of 6 mol of formaldehyde with 1 mol of melamine
yields a mixture of products, including all levels of methylolation and free formaldehyde
[1]. A side reaction is methylolation on oxygen, as shown in Scheme 11.1.
Scheme 11.1
11.1.2. Etherification Reaction
Following methylolation, the base catalyst is neutralized with excess acid and the appropriate
alcohol is added. The acid-catalyzed reversible reaction leads to formation of alkoxymethyl
groups. Nitric acid is commonly used as a catalyst because nitrate salt by-products are
relatively easily removed.
There is disagreement in the literature about the mechanism of the etherification reaction as well as of the closely related transetherification reaction, shown in Eq. 11.1. The
contending mechanisms for these substitution reactions are SN1 and SN2, both of which
are outlined in Scheme 11.2, where R is H and R0 OH is methyl or butyl alcohol. Note
that the conjugate base (A2) is omitted from the intermediate steps.
The distinguishing feature is whether the protonated methylol group dissociates to an
intermediate, resonance-stabilized carbocation, as in Eq. 11.5, before reacting with the
alcohol, characteristic of the SN1 mechanism, or reacts directly with the alcohol, as in
216
AMINO RESINS
Scheme 11.2
Eq. 11.6, characteristic of the SN2 mechanism. The controversy about the mechanism in MF
resin synthesis also applies to the transetherification reaction involved in cross-linking MF
resins with polyols, where R is methyl or butyl and R0 OH is the polyol. Experimental evidence bearing on this controversy has been acquired primarily in studies of coatings and is
discussed in Section 11.3.2. Based on the available evidence, we favor the SN1 mechanism.
Etherification of methylol groups on a singly substituted N (i.e., a N that also bears a H
atom) is proposed to follow the mechanism outlined in Scheme 11.3, where R is H and
R0 OH is methyl or butyl alcohol.
The distinguishing feature is formation of the uncharged imine intermediate, shown in
Eq. 11.7, which is possible due to the presence of the N22H group. Imines form by elimination of water catalyzed by relatively weak acids, such as carboxylic acids. Complexation of the methylolmelamine by an acid (A22H) is sufficient to yield the imine, either
in a concerted push-pull mechanism, as shown in Scheme 11.3, or a step-wise mechanism.
11.1. SYNTHESIS OF MELAMINE –FORMALDEHYDE RESINS
217
Scheme 11.3
Complexation of the reactive imine by acid is also sufficient to effect subsequent addition
of alcohol, resulting in overall substitution of water by alcohol. This reaction is also
reversible. Since catalysis in Scheme 11.3 involves complexation rather than protonation
by the acid, the rate is dependent on the nature of H22A as well as its concentration.
Indeed, both the more positive H end and the more negative A end of H22A participate
by facilitating breaking of both the C22O and the N22H bonds, respectively. By convention, this is called general acid catalysis. The mechanism in Scheme 11.3 also applies to
cross-linking of polyols by MF resins with N22H groups by transetherification, where R is
methyl or butyl and R0 OH represents the polyol.
In contrast, elimination of water from methylolmelamines lacking N22H groups, either by
the SN1 or SN2 mechanism (shown in Scheme 11.2), requires protonation of the methylol
group, necessitating a strong ionizing acid such as nitric, sulfuric, or a sulfonic acid. Such catalysis, requiring protonation, is called specific acid catalysis, signifying that the rate is dependent only on the concentration of protons (Hþ) and is independent of the conjugate base (A2).
11.1.3. Self-Condensation Reactions
Self-condensation refers to reactions that lead to the formation of bridges between triazine rings, resulting in dimers, trimers, higher oligomers, and ultimately, to cross-linked
218
AMINO RESINS
polymer. The extent of such reactions depends on process factors, including pH, ratio of
reactants, reaction temperature, rate of removal of water, and probably others. Two types
of linkages occur between triazines: methylene bridges (.NCH2N,) and dimethylene
ether bridges (.NCH2OCH2N,) [2,3]. Acid catalysis is said to favor dimethylene ether
bridges and base catalysis to favor methylene bridges; furthermore, formation of methylene
bridges may occur by base-catalyzed scission of dimethylene ether bridges [4].
Formation of methylene bridges has been demonstrated in the acid-catalyzed reaction
of a model compound N,N-dimethoxymethyl-N0 ,N0 ,N00 ,N00 -tetramethylmelamine (3) and
water [5]. In the presence of p-toluenesulfonic acid (pTSA) and water at elevated temperatures, a cyclic trimer (4) with methylene groups connecting the three nitrogens is formed.
Formation of the trimer is reversible [6]; when trimer 4 is heated with methyl alcohol and
formaldehyde at 1308C, compound 3 is regenerated.
The monomethoxymethyl model compound (5) is more reactive than 3 [7]. It rapidly
forms the cyclic trimer (4) in the absence of water at 1008C. In solution with an acid
catalyst, it equilibrates rapidly at 258C with the methylene ether bridge dimer (6).
11.2. TYPES OF MF RESINS
A variety of MF resins are made commercially, which differ by the ratio of functional
groups, the alcohol used for etherification, and the average degree of polymerization P̄.
To facilitate discussion, MF resins are classified into two broad groups. Class I resins
are made with relatively high ratios of formaldehyde to melamine, and hence most of
the nitrogens have two alkoxymethyl substituents. Class II resins are made with lower
ratios of formaldehyde, and many of the nitrogens have only one substituent.
From about 1940 through the 1950s, the predominant melamine resins were Class II
types used for cross-linking alkyds. These resins have enough bridging so that P̄ is 3 or
more, and the alcohol is usually n-butyl or isobutyl alcohol. Such resins are economical,
they are readily miscible in alkyd formulations, and they give a wide latitude; that is, coatings do not require exacting control of formulation to produce acceptable application
characteristics and film properties.
11.2. TYPES OF MF RESINS
219
Class I MF resins were commercialized in the 1950s. Methylated Class I resins are
more compatible than butylated resins with certain coreactant resins, and in some cases,
provide tougher films. Introduction of waterborne and high-solids coatings during the
1970s accelerated the shift toward Class I resins. Methylated MF resins are more miscible
with the water, solvent, and resin blend used in waterborne coatings; and Class I resins
with low P̄ values give lower viscosity to high-solids coatings. These generalities are
not universal; some Class II resins are used in low-solvent coatings.
A commercial methylated MF Class I resin has been reported to contain 62% monomers
(i.e., one triazine ring), 23% dimers, together with 15% trimers and higher oligomers [8].
The equivalent weight of pure hexamethoxymethylmelamine (HMMM) is 65 g equiv.21,
but the equivalent weight of this resin is 80 g equiv.21, resulting from the presence of
dimers, trimers, and higher oligomers. (In estimating equivalent weight, it is necessary to
make an assumption as to whether .NCH2N, or .NCH2OCH2N, bridging groups
react to form cross-links. Here, it is assumed that they do not.) Chromatographic analysis
of a different, but similar resin of this type is shown in Figure 11.1. The chromatograms
show at least 30 different chemical species present in significant amounts [2,9]. While
the predominant groups are 22N(CH2OCH3)2, various species containing incompletely
etherified groups [22N(CH2OCH3)(CH2OH)] and incompletely formylated groups
[22NH(CH2OCH3)] are present. These methylated MF resins are frequently called
HMMM resins, but the terminology high HMMM resin is more appropriate.
Figure 11.1. High-performance liquid chromatograms (HPLC) of a typical Class I high-HMMM resin:
(a) SEC chromatogram; (b) gradient HPLC chromatogram. (From Ref. [9], with permission.)
220
AMINO RESINS
A broad range of Class I resins is available with different degrees of polymerization P̄ and
with different extents of formylation and etherification. In addition to methylated derivatives, butylated as well as mixed methyl/n-butyl, methyl/isobutyl, methyl/ethyl, and
methyl/isooctyl derivatives are also available. Class I resins generally have lower P̄ and
therefore lower viscosity than Class II resins at the same percent solids content. The viscosity is also reduced by the presence of fewer polar groups. At a given P̄, butyl ethers
exhibit even lower viscosities, attributable to lower Tg values, than the corresponding
methyl ethers.
An even broader range of Class II resins is available because the ratio of formaldehyde/
melamine and of alcohol/formaldehyde can be varied through a wider range. It is harder to
suppress bridge-forming reactions during synthesis of Class II resins, and very low P̄
resins cannot be prepared. The predominant reactive group present in Class II resins is
22NHCH2OR. Hence, they are frequently called high NH resins. They also contain
22NH2, .NCH2OH, as well as 22N(CH2OR)2, groups. To minimize viscosity and
maximize reactivity, resin producers strive to reduce P̄ and maximize the amount of symmetrical TMMM (3), where R is Me. This is possible to a degree because the methylolation of 22NH2 groups, yielding 22NHCH2OH, is kinetically favored over methylolation
of NHCH2OH, yielding 22N(CH2OH)2. Owing to advances in HPLC analysis techniques,
it is possible to follow the effects of small changes in process conditions on composition.
Continuing refinement of processes has permitted commercial production of resins with
as high as 50% TMMM. A variety of alcohols are used to make different resin grades.
Amino resins usually contain small amounts of unreacted formaldehyde that may be
vaporized during application and baking of the coatings. Additional amounts of formaldehyde vapor may be formed during storage of formulated coatings and/or during the
baking process. These small amounts of formaldehyde vapor pose a potential hazard to
workers who are exposed daily. Formaldehyde exposure levels are strictly regulated in
the United States; and current technology appears adequate to satisfy these regulations.
Although the regulations already appear stringent, it is possible that allowable exposure
levels will be reduced further. Intensive research on ways to reduce formaldehyde emissions has been under way for some time. Besides improving ventilation, approaches
include changing resins, changing formulations, incinerating oven exhausts, and introducing formaldehyde scavengers. MF resin producers offer low-formaldehyde (,0.1 wt%)
grades of key resins.
11.3. MF – POLYOL REACTIONS IN COATINGS
Melamine – formaldehyde (MF) resins are used to cross-link coreactant resins having
hydroxyl, carboxylic acid, urethane (carbamate), and/or amide groups. Acrylic
(Chapter 8), polyester (Chapter 10), alkyd (Chapter 15), epoxy (Chapter 13), and polyurethane (Chapter 12) resins are the most important classes of coreactant resins. It is widespread practice to use combinations of cross-linkers (e.g., MF resins with other types of
amino resins) with (blocked) polyisocyanates (Chapter 12), with phenolic resins
(Section 13.6), or with resins bearing reactive silanes (Section 16.2).
Polyols (hydroxy-functional acrylics, polyesters, etc.) are the resins most commonly
cross-linked by MF resins. The hydroxyl groups of polyols react by either transetherification with the activated alkoxymethyl groups or by etherification of methylol groups
of MF resins to form new ether cross-links. The reaction with phenols, as in Eq. 11.4,
11.3. MF– POLYOL REACTIONS IN COATINGS
221
has the advantage that a C22C bond is formed so that the product is stable to hydrolysis.
Strong acid catalysts, such as sulfonic acids, are used for Class I resins, and weak acid catalysts, such as carboxylic acids, are used for Class II resins. The reactions are reversible
but are driven toward cross-linking by volatilization of the monofunctional alcohol or
water produced, as shown in Eq. 11.8, where R is alkyl or H and POH is the polyol.
ð11:8Þ
Rates of reaction with hydroxyl groups depend on the structure of the polyol and the
MF resin, the type and amount of catalyst, and the temperature. The rate at which MF
resins cross-link polyols increases with increasing volatility of the alcohol, according to
the following order: methyl . ethyl . n-butyl. These rates are probably influenced by
the rate of diffusion of the alcohol from the reactive site and by evaporation from the film.
A widely believed but mistaken conclusion is that steric hindrance limits the number of
active ether groups on high HMMM resins that can participate in cross-linking of films to
three, or at most four, of the six groups [1]. This conclusion was apparently based on the
need to use an excess over the stoichiometric amount of MF Class I resin relative to polyol
to achieve the desired film properties at baking schedules such as 30 minutes at 1208C,
using around 0.3 wt% (on the total weight of polymers) of pTSA. However, Hill and
Kozlowski provide strong evidence in support of an essentially complete reaction of
22NCH2OCH3 groups with certain polyols, showing that all 22NCH2OCH3 groups of
high HMMM resins can co-condense if there are enough hydroxyl groups present for
them to react with [8]. Their conclusions were based on studies of changes in mechanical
properties, as discussed in Section 4.2. These results further substantiate the hypothesis that
the excess MF resin used in most formulations is needed not to complete co-condensation
cross-linking, but to permit some degree of self-condensation cross-linking that may be
necessary to achieve the desired film properties.
Factors in addition to cure response must be considered when selecting the alcohol. For
example, viscosity at a given percent solids content can be reduced by substituting Class I
mixed ether resins for HMMM types. The alcohols in mixed ether resins are frequently
1 : 6 to 1 : 3 n-butyl or isobutyl alcohol to methyl alcohol. Only modest reductions of
VOC are attained because mixed ether resins contain a higher weight fraction of volatile
by-products; butyl groups represent a higher weight fraction of resins than do methyl
groups. Perhaps a greater advantage of mixed ether resins is that they impart lower
surface tension to coatings than do high HMMM types. High surface tension is often
associated with development of film defects such as crawling, cratering, leveling
(Chapter 24), and poor intercoat adhesion (Section 6.5) [10].
The rates of development of solvent resistance and film hardness when a fractionated
low NH Class I resin was used to cross-link polyester resins made with cyclohexanedimethanol (CHDM), neopentyl glycol (NPG), and 1,6-hexanediol have been reported
[7,11]. Film properties developed most rapidly with CHDM polyesters, followed
in order by NPG polyesters a close second, and hexanediol polyesters. It was also
found that film properties generally developed more rapidly with acrylic polyols than
with polyester polyols.
During acid-catalyzed cross-linking with polyols, self-condensation reactions of MF
resins also occur, including formation of methylene bridges and dimethylene ether
bridges. Self-condensation reactions result in cross-linking, as do reactions of MF resins
222
AMINO RESINS
with polyol (or another coreactant), which have been termed co-condensation reactions
[12]. Both self- and co-condensation reactions contribute to the structure of the crosslinked polymer network and to its film properties.
With strong acid catalysis, the apparent rate at which Class I resins react with most
polyols by co-condensation is faster than self-condensation. However, with Class II
resins, the apparent rates of co-condensation and self-condensation are similar. The relatively high levels of 22NHCH2OR and .NCH2OH groups in Class II resins promote selfcondensation reactions. The methylol groups can self-react analogously to reaction with a
polyol; and the 22NHCH2OR groups can eliminate alcohol to yield reactive melamine
imine (22N55CH2) groups, which can form dimethylene ether bridges by reaction with
22NCH2OH groups and methylene bridges by reaction with .NH groups. Based on the
time and temperature required to achieve hard films, Class I and II resins have also
been classified as high and low cure temperature resins, respectively, which reflects, at
least in part, the facility of self-condensation reactions by Class II resins.
Class I resins are used with an excess of alkoxymethyl groups over polyol hydroxyl
groups, so final development of film properties depends on the extent of self-condensation
reactions as well as on co-condensation. In some cases, this high stoichiometric ratio is
dictated by the lower cost of MF resins compared to that of some polyols. It may also
be that self-condensation cross-links give some advantage in film properties. The
amounts of MF resin and catalyst are determined empirically and are optimized for a
specified baking time and temperature. Formulations are designed to achieve optimum
properties when co-condensation is nearly complete and self-condensation is partly complete. If the use conditions differ from those for which the formulation was designed, the
extent of self-condensation and co-condensation will differ from the optimum levels.
Coating properties such as hardness, adhesion, exterior durability, and impact resistance
may be affected.
In high-solids coatings in which the hydroxy equivalent weight and the average functionality of the polyol are lower than those for low solids, higher molecular-weight
polyols, the results are particularly sensitive to variations in cure temperature and time.
These considerations have been addressed in terms of a cure window, corresponding to
the range of cure times and temperatures that provides films of acceptable properties
[13]. The cure window was shown to be smaller for high-solids MF –polyol coatings,
especially when Class II resins were used.
11.3.1. Catalysis of MF – Polyol Reactions
With strong acid catalysts, commonly aryl sulfonic acids, usually in the range 0.5 to
1wt% of MF resin, reactions of MF Class I resins with polyols give cured films in
10 to 30 minutes at 110 to 1308C. Co-condensation of MF Class I resins with
polyols can be catalyzed by weak acids, such as carboxylic acids; however, elevated
cure temperatures, generally greater than 1408C, are required [14]. In the case of
Class II resins, carboxylic acids are more effective in catalyzing co-condensation in
accordance with the mechanism provided in Scheme 11.3. Since many polyol resins
contain some carboxylic acid groups, which are present to promote adhesion and facilitate pigment dispersion, no added catalyst may be needed for cross-linking with Class
II resins. At the high temperatures (air temperatures as high as 3758C) used in coil
coating ovens, strong acid catalysts are needed even with Class II resins, owing to
the short cure times.
11.3. MF– POLYOL REACTIONS IN COATINGS
223
Cure time and/or temperature may be reduced by increasing the catalyst concentration.
However, storage (or package) stability is reduced by this approach, since the reaction at
ambient temperature is also catalyzed by acid. Another danger in increasing acid catalyst
concentration to reduce cure time and/or temperature relates to the durability of the
cured coating. In addition to catalyzing transetherification (and etherification) reactions
of MF resins with polyols, acid residues catalyze the hydrolysis of cross-links in cured coatings. The course of the hydrolysis reaction can be followed by reference to Schemes 11.2
and 11.3, where .N22CH222OR represents a cross-linked polymer and R0 OH represents
water. The hydrolysis reaction breaks cross-linked bonds and generates methylol groups
that, at least in part, eliminate formaldehyde as follows [12]:
ð11:9Þ
Free acids such as p-toluenesulfonic acid (pTSA) catalyze cross-linking under ambient
conditions at a rate that may increase the viscosity of a liquid coating above the range for
application in less than 6 months. One-package systems commonly use acid catalysts that
are deactivated: variously called acid precursors, latent acids, and blocked acids. Aryl sulfonic acids are strong acids (the pKa of pTSA ¼ 26); their tertiary amine salts are weak
acids (pKa ¼ 8 to 10) and do not catalyze the reactions of Class I MF resins. As shown in
Eq. 11.10, there is an equilibrium between pTSA amine salt and the protonated MF resin
that is the active species for catalysis.
ð11:10Þ
Although the equilibrium favors structures on the left, it may be shifted to the right by
volatilization of the amine, increasing the concentration of protonated MF resin, the initial
intermediate in the transetherification reaction (Scheme 11.2). Generally, the storage stability of a coating with a blocked catalyst approaches the stability of an uncatalyzed coating.
In some cases the cure rate approaches that of a coating catalyzed with free sulfonic acid,
whereas in others, the curing rate is somewhat reduced [15]. The balance of the storage
stability and cure rate is reported to be particularly favorable for N-benzylN, N-dimethylanilinium sulfonates [16]. In this case, the MF resin is probably activated
by O-alkylation with the benzyl cation, since quaternary anilinium sulfonates have no
available proton.
The acid strength of the medium can be no stronger than that of protonated melamine
resin, which exerts a leveling effect on acid strength, just as the strongest acid in water is
the hydronium ion. Apparently, pTSA (TsOH) is completely ionized by MF resins, which
explains why even stronger acids, such as hexafluorophosphoric acid, exhibit similar catalytic activity. On the other hand, pTSA is more effective for Class I MF resins than
weaker acids, such as butylphosphoric and carboxylic acids. With pTSA, the rate of
co-condensation of an HMMM resin and acrylic polyol is reported to be proportional to
the square root of the acid concentration [17].
224
AMINO RESINS
Although the acid strength of a strong acid catalyst makes little difference in cure rates, the
choice of acid (or blocked acid) catalyst can lead to differences in film properties of cured
coatings. The most widely used catalyst is pTSA. The grade of pTSA is important, since
some grades contain significant quantities of sulfuric acid, which can lead to pronounced yellowing of films during baking. It has been found that water resistance of films can be improved
by use of a more hydrophobic sulfonic acid, such as dinonylnaphthalene disulfonic acid
(DNNDSA), which is reported to produce films that are less likely to blister when exposed
to high humidity conditions [18]. DNNDSA is particularly effective in coatings applied
directly to metal, which, with pTSA are prone to blistering on humidity exposure.
p-Dodecylbenzenesulfonic acid (DDBSA) provides similar catalytic activity and film
properties to those of DNNDSA when used in coatings applied over a prime coat.
However, when a DDBSA catalyzed coating is used directly on steel, adhesion of the
coating is likely to be poor. It seems probable that the sulfonic acid group on DDBSA
is strongly adsorbed on the steel surface, causing the surface to become covered with
dodecyl groups. The low surface tension of the long hydrocarbon dodecyl groups on
DDBSA may cause dewetting by the rest of the coating or formation of a weak boundary
layer that reduces adhesion. The better adhesion with DNNDSA catalyzed coatings may
be related to the presence of two sulfonic acid groups or weaker surfactant properties.
The activity of acid (or blocked acid) catalysts can be affected by pigmentation of the
coating. Some grades of titanium dioxide pigments lead to loss of catalyst activity with
storage time. The loss is related to the composition of surface treatments on the TiO2.
Silicon dioxide – treated TiO2 is preferable to aluminum oxide – treated TiO2 in this
respect, probably owing to the basicity of alumina, which tends to neutralize the acid.
When alumina-treated TiO2 is required for other reasons, a higher concentration of acid
catalyst may be necessary to counteract the effect.
In waterborne MF– polyol coatings, the amine utilized to neutralize the solubilizing
carboxylic acid groups in the polyol (see Section 8.3 for a discussion) also neutralizes
the sulfonic acid. It was shown that commonly utilized hydroxy-functional amines such
as N,N-dimethylethanolamine (DMEA) may also participate in transetherification of MF
resins and/or transesterification of ester groups in the polyol during storage or initial
stages of curing, which may retard or prevent their volatilization during cure, causing
loss of cure response [19]. It was also shown that 2-amino-2-methylpropanol (AMP)
gives a faster cure rate since the primary amine can react with an acrylic resin to form
amides, reducing the basicity. Formation of an amide, together with transesterification
by the hydroxyl group of AMP, forms cross-links. Furthermore, AMP has been shown
to react with formaldehyde to form an oxazolidine, also reducing basicity so that it
does not inhibit cure as DMAE does [20]. Dynamic mechanical analysis (DMA) also
shows the lesser extent of inhibition by AMP and supports auxiliary cross-linking by
AMP [21].
11.3.2. Kinetics and Mechanism of MF –Polyol Co-condensation
Many studies have been directed at elucidating the mechanism(s) of reactions between MF
resins and polyols. Until recently, most of these studies were based on following the time
necessary to reach some degree of film hardness. However, both co- and self-condensation
reactions contribute to hardness [8]. Therefore, one must view with caution conclusions
from such studies applied specifically to the co-condensation reaction between hydroxyl
groups and MF resins.
11.3. MF– POLYOL REACTIONS IN COATINGS
225
It is desirable to develop methods for measuring the extent of co-condensation directly.
Infrared (IR) and Fourier transform infrared (FTIR) spectroscopy have been used to follow
changes in functional group concentrations using Class I MF resins as a function of time
and temperature [17]; particular care must be taken in interpreting results because of band
overlap in the hydroxy region. The rate of evolution of volatile reaction products, including formaldehyde and methyl alcohol, has been followed by gas chromatography [12,14]
and by thermogravimetric analysis [15]. Major contributions have been made by application of DMA to MF –polyol cross-linking coatings [8,22] (Section 4.2). Oscillating
plate rheometry has also been used [23].
A misconception resulting from using hardness results to study kinetics is that the
co-condensation reaction of Class I resins was thought to be much slower than it has
now been found to be. Disappearance of polyol hydroxyl groups, as followed by FTIR,
showed that the reaction is more rapid than the development of film properties
[14,23,24]. Relatively slow development of film properties is characteristic of step-growth
polymerization processes and may also reflect the slower rate of self-condensation reactions, which probably contribute to achievement of optimal cross-link density.
In Class I resins, it has been observed that reactivity is reduced by the presence of imino
(.NH) groups in the resin. Fractionation to remove essentially all of these groups from a
commercial HMMM resin yielded a surprising increase in reactivity with hydroxyl functional resins, enabling lower cure temperatures [7,25]. It is speculated that the relatively
basic 22NHCH2OCH3 groups present in most HMMM resins tie up a portion of the
catalyst.
The mechanism of co-condensation of MF Class I resins with polyols remains controversial, with experimental evidence being interpreted in support of (or against) the SN1 or
SN2 mechanisms, provided in Scheme 11.2. However, the two mechanisms are not
mutually exclusive and may occur simultaneously. Owing to the expectation that the activation parameters, Ea and A, are larger for the SN1 reaction (Section 2.3.2 and Figure 2.6),
the SN1 mechanism is favored with increasing temperature.
Based on studies with the model analog (4) and monofunctional alcohols, it was proposed that both SN1 and SN2 mechanisms occur [5]. The novel use of chiral monofunctional alcohols provided kinetic evidence for equal rates of reactivity of primary and
secondary alcohols with a high HMMM resin, but higher reactivity in the reverse reaction
with secondary alcohols [26]. These findings support the SN1 mechanism and are consistent with the general order of reactivity in coatings of primary . secondary alcohols, since
the curing reactions occur under reversible conditions. However, utilization of an oversimplified rate expression may compromise these conclusions.
Using an HMMM – acrylic composition and following the co-condensation reaction by
FTIR, Bauer provided support for the SN1 mechanism. He also interpreted the role of
methyl alcohol in the kinetic order of the reaction and the rate dependence on polyol in
support of the SN1 mechanism [17]. A key point is that the rate of co-condensation by
the SN1 mechanism is expected to depend on the concentration of polyol if the methyl
alcohol, formed from HMMM, competes with the polyol for the intermediate carbocation;
refer to Scheme 11.2. Water also competes with the polyol for this carbocation. This valid
interpretation of the SN1 mechanism is an important consideration, since dependence of
the rate on the polyol had mistakenly been interpreted as evidence against the SN1 mechanism. Competition for the intermediate carbocation by methyl alcohol is expected to gain
relative importance as the reaction proceeds because the methyl alcohol concentration
increases, at least temporarily, whereas the polyol concentration decreases with increasing
226
AMINO RESINS
conversion. Thus, kinetic studies based on hardness or other film properties, which tend to
develop only at high conversions, are expected to exhibit strong dependence on the concentration and nature of the polyol. Such results had generally been interpreted in favor of
the SN2 mechanism, but are equally consistent with the SN1 mechanism.
A further complication is the strong possibility that cross-linking with MF resins leads
to a gradient of compositions and properties within films. Although relatively few films
have been studied in this way, most studies detected a gradient. For example, Haacke
et al. used a microtome to separate layers in acrylic – MF clear coats and found that different layers had different compositions [27]. Furthermore, they found that Tg was as much as
158C higher at the surface than it was deep in the film and that cross-link density was
almost twice as high near the surface. They attributed the gradient to the differing
effects of escaping by-product alcohol on the cross-linking equilibrium reactions at different levels in the film. There is the additional possibility that catalyst concentrations may
not be uniform within the film, especially when catalysts blocked with volatile amines are
used. Other studies have suggested that there may be a very thin layer of material at the
surface that is quite different than the bulk in composition and properties [28]. These scattered reports suggest that to fully understand the performance of amino resin cross-linked
coatings (and perhaps of other types as well), it may be necessary to appreciate the effects
of composition and property gradients within the films.
Considerable progress has been made toward understanding the complex combinations
of reactions that occur during the curing of MF cross-linked coatings, but the process is not
yet fully understood; further research is needed.
11.3.3. Package Stability Considerations
Package stability of coatings containing MF resins is affected by other factors in addition
to catalyst levels, discussed in Section 11.3.1. Class II resins generally give poorer
package stability than Class I resins due to the preponderance of .NH and
.NCH2OH groups, which participate in slow, weak acid catalyzed reactions during
storage at room temperature. Stability is somewhat improved by addition of small quantities of tertiary amines. Primary or secondary amines, which react with formaldehyde,
should not be used with Class II resins. Free formaldehyde is in equilibrium with melamine methylol groups in these resins. Reaction of the formaldehyde with primary or
secondary amines displaces the equilibrium reaction in favor of formaldehyde
formation (i.e., demethylolation), which reduces the methylol group functionality of
the MF resins.
An important approach for increasing package stability is to use as much monofunctional alcohol as possible in formulations containing either class of MF resins. The
presence of monofunctional alcohol in the formulation extends the storage stability
since its reaction with the MF resin does not lead to cross-linking. On the other hand,
cross-linking and viscosity buildup occur when the MF resin self-condenses or reacts
with the polyol. Generally, it is desirable to utilize the same alcohol that is used to synthesize the MF resin. If a different alcohol is used, undesirable changes may occur. For
example, if n-butyl alcohol is used in the solvent with a methoxymethylmelamine resin,
the cure response gradually becomes slower as the proportion of butyl ether increases.
The viscosity of a coating may decrease during storage because butoxymethylmelamine
is less viscous (it has a lower Tg) than the corresponding methoxymethylmelamine.
Viscosity decrease may also result from breaking dimethylene ether bridges by excess
alcohol.
11.4. OTHER AMINO RESINS
227
11.3.4. MF Resin Reactions with Carboxylic Acids, Urethanes,
Carbamates, and Malonate-Blocked Isocyanates
Carboxylic acid– functional resins react with MF resins to form the corresponding ester
derivatives (Eq. 11.2); the reaction is slower than with hydroxyl groups. Carboxylic
acid groups are present in substantial levels in waterborne hydroxy-functional resins to
enhance dispersibility, as well as in some solventborne polyols, at low levels, to
enhance the stability of pigment dispersions. When significant levels of carboxylic acid
groups are present, it is desirable to cure under conditions which ensure that they react,
since residual carboxylic acid groups increase the water sensitivity of cured films.
As shown in Eq. 11.3, MF resins react with urethane groups. Polyurethanes containing
22OC(55O)NH22 structures (secondary carbamates), and no hydroxyl groups can be
cross-linked with MF resins at temperatures only slightly higher than those used to crosslink polyols [29]. This result indicates that when hydroxy-terminated urethanes are used
with MF resins, both groups are involved in cross-linking. Both groups react when polyurethane polyols are used as reactive diluents [30]. A variety of MF resins can cross-link
secondary carbamates; Class I resins are more reactive than Class II resins [29]. Secondary
carbamates are nearly as reactive as hydroxyl groups, despite their greater steric hindrance. A possible explanation of this surprising result is that the carbamate reaction is
irreversible [31]. Thus, the forward reaction might be substantially slower than the
forward reaction of hydroxyl groups, but the overall rate could be close to equal if the
reverse reaction does not occur with carbamates. The practice of cross-linking polyols
with mixtures of MF and polyisocyanate resins has become increasingly common. Formulators should keep in mind the strong possibility that some of the carbamate formed in the
isocyanate – hydroxyl reaction will react further with MF resin.
Polymers with pendant 22OC(55O)NH2 structures (primary carbamates) are less
hindered than secondary carbamates and would be expected to cross-link more rapidly.
Conclusive evidence for this expectation has not been published, but patent examples
suggest that it is true; polymers with primary carbamates are cured with Class I MF
resins under the same conditions as polymers with primary 22OH groups [32,33].
Primary carbamate-functional resins can be made by reacting isocyanate-terminated
resins with hydroxypropylcarbamate [32] or from acrylic resins [33]. The isocyanatebased resins can be cross-linked with Class I MF resins, but the cross-link density is
too high for some applications. Functionality can be reduced by blocking half of the functional groups of Class I resins with butyl carbamate [32]. Automotive clear coats made
from such combinations combine the usual high mar resistance of MF – polyol cross-linked
coatings with enhanced resistance to environmental etching.
Class I MF resins also react with malonic ester derivatives of isocyanates at rates
similar to their reactions with 22OH groups [34]. NMR study of a model compound confirmed that environmentally resistant carbon– carbon cross-links are irreversibly formed in
the process. As a result, environmental etch resistance of a coating made from a malonic
ester derivative of a triisocyanate and a Class I MF resin is excellent [34]. Malonic ester
derivatives of isocyanates can also function as cross-linkers for hydroxy-functional resins,
as discussed in Section 12.5.
11.4. OTHER AMINO RESINS
Other amino resins are used on a smaller scale. The structures of the other starting
materials are provided at the beginning of this chapter. Although the chemistry of these
228
AMINO RESINS
resins is similar to that of MF resins, there are differences, particularly as a result of differences in basicity of the resins [35].
11.4.1. Urea –Formaldehyde Resins
Substantial amounts of urea – formaldehyde resins are used because of their low cost and
high reactivity. Urea reacts with formaldehyde to form methylol derivatives. The first and
second formaldehyde units add easily, but trimethylol and tetramethylol derivatives of
urea form with more difficulty, if at all. Methylolated ureas can be etherified by reaction
with alcohols, analogously to methylolated melamines. Self-condensation reactions also
occur under both acidic and basic conditions. A variety of etherified urea – formaldehyde
(UF) resins are made with different ratios of formaldehyde to urea and different alcohols
for etherification.
In general, UF resins are the most economical amino resins; they are also the most reactive. With sufficient acid catalyst, coatings formulated with UF resins can cure at ambient
or mildly elevated temperatures. However, such coatings have poor exterior durability,
probably because the cross-links are relatively reactive and have relatively poor resistance
to hydrolysis. UF resins are used in coatings for temperature sensitive substrates such as
wood furniture, paneling, and cabinetry. In such applications, low temperature baking is
essential, and corrosion resistance (related to the hydrolytic stability of the cross-links)
is unimportant. Blends of UF and MF resins can be used to achieve fast cure, low cost,
and acceptable properties [36].
11.4.2. Benzoguanamine – Formaldehyde Resins
Using benzoguanamine as a starting material, a range of resins analogous to MF resins can
be prepared. As with melamine, benzoguanamine may be methoxymethylated and/or
butoxymethylated to various extents. The average functionality is lower because there
are only two 22NH2 groups per molecule. Etherified benzoguanamine – formaldehyde
(BF) resins yield cross-linked urethane films with greater resistance to alkali and to alkaline detergents, such as sodium tripolyphosphate, than that of MF resins. They also impart
excellent toughness. The exterior durability of BF-based coatings is poorer than that of
MF-based coatings. Thus, BF resins are used for applications such as washing machines
and dishwashers, in which resistance to alkaline detergents and toughness are more
important than exterior durability. The reduced exterior durability of BF-based coatings
probably reflects lower photostability arising from the presence of the phenyl group on
the triazine ring of benzoguanamine, although evidence supporting this reasonable
hypothesis does not appear to be available. Aside from the hydrophobicity of the
phenyl groups, a convincing explanation for enhanced alkali resistance is not intuitively
forthcoming.
11.4.3. Glycoluril – Formaldehyde Resins
Glycoluril reacts with formaldehyde to give tetramethylolglycoluril (TMGU) [35]. In the
pH range 4 to 8, aqueous solutions of TMGU show an equilibrium level of about 3.6
methylol groups per glycoluril, with 0.4 parts of free formaldehyde. Above pH 8.5,
demethylolation is more favored. For example, at pH 11 there are approximately equal
parts of methylol groups and free formaldehyde. Below pH 3, self-condensation occurs
REFERENCES
229
to form glycoluril dimers with methylene ether bridges:
ð11:11Þ
where .N22CH222OH represents methylolated glycoluril groups. This behavior is unlike
that of MF or UF resins. Whereas TMGU does not self-condense above pH 4, methylolated melamines and ureas self-condense at pH 7 and higher.
Tetramethylolglycoluril reacts with alcohols in the presence of strong acid catalyst to
form tetraalkoxymethylglycoluril (GF) resins. Tetramethoxymethylglycoluril is a relatively high melting solid and is used as a cross-linker in powder coatings (Section
28.1.3). In solution coatings, (dimethoxymethyldiethoxymethyl)glycoluril and tetrabutoxymethylglycoluril are used, since they are liquids and are handled more readily. The
mixed methyl –ethyl ether is water soluble.
Relative to other amino resins, GF resins produce coatings that exhibit greater flexibility at a similar cross-link density. Hence, GF resins are used in applications such as
coil coatings and can coatings, where flexibility is important. GF resins possess the
further advantage that less formaldehyde is evolved during cure than with MF resins.
Also, the glycoluril cross-linked polyols are more resistant than MF cross-linked
polyols to hydrolysis under acidic conditions and are said to be UV resistant [35]. The
higher cost of GF resins limits their use.
11.4.4. Poly(meth)acrylamide –Formaldehyde Resins
Acrylic copolymers of N-isobutoxymethylacrylamide can be prepared by two routes:
(1) synthesis of N-isobutoxymethylacrylamide monomer (from a step-wise reaction of
acrylamide with formaldehyde and isobutyl alcohol), followed by copolymerization
with acrylate monomers, and (2) copolymerization of acrylamide, followed by step-wise
reaction with formaldehyde and isobutyl alcohol. Analogous reactions can be carried out
with other alcohols and with methacrylamide.
Such alkoxymethyl(meth)acrylamide amino resins are used in coil coatings, where with
proper design they are weather resistant and more flexible than coatings based on MF–
acrylic polyol resins. One can speculate that their flexibility is enhanced by the absence
of densely cross-linked clusters of self-condensed MF resins with their rigid triazine rings.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
Santer, J. O., Prog. Org. Coat., 1984, 12, 309.
Chang, T. T., Prog. Org. Coat., 1996, 45, 211.
Subrayan, R. P.; Jones, F. N., J. Appl. Polym. Sci., 1996, 62, 1237.
Nastke, R.; et al., J. Macromol. Sci., 1986, A23, 579.
Wicks, Z. W., Jr.; Hsia, D. Y. Y., J. Coat. Technol., 1983, 55(702), 29.
Samaraweera, U.; Jones, F. N., J. Coat. Technol., 1992, 64(804), 69.
Jones, F. N.; et al., Prog. Org. Coat., 1994, 24, 189.
Hill, L. W.; Kozlowski, K., J. Coat. Technol., 1987, 51(751), 63.
van Dijk, J. H.; et al., FATIPEC Congress Book, 1980, Vol. II, p. 326.
230
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
AMINO RESINS
Albrecht, N., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA, 1986, p. 200.
Chu, G.; Jones, F. N., J. Coat. Technol., 1993, 65(819), 43.
Blank, W. J., J. Coat. Technol., 1979, 51(656), 61.
Bauer, D. R.; Dickie, R. A., J. Coat. Technol., 1982, 54(685), 57.
Lazzara, M. G., J. Coat. Technol., 1984, 56(710), 19.
Mijs, W. J.; et al., J. Coat. Technol., 1983, 55(697), 45.
Morimoto, T.; Nakano, S., J. Coat. Technol., 1994, 66(833), 75.
Bauer, D. R., Prog. Org. Coat., 1986, 14, 193.
Calbo, L. J., J. Coat. Technol., 1980, 52(660), 75.
Wicks, Z. W., Jr.; Chen, G. F., J. Coat. Technol., 1979, 50(638), 39.
Ferrell, P. E.; et al., J. Coat. Technol., 1995, 67(851), 63.
Hill, L. W.; et al., in Film Formation in Waterborne Coatings, Provder, T. A.; et al., Eds.,
American Chemical Society, Washington, DC, 1996.
Hill, L. W.; Kozlowski, K., Proc. Int. Symp. Coat. Sci. Technol., Athens, Greece, 1986, p. 129.
Yamamoto, T.; et al., J. Coat. Technol., 1988, 60(762), 51.
Nakamichi, T., Prog. Org. Coat., 1986, 14, 23.
Hill, L. W.; Lee, S.-B., J. Coat. Technol., 1999, 71(897), 127.
Meijer, E. W., J. Polym. Sci. A Polym. Chem., 1986, 24, 2199.
Haacke, G.; et al., J. Coat. Technol., 1995, 67(843), 29.
Jones, F. N.; et al., Prog. Org. Coat., 1998, 34, 119.
Higginbottom, H. P.; et al., J. Coat. Technol., 1999, 71(894), 49.
Hill, L. W., Polym. Mater. Sci. Eng., 1997, 77, 387.
Blank, W. J.; et al., Polym. Mater. Sci. Eng., 1997, 77, 391.
Rehfuss, J. W.; Ohrbom, W. H., U.S. patent 5,373,069 (1994).
Rehfuss, J. W.; St. Aubin, D. L., U.S. patent 5,356,669 (1994).
He, Z. A.; Blank, W. J., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
1998, p. 21.
Parekh, G. G., J. Coat. Technol., 1979, 51(658), 101.
Vaughan, G. D.; Jacquin, J. D., Proc. Int. Coat. Exhib., 1999, p. 288.
12
Binders Based on
Polyisocyanates:
Polyurethanes
Polyurethanes are polymers containing urethane (.NH22CO22O,) linkages; carbamate
is a synonym for urethane. Urethanes are usually formed by reaction of an alcohol with
an isocyanate, but they can also be made by other methods. Commonly, urethanes
made by a nonisocyanate route are called carbamates. To add to the confusing terminology, the terms urethane and polyurethane are applied to almost any binder derived from
isocyanates even though only part, if any, of the reaction products are urethanes. Isocyanates are used to make urethane-modified alkyds (uralkyds) as discussed in Section 15.7.
The market for urethanes in coatings in the United States in 2002 has been reported to be
about 250 million kilograms [1].
Urethane groups form intermolecular hydrogen bonds between polymer molecules;
they may be acyclic and/or cyclic:
Under mechanical stresses, energy (about 20 to 25 kJ mol21 of acyclic hydrogenbonds) may be absorbed by separation of hydrogen-bonds, which can re-form (probably
in different positions) when the stress is removed. Energy absorption by this reversible
bond-breaking/bond re-forming process reduces the likelihood of irreversible breaking
of covalent bonds leading to degradation. In addition, the urethanes can be designed
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
231
232
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
with “soft” and “hard” segments: that is, segments having low Tg and high Tg, respectively.
When stress is applied, the soft segments can extend between the hard segment anchors.
This factor, combined with the hydrogen-bonding, permits the preparation of coatings
with 2H hardness and 300% extension. These properties permit the design of polyurethanes
that are abrasion resistant while still resisting swelling with solvents. Polyurethanes tend to
absorb water by hydrogen-bonding with water, which plasticizes the coatings.
The isocyanate group is highly reactive, so polyisocyanates can be used to make coatings
that cure at ambient temperature or at moderately elevated temperatures. Coatings based on
aliphatic diisocyanates exhibit exceptional exterior durability when stabilized with hindered
amine light stabilizers (Section 5.3.3). The resistance of urethane coatings to environmental
etching is superior to that of many melamine–formaldehyde (MF) cross-linked coatings [2].
The principal limitations of isocyanates are cost and toxicity (particularly of relatively
low molecular weight compounds). Any cross-linker that reacts with hydroxyl, amine,
and/or carboxylic acid groups near room temperature is likely to be toxic, since the
body contains proteins and other materials having such substituents. The important
question is not toxicity as much as toxic hazard. Since higher molecular weight reduces
vapor pressure and permeability through body membranes, toxic hazard decreases, in
general, as molecular weight increases. With adequate ventilation and use of air masks
and protective clothing, as commonly done in chemical factories and resin manufacturing
plants, even relatively low molecular weight isocyanates can be handled safely. Most lowmolecular weight diisocyanates are sensitizers, however, which can result in an allergic
reaction on further exposure to isocyanate. The exposure leading to sensitization varies
over a wide time period for individuals and may occur only after years of repeated
exposure for some. The most common symptoms are hives and asthma. In extreme
cases sensitized people cannot be in the same room, or even the same building, in
which isocyanates are handled. Isocyanate suppliers supply safety instructions for handling and using isocyanates. Coating applicators, especially amateur users, are less likely
than professional chemists to have appropriate facilities or even to follow safety instructions. In using any highly reactive cross-linker, the competence of the user to handle toxic
materials should be taken into consideration in establishing formulations. To minimize the
hazard, most isocyanates used in coatings are oligomeric or polymeric derivatives.
12.1. REACTIONS OF ISOCYANATES
Isocyanates react with any active hydrogen compound. Alcohols and phenols react with
isocyanates to form urethanes, as in the following reaction, in which R and R0 can be
aromatic or aliphatic groups: The reaction is reversible at elevated temperatures.
In general, rates of urethane formation decrease in the following order: primary alcohols .
secondary alcohols . 2-alkoxyethanols . 1-alkoxy-2-propanols . tertiary alcohols. As a
rule, ease of reversion is the inverse of reactivity. Urethanes from tertiary alcohols are relatively unstable and may decompose with heat to give alkenes, carbon dioxide, and amines,
rather than alcohols and isocyanates.
12.1. REACTIONS OF ISOCYANATES
233
Urethanes react with isocyanates to form allophanates. This reaction is much slower
than the reaction of isocyanate with alcohol.
Isocyanates react rapidly with primary and secondary amines to form ureas. The reaction
is much faster than the reaction of isocyanates with alcohols. For many coating applications,
reactions are too rapid; however, hindered amines have been developed that react sufficiently
slowly to permit use in two package (2K) coatings [3,4] (Section 12.4.1):
Isocyanates react with ureas to form biurets. Biuret formation is slower than urethane
formation, but faster than allophanate formation.
Isocyanates react with water to form unstable carbamic acids, which dissociate into
carbon dioxide and an amine. The amine is so much more reactive than water that it
reacts with a second isocyanate (in preference to water) to form a urea. The reactivity
of water with isocyanates is somewhat slower than that of secondary alcohols but more
rapid than that of ureas.
Carboxylic acids react relatively slowly with isocyanates to form amides and CO2.
Hindered carboxylic acid groups such as in 2,2-dimethylolpropionic acid (DMPA) react
very slowly; accordingly, DMPA is commonly used to incorporate COOH groups into
polyurethanes.
234
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
Isocyanates also react with each other to form dimers (uretdiones) and trimers (isocyanurates). Formation of uretdiones is catalyzed by phosphines. Trimerization to form aliphatic isocyanurates is catalyzed by quaternary ammonium compounds; trimerization of
aromatic isocyanates is catalyzed by tertiary amines. Uretdiones decompose thermally
to regenerate isocyanates and are used as blocked isocyanates (Section 12.5). Isocyanurates are stable and are used extensively as building blocks for multifunctional isocyanates
(Section 12.3).
12.2. KINETICS OF REACTIONS OF ISOCYANATES WITH ALCOHOLS
The mechanism and catalysis of reaction of isocyanates with alcohols are not fully
understood. One must use caution when evaluating rate data from the literature, especially
in interpreting tables of rate constants compiled from different sources. The original
literature should be consulted to make sure that comparisons are based on studies done in
the same solvents and initial concentrations, since reaction rates depend on these variables.
Also, rates of urethane formation are commonly determined by following the disappearance
of isocyanate. If other products, such as allophanates and isocyanurates, are also formed, isocyanate disappearance does not correspond to urethane formation. Further, many of the
papers do not give the number of replicates, if any, or statistical confidence limits for the data.
Formulators often compare rates of cross-linking by determining the MEK doublerub resistance of films as a function of time and/or temperature and/or catalyst.
Usually, the test is stopped when the number of MEK double rubs without marring
of the surface exceeds 200. The test is sensitive to cross-link density in the relatively
early stages of the reaction but is not sensitive to variations in the high extent of reaction
obtained when films are cured. A film that reaches 200þ double rub resistance faster
than a second film will not necessarily reach full cross-linking more rapidly, or even
as rapidly as the second film. Another technique for following initial cross-linking is
determination of the time required for a solution of a triol and diisocyanates to gel.
Dynamic mechanical analysis (DMA) or solvent swelling are more reliable ways to
measure the cross-link density of more highly cross-linked coatings. Examples of the
fact that a 200þ MEK double rub test does not indicate full cure, or even give the
same rank order of time to reach 200þ double rubs as the time to reach full cure,
are given in Ref. [5] through comparisons of results with the measurement of cross-link
density by DMA analysis.
12.2. KINETICS OF REACTIONS OF ISOCYANATES WITH ALCOHOLS
235
12.2.1. Noncatalyzed Reactions
One might expect urethane formation to follow second-order kinetics, with the rate proportional to the concentration of each reactant: as shown in Eq. 12.1. Indeed, the reaction
of n-butanol with phenyl isocyanate in the dipolar aprotic solvent acetonitrile has been
reported to follow second order kinetics [6].
rate ¼ k½R22N55C55O½R022OH
(12:1)
Other kinetic studies show that the situation is more complex, and often the kinetic
order changes as the reaction progresses. In at least some nonpolar or low-polarity
solvents, the reaction follows third-order kinetics, at least in its early stages, first order
in isocyanate and second order in alcohol concentration. It seems reasonable to assume
that in most coating applications, reactions will be occurring in low polarity media.
Rate equation 12.2, would apply to such cases: whose implication of Eq. 12.2 is that
two molecules of alcohol are involved in the reaction with one molecule of isocyanate.
2
rate ¼ k½R22NCO½R022OH
(12:2)
Such results can be reasonably interpreted in terms of the mechanism provided in
Scheme 12.1. Reaction of isocyanate and alcohol (at a rate proportional to k1) produces
a zwitterionic reactive intermediate (RI), which can revert to starting materials (k21) or
proceed to product, urethane (k2). Product formation requires proton transfer from
oxygen to nitrogen. The second molecule of alcohol facilitates this transfer by way of
a six-membered cyclic activated complex (A).
Equation 12.3 is a rate expression that takes into account the sequence of reactions in
Scheme 12.1:
rate ¼ k1 ½RNCO½R0 OH
k2 ½R0 OH
k2 ½R0 OH þ k1
Scheme 12.1
(12:3)
236
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
If the term k2[R0 OH] is substantially greater than k21, Eq. 12.3 simplifies to Eq. 12.1.
This situation is favored when the alcohol concentration is high, such as at the start of the
reaction. If, on the other hand, the term k21 is substantially larger than k2[R0 OH],
Eq. 12.3 simplifies to Eq. 12.2 and second-order dependence on alcohol concentration
is observed. The latter circumstance is more likely, as alcohol is consumed during the
reaction, reducing [R0 OH]. As a result, the reaction rate diminishes rapidly as the reaction proceeds (since the rate is proportional to the square of a small, diminishing alcohol
concentration). The mechanism in Scheme 12.1 provides an explanation for the strong
solvent effects observed. Rates decrease as the hydrogen-bond acceptor potential of solvents increase in the order aliphatic hydrocarbons, aromatic hydrocarbons, esters and
ketones, ethers, and glycol diethers. Rates in aliphatic hydrocarbons can be two
orders of magnitude faster than in glycol diethers. Hydrogen-bonding of the zwitterionic
reactive intermediate RI and/or the alcohol with solvent may reduce the rate by
reducing the concentration of cyclic activated complex A. The mechanism in Scheme
12.1 is also consistent with the higher reactivity observed for aromatic isocyanates.
When R is an aromatic group, the negative charge on nitrogen in RI is delocalized in
the p-electron system of the aromatic ring, resulting in a lower energy (by resonance
stabilization) for RI and a faster rate of urethane formation relative to aliphatic
isocyanates.
Rate constants reported in the literature for reactions of alcohols with isocyanates are
commonly apparent initial second-order rate constants. Initial rate constants provide a
useful way of comparing reactivities without becoming involved in the complexities of
the kinetics, provided that comparisons are made in the same solvent and the same
initial concentrations of alcohols and isocyanates. Unfortunately, rate constants are sometimes compared which were determined in different solvents and/or at different initial
concentrations. Such comparisons can be misleading.
The kinetics of the isocyanate-alcohol reaction are even more complex than discussed
thus far, since the reaction is autocatalyzed by the urethane being formed. Another cyclic
transition state, B, could be involved, in which a molecule of urethane facilitates proton
transfer from oxygen to nitrogen in the zwitterionic reactive intermediate RI.
The autocatalyzed reaction would follow third-order kinetics but would be first order
in alcohol. The decline in alcohol concentration would be offset by the increase in
urethane concentration as the reaction proceeds. The apparent rate constant would
change through the course of the reaction depending on the rate constant for the
reaction involving activated complex A in comparison with that for complex B. Sato
studied the reaction of various isocyanates with methyl alcohol in di-n-butyl ether [7].
In the absence of added catalyst, his results can be expressed by Eq. 12.4, where a and
b are the initial concentrations of isocyanate and methyl alcohol, x is the concentration
12.2. KINETICS OF REACTIONS OF ISOCYANATES WITH ALCOHOLS
237
of product (urethane), and k2 represents the rate constant of the autocatalytic reaction.
dx
¼ k1 (a x)(b x)2 þ k2 (a x)(b x)
dt
(12:4)
It is assumed that there are no side reactions; therefore, a 2 x in Eq. 12.4 equals
[RNCO] in Eq. 12.2, and b 2 x equals [R0 OH]. In most cases studied by Sato, k2 was
larger than k1, but in a few cases, it was comparable. When k2 is substantially larger,
the second term in Eq. 12.4 dominates in the later stages of the reaction (as the urethane
concentration builds up). Sato’s studies were done in di-n-butyl ether; it would be of
interest to know what changes in the rate constants would result from changes in
solvent polarity. Sato reported that autocatalysis is more important for aliphatic than
for aromatic isocyanates, although the rate constants for autocatalysis were similar [7].
This results from the higher reactivity of aromatic isocyanates with alcohols (k1),
which reduces the relative importance of autocatalysis.
12.2.2. Catalysts
Reactions of isocyanates with alcohols are catalyzed by a variety of compounds, including
bases (tertiary amines, alkoxides, carboxylates), metal salts and chelates, organometallic
compounds, acids, and urethanes. Most primary aliphatic amines react so rapidly with
isocyanates at room temperature that their reaction rates and effect of catalysts have not
been determined. The reactions of amines with isocyanates are catalyzed by carboxylic
acids and water.
The most widely used catalysts in coatings are tertiary amines, commonly diazabicyclo[2.2.2]octane (DABCO1), and organotin(IV) compounds: most commonly, dibutyltin
dilaurate (DBTDL).
The mechanisms by which these and other catalysts operate are controversial. A reasonable explanation for catalysis by amines is that they facilitate proton transfer from the
alcohol to the isocyanate (and similarly to alcohols and urethanes). Proton removal from
the alcohol may occur during reaction with the isocyanate, thereby avoiding formation of
the positive charge on oxygen and lowering the energy of the reactive intermediate,
which may proceed to product by a proton transfer from the protonated amine, as shown
in Scheme 12.2 [8]. Proton removal by amines at an earlier stage than by alcohols and
urethanes is reasonable, owing to the greater basicity of amines.
Sato included catalysts in the study described in Section 12.2.1 [7]. For triethylamine
catalyzed reactions of isocyanates with methyl alcohol in dibutyl ether, Sato’s data fit
1
DABCO is Air Products general trademark for a series of different catalysts, including diazabicyclo[2.2.2]octane, which is DABCO L-33. However, DABCO has become widely used just for diazabicyclo[2.2.2]octane.
238
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
Eq. 12.5, where k3 and (cat) represent the rate constant of the catalyzed reaction and the
concentration of catalyst, respectively:
dx
¼ k1 (a x)(b x)2 þ k2 (a x)(b x) þ k3 (cat)(a x)(b x)
dt
(12:5)
If the catalyst is effective, k3 is larger than k1 or k2, and when the catalyst concentration
is sufficient, the rate is governed by the third term and is first order in [OH]. This is
generally observed in amine catalyzed reactions. The mechanism in Scheme 12.2 is
consistent with the third term of rate Eq. 12.5. Amine basicity is not the only factor
since DABCO, a weaker base, is a more active catalyst than triethylamine. The nitrogen
electron pairs are more readily accessible with DABCO than with triethylamine, which
could account for the higher catalytic activity. This possibility, that both nitrogen basicity
and electron accessibility are important factors, is supported by the even higher catalytic
activity (relative to DABCO) of 1-azabicyclo[2.2.2]octane (quinuclidine), which has both
the nitrogen accessibility of DABCO and the high basicity of triethylamine.
Catalysis of urethane formation by 1-8-diazabicyclo[5.4.0]undec-7-ene (DBU) is
reported to result from the formation of the ammonium ion and a hydrogen-bond type
complex with the polyol [9].
Comparisons are complicated because catalysts also catalyze allophanate formation
and trimerization of aromatic isocyanates to form isocyanurates. For example, reaction
of phenyl isocyanate with n-butyl alcohol (at 508C in acetonitrile) in the presence of
pentamethyldipropylenetriamine (PMPTA) yielded 30% urethane, while 70% of the isocyanate was converted into triphenylisocyanurate [10]. On the other hand, when
DABCO was used as a catalyst, the urethane was the principal product, with a small
amount of an allophanate also formed. Possibly, urethane formation is favored by sterically
accessible amines (e.g., DABCO) and isocyanurate formation is less sensitive to this factor.
The zinc complex of 2,4-pentanedione [Zn(AcAc)2], tin octanoate, and quaternary
ammonium compounds such as tetramethylammonium octanoate specifically catalyze
allophanate formation [11]. (Rate constants for the formation of urethanes, allophanates,
and isocyanurates for a variety of catalysts and co-catalysts are given in Ref. [11].)
Scheme 12.2
12.2. KINETICS OF REACTIONS OF ISOCYANATES WITH ALCOHOLS
239
Acids also catalyze the reaction, perhaps by protonating the isocyanate group.
Carboxylic acids act as catalysts [12]. Acetic acid has been reported to be a more effective
catalyst than a carboxylic acid –functional polyether polyester [13]. The effect of temperature on the reaction rate of m-chlorophenyl isocyanate with n-butanol in heptane catalyzed
with acetic acid has been studied and mechanisms proposed for the nonlinear increase in
reactivity with temperature [14]. The effect is related to the reduced dimerization of acetic
acid at higher temperature. Stronger acids are even more effective [15]. Phenyl acid phosphate is less effective as a catalyst than DBTDL at temperatures below 1008C, but is more
effective at 1308C. Blocking the acids by forming amine salts extends the pot life without
reducing the cure rate at 1308C.
Catalysis by organometallic compounds of many types has been reported. The majority
of these are organotin compounds. Of the many metal derivative catalysts, DBTDL is the
most widely used in coatings. It is soluble in a wide range of solvents, comparatively low
in cost, colorless, and in general, highly effective at levels on the order of 0.05 wt%.
DBTDL promotes urethane formation without promoting allophanate formation [16] or
trimerization [10]. Being hydrophobic, it is said to favor urethane formation over hydrolysis of the isocyanate when moisture is present. Dimethyltin diacetate (DMTDA) is usually
a somewhat more effective catalyst than DBTDL and is particularly useful with sterically
hindered isocyanates. Tin compounds are effective catalysts for reaction of alcohols with
aromatic isocyanates and are even more effective with aliphatic isocyanates. Although
aromatic isocyanates are more reactive than aliphatic isocyanates in uncatalyzed
reactions with alcohols, the reactivity of aliphatics and aromatics can be roughly equal
with DBTDL catalysis. On the other hand, amine catalysts are more effective with
aromatic than aliphatic isocyanates.
Many mechanisms have been proposed for the catalytic activity of tin compounds, but
none has been accepted universally. A plausible proposal is based on studies of reaction of
phenyl isocyanate with excess methyl alcohol using dibutyltin diacetate (DBTDA) as catalyst [17]; see also Ref. [18]. The reaction rate is first order in isocyanate concentration
and half order in both alcohol and catalyst concentration. Restating the kinetic equation
in the same form used thus far gives Eq. 12.6:
dx
¼ k3 (cat)1=2 (a x)(b x)1=2
dt
(12:6)
Based on kinetic results and the observation that the rate was suppressed by the addition
of acid, the mechanism in Scheme 12.3 was suggested. The mechanism involves sequential complexation of alcohol (with loss of Hþ) and isocyanate to the tin. The proposed
proton loss is consistent with the reduction observed in the activity of tin catalysts in
the presence of carboxylic acids. Presumably, addition of Hþ favors decomplexation of
alcohol and reversion to starting materials. This proposal is also consistent with the
observation that tin compounds are not effective catalysts for reactions of isocyanates
with amines.
Coordination of the isocyanate with tin is consistent with the roughly equivalent
reactivity of aliphatic and aromatic isocyanates observed with tin catalysts, since
stabilization of the reactive intermediate by electron delocalization with the aromatic
ring is eliminated as a major factor. Furthermore, tin activates both the alcohol and isocyanate, whereas amine catalysts probably activate only the alcohol by facilitating
proton removal.
240
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
Scheme 12.3
The rate dependence on alcohol concentration has important implications on the pot life
versus curing schedule with isocyanates. If the dependence changes from first to second
order (characteristic of the uncatalyzed reaction), the rate will diminish rapidly as the reaction nears completion. With less sensitive half-order dependence, as reported for tin catalysis, the rate at high concentrations (during storage) will be lower and will not slow
down as much as the reaction proceeds to completion, thereby favoring both a longer
pot life and a shorter cure time (or lower cure temperature). The first-order dependence
with amine catalysis results in an intermediate situation. Unfortunately, there are only
limited reports in the literature of the effect of solvent on rates of catalyzed reactions.
Another mechanism that has gained favor in recent years is via alcoholization of the
organotin ester to the organotin alkoxide:
Dibutyltin di(2-ethylhexanoate) interacts with an alcohol to form tin alkoxides (mono
or di), which then react with the isocyanate to give complexes, which in turn react with the
alcohol to give the urethane [19]. The effect of solvent on the catalysis of the reaction of
phenyl isocyanate with cyclopentanol in different solvents indicates that the principal
reaction pathway is through the dialkyltin monoxide monoester.
The catalytic effect of a series of dibutyltin derivatives on the rate of reaction of phenyl
isocyanate with ethylene glycol in THF has been published [20]. Catalytic activity
decreased in the order methoxide . laurate . perfluorobutyrate . benzene sulfonate .
12.2. KINETICS OF REACTIONS OF ISOCYANATES WITH ALCOHOLS
241
TABLE 12.1. Tensile Modulus (psi) (% Elongation-to-Break)
Catalyst
Initial
1 Week
4 Weeks
DMTDC
DBTDL
DBTM
2450 (270)
1900 (270)
1400 (280)
1900 (300)
450 (400)
450 (330)
800 (330)
200 (436)
250 (400)
chloride . hexafluoroarsenate; that is, it decreased with increasing acid strength of the
substituent. In the presence of dimethylbenzylamine, rates increased but the order
changed to perfluorobutyrate . methoxide ¼ laurate . chloride ¼ hexafluoroarsenate.
Cyclic organotin compounds such as 3,3-di-n-butyl-3-stanna-2,4-dioxa-8-thiospiro[5.3]nonane have been patented as catalysts for the reactions of sterically hindered
isocyanates with alcohols [21]. A model compound reaction example given was t-butyl
isocyanate with 2-ethoxyethyl alcohol.
Catalysts prepared by reacting DBTDL with p-toluenesulfonylisocyanate can be used
to make storage stable mixtures of polyisocyanate and catalyst in contrast to DBTDL
[22]. The catalyst is readily hydrolyzed with water or alcoholized with an alcohol. The
protected catalyst can be used in moisture curing coatings.
Metal chelates of 2,4-pentanedione [acetylacetone (AcAc)] also serve as catalysts. It
has been proposed that AcAc metal derivatives such as Zr(AcAc)4 activate the hydroxyl
group followed by insertion of isocyanate [23]. This catalyst is selective, significantly
favoring reaction with hydroxyl groups over reaction with water, so it is particularly
useful in two package (2K) waterborne urethane coatings. No significant synergy
was observed between Zr(AcAc)4 and tertiary amines [24]. The catalytic effect of
Zr(AcAc)4 on the reaction of butanol and butyl isocyanate is reduced by adding excess
acetyl acetone or acetic acid [25]. The effect of solvent composition on the reaction
rate is reduced substantially by metal AcAc catalysis.
Many metal salts have been shown to be catalysts. There is a wide variation in the extent
of catalysis of the NCO – OH reaction, depending on the diisocyanate and catalyst. Stannous
octanoate and lead octanoate are particularly effective although less so than DBTDL.
Increased catalyst concentration not only increases reaction rate but can also decrease
hydrolytic stability of films after application and curing. The effect is most evident in polyester–polyurethane films. Table 12.1 shows the effect on modulus (a measure of cross-link
density) of exposure of films of H12MDI–polyestertriol to 95% relative humidity at 708C
for various catalysts: DMTDC, DBTDL, and dibutyltin methoxide (DBTM) [26]. The
decreases in modulus observed seem reasonable, since the initial step in the hydrolysis of a
urethane may well be the reverse reaction to the reaction of an isocyanate with water.
Further studies are warranted with other catalysts, as well as with films cured without catalyst.
Amines can not only catalyze the formation of urethanes, they can also catalyze the
thermal decomposition of the polymer. For example, substantial reduction in the physical
properties of polyurethane elastomers have been reported when samples are exposed to
temperature above 1058C for long periods of time (400 hours), due to the presence of
DABCO catalyst [27].
12.2.3. Interrelationships in Catalysis
Combinations of DABCO and DBTDL often act synergistically; that is, the effect of the
combination is greater than would be predicted by the sum of the individual effects of the
two catalysts. It has been suggested in a study of the mechanism with DBTDL and
242
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
DABCO that a highly active complex forms between the tin and the amine. In the presence
of an alcohol it is suggested that this complex facilitates the alcoholysis of tin carboxylates
to the highly active tin alkoxides, which then rapidly adds isocyanate across the tin –
oxygen bond to form the carbamate adduct of tin. The adduct breaks up with another
molecule of alcohol to form the urethane and regenerate the tin alkoxide [28]. Metal
acetylacetone complexes act synergistically with DBTDL [29].
To increase pot life without reducing cure rate significantly, it is common to add volatile
inhibitors to a coating. The mechanism of the inhibition of organotin catalysis of urethane
formation by carboxylic acids has been investigated [30]. The pot life of DBTDL catalyzed
coatings can be extended by addition of small amounts of volatile carboxylic acids, such as
formic or acetic acids; after application the acids evaporate and there is little, if any, effect
on cure rate. Pot life can also be extended by the addition of 2,4-pentanedione (acetylacetone) with minimal effect on dry times. For example, coatings with improved pot life and
good curing are formulated with 2,4-pentanedione and DBTDL [31].
The use of amine salts as blocked catalysts for the reaction of aromatic isocyanates with
polyols in 2K systems has been patented [32]. For example, the triethyl amine (TEA) salt
of dibutyl phosphate in an MDI – polyether polyol 2K coating had a pot life of 45 minutes
at 258C, in contrast to 3 minutes with a stannous chloride – catalyzed system. Both coatings cured to the same degree at 1208C. Presumably, at the higher temperature there
was sufficient free amine and dibutyl phosphate for both to act as catalysts.
12.3. ISOCYANATES USED IN COATINGS
Aromatic and aliphatic isocyanates are used in coatings; the former are less expensive and
the latter provide films with better color retention and exterior durability.
12.3.1. Aromatic Isocyanates
The aromatic isocyanates most widely used in coatings are based on MDI. MDI is
available in several grades: bis(4-isocyanatophenyl)methane, a mixture of 55% of
the 2,40 isomer and 45% of the 4,40 isomer; and several oligomeric (frequently
called polymeric) MDI with longer chains of methylene phenyl groups. MDI is also
used as a prepolymer with polyether polyols. The volatility (particularly of the oligomeric grades) is low enough to reduce toxic hazard, especially as compared to toluene
diisocyanate (TDI).
12.3. ISOCYANATES USED IN COATINGS
243
The most common grade of commercial TDI consists of a mixture of about 80%
2,4- and 20% 2,6-diisocyanate isomers. Nearly pure 2,4-TDI is also available at a
premium price. Due to toxic hazards, TDI is not used as such in final coating formulations. For coatings in which unreacted isocyanate groups are needed, TDI is converted
into derivatives of higher molecular weight and higher functionality. Higher molecular
weight reduces the toxic hazard, and the higher functionality yields solvent resistant
films more rapidly.
2,4-TDI has the advantage of a differential in reactivity between the ortho- and the
para-isocyanate groups with alcohols, which makes possible synthesis of isocyanurates
and prepolymers with narrower molecular weight distribution than with diisocyanates,
in which the isocyanate groups are equally reactive. At 408C, the para-isocyanate
group of TDI is about seven times more reactive than the ortho group. Furthermore, no
matter which isocyanate group reacts first, the second group is less reactive than the
first. Overall, after reaction of the para-isocyanate, the remaining ortho group is 20
times less reactive than a para-isocyanate on a second TDI. The difference in reactivity
decreases as the temperature increases; at temperatures above 1008C, the ortho- and paraisocyanate groups have similar reactivities. Thus, for maximum selectivity a prepolymer
should be prepared at a low temperature; of course, this means that the reaction is slow.
Catalysts can be used, but the catalyst stays in the product and therefore sets a lower
limit on the amount of catalyst in the final formulation.
Studies of the reaction of TDI with alcohols catalyzed with tertiary amines and with
organotin compounds showed that the amines tended to equalize the reactivity of the
two NCO groups [33]. In a study of the ratio of para-to-ortho urethane formation in
the reaction of TDI with n-butanol in benzene with several catalysts, the lowest
ratio (2.8) was obtained with SnCl4, the highest ratio (9.9) with tin octanoate,
DBTDL, and triisopentylphosphine. For the uncatalyzed reaction at 208C the ratio was
11.5 [34].
Almost any polyhydroxy compound can be reacted with TDI to make prepolymers.
Low molecular weight hydroxy-terminated polyesters or mixtures of diols and triols are
commonly used. For safety, the levels of unreacted TDI in the prepolymer must be very
low. Low levels of TDI can be assured by using an NCO/OH ratio below 2 : 1 and
pushing the reaction to completion, but chain extension (caused by reaction of both
NCO groups of some TDI molecules) increases the molecular weight of the product.
An alternative process, used when low molecular weight prepolymers are needed, is to
react the polyhydroxy compound, often trimethylolpropane (TMP), with a large excess
of 2,4-TDI, then remove the excess TDI using a vacuum wiped-film evaporator. Very
low levels of free TDI and minimal chain extension are attainable, yielding low molecular
weight products suitable for high-solids coatings.
244
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
The isocyanurate derived from TDI made by trimerizing TDI (Section 12.1) has a lower
toxic hazard than that of monomeric TDI. The trimerization reaction occurs exclusively
with the para-isocyanate group.
12.3.2. Aliphatic Isocyanates
The principal aliphatic isocyanates used are 1,6-hexamethylene diisocyanate
(HDI), isophorone diisocyanate (IPDI), bis(4-isocyanatocyclohexyl)methane (H12
MDI), 1,3-xylylenediisocyanate (XDI), tetramethyl-m-xylidene diisocyanate (TMXDI),
m-isopropenyl-a,a-dimethylbenzylisocyanate (TMI), and 2,2,5-trimethylhexane diisocyanate (TMHDI). Diisocyanates are usually converted to derivatives before use in
coatings to increase functionality and reduce toxic hazard.
HDI is especially hazardous and is handled on a large scale only in chemical plants. The
first less hazardous derivative was a biuret, which can be made by reacting HDI with a
small amount of water and removing the excess HDI. The structure of HDI biuret
shown below is idealized; commercial products contain varying fractions of oligomeric
biurets. (The presence of oligomeric biurets makes the average functionality higher
than 3.) These polyfunctional isocyanates give coatings with good color retention and
weather resistance. The viscosity of an early commercial product was about 11.5 Pa.s at
208C. Grades with lower average molecular weights (and average functionalities nearer
12.3. ISOCYANATES USED IN COATINGS
245
to 3) are available with viscosities as low as 1.4 Pa.s.
HDI isocyanurates are used on a larger scale. The isocyanurate gives coatings with
greater heat resistance and even better long-term exterior durability than does HDI
biuret. Commercial products contain oligomeric material, and average functionality is
over 3. Grades with lower oligomer content with viscosities as low as 1 Pa.s are available.
With ammonium fluoride as a catalyst, an isomeric trimer, an iminooxadiazenedione
(called an unsymmetrical trimer) of HDI, is made as an approximately 50 : 50 mixture
with HDI isocyanurate. It has the advantage that viscosity at the same oligomer content
is lower than that of the corresponding HDI isocyanurate: 1 Pa.s versus 3 Pa.s [35]. A
very low viscosity unsymmetrical trimer with a viscosity of 0.7 Pa.s has been made available, it is useful in very high-solids coatings and because of the low viscosity is easily
incorporated in 2K waterborne urethane coatings [36].
Allophanate derivatives of HDI and IPDI are another type of polyfunctional isocyanate.
They are made by reacting an alcohol or diol with excess isocyanate, then removing
unreacted diisocyanate with a wiped film evaporator giving an isocyanate-terminated allophanate [37]. The properties can be varied by using different alcohols to make the starting
urethane, to make the compounds with different R substituents shown in the structure. For
example, the cetyl alcohol urethane from HDI yields an allophanate diisocyanate that is
soluble in aliphatic hydrocarbons. Derivatives with higher functionality are made by reacting a glycol with excess diisocyanate.
The uretdione dimer of HDI also has lower volatility with low viscosity and can be used
for cross-linking in ambient cure coatings. Very low viscosity grades (,100 mPa.s) are
reported to permit formulation of very low VOC coatings [38].
Commercial IPDI is a mixture of Z (cis) and E (trans) isomers in a 75 : 25 ratio. The
isomers are difficult to separate. Isophorone diisocyanate has two different types of NCO
groups. Studies performed under different conditions show that with DBTDL catalysis, the
246
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
secondary NCO group of both Z and E isomers are more reactive than the primary NCO group
[39–41]. The selectivity decreases with increasing temperature, and selectivity is greater
with sec-butyl alcohol than with n-butyl alcohol [41]. DABCO [39] promotes selective reaction of the primary isocyanate group of IPDI in contrast to other amines [40] and DBTDL
[41], where selectivity favors reaction with the secondary isocyanate groups. Selectivity
decreases with increasing temperature [41]. High selectivity is particularly important in
making IPDI prepolymers when low molecular weight and narrow distribution of molecular
weight are desired. It was concluded that the optimal conditions for prepolymer synthesis
based on IPDI are temperatures between 40 and 608C using DBTDL catalyst.
Isocyanurate derivatives of IPDI analogous to the HDI isocyanurates discussed above
are widely used cross-linkers. The rigidity of IPDI affords films with higher Tg. By blending IPDI and HDI isocyanurates, formulators can dial the Tg desired for a particular
application.
Bis(4-isocyanatocyclohexyl)methane) (H12MDI) is less volatile than HDI and IPDI and
is sometimes used as a free diisocyanate in coatings to be applied by roll coating but not by
spray coating. It is a mixture of stereoisomers; since both isocyanato groups are secondary,
reactivity is lower than HDI or IPDI.
XDI, TMXDI, and TMI have aromatic rings, but give color retention and exterior durability equivalent to those of aliphatic isocyanates. The exterior durability of their urethanes
probably results from the absence of isocyanate groups directly substituted on the aromatic
ring, as well as from the absence of abstractable hydrogens on the carbons adjacent to
nitrogen in the cases of TMXDI and TMI. Since the isocyanate group is on a tertiary
carbon, the reactivity is lower than that of less sterically hindered aliphatic isocyanates.
This difference can be offset by using higher catalyst levels and sterically accessible tin
catalysts such as DMTDA instead of DBTDL. TMXDI is offered as a low molecular
weight, essentially diisocyanate-free prepolymer with trimethylolpropane. TMI is used
as a comonomer with acrylic esters to make 2000 to 4000 MW copolymers with 40 to
50 mole% TMI; thus, each molecule has several isocyanate groups [42].
12.4. TWO PACKAGE SOLVENTBORNE URETHANE COATINGS
The largest volume of urethane coatings comprises two package (2K) coatings that are
mixed just before application. One package contains the polyol (or other coreactant),
pigments, solvents, catalyst(s), and additives; the other contains the polyisocyanate and
moisture-free solvents. Sometimes, the catalyst is in a separate third package so that the
cure rate can be adjusted for variations in ambient conditions. While the major reaction
is formation of urethane cross-links, some urea cross-links result from reaction of
atmospheric water with the isocyanate. 2K urethane coatings can be formulated for
ambient-temperature cure or baking coatings.
In many applications the components are mixed at the start of a work shift. Viscosity
starts to increase as soon as they are mixed, but the coatings must remain sprayable for
4 hours or for 8 hours after mixing. They are said to have a 4- or 8-hour pot life. For
factory application the pot life can be considerably shorter when special equipment is
used to mix the components just before spraying. If a very fast cure rate at relatively
low temperatures is needed, reactive coreactants and/or high catalyst levels are used,
and the pot life is very short. Such formulations can be applied using special spray equipment in which the two packages are fed to a spray gun by separate proportioning pumps
and mixed inside the gun just before they are sprayed or by impingement mixing as they
12.4. TWO PACKAGE SOLVENTBORNE URETHANE COATINGS
247
are sprayed. Such spray equipment is expensive, and care is required to assure that the
proper ratios are fed and thoroughly mixed.
An important variable in 2K polyisocyanate – polyol coatings is the NCO/OH ratio. The
ratio is commonly called indexing, or the index. In ambient cure systems, it is often found
that a ratio on the order of 1.1 : 1 gives better film performance than a 1 : 1 ratio. A probable
reason is that part of the NCO reacts with water from solvent, pigment, or air to give urea
cross-links. To the extent that this happens, two hydroxyls are unreacted for each water
molecule; use of excess NCO minimizes residual unreacted hydroxyl groups. Solvent
resistance is also improved. Since polyisocyanates are typically lower in viscosity than
the polyol, excess NCO gives a lower VOC [4]. Some finishes (e.g., aircraft finishes) are
formulated with NCO/OH ratios as high as 2 : 1. The resulting reduced hydroxyl concentration gives longer pot life. It is possible that the high mobility of water and the reactivity
of amine groups resulting from the water – isocyanate reaction favor faster reaction rates at
higher Tg values than are possible with only the hydroxyl –isocyanate reaction. Since the
rate of reaction of isocyanates generally decreases in the order primary alcohols . secondary alcohols . 2-hydroxyethyl ether alcohols . 2-hydroxypropyl ether alcohols, the
required excess of NCO over OH tends to increase in that order.
Almost any hydroxy-functional coreactant can be used; hydroxy-terminated polyester and
hydroxy-substituted acrylic resins are most common. In general terms, polyesters permit
higher solids and give films with greater solvent resistance and better adhesion to metals.
Acrylics provide faster dry, lower cost (since their equivalent weight can be made higher
than that of polyesters, resulting in the need for less of the more expensive isocyanate),
and better exterior durability, resulting from superior hydrolytic and photochemical stability.
Other resins are also used in 2K coatings. Alkyd resins have unreacted hydroxyl groups
and their rate of dry can be accelerated by adding a polyisocyanate such as the isocyanurate trimer of IPDI just before application. Nitrocellulose (shipped wet with plasticizer
instead of ethyl or isopropyl alcohol) is used in formulating cross-linking furniture lacquers with isocyanates. Bisphenol A (BPA) epoxy resins are also cross-linked with isocyanates through their hydroxyl groups. Although hydroxy-terminated polyethers are widely
used in urethane foams, they have limited use in coatings, since the resulting coatings
show high moisture vapor permeability, relatively poor exterior durability, and are soft
as a result of the low Tg of the polyethers.
As cross-linking proceeds, the Tg of the film increases. If the Tg of a partially reacted system
is near the temperature at which cross-linking must occur, the rate of urethane formation
becomes mobility controlled and the cure rate is slowed. If the Tg of the fully reacted
system is significantly above the cure temperature, the reaction virtually stops before the crosslinking reaction has gone to completion [43]. Since in many ambient cure coatings, one wants a
coating with a Tg somewhat above the curing temperature, selection of a combination of polyisocyanate and polyol that provides an appropriate final Tg is critical. A polyisocyanate with
relatively flexible aliphatic chains (such as HDI isocyanurate) can be used with a relatively
high Tg acrylic or polyester. On the other hand, a polyisocyanate that yields relatively rigid
cross-link segments [such as the TMP/a,a-tetramethylxylyl diisocyanate (TMXDI) prepolymer] requires a lower Tg acrylic or polyester. Mobility control of reactions is more likely to be
encountered in ambient temperature cure coatings than in baked coatings. Another approach to
balancing Tg is to use mixtures of IPDI and HDI isocyanurates [44].
With 2K coatings there is always a compromise between pot life and curing time (Section
2.3.2). Several formulating variables are available to increase pot life with little effect on cure
time. Concentrations of reactive groups should be kept as low as possible, but this becomes
more difficult as one formulates to increasingly higher solids. Using mixed polyol/moisture
248
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
curing systems helps. Since isocyanate–alcohol reactions proceed most slowly in media with
high levels of hydrogen-bond accepting groups, strong hydrogen-bond accepting solvents
should be selected to the extent possible. After application, as the solvent evaporates, the
medium becomes less hydrogen-bond accepting and the reaction rate increases accordingly.
This strategy serves the purposes both of extending pot life and of promoting reactivity after
application. Since solvents with greater hydrogen-bond accepting strength tend to have higher
viscosities, solvents having intermediate hydrogen-bond accepting strength usually give a
better compromise between solids and pot life [12].
Organotin catalysts are generally preferable to amine catalysts (Section 12.2.2) because
reaction rates often depend on alcohol concentration to the one-half power with organotin
catalysts and to the first power with amine catalysts. Since the effectiveness of tin catalysts
is reduced by carboxylic acids, one can add a volatile acid such as acetic or formic acid to
the formulation. The acid inhibits reaction during the pot life stage but evaporates during
application so that the inhibiting effect disappears. Similarly, 2,4-pentadione increases pot
life by chelating with tin compounds and evaporating after application (Section 12.2.2.1).
The availability of lower viscosity isocyanate cross-linkers (Section 12.3.2) makes
possible the formulation of higher-solids 2K coatings. However, one must consider not
just the viscosity of the cross-linker, but also that of the combination of polyol and crosslinker. In some cases, a lower viscosity cross-linker also has a lower equivalent weight;
this means that the weight ratio of cross-linker to polyol must be reduced to maintain
the same NCO/OH stoichiometric ratio [45]. Since the polyol often has a much higher
viscosity, this reduced ratio can give a higher viscosity, despite the lower viscosity of
the cross-linker. Higher solids can be obtained using a somewhat higher viscosity polyisocyanate with a higher equivalent weight.
Reactive diluents are useful in increasing solids. For example, polytrimethylene carbonate polyols have been recommended as reactive diluents for acrylic polyols in coatings
cross-linked with HDI isocyanurate. Best results were obtained with relatively low molecular weight triols made by reacting trimethylene carbonate with TMP. VOC was
reduced while increasing the film flexibility and maintaining hardness, weather resistance,
and mechanical properties [46].
Imines permit formulation of 2K coatings with very high solids, longer pot life, and
faster drying than with polyols [4]. They can be used alone or in combination with a
hydroxy-functional polyester or acrylic resin [47].
Ketimines react directly with isocyanates to yield a variety of products, depending on
the particular reactants and conditions. For example, isobutyl isocyanate reacts with the
ketimine derived from methylamine and acetone in 3 hours at 608C to yield isobutyl
methyl urea and a cyclic unsaturated substituted urea [4,48].
12.4. TWO PACKAGE SOLVENTBORNE URETHANE COATINGS
249
TABLE 12.2. Viscosities (100% Solids, mPa.s at 2388 C) and Gel
Times of Substituted Polyaspartic Acid Ethyl Esters
R
Viscosity
1500
1200
150
Gel Time
.24 h
2– 3 h
,5 min
Aldimines react analogously to ketimines with isocyanates to yield unsaturated substituted ureas [4]. Since aldimines are more stable than ketimines to hydrolysis, the fraction
undergoing direct reaction with isocyanate in the presence of water is greater than that
with ketimines.
Under conditions where imines can react directly with isocyanates or with water
(Section 12.6), the ratio of the two reactions depends on the relative humidity, the time
between application and cure, and the curing temperature [4]. The reaction with water
releases amine to react with isocyanate and carbonyl compound, which volatilizes. In
the direct reaction, carbonyl compound from the imine is not released, hence VOC emissions are lower. The direct reaction is catalyzed more than the water reactions by carboxylic acids, and water and tertiary amines reduce the rate of reaction. The imines
have very low viscosities; for example, the bis(methyl isobutyl ketone) ketimine of
ethylenediamine has a viscosity of 5 mPa.s and an equivalent weight of 112. Direct
reaction is more favored with aldimines, especially above 608C.
Hindered amines are also used for auto refinish clear coats. The coatings are called 2K
urethane coatings, although the cross-links are dominantly urea cross-links. Polyaspartic
esters [3,4] are hindered amines that are used in high-solids coatings.
As shown in Table 12.2, the gel time of polyisocyanate– polyaspartic ester combinations depends on the structure of the polyaspartic ester. The pot life of these combinations is increased by the addition of DBTDL [4]. On the other hand, carboxylic acids
and water accelerate reaction rates. By proper selection of isocyanate and aspartate and
adjustment of inhibitors and catalysts, it is possible to formulate very high solids 2K coatings with a reasonable pot life and fast curing.
12.4.1. 2K Urea Coatings
Polyurea coatings are those in which the final cross-linking reaction is between isocyanate
and amine groups to form ureas. The term 2K urea coatings is used in the maintenance
250
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
field. Other applications of coatings based on isocyanates and amines—for example, some
automotive refinish urethane clear coats—are called 2K urethane coatings (Section 12.4).
Moisture-curable urethanes cure by formation of urea bonds but are not called
polyurea coatings.
The pot life of 2K urea coatings depends on the amine(s) and polyisocyanate(s) used.
Primary aliphatic amines react too rapidly to be applied with any kind of equipment, aromatic amines can be applied with special equipment, and hindered amines can be used as
2K coatings. In applications such as roof coatings, floor coatings, traffic control paint, and
thick-film sag-resistant maintenance coatings, rapid cure is very desirable (Section 33.1).
This need has motivated development of equipment that makes it possible to spray apply
coatings having very short pot lives (as short as 1 second). 2K urea coatings based on
MDI – polyoxypropylene prepolymers with aromatic amines are used in maintenance coatings for which gloss and color retention are not critical. The coatings have the advantages
that they cure rapidly even at freezing temperatures, have zero VOC, and have low temperature flexibility [49]. Roofs and floors can be walked on within 30 seconds after they are
coated with such polyureas.
Another class of hindered amines is based on the addition of primary amines to polyacrylates such as TMP triacrylate and/or to BPA epoxy resins. A combination of a tertiary
C10–12 alkylamine and t-octylamine, added to TMP triacrylate, followed by the addition of
BPA epoxy resin, provide a hindered amine that is used with HDI isocyanurate in a 2K
traffic marking paint [50]. The same hindered amine is used with HDI isocyanurate in
maintenance top coats. The adhesion to properly prepared metal and concrete is such
that no primer is required. The coatings cure even at temperatures as low as 2188C and
provide high chemical resistance, which is superior to epoxy – polyamide coatings [49].
A hindered amine mixture based on amine adducts of BPA epoxy and TMP triacrylate
was used in formulating 2K coatings with HDI isocyanurate in the other package. The
coatings had adequate pot life, tack free time of 2.3 minutes at 258C, adhesion to a
variety of substrates, and good chemical resistance and exterior durability [51].
It is reported that the ureas formed can cyclize to hydantoins, losing alcohols. The resulting
film shrinkage could weaken adhesion. To overcome this problem, aspartic ester–amides
are recommended and used in 2K traffic marking paint with HDI isocyanurate [52].
12.5. BLOCKED ISOCYANATES
Blocked isocyanates are used as cross-linkers in baked coatings. The general idea is to
react an isocyanate with a blocking group that prevents it from cross-linking at coating
storage temperatures but that releases free isocyanate at bake temperatures. Blocked isocyanates are of major importance in cationic electrodeposition coatings, powder coatings,
coatings for electrical wire, and increasingly, 1K automotive coatings. Common blocking
agents include 2-ethylhexyl alcohol and 2-butoxyethanol in E-coats and e-caprolactam, 2butanone oxime [methyl ethyl ketone oxime (MEKO)], and uretdiones in powder coatings.
For detailed discussion of these topics, see Sections 27.2 and 28.1.3, respectively.
A series of extensive review articles on blocked isocyanates may be found in Ref. [53].
Although blocked isocyanate is the most widely used term, masked isocyanate and capped
isocyanate are sometimes used. The isocyanate group can be “blocked” through one of
several methods.
12.5. BLOCKED ISOCYANATES
251
12.5.1. Principles of Blocking and Deblocking
There are two possible modes of reaction of blocked isocyanates with a hydroxyl group: (1)
dissociation of the blocked isocyanate to blocking group and the isocyanate, which reacts
with the hydroxyl (i.e., elimination–addition), and (2) addition of the alcohol to the blocked
isocyanate to yield a tetrahedral intermediate followed by elimination of the blocking agent
(i.e., addition–elimination). Both pathways are shown in Scheme 12.4, in which B—H represents the blocking group. Evidence has been presented for both pathways, with elimination–addition being favored by increasing temperature, in accordance with a consideration
of the kinetic parameters Ea and A, as discussed in Section 2.3.2 and illustrated in Figure 2.6.
At storage temperatures the equilibrium is almost entirely on the side of the blocked
isocyanate, but as temperature increases, the equilibrium shifts to the right, liberating isocyanate to cross-link with the co-reactant. The temperature at which deblocking is rapid
depends on the blocking group. It also depends importantly on the relative reactivity
and stability of the product of the reaction with the coreactant. Generally, if the final
product is more stable than the blocked isocyanate, the reaction will proceed more
rapidly and at a lower temperature than it will if the final product is less stable. The rate
of cross-linking also depends on how rapidly the blocking agent evaporates from the film.
Kinetic studies of the reactions of blocked isocyanates are complicated by the fact that
all, or almost all, of the reactions are reversible, and several side reactions are generally
possible. Even the simplest case of heating a blocked isocyanates alone is complicated
by possible side reactions of the isocyanates. At high temperatures, there are trimerization
Scheme 12.4
252
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
reactions of the isocyanate or reaction with the original blocked isocyanate to form, for
example, an allophanate or biuret. These in turn can also thermally decompose through
different kinetic pathways. When used with a polynucleophile, the liberated isocyanate
can, in addition, react with the urethanes or ureas produced.
The curing schedule required for cross-linking of blocked polyisocyanates with polynucleophiles is dependent on many variables, among them the following:
.
.
.
.
.
.
.
.
Structures of the isocyanate, blocking agent, and nucleophile
Relative rate of reaction of the nucleophile with the isocyanate compared to the
reverse reaction rate of the isocyanate with the blocking agent
Rates of diffusion and evaporation of the blocking agent, and film thickness
Polarity and hydrogen bonding potential of the reaction medium (solvents or
coreactant)
Concentrations of reactive groups
Type and concentration of catalysts
Extent of side reactions and whether they lead to cross-linking or termination
Extent of cross-linking required to achieve optimum film properties
Typically, these variables have to be evaluated for each application. A combination that
leads to a satisfactory cure for wire enamel in a few seconds at 2658C but may not be
appropriate for an automotive clear coat cured in 20 minutes at 1408C.
In view of the many variables, comparisons of published deblocking temperatures for
different blocked isocyanates must be made with care. One should be sure to compare data
obtained by the same method. Furthermore, there is not a threshold temperature; rather,
there is a determination of the temperature at which the analytical method used detects
some extent of the deblocking and/or cross-linking reaction.
There are many different blocking agents, isocyanates, catalysts, catalyst concentrations,
and potential cure temperatures, and times. Studying all of these variables and their interactions would take a prohibitive amount of time and effort by conventional approaches to
formulations even with statistical experimental design techniques. As discussed in Chapter
34, such systems are potential candidates for high throughput experimental procedures
carried out with combinatorial chemistry apparatus. See Ref. [54] for a discussion of the
background of such experiments and their potential applicability to the coatings field. An
example of such a study is given in Ref. [55]. In one set of experiments, four isocyanates,
five blocking agents, and an acrylic polyol were evaluated with various catalysts at 12 temperatures from 80 to 2008C. With 3,5-dimethylpyrazole as a blocking agent, DBTDL gave the
greatest catalytic effect (curing at 1108C) with all four isocyanates. With MEKO-blocked
HDI isocyanurate, DBTDL and titanium(IV) 2-ethylhexanoate gave approximately equal
results (curing at 1408C). Diethyl malonate–blocked HDI isocyanurate showed no catalysis
(cure temperature 908C). The 128 experiments required to obtain these results were
completed in one day.
One needs to remember that reported deblocking temperatures frequently depend on
the method of analysis, heating rate, and other variables. Different analytical techniques
can give different deblocking temperatures for the same sample.
1. The most common methods for determining deblocking temperature follow some
change in physical properties: for example, gel time, the time required for a mixture of
12.5. BLOCKED ISOCYANATES
253
blocked isocyanate and coreactant to gel at a given temperature [56]. Another popular
method is to determine the time required for a coating film to achieve solvent resistance
sufficient to withstand 200 methyl ethyl ketone double rubs [57].
2. The IR spectra of isocyanates show a characteristic strong absorption band near
2250 cm21; deblocking temperatures have been reported as the temperature at which
this absorption is first detected. The result depends on the rate at which the sample is
heated; faster rates of heating give higher apparent deblocking temperatures. Fourier
transform infrared spectroscopy (FTIR) has been used to determine reaction kinetics
from a single dynamic temperature scan [58].
3. Use of FTIR in combination with dynamic mechanical analysis (DMA) is particularly effective in studying cure mechanisms by allowing the individual deblocking and
cross-linking reactions to be observed during the curing process [59]. FTIR has also
been used to determine the composition and rate of evolution of gases during cure; the
data on blocking agent evolution can be used to calculate kinetic constants.
4. Isothermal thermogravimetric analysis (TGA) has been used to determine kinetic
parameters for deblocking reactions [60]. The extent of reaction is followed by tracking
weight loss due to the release of the blocking agent.
5. The changes in heat flow associated with deblocking as measured by differential
scanning calorimetry (DSC) have been used to study the kinetics of the reaction [61].
6. The chemical species can be tracked directly using solid state NMR; solid state
13
C NMR can track the intensity of carbons within the blocking group, which become
significantly more intense after the deblocking reaction [62].
7. Since the reaction of amines with oxime-blocked isocyanates is first order, rate constants for deblocking of blocked isocyanates can be determined by following the rate of
reaction with amines [69].
8. The reaction of isocyanates with water leads to formation of CO2; blocked
isocyanates are heated in the presence of molecular sieves saturated with moisture,
and the lowest temperature at which CO2 is evolved is reported as the deblocking
temperature [63].
Cross-linking may continue after 200 MEK double rubs has been reached, having little
effect on MEK rub resistance but leading to further changes in the mechanical properties
of films. Such further cross-linking can be followed by film properties, changes in dynamic
mechanical analysis (DMA) results, and/or resistance to swelling by solvents. Evidence
that a 200 MEK double rub test does not indicate full cure (or even give the same rank
order of time to reach 200 double rubs as the time to reach full cure) is given in
Ref. [5] by comparisons of these results with the measurement of cross-link density by
DMA analysis.
Reaction rates and extents of reaction can be very dependent on volatilization of the
blocking agent. Care must be taken in making measurements to ensure that the measurement conditions reflect the use conditions, including film thickness and the ability of the
blocking agent to evaporate. When a sample of MEKO-blocked HDI isocyanurate was
heated at 1408C while taking a time resolved series of IR spectra, the spectra showed
that the NCO peak at 2250 cm21 grew steadily. However, when the same blocked isocyanate was heated in a thin film but in a covered ATR cell, there was insignificant generation
of isocyanate until after removal of the cover slide. Immediately the peak at 2250 cm21
began to increase [64].
254
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
This inhibition of deblocking reactions has clear implications for studying blocked
urethane systems in coatings. A related effect should be expected from increased film
thickness. In this case, slow diffusion of the blocking group through the reactive
medium could result in slower cure. Thus, analytical techniques used for tracking the
deblocking kinetics in thin films should (1) allow for evaporation of the blocking group
and (2) be of a film thickness comparable to that used in the final application. Not all
systems that use blocked isocyanates require that the blocking group leave as it does in
films. Studies related to shelf stability, adhesives, and thick section plastics require retention of the blocking group to predict field use performance.
In general, blocked aromatic isocyanates deblock at lower temperatures than blocked
aliphatic isocyanates. This results from the p-conjugation of the aromatic ring with the
isocyanate. Substitution of the aromatic ring with electron-withdrawing groups such as
Cl, NO2, and COOR increase deblocking rates, whereas electron donor groups such as
alkyl groups decrease deblocking rates. There are also steric effects on deblocking
rates. Blocked aliphatic isocyanates with the isocyanate group on a tertiary carbon (e.g.,
MEKO-blocked TMI) deblock at a lower temperatures [65]. MEKO blocked on the
ortho position of TDI deblocks 3.2 times faster than the less sterically crowded para
isomer at 1108C; MEKO-blocked TMI deblocks 7.9 times faster than less sterically
crowded MEKO-blocked cyclohexyl isocyanate at 1108C; and bis(isopropylketoxime)blocked TMI deblocks 5.5 times faster than MEKO-blocked TMI at 1108C [66].
12.5.2. Blocking Groups
The structures of blocking groups have a major effect on deblocking temperatures and cure
rates of coatings.
Many alcohols have been used as blocking agents; generally, they give high deblocking
temperatures; but because of low reactivity, alcohol-blocked isocyanates have excellent
stability in waterborne coatings. The stability is a major reason that 2-ethylhexyl
alcohol was once used widely in electrodeposition primers (E-coats), where long-term
tank stability is essential. The need to reduce curing temperatures for cationic E-coat
primers has led to the use of monoethers of ethylene glycol such as 2-butoxyethyl
alcohol [67].
Phenol-blocked isocyanates deblock at lower temperatures than alcohol-blocked
urethanes, in line with the slower rate of the forward reaction. Various alkyl phenols
are also used. The blocking agent can be designed to have more than one function. For
example, the reaction of 2-[(dimethylamino)methyl]phenol with HDI isocyanurate gives
a blocked isocyanate that is also water-dispersible as its amine salt [68].
Oximes have been widely used due to their low deblocking temperatures compared to
alcohols, phenols, and caprolactam. They have been the most favored of blocking groups
for aliphatic polyisocyanates in solvent borne coatings and were the first to see use in automotive top coats. A variety of potential structures are available, since almost any aldehyde
or ketone can be reacted with hydroxylamine to form the oxime.
Among the advantages of the oxime groups is their high reactivity toward isocyanates,
which allows the blocked products to be readily made without catalyst. The reaction rate
12.5. BLOCKED ISOCYANATES
255
of MEKO with cyclohexyl isocyanate was found to be between 50 and 75 times faster than
octanol with the isocyanate [69]. However, the high reactivity of MEKO is disadvantageous with regard to the deblocking reaction, because it competes more effectively than
the polyol reactant for the intermediate isocyanate, thereby slowing the forward reaction;
there is also concern about possible toxic hazards of MEKO.
Introduction of an electron withdrawing group in the alpha position of an oxime group
is reported to lower deblocking temperatures [70]. Oximes of alkyl esters of pyruvic acid –
blocked isocyanates have lower deblocking temperatures than MEKO and do not yellow
when overbaked. Coatings cured with these oximes have good acid resistance. The esters
are said to be non-toxic.
Caprolactam has been used widely as a blocking agent, particularly in powder coatings.
It has a high deblocking temperature compared to oximes, and 2-pyrrolidone (also called
pyrrolidinone) deblocks at even higher temperatures [71]. The bond angles in the sevenmember strained ring of caprolactam may result in easier formation of a six-member ring
between the urethane hydrogen and the carbonyl oxygen of the amide and possibly in a
greater steric effect than with the five-member ring of pyrrolidone. The level of volatiles
from blocked isocyanates can be quite substantial. In the case of IPDI monomer blocked
with caprolactam, 50% of the adduct’s mass will be lost through volatilization of the
blocking group. Furthermore, caprolactam condenses in cooler parts of the oven, requiring
extra maintenance.
Noncyclic acylureas are also used. Acetanilide-blocked HDI isocyanurate has been
reported to have a deblocking temperature of 1008C, compared with the corresponding
MEKO-blocked isocyanate deblocking temperature of 1308C [72]. Deblocking is promoted by having the carbonyl oxygen in a position to form an intermediate six-member
ring with the proton on the nitrogen from the isocyanate.
Pyrazoles [73] and 1,2,4-triazoles [74] deblock at relatively low temperatures, which is
probably facilitated by hydrogen-bonding, as illustrated below for 3,5-dimethylpyrazole.
Consistent with this hypothesis, the deblocking reaction is enhanced by increasing basicity
of pyrazoles. Hydrogen-bonding is expected to enhance elimination of the pyrazole by a
cyclic proton transfer mechanism. The addition of DABCO interferes with this protonation. 3,5-Dimethylpyrazole and 1,2,4-triazole also have the advantages of less yellowing
than oximes and have not been reported to have any toxic hazards. Furthermore, in contrast to phenol-blocked isocyanates, the cross-linking reactions are catalyzed by DBTL. It
256
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
is suggested that the reaction proceeds by the elimination – addition pathway.
Secondary amines are useful blocking agents. However, primary amines are not. The
urea from a primary amine and an isocyanate can cleave on either side of the central
C55O group; cleavage to form the primary amine derived isocyanate releases a volatile,
toxic monoisocyanate.
A study of a large number of hindered amines as blocking agents for powder coatings
discloses that the most preferred amine is bis(2,2,6,6)-tetramethylpiperidinyl)amine, since
it remains in the film to a large extent after baking and contributes to the UV stabilization
of a coating [75]. It has been reported that highly hindered amines such as t-butylbenzylamine can be used as a blocking agent in coatings for plastics that cure at 1208C.
In contrast, DMP used as a blocking agent required 1358C [76].
Any active methylene compound will react with an isocyanate by addition to the carbon
of the active methylene group. By far the most useful active methylene compound for
blocking is diethyl malonate (DEM). Although classified as blocked isocyanates, DEMblocked isocyanates do not unblock in the usual way to give free isocyanate. Instead, a
model compound study showed that reaction of DEM-blocked cyclohexyl isocyanate
with n-hexyl alcohol yields primarily transesterification products, as well as products
derived from cleavage of the CO22NH amide bond (plus a small amount of dicyclohexyl
urea resulting from reaction of cyclohexyl amine and the small amount of cleavage to
isocyanate) (Scheme 12.5) [77]. It was proposed that the enol form dissociated to give
ketenes, which formed the various products (Scheme 12.6). Suppliers have recommended
that formulations be based on a ratio of one blocked isocyanate to one coreactant group.
However, there are potentially two cross-link sites for each isocyanate group, discounting
side reactions.
There is a marked advantage in using a monoalcohol as part of the solvent in a diethyl
malonate coating since this provides good package stability by providing excess monoalcohol for transesterification, as discussed in Section 2.3.2. Under storage conditions,
the monoalcohol minimizes molecular weight buildup due to cross-linking, but when
the alcohol evaporates after application, the cross-linking reaction proceeds rapidly.
Package stability is also said to be improved by neutralizing the NaOCH3 catalyst used
in carrying out the blocking reaction with acids such as p-toluenesulfonic acid or
2-ethylhexyl acid phosphate [78].
12.5. BLOCKED ISOCYANATES
257
Scheme 12.5 Products of diethyl malonate– blocked cyclohexyl isocyanate reacting with hexyl alcohol.
Although the cure temperatures are low, hydrolytic stability is adequate to use
malonate-blocked isocyanates in 1K waterborne coatings [79]. The active hydrogen of
DEM-blocked isocyanates also reacts with MF resins [80]. Clear coats for automobiles
have been formulated with a combination of a hydroxy-functional acrylic resin,
DEM-blocked HDI and IPDI trimers, and an MF resin [81]. Cross-linking results from
reactions of the DEM-blocked isocyanate with hydroxyl groups, MF resin cross-linking
of hydroxyl groups, and MF resin reaction with the malonic ester blocking groups.
There has been a trend to replace conventional blocked isocyanates in powder coatings
with compounds that do not release a volatile blocking group on curing. Uretdioneblocked isocyanates are attractive because they do not generate volatile blocking
groups. Because of their superior exterior durability, there is more interest in uretdiones
from aliphatic isocyanates. When one attempts to make HDI uretdione, a major portion
of the HDI is converted to isocyanurate. However, with IPDI it is possible to prepare
the uretdione with only minor amounts of isocyanurate [82]. The free isocyanate groups
on the uretdione must be reacted before use; reaction with polyols have been recommended for IPDI uretdione [83]. IPDI uretdione is at best only difunctional, so that
Scheme 12.6 Proposed ketene formation of maleate-blocked isocyanate.
258
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
cross-linking cannot be achieved with a difunctional coreactant. It can be reacted with a
mixture of diols and triols to increase average functionality in the range 2.3 to 2.5 [82].
A mixture of IPDI and HDI can be dimerized; in the process some of the HDI is dimerized
but some is trimerized; thus, the overall product has a functionality somewhat over 2, and
the final films can be more flexible because of the hexamethylene groups [82].
A different approach to 1K urethane coatings is the use of encapsulated particles of
isocyanates. These materials have been surface reacted so that they are insoluble in the
rest of the vehicle at storage temperatures but dissolve in the coating during heating,
releasing free isocyanate that reacts with a hydroxy-functional polymer. An approach to
carrying out this idea is to make a fine particle size dispersion of isocyanate in water, in
which surface isocyanate groups react with water to form a urea rich surface. The
powder is filtered and dried, then dispersed in a coating [84]. Various methods have
been used to make such particles. TDI isocyanurate dispersed in dioctyl phthalate with
a small amount of 4,40 -diamino-3,30 -dimethyldicyclohexylamine added gives particles
of about 10 mm that are used in plastisols [85]. TDI dimer dispersed in polypropylene
glycol with a small amount of 2,5-dimethyl-2,5-hexanediamine gives a dispersion that
is stable for over 3 months and cures at 1208C [86].
12.5.3. Catalysis of Blocked Isocyanate Coatings
Catalysts are usually included in blocked isocyanate formulations but commonly without
consideration of what reaction or reactions they are involved in. Typically, the same
catalysts that are used in unblocked 2K polyurethanes are used with the blocked
systems, although at higher levels. As indicated earlier, many reactions are taking place
and the catalyst can be involved in any one or more of these: the deblocking reaction,
and/or the reaction of the free isocyanate with the other nucleophile, and/or an
addition –elimination reaction, and/or side reactions.
Carlson et al. studied the effect of dibutyltin dilaurate (DBTL) on the deblocking of
IPDI-based cross-linkers without a coreactant. In this case the DBTL showed no catalytic
effect on the actual deblocking reaction but did, over a longer time, lead to a decrease in
isocyanate concentration as a result of side reactions [87]. DBTL can catalyze the reaction
of the unblocked isocyanate and a hydroxyl group and can be a mild allophanate catalyst.
A series of catalysts were studied in the reaction of MEKO-blocked HDI isocyanurate with
a hydroxy-functional acrylic resin [88]. The films were intentionally undercured (20 minutes
at 1308C) to emphasize the differences in catalytic activity. Several other blocked isocyanates
were also tested, and the authors concluded that of the catalysts tested, bismuth tris(2-ethyl
hexanoate), cobalt bis(2-ethyl hexanoate), and Ti(AcAc)4 gave the most catalytic response.
A study of the catalytic effect of DBTDL on the deblocking temperatures of an IPDI
isocyanurate/triol coating with a series of blocking agents used DMA to track the
12.6. MOISTURE-CURABLE URETHANE COATINGS
259
TABLE 12.3. Minimum Temperatures (88 C) for a Detectable Reaction of Blocked IPDI
Isocyanurate with TMP Polyether Triol with and Without DBTDL Catalyst
Blocking Group
Caprolactam
MEKO
1,2-Pyrazole
Diisopropylamine
3,5-Dimethylpyrazole
With DBTDL
Without DBTDL
163
137
122
115
112
175
156
144
136
158
minimum temperatures at which cross-linking is detectable. As seen in Table 12.3, there
were significant differences between the different blocking groups’ response to catalysis
[89]. In the same study, diethylmalonate showed the same temperature (1048C) with
and without catalyst.
Catalysis of reactions of blocked isocyanate reactions with polyols in waterborne systems
requires special considerations that have not been widely recognized except in cationic
E-coats. It is well known that carboxylic acids inhibit catalysis by organotin compounds,
but most waterborne systems include coreactants that are dispersed in water by amine salts
of carboxylic acid groups. Blocked isocyanate anionic coatings are catalyzed to a lesser
degree by DBTDL than the corresponding 2K solventborne coating and require higher
curing temperature (or longer time) to reach the same extent of cross-linking [90]. The cure
rate of such coatings was affected by the acid number of the coreactant coatings with
higher acid number coreactants curing more slowly with the same level of catalyst. Another
factor here is the hydrolytic stability of the catalyst. DBTDL is relatively easily hydrolyzed
and hence is a questionable choice, although it is commonly reported as the catalyst.
12.6. MOISTURE-CURABLE URETHANE COATINGS
Moisture-curable urethane coatings cross-link by reaction of isocyanate resins with atmospheric water. Not only are they called moisture-curable but also commonly moisture-cure
and occasionally, hydrocure, moisture-hardening, or moisture-tempered. They are onepackage (1K) formulations that can cure at temperatures as low as 08C. Such coatings
are based mostly on isocyanate-terminated resins made from hydroxy-terminated polyesters or polyethers by reacting the terminal hydroxyl groups with excess polyisocyanate.
To minimize the amount of unreacted diisocyanate, the resin is prepared utilizing an
NCO/OH ratio that is significantly less than 2 : 1. Ratios higher that 2 : 1 are used with
oligomeric MDI since the toxicity of MDI is low and lower viscosity resins can be obtained.
Almost any type of polyol can be used as the backbone of moisture-curable resins,
although if functionality is too high, it becomes difficult to avoid gelation. Polyether
polyols are widely used; the ether backbone is completely stable to hydrolysis; the
resins tend to have low Tg, thereby enhancing flexibility; and they have relatively low
costs. However, their exterior durability is relatively low. Polyester polyols are widely
used, although the coatings may have poor saponification resistance. Depending on the
components, resins with a broad range of Tg are available. Polycarbonate polyols give
resins with hydrolytic stability superior to polyester polyols. Block copolymers of
polyols and siloxane segments are prepared by reacting excess isocyanate with a combination of polyols and hydroxy-terminated siloxanes [91]. These resins cure by reaction of
terminal isocyanate groups with water.
260
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
Hydrolytic resistance of coatings is affected by several variables, including cross-link
density and free volume availability as well as the functional groups present. Model compound studies suggest that substituted ureas are somewhat more easily hydrolyzed under
both neutral and acidic (pH 1.34) conditions than are analogous urethanes [92]. Ureas and
urethanes from aliphatic isocyanates are somewhat more resistant to hydrolysis than those
from aromatic isocyanates. Both urea and urethane groups are generally more stable to
hydrolysis than are unhindered ester groups.
The coatings are stable when stored in the absence of water but cross-link after application by reaction of the isocyanates with water to form amines, which react with other
isocyanates to form substituted urea cross-links. Cure rates depend on the water content
of the air; at low temperatures, higher relative humidity is required than at higher temperatures, since relative humidity increases with decreasing temperature for the same
water content. At high humidity and temperature, cure is rapid, but the carbon dioxide
released by the reaction of isocyanate with water can be trapped as bubbles, especially
in thick films. (See Ref. [93] for a discussion of the effects of temperature and humidity
and other application considerations.)
Although the most common moisture-curable urethanes are based on the direct reaction
of NCO groups and water, several other approaches are also used. Oxazolidines hydrolyze
with water to yield free amine and hydroxyl groups: These in turn react with isocyanates to
form the expected urea and urethane linkages. Difunctional oxazolidines yield higher
functionality [94]. They are, in effect, blocked amines.
Other “blocked amine” reactants have been used in preparing moisture-curable
systems. In general, such systems have the advantage of faster cure than obtained with
isocyanate-terminated urethanes. When exposed to water, imines act as blocked amines
because they hydrolyze to yield free amines, which react with an isocyanate. For
example, a urethane from a polyester diol, polyester polyol, and IPDI with the diketimine
from isophoronediamine and methyl isobutyl ketone (MIBK) is dry-to-sand in 30 minutes
and through dry in 60 minutes [95]. Aldimines give more stable systems. For example, a
moisture-curable coating is based on an MDI prepolymer and the aldimine from isobutyraldehyde and bis(4-aminocyclohexyl)methane [96]. Isocyanate-terminated urethane resins
can be partially reacted with a hydroxy-functional oxazolidine. For example, a coating
resin with excellent color and exterior durability is made from caprolactone polyester
triol, tris(2-hydroxyethyl)isocyanurate, IPDI, and 2-hydroxyethyloxazolidine [97]. After
application of such a system, moisture hydrolyzes the oxazolidine to an aminoalcohol
that reacts more rapidly with isocyanate groups than water reacts.
The isocyanate of a moisture-curable urethane can be reacted with 3-aminopropyltriethoxysilane to give a resin with terminal triethoxysilyl groups that will also moisture
cure. An advantage of the triethoxysilyl resin is that part of the solvent can be ethyl
alcohol. The ethyl alcohol permits a reasonable pot life in the presence of water, such as
comes into a coating from pigment surfaces. If free isocyanate groups are present, the
coating is not stable in the presence of either ethyl alcohol or water. The trialkoxysilyl
12.7. WATERBORNE URETHANE COATINGS
261
approach also avoids the formation of CO2, which can lead to film imperfections, such as
pinholing, in moisture-cure urethanes. (See Section16.2 for further discussion.)
12.7. WATERBORNE URETHANE COATINGS
12.7.1. Polyurethane Dispersions
Polyurethane dispersion resins are predominantly linear or lightly branched relatively high
molecular weight polyurethane polymers dispersed in water. Such materials are latexes,
but they are called polyurethane dispersions (PUDs). Since urethanes hydrogen-bond
strongly with water, the particles are swollen with water, plasticizing the polymer,
thereby permitting film formation with higher Tg polymers than is possible with acrylic
latexes [98]. It is common practice for formulators to blend PUDs with less expensive
azrylic latexes to upgrade film properties (Section 12.7.2).
PUDs are made with aliphatic diisocyanates, polyols, and a monomer designed to
provide stability to the dispersion in water. To increase molecular weight they are chain
extended after dispersion in water. The chain extension can be accomplished by allowing
the terminal NCO groups to react with water or by adding a diamine to the initial dispersion that will react more rapidly than water with the isocyanate groups.
A variety of polyols are used. Polypropylene glycol has the advantage of imparting
flexibility, hydrolytic stability, and low cost, but the films tend to have poor adhesion,
poor exterior durability, and poor solvent resistance. Polytetramethylene glycol gives
improved solvent resistance but at higher cost. Polyesters have the advantages of adhesion,
toughness, abrasion resistance, and exterior durability and can be designed to give a wide
variety of mechanical properties, but they suffer from limited hydrolytic stability. Polycarbonates provide hydrolytic stability, flexibility, and abrasion resistance but at higher cost.
2,2-Dimethylolpropionic acid (DMPA) is used as an ionic monomer precursor; the carboxylic acid groups are neutralized with a tertiary amine before dilution of the prepolymer
with water to impart dispersibility. The hindered carboxylic acid group on DMPA is almost
completely unreactive with isocyanates, so that the reaction is limited to the hydroxyl
groups. Carboxylic acid groups also serve as sites for cross-linking by H-bonding as well
as by covalent bonding. Polyglycol monoethers are used as nonionic monomers for stabilizing the dispersion. They give lower viscosity prepolymers; and the dispersions are more
freeze–thaw stable as well as more stable to changes in pH and soluble ion concentration.
However, films are more hydrophilic and exterior durability is poor.
One way to prepare PUDs is by the prepolymer mixing process. A polyol and DMPA
are reacted with excess diisocyanate to form an isocyanate-terminated prepolymer. A tertiary amine is added, the solution is added to water, leading to a dispersion of polymer
particles, and the polymer is chain extended by reacting with a diamine, such as hydrazine
or ethylenediamine, giving a dispersion of a higher molecular weight polymer [99]. The
dispersing step must be carried out in a short time at temperatures low enough so that
the NCO groups do not react appreciably with water. The amine (or hydrazine) reacts
much more rapidly than water. Aromatic isocyanates react with water too rapidly to be
used in this process; generally, cycloaliphatic isocyanates such as IPDI are used. Monomeric HDI is not used since it is somewhat soluble in water, resulting in the formation
of precipitates outside the dispersion particles. The final polymer contains carboxylic
acid salts of the tertiary amine. Since the salts are very hydrophilic, they concentrate at
262
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
the surface of the dispersed particles, stabilizing them against coalescence. Molecular
weight is controlled by the original ratio of isocyanate to hydroxyl groups and by the
ratio of extending diamine to terminal isocyanate groups. Frequently, the reaction is
carried out with a solvent, N-methylpyrrolidone (NMP) most commonly, to dissolve
DMPA and to reduce the viscosity for ease of dispersion.
Lower viscosity prepolymers for PUDs can be made using a two-step process. First a
linear isocyanate-terminated polyurethane is made from a diisocyanate, a diol, and
DMPA; in the second step the terminal isocyanate groups are reacted with a triol [100].
A variety of other techniques for making PUDs have been used [99].
PUDs with polyester backbones and DMPA are subject to hydrolysis, which can lead to
breaking the polymer backbone and separation into two phases, especially at elevated
temperatures [101]. The stability is lowest for PUDs not chain extended; as the amount
of chain extension (with ethylenediamine) is increased, stability improves. The effect of
using different chain extenders, including tri- and tetrafunctional amines, on polyester –
DMPA PUD properties, including hydrolytic stability, has also been studied [102].
Hydrolytic stability, MEK rub resistance, and peel strength of films are higher using,
for example, melamine as chain extender rather than ethylenediamine; viscosity is
said to be unaffected by changing to trifunctional amine from difunctional amine extender.
One can incorporate a small amount of trifunctional isocyanate into an isocyanateterminated polymer, resulting in some branching on dispersion in water. Also, one can
chain extend with diethylene triamine to introduce branching or a low degree of crosslinking—low enough that the particles can still coalesce to a continuous film.
Cross-linkable polyurethane dispersions are also prepared. Diethanolamine or ethanolamine are added to the water in which the prepolymer is added, resulting in terminal
hydroxy groups that can be cross-linked with MF resins or polyisocyanates. Polyunsaturated polyurethane dispersions (PUDs) are also being used (Section 15.7.2).
12.7.2. Acrylic – Polyurethane Dispersions
There have been three major approaches to preparing aqueous acrylic – urethane systems:
(1) blending of PUDs and latexes (or water-reducible acrylic resins); (2) polymerizing
vinyl monomers, such as acrylates, in PUDs to make hybrid acrylic – urethane resins;
and (3) and cross-linking acrylic –urethane systems.
When combined aqueous polyurethane dispersions (PUDs) and acrylic latexes offer
several potential advantages compared to each of the two separate types of polymers. In
general terms, polyurethane based polymers offer superior abrasion resistance due to intermolecular hydrogen bonding. Acrylic resins have lower costs, due to both raw material
cost and lower processing costs. Many latexes contain surfactants that can bloom to the
surface of films reducing gloss, whereas PUDs can be made without surfactants. PUDs
have a low minimum film formation temperature (MFFT) relative to their dry film Tg
owing to the plasticizing effect of hydrogen bonding with water. Also, owing to hydrogen
bonding of urethane groups, film properties are not as dependent on Tg as in acrylic
latexes. Accordingly, low MFFT can be achieved without reducing hardness. For
example, a paint made with an acid-functional acrylic latex and a PUD with a Tg of
2408C has a longer open time, permitting rebrushing to assure hiding and lapping
[103]. Coalescing solvents are required to permit film formation with latexes at low temperatures, whereas they are not required with PUDs. Films of PUDs alone have higher
moisture vapor transmission than that of films from many acrylic latexes.
12.7. WATERBORNE URETHANE COATINGS
263
Although blends of PUDs with relatively inexpensive latexes can offer significant cost
reductions and improvement in some properties, improvement of latex properties is not
directly proportional to the ratio of the two products. For example, blending equal parts
of an acrylic latex and a PUD gives films with a tensile strength of 2900 psi, compared
to 2000 and 6400 psi for the individual films, respectively, corresponding to an average
of 4200 psi. Similar nonlinear effects were observed with other mechanical and chemical
properties. It is suggested that inhomogeneity caused by separate acrylic and urethane
domains contributes to the less than expected improvement in properties. These areas
may have excessive internal strains and/or incomplete coalescence, which leads to
increased permeability and reduced cohesive strength [100].
Better film properties can be obtained by polymerizing the acrylate esters in the urethane.
The earliest approach was to use the particles of the PUD as “seed” for polymerization of the
acrylates. The effect of various reaction parameters on the rate of polymerization of MMA,
BA, and a 50 : 50 mix of MMA and BA in an anionic polyester PUD, dissolved in
N-methlpyrrolidone (NMP), has been reported [104].
To eliminate the NMP, another procedure for preparing hybrid polymers has been
developed in which acrylic monomers are used as solvent. For example, a PUD,
derived from poly(NPG adipate) diol, DMPA, and H12MDI, is dissolved in BA,
styrene, and 1,6-hexanediol diacrylate, neutralized with TEA, dispersed in water, chain
extended with ethylenediamine, and then polymerized with AIBN [105].
12.7.2.1. Cross-Linked PUD – Acrylate Systems There are two types of crosslinked acrylic – urethane dispersions: (1) reaction of a functional group on one or both
of a urethane and a vinyl resin that reacts with another functional group on the other
resin; and (2) use of a separate cross-linker that can react with functional groups on
both the urethane and the acrylic resin.
Several compositions and applications follow. A UV cure vehicle for wood coatings is
prepared from a carboxy-functional polyol, adding glycidyl methacrylate (GMA) to react
with COOH groups, adding DMPA and HDI, neutralizing with TEA and diisopropylethylamine, dispersing in water, chain extending with ethylenediamine, and adding a photoinitiator [106].
An autoxidizable coating is based on a water-reducible acrylic resin with a PUD made
from a soy alkyd with HDI, neutralized with NH4OH, and dispersed in water [107].
Hydrazine-terminated PUDs have been used with keto-functional latexes to make crosslinking systems. A study of cross-linking of blends of a keto-functional acrylic latex
and hydrazine-terminated PUD has been published [108]. When the water evaporates
from the film after application, cross-linking occurs by reaction of the hydrazine groups
with keto groups to form hydrazone linkages.
External cross-linkers are used to react with functional groups on acrylic – PUDs. For
example, a chip resistant automobile primer is formulated with an H12MDI/HDI/polyester polyol PUD, a carboxy- and hydroxy-functional acrylic latex, 2-amino-2-methylpropanol, and an MF resin [109].
Polyepoxides are used as cross-linkers. A top coat with exterior durability for application over stains on plastic composite surfaces is formulated with a urethane – acrylic
hybrid PUD, an acrylic resin, and a methylpolysiloxane microemulsion. An aqueous aliphatic epoxy emulsion, which cross-links by reacting with COOH groups on the PUD, is
added shorthly before application. The pot life of the combined system is over a
week [110].
264
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
12.7.3. 2K Waterborne Urethanes
Since isocyanates react relatively readily with water, it was assumed for many years that
they could not be used directly in waterborne coatings. By the late 1980s this assumption
was in question and development of 2K waterborne coatings began. In the last few years,
2K waterborne systems have been adopted commercially on a large scale [111]. Bayer
Corporation received a Presidential Green Award in 2000 for its work on these systems.
Cross-linkers are polyisocyanates and to a lesser extent, polycarbodiimides and polyaziridines. See Ref. [112] for an extensive review of 2K urethane systems.
12.7.3.1. Components of Waterborne 2K Coatings Two types of polyisocyanate
cross-linkers are used: conventional polyisocyanates and polyisocyanates that have been
modified to be more hydrophilic. In general terms, hydrophilically modified isocyanates
are easier to incorporate into a coating, but unmodified isocyanates are less expensive
and give films that have better exterior durability. Modifying isocyanates so that they
are more readily dispersible in water permits easier mixing of the two packages. Nonionic
hydrophilically modified polyisocyanates made by reacting a fraction of the NCO groups
on a polyisocyanate (such as HDI or IPDI isocyanurate) with a polyethylene glycol monoether are more easily stirred into an aqueous dispersion of a coreactant to form a heterogeneous dispersion, in which the polyisocyanate and coreactant are in separate dispersed
particles [113]. It is critical that the adducts be non-crystalline. The ease of mixing and
stability of the dispersions increases as the length or amount of the polyether modifier
is increased [114]. Adequate water dispersibility without crystallization can be achieved
using ethers with n . 5 but ,10 [115]. Use of a polyester or polyether as a modifier
for isocyanate overcomes the crystallization problem of polyethers and decreases the
water sensitivity of final films. For example, reacting a polyethylene glycol monomethyl
ether (n ¼ 7) with e-caprolactone gives a polyester or polyether that is used to modify HDI
isocyanurate [116].
Water sensitivity of films is reduced by partially reacting a hydrophilically-modified
polyisocyanate with an aminoalkyltrialkoxysilane. For example, HDI is reacted with the
monomethyl ether of a polyethylene glycol (n . 10) and an amine-functional silane
made by reacting diethyl maleate with 3-aminopropyltrimethoxysilane. Primers and top
coats formulated with trialkoxysilylated isocyanate and water-reducible acrylic resins
show substantial advantages in gloss retention and reduced blistering on water immersion
than corresponding films without the silane [117].
PUDs with polyester backbones and DMPA have been widely used. However, they are
subject to ester hydrolysis, or saponification at basic pH, which can lead to cleaving the
polymer backbone and separation into two phases, especially at elevated temperatures.
The hydrolysis is autocatalytic since carboxylic acids formed from ester hydrolysis catalyze the hydrolysis, as does DMPA. Chain extension with ethylenediamine improves stability, which supports acid catalysis. Polycarbonate diols are reported to be more resistant to
hydrolysis than are polyester diols [118].
Hybrid PUDs containing hydroxy-functional acrylic – polyurethanes have been
used [119], including their coreaction with a nonionic hydrophilically modified HDI
isocyanurate [120].
Acrylic latexes have also been used in 2K waterborne urethane coatings. A statistical
study of the effect of several variables on the performance of finishes for wood kitchen
cabinets, based on hydroxy-functional acrylic latexes cross-linked with a nonionic
12.7. WATERBORNE URETHANE COATINGS
265
hydrophilically modified polyisocyanate, has demonstrated that a high hydroxy content of
the latex (hydroxyl number 52), small particle size of the latex, core –shell preparation
of the latex, and low Tg (obtained by increasing level of coalescing solvent) enhanced
performance [121].
An extensive study of the effect of carboxylic acid structure and level in acrylic latexes
has been reported. It was shown that reaction of isocyanate groups with COOH groups
derived from acrylic acid resulted in undesirable bubble formation to a greater extent
than occurred with COOH groups derived from b-carboxyethyl acrylate. The latexes
were made by emulsion polymerization of methyl methacrylate, butyl acrylate, and hydroxyethyl methacrylate, together with acrylic acid, methacrylic acid, or b-carboxyethyl
acrylate, and neutralized with DMAE in 2K urethane coatings [122].
Isocyanates react with water as well as with the coreactant, which is usually hydroxyfunctional. Although the rate constants of reaction with water are lower than with primary
alcohols, the large excess of water means that a significant amount of the isocyanate will
react with water. Since the reaction with water also gives cross-links (urea instead of
urethane), the problem can be minimized by use of a large excess of isocyanate; this indexing is frequently a 2 : 1 or even higher isocyanate/hydroxyl ratio. However, urea linkages
can lead to turbidity in films, due to the relatively low compatibility and crystallinity of
urea compared to urethane groups in the polymer matrix, and the cost of the isocyanate
is almost always higher than that of the polyol. Also, reaction with water leads to the evolution of CO2, which may give foaming or bubbling of applied films. Hence, in most cases
it is desirable to reduce the reaction with water so that indexing need be only a little greater
than 1. The required excess of water is affected by the hydroxy-functionality, f̄n, of the
coreactant.
Relative reaction rates of NCO with hydroxyl groups and water depend on the catalyst.
Zr(AcAc)4 (King Industries, ZrCAT) catalyst shows better selectivity than DBTDL
[23,24]. See Section 12.2.2 for a further discussion of the effect of various catalysts on
the relative rates of reaction of isocyanates with water and alcohols.
12.7.3.2. Mixing and Application Considerations It is relatively easy to make 2K
waterborne coatings with good properties in the laboratory; however, production use is
more difficult. There are several potential problems. In some systems, it is difficult to
assure that uniform stoichiometric ratios are obtained throughout the film. If reaction
occurs to a significant extent before application, coalescence will be inhibited, and film
properties may therefore, be poor. As with solventborne 2K coatings, waterborne 2K coatings can be expected to have limited pot life. In solventborne coatings, pot life can be
determined by monitoring viscosity increases. In many 2K waterborne coatings, viscosity
does not increase as reactions between NCO and OH take place, since they change the viscosity inside the aggregate particles but not the bulk viscosity. In fact, the generation of
CO2 by the reaction of isocyanate with water decreases the pH, leading to shrinking of
the aggregate particles, which can result in viscosity reduction [123].
To minimize mixing problems, the two dispersions should start out with similar viscosities. Generally, it is desirable to use as low viscosity isocyanates as possible. The viscosity of HDI isocyanurate with a functionality of 3.3 is 1.7 Pa.s at 288C, compared to 8.5
Pa.s for HDI biuret with the same functionality, making it easier to disperse the isocyanurate [124]. The viscosity of a combination of HDI uretdione and isocyanurate is lower than
that of HDI isocyanurate by itself, which makes it easier to incorporate the combination
into dispersions [125]. Very low viscosity HDI unsymmetrical trimer (0.7 Pa.s) has
266
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
been recommended for use in 2K urethane coatings [36]. Use of a proprietary waterreducible acrylic resin with conventional HDI based isocyanates is reported to be satisfactory [126]. 4-Isocyanatomethyl-1,8-diisocyanatononane is reported to stir into water and
aqueous systems more easily than other polyisocyanates [127]. The polyisocyanate component is sometimes diluted with solvent such as butyl acetate to reduce its viscosity. Solvents such as cyclic carbonates or lactones give smaller particle size and more uniform
dispersions in water than those for butyl acetate [128]. Results also depend on the
process used to mix the two packages together. High but not excessive shear is required
to obtain relatively uniform particle size with an average diameter of about 150 nm [129].
In mixing two dispersions, one containing isocyanate, the other containing coreactant,
the maximum physical stability of the dispersion will generally be reached with the smallest particle size. However, a smaller particle size requires high intensity agitation to break
up the initial particles, which increases the possibility of contact of isocyanate groups with
water. The probability of new particles being formed that contain both reactants also
increases, resulting in non-uniform films, with some parts being high in isocyanate and
others high in coreactant. In making dispersions of hydrophilically-modified polyisocyanates in dispersions of water-reducible acrylic resins, it has been shown that bimodal particle size distribution of small particles and larger particles form at low shear rates; at high
shear rates there was a single broader small particle size distribution; with excess shearing,
the particle size increased and distribution broadened [130].
Carrying out the high intensity mixing in line with application can minimize the problems. Spray equipment designed to provide in-line intensive mixing of the two components has been used to apply polyisocyanate 2K water-reducible coatings [131]. To
assure proper mixing of 2K coatings, a specially designed jet mixer is recommended [132].
Performance can be affected by the rate of loss of water from films after application. If
water is lost slowly, there will be a larger extent of reaction of isocyanate, with water
requiring higher indexing to achieve full cross-linking of coreactant groups. Thicker
films tend to increase the relative reaction with water by decreasing the rate of water
loss through the film [133]. For ambient cure coatings, relative humidity affects the rate
of water loss as well as surface cure. Above about 70% RH, gloss is reduced due to
bubble formation caused by increased evolution of CO2. FTIR ATR analysis of films of
2K waterborne urethane coatings shows that reaction of isocyanate with water to form
urea groups increases near the surface as RH increases [134].
Baking coatings can generally be formulated with lower indexing than ambient cure
coatings because of the faster removal of water from the films. A large fraction of the
water evaporates during flash off and initial baking, reducing the extent of isocyanate reaction with water, thereby minimizing CO2 evolution and permitting lower NCO/OH ratios.
In one system, indexing of 1.3 to 1.8 for ambient cure coatings could be reduced to 1.1 to
1.3 for heat cure coatings [135].
12.8. HYDROXY-TERMINATED POLYURETHANES
Diisocyanates can be reacted with diols and triols at an N55C55O/OH ratio of less than 1
to make hydroxy-terminated polyurethanes, which can be cross-linked with MF resins or
with other cross-linkers that react with hydroxyl groups. Urethane groups also cross-link
with MF resins (Section 11.3.4). Compared to polyesters, hydrolytic stability is better and
coatings provide the toughness and abrasion resistance associated with urethane coatings.
REFERENCES
267
However, solids of equal molecular weight and viscosity are lower, due to intermolecular
hydrogen bonding, and the residual urethane groups increase moisture absorption. Compared to hydroxy-functional acrylic resins, they offer the advantage of abrasion resistance
and the possibility of using lower molecular weight resins. Whereas molecular weight
reduction of acrylic resins made by conventional free radical –initiated polymerization
is limited by the problem of ensuring that at least two hydroxyl groups are present on
each oligomer molecule (Section 8.2.1), hydroxy-terminated polyurethanes have two
(or more) terminal hydroxyl groups, even at very low molecular weight (as with polyesters). Furthermore, the urethane (carbamate) groups may participate in cross-linking
with amino resins, in effect increasing functionality.
REFERENCES
1. http://www.adhesivesmag.com (2003).
2. Nordstrom, J. D.; Dervan, A. H., Proc. Waterborne High-Solids Coat. Symp., New Orleans LA,
1993, p. 3.
3. Zweiner, C.; et al., Farbe Lack, 1991, 97, 1052.
4. Wicks, D. A.; Yeske, P. E., Prog. Org. Coat., 1997, 30, 265.
5. Higginbottom, H. P.; et al., Prog. Org. Coat., 1998, 34, 27.
6. Schwetlick, K.; et al., J. Chem. Soc. Perkin Trans., 1994, 599.
7. Sato, M., J. Am. Chem. Soc., 1960, 82, 3893.
8. Thiele, L.; Becker, R., in Advances in Urethane Science and Technology, Frisch, K. C.;
Klempner, D., Eds., Technomic, Lancaster, PA, 1993, Vol. 12, p. 59.
9. Hira, Y.; et al., Mater. Sci. Eng., 1983, 49, 336.
10. Wong, S. W.; Frisch, K. C., J. Polym Sci. A Polym. Chem., 1986, 24, 2867, 2877.
11. Schwetlick, K.; Noack, R., J. Chem. Soc. Perkin Trans., 1995, 2, 395.
12. Hazel, N.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1997,
p. 237.
13. Berlin, P. A.; et al., Kinet. Katal., 1993, 34, 640.
14. Yu, N.; et al., Kinet. Catal. (Engl. Ed.), 1995, 36, 612.
15. Nordstrom, J. D.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
1997, p. 70.
16. Yilgor, I.; McGrath, J. E., J. Appl. Polym. Sci., 1985, 30, 1733.
17. Van der Weij, F. W., J. Polym. Sci. A Polym. Chem., 1981, 19, 381.
18. Luo, S.-G.; et al., J. Appl. Polym. Sci., 1997, 65, 1217.
19. Draye, A.-C.; et al., Main Group Met. Chem., 1999, 22, 367.
20. Robins, J.; et al., Adv. Urethane Sci. Technol., 1984, 9, 65.
21. Throckmorton, P.; et al., U.S. patent 3,828,007 (1974).
22. Richter, R.; et al., U.S. patent 5,045,226 (1991).
23. Blank, W. J.; et al., Prog. Org. Coat., 1999, 35, 19.
24. He, Z. A.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1999,
p. 157.
25. Oberth, A. E.; Bruenner, R. S., Ind. Eng. Chem. Fundam., 1969, 8, 383.
26. Squiller, E. P.; Rosthauser, J. W., Mod. Paint Coat., 1987, June, 26.
27. Okumoto, T.; et al., Nippon Gomu Kyokaishi, 1995, 68(4), 244.
268
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
28. Bechara, I. S., in Urethane Chemistry Applications, ACS Symp. Ser. 172, American Chemical
Society, Washington, DC, 1981, pp. 393 – 402.
29. Yamada, M.; et al., Jpn. Kokai Tokkyo Koho JP 07,258,586 (1995).
30. Draye, A.-C.; Tondeur J.-J., J. Mol. Catal. A Chem., 1999, 140, 31.
31. Ohira, T.; et al., U.S. patent 6,180,714 (2001).
32. Kerimis, D.; et al., U.S. patent 4,521,545 (1985).
33. Palyutkin, V. G.; Zharkov, G. M., Kinet. Katal., 1985, 26, 476.
34. Korzyuk, E. L.; Zharkov, V. V., Kinet. Katal., 1981, 22, 522.
35. Richter, F.; et al., U.S. patent 5,717,091 (1998).
36. Richter, F.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2003,
pp. 71–85.
37. Potter, T. A.; Slack, W. E., U.S. patent 5,124,427 (1992).
38. Wojcik, R. T., Polym. Mater. Sci. Eng., 1994, 70, 114.
39. Ono, H.-K.; et al., J. Polym. Sci. C Polym. Lett., 1985, 23, 509.
40. Hatada, K.; et al., J. Polym. Sci. A Polym. Chem., 1990, 28, 3019.
41. Lomoelder, R.; et al., J. Coat. Technol., 1997, 69(868), 51.
42. Alexanian, V.; et al., U.S. patent, 5,254,651 (1993).
43. Fiori, D. E.; Dexter, R. W., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA,
1986, p. 186.
44. Wojick, R. T.; et al., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA, 1995,
p. 287.
45. Jorissen, S. A.; et al., Proc. Waterborne Higher-Solids Powder Coat. Symp., New Orleans, LA,
1992, p. 182.
46. Zhou, L.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2000,
pp. 262 – 281.
47. Mormilem, P. J.; et al., U.S. patent 5,214,086 (1993).
48. Bock, M.; Halpaap, R., J. Coat. Techol., 1987, 59(755), 131.
49. Takas, T. P., JCT Coat. Tech, 2004, 1(5), 40.
50. Naderhoff, B.; et al., U.S. patent 6,369,189 (2000).
51. Giles, D.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2003,
pp. 167 – 182.
52. Hansen, R. G.; et al., U.S. patent 6,005,062 (1999).
53. Wicks, Z. W., Jr., Prog. Org. Coat., 1975, 3, 73; 1981, 9, 3. Wicks, D. A.; Wicks, Z. W., Jr.,
Prog. Org. Coat., 1999, 148 – 172; 2001, 41, 1 – 85.
54. Wicks, D. A., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002, p. 1.
55. Bach, H.; et al., Farbe Lack, 2002, 108(4), 30.
56. Katsamberis, D.; Pappas, S. P., J. Appl. Polym. Sci., 1990, 41, 2059.
57. Huang, Y.; et al., J. Coat. Technol., 1995, 67(842), 33.
58. Carlson, G. M.; et al., in Advances in Urethane Science and Technology, Frisch, K. C.;
Klempner, D., Eds., Technomic, Lancaster, PA, 1984, Vol. 9, p. 47.
59. Kuo, C.; Provder, T., Polym. Mater. Sci. Eng., 1988, 59, 474.
60. Kordomenos, P. I.; et al., J. Coat. Technol., 1982, 54(687), 43.
61. Provder, T.; et al., in Analytical Calorimetry, Johnson, J. F.; Gill, P. S., Eds., Plenum,
New York, 1984, Vol. 5, p. 377.
62. Cholli, A.; et al., J. Appl. Polym. Sci., 1983, 28, 3497.
63. Mohanty, S.; Krishnamurti, N., Eur. Polym. J., 1998, 34, 77.
REFERENCES
269
64. Wicks, D. A.; Wicks, Z. W., Jr., Prog. Org. Coat., 1999, 36, 148 – 172.
65. Lucas, H. R.; Wu, K.-J., J. Coat. Technol., 1993, 820, 59.
66. Pappas, S. P.; Urruti, E. H., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA,
1986, p. 146.
67. Hoppe-Hoeffler, M.; et al., Ger. Offen. DE 3,938,883 (1991).
68. Nakabayashi, A.; Kamiyanagi, K., Jpn. Kokai Tokkyo Koho JP 02,304,056 (1990).
69. Witzeman, J. S., Prog. Org. Coat., 1996, 27, 269.
70. Bernard, J.-M.; Perroud, E., PCT int. appl., WO 97/24,386 (1997).
71. Regulski, T.; Thomas, M. R., Org. Coat. Appl. Polym. Sci. Proc., 1983, 48, 1003.
72. Muramatsu, I.; et al., Prog. Org. Coat., 1993, 22, 279.
73. Cooray, B.; Spencer, R., Paint Resin, 1998, October, 18.
74. Keck, M. T.; et al., U.S. patent 5,688,598 (1997).
75. Gras, R.; Wolf, E., Cana. patent appl. 2,186,089 (1997).
76. K. Martin, http://specialchem4coatings.com/resources/articles/article.aspx?id¼1764 (2004).
77. Wicks, Z. W., Jr.; Kostyk, B. F., J. Coat. Technol., 1977, 49(634), 77.
78. Usui, T.; Asahina, Y., U.K. patent appl. GB 2,311,993 (1996).
79. Daude, G.; Girard, P., U.S. patent 4,623,592 (1996).
80. Blank, W. J.; et al., Polym. Mater. Sci. Eng., 1997, 77, 391.
81. Rockrath, U.; et al., U.S. patent 5,516,559 (1996).
82. Schmitt, F.; et al., Prog. Org. Coat., 1998, 34, 227.
83. Guida, R. M., Mod. Paint Coat., 1996, July, 34.
84. Kopp, R.; et al., Angew. Makromol. Chem., 1994, 223, 61.
85. Hurnick, H.; et al., Can. patent 1,255,042 (1989).
86. Groegler, G.; et al., U.S. patent 4,483,974 (1984).
87. Carlson, G. M.; et al., Polym. Sci. Technol., 1987 36, 197.
88. Blank, W. J.; et al., Polym. Mater. Sci. Eng., 1998, 79, 399.
89. Engbert, Th.; et al., FATIPEC Congr. 23B, 1996, pp. B115 –B143.
90. Lomoelder, R.; Reichel, D., FATIPEC Congress Book, 1998, Vol. D, p. D-25.
91. Kozakiewicz, J., Prog. Org. Coat., 1996, 27, 123.
92. Chapman, T. M., J. Polym. Sci. A Polym. Chem., 1989, 27, 1983.
93. Gardner, G., J. Prot. Coat. Linings, 1996, February, 81.
94. Howartl, G. A., Proc. Waterborne Higher-Solids Powder Coat. Symp., New Orleans LA, 1996,
p. 92.
95. Nachtkamp, K.; et al., U.S. patent 4,481,345 (1984).
96. Yonek, K. P., U.S. patent 5,726,272 (1998).
97. Petschke, G. H., U.S. patent 4,471,102 (1984).
98. Noble, K. L., Prog. Org. Coat., 1997, 32, 131.
99. Satguru, R.; et al., J. Coat. Technol., 1994, 88(830), 47.
100. Hegedus, C. R.; Kloiber, K. A., J. Coat. Technol., 1996, 68(860), 39.
101. Lo, H. H.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans
LA, 1997, p. 192.
102. Li, H.-C.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 1997, p. 204.
103. Gray, R. T.; Lee, J., U.S. patent 6,303,189 (2001).
104. Kukanja, D.; et al., J. Polym. Sci., 2002, 84, 2639.
270
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
BINDERS BASED ON POLYISOCYANATES: POLYURETHANES
Vijayendran, B. R.; et al., U.S. patent 5,173,526 (1992).
Weikard, W.; et al., U.S. patent 6,521,702 (2003).
Schneider, V.; et al., U.S. patent 5,688,859 (1997).
Okamoto, Y.; et al., Prog. Org. Coat., 1996, 29, 175.
Gessner, M. A.; Kandow, T. P., U.S. patent 6,437,036 (2002).
Porter, C. A., U.S. patent 6,475,615 (2002).
Melchiors, M.; et al., Prog. Org. Coat., 2000, 40, 59.
Wicks, Z. W., Jr.; et al., Prog. Org. Coat., 2002, 44, 161.
Renk, C. A.; Schwartz, A. J., Proc. Waterborne High-Solids Powder Coat. Symp., New
Orleans, LA, 1995, p. 266.
Shaffer, M.; Bui, H., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
1998, p. 93.
Laas, H.-J.; et al., U.S. patent 5,252,696 (1993).
Laas, H.-J.; et al., U.S. patent 5,731,396 (1998).
Hovestadt, W.; et al., U.S. patent 5,854,338 (1998).
Zermani, J. T., J. Water Borne Coat., 1984, August, 2.
Hegedus, C. R.; et al., J. Coat. Technol., 1996, 68(852), 51.
Tien, C. T.; et al., U.S. patent 5,594,065 (1997).
Feng, S. X.; et al., J. Coat. Technol., 1999, 71(899), 51.
Nabuurs, T.; et al., Prog. Org. Coat., 1999, 35, 129.
Jacobs, P.; et al., Paint Coat. Ind., 1998, October, 116.
Reiff, H., U.S. patent 5,258,452 (1993).
Hombach, R.; et al., U.S. patent 4,663,377 (1987).
Bassner, S. L.; Hegedus, C. R., J. Prot. Coat. Linings, 1996, September, 52.
Richter, F.; et al., U.S. patent 6,100,326 (2000).
Weyland, P., U.S. patent 5,587,421 (1996).
Bui, H.; et al., Eur. Coat. J., 1997, 97, 476.
Dvorchak, M. J., J. Coat. Technol., 1997, 69(866), 47.
Dvorchak, M. J.; Bui, H., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 1998, p. 80.
Kahl, L.; et al., Eur. Coat. J., 1996, 96, 563.
Dewhurst, J. E.; et al., Polym. Mater. Sci. Eng., 1999, 81, 195.
Urban, M. W., Prog. Org. Coat., 2000, 40, 195.
Hegedus, C. R.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
1998, p. 391.
13
Epox y and Phenolic Resins
Epoxy and phenolic resins are two other classes of step-growth resins important in
coatings.
13.1. EPOXY RESINS
Terminology of epoxies can be confusing. Epoxy groups (also called epoxides) are threemembered cyclic ethers; in IUPAC and Chemical Abstracts nomenclature, they are called
oxiranes. Most commercially important epoxy resins are derived from (chloromethyl)
oxirane, more commonly known as epichlorohydrin (ECH). The resins generally
contain oxiranylmethyl ethers or esters, usually called glycidyl ethers or esters. In addition
to the uses discussed in this chapter, epoxy esters are discussed in Section 15.8 because of
their close relationship to the chemistry of alkyds; acrylated epoxy resins are discussed in
Section 29.2.4.
13.1.1. Bisphenol A Epoxy Resins
The first epoxy resins used in coatings, and still the largest volume, are bisphenol A (BPA)
epoxies made by reacting BPA with ECH. Under basic conditions, the initial reaction
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
271
272
EPOX Y AND PHENOLIC RESINS
is formation of a BPA anion (BPA2), which attacks ECH and results in the formation
of a new oxirane ring with elimination of chloride anion (Cl2), as shown in Scheme 13.1.
The initial product is the monoglycidyl ether of BPA (MGEBPA). Analogous reaction
of the phenolic group of MGEBPA with NaOH and ECH gives the diglycidyl ether of BPA
(DGEBPA). The epoxy groups of MGEBPA and DGEBPA react with BPA2 to extend the
chain, as shown for DGEBPA; these reactions introduce alcohol groups on the backbone.
Continuation of these reactions results in linear polymers, since both the BPA and ECH are
difunctional. Bisphenol A epoxy resins are made with excess ECH, so the end groups are
glycidyl ethers. The polymers may be represented by the following general formula, where
the molar ratio of ECH to BPA determines the average n value.
Molecular weight (MW) is controlled by the ratio of ECH to BPA. With a large excess of
ECH, it is possible to make a resin that is dominantly DGEBPA, that is, where n ¼ 0 in the
general formula. The pure n ¼ 0 compound is a crystalline solid, but the largest volume
commercial grades are liquids with n values of 0.11 to 0.15 (called standard liquid
resin). As the ratio of ECH to BPA is reduced, MW and the n value of the epoxy resin
increase. Viscosity also increases with MW. Above an average n value of 1, the resins
are amorphous solids with increasing Tg. Although the resins are said to have melting
points, they do not melt in the sense that a crystalline solid does. Rather, under specified
test conditions, the resins flow to some standard extent at the melting point (the term
softening point is sometimes used). Commercially available higher MW resins are often
designated as types 1001, 1004, 1007, and 1009. Table 13.1 gives average n values,
epoxy equivalent weights (EEW), and melting points for commercial BPA epoxy resins
[1]. EEW is also often called weight per epoxide (WPE). Different suppliers offer subtypes
and quote significantly different ranges for EEW and WPE of the various types. As MW
increases, EEW and average hydroxy functionality also increase. In some even higher
MW (n510) epoxy resins, the amount of epoxy groups present is so small that the resins
approach being just polyfunctional alcohols, commonly called phenoxy resins.
Although theoretically, there should be two epoxy groups on each molecule, a variety
of other end groups are present to a small extent. To a minor degree, unreacted phenol and
1,2-chlorohydrin terminal groups are present; additional non-epoxy end groups result
from side reactions. About 2% of the terminal groups are 1,2-glycol derivatives. These
can result from hydrolytic side reactions of epoxy groups, and of chlorohydrin groups.
TABLE 13.1. Characterization of Commercial BPA Epoxy Resins
Resin
Standard liquid
1001 Type
1004 Type
1007 Type
1009 Type
n Value
EEW
Melting Point (8C)
0.13
2
5.5
14.4
.16
182 –192
500 –560
875 –950
1600–2300
2500–5500
Liquid
65– 85
90– 110
120– 135
130– 160
13.1. EPOXY RESINS
273
ð13:1Þ
Another side reaction is ring opening by a phenoxide anion polymer end group with
ECH at the less favored, more sterically hindered position: as shown in Eq. (13.2). The
resulting 1,3-chlorohydrin derivative (after proton transfer) cannot ring close to give an
oxirane and is relatively stable under the reaction conditions. This group is the main
site of non-hydrolyzable chlorine that is commonly reported in specifications of commercial epoxy resins.
ð13:2Þ
Instead of ring closing to give an epoxy group, the hydroxyl group of the
1,2-chlorhydrin can react with another ECH molecule to give an addition product:
ð13:3Þ
A hydroxyl group on the polymer backbone can also react with ECH to yield branched
molecules with an epoxy functionality ( f̄n) of 3:
ð13:4Þ
If there were no side reactions, the epoxy f̄n of BPA epoxy resins would be 2; however, the net
effect of side reactions is that commercial BPA epoxy resins generally have an f̄n value below
2, commonly about 1.9. In some cases, this lower f̄n can have an important effect on film properties. The presence of terminal glycol groups can give lower viscosity resins since they result
from chain termination, reducing MW. It has been shown that the presence of small controlled amounts of terminal glycol groups can have beneficial effects on adhesion [2].
The procedure shown in Scheme 13.1 for making BPA epoxy resins is called the taffy
process. Stoichiometric amounts of NaOH are required, resulting in formation of a large
amount of NaCl, which must be removed from the resin by washing with water. The
washing step is not difficult for the standard liquid resin, due to its low viscosity.
274
EPOX Y AND PHENOLIC RESINS
Scheme 13.1
However, as the ratio of ECH to BPA is decreased, higher average MW products are produced, the reaction mixture becomes highly viscous, and water washing to remove NaCl
becomes difficult. Also, as MW and viscosity increase, the probability of branching
increases. The taffy process is now used only for liquid resins.
Higher MW epoxy resins are prepared by reacting the standard liquid epoxy resin
(n ¼ 0.13) with BPA in the presence of a catalyst such as ethyltriphenylphosphonium
hydroxide. The catalyzed reaction of BPA with both epoxide groups of the standard
resin results in a higher MW resin with BPA end groups, which react further with standard
resin to give epoxide end groups. The average MW depends on the ratio of liquid resin to
BPA. There can be variations in the resins obtained, depending on the catalyst used. This
procedure for making epoxy resins is called the advancement process or upgrade process.
The advancement process is carried out at higher temperature than the taffy process;
hence, the viscosity is lower, agitation is better, and there is less branching. No NaCl is
produced; therefore, the difficult salt removal step is unnecessary.
Development of advanced HPLC analytical techniques has permitted improvements in
process development and control. Not only can the individual oligomers (n ¼ 0, 1, 2, 3,
etc.) be separated, but also oligomers in which one or both ends have 1,3-chlorohydrin
or 1,2-dihydroxy groups can be separated [3] (Eqs. 13.1 and 13.2). Analysis of the products resulting when minor process changes are made in reaction conditions permits establishing process parameters to meet relatively narrow product specifications. Tight process
control is particularly critical for epoxy resins to be used in electronic applications and
powder coatings (Chapter 28). Resins made by the taffy process consist of oligomers
with n ¼ 0, 1, 2, 3, 4, 5, and so on, whereas resins made by the advancement process
have largely even number n value oligomers, that is, n ¼ 0, 2, 4, 6, and so on. The predominance of oligomers with even number n values in resins from the advancement process
13.1. EPOXY RESINS
275
follows from starting with the diglycidyl ether of BPA. Roughly 10 wt% of odd n molecules result from the presence of about 10% of n ¼ 1 resin in the liquid resin.
Bisphenol A epoxy resins perform especially well in coatings applications in which
excellent adhesion, electrical properties, and corrosion resistance are required. A limitation is poor exterior durability, primarily resulting from direct absorption of UV radiation
by the aromatic ether groups, which ultimately leads to photoxidative degradation. Lower
MW resins require less VOC and give better chemical resistance when cross-linked with
amines since the cross-link density of films is higher. However, when higher MW resins
are used with cross-linkers that react with hydroxyl groups (e.g., phenolic or amino resins),
chemical resistance is better. Cross-linked higher MW resins also give films that dry more
rapidly and have improved corrosion resistance.
13.1.2. Other Epoxy Resins
A variety of other epoxy resins is available. Bisphenol F (BPF) epoxy resins (from reaction of
ECH with BPF in place of BPA) have the advantage of lower viscosities at the same n value.
Standard liquid BPA epoxies have viscosities on the order of 12 to 14 Pa.s at 258C, whereas the
viscosities of comparable BPF epoxies are 2.5 to 4.5 Pa.s. Solid BPF resins are seldom used.
Still lower viscosities can be obtained by reacting BPA epoxies with an alcohol, such as
n-butyl alcohol, followed by reacting the resultant hydroxyl groups with ECH, using
(CH3)4NCl as catalyst [4]. A new epoxy resin is obtained, a major component of which is
2,20 -bis[ p-(3-n-butoxy-2-glycidyloxypropyloxy)phenyl]propane. Viscosities on the order of
1 Pa.s are reported [5]. The effect of the butyl ether group is to lower the Tg and hence the viscosity. Cross-linked films are softer and more impact resistant than comparable films using
BPA resins.
Epoxy resins with lower viscosity than BPA epoxy resins can be made by substituting a
flexible diol for BPA in the advancement process to make what might be called copolymer
epoxy resins. Examples are advancement resins made from combinations of propylene or
dipropylene glycol and BPA [6]. When cross-linked with phenolic resins, these resins
form films that are more flexible than films from homopolymer BPA epoxy resins but still
provide excellent adhesion. Lower Tg, lower viscosity, and more flexible epoxy resins can
also be prepared from bisphenols with longer chain links between the two phenol rings,
especially multiple 1,2-diether groups [7]. Experimental resins to demonstrate the idea
were prepared by reacting a series of such bisphenols with DGEBPA, using
276
EPOX Y AND PHENOLIC RESINS
ethyltriphenylphosphonium acetate/acetic acid in methyl alcohol as a catalyst. The polyglycidyl ether of castor oil is used with BPA to improve flexibility and water resistance.
Hydrogenated BPA epoxy resins have lower Tg and lower viscosity than do BPA resins
having the same n value. They also exhibit better exterior durability relative to both BPA
and BPF resins, resulting from the absence of the UV absorbing aromatic ether groups.
Brominated BPA epoxy resins are used in flame retardant coatings.
Epoxy resins are also prepared by reaction of ECH with novolac phenolic resins
(Section 13.6.2). The resulting novolac epoxy resins are useful in applications in which
more than two epoxy groups per molecule are desirable, especially in powder coatings
where solid epoxy resins are needed. Epoxy resins derived from the reaction of o- or
p-cresol –formaldehyde novolacs and ECH are available, with an epoxy f̄n of 2.2 to 5.5.
Other novolac epoxy resins are prepared from novolac resins made with phenol or
bisphenol A. Epoxy resins with a functionality of 4 are obtained from tetraphenolethane.
Cross-linked novolac epoxy resin coatings are more chemically resistant than BPA epoxy
resins, due to the higher cross-linked density resulting from the higher functionality. A
general structure of novolac epoxies is shown below.
Triglycidylisocyanurate (TGIC) is a solid trifunctional epoxy cross-linker used in
powder coatings. The presence of three functional groups gives higher cross-link density
than that obtained with BPA epoxy resins, and photochemical stability of the cured coatings is superior. However, there is concern that use of TGIC may present toxic hazards.
In order to reduce VOC, reactive diluents are being used. A coating formulated with the
triglycidyl ether of TMP and a phenol – formaldehyde novolac resin showed better resistance to methylene chloride, acetic acid, and sulfuric acid than one formulated with a
13.1. EPOXY RESINS
277
bisphenol F resin [8]. Neopentyl glycol, butanediol, and cyclohexanedimethanol diglycidyl
ethers are also used. Monofunctional epoxies such as the glycidyl ether of n-butyl alcohol,
o-cresol glycidyl ether, and the glycidyl ester of neodecanoic acid are used to make very
high solids coatings. Being monofunctional, they reduce cross-link density, generally
resulting in some loss of properties when used with BPA epoxy resins. They are more
useful for reducing VOC of novolac epoxies, since novolacs have higher functionality
as well as higher viscosity than BPA epoxies.
A variety of other aliphatic epoxy products is available. Epoxidized soy and linseed oils
are used in making acrylate derivatives for UV cure resins (Section 29.2.4) and thermal
cationic cure resins (Section 13.3.6). Others are made by the reaction of polyols, such
as glycerol, sorbitol, and polyethylene or polypropylene glycols, with ECH and a base catalyst. Such aliphatic epoxies are used as reactive diluents to reduce viscosity and give films
with better exterior durability.
Also available are low MW cycloaliphatic diepoxy compounds such as 3,4epoxycyclohexylmethyl-30 ,40 -epoxycyclohexylcarboxylate (1), prepared by epoxidation
of the corresponding alkenes, generally with peracetic acid. Another is diglycidyl
1,2-cyclohexanedicarboxylate (2). Such low MW epoxy derivatives are particularly
useful as reactive diluents in cationic coatings (Sections 13.3.6 and 29.3.1). They can
also be used as cross-linking agents for polyols (Section 13.3.3), carboxylic acids, and
anhydrides (Section 13.3.2).
A widely used way to make epoxy-functional resins is by free radical polymerization
of acrylic esters with glycidyl methacrylate (GMA) as a co-monomer. By varying
co-monomers, GMA content, and MW, a range of materials can be made. Exterior durability and acid resistance can be excellent. GMA modified acrylic resins are used in clear
coats for automobiles, among other applications.
278
EPOX Y AND PHENOLIC RESINS
13.2. EPOXY – AMINE SYSTEMS
Epoxy groups react at ambient temperatures with primary amines to form secondary
amines and with secondary amines to form tertiary amines. Tertiary amines react at
higher temperatures to form quaternary ammonium compounds.
Reaction rates depend on epoxy and amine structure, concentration, catalysis, and
media effects. Terminal epoxy groups such as glycidyl ethers and esters are more reactive
than internal epoxy groups [e.g., cycloaliphatic diepoxide (1)], which are more sterically
hindered.
Reactivity of amines tends to increase with base strength and decrease with steric
crowding. The general order of reactivity, primary . secondary tertiary amines, can
be attributed to steric effects as well as to the absence of a transferable proton in the
case of tertiary amines. Cycloaliphatic amines have reduced reactivity; the second reaction
of such an amine is particularly slow. Aliphatic amines are more reactive than aromatic
amines, which are less basic. The reaction is catalyzed by water, alcohols, tertiary
amines, and weak acids (most notably by phenols), which promote ring opening by
proton complexation with the epoxide oxygen. Strong acids are not effective catalysts
for the reaction, since they protonate the amine in preference to the epoxide. Hydrogenbond acceptor solvents tend to reduce reaction rates, probably by complexing with
hydrogen donors in competition with the epoxy group.
Scheme 13.2 shows the attack on an epoxy group by the nonbonded electron pair of a
secondary amine. Attack is primarily at the less sterically hindered terminal end of the
epoxy group. Scheme 13.2 also shows catalysis by a weak acid (HA) by hydrogen complexation and hydrogen donation to the epoxide oxygen, which facilitates ring opening.
The catalyst is regenerated by removing the proton from the amine nitrogen, which
takes on a positive charge during the reaction. The conjugate base A2 could also participate in the catalysis by assisting removal of the amine proton during the ring-opening reaction. Based on this reasonable possibility, an explanation for the more effective catalysis
by weak acids relative to strong acids noted earlier can be hypothesized. A strong acid
2
(e.g., HCl) exists predominately as the amine salt (R2NHþ
2 Cl ). Because of the large
Scheme 13.2
13.2. EPOXY –AMINE SYSTEMS
279
excess of amine reactant compared to the catalyst, even weaker acids such as phenols are
2
also expected to exist largely as the corresponding salts (R2NHþ
2 ArO ). This suggests that
the proton donor in the reaction is the same amine ion in both cases. On the other hand, the
proton acceptor, ArO2, is a substantially stronger base than Cl2 because conjugate base
strength increases as acid strength decreases. The proton acceptor ArO2 more effectively
promotes the second reaction shown in Scheme 13.2. Although we are unaware of experimental evidence that directly supports this hypothesis in epoxy –amine reactions, the
general importance of concerted weak acid– weak base catalysis is well established [9].
2,4,6-[Tris(dimethylaminomethyl)]phenol, which has both phenolic and tertiary amine
groups, is an important catalyst for epoxy – amine reactions.
Epoxy –amine reactivity is too high at ambient temperatures to allow sufficient storage
stability of a coating containing polyamine and polyepoxide in the same package; thus,
two package (2K) coatings are required. With many aliphatic amines, the pot life is
limited to a few hours, and the coating generally takes about a week to cure at ambient
temperature.
13.2.1. Pot Life and Cure Time Considerations
Epoxy – amine coatings are formulated to maximize pot life and minimize cure time. (See
Section 2.3.2 for a discussion of kinetic limitations that prevent both long-term
one-package stability and moderate cure temperatures and of ways to obviate these limitations.) Many factors must be considered, including reactive group concentrations; structural effects of amine, epoxy, and solvents on reaction rates; the equivalent and
molecular weights; and the f̄n of the reactants. As the MW of a BPA epoxy resin is increased,
the number of equivalents per liter of epoxy groups decreases; therefore, the reaction rate is
slower. Furthermore, as the MW increases, viscosity increases. Thus, to formulate to the
same viscosity, the amount of solvent must be increased, which decreases the concentration
of both the amine and epoxy groups and lengthens pot life. Unfortunately, the need to reduce
VOC emissions forces formulation of higher solids coatings, which have shorter pot lives.
An approach to the important problem of lengthening pot life without reducing cure
time is to use blocked amine cross-linkers. Ketones react with primary amines to give ketimines, which do not readily react with epoxy groups. However, ketimines hydrolyze with
water to release free amine plus ketone: the reverse reaction of ketimine formation.
Ketimine – epoxy moisture-curable coatings are stable in a moisture-free environment
but cure after application and exposure to ambient moisture. Most commonly, methyl
ethyl ketone (MEK) is used. The high volatility of MEK from thin coating films minimizes
reverse reaction with amine.
Ketimine–epoxy systems are stable indefinitely in the absence of water and should therefore permit formulation of one package coatings. Nevertheless, they are most commonly
used in long pot life two package coatings, because of the difficulty of drying all of the components in a coating. Many epoxy–amine coatings are pigmented, and pigment surfaces have
a layer of water on them before they are dispersed. The water remains in the coating and
280
EPOX Y AND PHENOLIC RESINS
hydrolyzes the ketimine. Although removal of water from the surface of the pigment is possible, it adds to the cost. The amount of water is not high, usually less than 1% of the pigment
weight, but the MW of water is so low that a little water will hydrolyze a larger amount of
ketimine. Solvents also contain water, and use of anhydrous solvents would also increase
costs. A further difficulty is that, cure rate depends on relative humidity as well as temperature.
An interesting approach to increasing storage stability of epoxy coatings sufficiently to
permit formulation of one package coatings that can be cured at moderately elevated
temperatures is to use a cross-linker that must undergo a phase change in order to react.
A solid polyamide prepared from phthalic anhydride and diethylenetriamine [DETA;
H2N(CH2CH2NH)2H] is an example [10]. When dispersed in an epoxy resin and heated
above 1008C, the solid polyamide liquefies and reacts to form imides, releasing amine
cross-linkers, including free DETA. Because the polyamide is a solid during storage, its
reaction rate is not controlled by the kinetics of the reaction, so Arrhenius restrictions
on package stability versus cure schedule (Section 2.3.2) do not apply.
Tertiary alcohol (e.g., t-butyl) urethanes are relatively unstable and may thermally decompose to give alkenes, carbon dioxide, and amines. Advantage is taken of this reaction by using
t-butyl alcohol–blocked isocyanates as sources of amines to cross-link epoxy resins. It has
been shown that reaction of an oligomer obtained from reaction of t-butyl alcohol–blocked
IPDI with an epoxy resin liberates amine that reacts with epoxy groups to yield amine crosslinks, typical of epoxy amine coatings, rather than urethane cross-links [11].
A commonly used cross-linker for epoxy powder coatings is dicyandiamide (DICY), a
crystalline compound (mp 2058C) that also provides latency by insolubility. The complex
reactions of epoxies with DICY have been the subject of numerous studies [12] but are not
fully understood.
Dicyandiamide decomposes into soluble reactive products below its melting point, and
cure temperatures can be reduced substantially by using as accelerators tertiary amines
(e.g., benzyldimethylamine), imidazoles (e.g., 2-methylimidazole), or a substituted urea
[e.g., N-3-( p-chlorophenyl)-N 0 , N 0 -dimethylurea] that releases free amine at elevated
temperatures. Imidazole-blocked 1,5-naphthalene diisocyanate (an imidazole urea) has
been patented for use as a blocked catalyst for epoxy –dicyandiamide coatings [13].
The blocked catalyst provides stability required for one package products.
13.2. EPOXY –AMINE SYSTEMS
281
13.2.2. Toxicity and Stoichiometric Considerations
Factors other than reaction rates are involved in selecting amines. Many amines are toxic.
Although they are easily handled safely in a chemical plant, toxic hazards can arise if
relatively inexperienced, careless, or uninformed personnel mix and apply 2K epoxy coatings incorporating some amines. For example, diethylenetriamine (DETA) is an efficient
cross-linker for epoxies, but the handling hazard is high. In general, toxic hazards are
reduced by increasing MW and reducing water solubility. As MW increases, volatility
decreases, which reduces the chances of inhaling dangerous amounts of amine. Also, as
water solubility decreases and MW increases, permeability through body membranes
such as skin decreases, generally reducing the toxic hazard. There are, of course, exceptions to these broad generalizations. Safety data sheets supplied by the manufacturer
should always be read for safe handling recommendations.
Other disadvantages of using low MW, highly functional amines are their low equivalent weights (EW) and viscosities. The EW of pure DETA is 21. If DETA were used
with an epoxy resin with an EW of about 500, the stoichiometric weight ratio of the
two components would be about 25 :1. This disparate amount would reduce the probability
of mixing stoichiometric amounts in two package coatings and increase the difficulty of
uniform mixing.
One approach to designing amine cross-linkers with a higher EW and lower toxic
hazard is to make amine adducts. Standard liquid BPA epoxy (n ¼ 0.13) is reacted with
an excess of a multifunctional amine such as DETA. The excess amine is removed by
vacuum distillation, leaving the amine-terminated adduct, as shown in the idealized
structure below. The low MW DETA is handled in the chemical factory only with
proper precautions. Similar adducts are prepared using a variety of amines to provide
adducts with a range of cure rates and pot lives.
Another approach is to react a multifunctional amine (e.g., DETA) with aliphatic
mono- or dicarboxylic acids to form amine-terminated amides. Dimer fatty acids are
widely used; they are complex mixtures, predominately C36 dicarboxylic acids, made
by acid catalyzed dimerization of unsaturated C18 fatty acids from linseed or soy oil
and tall oil fatty acids (Section 14.3.1). The reaction yields amine-functional polyamides.
Terminology is confusing; amides derived from dimer acids are frequently called simply
polyamides. The amide nitrogen groups do not react with epoxy groups; only the terminal
amine groups do. (There are similar polyamides that are not amine-functional and not
useful as cross-linkers for epoxy resins.) An example is made by reacting dimer acids
with excess DETA. A mixture of products is obtained, including the simplest polyamide,
shown below. The amine EW is the corresponding MW divided by the number of reactive
NH groups (six in the polyamide shown). A range of polyfunctional amines is used in
making polyamides, including DETA, triethylenetetramine (TETA), aminoethylpiperazine, and aromatic diamines such as m-phenylenediamine. The long aliphatic chains
from the dimer acids improve flexibility, wetting, and adhesion as compared to use of
282
EPOX Y AND PHENOLIC RESINS
an amine such as DETA.
In the preparation of polyamides using an amine such as DETA, the polyamide can
react further by eliminating water and forming a terminal imidazoline group. The extent
of this reaction varies, depending upon reaction conditions. The fraction of imidazoline
affects solubility, compatibility, and pot life (and inversely reactivity). Imidazoline formation reduces average functionality and cross-link density of cured films [14,15]. Commercial grades of polyamides contain from 35 to more than 85% imidazoline, from which
the formulator may choose in designing coatings for a particular application.
A third approach is to react DETA with monocarboxylic acids. The products, frequently called amidoamines, are commonly used with a BPA epoxy. This, of course,
increases the viscosity and the adducts are generally used with a high-boiling solvent
such as benzyl alcohol. Compatibility with the cured film and its high boiling point
result in part of the benzyl alcohol staying in the film.
Conventional amine curing agents are unsatisfactory when it is necessary to cure an
epoxy coating at temperatures below 58C—the cure rate is too slow. This problem has
been addressed by using amine Mannich bases, prepared by reacting a methylolphenol
with excess polyamine [16]. Although the functionality of the amine is reduced, the
presence of the phenolic hydroxyl accelerates the epoxy – amine reaction.
13.2.3. Graininess and Blushing
BPA epoxy resins and polyamides are mutually soluble in the solvents used in epoxy –
amine coatings, but most are not compatible in the absence of solvents. Thus, as solvent
evaporates, phase separation can occur, resulting in a rough surface, called graininess.
Graininess can be avoided by mixing the two packages 30 minutes to an hour prior to application. Partial reaction of the two components takes place during the waiting period. The
elimination of graininess can be attributed to increased viscosity and/or the formation of
reaction products that improve the compatibility of the mixture. When liquid standard resin
is used, longer times are required after mixing to avoid phase separation. Proprietary amine
cross-linkers are available that exhibit better compatibility, minimizing the problem [15].
Blushing or scumming is the appearance of a grayish, greasy deposit on the surface of
films, usually accompanied by incomplete surface cure. Low-temperature high-humidity
conditions increase the probability of blushing. Blushing decreases gloss, increases
13.2. EPOXY –AMINE SYSTEMS
283
yellowing, gives poor recoatability, and may interfere with intercoat adhesion. Blushing is
said to result from formation of relatively stable carbamate salts of some amine groups on
exposure to carbon dioxide and water vapor in the atmosphere [17].
As with graininess, it is often possible to minimize blushing by mixing the epoxy and
amine components one hour or so before application. During the waiting period, some of
the most reactive amine groups (those most likely to form carbamates) react with epoxy
groups, so when the film is applied after this waiting period, blushing is less likely to
occur. If the waiting period is too long, viscosity increase is excessive. Amine adduct
and Mannich base epoxy coatings show little, if any, blushing.
13.2.4. Tg Considerations
It is critical to select a combination of epoxy resin and amine cross-linker that will give a
final Tg that permits relatively complete reaction of the amine and epoxy groups at the
temperature encountered during application. As polymerization and cross-linking
proceed, the Tg increases. As the Tg of a homogeneous network approaches the cure
temperature T, T 2 Tg and free volume decrease, and the reaction rate becomes
limited by the mobility of the reactants rather than by their reactivity. If the Tg
approaches a value of about 40 to 508C above the reaction temperature, reaction essentially ceases and functional groups remain unreacted (Section 2.3.3). Unreacted functional groups cause adverse effects on mechanical properties and solvent resistance. If
a coating that cures well when applied to an offshore oil rig in the Caribbean is
applied to the support structure of an oil rig in the North Sea, where the water temperature even in summer does not exceed 48C, adequate curing may not occur. A possible
example of the effect of Tg is a report that after 7 days curing at 258C, films from
butyl ether – modified BPA epoxy resin (Section 13.1.2), cross-linked with an amine
adduct, had better methyl alcohol resistance than films made with the same amine
adduct and unmodified BPA epoxy resin [5]. There is no evident explanation for
these results based on the chemical compositions. The superior methyl alcohol resistance may have resulted from a greater extent of reaction of the lower Tg butylated
derivative before mobility limitation slowed the reaction.
13.2.5. Other Formulating Considerations
Epoxy coatings can be formulated for application under water to oil rigs, bridges, and so
on. In such formulations it is essential that the polyamine be insoluble in water and that
solubility of water in the polyamine is minimal. Water plasticizes epoxy –amine coatings,
lowering the Tg.
Epoxy –amine coatings tend to exhibit limited solvent resistance and are particularly
susceptible to attack by acidic solvents such as acetic acid. At least in part, this sensitivity
results from diffusion of acetic acid into the film, followed by formation of acetate salts
with amine groups. The hydrophilic salt groups increase the solubility of water in the
film, increasing water permeability, which lowers Tg by plasticizing the film, making it
more susceptible to damage. This situation is exacerbated when cross-link density is
284
EPOX Y AND PHENOLIC RESINS
low. Bisphenol A epoxy resins have an f̄n of about 1.9, so even with highly functional
amine components, cross-link density is limited, especially if there is deviation from stoichiometry. Since the amines are usually polyfunctional, it is generally best to formulate
with a small excess (about 10%) of amine cross-linker to assure that the epoxy groups
are fully reacted. The problem can be alleviated by use of higher functionality epoxy
resins such as novolac epoxies, which have f̄n up to 5. With the higher functionality
and more viscous novolac epoxies, care must be exercised in selection of an amine crosslinker to ensure adequate reaction. Sometimes, blends of BPA and novolac epoxy resins
are used.
Addition of aminoethylaminopropyltrimethoxysilane to BPA epoxy – polyamide coatings improves heat resistance, decreases color change with heat, increases hardness, and
increases solvent resistance as a result of the additional cross-linking [18].
Siloxane-modified epoxy resins are also reported to give superior coating properties.
Formulation of a BPA epoxy resin modified by reaction with hydroxyl-terminated
dimethylsiloxane, with a polyamide cross-linker, resulted in improved resistance to
impact, mandrel bending, scratching, humidity, and NaOH. Due to the lower surface
tension, films picked up much less dust [19].
Another factor to be considered when formulating epoxy – amine coatings is the effect
of solvent composition on the coating. As noted earlier, hydrogen-bond acceptor solvents
extend pot life. However, ketones and esters should be avoided, since they form ketimines
or undergo aminolysis, respectively, especially with primary amines, at room temperature,
which reduces the concentration of reactive amine groups. An exception is t-butyl acetate,
in which case amine loss is essentially as slow as when xylene is used [20]. Another
important advantage of t-butyl acetate is its negligible photochemical reactivity, resulting
in its exemption from VOC regulations in the United States in 2005 and its absence on the
HAP list.
Alcohols and water catalyze the reaction of amines with epoxy resins and can also react
with epoxy groups, affecting the stability of the epoxy component package. Alcohols react
very slowly with epoxy groups at room temperature. When the reactant is a monoalcohol,
there is little change in viscosity. Formulators generally judge package stability by
changes in viscosity, but in this case, there could be a change in epoxy functionality
without a significant change in viscosity. To the extent that epoxy groups react with an
alcohol, the potential for cross-linking is decreased. Since the f̄n of BPA epoxy resins is
somewhat less than 2, further loss of functionality could cause the final film to have inferior
properties. Apparently, in many systems, there is no difficulty. For example, some epoxy
resins are sold as solutions in glycol monoethers. However, in other cases, a decrease in
epoxy content has been observed after storage over a period of several months.
A similar situation can arise with water. Again, reaction of an epoxy group with water
consumes epoxy groups. In some cases, in which the epoxy package is pigmented with
TiO2, it has been shown that the epoxy content decreases with storage time, presumably
as a result of the reaction of water from the surface of the TiO2. This reaction, as well as
that with alcohols, may be catalyzed by basic and/or acidic impurities or by alumina
(basic) and silica (acidic), generally present as surface treatments on TiO2. In most
cases epoxy-package pigmented coatings have been used without problems. Apparently,
the effects of both alcohols and water on package stability are system dependent.
In view of such reports, it is recommended that package stability be checked
periodically by epoxy group analysis rather than by relying solely on monitoring
viscosity changes.
13.2. EPOXY –AMINE SYSTEMS
285
13.2.6. Waterborne Epoxy – Amine Systems
To reduce solvent content, waterborne epoxy –amine coatings have been widely investigated. One approach has been to make emulsions. Incorporation of emulsifying agents
in either or both the amine and the epoxy package permits the addition of water during
mixing. Application solids of a waterborne coating are lower than those of an organic
solvent-soluble system, but the VOC content is lower, since a substantial fraction of the
volatiles is water. Some loss of epoxy groups can be expected during the pot life of the
mixed coating, due to the high water content. Proprietary “self-emulsifiable” epoxy
resins and polyamides are available, which are made by reacting a surfactant with the
epoxy resin or polyamide; properties approaching those of solventborne coatings can be
achieved [21]. Properties equal to solventborne coatings have been disclosed for waterborne epoxy – amine coatings, derived from an advanced BPA epoxy resin emulsified
with an epoxy-terminated surfactant together with a glycidyl ether of C8 – 14 alcohols
and an amino-amide cross-linker [22].
Another approach to waterborne systems is to use salts of resins having primary amine
groups. The amine groups of a concentrated solution of the resin in organic solvents are
neutralized with hydrochloric acid. When the solution is diluted with water, polymer
aggregates, swollen with solvent and water, are formed having the amine salt groups on
the outer periphery of the aggregates suspended in a continuous water phase. The behavior
is analogous to that of water-reducible resins discussed in Section 8.3. When an organic
solution of epoxy resin is mixed into the system, it goes inside the resin aggregates.
The epoxy groups are thus kept separate from the amine hydrochloride groups, permitting
a pot life of several days. After the coating is applied, the water and solvent evaporate,
leaving the amine hydrochloride and epoxy groups in the same phase. They react to
yield a chlorohydrin and a free primary amine. Then the amine can react twice with
two additional epoxy groups. Since BPA epoxy resins have less than two epoxy groups
per molecule and about one in three of these groups will be converted to chlorohydrins,
it is desirable to use at least some novolac epoxy with an f̄n up to 5. Then when a third
of the groups are converted to chlorohydrins, there is still an average of more than two
epoxy groups per molecule to cross-link with amine groups.
A novel approach is to use weakly acidic solvents, such as nitroalkanes [23,24].
Nitroalkanes form salts of amines: as shown in the equation below; the salt groups stabilize epoxy – amine emulsions and allow the system to be reduced with water. Following
application, the nitroalkane solvent evaporates, shifting the acid– base equilibrium to
the free amine, that is, to the left in the equation. Thus, the amine –nitroalkane combination functions as a transient emulsifying agent that stabilizes the emulsion during
storage and yet is not present to affect the final film properties adversely. Conversion of
amine groups to salts prolongs the pot life of the mixed composition, since polar salt
groups are oriented outward into the water phase while the epoxy groups are in the interior
of the emulsion particles.
Resins used in automotive primers applied by cathodic electrodeposition are prepared
by the reaction of epoxy and amine resins. The use of aqueous dispersions of carboxylate
salts of such resins is discussed in Section 27.2.
286
EPOX Y AND PHENOLIC RESINS
13.3. OTHER CROSS-LINKING AGENTS FOR EPOXY RESINS
13.3.1. Phenols
Epoxy resins, including BPA and novolac epoxies, can be cross-linked with phenolic
resins; both resole and novolac phenolic resins, discussed in Sections 13.6.1 and 13.6.2,
respectively, can be used. The reaction with phenols occurs predominantly at the less hindered CH2 site of unsymmetrical epoxides (e.g., glycidyl derivatives): The reaction is acid
catalyzed; pTSA and phosphoric acids are commonly used.
The phenolic hydroxyl groups of both resole and novolac phenolics react with epoxy
groups. In addition, the methylol groups of the resole phenolics undergo self-condensation
and probably also react with hydroxyl groups on the epoxy resin. Thus, cross-link density
is higher with resole phenolics. There is also a reaction with the hydroxy groups formed by
the reaction with phenols. The coatings require baking, and package stability is relatively
limited. Package stability is enhanced with etherified resole resins. Increased solids and
high functionality are reported using butoxymethylolated BPA as the phenolic resin [5].
For other examples, see Section 13.6.3.
Unpigmented epoxy – phenolic coatings are used as linings for beverage cans and for
some types of food cans. Concern has been raised because of the possible endocrine
disruption by free bisphenol A, an estrogen mimic. Studies are under way to determine
whether trace amounts of BPA are extracted in food or beverage cans from BPA epoxy
containing can linings [25]. Pigmented epoxy –phenolic coatings are used as high performance primers. In both cases, the major advantages are adhesion to metals, even in
the presence of water, and complete resistance to hydrolysis. In both of these applications,
neither the discoloration that occurs on baking nor poor exterior durability is important.
Waterborne epoxy – phenolic emulsion coatings have been developed. Since conventional surfactants remain in the final film and detract from water resistance, water-soluble
amine salts of methacrylic acid copolymers have been recommended [26]. For example,
an amine salt of a copolymer of 40 : 20 : 20 : 20 methacrylic acid/methyl methacrylate/ethyl acrylate/styrene was shown to be an effective emulsifying agent. A further
approach to waterborne epoxy –phenolic coatings is to prepare carboxylic acid– functional
derivatives [27]. For example, a polyalkylidene phenol can be reacted with a BPA epoxy
resin and then with formaldehyde and a haloacetic acid. The product has carboxylic acid,
epoxy, and methylol groups. Aqueous dispersions are prepared using a tertiary amine.
13.3.2. Carboxylic Acids and Anhydrides
Carboxylic acids are effective cross-linkers for epoxy coatings. Reference [28] provides a
review of the patent literature. The reaction of a carboxylic acid and an epoxy group yields
a hydroxy ester. Ring opening occurs predominantly at the less hindered CH2 carbon,
although reaction at the more hindered CH22R site is significant.
ð13:5Þ
13.3. OTHER CROSS-LINKING AGENTS FOR EPOXY RESINS
287
The reaction has been shown to be second order in carboxylic acid. Most likely, one of
the acid groups functions by nucleophilic attack at the CH2 group, whereas the other functions as an electrophile and assists ring opening by complexation with the epoxy oxygen
(Scheme 13.2). The second-order dependence of reaction rate on acid concentration also
results in a rapidly decreasing rate with conversion, exacerbating the difficulty of achieving high conversions:
rate ¼ k½epoxy½RCOOH2
The third-order dependence of the reaction rate (first order in epoxy and second order in
carboxylic acid) also results in a small Arrhenius A value owing to the high molecular
ordering of the epoxide and two carboxylic acid groups required in the transition state,
reducing reactivity at all temperatures. Glycidyl methacrylate copolymers and cycloaliphatic epoxides such as 3,4-epoxycyclohexylmethyl-30 ,40 -epoxycyclohexylcarboxylate
(1) and diglycidyl 1,2-cyclohexanedicarboxylate (2) react more rapidly than BPA epoxies.
Tertiary amines catalyze the reaction of carboxylic acids with epoxies. Triphenylphosphine is also reported to be a particularly effective catalyst. With triphenylphosphine catalysis and an excess of epoxy groups to carboxylic acid groups, coatings can be formulated
that cross-link at 258C [29].
An extensive study of catalysis selection on curing temperature and film properties in
triglycidylisocyanurate/carboxylic acid– functional polyesters for potential use in powder
coatings has been published [30]. Benzyltrimethylammonium chloride was found to give
the best combination of catalyst effects: fastest cure at 1208C, slowest cure at 80 to 908C,
and least yellowing on overbake. The slow cure at 80 to 908 is indicative of stability during
extrusion. The results in actual powder coatings will be of interest.
Latent amine catalysts have been designed for powder coatings formulated with BPA
epoxy resins and carboxylic acid –functional polyesters [31]. The catalysts are crystalline
solid amic acids (e.g., the amic acid derived from the reaction of 3-methylphthalic anhydride and N, N-dimethylaminopropylamine). The amic acids are zwitterions, in which the
tertiary amine is protonated, minimizing both their solubility in the powder coating and
their catalytic activity. On heating, the amic acid melts and undergoes intramolecular
cyclization to an imide with the elimination of water and deprotonation of the tertiary
amine, which catalyzes the carboxylic acid– epoxy reaction.
Hydroxyl groups compete with carboxylic acid groups in reactions with epoxides.
When carboxylic acids are used as cross-linking agents, part of the cross-links result
from the reaction of epoxies with hydroxyl groups originally present on the epoxy resin
or generated in the epoxy – carboxylic acid reaction. Esterification of carboxylic acid
groups with hydroxyls may also occur.
Carboxylic acid – functional acrylics can be cross-linked with BPA epoxy resins, as
well as with acrylic copolymers, having pendant epoxy groups derived from glycidyl
methacrylate (GMA) as a co-monomer [28]. Self cross-linking acrylics can be made by
incorporating both (meth)acrylic acid and GMA in the same polymer. Although the reaction rate of epoxy groups with carboxylic acid groups at storage temperatures is slow, it is
not zero. When self cross-linking resins are made, the time of storage stability starts when
the resin is made. On the other hand, when the functional groups are on two different
resins, the time of storage stability does not start until the resins are mixed to make the
liquid coating.
288
EPOX Y AND PHENOLIC RESINS
A waterborne epoxy coating cross-linked with carboxylic acid groups has been
reported to give films with a combination of flexibility, hardness, solvent resistance,
and light color [32]. The cross-linker was prepared by reacting 2 mol of trimellitic anhydride (TMA) with 1 mol of ethylene glycol to give a combination of 60% ethylene
glycol bis(anhydromellitate) and 11% unreacted TMA, the balance being oligomers.
Aqueous dispersions of a BPA epoxy resin were reacted with aqueous solutions of
the cross-linker which were stable and in which the anhydride groups hydrolyzed to
acids.
Cyclic anhydrides are also used as cross-linkers for epoxy resins. (See Ref. [28] for a
review of the patent literature.) Anhydrides react initially with the epoxy resin hydroxyl
groups, yielding esters and carboxylic acids. The resulting carboxylic acid groups then
react with epoxy groups. This reaction generates a new hydroxyl group, and so on.
Epoxy groups can also react directly with anhydrides. Tertiary amines are generally
used as catalysts; they probably function primarily by reacting with the epoxy to form a
transient zwitterion, which then reacts with the anhydride.
13.3.3. Hydroxyl Groups
Hydroxy-functional resins are not sufficiently reactive with BPA epoxy resins to be useful
as cross-linkers, although when BPA epoxies react with resole phenolic resins, carboxylic
acids, and anhydrides, reaction also occurs with hydroxyl groups. However, with proper
catalysis, cycloaliphatic epoxies serve as cross-linking agents for polyols for films baked
at 1208C. This reactivity has been used to formulate waterborne coatings based on caprolactone polyols and 3,4-epoxycyclohexylmethyl-30 ,40 -epoxycyclohexylcarboxylate (1)
with diethylammonium triflate as a blocked catalyst [33]. The active catalyst, triflic
acid, is sufficiently strong that it is not consumed by direct addition to epoxide groups.
(See a related discussion in Section 13.3.6.)
BPA epoxy resins can be cross-linked by reactions of their hydroxyl groups. Both MF
and UF amino resins are used; cross-linking occurs mainly between the activated ether
groups of the MF or UF resin and the hydroxyl groups of the epoxy resin by transetherification (Chapter 11). Generally, amine salts or esters of pTSA or another sulfonic acid
are used as latent catalysts.
BPA epoxy resin emulsions in water have been cationically polymerized by perchloric
acid. The product is higher MW polyols, although some epoxy groups remain. Coatings
are formulated with MF resins to yield cross-linked films. The resins have the advantage
that the amount of free BPA and n ¼ 0 BPA epoxy resin is much lower than when conventional epoxy resins are used, such as in can linings [34].
Polyisocyanates also cross-link the hydroxyl groups of epoxy resins. Blocked isocyanates are generally preferred because they permit one-package stability (Section 12.5).
13.3.4. Mercaptans
Mercaptans (thiols, RSH) react with epoxies to yield sulfides. The reaction is strongly
catalyzed by tertiary amines, which convert the mercaptan into the more highly reactive
mercaptide anion RS2. Following addition of RS2, the intermediate alkoxide anion is protonated by the ammonium cation, thereby regenerating the catalyst, as shown.
13.3. OTHER CROSS-LINKING AGENTS FOR EPOXY RESINS
289
Reactivity of the mercaptide anion is sufficient for ambient temperature cure 2K coatings. Polysulfide rubbers are low MW mercaptan-terminated polymers; they have been
used to cross-link with BPA epoxy resins in primers for aircraft. The unpleasant odor
of mercaptans is a drawback for some applications.
13.3.5. Homopolymerization
Epoxy groups undergo homopolymerization to polyethers in the presence of tertiary
amines, Lewis acids, and very strong protic acids (super acids). Acid precursors are
most commonly used as initiators. There are two types: blocked or latent acids that
undergo thermal decomposition to give the free acid and photoinitiators that release
acid on exposure to UV. UV curable epoxy coatings are discussed in Section 29.3.1.
The following equation shows initiation and the first step of polymerization:
Suitable super acids are trifluoromethylsulfonic acid (triflic acid; F3CSO3H), hexafluoroantimonic acid (HSbF6), hexafluoroarsenic acid (HAsF6), hexafluorophosphoric acid
(HPF6), and boron trifluoride (BF3) etherate. Since strong acids have weak conjugate
bases, the corresponding counter ions are nonnucleophilic. Only super acids are effective
for homopolymerization of epoxies. Relatively strong acids such as HCl and pTSA are ineffective because the conjugate bases of such acids are nucleophilic enough to add to the
protonated epoxy group, preventing addition of a second epoxy group, as in the homopolymerization reaction. The result is addition to form a chlorohydrin (or a sulfonate ester) rather
than polymerization.
Homopolymerization can also be used for thermosetting coatings. a,a-Dimethylbenzylpyridinium hexafluoroantimonate is a blocked catalyst that permits curing at 1208C
of a GMA copolymer using 3,4-epoxycyclohexylmethyl-30 ,40 -epoxycyclohexylcarboxylate
(1) as a reactive diluent while retaining adequate pot life [35]. Cycloaliphatic epoxides
290
EPOX Y AND PHENOLIC RESINS
such as 1 have been used along with polyols in thermal cationic cure coatings (Section 13.3.3);
part of the cross-linking is from homopolymerization and part from reaction with hydroxyl
groups [36]. Epoxidized linseed oil can be added to increase impact resistance.
13.4. WATER-REDUCIBLE EPOXY ACRYLIC GRAFT COPOLYMERS
A large scale use of epoxy resins is to make water-dispersible acrylic graft copolymers
[37,38]. One way to prepare such graft copolymers is to react an acrylic copolymer
derived from ethyl acrylate, styrene, and methacrylic acid with a BPA epoxy resin in a
glycol ether solvent under conditions where only a fraction of the carboxyl groups react.
A phenolic resin or Class I MF resin is added as a cross-linker, the acid groups are neutralized with an amine, and the system is diluted with water. The result is a dispersion that is
used as a spray applied coating for the interior linings of two piece beverage cans.
A second method is to polymerize ethyl acrylate, styrene, and methacrylic acid using
benzoyl peroxide as initiator in the presence of a BPA epoxy resin in a glycol ether
solvent. The reaction is carried out at about 1308C, at which both benzoyloxy and
phenyl radicals are generated. These radicals can initiate polymerization as well as
abstract hydrogens from the epoxy resin [39]. Abstraction of a hydrogen results in a
free radical on the epoxy resin backbone, which serves as an initiating site for polymerization of the vinyl co-monomers. Thus, a graft copolymer is formed with acrylic–styrene side
chains substituted with carboxylic acid groups (from the methacrylic acid). The product is a
mixture of epoxy–acrylic graft copolymer, nongrafted acrylic copolymer, and unreacted
epoxy resin. The complex resin mixture is mixed with amine, cross-linker, and water as
above to form a dispersion. Since the water-solubilizing groups are attached by C22C
bonds rather than by ester groups, as in the first method, the resins are resistant to hydrolysis
during storage. Sometimes, a latex is blended with epoxy–acrylic dispersions to reduce cost.
13.5. EPOXY RESIN PHOSPHATE ESTERS
Phosphoric acid reacts with BPA epoxy resins to generate phosphate esters. Complex reactions occur; the predominant product is the monophosphoric acid ester of the primary
alcohol, but other products are also present [30]. Most epoxy groups are hydrolyzed
during the reaction to give the corresponding diols.
Low MW epoxy phosphates are used as adhesion promoters. Higher MW epoxy resins
can also be modified by reacting with minor amounts of aqueous phosphoric acid. Their
use in epoxy –phenolic formulations gives coatings with better adhesion and flexibility,
without the need for a phosphoric acid catalyst, than those with corresponding unmodified
epoxy – phenolic coatings [40].
13.6. PHENOLIC RESINS
Although their importance has waned, phenolic resins still have significant uses. Phenolics
are made by reaction of formaldehyde with phenol and substituted phenols. The products
depend on the phenol(s) used, the stoichiometric ratio of phenol to formaldehyde, and the
pH during the reaction. Phenolic resins are divided into two broad classes: resole phenolics, which are made using alkaline catalysts and high ratios of formaldehyde to phenol,
13.6. PHENOLIC RESINS
291
and novolac phenolics, which are made using acid catalysts and low ratios of formaldehyde to phenol. Phenolic resins are used on a large scale in plastics and adhesives applications. (See Ref. [41] for details of the chemistry and the wide range of applications.)
13.6.1. Resole Phenolic Resins
Under alkaline conditions, the initial reaction product of phenol and formaldehyde is a
mixture of ortho and para methylolated phenols. The methylolated phenols are more reactive with formaldehyde than the unsubstituted phenol, resulting in the rapid formation of
2,4-dimethylolphenol and subsequently, 2,4,6-trimethylolphenol; the latter is the predominant product, with a large excess of formaldehyde and a relatively short reaction time.
With lower ratios (but still a molar excess) of formaldehyde to phenol and longer reaction
times, formation of higher MW resole phenolic resins is favored. Polymerization occurs
primarily by a methylol group on one phenol reacting at the ortho or para position of
another phenol to form a methylene bridge connecting the two phenols. Dibenzyl ether
bridges connecting two phenols also form by reaction of two methylol groups with each
other. With excess formaldehyde, methylol groups are present on the terminal phenol
groups of resole resins. Although not shown in the general structure, some of the aromatic
rings have three substituents.
Such phenol-based resole resins cross-link on heating and are used in adhesive and
plastics applications. However, they are not suitable for coatings applications, primarily
because their cross-link density is higher than appropriate for any coating. Furthermore,
the package stability of the resins is limited. Resole phenolics useful in coatings applications
are made from monosubstituted phenols and mixtures of monosubstituted phenols with
phenol. The use of substituted phenols reduces the potential cross-link density. There are
two broad categories of such resins: (1) those that are soluble in alcohol and other low
MW oxygenated solvents, commonly called alcohol-soluble heat-reactive phenolics, and
(2) those that are soluble in vegetable oils and are called oil-soluble heat-reactive phenolics.
Alcohol-soluble heat-reactive resole resins are prepared by reacting phenol, o- or p-cresol,
and formaldehyde in the presence of a base catalyst at less than 608C while removing water
under vacuum. The catalyst is neutralized, alcohol is added, and the salt resulting from catalyst neutralization is removed by filtration. Potential cross-link density is controlled by the
ratio of phenol to cresol; MW is controlled by the ratio of formaldehyde to phenols and by
reaction time.
Such resole phenolic resins are used in interior can coatings and tank linings. They
require baking using an acid catalyst to cure in short times. To enhance flexibility and
adhesion, they are commonly blended with low MW poly(vinyl butyral) as a plasticizer.
The films are resistant to swelling by oils, such as encountered in canned fish, and are
292
EPOX Y AND PHENOLIC RESINS
completely resistant to hydrolysis. These resins and other heat-reactive phenolics discolor
during baking due to quinone methide formation, which restricts their use to applications
for which development of a yellow-brown color is permissible. These resins are also
blended with epoxy resins in thermosetting coatings for applications such as primers
and can coatings (Section 13.3.1). The absence of hydrolyzable bonds and generally excellent adhesion properties are their chief advantages.
Oil-soluble heat-reactive phenolics are prepared by reacting a para-substituted phenol
(e.g., p-phenylphenol, p-t-butylphenol, or p-nonylphenol) with somewhat less than 2 mol
of formaldehyde per mole of substituted phenol. The resulting resole phenolics are cast
from the reactor after neutralization of the catalyst. The resulting resole phenolics are
solid, linear resins with terminal methylol groups. The most common use for such resins
is to make varnishes with linseed oil and/or tung oil. However, since varnishes have, in
large measure, been replaced by other vehicles, the consumption of these resins has
declined markedly.
13.6.2. Novolac Phenolic Resins
Novolac phenolics of interest for coatings are made with acid catalysts and o- or
p-substituted phenols. MW is controlled by the molar ratio of phenol and formaldehyde,
which is always greater than 1. In contrast to resole phenolics, the terminal phenol groups
are not metholylated, as shown in the idealized structure.
Three types of novolac resins are used in coatings:
1. Alcohol-soluble nonheat-reactive low MW phenolics are derived from o- or
p-cresol. An important use for these resins is in the preparation of novolac epoxy resins
by reaction with epichlorohydrin (Section 13.1.2).
2. Oil-soluble nonheat-reactive novolac phenolic resins are made using a low ratio of
formaldehyde, an acid catalyst, and a substituted phenol (e.g., p-phenylphenol,
p-t-butylphenol, or p-nonylphenol). They are used together with drying oils, particularly
tung oil or tung –linseed oil mixtures, to make varnishes. Such varnishes are still used
to a small extent as marine spar varnish, for which their reputation for durability maintains
their position in the do-it-yourself marine yacht market. The durability of phenolic
varnishes may well result, at least in part, from the antioxidant activity of the phenolic
groups. (See Section 5.2.2 for a discussion of phenolic antioxidants.)
3. Rosin-modified phenolic resins are the principal type of modified phenolics still in
use. Their use in coatings is limited to low cost varnishes. The largest volume use is in
printing inks. For inks, the phenolic resin is prepared in the presence of rosin esters
and/or zinc or calcium salts of rosin. The structures of the reaction products, which are
REFERENCES
293
high melting hydrocarbon-soluble resins, are not completely known. An important
example of their use is in publication gravure inks such as those used in mail-order catalogs and some magazines. They are also used, to a degree, in heat-set letterpress printing
inks for magazine and paperback book applications.
13.6.3. Ether Derivatives of Phenolic Resins
The package stability of alcohol-soluble resole resins and their compatibility with epoxy
resins can be improved by partial conversion of the methylol groups to ethers. The ether
groups undergo exchange reactions with hydroxyl groups in the presence of acid catalysts.
Allyl ethers have been used for many years to cross-link epoxy resins in interior can
coatings.
Low MW n- and isobutyl ethers are also used for cross-linking epoxy resins and other
hydroxy-substituted resins, primarily by etherification and transetherification reactions
[42]. A typical resin has an average of 2.2 aromatic rings per molecule. Due to its low
average MW (ca. 320) the resin has a moderate viscosity, as supplied in butyl alcohol solution. The reactive groups in the resins are primarily butoxymethyl groups, but there are
also benzyloxy groups and some free methylol groups. Furthermore, the phenol groups can
react with epoxy groups. Acid catalysts such as phosphoric or sulfonic acids are generally
used. Blocked acids may be utilized to extend shelf life.
GENERAL REFERENCES
Ellis, B., Ed., Chemistry and Technology of Epoxy Resins, Blackie Academic & Professional,
London, 1993.
May, C. A., Ed., Epoxy Resins: Chemistry and Technology, Marcel Dekker, New York, 1988.
REFERENCES
1. McAdams, L. V.; Gannon, J. A., in Encyclopedia of Polymer Science and Engineering, 2nd ed.,
Wiley, New York, 1986, Vol. 6, pp. 322 – 382; Epoxy Products and Applications, Dow Chemical
Co., Midland, MI, 2004, www.dow.com/epoxy/products. Physical Properties Guide for Epoxy
Resins and Related Products, Resolution Chemical Co., Houston, TX, 2001, www.resins.com.
2. Sheih, P. S.; Massingill, J. L., J. Coat. Technol., 1990, 62(781), 25.
3. Scheuing, D. R., J. Coat. Technol., 1985, 57(723), 47.
4. Bagga, M. M., Eur. patent appl. EP22,073, 1981.
5. Bozzi, E. G.; Helfand, D., FSCT Symposium, Louisville, KY, May 1990. See also Payne, K. L.;
Puglisi, J. S., J. Coat. Technol., 1987, 59(752), 117.
6. Massingill, J. L.; et al., J. Coat. Technol., 1990, 62(781), 31.
7. Dubois, R. A.; Sheih, P. S., J. Coat. Technol., 1992, 64(808), 51.
8. Kincaid, D. S.; Schulte, J. A., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA,
2001, pp. 127 – 141.
9. Jencks, W. P., Catalysis in Chemistry and Enzymology, McGraw-Hill, New York, 1969,
pp. 199 – 211. Jencks, W. P., Chem. Rev., 1972, 72, 705.
10. Agostinho, M.; Brytus, V., J. Coat. Technol., 1988, 60(764), 61.
11. Holubka, J. W., in Reactive Oligomers, ACS Symp. Ser. 282, American Chemical Society,
Washington, DC, 1985.
294
EPOX Y AND PHENOLIC RESINS
12. Gilbert, M. D.; et al., Macromolecules, 1991, 24, 360. Fedtke, M.; et al., Polym. Bull., 1993, 31,
429.
13. Lee, F. W. C.; Baron, K. S., U.S. patent 4,533,715 (1985).
14. Brytus, V., J. Coat. Technol., 1986, 58(740), 45.
15. Brady, R. F.; Jr., Charlesworth, J. M., J. Coat. Technol., 1993, 65(816), 81.
16. Weinmann, D. J.; et al., J. Coat. Technol., 1996, 68(863), 29.
17. Tess, R. W., in Epoxy Resins: Chemistry and Technology, May, C. A., Ed., Marcel Dekker,
New York, 1988, p. 743.
18. Kiatkamjornwong, S.; Yusabai, W., Surf. Coat. Int. B Coat. Trans., 2004, 87-B3, 149.
19. Ahmad, S.; et al., Prog. Org. Coat., 2005, 54, 248.
20. Cooper, C.; et al., J. Coat. Technol., 2001, 73(922), 19.
21. Wegmann, A., J. Coat. Technol., 1993, 65(827), 27.
22. Stark, C. J.; et al., U.S. patent 6,221,934 (2001).
23. Albers, R., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA, 1983, pp. 130 – 143.
U.S. patent 4,352,898 (1982).
24. Lopez, J. A., U.S. patent 4,816,502 (1989).
25. Howe, S. R.; et al., J. Coat. Technol., 1998, 70(877), 69. Wingender, R. J.; et al., J. Coat.
Technol., 1998, 70(877), 75.
26. Kojima, S.; et al., J. Coat. Technol., 1993, 65(818), 25.
27. Oberressl, P.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
1997, p. 296.
28. Ooka, M.; Ozawa, H., Prog. Org. Coat., 1994, 23, 325.
29. Shalati, M. D.; et al., Proc. Int. Conf. Coat. Sci. Technol., Athens, Greece, 1990, p. 525.
30. Merfeld, G.; et al., Prog. Org. Coat., 2005, 52, 98.
31. Pappas, S. P.; et al., J. Coat. Technol., 1991, 63(796), 39.
32. Anderson, R. L.; Bohr, T., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 1999, pp. 515 –526.
33. Eaton, R. F.; Lamb, K. T., J. Coat. Technol., 1996, 68(860), 49.
34. Walker, F. H.; et al., Prog. Org. Coat., 2002, 45, 291.
35. Nakano, S.; Endo, T., Prog. Org. Coat., 1996, 28, 143.
36. Eaton, R. F., Polym. Mater. Sci. Eng., 1997, 77, 381.
37. Woo, J. T. K.; et al., J. Coat. Technol., 1982, 54(689), 41.
38. Woo, J. T. K.; Toman, A., Polym. Mater. Sci. Eng., 1991, 65, 323.
39. Woo, J. T. K.; Toman, A., Prog. Org. Coat., 1993, 21, 371.
40. Massingill, J. L., J. Coat. Technol., 1991, 63(797), 47.
41. Kopf, P. W., in Encyclopedia of Polymer Science and Engineering, 2nd ed., Wiley, New York,
1988, Vol. 11, pp. 45– 95.
42. Santolink EP 560, Monsanto Chemical Co., Springfield, MA, 1990.
14
Drying Oils
Among the oldest binders for paints are drying oils, which are liquid vegetable or fish
oils that react with oxygen to form solid films. They have been used since prehistoric
times. Leonardo da Vinci, Rembrandt van Rijn, and most other famous painters used
drying oils. In the nineteenth and early twentieth centuries, binders of most paints
were drying oils. Their use has decreased; however, they still have applications. Most
important, they are raw materials for other binders, such as alkyd, epoxy ester, and
uralkyd resins (Chapter 15). These resins can be considered to be synthetic drying
oils; an understanding of drying oil chemistry is a necessary foundation for understanding them, which is presented herein; general references are also provided for more
specific details.
14.1. COMPOSITION OF NATURAL OILS
Naturally occurring oils are triglycerides, triesters of glycerol and fatty acids. Some
triglycerides are drying oils, but many are not. The reactivity of drying oils with
oxygen results from the presence of 1,4-dienes (22CH55CHCH2CH55CH22), which
comprise diallylic methylene groups, or 1,3-conjugated double bonds. Esters of many
different fatty acids occur in nature. Fatty acids with 18 carbon atoms are most
common; those found in oils that are most important in coatings are provided below.
The letters c and t represent cis and trans orientation, corresponding to zusammen
and entgegen, respectively; the numbers designate the position of the first
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
295
296
DRYING OILS
carbon of the double bond.
The oils are mixtures of mixed triglycerides with different fatty acids distributed
among the triglyceride molecules. Separation of the tens or hundreds of different triglyceride molecules present in an oil borders on the impossible. Therefore, oils are characterized by high performance liquid chromatography (HPLC) or gas chromatography
(GC) of the methyl esters obtained by transesterification [1]. Typical fatty acid contents
of some oils are given in Table 14.1. Compositions of the oils vary, sometimes quite
widely, with variations in plant strain, climate, soil, and other growth conditions. Differences in sunflower oils are especially large, as illustrated in Table 14.1 for sunflower oils
from Minnesota and Texas. In general, oils derived from seeds grown in colder climates
have lower freezing points, due to higher levels of unsaturation.
Included in Table 14.1 are tall oil fatty acids (TOFA). The word tall is the Swedish
word for pine. Tall oil fatty acids are obtained as a by-product of the sulfate pulping
process for making paper. Oils present in the trees are saponified to soaps in the
process. Tall oil fatty acids are obtained by acidification, followed by fractional distillation
to separate the fatty acids from rosin acids and other by-products also present. These
important fatty acids have a range of compositions, especially between acids of North
American and European origin, as shown. They are often chemically modified to give
conjugated bonds (Section 14.3.3).
Animal oils and fats are also triglycerides, but the only animal oils used to any degree in
coatings are refined fish oils. They contain triglycerides of a wide spectrum of highly
unsaturated fatty acids, including C18 to C26 fatty acids having up to five nonconjugated
double bonds each separated by single methylene groups.
297
14.2. AUTOXIDATION AND CROSS-LINKING
TABLE 14.1. Typical Fatty Acid Compositions of Selected Oils
Fatty Acid
a
Oil
Linseed
Safflower
Soybean
Sunflower, MN
Sunflower, TX
Tung
Tall oil fatty acidsc
Tall oil fatty acidsf
Castor
Coconut
Saturated
Oleic
Linoleic
Linolenic
10
11
15
13
11
5
8
2.5
3
91
22
13
25
26
51
8
46
30
7
7
16
75
51
61
38
43
41d
45
3
2
52
1
9
trace
trace
80b
3
1
Other
—
2e
14g
87h
a
Saturated fatty acids are mainly mixtures of stearic (C18) and palmitic (C16) acids; coconut oil also contains C8,
C10, C12, and C14 saturated fatty acids.
b
a-Eleostearic acid.
c
North American origin.
d
Linoleic plus geometric and conjugated isomers.
e
Rosin.
f
European origin.
g
Pinolenic acid.
h
Ricinoleic acid.
14.2. AUTOXIDATION AND CROSS-LINKING
Drying, semi-drying, and non-drying oils are often defined based on their iodine value, that
is, grams of iodine required to saturate the double bonds of 100 g of an oil. Some authors
classify oils as follows: drying oils, iodine value greater than 140; semi-drying oils, iodine
value 125 to 140; and non-drying oils, iodine value less than 125 [2]. Although iodine
values can serve as satisfactory quality control specifications, they are not useful and
can be misleading in defining a drying oil or for predicting reactivity.
14.2.1. Nonconjugated Drying Oils
A useful empirical relationship is that nonconjugated oils are drying oils if their drying
index is greater than 70 [3]. The drying index is calculated as follows:
drying index ¼ (%linoleic acid) þ 2(%linolenic acid)
Using this formula with the data from Table 14.1, the drying index of linseed oil is
120, so it is a drying oil; the drying index of soybean oil is 69, so it is a semi-drying
oil. Linoleic and linolenic acids have one and two diallylic methylene groups
(22CH55CHCH2CH55CH22) per molecule, respectively, which are the active groups
for initiating drying. Accordingly, drying is related more specifically to the average
number of diallylic groups per molecule. If this number is greater than about 2.2, the
oil is a drying oil; if it is moderately below 2.2, the oil is a semi-drying oil; there is no
sharp dividing line between semi-drying oils and non-drying oils. These statements
apply to synthetic drying oils as well as to natural oils. Since diallylic groups are the
sites for cross-linking, it is convenient to relate the average number of such groups per
molecule to the number average functionality f̄n of the triglyceride or synthetic drying oil.
298
DRYING OILS
The diallylic methylene groups, activated by their allylic relationship to two double
bonds, are much more reactive than methylene groups allylic to only one double bond,
as reflected by the relative rates of autoxidation of synthetic triolein (glyceryl trioleate),
trilinolein, and trilinolenein, which are 1 : 120 : 330, respectively [4]. The number of diallylic groups ( f̄n) of the three triglycerides are 0, 3, and 6, respectively; the theoretical
iodine values are 86, 173, and 262. The autoxidation rates are more closely related to
the number of diallylic groups between double bonds, f̄n, than to iodine values, which
are proportional to the average number of double bonds per molecule. Based on the data
in Table 14.1, the f̄n of the linseed oil is 3.6; it is a drying oil. The f̄n of soybean oil is
2.07; it is a semi-drying oil. The higher the f̄n of a drying oil, the more rapidly a
solvent-resistant cross-linked film forms on exposure to air.
The reactions taking place during drying are complex. Early studies of the chemistry
of drying were done without benefit of modern analytical instrumentation, which has
recently been brought to bear on this complex problem [5 – 7]. Cross-linked films
form from linseed oil in the following stages: (1) an induction period during which naturally present antioxidants (mainly tocopherols) are consumed, (2) a period of rapid
oxygen uptake with a weight gain of about 10% (FTIR shows an increase in hydroperoxides and appearance of conjugated dienes during this stage), and (3) a complex
sequence of autocatalytic reactions in which hydroperoxides are consumed and crosslinked film is formed. In one study, steps 1, 2, and 3 were far along in 4, 10, and 50
hours, respectively, when catalyzed by a drier (Section 14.2.2) [5]. Cleavage reactions
to form low molecular weight by-products also occur during the later stages of film formation. Slow continuing cleavage and cross-linking reactions through the lifetime of the
film lead to embrittlement, discoloration, and formation of volatile by-products. Oils
with significant quantities of fatty acids having three double bonds, such as linolenic
acid, discolor to a particularly marked degree.
The following scheme illustrates some of the many reactions that occur during
cross-linking. Initially, naturally present hydroperoxides decompose to form free
radicals.
At first, these highly reactive free radicals react mainly with the naturally present antioxidant. But as the antioxidant is consumed, the free radicals react with other compounds.
The diallylic methylene hydrogens are particularly susceptible to abstraction, yielding the
resonance-stabilized free radical (1):
Free radical 1, in which the free electron is delocalized with high electron density on the
terminal carbons, reacts with oxygen to give predominantly a conjugated peroxy free
14.2. AUTOXIDATION AND CROSS-LINKING
299
radical such as 2:
The peroxy free radicals can abstract hydrogens from other diallylic methylene groups
to form additional hydroperoxides and generate free radicals such as 1. Thus, a chain reaction is established, resulting in autoxidation. At least part of the cross-linking occurs by
radical – radical combination reactions forming C22C, ether, and peroxide bonds. These
reactions correspond to termination by combination reactions in free-radical chain-growth
polymerization (Section 2.2.1).
Reactions analogous to the addition step in chain-growth polymerization could also
produce cross-links. For example, free radicals may add to conjugated double bonds,
either initially present or formed from oxidation of diallylic groups, as in intermediate
2, to give 3.
Such addition reactions could yield C22C or C22O linkages, depending on the structure
of R.. Subsequently, free radical 3 can add oxygen to form a peroxy free radical, abstract a
hydrogen from a diallylic methylene group, combine with another free radical, or add to a
conjugated double bond system.
Studies of the catalyzed reactions of ethyl linoleate with oxygen (Section 14.2.2) by 1H
and 13C NMR indicated that the predominant cross-linking reactions were those that
formed ether and peroxy cross-links [6,7]. Mass spectroscopic studies showed that only
about 5% of the cross-links were new C22C bonds [6]. However, Fourier transform
infrared spectroscopy (FTIR) and FT Raman analysis of drying linseed oil lead to the
conclusion that only C22O22C and C22C bonds were formed [8], which illustrates the
difficulty of analyzing products in these systems. Substantial levels of epoxy groups
were detected in the reaction mixture, rising to a maximum in about 5 days and virtually
disappearing in 100 days [8]. It is suggested that epoxy groups may react with carboxyl
groups from oxidation of aldehydes to form ester cross-links.
Rearrangement and cleavage of hydroperoxides to aldehydes and ketones, among other
products, lead to low molecular weight by-products. It has also been shown that low MW
carboxylic acids are released from linseed oil films as they age [9]. The characteristic odor
300
DRYING OILS
of oil and alkyd paints during drying is attributable to such volatile by-products, as well as
to the odor of organic solvents.
Undesirable odor has been a factor motivating replacement of oil and alkyds in paints
with latex, particularly for interior applications. The reactions leading to these odors
have been studied extensively in connection with flavor changes of vegetable cooking
oils [10]. Aldehydes have been shown to be major by-products from the catalyzed
autoxidation of methyl oleate, linoleate, and linolenate, as well as from curing of
drying oil-modified alkyd resins [5,11]. It has also been shown that C9 acid esters remain
in the nonvolatile reaction mixture [11]. It has been proposed that formation of aldehydes, ketones, and carboxylic acids in the aging of films of linseed oil result from
photoxidation [12].
Dried films, especially of oils with three double bonds, yellow with aging. The yellow
color bleaches significantly when exposed to light; hence, yellowing is most severe when
films are covered, such as by a picture hanging on a wall. The reactions leading to color are
complex and are not fully understood. Yellowing has been shown to result from incorporation of nitrogen compounds and is increased markedly by exposure to ammonia. It has
been proposed that ammonia reacts with 1,4-diketones formed in autoxidation to yield
pyrroles, which oxidize to yield highly colored products [13]. A review of the causes of
yellowing concludes that previous suggestions do not correspond to expectations from
the mechanisms of drying of oils. The authors also show that yellowing of poppy seed
oil is almost as severe as that of linseed oil, yet poppy seed oil contains very few fatty
acids with three double bonds. It was proposed that yellowing results from contaminants
in the oils but specific contaminants were not identified [14].
14.2.2. Catalysis of Autoxidation and Cross-Linking
The rates at which uncatalyzed nonconjugated drying oils dry are slow. Many years ago, it
was found that metal salts (driers) catalyze drying. The most widely used driers are oilsoluble cobalt, manganese, lead, zirconium, and calcium salts of 2-ethylhexanoic acid
or naphthenic acids. Salts of other metals, including rare earths, are also used. Despite
many studies, there is not agreement as to the mechanisms by which driers work.
Cobalt and manganese salts, called top driers or surface driers, primarily catalyze
drying at the film surface. Lead and zirconium salts catalyze drying throughout the film
and are called through driers. Calcium salts show little, if any, activity alone, but reduce
the amount of other driers needed. The surface-drying catalysis by cobalt is reported to
accelerate the reaction of oxygen with methylene groups to form hydroperoxides;
whereas manganese salts are reported, primarily, to accelerate decomposition of hydroperoxides to free radicals [15].
The net result is formation of water and free radicals. Note that cobalt cycles between the
two oxidation states. The activity of through driers has not been adequately explained.
14.2. AUTOXIDATION AND CROSS-LINKING
301
Combinations of metal salts are almost always used. Mixtures of lead with cobalt and/or
manganese are particularly effective, but as a result of toxicity control regulations, lead driers
can no longer be used in consumer paints sold in interstate commerce in the United States.
Combinations of cobalt and/or manganese with zirconium, frequently including calcium, are
commonly used. Zirconium may inhibit the deactivation of cobalt by carboxylic acids.
A cobalt–zirconium combination is said to be particularly effective. Cobalt is said to primarily catalyze oxidation rather than polymerization, whereas a 1 : 3 Co–Zr complex increased
both oxidation and polymerization. Zirconium was reported to be more effective than lead
[16]. In a study of driers for high-solids alkyds, a combination of cobalt and neodymium
salts with bipyridyl was found to be most suitable [17]. 1,10-Phenanthroline is also used
to accelerate the activity of cobalt and manganese driers.
It has been reported that cobalt and cobalt salts may be carcinogenic and genotoxic.
Model compound studies were carried out to find replacements. A tetranuclear cluster
[Mn4O2(2-ethylhexanoate)6(bipyridine)2] and a complex of manganese acetylacetonate/
2,20 -bipyridine were found to be very effective [18,19].
Calcium does not undergo redox reactions; it has been suggested that it may promote
drying by adsorbing preferentially on pigment surfaces, minimizing the adsorption of
active driers. The amounts of driers needed are system specific. Their use should be
kept to the minimum possible level, since they not only catalyze drying, but also catalyze
the reactions that cause postdrying embrittlement, discoloration, and cleavage.
Skinning occurs when coatings prepared from drying oils or oxidizing alkyds are
exposed to the atmosphere, especially when driers have been added. To control antiskinning, additives are required. (Antiskinning additives and a comparison of their effectiveness are reviewed in Ref. [20].) The most effective antiskinning agent was found to be
methyl ethyl ketone oxime (MEKO). MEKO forms a complex with cobalt ions,
[Co(MEKO)1– 8]3þ. The complex is in equilibrium with the components so that while it
is present it reduces the effectiveness of the cobalt, but when paint is applied, the
MEKO evaporates relatively rapidly, reversing the equilibrium releasing the cobalt. The
MEKO is added to the coating just before filling into containers. In some cases a small
amount of a solution of MEKO is added after the containers are filled and just before
they are sealed, to prevent skinning during storage. Addition of 0.2% of MEKO prevents
skinning for more than 250 days. In one study, MEKO increased the dust-free time of an
alkyd coating to 2 hours from 1 hour 45 minutes, but reduced the fully dried time from
4 hours 45 minutes to 4 hours. The initial effect results from the time for the MEKO to
evaporate completely. The fast full drying may result from faster permeation of oxygen
in the earlier stages after application while MEKO is inhibiting surface cure.
14.2.3. Conjugated Drying Oils
Oils containing conjugated double bonds, such as tung oil, dry more rapidly than any nonconjugated drying oil. Drying of the ethyl ester of conjugated linoleic acid has been
studied. Higher oligomers were formed than under comparable conditions with the nonconjugated ethyl linoleate. It was found that the initial cross-links were peroxide links,
which, over time, became ether cross-links [21]. In general, the water and alkali resistance
of films derived from conjugated oils are superior, presumably because more of the crosslinks are stable ether bonds. However, since the a-eleostearic acid in tung oil has three
double bonds, with baking and aging, discoloration is severe.
302
DRYING OILS
14.3. SYNTHETIC AND MODIFIED DRYING OILS
Several types of chemical modification of drying oils are in use. In addition to those discussed here, drying oil-modified alkyds, epoxy esters, and uralkyds are discussed in
Chapter 15.
14.3.1. Heat-Bodied Oils, Blown Oils, and Dimer Acids
Both nonconjugated and conjugated drying oils can be thermally polymerized by heating
under an inert atmosphere to form bodied oils. Bodied oils have higher viscosities and are
used in oil paints to improve application and performance characteristics. Process temperatures may be as high as 300 to 3208C for nonconjugated oils and 225 to 2408C for
conjugated oils, although the reactions occur at an appreciable rate at somewhat lower
temperatures. At least in part, bodying may result from thermal decomposition of hydroperoxides, always present in natural oils, to yield free radicals, resulting in a limited degree
of cross-linking. It has also been shown that thermal rearrangement to conjugated systems
occurs followed by Diels –Alder reactions, which lead to formation of dimers [22]. Since
tung oil has a high concentration of conjugated double bonds, it undergoes thermal
polymerization more rapidly than nonconjugated oils (e.g., linseed oil). Unless one is
careful to control the heating of tung oil, the polymerization will lead to gelation.
Viscosity can also be increased by passing air through drying oils at relatively moderate
temperatures, 140 to 1508C, to produce blown oils. Presumably, reactions similar to those
involved in cross-linking cause autoxidative oligomerization of the oil. The effects of
temperature on blowing have been investigated. It is reported that the blown oils are
shear thinning [23].
Polyunsaturated acids dimerize or oligomerize by heat treatment; the reactions are acid
catalyzed. For example, the doubly unsaturated fatty acids of TOFA can be dimerized or
oligomerized followed by removal of the residual monobasic acids by distillation. Under
the high reaction temperatures, some decarboxylation occurs. Decarboxylation can be
minimized by heating under pressure in the presence of a small percentage of water and
activated clay. The products obtained are called dimer acids. They are predominantly
C36 dicarboxylic acids, with small fractions of monocarboxylic acid and some C54 tricarboxylic acids. For some uses it is desirable to eliminate residual double bonds by hydrogenation. Dimer acids are used to make polyesters (Section 10.1.2) and polyamides
(Section 13.2.2).
14.3.2. Varnishes
The drying rate of drying oils can be increased by dissolving a solid resin in the oil and diluting with a hydrocarbon solvent. Such a solution is called a varnish. The solid resin serves to
increase the Tg of the solvent-free film so that film hardness is achieved more rapidly. The
rate of cross-linking does not increase, so the time required for the film to become solvent
resistant is not shortened. Any high melting thermoplastic resin soluble in drying oil will
serve the purpose. The higher the “melt point” of the resin, the greater the effect on
drying time. Naturally occurring resins such as congo, copal, damar, and kauri resins, synthetic resins such as ester gum (glyceryl esters of rosin), phenolic resins (Section 13.6), and
coumarone–indene resins have been used.
14.3. SYNTHETIC AND MODIFIED DRYING OILS
303
In varnish manufacture, the drying oil (usually linseed oil, tung oil, or mixtures of the
two) and the resin are cooked together to high temperature to obtain a homogeneous solution of the proper viscosity. The varnish is then thinned with hydrocarbon solvents to
application viscosity. During cooking, some dimerization or oligomerization of the
drying oil occurs; in some cases, reaction between the oil and resin has been demonstrated.
Varnishes were widely used in the nineteenth and early twentieth centuries, but have
been mostly replaced by a variety of other products, especially alkyds, epoxy esters,
and uralkyds. The term varnish has come to be used more generally for transparent coatings, even though few of them today are varnishes in the original meaning of the word.
14.3.3. Synthetic Conjugated Oils
Tung oil dries rapidly but is expensive; and its films discolor rapidly due to the presence of
three double bonds. These shortcomings led to efforts to synthesize conjugated oils,
especially those containing esters of fatty acids with two conjugated double bonds. One
approach is to dehydrate castor oil using acid catalysis. A major component (87%) of
the fatty acid content of castor oil triglycerides is ricinoleic acid, 12-hydroxy-(Z )-9octadecenoic acid, which dehydrates to mixed geometric isomers of 9,11-conjugated
and 9,12-non-conjugated fatty acid esters. Dehydrated castor oil dries relatively
rapidly at room temperature, but on further exposure to air, the surface becomes tacky.
This after tack has been attributed to the presence of various geometric isomers
formed during dehydration. Dehydrated castor oil and its fatty acids are used
primarily to prepare alkyds and epoxy esters for baking coatings, in which after tack
does not occur.
Non-conjugated oils can be partially isomerized to conjugated oils by heating with a
variety of catalysts, mainly alkaline hydroxides. A similar process can be used to partially
conjugate double bonds of TOFAs. Synthesis of conjugated fatty acids by treatment of oils
at high temperature with aqueous alkali hydroxides accomplishes isomerization and saponification simultaneously [24]. The principal use of such conjugated oils and fatty acids
has been to make alkyds and epoxy esters.
14.3.4. Esters of Higher Functionality Polyols
When oil-derived fatty acids are reacted with polyols having more than three hydroxyl
groups per molecule, the number of cross-linking sites per molecule increases relative
to the corresponding natural triglyceride oil. Whereas soybean oil is a semi-drying oil
since f¯n is 2.07, the pentaerythritol (PE) tetraester of soybean fatty acids has an f¯n value
of 2.76 and is a drying oil. The PE ester of linseed fatty acids has an f¯n value of about
5 and gives dry, solvent-resistant films more rapidly than does linseed oil. Still faster
drying rates can be achieved with still higher functionality polyols, such as di- and tripentaerythritol. Alkyds, epoxy esters, and uralkyds (Chapter 15) made with fatty acids from
such oils as soybean and linseed oils can be considered as higher functionality synthetic
drying oils.
14.3.5. Maleated Oils
Both conjugated and nonconjugated oils react with maleic anhydride to form adducts.
Conjugated oils such as dehydrated castor oil react at moderate temperatures by a
304
DRYING OILS
Diels –Alder reaction:
Non-conjugated oils such as soybean and linseed oils require higher temperatures
(.1008C) and form a variety of adduct structures. Model compound studies using
methyl linoleate demonstrate that maleic anhydride undergoes an ene reaction to give
succinyl anhydride adducts. The ene reaction results in conjugated bonds, as shown in
general structure (4); subsequent Diels –Alder reaction with a second maleic anhydride
gives a dianhydride, as shown in general structure (5) [25].
The products of these reactions, termed maleated oils, or sometimes, maleinized oils,
react with polyols to give moderate molecular weight derivatives that dry faster than
unmodified drying oils. For example, maleated soybean oil esterified with glycerol dries
at a rate comparable to that of a bodied linseed oil with a similar viscosity.
Maleated linseed oil can be made water-reducible by hydrolysis with aqueous ammonia
to convert the anhydride groups to the ammonium salts of the diacid. Such products have
not found significant commercial use, but the process is used to make water-reducible
alkyds and epoxy esters (Sections 15.3 and 15.8).
14.3.6. Vinyl-Modified Oils
Both conjugated and non-conjugated drying oils react in the presence of a free radical
initiator with such unsaturated monomers as styrene, vinyltoluene, and acrylic esters. High
degrees of chain transfer result in the formation of a variety of products: for example, low
molecular weight polymers of the monomer or mixture of monomers, short-chain graft
copolymers, and dimerized drying oil molecules. The reaction of drying oils with such
monomers is not commercially important, but the same principle is used to make modified
alkyds (Section 15.6). Linseed oil modified with cyclopentadiene has found fairly sizable
commercial use. This product is made by heating a mixture of linseed oil and dicyclopentadiene above 1708C, at which monomeric cyclopentadiene is liberated at an appreciable rate
by the reverse Diels–Alder reaction. The product is inexpensive and dries faster than linseed
oil. However, its odor and dark color limit its applications.
REFERENCES
305
GENERAL REFERENCES
Fox, F. L., Oils for Organic Coatings, Federation of Societies for Coatings Technology, Blue Bell,
PA, 1965.
Rheineck, A. E.; Austin, R. O., Drying oils, in Myers, R. R.; Long, J. S., Eds., Treatise on Coatings,
Marcel Dekker, New York, 1968, Vol. I, No. 2, pp. 181 – 248.
Wicks, Z. W., Jr., Drying oils, in Kirk– Othmer Encyclopedia of Chemical Technology, 5th ed.,
Wiley, New York, 2003.
REFERENCES
1. Ackman, R. G., Prog. Chem. Fats Other Lipids, 1972, 12, 165. Khan, G. R.; Scheinmann, F.,
Prog. Chem. Fats Other Lipids, 1977, 15, 343. King, J. W.; et al., J. Liq. Chromatogr., 1982,
5, 275.
2. Rheineck, A. E.; Austin, R. O., in Treatise on Coatings, Myers, R. R.; Long, J. S., Eds., Marcel
Dekker, New York, 1968, Vol. 1, No. 2, pp. 181 – 248.
3. Greaves, J. H., Oil Colour Trades J., 1948, 113, 949.
4. Chipault, J. R.; et al., Off. Digest, 1951, 23, 740.
5. Hartshorn, J. H., J. Coat. Technol., 1982, 54(687), 53.
6. Muizebelt, W. J.; et al., J. Coat. Technol., 1998, 70(876), 83.
7. Falla, N. A. R., J. Coat. Technol., 1992, 64(815), 55.
8. Mallegol, J.; et al., J. Am. Oil Chem. Soc., 1999, 76, 967; 2000, 77, 249.
9. van den Berg, J. D. J.; et al., Prog. Org. Coat., 2001, 41, 143.
10. Frankel, E. N., Prog. Lipid. Res., 1980, 19, 1.
11. Hancock, R. A.; et al., Prog. Org. Coat., 1989, 17, 321, 337.
12. Mallegol, J.; et al., J. Am. Oil Chem. Soc., 2000, 77, 257.
13. Robey, T. L. T.; Rybicka, S. M., Paint Res. Sta. Tech. Paper 217, 1962, 13(1), 2.
14. Mallegol, J.; et al., Stud. Conserv., 2001, 46, 121.
15. Vereholantsev, V., Eur. Coat. J., 2000, (1 – 2), 120.
16. Meneghetti, S. M. P.; et al., Prog. Org. Coat., 1998, 33, 219.
17. Hein, R. W., J. Coat. Technol., 1999, 71(898), 21.
18. Warzeska, S. T.; et al., Prog. Org. Coat., 2002, 44, 243.
19. Oyman, Z. O.; et al., Polymer, 2005, 46, 1731.
20. Bieleman, J., Surf. Coat. Int. A, 2003, (10), 411.
21. Muizebelt, W. J.; et al., Prog. Org. Coat., 2000, 40, 121.
22. Wheeler, D. H.; White, J., J. Am. Chem. Soc., 1967, 44, 298.
23. Gueler, O. K.; et al., Prog. Org. Coat., 2004, 51, 365.
24. Bradley, T. F.; Richardson, G. H., Ind. Eng. Chem., 1942, 34, 237.
25. Rheineck, A. E.; Khoe, T. H., Fette Seifen Anstrichm., 1969, 71, 644.
15
Alkyd Resins
Although no longer the largest volume vehicles in coatings, alkyds are still of major
importance. In 2002, $410 million worth of alkyds were used in U.S. coatings [1].
Alkyds are prepared from polyols, dibasic acids, and fatty acids. They are polyesters,
but in the coatings field the term polyester is reserved for “oil-free polyesters” (i.e.,
those free of fatty acids), discussed in Chapter 10. The term alkyd is derived from
alcohol and acid. Alkyds tend to be lower in cost than most other vehicles and tend
to give coatings that exhibit fewer film defects during application (Chapter 24).
However, the durability of alkyd films, especially outdoors, tends to be poorer than that
of films from acrylics, polyesters, and polyurethanes, although better than that of
BPA epoxies.
There are many types of alkyds. One classification is into oxidizing and nonoxidizing
types. Oxidizing alkyds cross-link by the same mechanism as drying oils, as discussed
in Chapter 14. Nonoxidizing alkyds are used as polymeric plasticizers or as hydroxyfunctional resins, which are cross-linked by melamine – formaldehyde (MF), or urea –
formaldehyde (UF) resins or by isocyanate cross-linkers. A second classification is
based on the ratio of monobasic fatty acids to dibasic acids utilized in their preparation,
which was adapted from the terminology used to classify varnishes. Varnishes with
high ratios of oil to resin were called long oil varnishes; those with a lower ratio,
medium oil varnishes; and those with an even lower ratio, short oil varnishes. The oil
length of an alkyd is calculated by dividing the amount of “oil” in the final alkyd by
the total weight of the weight of all ingredients minus water evolved in reaction, expressed
as a percentage. The amount of oil is defined as the triglyceride equivalent to the amount of
fatty acids in the alkyd. The 1.04 factor in the second equation converts the weight of fatty
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
306
15.1. OXIDIZING ALKYDS
307
acids to the corresponding weight of triglyceride oil:
oil length ¼
weight of “oil”
100
weight of all ingredients water evolved
oil length ¼
1:04 weight of fatty acids
100
weight of all ingredients water evolved
Alkyds with oil lengths greater than 60 are long oil alkyds, those with oil lengths from 40
to 60 are medium oil alkyds, and those with oil lengths less than 40 are short oil alkyds.
There is some variation in the literature in the dividing lines between these classes.
Another classification is unmodified or modified alkyds. Modified alkyds contain other
monomers in addition to polyols, polybasic acids, and fatty acids. Examples are styrenated
alkyds (Section 15.6) and silicone alkyds (Section 16.1.2).
15.1. OXIDIZING ALKYDS
Oxidizing alkyds can be considered as synthetic drying oils. They are polyesters of one or
more polyols, one or more dibasic acids, and fatty acids from one or more drying or semidrying oils. The most commonly used polyol is glycerol, the most commonly used dibasic
acid is phthalic anhydride (PA), and a widely used oil is soybean oil. Let us consider a
simple idealized example of the alkyd prepared from 1 mol of PA, 2 mol of glycerol,
and 4 mol of soybean fatty acids. Using the fatty acid composition data for soybean oil
from Table 14.1, we can calculate that this alkyd would have an f̄n of 2.76 [i.e., it
would have an average of 2.76 activated diallylic groups (22CH55CHCH2CH55CH22)
per molecule (see Sections 14.2.1 and 14.3.4 for a discussion of f̄n)] and therefore
would dry to a solid film. The alkyd would form a solvent resistant film in about the
same time as a pentaerythritol (PE) ester of soybean fatty acids (Section 14.3.4) since
they have the same f̄n. However, the alkyd would form a tack-free film faster because
the rigid aromatic rings from PA increase the Tg of the film.
If the mole ratio of PA to glycerol were 2 : 3, corresponding to an equivalent ratio of 4 : 9,
5 mol of soybean fatty acid could be esterified to yield an alkyd with an f̄n of 3.45. This alkyd
would cross-link more rapidly than the 1 : 2 : 4 mole ratio alkyd and would also form tackfree films even faster because the ratio of aromatic rings to long aliphatic chains would be
2 : 5 instead of 1 : 4. As the ratio of PA to glycerol is increased further, the average functionality for autoxidation increases and the Tg after solvent evaporation increases because of the
increasing ratio of aromatic to long aliphatic chains. For both reasons, films dry faster.
A theoretical alkyd prepared from 1 mol each of glycerol, PA, and fatty acid would
have an oil length of about 60. However, if one were to try to prepare such an alkyd,
the resin would gel prior to complete reaction. Gelation would result from reaction of a
sufficient number of trifunctional glycerol molecules with three difunctional PA molecules
to form cross-linked polymer molecules, swollen with partially reacted components.
Gelation can be avoided by using a sufficient excess of glycerol to reduce the extent of
cross-linking. When the reaction is carried to near completion with excess glycerol,
there are few unreacted carboxylic acid groups, but many unreacted hydroxyl groups.
Alkyds are manufactured in batch step-growth polymerization processes (Section
15.5), usually at temperatures above 2008C. It is often desired to adjust monomer
308
ALKYD RESINS
compositions and process conditions so that the polymer is approaching gelation at the end
of the process, but if it gels in the kettle, a very expensive cleanup job is required. There
have been many attempts, none fully successful, to calculate the ratios of functional groups
and the extent of reaction that can be reached without encountering gelation. The problem
is complex. The reactivity of the hydroxyl groups can be different; for example, glycerol
contains both primary and secondary alcohol groups. Under esterification conditions,
polyol molecules can self-condense to form ethers and in some cases can dehydrate to
form volatile aldehydes (Section 15.5.2). Reactivity of the carboxylic acids also varies.
The rate of formation of the first ester from a cyclic anhydride is more rapid than formation
of the second ester. Aliphatic acids esterify more rapidly than aromatic acids. Polyunsaturated fatty acids and their esters can dimerize or oligomerize to form cross-links. Of
the many papers in the field, that by Blackinton recognizes the complexities best [2]. In
addition to the foregoing complexities, particular emphasis is placed on the extent of formation of cyclic compounds by intramolecular esterification reactions.
In practice, alkyd resin formulators have found that the mole ratio of dibasic acid to
polyol should be less than 1 to avoid gelation. How much less than 1 depends on many
variables. Composition variables are discussed in Sections 15.1.1 through 15.1.3; the
effect of variables in reaction conditions is discussed in Section 15.5.
For medium oil alkyds, the ratio of dibasic acid to polyol is not generally changed much
relative to alkyds with an oil length of about 60, but the fatty acid content is reduced to the
extent desired. This results in a larger excess of hydroxyl groups in the final alkyd. It is
commonly said that as the oil length of an oxidizing alkyd is reduced below 60, the
drying time decreases to a minimum at an oil length of about 50. However, this conventional wisdom must be viewed cautiously. The ratio of aromatic rings to aliphatic chains
continues to increase, increasing Tg after the solvent evaporates from the film tending to
shorten the time to form a tack-free film. However, at the same molecular weight (MW),
the number of fatty acid ester groups per molecule decreases as the oil length decreases
below 60, since more hydroxyl groups are left unesterified. Therefore, the time required
to achieve sufficient cross-linking for solvent resistance increases.
Long oil alkyds are soluble in aliphatic hydrocarbon solvents. As the oil length decreases,
mixtures of aliphatic and aromatic solvents are required, and oil lengths below about 50
require aromatic or oxygenated solvents, which are more expensive than aliphatics. The viscosity of solutions of long oil alkyds, especially of those with oil lengths below 65, is higher
in aliphatic than in aromatic solvents; in medium oil alkyds, which require mixtures of aliphatic and aromatic solvents, viscosity decreases as the proportion of aromatic solvents
increases. In former days, and to some extent still, it was considered desirable to use a
solvent mixture that gave the highest possible viscosity; then at application viscosity, the
solids were lower and the raw material cost per unit volume was less. Accordingly,
alkyds were designed to have high dilutability with aliphatic solvents. This was false
economy, but it was a common practice and is still being practiced to some extent. Increasingly, the emphasis is on reducing VOC, so the question becomes how to design alkyds with
low solvent requirements rather than high dilutability potential. Furthermore, the aromatic
solvents are on the HAP list. High-solids alkyds are discussed in Section 15.2.
15.1.1. Monobasic Acid Selection
Drying alkyds can be made with fatty acids from semi-drying oils, since the f̄n can be well
above 2.2. For alkyds made by the monoglyceride process (Section 15.5.1), soybean oil is
15.1. OXIDIZING ALKYDS
309
used in the largest volume. Soybean oil is economical and supplies are dependable because
it is a large scale agricultural commodity; alkyd production takes only a few percent of the
world supply. For alkyds made by the fatty acid process (Section 15.5.1), tall oil fatty acids
(TOFA) are, at times, more economical than soybean fatty acids. Both soybean oil and
TOFA contain roughly 40 to 60% linoleic acid and significant amounts of linolenic
acid (Table 14.1). White coatings containing linolenic acid esters gradually turn yellow.
Premium cost “nonyellowing” alkyds are made with safflower or sunflower oils, which
are high in linoleic acid but contain very little linolenic acid.
Applications in which fast oxidative drying and high cross-link density are important
require alkyds made with drying oils. The rate of oxidative cross-linking is affected by
the functionality of the drying oils used. At the same oil length and MW, the time required
to achieve a specific degree of cross-linking decreases as the average number of diallylic
groups ( f̄n) increases. Linseed long oil alkyds therefore cross-link more rapidly than
soybean long oil alkyds. The effect is especially large in very long oil alkyds and less
noticeable in alkyds with oil lengths around 60, where f̄n is very high even with soybean
oil, and the effect of further increase in functionality by using linseed oil is small.
Because of the large fraction of esters of fatty acids with three double bonds in linseed
alkyds, their color and color retention is poorer than that of soybean alkyds. Tung oil –
based alkyds, because of the high proportion of esters with three conjugated double
bonds, dry still faster. Tung oil alkyds also exhibit a high degree of yellowing. Dehydrated
castor alkyds have fairly good color retention, since they contain only a small proportion of
esters of fatty acids with three double bonds; they are used primarily in baking coatings.
As discussed in Section 14.3.1, drying oils and drying oil fatty acids undergo dimerization at elevated temperatures. Dimerization occurs concurrently with esterification during
alkyd synthesis; it generates difunctional acids, increasing the mole ratio of dibasic acids
to polyol. The rate of dimerization is faster with drying oils having a higher average
number of diallylic groups per molecule and with those having conjugated double
bonds. Thus, the MW, and therefore the viscosity, of an alkyd made with the same ratio
of reactants depend on the fatty acid composition. The higher the degree of unsaturation,
the higher the viscosity, due to the greater extent of dimerization. Linseed alkyds have
higher viscosities than soybean alkyds made with the same monomer ratios under the
same conditions. The effect is particularly marked with tung oil. It is difficult to prepare
straight tung alkyds because of the risk of gelation; commonly, mixed linseed-tung
alkyds are made when high oxidative cross-linking functionality is desired.
A critical factor involved in the choice of fatty acid is cost. Drying oils are agricultural
products and hence tend to be volatile in price. By far the major market for vegetable oils
is for foods. Depending on relative prices, one drying oil is often substituted for another in
certain alkyds. By adjusting for functionality differences, substitutions can frequently be
made without significant changes in properties.
Benzoic acid is also used, especially to esterify some of the excess hydroxyl groups
remaining in the preparation of medium oil alkyds. The benzoic acid increases the ratio
of aromatic to aliphatic chains in the alkyd, thus contributing to a higher Tg of the solventfree alkyd and more rapid formation of a tack-free film. At the same time, the reduction in
the free hydroxyl content may somewhat reduce the water sensitivity of the dried films.
Rosin can also be used in the same fashion. Although rosin is not an aromatic acid,
its polynuclear ring structures are rigid enough to increase Tg. If the critical requirement
in drying is rapid development of solvent resistance, such benzoic acid and rosin modifications do not serve the purpose; they only reduce tack-free time. Frequently, benzoic
310
ALKYD RESINS
acid –modified alkyds are called chain-stopped alkyds. The implication of the terminology
is that the benzoic acid stops chain growth. This is not the case; the benzoic acid simply
esterifies hydroxyl groups that would not have been esterified if the benzoic acid were
absent. The effect on degree of polymerization is negligible.
15.1.2. Polyol Selection
Glycerol is the most widely used polyol because it is present in naturally occurring oils from
which alkyds are commonly synthesized (Section 15.5.1). The next most widely used polyol
is pentaerythritol (PE). To avoid gelation, the tetrafunctionality of PE must be taken into
account when replacing glycerol with PE. If the substitution is made on a mole basis
rather than an equivalent basis, chances for gelation are minimized. As mentioned earlier,
the ratio of moles of dibasic acid to polyol should be less than 1, and generally, a slightly
lower mole ratio is required with PE than with glycerol. At the same mole ratio of dibasic
acid to polyol, more moles of fatty acid can be esterified with PE. Hence, in long oil
alkyds, the average functionality for cross-linking is higher, and the time to reach a given
degree of solvent resistance is shorter for a PE alkyd than for a glycerol alkyd. Because of
this difference, one must be careful in comparing oil lengths of glycerol and PE alkyds.
When PE is synthesized, di- and tripentaerythritol are by-products, and commercial PE
contains some of these higher polyols. Consequently, care must be exercised in changing
sources of PE, since the amount of the higher polyols may differ. Because of the very high
functionality, diPE and triPE (F ¼ 6 and 8, respectively) are useful in making fast drying
low MW alkyds (Section 15.2).
Styrene and allyl alcohol can copolymerize to form a high-functionality polyol for
alkyds. For example, it can be used in combination with PE to make long-oil alkyds
with relatively low viscosities and good hardness and humidity resistance [3].
To reduce cost, it is sometimes desirable to use mixtures of PE and ethylene or propylene glycol. A 1 : 1 mole ratio of tetra- and difunctional polyols gives an average functionality of 3, corresponding to glycerol. The corresponding alkyds can be expected to be
similar, but not identical. Trimethylolpropane (TMP) is useful, but the rate of esterification
15.1. OXIDIZING ALKYDS
311
is slower than with glycerol. Although all of TMP’s alcohol groups are primary, they are
somewhat sterically hindered by the neopentyl structure [4]. Trimethylolpropane however,
gives a narrower MW distribution, which provides alkyds with a somewhat lower viscosity
than the comparable glycerol-based alkyd. A kinetic study demonstrated that esterification
of one or two of the hydroxyl groups of TMP has little effect on the rate constant for
esterification of the third hydroxyl group [5]. It can be speculated that PE behaves
similarly.
15.1.3. Dibasic Acid Selection
Dibasic acids used to prepare alkyds are usually aromatic. Their rigid aromatic rings
increase the Tg of the resin. Cycloaliphatic anhydrides, such as hexahydrophthalic anhydride, are also used. Although they are not as rigid as aromatic rings, the cycloaliphatic
rings also increase Tg.
By far the most widely used dibasic acid is PA. It has the advantage that the first esterification reaction proceeds rapidly by opening the anhydride ring. The amount of water
evolved is lower, which also reduces reaction time. The relatively low melting point
(the pure compound melts at 1318C) is desirable, since the crystals melt and dissolve
readily in the reaction mixture. In large scale manufacturing, molten PA is used, which
reduces packaging, shipping, and handling costs.
The next most widely used dibasic acid is isophthalic acid (IPA). As discussed in
Section 10.1.2, esters of IPA are more resistant to hydrolysis than are those of PA in
the pH range 4 to 8, the most important range for exterior durability. On the other
hand, under alkaline conditions, esters of phthalic acid are more resistant to hydrolysis
than isophthalic esters. The raw material cost for IPA is not particularly different from
PA (even after adjusting for the extra mole of water that is lost), but the manufacturing
cost is higher. The high melting point of IPA (3308C) leads to problems getting it to dissolve in the reaction mixture so that it can react. High temperatures are required for longer
times than with PA; hence more dimerization of fatty acids occurs with TPA, resulting in
higher viscosity. The longer time at higher temperature also leads to greater extents of side
reactions of the polyol components, which are discussed in Section 15.6.2 [6]. Thus, when
substituting IPA for PA, one must use a lower mole ratio of IPA to polyol in order to make
an alkyd of similar viscosity.
Terephthalic acid is not useful in making alkyds because it dissolves very slowly even
at the highest feasible process temperatures. However, it can be used by transesterification
of its esters. Most notable is the use of recycled poly(ethylene terephthalate) (PET) in the
312
ALKYD RESINS
synthesis of alkyds. In a typical process [7], PE, a small amount of ethylene glycol, part of
the soy fatty acids, and dibutyltin oxide catalyst are heated until molten recycled PET is
added and the temperature raised to 2408C until the viscosity and acid number meet specifications. Properties of films from this resin and a comparable standard alkyd were essentially equal. Dimethyl terephthalate is also used.
15.2. HIGH-SOLIDS OXIDIZING ALKYDS
The need to minimize VOC emissions has led to efforts to increase the solids content of
alkyd resin coatings. Since xylene is on the HAP list, its use is being reduced. Some
increase in solids can be realized by a change of solvents. Aliphatic (and to a somewhat
lesser degree, aromatic) hydrocarbon solvents promote intermolecular hydrogen
bonding, especially between carboxylic acids, but also between hydroxyl groups,
thereby increasing viscosity. Use of at least some hydrogen-bond acceptor solvent, such
as a ketone or an ester, or hydrogen-bond acceptor-donor solvent such as an alcohol,
gives a significant reduction in viscosity at equal solids.
An approach to increasing solids is to decrease the MW, which is easily accomplished
by decreasing the dibasic acid/polyol ratio and going to longer oil length alkyds.
However, making a significant reduction in VOC by this route results in an alkyd with
lower functionality for cross-linking and a lower ratio of aromatic to aliphatic chains.
Both changes increase the time for drying. The effect of longer oil length on functionality
can be minimized by using drying oils with higher average functionality. Use of oils containing linolenic or a-eleostearic acid is limited by their tendency to discolor. One can use
safflower oil, which has a higher linoleic acid content, and less linolenic acid than soybean
oil. Proprietary fatty acids with 78% linoleic acid are commercially available. Increasing
the concentration of driers (Section 14.2.2) accelerates drying, but also accelerates yellowing and embrittlement.
Solids can be increased by making resins with narrower MW distributions. For example,
one can add a transesterification catalyst near the end of the alkyd cook; this gives a more
uniform MW and a lower viscosity product. To study the effect of distribution, model
alkyds with very narrow MW distribution were synthesized by using dicyclohexylcarbodiimide, which allows low temperature esterification [4]. With the same ratio of reactants, the
M̄n and polydispersity were lower than for conventional alkyd control. These differences
resulted from less dimerization through reactions of the double bond systems of the fatty
acids and avoidance of self-etherification of polyol in low temperature preparation. It
was found that the solids could be 2 to 10% higher than with the conventionally prepared
alkyd of the same raw material composition. The model alkyds dried more rapidly, but their
film properties, especially impact resistance, were inferior to those obtained with control
resins with the usual broad MW distribution [4]. Conventionally prepared TMP alkyds
had lower MWs and viscosities than the glycerol alkyds. This difference may result from
less self-etherification of TMP than with glycerol.
High-solids alkyds for baking applications have been made using tripentaerythritol.
The high functionality obtained using this polyol (F ¼ 8) gives alkyds that cross-link as
rapidly as shorter oil length higher viscosity glycerol alkyds [8]. However, for air dry
applications, the lower aromatic/aliphatic ratio lengthens the tack-free time.
Another approach to high-solids alkyds is to use reactive diluents in place of part of the
solvent. The idea is to have a component of lower MW and much lower viscosity than the
15.3. WATER-REDUCIBLE ALKYDS
313
alkyd resin, which reacts with the alkyd during drying, so it is not part of the VOC emissions. Use of reactive diluents in alkyd coatings is reviewed in Ref. [9]. The authors give a
list of key properties of a reactive diluent: low viscosity, good compatibility with alkyds,
low volatility (bp . 3008C), nontoxic, low color, economic replacement of solvent, and
of course, ability to participate in oxidative cure.
Mixed acrylic and drying oil fatty acid amides of hexa(aminomethoxymethyl)melamine are used [10,11]. Such reactive diluents contain high functionalities of .NCH2
NHCOCH55CH2 and .NCH2NHCOC17Hx moieties and promote fast drying. A combination of 2,7-octadienyl maleate and fumarate and 2-(2,7-octadienyloxy)succinate has
been reported to be a particularly effective reactive diluent [9]. Use of a reactive
diluent prepared by reacting drying oil fatty acids with excess dipentaerythritol and
then with isophorone diisocyanate has been patented [12]. Epoxidized soybean oil can
be used as a reactive diluent for high-solids alkyds using a combination of cobalt and
neodymium driers [13].
Using optimized resins and in some cases, reactive diluents, good quality air dry and
baking alkyd coatings can be formulated with VOC levels of 280 to 350 g L21 of
coating. A 250 g L21 level is attainable only with some sacrifice of application and film
properties; still lower limits of permissible VOC are projected. Lower VOC targets can
possibly be achieved using exempt solvents (18.9.1).
15.3. WATER-REDUCIBLE ALKYDS
As with almost all other resin classes, work has been done to make alkyd resins for coatings that can be reduced with water. One approach that has been studied more extensively
in Europe than in the United States is the use of alkyd emulsions [14,15]. (See Section 26.3
for further discussion.) The emulsions are stabilized with surfactants and can be prepared
with little, if any, volatile solvent. Solid alkyds can be emulsified in hot water with emulsifying agents. Pigmented coatings formulated with alkyd emulsions lose dry time on
storage because of absorption of cobalt drier on pigment surfaces and precipitation of
cobalt hydroxide. Possible solutions to the problem are to use a combination of cobalt neodecanoate and 2,20 -bipyridyl [16] or to emulsify the driers and the alkyd separately and
mix the emulsions so the alkyd and the drier are in separate phases [17].
It is common to add a few percent of an alkyd –surfactant blend to latex paints to
improve adhesion to chalky surfaces (Section 32.1) and in some cases, to improve
adhesion to metals (see Section 33.1.3). It is important to use alkyds that are as resistant
as possible to hydrolysis. Hybrid alkyd– acrylic latexes have been prepared by dissolving
an oxidizing alkyd in the monomers used in emulsion polymerization, yielding a latex with
an alkyd grafted on the acrylic polymer [18,19]. Hybrid alkyd – acrylic latexes have been
prepared by emulsion polymerization. Hydroperoxidized sunflower oil was used as the
initiator to polymerize a combination of sunflower oil, a long oil alkyd, and ethyl methacrylate. Films formed from the latex gave the fast dry expected from a latex followed by
autoxidation to give cross-linked films [20].
Another approach to hydrolytically stable waterborne alkyds is to make dispersions
with an alkyd core and an acrylic shell. They are made by polymerizing unsaturated
fatty acids with acrylic monomers, including acrylic acid, then reacting the product
with a hydroxy-terminated alkyd resin and neutralizing with NH4OH. The acrylic shell
reduces contact between the alkyd and the water, increasing hydrolytic stability. To
314
ALKYD RESINS
increase the hydrolytic stability further, it is recommended to use at least some polyol with
secondary hydroxyl groups [21]. Such dispersions can be essentially solvent-free.
Yet another approach is to make alkyds with an acid number in the range 35 to 50 using
secondary alcohols or ether alcohols as solvents. The acid groups are neutralized with
ammonia or an amine. The resulting solution can be diluted with water to form a dispersion of solvent swollen aggregates in water. Such resins are analogous to the waterreducible acrylics discussed in Section 8.3, and their behavior during water dilution is
similar. Note that the use of primary alcohol solvents must be avoided because they can
more readily transesterify with the alkyd during resin production and storage, leading to
reduction in MW and f̄n [22].
Hydrolytic instability can be a problem with water-reducible alkyds. If the carboxylic
acid groups are half esters of phthalic or trimellitic acid, the hydrolytic stability will be
poor and probably inadequate for paints that require a shelf life of more than a few
months. Due to the anchimeric effect of the neighboring carboxylic acid group, such
esters are relatively easily hydrolyzed. As hydrolysis occurs, the solubilizing acid salt is
detached from the resin molecules, and the aqueous dispersion loses stability. A more
satisfactory way to introduce free carboxylic acid groups is by reacting a completed
alkyd with maleic anhydride. Part of the maleic anhydride adds to the unsaturated fatty
acid esters as discussed in Section 14.3.5. The anhydride groups are then hydrolyzed
with amine and water to give the desired carboxylate salt groups, which are attached to
resin molecules with C22C bonds and cannot be hydrolyzed off. There is still a hydrolytic
stability problem with the alkyd backbone, but hydrolysis does not result in destabilization
of the dispersion. Similarly, acrylated fatty acids can be used to synthesize water-reducible
alkyds with improved hydrolytic stability [23].
After the film is applied, the water, solvent, and amine evaporate, and the film crosslinks by autoxidation. Since there are a fairly large number of residual carboxylic acid
groups left in the cross-linked binder, the water resistance, and particularly the alkali
resistance of the films are reduced but are still satisfactory for some applications [24].
Early water resistance can be a problem if, for example, a freshly painted surface is
rained on before all the amine has evaporated from the film. Commonly, ammonia is
used as the neutralizing amine because it is assumed that ammonia volatilizes faster
than any other amine. This assumption is not necessarily so; if the Tg of the alkyd film
is sufficiently high before all of the amine has volatilized, loss of amine becomes controlled by diffusion rate. The rate of diffusion of amine through the carboxylic acid–
functional film is affected by the base strength of the amine. A less basic amine, such
as morpholine, may leave the film before ammonia even though its volatility is considerably lower.
Superior resistance to hydrolysis during storage can be achieved by reacting the
hydroxyl groups of an alkyd with isophorone diisocyanate and dimethylolpropionic acid
to form an aqueous polyurethane dispersion [25]. (See Section 15.7.2 for further
discussion of autoxidizable polyurethane dispersions.)
15.4. NONOXIDIZING ALKYDS
Certain low MW short – medium and short oil alkyds are compatible with such polymers
as nitrocellulose and thermoplastic polyacrylates. Therefore, such alkyds can be used as
plasticizers for these polymers. They have the advantage over monomeric plasticizers
15.5. SYNTHETIC PROCEDURES FOR ALKYD RESINS
315
(e.g., dibutyl or dioctyl phthalate) that they do not volatilize appreciably when films are
baked. It is generally not desirable to use oxidizing alkyds, which would cross-link and
lead to embrittlement of the films, especially on exterior exposure. Therefore, non-drying
oil fatty acids (or oils) are used in the preparation of alkyds for such applications. For
exterior acrylic lacquers, pelargonic acid (n-C8H17COOH) alkyds combine excellent
resistance to photodegradation with good compatibility with thermoplastic acrylic
resins. An interesting sidelight on terminology is that these pelargonic alkyds have been
called polyesters rather than alkyds because the word polyester connotes higher quality
than the word alkyd. Castor oil– derived alkyds are particularly appropriate plasticizers
for nitrocellulose lacquers for interior applications, since the hydroxyl groups on the ricinoleic acid promote compatibility.
All alkyds, particularly short-medium oil and short oil alkyds, are made with a large
excess of hydroxyl groups, to avoid gelation. These hydroxyl groups can be cross-linked
with MF resins or with polyisocyanates. In some cases, relatively small amounts of MF
resin are used to supplement the cross-linking during baking of medium oil oxidizing
alkyds. To achieve compatibility, butylated MF resins are used. Such coatings provide
somewhat better durability and faster curing than those provided by alkyd resins alone,
with little increase in cost. The important advantage of the relative freedom from film
defects common to alkyd coatings can be retained (Chapter 24). However, the high levels
of unsaturation remaining in the cured films impart a tendency to discoloration on overbake
or exterior exposure and cause loss of gloss and embrittlement on exterior exposure. These
difficulties can be reduced by using non-drying oils with minimal levels of unsaturated fatty
acids. Coconut oil has been widely used; its performance can be further enhanced by hydrogenation of the small amount of unsaturated acids present in it.
Since isophthalic (IPA) esters are more stable to hydrolysis in the pH range 4 to 8, than
phthalate esters, the highest performance exterior alkyd– MF enamels use nonoxidizing
IPA alkyds. For many applications, alkyd –MF coatings have been replaced with
acrylic – MF or polyester –MF coatings to improve the overall balance of film properties.
Polyamide-modified alkyd resins are used as thixotropic agents to increase the low
shear viscosity of alkyd paints. For example, high-solids thixotropic alkyds based on polyamides made with aromatic diamines have been developed that give superior performance
in high-solids alkyd coatings [26].
15.5. SYNTHETIC PROCEDURES FOR ALKYD RESINS
Various synthetic procedures, each with many variables, are used to produce alkyd resins;
the general reference and Refs. [27] and [28] provide useful reviews of manufacturing
procedures. Alkyds can be made directly from oils or by using free fatty acids as raw
materials.
15.5.1. Synthesis from Oils or Fatty Acids
15.5.1.1. Monoglyceride Process In the case of glycerol alkyds, it would be absurd
to first saponify an oil to obtain fatty acids and glycerol, and then esterify the same groups
in a different combination. Rather, the oil is first reacted with sufficient glycerol to give the
total desired glycerol content, including the glycerol in the oil. Since PA is not soluble
in the oil but is soluble in the glycerol, transesterification of oil with glycerol must be
316
ALKYD RESINS
carried out as a separate step before the PA is added; otherwise, glycerol phthalate gel
particles would form early in the process. This two-stage procedure is often called the
monoglyceride process. The transesterification reaction is run at 230 to 2508C in the presence of a catalyst; many catalysts have been used. Before the strict regulation of lead in
coatings, litharge (PbO) was widely used; the residual transesterification catalyst also
acted as a drier. Examples of catalysts now used in the United States are tetraisopropyltitanate, lithium hydroxide, and lithium ricinoleate. The reaction is run under an inert atmosphere such as CO2 or N2 to minimize discoloration and dimerization of drying oils.
Although the process is called the monoglyceride process, the transesterification reaction actually results in a mixture of unreacted glycerol, monoglycerides, diglycerides, and
unconverted drying oil. The composition depends on the ratio of glycerol to oil and on
catalyst, time, and temperature. In general, the reaction is not taken to equilibrium. At
some relatively arbitrary point, PA is added, beginning the second stage. The viscosity
and properties of the alkyd can be affected by the extent of reaction before the PA addition.
Although many tests have been devised to evaluate the extent of transesterification, none is
very general because the starting ratio of glycerol to oil varies over a considerable range,
depending on the oil length of the alkyd being made. (In calculating the mole ratio of
dibasic acid to polyol, the glycerol already esterified in the oil must also be counted.) A
useful empirical test is to follow the solubility of molten PA in the reaction mixture.
This test has the advantage that it is directly related to a major requirement that must
be met. In the first stage, it is common to transesterify the oil with less expensive
pentaerythritol to obtain mixed partial esters. The second stage, esterification of the
“monoglyceride” with PA, is carried out at a temperature of 220 to 2558C.
15.5.1.2. Fatty Acid Process It is often desirable to base an alkyd on a polyol (e.g.,
PE) other than glycerol. In this case, fatty acids must be used instead of oils, and the
process can be performed in a single step with reduced time in the reactor. Any drying,
semi-drying, or non-drying oil can be saponified to yield fatty acids, but the cost of separating fatty acids from the reaction mixture increases the cost of the alkyd. A more economical alternative is to use tall oil fatty acids (TOFA), which have the advantage that they
are produced as fatty acids. TOFA compositions are roughly similar to those of soybean
fatty acids (Table 14.1). Specially refined tall oils with higher linoleic acid content are
available, as are other grades that have been treated with alkaline catalysts to isomerize
the double bonds partially to conjugated structures. Generally, when fatty acids are
used, the polyol, fatty acids, and dibasic acid are all added at the start of the reaction,
and the esterification of both aliphatic and aromatic acids is carried out simultaneously
in the range 220 to 2558C.
15.5.2. Process Variations
Esterification is a reversible reaction; therefore, an important factor affecting the rate of
esterification is the rate of removal of water from the reactor. Most alkyds are produced
using a reflux solvent such as xylene to promote the removal of water by azeotroping.
Since the reaction is run at a temperature far above the boiling point of xylene, less
than 5% xylene is used. The amount is dependent on the reactor and is set empirically
such that there is enough to reflux vigorously but not so much as to cause flooding of
the condenser. Some of the xylene is distilled off along with the water; water is separated
and xylene is returned to the reactor. Methyl isobutyl ketone (MIBK) has also reported as a
15.5. SYNTHETIC PROCEDURES FOR ALKYD RESINS
317
Figure 15.1. Effect of temperature and time on viscosity. (Adapted from Ref. [28], with permission.)
reflux solvent; it reduces the viscosity more than xylene. The presence of solvent is desirable for other reasons: Vapor serves as an inert atmosphere, reducing the amount of inert
gas needed, and the solvent serves to avoid accumulation of sublimed solid monomers,
mainly PA, in the reflux condenser.
Reaction time is affected by reaction temperature. Figure 15.1 shows the relation among
temperature, time, and viscosity in preparing an alkyd. Figure 15.2 shows the relationship
of the variables with acid number (i.e., milligrams of KOH required to titrate the acid in 1 g
Figure 15.2. Effect of temperature and processing time on the acid value of a typical medium oil linseed
alkyd. (Adapted from Ref. [28], with permission.)
318
ALKYD RESINS
of resin solids). There are economic advantages to short reaction times. Operating costs are
reduced, and the shorter times permit more batches of alkyd to be produced in a year,
increasing capacity without capital investment in more reactors. Therefore, it is desirable
to operate at as high a temperature as possible without risking gelation.
A critical aspect of alkyd synthesis is deciding when the reaction is completed. Disappearance of carboxylic acid is followed by titration, and increase in MW is followed
by viscosity. Determination of acid number and viscosity both take some time. Meanwhile, in the reactor, the reaction is continuing. After it is decided that the extent of reaction is sufficient, the reaction mixture must be “dropped” into a larger tank containing
solvent. When a 40,000-L batch of alkyd is being made, a significant time is required
to get the resin out of the reactor into the reducing tank; meanwhile, the reaction is continuing. The decision to start dropping the batch must be made so that the acid number and
viscosity of the batch will be right after the continuing reaction that occurs between the
time of sampling, determination of acid number and viscosity, and discharging of the
reactor. The time for these determinations becomes the rate controlling step in production.
If they control tests can be done rapidly enough, the reaction can be carried out at 2408C or
even higher without overshooting the target acid number and viscosity. On the other hand,
if the tests are done slowly, it may be necessary to run the reaction at only 2208C, which
may require 2 hours or more of additional reaction time (Figures 15.1 and 15.2). Automatic titration instruments permit rapid determination of acid number, so the usual limit
on time required is viscosity determination. While attempts have been made to use the viscosity of the resin at the reaction temperature to monitor change in the MW, the dependence of viscosity on MW at that high temperature is not sensitive enough to be very
useful. The viscosity must be determined on a solution at some lower standard temperature. Since viscosity depends strongly on solution concentration and temperature, these
variables must be carefully controlled.
In alkyd production, viscosity is commonly determined using Gardner bubble tubes, as
discussed in Section 3.3.4. The cook is continued until the viscosity is high enough so that
by the time the resin batch is dropped into the solvent and the batch cooled, its viscosity
will be what is called for in the specification. This means starting to discharge the reactor
when the test sample is at some lower viscosity. It is not possible to generalize how large
this difference should be; it depends on the specific alkyd composition, the temperature at
which the reaction is being run, the time required to do the determination, the time
required to empty the reactor, and so on. Viscosities can be determined more rapidly
using a cone and plate viscometer (Section 3.3.2) than with bubble tubes; the very
small sample required for a cone and plate viscometer can be cooled and equilibrated at
the measurement temperature more quickly.
Many variables affect the acid number and viscosity of alkyds. One is the ratio of reactants: The closer the ratio of moles of dibasic acid to polyol approaches 1, the higher the
MW of the backbone of the resin, but also the greater the likelihood of gelation. A useful
rule of thumb for a starting point is to use a mole ratio of 0.95. The final ratio is determined
by adjustments such that the combination of acid number and solution viscosity come out
at the desired levels. The greater the ratio of hydroxyl groups to carboxylic acid groups,
the faster the acid groups are reduced to a low level. The degree of completion of the reaction is an important factor controlling the viscosity, as is the acid number. It is usually
desirable to have a low acid number, typically in the range 5 to 10.
The composition of the fatty acids is a major factor affecting the viscosity, and compositions of an oil or grade of TOFA can be expected to vary somewhat from lot to lot.
15.5. SYNTHETIC PROCEDURES FOR ALKYD RESINS
319
Dimerization and oligomerization of the unsaturated fatty acids occur in the same temperature range at which the esterification is carried out (Section 14.3.1). Fatty acids with
conjugated double bonds dimerize more rapidly than those with nonconjugated bonds,
and dimerization rates increase with the level of unsaturation. At the same ratio of phthalic
to polyol to fatty acids, alkyds of the same acid number and solution concentration will
increase in viscosity in the order soybean , linseed , tung.
Some volatilization of polyol, PA, and fatty acids out of the reactor will occur depending
on the design of the reactor, the rate of reflux of the azeotroping solvent, the rate of inert gas
flow, and the reaction temperature, among other variables; the amount and ratio of these
losses affect the viscosity at the standard acid number. The exact ratio of reactants must
be established in the reactor that is used for synthesis. Since gelation can occur if the
ratio of dibasic acid to polyol is too high, it is better not to put all of the PA into the
reactor in the beginning. If the viscosity is too low when the acid number is getting
down near the standard, more PA can easily be added. The amount of PA held back can
be reduced as experience is gained cooking a particular alkyd in a particular reactor.
Side reactions can affect the viscosity –acid number relationship. Glycerol and other
polyols form ethers to some degree during the reaction. Glycerol can also form toxic acrolein by successive dehydrations.
When these reactions occur, the mole ratio of dibasic acid to polyol increases and the
number of hydroxyl groups decreases; therefore, at the same acid number, the MW will be
higher. Excessively high viscosity and even gelation can result. Ether formation is catalyzed by strong protonic acids; therefore, it is desirable to avoid them as catalysts for
the esterification. Monobutyltin oxide has been used as an esterification catalyst; presumably, it does not significantly catalyze ether formation. As noted earlier, PE and TMP seem
less vulnerable than glycerol to undesirable side reactions such as ether formation, and
glycerol is the only polyol that can decompose to form acrolein. A hydroxyl group on
one end of a growing polyester chain can react with a carboxylic acid group on another
end of the same molecule, leading to ring formation. Transesterification of chain linkages
can have the same result. Since cyclization reactions reduce chain length, their net effect is
to reduce viscosity.
Many alkyd resins have broad, uneven MW distributions (Figure 2.3). Kumanotani has
shown that even modest changes in reaction conditions can cause large differences in MW
distribution, which can have significant effects on final film properties [29]. In many
alkyds, very small gel particles, microgels, are formed. Kumanotani has shown that
these microgels play an important role in giving greater strength properties to final
films. Process changes that may make the alkyd more uniform may be undesirable. For
example, allowing glycerolysis to approach equilibrium before addition of PA and
using transesterification catalysts in the final stages of esterification both favor narrower
MW distributions and lower viscosities, but films made from the more uniform alkyds
may exhibit inferior mechanical properties.
320
ALKYD RESINS
15.6. MODIFIED ALKYDS
Oxidizing alkyds can be modified by reaction with vinyl monomers. The most widely used
monomers are styrene, vinyltoluene, and methyl methacrylate, but essentially any vinyl
monomer can be reacted in the presence of an alkyd to give a modified alkyd. Methyl
methacrylate imparts better heat and weather resistance than styrene but at higher cost.
In making styrenated alkyds, an oxidizing alkyd is prepared in the usual way; then
styrene and a free radical initiator such as benzoyl peroxide are added. The resulting
free radical chain process leads to a variety of reactions, including formation of low
MW homopolymer of styrene, grafting of polystyrene onto the alkyd, and dimerization
of alkyd molecules. The reaction is generally carried out at about 1308C, which favors
decomposition of benzoyl peroxide to form phenyl free radicals; phenyl radicals have a
greater tendency to abstract hydrogen, which favors grafting. After the reaction is complete, the resin is diluted with solvent. The ratio of alkyd to styrene can be varied over
a wide range; commonly 50% alkyd and 50% styrene is used. The ratio of aromatic
rings to aliphatic chains is greatly increased, and as a result, the Tg of styrenated alkyds
is higher and the tack-free time is shorter. Styrenated alkyds give a “dry” film in 1 hour
or less versus 4 to 6 hours for the counterpart non-styrenated alkyd. However, the
average functionality for oxidative cross-linking is reduced, not just by dilution with
styrene, but also because the free radical reactions involved in the styrenation consume
some activated methylene groups. As a result, the time required to develop solvent resistance is longer than for the counterpart alkyd. The fast drying and low cost make styrenated
alkyds very attractive for some applications, but in other cases, the longer time required for
cross-linking is more critical, in which case styrenated alkyds are not appropriate.
Styrenated alkyd vehicles are often used for air dry primers. One must be careful to
apply the top coat almost immediately, or not until after the film has had ample time to
cross-link. During the intermediate time interval, application of the top coat is likely to
cause nonuniform swelling of the primer, leading to what is called lifting of the primer.
The result of lifting is the development of wrinkled areas in the surface of the dried
film. End users who are accustomed to using alkyd primers, which do not give a hard
film until a significant degree of cross-linking has occurred, are particularly likely to
encounter problems of lifting if they switch to styrenated alkyd primers.
Acrylate-functional alkyds are reported to dry more rapidly than the corresponding
unmodified resin. For example, the carboxylic acid groups on an alkyd prepared from
PE, TMP, soy oil, and IPA with an acid number of 8.6 is reacted with glycidyl methacrylate (GMA). Using zirconium and cobalt driers, films through dried in 9 hours, in contrast
to .11 to ,20 hours for the unmodified resin [30].
15.7. URALKYDS AND OTHER AUTOXIDIZABLE URETHANES
Diisocyanates are used to make a variety of resins that share similar characteristics with
alkyds, as discussed in this section. For a description of the isocyanates, see Section 12.3.
15.7.1. Uralkyds
About 39,000 metric tons of uralkyds were used in the United States in 1997 [31].
Uralkyds are also called oil-modified urethanes. The largest use for uralkyds is as the
15.7. URALKYDS AND OTHER AUTOXIDIZABLE URETHANES
321
vehicle for urethane varnishes for use by contractors and the do-it-yourself market. The
abrasion and hydrolytic resistance of such coatings are greatly superior to that obtained
with conventional varnishes or alkyd resins. Many “varnishes” sold to consumers today
are based on uralkyds; they are not really varnishes in the original sense of the word.
They are used as transparent coatings for furniture, woodwork, gymnasium floors, and
hardwood floors; applications in which good abrasion, mar, scuff, chemical, and saponification resistance are important. They are also widely used in maintenance paints.
To prepare uralkyds, one transesterifies a drying oil with a polyol such as glycerol or PE
to make a “monoglyceride,” which is then reacted with somewhat less diisocyanate than
the equivalent amount of NCO based on the free OH content. To assure that no NCO
groups remain unreacted, methyl alcohol (or other low MW alcohol) is commonly
added at the end of the process. For lower cost, one can use a combination of phthalic
anhydride and diisocyanate. Color retention is superior with aliphatic diisocyanates, but
Tg is higher and cost is lower with aromatic diisocyanates. An extensive review paper
is available [32].
A comparison of the properties of uralkyds, alkyds, and 2K urethane coating has been
published [33]. Four uralkyds prepared from linseed oil and safflower oil, each with TDI or
MDI, were compared with linseed and safflower alkyds. All of the alkyds and uralkyds had
an oil length of 60%. They were compared with two 2K polyester – MDI resin systems.
Primers and pigmented top coats were formulated with each resin. The cohesive strengths
of the uralkyd paint films were greater than those of the alkyd films and approached those
of the 2K coatings. Of the uralkyds, the highest cohesive strength was with films made
using the linseed oil– TDI resin. Scratch resistance and tensile strength tests showed the
same ranking of the films made from the eight resins. Salt spray resistance and humidity
resistance of the uralkyd coatings were superior to those of the alkyd coatings but inferior
to the 2K urethane coatings. Particularly notable, the chemical and solvent resistance of
the uralkyd films were markedly better than the alkyd films, although somewhat inferior
to the 2K films. Studies of the permeability rate of water into uralkyd and alkyd films indicated that alkyd films are more permeable than uralkyd films [34].
By far the most widely used diisocyanate is TDI. A uralkyd with improved color retention compared to that of a uralkyd made with TDI can be made by reacting the polyol made
by transesterification of linseed oil with PE, then reacting with H12MDI [35].
Another advantage of uralkyds over alkyds is ease of manufacture. The time for the
reaction of the isocyanate groups is substantially shorter and the reaction temperature is
on the order of 708C, compared to 220 to 2458C for the esterification of alkyds. This
results not only in lower production costs and higher equipment utilization but also in
less dimerization of the unsaturated chains, hence reduced viscosity at the same molar
ratios.
Traditionally VOC of uralkyd varnishes has been over 500 g L21. As of 2005, EPA
VOC regulations for varnishes require less than 450 g L21, California south coast as
well as northeastern states regulations require less than 350 g L21, and the California regulations will be further reduced to 275 g L21 in 2006. It is reported that uralkyds with a
VOC of 345 g L21 can be made by reducing the MW. Meeting the 275 g L21 regulation
will require greater changes. With careful balancing of driers and accelerators, the dry
time and performance of higher VOC uralkyds can be approached. One approach is to
use the exempt solvent 4-chlorobenzotrifluoride (PCBTF) to replace part of the mineral
spirits solvent (Section 18.9.1). The higher-solids varnishes have the advantage that two
coats are sufficient instead of the usual three; however, great care must be taken to
322
ALKYD RESINS
avoid too great a film thickness. Water-reducible urethane varnishes can be used (see the
next paragraph). Autoxidizable acrylic – urethane hybrid dispersions permit preparation of
high quality products at a higher cost (Section 15.7.2). For the contractor market, 2K
waterborne urethanes have many advantages but are not suitable for the do-it yourself
market (Section 12.7.3) [36].
As with other types of coatings, VOC regulations stimulated development of waterreducible uralkyds. An example is a maleated uralkyd prepared from a partial ester of
linseed oil fatty acid with PE, which is reacted with MDI and then with maleic anhydride
[37]. The resin is diluted in 2-butoxyethanol, neutralized with TEA, and diluted with
water. The resin is much more resistant to hydrolysis than is a corresponding alkyd.
15.7.2. Autoxidizable Polyurethane Dispersions
Water-reducible polyurethane dispersions (PUDs) (Section 12.7) can be made autoxidizable by introducing appropriate fatty acids. For example, a PUD made by reacting a diisocyanate with a diol, a fatty acid monoglyceride, and dimethylolpropionic acid (DMPA)
is neutralized with an amine and then dispersed in water. Water acts as a chain extender for
terminal isocyanate groups [38]. Hydrolytic stability of the aqueous dispersions is excellent, and the films also have the expected abrasion resistance. Drying speed, color, and cost
are dependent on the drying oil and diisocyanate used. For example, a PUD is made from a
dehydrated castor oil, pentaerythritol, phthalic alkyd reacted with DMPA, dimethylisopropylamine, and isophorone diisocyanate (IPDI) [39]. A PUD for wood flooring coatings is
prepared from an unsaturated polyol prepared by transesterification of castor oil and soy
oil, DMPA, H12MDI, neutralized with TEA, dispersed in water, and chain-extended with
hydrazine hydrate and ethylenediamine [40]. The coatings are said to resist solvents,
abrasion, and heel marking.
Acrylic – urethane hybrid dispersions have lower cost. For example, a drying oil fatty
acid amide of diethanolamine is reacted with DMPA, cyclohexanedimethanol, a monomethyl ether of polyethylene glycol, and TDI in N-methylpyrrolidone (NMP), methyl
methacrylate is added, the product is diluted with dipropylene glycol monomethyl
ether, neutralized with dimethylaminoethanol (DMAE), and then diluted with water
[41]. After dilution with water, the methacrylate monomer is polymerized; the resulting
PUD gives films that cross-link by autoxidation.
15.8. EPOXY ESTERS
BPA epoxy resins can be converted to what are commonly called epoxy esters by reacting
with fatty acids. Drying or semi-drying oil fatty acids are used so that the products crosslink by autoxidation. The epoxy groups undergo a ring-opening reaction with carboxylic
acids to generate an ester and a hydroxyl group (Section 13.3.2). These hydroxyl groups,
as well as the hydroxyl groups originally present on the epoxy resin, can esterify with fatty
acids. They are generally made by starting with a low MW epoxy resin (i.e., the standard
liquid resin, n ¼ 0.13) and extending with BPA by the advancement process (Section
13.1.1) to the desired MW. Off-specification epoxy resin is often used to reduce cost.
The fatty acids are added to the molten, hot resin, and the esterification reaction is continued until the acid number is low, usually less than 7 mg of KOH per gram of resin. In the
esterification reaction with fatty acids, the average number of sites for reaction is the n
15.8. EPOXY ESTERS
323
value, corresponding to the number of hydroxyl groups on the resin, plus twice the number
of epoxy groups. The esterification is carried out at high temperatures (220 to 2408C). The
rate of esterification slows as the concentration of hydroxyl groups diminishes, and side
reactions occur, especially dimerization of the drying oil fatty acids (or their esters). It
is not practical to esterify more than about 90% of the potential hydroxyl groups, including
those from ring opening the epoxy groups.
Tall oil fatty acids are commonly used because of their low cost. Linseed fatty acids
give faster cross-linking coatings due to higher average functionality. However, their viscosity is higher because of the greater extent of dimerization during esterification, and
their cost is higher. For still faster cross-linking, part of the linseed fatty acids can be
replaced with tung fatty acids, but the viscosity and cost are still higher. The color of
epoxy esters from linseed and linseed– tung fatty acids is darker than the tall oil esters.
Dehydrated castor oil fatty acids give faster curing epoxy esters for baked coatings. The
rate of formation of a dry film from epoxy esters depends on two factors: the average
number of diallylic groups f̄n and the ratio of aromatic rings to long aliphatic chains.
The f̄n can be maximized by using higher MW BPA epoxy resin and by using enough
fatty acid to react with a large fraction of the epoxy and hydroxyl groups. The ratio of aromatic rings to fatty acids can be maximized by using high MW epoxy resin and esterifying
a smaller fraction of epoxy and hydroxyl groups.
Epoxy esters are used in coatings in which adhesion to metal is important. Although the
reasons are not completely understood, it is common for epoxy coatings, including epoxy
esters, to have good adhesion to metals and to retain adhesion after exposure of the coated
metal to high humidity, a critical factor in corrosion protection. A distinct advantage of
epoxy esters over alkyd resins is their greater resistance to hydrolysis and saponification.
The backbone of alkyds is held together with esters from PA and the polyol, whereas in
epoxy esters, the backbone is held together with C22C and ether bonds. Of course, the
fatty acids are bonded to the backbone with ester groups in both cases, but the fraction
of polymer bonds in a dry film subject to hydrolysis is substantially lower in the case of
epoxy esters. On the other hand, exterior durability of epoxy ester coatings is poor, as
is the case with all films made with BPA epoxy resins. As a result of these advantages
and disadvantages, the major uses for epoxy resins are in primers for metal and in can coatings, such as for crowns (bottle caps), in which the important requirements are adhesion
and hydrolytic stability. In baking primers, it is sometimes desirable to supplement the
cross-linking through oxidation by including a small amount of MF resin in the formulation to cross-link with part of the free hydroxyl groups on the epoxy ester.
Epoxy ester resins with good exterior durability (better than alkyds) can be prepared by
reacting epoxy-functional acrylic copolymers (made with glycidyl methacrylate) with
fatty acids. The product is an acrylic resin with multiple fatty acid ester side chains. By
appropriate selection of acrylate ester comonomers and MW, the Tg of the resin can be
designed so that a tack-free film is obtained by solvent evaporation; then the coating crosslinks by autoxidation. For an application such as repainting an automobile at ambient
temperatures, the cross-linking can proceed relatively slowly and need not be catalyzed
by metal salt driers. The rate of cross-linking is slower without driers, but exterior durability is better.
Epoxy esters can also be made water-reducible: for example by reacting maleic anhydride with epoxy esters prepared from dehydrated castor oil fatty acids. Subsequent
addition of a tertiary amine such as 2-(dimethylamino)ethanol in water results in ring
opening of the anhydride to give amine salts. Like other water-reducible resins
324
ALKYD RESINS
(Section 8.3), these resins are not soluble in water, but form a dispersion of resin aggregates swollen with water and solvent in an aqueous continuous phase. The hydrolytic stability of these epoxy esters is better than corresponding alkyds and is sufficient for use in
anionic electrodeposition primers (Section 27.1), now largely replaced by cationic
primers. Water-reducible epoxy esters are still used in spray applied baking primers and
primer-surfacers. They are also used in dip coating primers, in which nonflammability
is an advantage.
GENERAL REFERENCE
Wicks, Z. W., Jr., Alkyd Resins, in Encyclopedia of Polymer Science and Technology, 3rd ed.,
Wiley, New York, 2002, pp. 318 – 340.
REFERENCES
1. U.S. Department of Commerce, Bureau of Census, Current Industrial Reports: Paint and
Coating Manufacturing, http://www.census.gov.
2. Blackinton, R. J., J. Paint Technol., 1967, 39(513), 606.
3. Porreau, D. R.; Smyth, S. E., JCT Coat. Tech, 2004, 1(2), 40.
4. Kangas, S. L.; Jones, F. N., J. Coat. Technol., 1987, 59(744), 89.
5. Bacaloglu, R.; et al., Angew. Makromol. Chem., 1988, 164, 1.
6. Brown, R.; et al., Off. Dig., 1961, 33, 539.
7. Kawamura, C.; et al., Prog. Org. Coat., 2002, 45, 185.
8. Kass, P.; Wicks, Z. W., Jr., U.S. patent 2,577,770 (1951).
9. Zabel, K. H., et al., Prog. Org. Coat., 1999, 35, 255.
10. Resimene AM-300 and AM-325, tech. bull., Monsanto Chemical Co. (now Solutia, Inc.),
St. Louis, MO, January 1986.
11. Strazik, W. F.; et al., U.S. patent 4,293,461 (1981).
12. Bracken, J., U.S. patent 6,075,088 (2000).
13. Kuang, Z.; et al., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA, 1999,
pp. 126 –141.
14. Osterberg, G.; et al., Prog. Org. Coat., 1994, 24, 281.
15. Osterberg, G.; Bergenstahl, B., J. Coat. Technol., 1996, 68(858), 39.
16. Weissenborn, P. K.; Motiejauskaite, A., Prog. Org. Coat., 2000, 40, 253.
17. Cerkholantsev, V. C., Eur. Coat. J., 2000, (1– 2), 120.
18. Nabuurs, T.; et al., Prog. Org. Coat., 1996, 27, 163.
19. Gooch, J. W.; et al., Proc. Waterborne High Solids Powder Coat. Symp., New Orleans, LA,
1997, p. 366.
20. van Hamersfeld, E. M. S.; et al., Prog. Org. Coat., 1999, 35, 235.
21. Kim, K.-J.; et al., Proc. Waterborne Higher Solids Coat. Symp., New Orleans, LA, 2004, Paper 36.
22. Bouboulis, C. J., Proc. Water-Borne Higher-Solids Coat. Symp., New Orleans, LA, 1982, p. 18.
23. Zuchert, B; Biemann, H., Farg Lack Scand., 1993, (2), 9. Weger, W., Fitture Vernici, 1990,
B66(9), 25.
24. Hurley, R.; Buona, F., J. Coat. Technol., 1982, 54(694), 55.
25. Wilke, G.; et al., U.S. patent 6,187,384 (2001).
REFERENCES
325
26. Bakker, P. J.; et al., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA, 2001,
pp. 439 – 453.
27. Kaska, J.; Lesek, F., Prog. Org. Coat., 1991, 19, 283.
28. The Chemistry and Processing of Alkyd Resins, Monsanto Chemical Co., St. Louis, MO, 1962.
29. Kumanotani, J.; et al., Adv. Org. Coat. Sci. Technol. Ser., 1984, 6, 35.
30. Kuo, T.; et al., U.S. patent 6,911,493 (2005).
31. Skeist Report VI, Skeist Inc., Whippany, NJ, 1998, pp. 805 – 826.
32. Wicks, D. A.; Wicks, Z. W., Jr., Prog. Org. Coat., 2005, 54, 141.
33. Narayan, R.; et al., Paint India, 1998, September, 116.
34. Xu, Y.; et al., Prog. Org. Coat., 2002, 45, 331.
35. Mountfield, B. A.; et al., U.K. patent GB 1,059,936 (1967).
36. Caldwell, R. A., JCT Coat. Tech, 2004, 2(3), 30.
37. Reichold Albert Chemie AG, British patent GB 1,199,810 (1970).
38. Liu, W.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1999,
p. 202.
39. Prantl, B.; et al., U.S. patent 5,319,052 (1994).
40. Irle, C.; et al., U.S. patent 6,559,225 (2003).
41. Coogan, R. G.; Damery, S., U.S. patent 6,548,588 (2003).
16
Silicon Derivatives
Three classes of silicon derivatives are used in coatings: silicones, reactive silanes, and
orthosilicates. Orthosilicates are also covered for their use in organic/inorganic coatings.
16.1. SILICONES
Silicones, more properly called polysiloxanes, are polymers with backbones consisting of
22Si(R)222O22 repeating units. Polysiloxanes are prepared from chlorosilanes. The principal commercial monomers are the following methyl and phenyl substituted silanes.
Me3SiCl
Ph2SiCl2
PhSi(Me)Cl2
Me2SiCl2
PhSiCl3
MeSiCl3
trimethylchlorosilane
diphenyldichlorosilane
phenylmethyldichlorosilane
dimethyldichlorosilane
phenyltrichlorosilane
methyltrichlorosilane
Chlorosilanes react with water to form silanols, which in turn can condense to form
siloxanes. For example, dimethyldichlorosilane reacts with water to form a polysiloxane,
often represented as a linear polysiloxane.
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
326
16.1. SILICONES
327
Structures of the polymers can be more complex than represented by this linear structure. Siloxanes are more apt than hydrocarbons to form large ring cyclic structures such as
(Me2SiO)4. Such cyclic compounds are intermediates in the polymerization and are
present to varying degrees in the finished product.
Oligomers are made using monochlorosilanes as chain-terminating groups; molecular
weight (MW) is reduced in proportion to the monochlorosilane/dichlorosilane ratio. The
oligomeric products are called silicone fluids or silicone oils. As discussed in Chapter 23,
small amounts of silicone fluids are used as additives in coatings. Poly(dimethylsiloxane)
fluids reduce surface tension because the ease of rotation around the Si22O22Si bonds
leads to rapid orientation of methyl groups at the surface. The methyl groups give a
surface with very low surface tension. Low MW silicone fluids can be used as VOC
exempt solvents.
Poly(dimethylsiloxanes) have limited compatibility with many coatings resins. Chemically modified silicone fluids with broader ranges of compatibility have been described
[1]. Examples are polysiloxane–polyether block copolymers, polyether copolymers that
also have some longer alkyl chains on the siloxane groups, and poly(dimethylsiloxanes)
modified with esters and aryl-substituted alkyl groups. Siloxanes with a polyether terminal
group or side chain have been reported to be excellent surface tension additives to prevent
film defects in waterborne coatings. They are prepared by reacting allyl-terminated polyethers with poly(dimethylsiloxanes) having Si22H bonds [2].
16.1.1. Silicone Rubbers and Resins
Silicone rubbers are cross-linked polysiloxanes, and silicone resins are usually branched
polysiloxanes. In both cases, trichlorosilanes are used to provide chain branching and ultimately, cross-linking. They can be copolymerized with dichlorosilanes to make silicone
rubbers. Although silicone rubbers are expensive, they have unique and useful properties;
they exhibit excellent resistance to oxidation accompanied by retention of flexibility at low
temperatures. They have a low Tg as a result of the ease of rotation around the Si22O22Si
bonds.
Chain branching is not the only possible outcome when trichlorosilanes are copolymerized. They can react with water to form three-dimensional cluster compounds (silsesquioxanes), exemplified by the cubic octamer (C6H5)8Si8O12.
Polymerization of a mixture of mono-, di-, and trichlorosilanes, carried out by hydrolyzing a solution of the monomers in a water-immiscible solvent, results in a silicone resin
with some unreacted hydroxyl groups. The silicone resin is separated from the acidic water
phase, and residual acid is removed. A stable solution of a silicone resin is obtained if the
last water is removed from the reaction mixture at relatively low temperatures in the
absence of catalyst. The M̄n of such resins is generally in the range 700 to 5000. Different
copolymerization processes yield different proportions of linear, branched, cyclic, and
three-dimensional cluster structures; the composition affects properties.
328
SILICON DERIVATIVES
Silicone resins can be baked to form cross-linked films. The cross-linking process is
reversible. A typical cure schedule is 1 hour at 2258C, but the time and temperature can
be reduced with catalysts such as zinc octanoate. Since the cross-linking reaction is reversible, silicone films are sensitive to water, especially under basic conditions. Ammonia and
amines are especially destructive to such films. Evidence has been presented that aminolysis occurs selectively at Si groups with three oxygens (i.e., at cross-linking sites), probably resulting from the greater electrophilicity of such Si groups with three electronegative
oxygen atoms [3].
The MW and viscosity of silicone resins depend on the proportion of monochloro-,
dichloro-, and trichlorosilane monomers and on the process. The ultimate cross-link
density depends on the fraction of trichlorosilane in the recipe and on the fraction of this
monomer that forms branched rather than cyclic or cluster structures during the process.
Most silicone resins are copolymers of methyl and phenyl substituted monomers; other
alkyl monomers are sometimes included. Properties depend on the phenyl/methyl ratio, as
summarized in Table 16.1. The rate of the cross-linking reaction is faster with high-methyl
substituted silicone resins. Consequently, a larger amount of material is lost during heat
curing of high-phenyl resins, since low MW components evaporate in the oven to a
larger degree before polymerizing. The faster cure rate of the high-methyl substituted silicones and the longer package stability of the high-phenyl substituted silicones follow from
the same factor. The greater UV resistance of high-methyl relative to high-phenyl substituted silicone resins gives greater gloss retention during exterior exposure. The exterior
durability of well-formulated silicone coatings is better than that of acrylic – MF,
polyester –MF, or urethane coatings, and approaches that of highly fluorinated polymers.
On the other hand, high-phenyl silicones are superior to high-methyl silicones for applications requiring high temperature resistance, and far superior to other organic coatings
except certain fluoropolymers. The thermal stability of high-methyl silicones, although
inferior to that of high-phenyl silicones, is much greater than that of silicones with
longer alkyl groups. The half-lives at 2508C, based on weight loss studies, are reported
to exceed 100,000 hours for phenylsilicone films compared with over 10,000 hours for
methyl silicone films and only 2 hours for propyl silicone films [4]. The useful life of a
methyl silicone film at 3508C is given as 1000 hours; for comparison a “polyester” film
(presumably cross-linked with a MF resin) has a useful life of 1000 hours at 2238C.
When silicones are thermally decomposed, the ultimate product is silicon dioxide,
which, though brittle, can serve as a temperature-resistant coating binder. For example,
chimney paints are made from silicone resins pigmented with aluminum flake for use at
over 5008C for years. At the high service temperature, the organic substituents burn off,
leaving behind a film of the aluminum pigment in a matrix of silicone dioxide with
perhaps some aluminum silicate—essentially glass. Although it is brittle, this film
TABLE 16.1. Properties of High-Methyl Versus
High-Phenyl Silicone Resins
High-Methyl
Lower weight loss in curing
Faster cure rate
Higher UV stability
Lower temperature flexibility
Source: Adapted from Ref. [4].
High-Phenyl
Greater weight loss in curing
Longer package stability
Higher heat stability
Greater oxidation resistance
16.1. SILICONES
329
continues to provide protection if it is not damaged mechanically. (See Ref. [4] for discussion of coatings with varying heat resistance.)
As expected from the greater rigidity of aromatic rings, coatings from methyl silicone
resins have low temperature flexibility superior to those from phenylsilicone and also
superior to most other organic coatings. They also combine the properties of being repellant to liquid water but permeable to water vapor.
16.1.1.1. Modified Silicone Resins High-solids silicone resins have been made
available that cure either directly using zinc octanoate catalysts or by cross-linking with
trialkoxysilanes using a titanate catalyst [5]. Waterborne silicone resins have also been
developed. One approach is to emulsify a silicone resin in water using an emulsifying
agent [6]. Emulsions of methyl silicone resins, an amino-functional polysiloxane emulsion, and a styrene – acrylic latex are used to formulate exterior house paints. Such coatings
have been used on a large scale in Europe for many years but are just being introduced into
the United States. They have outstanding exterior durability [7].
Silicone resins can be modified with a variety of functional groups. Epoxy-functional
resins are prepared by reacting methylphenyldimethoxysilane, phenyltrimethoxysilane,
and dimethyldimethoxysilane with water with KOH as catalyst followed by reacting the
resulting resin with 3-glycidyloxypropyltriethoxysilane. Thermosetting coatings can be
formulated by cross-linking the epoxy silicone with a diaminodialkoxysilane [8]. Epoxyfunctional silicone resins cross-linked with carboxylic acid –functional acrylic resins are
used to formulate marine finishes with excellent durability [9].
Carbamate-functional siloxanes are prepared from hydroxy-terminated siloxanes.
Dimethylsiloxane oligomers with 4 to 5 units and two silicon hydride groups are
reacted with the allyl group of trimethylolpropane monoallyl ether using chloroplatinic
acid as the catalyst to give the hydroxy-functional siloxane. The hydroxy-functional
oligomer is then reacted with 1-methoxy-2-propylcarbamate using butyl stannoic acid
and triphenylphosphite as catalysts to give the carbamate-functional oligomer. The resin
is combined with a carbamate-functional acrylic resin and an MF resin to formulate automobile clear coats with excellent scratch resistance [10].
16.1.2. Silicone-Modified Resins
The high cost of pure silicone coatings and the long, high temperature cures can be
reduced by combining silicones with other coating resins. The earliest approach was
simply to add a silicone resin to an alkyd resin in the reactor at the end of the alkyd
cook. Although some covalent bonds between silicone resin and alkyd might form, probably most of the silicone resin simply dissolves in the alkyd. The exterior durability of
silicone-modified alkyd coatings made this way is significantly better than unmodified
alkyds. The improvement in durability is roughly proportional to the amount of added silicone resin; 30% silicone resin is a common degree of modification. Silicone resins
designed for this purpose may contain higher alkyl as well as methyl and phenyl groups
to improve compatibility. Alkyd coatings modified with predominantly phenylsilicone
resins are reported to have greater thermoplasticity, faster air drying, and higher solubility
than predominantly methyl silicone – modified alkyds. These results can be understood
when we consider that the higher rigidity of the aromatic rings leads to a “solid” film at
an earlier stage of cross-linking. Less cross-linking in the phenylsilicone-modified coatings makes them more thermoplastic and soluble.
330
SILICON DERIVATIVES
Further improvements in exterior durability are obtained by coreacting a silicone intermediate during synthesis of the alkyd. Such intermediates react readily with free hydroxyl
groups. Silicone-modified alkyds made this way are said to impart better durability than
alkyds in which the silicone resin is cold blended. They are used mainly in outdoor air
dry coatings where application is expensive (e.g., in a top coat for steel petroleum
storage tanks) and a longer coating service life more than offsets the extra cost.
Silicone-modified polyester and acrylic resins are made by reacting silicone intermediates with the hydroxy-functional resins; in some cases, the silicone intermediates have
silanol (Si22OH) groups. These can undergo co-condensation with hydroxyl groups on
the resin to be modified as well as self-condensation with other Si22OH groups. The
ratio of the two reactions is controlled by catalyst choice. The best catalysts for promoting
the desirable co-condensation between the resin and the silicone intermediate are titanates
such as tetraisopropyl- or tetraisobutyltitanate. In model compound studies, tetraisopropyltitanate was shown to favor co-condensation over self-condensation by 3.4 : 1 as compared to a 0.23 : 1 ratio with no catalyst. Co-condensation between a hydroxy-functional
silicone intermediate and acrylic polyol is shown below.
Self-condensation between silicone intermediates can also occur resulting in some
Si22O22Si cross-links. Excessive self- or co-condensation of a polyfunctional silicone
intermediate and a polyfunctional resin can result in gelation.
It is easier to control the reaction when the reactive functional groups on the silicone
intermediates are silylmethoxy (Si22OMe) rather than silanol (Si22OH) groups. A methoxylated silicone intermediate has been described which has a weight average MW of 470
and an equivalent weight of 155 [4]. An idealized structure of the intermediate is
A solution of the polyester or acrylic is heated with the desired amount of silicone intermediate and the titanate catalyst at 1408C until a predetermined viscosity is reached. The
reaction is relatively slow at 1408C; higher temperatures reduce reaction time and hence
cost but increase the risk of overshooting the desired viscosity as well as the risk of gelation. Typical silicone-modified polyesters and acrylics have 30 to 50 wt% silicone. Modification with less than about 25% silicone provides little improvement in exterior
durability. Exterior durability is enhanced with increasing silicone content above 30%,
but cost also increases.
The same reactions occur during cross-linking after the coating is applied. Usually, zinc
octanoate is used as the catalyst for cross-linking the coating since titanate esters are
hydrolyzed by the water brought into the coating by pigments. The principal applications
are in baking coatings, especially in coil coatings for metal siding (Section 30.4), where a
typical cure schedule is 90 seconds at peak metal temperatures of 3008C. Such coatings
tend to soften when exposed to high humidity for prolonged periods. The softening is
16.2. REACTIVE SILANES
331
called reversion. It is speculated that it is caused by reversible hydrolysis of the crosslinks, which, as discussed above, have three oxygen atoms bound to a single silicon
atom. The film hardens again if the ambient humidity decreases, but the film is vulnerable
to physical damage if scraped while it is soft. To minimize this problem, a small amount of
MF resin can be added as a secondary cross-linker.
Waterborne silicone-modified resins can be prepared from water-reducible acrylic and
polyester resins. Also, acrylic latexes prepared with hydroxyethyl (meth)acrylate as a
comonomer can be modified with silicone intermediates [6]. Silicone-modified epoxy
resins have also been made [11]. See Ref. [4] for examples of formulations and preparation
of several examples of silicone-modified resins.
16.2. REACTIVE SILANES
Reactive silane is the term used to identify compounds having a trialkoxysilyl group and
an alkyl group bearing a reactive substituent. They have been used in coatings in several
different ways (Refs. [11] and [12] are review papers). Commercially available reactive
silanes include 3-aminopropyltriethoxysilane, N-(2-aminoethyl)-3-aminopropyltrimethoxysilane, 3-glycidyloxypropyltrimethoxysilane, and 3-methacryloxypropyltrimethoxysilane, among others.
Trialkoxysilyl groups can react directly, or indirectly in the presence of water, with
hydroxyl groups; the other functional group can participate in the cross-linking reaction
with the resins in a coating. For example, the trialkoxysilyl group on 3-aminopropyltriethoxysilane can react with hydroxyl groups on the surface of glass, while the amine group can react
with epoxy groups in an epoxy resin. The interactions with surfaces are complex [11]. The
relative rates of hydrolysis of the alkoxysilyl groups to silanol groups, self-condensation reactions, and reactions with substrate hydroxyl groups vary with pH, water concentration,
alcohol concentration, reactivity of the substrate hydroxyl groups, and rates of reverse reactions. In the case of amino-functional silanes, the situation can be further complicated by the
competing possibility of adsorption of amine groups on the substrate. An idealized picture of
the interaction of 3-aminopropyltrimethoxysilane with the surface of glass and an epoxy resin
is shown in Scheme 6.2 in Section 6.4.3. Before the advent of reactive silanes it was difficult
to achieve adhesion to glass that would resist displacement with water. Use of reactive silanes
has made possible coatings and sealants with excellent adhesion to glass in the presence of
water. Although water can hydrolyze some of the linkages, the reaction is reversible. With
multiple attachments, adhesion is not lost. (See Section 6.4.3 for further discussion.)
Reactive silanes also form stable condensation products with other oxides, such as
those of aluminum, zirconium, tin, titanium, and nickel; less stable bonds are formed
with oxides of carbon, iron, and boron [12]. Reactive silane additives are used not only
as adhesion promoters but also to treat the surface of oxide pigments to improve dispersion
properties.
Bis(trisilylalkoxy)alkanes are being investigated to treat the surface of steel to increase
adhesion [13]. Clean steel is rinsed with water and then the wet steel is dipped in an
aqueous solution of bis(trimethoxysilyl)ethane (BTSE) followed by dipping in an
aqueous solution of a reactive silane. The BTSE reacts with water and hydroxyl groups
on the steel and silanols from other molecules of BTSE to give a water-resistant anchor
to the steel. The reactive silane reacts with other silanol groups from the BTSE, and the
reactive group can react with a coating binder. There are many variables in the treatment,
332
SILICON DERIVATIVES
excellent results have been obtained in laboratory work, and commercial use is anticipated. A mixture of bis(trimethoxysilylpropyl)amine and vinyltriacetoxysilane has been
reported to provide corrosion protection equal to chromate primers on aluminum, galvanized steel, stainless steel, steel, and cobalt – chromium alloy [14]. (See Sections 6.4.3
and 7.3.2 for further discussion.)
Resins with multiple trialkoxysilyl groups can be used as binders for moisture-curable
coatings. For example, an isocyanate-terminated resin can be reacted with 3-aminopropyltrialkoxysilane to give a resin with terminal trialkoxysilyl groups. Coatings made using such
resins cross-link to a polymer network after application and exposure to humid air. Multifunctional isocyanates may be used to prepare resins with increased cross-link density. An
advantage of the triethoxysilyl resin over the isocyanate resin is that part of the solvent can
be ethyl alcohol. The ethyl alcohol permits a reasonable pot life in the presence of water
such as comes into a coating from pigment surfaces. If free isocyanate groups are
present, the coating is not stable in the presence of either ethyl alcohol or water. The trialkoxysilyl approach also avoids formation of CO2, which can lead to film imperfections such
as pinholing in moisture-curable urethanes. After the alcohol evaporates, the reaction with
water proceeds rapidly at ambient temperature. For example, isocyanate-terminated
urethanes prepared by reacting polyols and MDI are end-capped with 3-aminotrimethoxysilane to prepare a resin used in moisture-curable coatings [15].
Trialkoxysilyl-functional acrylic resins can be made by copolymerizing a trialkoxysilylalkyl methacrylate with other acrylic monomers [16 – 18]. Coatings based on such
resins cure on exposure to atmospheric moisture; the reaction is catalyzed with organotin
compounds or organic acids. The resulting coatings are reported to have excellent exterior
durability, resistance to environmental etching and marring, and adhesion to aluminum.
They are used in automobile OEM and refinish clear coats. In baking automotive clear
coats for plastics, trimethoxypropylsilyl-functional acrylic resins; a methoxy group hydrolyzes at room temperature and then reacts with another silanol group during baking to
cross-link the coating. The scratch and mar resistance before and after weathering are
superior to MF cross-linked coatings [19].
Automotive clear coats are being made by combining trialkoxysilylalkyl-functional
and hydroxy-functional acrylic resins with MF resins or blocked isocyanates [18]. No
detailed studies of the cross-linking reactions have been published, but one can speculate
that they include hydrolysis and condensation of the trialkoxysilyl groups to form siloxane
cross-links, reaction of hydroxyl groups with MF resins or blocked isocyanates, and transetherification of the hydroxyl groups with the trialkoxysilyl groups. The resulting coatings
are reported to have excellent exterior durability and very good resistance to environmental etching and marring. Another advantage is that formaldehyde emissions are
reduced because lower levels of MF resin are used.
Trialkoxysilylated acrylic and vinyl acetate latexes can be prepared using 3methacryloxypropyltriisobutoxysilane as a comonomer in emulsion polymerization
[20]. The isobutoxysilyl derivatives are used since they are more resistant to hydrolysis
at polymerization temperatures. The latexes are storage stable; they do not cross-link
until water evaporates from a film. Possibly the alkoxysilyl groups hydrolyze during
storage, but the silanols formed do not condense until after a film is formed because of
the large excess of water present. Coatings made from them are reported to have
superior adhesion as well as high chemical, solvent, and mar resistance. Carboxylic
acid –functional acrylic latexes can be cross-linked using b-(3,4-epoxycyclohexyl)ethyltriethoxysilane [21].
REFERENCES
333
16.3. ORTHOSILICATES
Tetraethylorthosilicate (TEOS), Si(OEt)4, is used in making binders for zinc-rich primers,
which are widely used primers for corrosion protection of steel (Sections 7.4.3 and 33.1.2).
When exposed to water vapor in the atmosphere, Si(OEt)4 hydrolyzes, with acid or base
catalysis, to give a complex polysilicic acid network that forms the coating binder.
To make the primers, an ethyl alcohol solution of Si(OEt)4 is partially polymerized
by the addition of a small amount of water (just enough to increase the MW sufficiently to give the desired viscosity), and the zinc powder pigment is dispersed in
the resulting oligomer solution. Note that the reaction is reversible; the ethyl alcohol
solvent retards polymerization by small amounts of water in the coating formulation
during storage. When the coating is applied and the alcohol evaporates, water is
absorbed from the atmosphere, and the cross-linking reaction continues to completion
at ambient temperatures. Although the cross-linked binder is predominantly polysilicic
acid, zinc salts are incorporated in the matrix. They form by reaction with the zinc
hydroxide and zinc carbonate always present on the surface of zinc metal pigments.
Iron salts from the steel surface may also be incorporated in the matrix. (The chemistry
and properties of zinc-rich primers based on Si(OEt)4 are reviewed in Ref. [22].) For
waterborne zinc-rich primers, aqueous solutions of potassium, sodium, and/or lithium
silicate are being used.
16.3.1. Sol-Gel Coatings
Tetraethyl orthosilicate (TEOS) is used with various other components to prepare sol-gel
hybrid organic – inorganic vehicles for coatings. Such systems are commonly called
sol-gel coatings; another name is ormosil. For example, a sol-gel composed of the condensate of ethylsilane and tetramethyltetravinyl-cyclotetrasiloxane with water and a base
gives a tetrasilanol, which when combined with TEOS and an acid catalyst gives a
clear coat to be applied over a base coat –clear coat automotive coating. The resulting
finish had a better resistance to automatic car washing and sulfuric acid spot tests than
that of a control without the sol-gel coating [23]. Another application is treatment of
metals for corrosion protection replacing chromate pigmented primers. For example,
TEOS is added to water acidified with nitric acid; after hydrolysis, glycidoxypropyltrimethoxysilane is added and the resulting ormosil was applied to aluminum. Various
tests indicate that corrosion protection can be expected to exceed other approaches [24].
REFERENCES
1. Fink, F.; et al., J. Coat. Technol., 1990, 62(791), 47.
2. Spiegelhauer, S., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 161 – 170.
3. Hsiao, Y.-C.; et al., J. Appl. Polym. Sci., 1975, 19, 2817. See also Pappas, S. P.; Just, R. L.,
J. Polym. Sci. Polym. Ed., 1980, 18, 527.
334
SILICON DERIVATIVES
4. Finzel, W. A.; Vincent, H. L., Silicones in Coatings, Federation of Societies for Coatings
Technology, Blue Bell, PA, 1996.
5. Finzel, W. A., J. Coat. Technol., 1992, 64(809), 47.
6. Laubender, T., Proc. Int. Conf. Waterborne Coat., Milan, Italy, 1992, Paper 21.
7. Mangio, R.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 41–56; Paint Coat. Ind., 2001, January, 50.
8. Greene, J. D., U.S. patent 6,344,520 (2002).
9. Fransehn, P.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 2004, Paper 7.
10. Anderson, L. G.; et al., U.S. patent 6,803,408 (2004).
11. Ogarev, V. A.; Selector, S. L., Prog. Org. Coat., 1992, 21, 135.
12. Witucki, G. L., J. Coat. Technol., 1993, 65(822), 57.
13. van Ooij, W. J.; Child, T., Chemtech, 1998, February, 26.
14. Zhu, D.; van Ooji, W. J., Prog. Org. Coat., 2004, 49, 42.
15. Baghdachi, J.; et al., J. Coat Technol., 2002, 74(932), 81.
16. Ooka, M.; Ozawa, H., Prog. Org. Coat., 1993, 23, 325.
17. Furukawa, H.; et al., Prog. Org. Coat., 1994, 24, 81.
18. Nordstrom, J. D., Proc. Waterborne Higher-Solids Powder Coat. Symp., New Orleans,
LA, 1995, p. 192.
19. Yaneff, P. V.; et al., J. Coat. Technol., 2002, 74(933), 135.
20. Chen, M. J.; et al., J. Coat. Technol., 1997, 69(870), 43.
21. Chen, M. J.; et al., J. Coat. Technol., 1997, 69(875), 49.
22. Parashar, G.; et al., Prog. Org. Coat., 2001, 42, 1.
23. Hofacher, S.; et al., Prog. Org. Coat., 2002, 45, 159.
24. Parkhill, R. L.; et al., Prog. Org. Coat., 2001 41, 254.
17
Other Resins and
Cross-Linkers
The types of resins and cross-linkers used in the largest volumes in coatings were
described in Chapters 8 to 16. In this chapter, other commercially important materials
and selected promising developments are discussed.
17.1. HALOGENATED POLYMERS
Halogenated polymers have desirable properties for certain applications, including low
water permeability. Some are used in top coats for corrosion protection. Others serve as
binders for adhesion promoting primers for plastics.
17.1.1. Solution Grade Thermoplastic Vinyl Chloride Copolymers
Solution grade vinyl chloride resins are copolymers of vinyl chloride and vinyl acetate,
frequently with a small amount of a third monomer to impart special properties. The
vinyl acetate lowers Tg and broadens the range of applicable solvents. For example, a terpolymer of vinyl chloride, vinyl acetate, and maleic acid with a weight ratio of 86 : 13 : 1
(mole ratio 81 : 17 : 1) and an M̄w on the order of 75,000 has been used in interior coatings
for beverage cans. The maleic acid promotes adhesion; other solution vinyl resins are
made with hydroxy functionality. Use of solution vinyl resins has declined because of
the low solids required for application, about 10 to 12 nonvolatile volume solids (NVV).
Vinyl chloride copolymers require stabilization to prevent thermal and photochemical
degradation. As discussed in Section 5.3, the polymers undergo dehydrochlorination in an
autocatalytic chain reaction. A variety of stabilizing agents are used, including organotin
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
335
336
OTHER RESINS AND CROSS-LINKERS
esters such as dibutyltin dilaurate; barium, cadmium, and strontium soaps; maleates; and
oxirane compounds.
17.1.2. Vinyl Chloride Dispersion Copolymers
Dispersion grade vinyl chloride copolymers having high molecular weights (MW) can be
used in high-solids coatings. They are prepared by suspension polymerization as particles
with an average diameter of a few micrometers. In this process, a solution of an initiator in
the monomer mixture is added with vigorous stirring to a hot dilute solution of poly(vinyl
alcohol) (PVA) in water. The monomers are dispersed into small droplets, inside of which
the polymerization is initiated and completed, resulting in the formation of particles.
A monomer-soluble initiator is used in suspension polymerization in contrast to the watersoluble initiators used in emulsion polymerization. The PVA stabilizes the suspension by
minimizing coalescence of the particles when the temperature of the reaction is above the
Tg. Since the PVA has many easily abstractable hydrogens, graft side chains grow on the
PVA. After polymerization is complete, the polymer particles are separated by filtration,
washed, and dried. The MW of polymers made by suspension polymerization depends primarily on initiator structure and on the combination of reaction temperature and initiator
concentration. Commonly, M̄w is on the order of 100,000.
In plastics, dispersion grade vinyl chloride copolymers are used in plastisols. Plastisols
are polymer particles dispersed in a plasticizer. Since the Tg of the polymer is well above
room temperature, and also because the polymer is partially crystalline, the polymer does
not dissolve in the plasticizer at room temperature at an appreciable rate. When the plastisol is heated above the polymer Tg and above the melting point of the crystalline regions
in the polymer, the polymer dissolves in the plasticizer and the particles coalesce to a
molten state. When cooled, the product is a plastic consisting of a homogeneous solution
of polymer and plasticizer. The viscosity of plastisols is generally too high for application
as coatings. The viscosity can be lowered by addition of solvent. Solvents must be chosen
that dissolve the plasticizer without significantly swelling the polymer particles.
Application viscosities can generally be achieved with 20% solvent or less. These systems
are properly called organosols, but they are also frequently called plastisols even though
there is volatile solvent present.
17.1.3. Chlorinated Rubber, Chlorinated Ethylene Vinyl Acetate
Copolymers, and Chlorinated Polyolefins
Chlorinated rubber is used in top coats for heavy duty maintenance paints because of the
barrier properties resulting from its low water permeability. It is also used in tie coats on
polyolefin plastics. Chlorinated rubber is compatible with some alkyd resins and is used to
impart fast drying qualities to alkyd-based traffic paints. Like PVC, chlorinated rubber
dehydrochlorinates and requires stabilizers similar to those used with PVC. Some metal
salts, especially iron, tend to promote degradation of chlorinated rubber so it degrades
when applied over rusty steel [1].
Chlorinated rubber is prepared from natural rubber that has been milled to reduce its
MW. A solution of milled rubber in CCl4 is reacted with chlorine or some other chlorinating reactant. The reactions are complex: Addition to the double bonds, substitution, and
cyclization reactions all occur. To eliminate most of the double bonds from the rubber,
the final product contains 65 to 68 wt% chlorine. Various MW grades are manufactured.
17.1. HALOGENATED POLYMERS
337
Polymer strength increases with MW, but the viscosity of solutions also increases. The
formulator makes an appropriate compromise between coating solids and film properties
for a particular application. Environmental regulations in the United States have increased
the cost of producing chlorinated rubber to prohibitive levels; production has been discontinued, and only imported material is available.
Chlorinated ethylene/vinyl acetate copolymers have been developed which can be used
to replace chlorinated rubber in at least some applications [2]. Since the initial resin does
not contain double bonds, the chlorine content need not be as high as in chlorinated rubber;
grades with 52, 55, and 58% chlorine are available. The higher the chlorine content, the
better the barrier properties but the lower the flexibility and impact resistance. Performance is said to be comparable to chlorinated rubber, with the advantage of better
storage stability.
Chlorinated polyolefins (CPOs) are an important class of adhesion promoters in primers
for plastics such as “thermoplastic polyolefin” (TPO) (Section 31.2.2). TPO is used in a
large volume by the auto industry: for example, for bumpers. The predominant CPO is
a chlorinated polypropylene that has been modified with maleic anhydride.
17.1.4. Fluorinated Polymers
) n, has the greatest exterior durability and heat
Polytetrafluoroethylene, PTFE, 22
( CF2CF222
resistance of any polymer used in coatings. However, PTFE is insoluble in solvents and its
fusion temperature is so high that its coating uses are limited to applications in which the
substrate can withstand high temperatures [3]. For example, aqueous dispersions of PTFE
are used for coating the interior of chemical processing equipment and cookware [4]. After
application the polymer particles are sintered at temperatures as high as 4258C. Like all
highly fluorinated compounds, PTFE has such a low surface free energy that it is not
wet by either water or oils; hence, it provides a “non-stick” cooking surface. Adhesion
of PTFE is also problematic; in general, the metal surface is roughened so that “dovetail
joints” can be formed when the coating and substrate are baked (Section 6.1).
Since its fusion temperature is somewhat lower, poly(vinylidene fluoride), PVDF,
) n, can be used in additional applications. PVDF is used in coil coatings as
22
( CH2CF222
a plastisol-like dispersion in a solution of acrylic resin [3,5]. The fusion temperature of
the films is reported to be 2458C. The exterior durability is outstanding, but the cost is
high and only low gloss coatings are possible. Copolymers of vinylidene fluoride
(VDF) are also finding use in powder coatings.
Vinylidene fluoride copolymer latexes are said to permit formulation of paints that form
films at ambient temperatures. The latex is made from a seed latex that is a copolymer of
vinylidene fluoride and another perfluorinated monomer, resulting in a less crystalline
copolymer. With this seed, polymerization is continued with acrylic monomers, resulting
in latex particles with uniform distribution of acrylate and fluorinated polymers that are
film formers at ambient temperature. Film properties are reported to be excellent with
exterior durability with the same rating as solventborne PDVF [6].
Fluorinated copolymers with functional groups such as hydroxyl groups can be crosslinked after application. Copolymers of VDF with a hydroxy-functional monomer crosslinked with a polyisocyanate give coatings with superior wet adhesion and corrosion as
compared with PVDF homopolymer [7]. Perfluoroalkyl acrylate esters have been copolymerized with hydroxyethyl methacrylate (HEMA) to form solvent-soluble resins;
however, the monomers are expensive. Halofluoroethylene CF255CFX/vinyl ether
338
OTHER RESINS AND CROSS-LINKERS
copolymers have been used in steel building panels and to a limited extent in clear coats
for automobiles [3]. Vinyl ethers and CF255CFX form alternating polymers. Functional
groups can be introduced by copolymerizing hydroxy- and/or carboxy-substituted vinyl
ether comonomers. The Tg is controlled by the ratio of fluorinated monomer to vinyl
ether monomers and by the chain length of the alkyl group on the vinyl ether. Copolymers
with hydroxyl groups can be cross-linked with MF resins or polyisocyanates. Gloss retention on exterior exposure is reported to be superior to that of acrylic coatings.
Acrylic monomers that have fluorinated side chains (e.g., 2,2,2-trifluoroethyl methacrylate and perfluoroalkyl acrylate) are available. Polymers containing such monomers tend
to have very low surface energies and have been studied for applications such as antifouling and antigraffiti coatings. Concern is being expressed about environmental contamination by perfluoroalkylcarboxylic acids that apparently are derived by photoxidation of
perfluoroalkanols and are accumulating in fish and animals, especially in the arctic [8].
17.2. CELLULOSE DERIVATIVES
Cellulose is a naturally occurring polymer of repeating anhydroglucose units; as the main
building block of plant structures, it is insoluble in water and organic solvents. Cellulose
derivatives that are soluble in water and others that are soluble in organic solvents can be
made by reactions of the hydroxyl groups.
Many solvent soluble cellulose derivatives have been prepared. Those of most
importance in coatings are nitrocellulose (NC) and cellulose acetobutyrate (CAB); ethyl
cellulose, cellulose acetate, and cellulose propionate have been used to some extent.
Water-soluble cellulose derivatives are used to modify the flow properties of latex
paints and are discussed in Chapter 32.
17.2.1. Nitrocellulose
Nitrocellulose (NC) is another example of poor nomenclature since it is not a nitro compound. It is a nitrate ester and should be called cellulose nitrate. Three types of NC are
available with differing extents of esterification. Their compositions are given in
Table 17.1. An isomeric dinitroanhydroglucose unit of nitrocellulose is
17.2. CELLULOSE DERIVATIVES
339
TABLE 17.1. Composition of Nitrocellulose Types
Type
SS (spirit soluble)
RS (regular solubility)
Explosive
Nitrogen
2
2ONO2 Groups (%)
per Anhydroglucose Unit
10.7– 11.2
11.8– 12.3
12.3– 13.5
1.9–2.0
2.15–2.25
2.25–2.5
The SS type is alcohol soluble and is used in flexographic printing inks but has few uses in
coatings. The RS type is soluble in esters and ketones and is generally used in coatings.
The explosive type is included in Table 17.1 to emphasize the hazards of nitrocellulose.
The RS type used in coatings is not explosive but is highly flammable. To reduce the handling hazard, nitrocellulose is shipped wet with a nonsolvent. In most cases for the RS type,
alcohols are used and the NC is shipped at 70 NVW, with the balance being ethyl or isopropyl alcohol. Although RS nitrocellulose is not soluble in alcohol, it is soluble in mixtures of ketones and esters with alcohols and hydrocarbons (Section 18.2).
Nitrocellulose is made by reacting cellulose with nitric acid. Small amounts of water
are present, and some hydrolysis of the cellulose occurs, reducing the MW. A range of
MW grades is available. The higher the MW, the greater the polymer strength but the
lower the solids of coatings that can be made. Viscosity of nitrocellulose solutions is
tested with a falling ball viscometer, in which the time for a ball to fall a specified distance
through a nitrocellulose solution of standard concentration is measured. The various MW
grades are designated as 14, 12, 1 second, and so on. The times are proportional to the time in
the falling ball viscosity test. The larger the number of seconds, the higher the MW.
Nitrocellulose was used in lacquers for automobile top coats before acrylic lacquers,
which have better exterior durability, were developed. They continued to be used for
some years in refinish lacquers for repairing automobiles because they polished more
easily and gave a deeper gloss appearance than room temperature dried acrylic lacquers.
Nitrocellulose lacquer primers continue to be used to some degree in refinishing automobiles. The principal remaining use of NC lacquers is in wood finishing (Section
31.1). These lacquers have relatively low solids but continue to be used to a significant,
if decreasing, extent because they enhance the appearance of wood grain to a greater
extent than any other coating. Other desirable characteristics for NC lacquers for wood
finishing are rapid drying, which permits handling and shipment soon after finishing,
and easy damage repair, since the films are thermoplastic. Increasingly stringent VOC
emission regulations can be expected to force further reductions in use of NC.
Use of nitrocellulose lacquers has declined substantially in the United States and
Europe because of their high VOC, but use in other parts of the world is still substantial.
For example, it is reported that worldwide use in 2001 in furniture finishes was 533,000
metric tons [9]. Similarly, substantial amounts of automobile refinish nitrocellulose
lacquer continue to be used.
17.2.2. Cellulose Acetobutyrate
A variety of organic esters of cellulose are used in fibers, plastics, and to a degree, in coatings. The only one still used to a significant extent in coatings is a mixed acetate, butyrate
ester (CAB). CAB lacquers are used to some extent; they exhibit lighter color and better
color retention with reduced handling risk relative to NC lacquers. The principal use has
340
OTHER RESINS AND CROSS-LINKERS
been as a component in acrylic automotive coatings. CABs help control flow, particularly
to facilitate orientation of aluminum flakes parallel to the surface of the film during application and film formation. Orientation of the flakes in the final film enhances the appearance of metallic finishes. (See Ref. [10] and Chapter 29.)
Several types of CAB are available with different ratios of acetate/butyrate/unesterified hydroxyl groups and with differing MWs. The ratio affects the solubility and compatibility of the polymer. The larger the ratio of butyrate ester, the larger the amount of
aromatic hydrocarbon that can be tolerated in a solvent mixture. Also, as the butyrate
content increases, the Tg tends to be lower. Compatibility with acrylic resins is dependent
on both the CAB and the acrylic polymer. An example of a CAB grade used in acrylic
lacquers has an average of 2.2 acetate group, 0.6 butyrate group, and 0.2 unreacted
hydroxyl group per anhydroglucose unit. This corresponds to about 30% acetyl content,
17% butyryl content, and 1.5% hydroxyl content, based on the total weight of polymer.
17.3. UNSATURATED POLYESTER RESINS
Unsaturated polyester resins are maleic acid– containing polyesters dissolved in styrene.
They contain unsaturated maleate [Z-(cis) configuration] and fumarate [E-(trans) configuration] groups in the backbone. These resins are used on a large scale in glassreinforced plastics, commonly called fiberglass plastics. They are used on a smaller
scale in some coating applications.
A variety of diacids and diols can be used to make unsaturated polyesters; the most
common are phthalic anhydride (PA), maleic anhydride, and propylene glycol (or a
mixture of propylene glycol and propylene oxide). These are the lowest cost raw materials;
furthermore, the initial reaction can be carried out at moderate temperatures, and a
minimum amount of water is formed. During esterification, some of the maleate ester
groups isomerize to fumarate ester groups. The fraction isomerized depends on processing
time and temperature and can also vary depending on the other components of the polyester. Fumarate groups are more reactive during cross-linking than maleate groups. Reaction conditions during esterification must be carefully controlled to ensure that the extent
of isomerization is reproducible from batch to batch. Use of fumaric acid instead of maleic
anhydride results in a higher fumarate diester content; such polyesters generally give
harder plastics than those made with maleic anhydride. More flexible plastics can be
obtained by substituting aliphatic diacid (e.g., adipic acid) for part of the PA to reduce
Tg. Although other glycols can be used, economics favor use of propylene glycol (or
propylene oxide). The properties of the cured materials are also affected by the ratio of
polyester to styrene; typically, a 70 : 30 ratio is used.
The resin/styrene solution is cross-linked using free-radical initiators. The product is a
complex mixture derived from copolymerization of styrene with the maleate or fumarate
double bonds, graft copolymerization, and homopolymerization of styrene. Cross-link
density is affected by the MA/PA ratio in the resin, commonly 15 to 40 mol% of
maleic anhydride. Peroxides such as benzoyl peroxide are used at temperatures of 70 to
1008C. Many applications require cross-linking at ambient temperatures. In these cases,
an initiator such as methyl ethyl ketone (MEK) peroxide, and promoters such as a
mixture of dimethylaniline and cobalt naphthenate, are added to the resin/styrene solution
just before use. The combination reacts rapidly at room temperature to generate free radicals, which initiate cross-linking. MEK peroxide is a complex mixture in which the major
17.3. UNSATURATED POLYESTER RESINS
341
components contain both peroxide and hydroperoxide groups [11], as shown. The cobalt
salt acts as a redox catalyst for the decomposition of the peroxide and hydroperoxide
groups into free radicals; dimethylaniline further promotes that reaction.
Gel coats are pigmented unsaturated polyester/styrene coatings (with initiator and promoter); they are sprayed on the inside of a mold surface (Section 31.2.1). The exposed
(back) surface of the gel coat is then sprayed with glass fiber –loaded unsaturated polyester/styrene compound and then covered with plastic film and cross-linked. After the part
is taken from the mold, the surface is the gel coat that cured in the absence of oxygen.
Many glass-reinforced plastic objects, ranging from prefabricated shower stalls to boat
hulls, are made this way. In the case of boat hulls, hydrolytic stability is particularly
important, so the surface retains gloss after outdoor exposure and immersion in water.
Unsaturated polyesters made using neopentyl glycol, MA, and isophthalic acid provide
better gloss retention than those made from propylene glycol and PA, but the cost
is higher.
The free radical polymerization is oxygen inhibited. This is not a problem for many
plastics applications since the reaction is commonly carried out in a mold, so the resin
surface is not exposed to air. In most coating applications, the top surface is exposed to
the air and will remain sticky after the polymerization is complete below the surface. A
further problem is the volatility of styrene. These problems can be minimized by incorporating some insoluble semicrystalline paraffin wax in the formulation. After the coating is
applied, the low surface tension wax particles come to the surface preferentially. The wax
layer reduces the rate of styrene loss and also reduces oxygen concentration in direct
contact with the polymerizing coating, thereby minimizing the difficulty of surface
cure. The wax layer, however, results in a relatively uneven low gloss surface, suitable
for some applications but not for others.
Unsaturated polyesters can be used in UV cure coatings (Section 29.2.4). A photoinitiator is used that generates free radicals on exposure to UV radiation. High intensity radiation sources are used which generate very large numbers of free radicals sufficiently
rapidly at the surface so the oxygen in the air at the surface is depleted (by reaction
with radicals), and polymerization can proceed. Acrylate systems are more widely used
than styrene/unsaturated polyesters.
Oxygen inhibition can be minimized using coreactants having allyl groups with
styrene/unsaturated polyesters [12]. A free radical abstracts a hydrogen atom from the
methylene group activated by both the adjacent double bond and the ether oxygen. The
resulting free radicals from this reaction then react with oxygen to form peroxy free radicals, which in turn abstract hydrogen from the activated methylene groups to form hydroperoxides. The reactions consume some of the oxygen at the surface and generate new
hydroperoxide in a chain reaction; as a result, oxygen inhibition is reduced.
Use of allyl ether reactions has been extended to waterborne unsaturated polyester
resins. A low MW ester is prepared by reacting 2 mol of maleic anhydride with 1 mol
of a mixture of low MW diols and polyalkylene glycols. The resulting partial ester is
342
OTHER RESINS AND CROSS-LINKERS
further esterified with 2 mol of trimethylolpropane diallyl ether [13]. The polyalkylene
glycol ester segments are efficient emulsifying agents so the polyester can be emulsified
in water for application. Films can be cured either with hydroperoxide/cobalt initiators
or by photoinitiators and UV radiation.
17.4. (METH)ACRYLATED OLIGOMERS
Thermosetting resins can be made with (meth)acrylate double bonds as cross-linkable
groups. Such resins are used in radiation cured coatings (Section 29.2.4). They can also
be used in ambient cure and at somewhat elevated temperatures (force-dry) with free
radical initiators. Both acrylate and methacrylate polymerizations are air inhibited, particularly methacrylate systems, which generally require higher cure temperatures to minimize the effect of oxygen inhibition.
Acrylated oligomers can also be cross-linked with polyfunctional primary amines by a
Michael addition reaction, as illustrated for monofunctional reactants. This reaction is fast,
so practical systems are based on blocked amines, commonly ketimines [14]. When a
coating film is exposed to atmospheric moisture, the ketimine reacts with water to
release the free primary amine.
Acrylated oligomers can also be cross-linked by a Michael reaction with acetoacetylated resins and their enamine derivatives, as discussed in Section 17.6.
17.5. 2-HYDROXYALKYLAMIDE CROSS-LINKERS
Esterification of most alcohols with carboxylic acids is too slow for practical use as a
cross-linking reaction. However, 2-hydroxyalkylamides undergo esterification reactions
more rapidly than simple alcohols. Polyfunctional 2-hydroxyalkylamides (e.g., the tetrafunctional hydroxyalkylamide derived from aminolysis of dimethyl adipate with diisopropanolamine) can serve as cross-linkers for carboxylic acid-functional acrylic or polyester
resins [15].
The properties of coatings obtained by cross-linking carboxylic acid– functional acrylic
resins with hydroxyalkylamides compare favorably with those obtained using MF resins as
cross-linkers with the same resins. An advantage relative to MF cross-linkers is the
absence of formaldehyde, which is emitted in low concentrations when MF based coatings
17.6. ACETOACETATE CROSS-LINKING SYSTEMS
343
are baked. A disadvantage is that high baking temperatures are required; the lowest
reported satisfactory bake is 1508C for 30 minutes.
Hydroxyalkylamides are soluble in both water and common coating solvents and hence
are useful as cross-linkers in either waterborne or solventborne coatings. Tetra-N,N,
N0 ,N0 -(2-hydroxyethyl)adipamide is a solid well suited for use in powder coatings [16]
(Section 28.1.3). However, some discoloration can occur on baking [17], presumably
due to formation of some free amine groups by thermal rearrangement of the
hydroxylalkylamides.
The cross-linking reaction exhibits unique features, compared to general esterification
reactions. It is not catalyzed by acid; also, 2-hydroxyalkylamides with secondary
hydroxyls esterify more rapidly than a corresponding primary alcohol derivative. Aromatic
carboxylic acids esterify faster than aliphatic carboxylic acids, the reverse of other alcohols. In addition to the faster reaction of 2-hydroxyalkylamides with aromatic acids, the
saponification resistance of the products is greater than of the corresponding esters from
aliphatic acids. Mechanistic studies of the hydroxyalkylamide/carboxylic acid esterification reaction, as well as related reactions, have been reported. An explanation of
the results, involving participation of the amide group and intermediate formation of
oxazolinium groups, has been advanced that is consistent with the experimental observations [18]. Direct evidence for formation of oxazolinium intermediates has been
obtained [17].
It has also been shown that 2-hydroxyalkylamides react with epoxies. At elevated
temperatures, the amide rearranges to the aminoester, and the resulting amine group
then reacts with the epoxy [19].
17.6. ACETOACETATE CROSS-LINKING SYSTEMS
The chemistry of b-keto esters such as acetoacetic esters has been widely studied. Acetoacetic esters are weak acids and exist to a significant degree in the tautomeric enol form.
They undergo a wide range of reactions, some of which can be adapted for coatings
purposes. Several methods are available to prepare resins bearing acetoacetate groups.
Acetoacetoxy-functional acrylic solution resins can be made by copolymerizing acetoacetoxyethyl methacrylate (AAEM) with other acrylate monomers [20,21]. One can also use
AAEM to make acrylic latexes [22]. However, the long-term stability of acetoacetic ester
polymers in aqueous media has been questioned. Hydroxy-functional resins can be reacted
with diketene or transesterified with methyl acetoacetate to form acetoacetylated resins.
Replacement of the hydroxyl group with the less polar acetoacetate group leads to a
reduction in viscosity, permitting higher solids at the same application viscosity.
344
OTHER RESINS AND CROSS-LINKERS
A variety of coreactants with acetoacetylated resins have been investigated.
Melamine – formaldehyde resins react with acetoacetate groups in the presence of an
acid catalyst, although somewhat less rapidly than with hydroxyl groups [20]. Film properties of the resulting coatings using acetoacetylated resins are reported to be comparable
with those made from hydroxy-functional resins. There are indications of improved wet
adhesion, perhaps resulting from chelating interactions with metal surfaces.
Isocyanates also react with acetoacetate groups; the cure rate is slower than with
hydroxyl groups, but the pot life is longer. Film properties are reported to be similar to
those of conventional polyurethanes [20]. In other examples of 2K coatings, based on
MDI, the solids at equal viscosity were about 10% higher with the acetoacetylated resin
than with the original hydroxy resin, while retaining excellent film properties [23].
Polyacrylates (e.g., trimethylolpropane triacrylate) undergo Michael reactions with
acetoacetate groups at ambient temperatures in the presence of strongly basic catalysts
such as tetramethylguanidine (TMG) [22]. Formic acid has been reported to be an effective
volatile blocking agent to extend pot life when 1,8-diazabicyclo[5.4.0]undec-7-ene is used
as a catalyst [21].
Amines react rapidly with acetoacetic esters; hence, polyamines can be used as crosslinkers for acetoacetate-functional latexes [22]. Amines have also been reported to be
excellent cross-linkers for acetoacetate-functional water-reducible acrylic resins. Reactions are too rapid under many conditions to have sufficient pot life even as 2K coatings.
However, it has been found that if the resin is neutralized with ammonium hydroxide and
17.7. POLYAZIRIDINE CROSS-LINKERS
345
the solids of the vehicle are restricted to 15 to 20%, stability is improved. For example, a
copolymer of AAEM, MAA, BA, and MMA, neutralized with an equal equivalent of
NH4OH at 18% solids, blended with 1,6-hexamethylenediamine, is storage stable for a
year. It was proposed that part of the acetoacetoxy groups react with the NH4OH to
form the amino-enamine, thus minimizing cross-linking by the hexamethylenediamine.
After application, the ammonia evaporates and the resin is cross-linked by the hexamethylenediamine [24].
In solution coatings, the reaction is so fast at ambient temperatures that pot life is
limited. This problem can be minimized by blocking the amine groups with ketones.
The resulting ketimine hydrolyzes in the presence of water, permitting the cross-linking
reaction to proceed [25]. The cross-links are tautomeric ketimine –eneamine groups that
are thought to interact strongly with metal surfaces [26]. A ketimine –acetoacetate crosslinking primer is reported to give excellent adhesion and corrosion resistance when
applied to an aircraft grade aluminum alloy with a chromate-free pretreatment.
Ketimine/eneamine tautomers can also serve as Michael reactant cross-linkers with
polyacrylates [27]. Acetoacetate pendant resins also can be cross-linked by reaction
with aldimines at room temperature [28].
17.7. POLYAZIRIDINE CROSS-LINKERS
Aziridine, the nitrogen three-membered ring counterpart of oxirane, and its derivatives
have been studied for many years; in some cases polyaziridines have been used as crosslinkers. The common name of aziridine is ethyleneimine. Ethyleneimine is highly toxic
and may be carcinogenic. Propyleneimine is somewhat less toxic. Polyaziridines are
skin irritants, and some individuals may become sensitized. The mutagenicity of polyaziridines is controversial; however, dilution by coating vehicles reduces possible toxic effects
[29]. Ethyleneimine is even more reactive with acids than is ethylene oxide. In the presence of relatively strong acids it polymerizes very rapidly to yield poly(ethyleneimine),
) n.
22
( CH2CH2NH22
Among the many reactions of aziridines, the one of greatest interest in coatings applications is the reaction of a polyaziridine with a polyfunctional carboxylic acid to form
2-aminoester cross-links. Some 2-aminoesters can spontaneously rearrange to the corresponding 2-hydroxyamides, but this reaction does not break the cross-link.
A variety of polyfunctional aziridines have been investigated. An example is the
trifunctional Michael addition product of 3 mol of aziridine to 1 mol of
346
OTHER RESINS AND CROSS-LINKERS
trimethylolpropane triacrylate. A review paper covers tris-3-(1-aziridino)propionates
and their uses. Trimethylolpropane tris[3-(2-methylaziridino)propionate], trimethylolpropane tris(3-aziridinopropionate), and pentaerythritol tris(3-aziridinopropionate) are
evaluated [30].
Carboxy groups, amines, alcohols, and mercaptans react by opening the ring, giving a
secondary amine as a cross-link. The reactions are acid catalyzed. The main use of such
polyaziridines is to cross-link with carboxylic acid groups on latexes and waterborne polyurethanes. They are particularly useful for cross-linking COOH-functional latexes in
coatings for factory-coated wood cabinetry and flooring, where a modest degree of crosslinking improves film properties over those attainable with un-cross-linked latexes. Reaction with the carboxylic acid is much faster than the reaction of water with aziridine
groups, but the reaction rate with water is such that the pot lives are 48 to 72 hours
[29]. With water the aziridine hydrolyzes to an amino alcohol, but it is said that there is
no indication that the hydrolyzed aziridine affects film properties adversely. Additional
cross-linker can be added to restore reactivity. Solvent swelling tests on cured films
show that cross-linking is more complete with the aziridine than comparison tests with carbodiimide or MF resin cross-linked PUD. (See Section 12.7.1 for further discussion.) In
view of the potential toxic hazards, manufacturers’ recommendations for safe handling
should be followed carefully.
17.8. POLYCARBODIIMIDE CROSS-LINKERS
Carbodiimides react with carboxylic acids and react slowly enough with water so they can
be used in waterborne coatings, including latexes. The product of the reaction with a carboxylic acid is an N-acylurea:
Multifunctional carbodiimides are available to cross-link carboxylic acid– functional
resins, including latexes, polyurethane dispersions, and blends of the two [31]. Crosslinking occurs within several days at ambient temperature and much faster with heat
[32]. In one study, with latexes, curing conditions ranged from 60 to 1278C for 5 to
30 minutes, and the higher temperatures gave better films [31]. Apparently, film properties depend on physical film formation as well as on the extent of chemical
cross-linking.
A water-emulsifiable polycarbodiimide that is terminated with a mixture of the methyl
ether of diethylene glycol and a methyl ether of a polyethylene glycol is used as a crosslinker for COOH-functional latexes. The design of the latex markedly affected properties.
The best properties were obtained with a BA/MMA/MAA latex that was prepared with
the core containing 2% MAA and the shell 1% MAA [33].
REFERENCES
347
17.9. POLYCARBONATES
Polytrimethylene carbonate polyols have been recommended as reactive diluents for
acrylic polyols in coatings cross-linked with HDI isocyanurate. Best results were obtained
with relatively low MW triols made by reacting trimethylene carbonate with TMP. VOC
was reduced while increasing film flexibility and maintaining hardness, weather resistance, and mechanical properties [34].
Semi-crystalline polycarbonate resins are prepared by cationic ring-opening polymerization of dimethyl dimethylene carbonate with polyols using fumaric acid as a
catalyst. Various polyols can be used. There was particular interest in low MW resins
prepared using neopentyl glycol. As discussed in Section 28.1.6, methacrylate esters are
of interest in UV cure powder coatings [35].
The possibility of using polycyclic carbonates as resins that can be cross-linked by
polyamines to polyhydroxyurethanes is being explored. A paper reviewing the preparation
and reactions of polycarbonate resins has been published [36]. The cyclic carbonate polymers can be synthesized by transesterification of glycerol carbonate with MMA, copolymerizing the resulting carbonate methacrylate, and reaction of glycidyl methacrylate with
CO2, and copolymerizing. The polyfunctional carbonates can then be cross-linked with
polyamines.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Morcillo, M.; et al., Prog. Org. Coat., 1993, 21, 315.
Hypalon CP Chlorinated Polyolefins, tech. bull., Du Pont Company, Wilmington, DE, 1989.
Munekata, S., Prog. Org. Coat., 1988, 16, 113.
Batzar, K., Proc. Int. Conf. Org. Coat., Athens, Greece, 1989, p. 13.
Gaske, J. E., Coil Coatings, Federation of Societies for Coatings Technology, Blue Bell, PA,
1987.
Wood, K. A.; et al., http://specialchem4coatings.
Barbucci, A.; et al., Prog. Org. Coat., 1996, 29, 7.
Ritter, S. K., Chem. Eng. News, 2006, 84(5), 37.
Irle, C.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2003, pp.
357 – 374.
McBane, B. N., Automotive Coatings, Federation of Societies for Coatings Technology, Blue
Bell, PA, 1987.
Sheppard, C. S., in Encyclopedia of Polymer Science and Engineering, 2nd ed., Wiley,
New York, 1988, Vol. 11, pp. 1 – 21.
Traencknerm H.-J.; Pohl, H. U., Angew. Makromol. Chem., 1982, 108, 61.
Dvorchakm, M. J.; Riberi, B. H., J. Coat. Technol., 1992, 64(808), 43.
Noomen, A., Prog. Org. Coat., 1989, 17, 27.
Lomax, J.; Swift, G. F., J. Coat. Technol., 1978, 50(643), 49.
Mercurio, A., Proc. Int. Conf. Org. Coat., Athens, Greece, 1990, p. 235.
Stanssens, D.; et al., Prog. Org. Coat., 1993, 22, 379.
Wicks, Z. W., Jr.; et al., J. Coat. Technol., 1985, 57(726), 51.
Jones, F. N.; Lin, I-C., J. Appl. Polym. Sci., 1984, 29, 3213.
Rector, F. D.; et al., J. Coat. Technol., 1989, 61(771), 31. Witzeman, J. S.; et al., J. Coat.
Technol., 1990, 62(789), 101.
348
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
OTHER RESINS AND CROSS-LINKERS
Li, T.; Graham, J. C., J. Coat. Technol., 1993, 65(821), 64.
Geurnik, P. J. A.; et al., Prog. Org. Coat., 1996, 27, 73.
Narayan, R.; Raju, K. V. S. N., Prog. Org. Coat., 2002, 45, 59.
Esser, R. J.; et al., Prog. Org. Coat., 1999, 36, 45.
Clemens, R. J.; Rector, F. D., J. Coat. Technol., 1989, 61(770), 83.
Zabel, K. H.; et al., Prog. Org. Coat., 1998, 34, 236.
Schubert, W., Prog. Org. Coat., 1993, 22, 357.
Chen, N.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1996,
p. 198.
Pollano, G., Polym. Mater. Sci. Eng., 1997, 77, 383.
Roesler, R. R.; Danielmeier, K., Prog. Org. Coat., 2004, 50, 1.
Taylor, J. W.; Basset, D. R., in Technology for Waterborne Coatings, Glass, J. E., Ed., American
Chemical Society, Washington, DC, 1997, p. 137.
Ucarlink Crosslinkers, tech. bull., Dow Chemical Company, Midland, MI, 2001.
Porzio, R. S.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2003, pp. 129 – 143.
Zhou, L.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2000,
pp. 262 –281.
Löwenhielm, P.; et al., Prog. Org. Coat., 2005, 54, 269.
Webster, D. C., Prog. Org. Coat., 2003, 47, 77.
18
Solvents
Most coatings contain volatile material that evaporates during application and film formation. The volatile components reduce viscosity for application and control viscosity
changes during application and film formation. In some cases, the volatile material must
be a solvent for the resins in a formulation; in other cases, nonsolvent volatile components
are desired. Usually, any volatile organic material is called a solvent whether or not it dissolves the resin. Too often, formulators do not consider the effects that solvent choice can
have on coating performance. Selection of volatile components affects popping, sagging,
and leveling and can affect adhesion, corrosion protection, and exterior durability.
As discussed in Sections 18.7, 18.8, and 18.9 almost all solvents are classified by the
U.S. Environmental Protection Agency (EPA) as photochemically reactive volatile
organic compounds (VOCs), and their use has been regulated to reduce air pollution
since the 1970s. In 1990, the U.S. Congress listed certain common solvents as hazardous
air pollutants (HAPs), further limiting their use.
18.1. SOLVENT COMPOSITION
Various organic compounds and mixtures are used as solvents. They can be classified in
three broad categories: weak hydrogen-bonding, hydrogen-bond acceptor, and hydrogenbond donor – acceptor solvents.
Weak hydrogen-bonding solvents are aliphatic and aromatic hydrocarbons. Commercial aliphatic solvents are mixtures of straight chain, branched chain, and alicyclic hydrocarbons. They vary in volatility and solvency. Varnish makers and painters (VM&P)
naphthas are aliphatic solvents with high volatility. Mineral spirits are slower-evaporating
aliphatic hydrocarbons; special grades with low aromatic content, and hence less odor,
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
349
350
SOLVENTS
are available. An advantage of aliphatics is low cost, especially on a volume basis since
their densities as well as their price per unit weight are low. Aromatic hydrocarbon
solvents are more expensive than aliphatic ones but dissolve a broader range of resins.
Toluene, xylene, and high flash aromatic naphthas have been in widely use. Commercial
xylene is a mixture of isomeric xylenes and ethyl benzene. Higher-boiling mixtures of
aromatic hydrocarbons (high flash aromatic naphthas) are predominantly mixed alkyl
and dialkyl benzenes with three to five carbons in the substituent groups. Use of
benzene is prohibited because of toxicity. Toluene and xylene have been used on a
large scale; however, they are listed as HAPs (Section 18.8).
Chlorinated solvents, in general, are no longer used in coatings, primarily because most
of them lead to ozone depletion in the stratosphere. However, p-chlorobenzotrifluoride
(PCBTF) is inert and is exempt as a VOC and HAP. The solvency is similar to that of
xylene. Ostrowski discusses formulation with PCBTF [1].
Esters and ketones are hydrogen-bond acceptor solvents. Ketones are generally less
expensive than esters with corresponding vapor pressures. The cost differential is particularly marked on a volume basis since ketones have lower densities. Use of methyl ethyl
ketone (MEK) and methyl isobutyl ketone (MIBK) is being reduced because they are
on the HAP list; use of acetone is increasing because it is exempt as either a VOC or a
HAP. Use of esters is increasing. Slow evaporating esters such as 1-methoxy-2-propyl
acetate and 2-butoxyethyl acetate are often preferred over slow evaporating ketones
such as isophorone and methyl n-amyl ketone because of odor. Methyl acetate and
t-butyl acetate are exempt as either VOC or HAP. Esters of 2-ethoxyethanol were
widely used at one time but are on the HAP list, but no control is required if emissions
are less than 10 tons per year. (See Section 18.8 for a list of HAPs.)
Esters should not be used as solvents for resins bearing primary or secondary amine
groups because aminolysis converts esters to amides. An exception to this statement is
t-butyl acetate. It has been shown that epoxy – amine coatings lose amine almost as
slowly in t-butyl acetate as in xylene [2].
Nitroparaffins such as 2-nitropropane are highly polar hydrogen-bond accepting
solvents. Their high polarity results in increased electrical conductivity, which is useful
for adjusting solvent combinations in electrostatic spraying (Section 23.2.3).
Alcohols are strong hydrogen-bond donor – acceptor solvents. The most widely used
volatile alcohols are methyl, ethyl, isopropyl, n-butyl, sec-butyl, and isobutyl
alcohols. Most latex paints contain a slow evaporating water-soluble solvent such as propylene glycol that does not dissolve in the polymer particles; it serves as antifreeze
(Section 32.1). In water-reducible acrylic and polyester resin coatings, low volatility
ether alcohols such as 1-propoxypropan-2-ol and 2-butoxyethanol are used.
The value of solvents in the U.S. manufacture of paints and coatings in 1997 and 2002
are shown in Table 18.1. The comparisons may be affected to some extent by relative
TABLE 18.1. Value of Solvents Used in Paint and
Coatings (Millions of Dollars)
Solvent Class
1997
2002
Hydrocarbons
Alcohols
Esters
Ketones
Glycol and derivatives
223.5
77.7
103.4
106.7
108.0
206.4
62.5
136.0
101.9
141.8
Source: Ref. [3].
18.2. SOLUBILITY
351
differences in prices. Overall use has declined. The use of hydrocarbons and ketones was
particularly affected by the HAP limitation. The increase in glycol derivatives results
from their use in the growing field of waterborne coatings.
18.2. SOLUBILITY
In the early days of the paint industry, there was little problem selecting solvents—
almost all resins dissolved in hydrocarbons. An exception was shellac, which dissolved
in ethyl alcohol. Early in the twentieth century, nitrocellulose lacquers were introduced,
and solvent selection became a greater challenge. RS nitrocellulose (Section 17.2.1) is
soluble in esters and ketones but not in hydrocarbons or alcohols. However, RS nitrocellulose does dissolve in a mixture of ketones or esters and hydrocarbons, which
reduces solvent cost. If ethyl alcohol is included in the solvent mixture, a higher
ratio of hydrocarbons can be used, further reducing cost. In those days, esters and
ketones were classified as true solvents, hydrocarbons as diluents, and alcohols as
latent solvents.
To control the drying rate of nitrocellulose lacquers under various conditions, mixtures
of solvents are required. At a minimum, the solvent blends contain two esters or ketones,
two hydrocarbons, and an alcohol. Solvent mixtures must be selected to maintain solubility throughout solvent evaporation from films. If the slowest evaporating solvent were a
hydrocarbon, nitrocellulose would precipitate before all the solvent was gone to give an
uneven film with poor appearance and physical properties. As many as 10 solvents
are blended to control drying and solubility through the drying cycle at the lowest
volume cost.
18.2.1. Solubility Parameters
After 1930, more and more types of resins were adopted for coatings, and empirical
selection of solvents and solvent mixtures became more difficult. The general rule of
“like dissolves like” was broadened by the experience with nitrocellulose that mixtures
of weak hydrogen-bonding hydrocarbons with strong hydrogen-bonding alcohols give
solvency similar to medium hydrogen-bonding esters and ketones. In the 1950s, Burrell
initiated studies to develop a more scientific basis for selecting solvents and formulating
solvent mixtures [4]. He turned to the work of Hildebrand, who analyzed miscibility of
small molecule organic compounds in terms of thermodynamics [5]. Hildebrand
showed that the tendency of a pair of chemicals to mix spontaneously can be described
in terms of Gibbs free energy equation:
DGm ¼ DHm T DSm
Miscibility is thermodynamically favored when the change in free energy of mixing,
DGm, is negative. Hildebrand pointed out that the change in entropy DSm is usually positive, since a solution is usually less ordered after mixing than before. (There are exceptions
when strong interactions increase the entropic order of mixtures relative to the separate
materials, resulting in a negative entropy change. But in most cases, DSm is positive
and the entropy change favors mixing.) Hildebrand focused on DHm, the enthalpy
of mixing, since it usually determines the sign of DGm. Change in DHm is in turn
352
SOLVENTS
related to change in the energy of mixing DEm; R is the gas constant and T is temperature
(kelvin):
DEm ¼ DHm RT
Liquids have intermolecular attractive forces strong enough to hold the molecules
together; otherwise, they would be gases. The forces can be measured by determining
the energy needed to vaporize a liquid at that temperature. Results are expressed in
terms of the molar energy of vaporization divided by the molar volume V. Hildebrand
reasoned that the energy required to separate molecules during mixing is related to
cohesive energy density DEv/V. He expressed the change in DEm of ideal solvents by
the following equation, where Vm is the average molar volume and f1 and f2 are the
volume fractions of two components:
"
#2
DEv 1=2
DEv 1=2
DEm ¼ Vm f1 f2
V 1
V 2
The square root of cohesive energy density of a solvent was defined as the solubility
parameter d. A pair of solvents are miscible if the differences in solubility parameters
approach zero. Then the DEm of the pair approaches zero, the DHm is small, and the
free energy of mixing is controlled by DSm. When DSm is positive, the solvents are
miscible. DEv and Vm vary with temperature; therefore, d varies with temperature. Most
tables of solubility parameters give the values at 258C, although commonly, no temperature is specified.
There is risk of confusion about units. The older units of solubility parameter were
(cal cm23)1/2, sometimes designated as hildebrands. In the SI system, the proper units
are (MPa)1/2; 1(cal cm23)1/2 ¼ 0.488(MPa)1/2. The SI units have not been widely
adopted and (cal cm23)1/2 units are still used, although frequently not stated.
When data are not available for energy of vaporization, solubility parameters can be
estimated by using empirical equations relating boiling points, vapor pressure data, or
surface tension data [6]. They can also be estimated by summation of Small’s molar attraction constants, G, using the following equation, where r is density and M is molecular
weight (MW):
d¼
rX
1X
G¼
G
M
V
Selected Small’s constants are listed in SI units in Table 18.2. As with solubility
parameters, the units for Small’s constants are commonly not specified and are frequently
the old units. Hoy [7] used Small’s constants to calculate solubility parameters of 640
compounds.
Both Small’s constants and solubility parameters of hydrogen-bond donor – acceptor
molecules vary with the environment. The value for the alcohol hydroxyl group included
in Table 18.2 depends on other groups in the solvent, with which the hydroxyl group might
hydrogen bond, and on the polarity of other components in mixed solutions. Water is
the most extreme case. Although specific values of a solubility parameter of water are
18.2. SOLUBILITY
353
TABLE 18.2. Small’s Molar Attraction Constants, (MPa)1/2 (cm3 mol21) at 2588 C
Hydrocarbon Group
2
2CH3
2
2CH2
2
2CH2
2
5
5CH2
2
5
5CH2
Phenyl
Phenylene
G
Other Groups
G
284
270
176
249
259
1400
1370
O (ethers)
O (oxiranes)
Cl
CO (ketones)
COO (esters)
OH
236
361
420
539
668
463
sometimes encountered, the values vary widely, since they are very dependent on the
medium. Their usefulness is limited to comparing systems with similar compositions.
Burrell tried to apply solubility parameters to prediction of the solubility of resins [4].
The volatility of resins is so low that DEv at 258C cannot be directly measured. Burrell got
around this by determining the solubility of a resin in a series of solvents with known
solubility parameters. The solubility parameter range of a resin was taken as the range
of solubility parameters of the solvents that would dissolve the resin. He found that in
many cases, resins were not soluble in all the solvents with solubility parameters that
fell within the solubility parameter range determined for the resin. Most of the cases, in
which there were discrepancies, involved solvents and resins having markedly different
potential for hydrogen-bonding. He divided solvents into poor, medium, and strong
hydrogen-bonding, then determined the solubility of the resins in a series of each class
of solvents. The resulting ranges of solubility parameters for a resin in the three groups
of solvents permitted fairly good predictions of solubility. Burrell found that one can
predict with some confidence whether a solvent mixture would dissolve a resin. One
can calculate the weighted average solubility parameter of a mixture known to dissolve
a resin using the following relationship, where the x values are the mole fraction of
solvents in the mixture:
dmix ¼
x1 v1 d1 þ x2 v2 d2 þ x3 v3 d3 þ
x1 v1 þ x2 v2 þ x3 v3 þ
Since molar volumes for most solvents are fairly similar, a simplifying approximation
is to use volume fractions to calculate the average solubility parameter:
dmix ¼ f1 d1 þ f2 d2 þ f3 d3 þ
One does a similar calculation to determine the average degree of hydrogen bonding.
Using this procedure, it was almost always possible to predict whether an alternative
solvent mixture would also dissolve the resin. Next, Burrell applied the idea to prediction
of solvents or solvent mixtures for new resins using Small’s molar constants to estimate
the solubility parameter of the new resin. Gram equivalent weight of the average repeating
unit of the resin was used for the MW in Small’s equation. Results were fair; in a majority,
but not in all cases, the resins were soluble in solvents and solvent mixtures, with solubility
parameters similar to those calculated for the resin, provided that the need for similar
levels of hydrogen-bonding was taken into consideration.
354
SOLVENTS
A variety of refinements and extensions of Burrell’s techniques have been proposed.
Lieberman mapped solubility of polymers in a grid in which the axes are solubility parameter and hydrogen-bonding index [8]. The plot predicts that any solvent or solvent
blend with solubility parameter and hydrogen-bonding index values that fall within
the boundaries of the plot will dissolve the resin. Such plots can be useful, but their
preparation requires substantial experimental effort [9].
18.2.2. Three-Dimensional Solubility Parameters
Other investigators considered this two-dimensional approach to be too simplistic and proposed three-dimensional systems. The system of Hansen [10] came to be most widely
accepted. Hansen reasoned that since there are three types of interactive forces between
molecules, there should be three types of solubility parameters: dispersion, dd; polar,
dp; and hydrogen bond, dh. The total solubility parameter was arbitrarily set equal to
the square root of the sum of the squares of the partial solubility parameters:
d ¼ (d2d þ d2p þ d2h )1=2
For mixed solvents, one can calculate a weighted average of the three partial solubility
parameters:
dd(blend) ¼ (wdd )1 þ (wdd )2 þ þ (wdd )n
dp(blend) ¼ (wdp )1 þ (wdp )2 þ þ (wdp )n
dh(blend) ¼ (wdh )1 þ (wdh )2 þ þ (wdh )n
Three-dimensional solubility parameters can be determined or calculated by a variety
of methods. Although values are commonly given as three figures, it should not be inferred
that the third figure is significant. Table 18.3 gives solubility parameters for some representative solvents selected from tables in Ref. [10] and the Polymer Handbook [11].
TABLE 18.3. Three-Dimensional Solubility Parameters, (MPa)1/2
Solvent
Total
dd
dp
dh
n-Hexane
Toluene
o-Xylene
Acetone
Methyl ethyl ketone
Methyl isobutyl ketone
Isophorone
Ethyl acetate
Isobutyl acetate
n-Butyl acetate
Methyl alcohol
Ethyl alcohol
Isopropyl alcohol
n-Butyl alcohol
2-Butoxyethanol
14.9
18.2
18.0
19.9
19.0
17.0
19.8
18.2
16.8
17.4
29.7
26.6
23.5
23.1
20.9
14.9
18.0
17.0
15.5
16.0
15.3
16.6
15.8
15.1
15.8
15.1
15.8
16.4
16.0
16.0
0
1.4
1.4
10.4
9.0
6.1
8.2
5.3
3.7
3.7
12.3
8.8
6.1
5.7
5.1
0
2.0
3.1
7.0
5.1
4.1
7.4
7.2
6.3
6.3
22.3
19.4
16.4
15.8
12.3
18.2. SOLUBILITY
355
Hansen determined three-dimensional solubility parameters for a group of resins by
experimentally testing solubilities of 34 polymers in 90 solvents [10]. As an alternative
to this laborious procedure, Hoy calculated three-dimensional solubility parameters of
resins by a method analogous to use of Small’s constants [12]. The values determined
or calculated by various methods are frequently not the same; tables of (presumably)
self-consistent values have been published and are available in computer databases.
Three-dimensional solubility parameters are based on thermodynamic laws, but several
assumptions and arbitrary choices are involved in their derivation and use. Assertions
that they are “theoretically sound” can be misleading [6]. They are best regarded as an
empirical method that has proven useful for finding alternative solvent mixtures with
similar solvency characteristics.
Attempts have been made to apply solubility parameters to other problems, but the
theoretical foundations are shaky and results have been erratic. For example, use of solubility parameters to predict solubility of one polymer in another (compatibility) frequently
gives erroneous predictions. It has been increasingly recognized that three-dimensional
solubility parameters are an oversimplification of the complex factors involved in
solubility. The difficulties result from at least two interrelated factors. First, they
neglect entropy changes, and second, the hydrogen-bond solubility parameter combines
the effects of donors and acceptors. Entropy changes are particularly likely to be significant in hydrogen-bonding systems. Hansen has recognized the problem and in some cases
suggests giving increased weight to the hydrogen-bond parameter [10].
The difficulties increase as polymer MW increases. As molecules get larger, interaction
between solvent and polymer molecules must become greater to overcome intermolecular
polymer – polymer interactions. Intramolecular interactions may also play a role.
Estimates of solubility parameters of polymers by use of group attraction constants do
not take MW into consideration.
People first working with polymers may be surprised by some of the solubility effects.
They are accustomed to upper limits of solubility; for example, 36.1 g of NaCl dissolves
in 100 g of water at 258C. In general, there is no upper limit to the solubility of a
polymer. If a small amount is entirely soluble, it is quite safe to conclude that any
larger amount will be soluble in that solvent. However, it is common for there to be a
lower limit of solubility. A high concentration of polymer may be soluble in some
solvent(s), but on dilution, part of the polymer may precipitate. This phenomenon can
be used to fractionate polymers by dilution. The fractions precipitating first are generally
the highest-MW components. As dilution is continued further, progressively lower MW
fractions are precipitated. In some cases, the fractionation is based on the polarity of
parts of the resin. Many alkyds are soluble at high concentrations in aliphatic solvents,
but precipitate partially on dilution. The first fractions precipitating are high molecular
weight (MW) molecules with a larger than average number of hydroxyl and/or carboxylic
acid groups. One can think of the situation in reverse: Some solvent is soluble in the
polymer (resin), and all of the polymer is soluble in the combination of polymer and
solvent. As more solvent is added, the solvency of the system changes, and parts of the
polymer are not soluble in the more dilute polymer solution. The precipitate is highly
swollen with solvent.
Hoy showed that glycol ethers can change apparent polarity to assume the polar nature
of their surrounding environment [7]. In polar solvents, they behave as polar solutes.
However, in nonpolar solvents, they undergo intramolecular bonding or intermolecular
hydrogen bonding, to form dimers. In view of this consideration, it is not surprising
356
SOLVENTS
that solubility predictions often fail, since the resins are often polydisperse, polyfunctional
materials with multiple hydrogen-bond donor and acceptor sites.
18.2.3. Other Solubility Theories
Numerous refinements of solubility parameters have been proposed. Some have involved
adding a fourth parameter, for example by dividing dh into hydrogen-bond donor and
acceptor terms. Huyskens and Haulait-Pirson proposed equations that reflect changes in
entropy and attempt to account for the differences resulting from both hydrogen-bond
acceptor and donor groups [13]. They provide somewhat better predictions, but only the
relatively simple cases of poly(vinyl acetate), poly(methyl methacrylate), and poly(ethyl
methacrylate) were studied. They explain the shortcomings of the three-dimensional
system but fall short of giving a broadly applicable alternative.
A major series of studies involving cooperative work in several laboratories led to the
development of what a technical news story called a “universal solubility equation” [14].
The researchers themselves “believe this to be somewhat of an overstatement” [15]. The
complexity of polymer solubility can be seen from the fact that this universal equation
has 13 factors in a five-term equation and still does not take into consideration several
important situations. Among the limitations critical to coatings is that thus far, the
equation has been applied primarily to monofunctional solutes.
18.2.4. Practical Considerations
Meanwhile, what can formulators do? Solvent changes are frequent, motivated by new
toxic hazard information, changes in relative cost, new regulations, temporary shortages,
and so on. This is the type of problem for which solubility parameters are most useful,
recognizing that the results may be imperfect. One can use solvent computer programs
with data banks that include solubility parameters, evaporation rates, cost, density, and
so on. This permits calculation of alternative solvent mixtures with solubility characteristics that do not change substantially during evaporation, while minimizing cost per
unit volume. Solvent suppliers have developed sophisticated computer programs that
can be used to calculate alternative solvent blends on a customer service basis. The
result of such a calculation is given in Table 18.4, in which solubility parameter units
are converted to (MPa)1/2 and rounded off [16]. Solvent blend 1 contained 2-ethoxyethyl
acetate. which because of health concerns had to be removed from the formulation.
Blends 2 and 3 were calculated to have close to equal average solubility parameters and
evaporation rates. Blend 2 was unsatisfactory because the slowest evaporating solvent,
xylene, was not a solvent for the resin. Blend 3 was a better starting point since the
slowest evaporating solvent, 1-methoxy-2-propyl acetate, was a true solvent for the
resin. Relative evaporation rates are also given in Table 18.4. Relative evaporation rates
of mixed solvents must be viewed with caution (Section 18.3.3).
A second type of problem faced by formulators is to predict solvents and solvent
combinations for a new resin. Solubility parameters may be useful, but in most cases,
qualitative application of the idea that like dissolves like is just as effective for finding
a starting point and takes less time. The problem of selecting solvents for new resins
has become easier because, to increase solids, lower MW resins are being used. A
wider range of solvents dissolves a resin as the resin’s MW gets lower.
18.3. SOLVENT EVAPORATION RATES
357
TABLE 18.4. Solvent Blends, Weight Percents
Component
MEK
MIBK
Xylene
Toluene
n-Butyl alcohol
2-Ethoxyethyl acetate
1-Methoxy-2-propyl acetate
Solubility parameter, da
dd
dp
dh
Relative evaporation rate
Costb
a
b
1
2
3
9.9
29.7
21.9
20.1
13.5
4.9
—
6.1
32.8
25.2
18.4
17.5
—
—
14.7
19.7
24.3
18.2
17.3
—
5.7
Properties
19
18
5.7
5.7
1.16
0.70
19
18
5.7
5.1
1.17
0.68
19
18
5.3
5.3
1.14
0.68
(MPa)1/2.
At the time of study, in dollars per kilogram.
18.3. SOLVENT EVAPORATION RATES
During application and film formation, the volatile material evaporates out of the coating.
The rate at which evaporation occurs affects not only the time required to convert a coating
to a dry film, but also the appearance and physical properties of the final film. Like so
many topics in the coatings field, evaporation rate appears simple but turns out to be
complex.
18.3.1. Evaporation of Single Solvents
The rate of evaporation of a solvent is affected by four variables: temperature, vapor
pressure, surface/volume ratio, and rate of air flow over the surface. The rate of evaporation of water is also affected by relative humidity.
The important temperature is the temperature at and near the surface. Although this
temperature may initially be that of the surrounding air, it decreases as solvent evaporates.
Whereas the surface is cooled by evaporation, it is warmed by thermal diffusion from
within the sample and its surroundings. Depending on the circumstances, thermal diffusion
may occur rapidly, so that surface temperature does not fall much during evaporation, or
slowly, resulting in a sharp drop in surface temperature. The cooling effect is largest for
those solvents and situations in which the solvent evaporates most rapidly. Also, the
higher the heat of vaporization of the solvent, the greater the temperature drop if all
other variables are equal.
Vapor pressure of the solvent at the temperature(s) at which evaporation occurs is the
important vapor pressure. A common error is to assume that boiling points [the temperatures at which solvents have a vapor pressure of 101.3 102 kPa (1 mmHg ¼ 0.1333 kPa;
1 atm ¼ 101.3 kPa) are directly related to vapor pressures at other temperatures. However,
boiling points are poor indicators. For example, benzene has a boiling point of 808C and
ethyl alcohol has a boiling point of 788C, but at 258C, their respective vapor pressures are
1.3 and 0.79 kPa. Consequently, benzene evaporates more rapidly than ethyl alcohol at
358
SOLVENTS
258C under the same conditions. Similarly, at 258C n-butyl acetate (bp 1268C) evaporates
more rapidly than n-butyl alcohol (bp 1188C).
The ratio of surface area to volume is important because solvent evaporation occurs at
the solvent –air interface. If 10 g of solvent is spread out over an area of 100 cm2, it evaporates more rapidly than if the surface area is 1 cm2. Accordingly, the concentration of the
resin solution and the viscosity increase more slowly during solvent evaporation from
thicker films having the same surface area. Solvent evaporates more slowly from a
coating in an open can than after application as a film. When a coating is applied by a
spray gun, it is atomized to small particles as it comes out of the orifice of the gun. The
ratio of surface area to volume is very high, as is the rate of solvent loss. Thus, a major
fraction of solvent evaporates after the spray droplets leave the orifice of the spray gun
and before they arrive at the surface to which the coating is being applied.
The rate of air flow over the surface is important because the rate of evaporation
depends on the partial pressure of the solvent vapor in the air at the air –solvent interface.
If vaporized solvent molecules are not carried away from the surface quickly, the partial
pressure of solvent builds up and evaporation is suppressed. Air flow rates vary substantially depending on the application method, and therefore the solvents used in a coating
must be selected for the particular application conditions. For example, spraying a
coating with an air spray gun results in significantly more loss of solvent than when the
coating is sprayed from an airless gun because of the greater air flow over the surface
of the droplets by the former method (Sections 23.2.1 and 23.2.2). The rate of solvent
loss from a freshly coated surface depends on the rate of air flow through the spray
booth. If the same coating is applied to the outside and the inside of a pipe, solvent
evaporates more rapidly from the outside unless the inside is ventilated. Air flow effects
cause nonuniform evaporation from coated objects; solvent evaporates more rapidly
near the edges of a coated panel than from its center.
Relative humidity has little effect on the evaporation rates of most solvents; however,
it has a major effect on the evaporation rate of water. The higher the relative humidity,
the more slowly water evaporates when all other conditions are equal. As relative humidity approaches 100%, the rate of evaporation of pure water approaches zero. The evaporation rate of water from solutions can be negative at high humidity. Relative humidity
decreases as air is warmed; sometimes it is feasible to compensate partially for the
effects of humidity on drying by modestly increasing the air temperature, since
warming the air both increases the vapor pressure of the water and decreases the relative
humidity.
18.3.2. Relative Evaporation Rates
As mentioned in Section 18.2, the introduction of nitrocellulose lacquers led to the need to
formulate complex solvent combinations. An important criterion involved in solvent
selection is evaporation rate. From the standpoint of solvent selection, vapor pressure is
an important variable that affects evaporation rate. However, it is difficult to look at
vapor pressure data and judge how much more rapidly one solvent will evaporate than
another. Procedures for measuring absolute evaporation rates were developed; and
because of the difficulties of controlling air and heat flows in different types of measuring
devices, relative evaporation rates were determined. Rates of solvent evaporation were
related to the evaporation rate of n-butyl acetate, which at the time was a standard
nitrocellulose solvent. Relative evaporation rate E is defined by Eq. 18.1, where t90 is
18.3. SOLVENT EVAPORATION RATES
359
the time for 90 wt% of a sample to evaporate in a given type of apparatus under controlled
conditions:
E¼
t90 (n-butyl acetate)
t90 (test sovlent)
(18:1)
Using Eq. 18.1, the relative evaporation rate of n-butyl acetate is 1 by definition.
Some authors express E as a percentage figure, corresponding to 100 for n-butyl
acetate. In either case, the higher the E value, the faster the evaporation rate of the
test solvent. One should be sure which reference point is being used; there is at least
one paper in the literature in which an author used data from two different tables
without realizing that he was mixing data that differed by a factor of 100; his conclusions
were absurd.
Determination of relative evaporation rates requires measurement under carefully
controlled standardized conditions. The Shell Thin Film Evaporometer is an example of
an apparatus. A top loading balance is placed in an enclosure designed to minimize variation in air flow over the surface of the balance pan. A flow of 258C air with a relative
humidity of less than 5% at a rate of 21 L min21 is maintained. A sample of 0.70 mL
solvent is dispensed onto a piece of filter paper on the balance pan, and the time required
to lose 90% of the sample weight is determined. Rates determined this way are volumebased relative evaporation rates. Although volume-based evaporation rates are the most
common in the literature, some experimental procedures use a fixed weight rather than
a fixed volume of solvent. Care must be taken not to mix data obtained by different
methods. One should also avoid comparing relative evaporation rates determined using
different instruments because different evaporation conditions can change not only the
absolute evaporation rates, but also the relative evaporation rates.
A study by Rocklin illustrates the effects of changes in conditions on relative evaporation rates [17]. He compared the relative evaporation rates of 66 solvents measured by the
standard procedure (evaporation from filter paper) with rates measured by evaporation
directly from the flat aluminum pan of the balance. Table 18.5 gives his data for
several solvents. Note that both sets of data are relative to n-butyl acetate. As can be
seen by comparing the Epaper/Emetal ratios, significant differences exist between relative
evaporation rates when the solvents evaporate from filter paper and the corresponding
relative rates when the same solvents evaporate from a smooth metal surface in the
same instrument at the same settings.
The most notable differences between Epaper and Emetal occur with faster evaporating
solvents and with water and alcohols. In the case of evaporation from filter paper, the
ratio of surface area to volume is much higher than in the case of the evaporation from
the metal. Therefore, at first solvents evaporate more rapidly from filter paper, resulting
in a sharp drop in temperature, which leads to a decrease in vapor pressure that slows
evaporation. In the case of evaporation from a metal surface, the surface area is much
smaller and the thermal conductivity is much higher. Both factors minimize the temperature drop, leading to less decrease in vapor pressure and evaporation rate. The difference is
greater with fast evaporating solvents such as n-pentane and acetone than with n-butyl
acetate, which evaporates more slowly, resulting in their low ratios of Epaper to Emetal.
The experimental temperature was 258C, but that is the air temperature. What controls
the rates of evaporation and hence the relative rates is the actual temperature(s)
of the surfaces.
360
SOLVENTS
TABLE 18.5. Volume-Based Relative Evaporation Rates at 2588 C
Solvent
Epaper
Emetal
Ratio
n-Pentane
Acetone
Ethyl acetate
Methyl ethyl ketone
n-Heptane
Toluene
Ethyl alcohol
Methyl isobutyl ketone
Isobutyl acetate
n-Butyl acetate
sec-Butyl alcohol
m-Xylene
n-Butyl alcohol
2-Ethoxyethanol
Water
Methyl n-amyl ketone
2-Ethoxyethyl acetate
n-Decane
2-Butoxyethanol
Isophorone
Diethylene glycol monoethyl ether
12
5.7
4.0
3.9
3.6
2.0
1.7
1.7
1.5
1
0.93
0.71
0.44
0.37
0.31
0.34
0.20
0.18
0.077
0.023
0.013
38
10
6.0
5.3
4.3
2.1
2.6
1.7
1.5
1
1.2
0.71
0.48
0.38
0.56
0.35
0.19
0.16
0.073
0.026
0.014
0.32
0.55
0.67
0.74
0.83
0.92
0.65
1.0
1.0
1
0.81
1.0
0.92
0.98
0.56
0.96
1.1
1.1
1.1
1.0
0.99
The Epaper/Emetal ratios are also low for water and alcohols. This results from the
greater extent of hydrogen-bond interaction of the hydroxyl groups with the very large
surface area of cellulose of the paper compared with the smooth aluminum surface.
This retards evaporation of water and alcohols relative to n-butyl acetate. Note that relative
to n-butyl acetate, sec-butyl alcohol evaporates more rapidly from the smooth metal
surface and more slowly from the filter paper.
When formulating baking coatings for spray application, it is common to use a mixture
of fast and very slow evaporating solvents. A significant fraction of the fast evaporating
solvent evaporates before the spray droplets reach the object being coated, raising viscosity and reducing the tendency of the coating to sag, while the slow evaporating
solvent keeps the viscosity low enough to promote leveling and to minimize the probability of popping when the coated object is put into a baking oven. (See Chapter 23
for a further discussion of sagging, leveling, and popping.) In selecting slow evaporating
solvents, formulators have generally used tables of relative evaporation rates at 258C and
boiling points, which are available for a large number of solvents. Evaporation rate data
have been published for only a limited number of slow evaporating solvents over the
range of 75 to 1508C [18]. The rates were determined using a thermogravimetric analyzer
(TGA) isothermally at a series of furnace temperatures. The rate relationships can be
quite different at different temperatures. For example, at 258C, the evaporation rate for
a commercial mixture of the dimethyl esters of succinic, glutaric, and adipic acids is
about five times slower than for isophorone. However, at 1508C, their evaporation rates
are approximately equal.
Which evaporation rate data are “correct”? All of them are; they depend on the particular circumstances under which they were determined. However, we do not apply coatings
to filter paper (or to aluminum balance pans or to TGA pans), nor do we dry them in a Shell
Evaporometer (or a TGA furnace). The rates at which solvents evaporate in actual use
18.3. SOLVENT EVAPORATION RATES
361
depend on the particular situation. Solvents evaporate more rapidly from coatings applied
by a spray gun than from coatings applied by roller coating. The type of spray gun can
make a substantial difference. Other significant variables include the rate of air flow
through the spray booth, the shape and mass of the object coated, the film thickness
applied, the flash-off time before entering an oven, and the way that heat is zoned in
the oven.
18.3.3. Evaporation of Mixed Solvents
Evaporation of solvent blends rather than pure solvents adds further complications. In
ideal homogeneous solutions, vapor pressure is governed by Raoult’s law, which predicts
that the vapor pressure Pi of the ith component of the solution is reduced from the vapor
pressure of the pure liquid P0i in proportion to its mole fraction xi:
Pi ¼ xi P0i
Since the vapor pressure of each solvent in the blend is different, the composition of the
solvents that evaporate is different from that of the solvent blend. Therefore, the partial
vapor pressures change continuously as solvent evaporates from a mixture. Raoult’s
law provides a good approximation for many combinations of solvents, especially when
structures are similar and intermolecular interactions are minimal. However, other
solvent mixtures are nonideal as a result of interaction effects. These effects change as
ratios of solvents change. The vapor pressure Ptotal of any mixture of miscible solvents
can be calculated by the following equation, where x represents an empirical adjustment
factor for the interaction effects, usually given the more dignified-sounding title of activity
coefficient:
Ptotal ¼ P1 þ P2 þ þ Pi ¼ x1 P01 x1 þ x2 P02 x2 þ þ xi P0i xi
Activity coefficients for many solvents have been evaluated and incorporated into computer programs that can calculate partial vapor pressures of each solvent in a mixture
throughout its evaporation. One such program is called UNIFAC [19]. Results of such
calculations are often presented as the partial vapor pressures at each 10% interval
through the evaporation. Several assumptions are involved in calculating vapor pressures,
and in order to relate vapor pressure to evaporation rate, it is assumed that all other factors
(temperature, surface/volume ratio, and air flow) affecting evaporation rate are fixed.
Thus, the results of such calculations are only approximations. However, in view of the
uncertainties of relative evaporation rates, a high level of accuracy is not needed. The
evaporation of solvent blends is discussed in an extensive review paper [20].
Relative evaporation rates of mixed solvent compositions can be determined experimentally in an evaporometer. The relative evaporation rate of a mixed solvent ET is
calculated from volume fractions c, activity coefficients a, and relative evaporation
rates E of the individual solvents [21]:
ET ¼ (caE )1 þ (caE )2 þ þ (caE )n
Such calculated ET values are of dubious accuracy: The composition changes over
time, so the value of ET changes. The difference between experimental and calculated
362
SOLVENTS
relative evaporation rates is especially large when comparing a solvent mixture with a
narrow range of E values with another with a wide range of E values. For example, one
could calculate ET for a mixture of two solvents with high and low E values that would
be equal to the E value of a single solvent with an intermediate E value. However, it is
evident that the mixed solvent would actually have a smaller relative evaporation rate
than the single solvent since after the fast evaporating solvent was gone, the slow
evaporating solvent would evaporate more slowly than the single intermediate E
value solvent.
This description makes it sound almost impossible to formulate a satisfactory solvent
mixture for a coating. However, formulation does not start in a vacuum of knowledge.
Experience with other formulations that proved useful under somewhat similar circumstances provides guidance. Using such formulations together with any reasonable table
of relative evaporation rates, one can make a first attempt at a new formulation. The
coating is then applied under the particular circumstances, and adjustments are made as
needed. The final adjustment is almost always made by an experienced person in the
user’s factory with production spray guns and operating conditions. The use of relative
evaporation rate tables assists development of a formulation that is in the ball park of
what is needed; because final in-factory adjustments are needed anyway, it may not
matter much which relative evaporation rate table was used.
Evaporation rates of water – organic solvent mixtures require special consideration for
at least four reasons. First, strong interactions often cause deviations from Raoult’s law.
Second, relative humidity (RH) affects the evaporation rate of water but has little or no
effect on the evaporation rates of organic solvents. Third, azeotropic effects may occur.
Fourth, the heat capacity and heat of vaporization of water are unusually high.
The humidity effect is illustrated by the fact that the relative evaporation rate (E) of
water at 0 to 5% RH and an air temperature of 258C is 0.31; but at 100% RH, E is 0. If
a solution of 2-butoxyethanol (E ¼ 0.077) in water evaporates at low RH, water evaporates more rapidly, and the remaining solution becomes enriched in 2-butoxyethanol.
At high RH, 2-butoxyethanol evaporates more rapidly and the remaining solution
becomes enriched in water. At some intermediate RH, the relative evaporation rates of
water and 2-butoxyethanol are equal. When an aqueous solution evaporates at this RH,
the composition of the remaining solution is constant. This RH has been called the critical
relative humidity (CRH) [22]. The CRH for 2-butoxyethanol solutions in water is estimated at about 80%. If the relative evaporation rate of the solvent were larger than that
of water, even at 0 to 5% RH, there would be no CRH, since the solvent would evaporate
more rapidly than water at any RH. At the other end of the scale, if the relative evaporation
rate of the solvent were very low, the CRH would approach 100%.
Azeotropic behavior is likely to occur with water solutions. Rocklin has studied the role
of azeotropy in speeding up water – solvent evaporation in humid air [23]. He reports
development of a computer model, the AQUEVAP program, that permits calculation of
the fastest evaporating water – solvent blends at various RH values. For example, at
40% RH, the time required for 90% evaporation of a 20 wt% solution of 2-butoxyethanol
in water is 1820 seconds compared with 2290 for water alone. Thus, co-evaporation of
water and 2-butoxyethanol accelerates evaporation.
The high heat capacity and heat of vaporization of water can also affect the evaporation
rates of water and water –solvent blends in an oven. For example, the times for 99% weight
loss of 2-butoxyethanol (bp 1718C), water, and a 26 : 74 blend of 2-butoxyethanol/water in
a TGA when room temperature samples were put into the furnace at 1508C were 2, 2.6, and
18.3. SOLVENT EVAPORATION RATES
363
2.5 minutes, respectively [24]. While the air temperature in the furnace was 1508C, the
samples took some time to heat up. The higher heat of vaporization of water (2260 J g21
at its boiling point) compared to 2-butoxyethanol (373 J g21 at its boiling point) slowed
the rate of heating of the water and water –solvent blend enough to more than offset the
expected evaporation rates based on boiling points or E values.
18.3.4. Evaporation of Solvents from Coating Films
Except in high-solids coatings (Section 18.3.5), the resin or other coating components
have little effect on the initial rate of solvent evaporation when coating films are
applied. Initial rates of evaporation of solvent from resin solutions are, within experimental error, the same as or close to the rates of evaporation of the solvents alone
under the same conditions. Such observations are not inconsistent with Raoult’s law,
which predicts that dissolved resins will have little effect on vapor pressure because
of their high MWs. However, as solvent loss from a coating continues, a stage is
reached at which the rate of evaporation slows sharply. As the viscosity of the remaining
coating increases, availability of free volume decreases and the rate of solvent loss
becomes dependent on the rate of diffusion of solvent through the film to the surface
rather than on the rate of evaporation from the surface. The solids level at which the transition from evaporation rate control to diffusion rate control occurs varies widely but is
often in the range 40 to 60 NVV.
Hansen described the situation as first stage and second stage losses of solvent [25]. In
the first stage, the rate is governed by the factors that govern evaporation of solvent
mixtures: vapor pressure, surface temperature, air flow over the surface, and surface/
volume ratio; at this stage, there is a first-power dependence on wet film thickness.
After a transitional stage, evaporation slows and the rate of solvent loss becomes
dependent on the rate of diffusion of the solvent molecules through the film. During
this stage, the evaporation rate depends on the square of film thickness. In line with the
changes in viscosity, the first stage is sometimes called the wet stage and the second
stage, the dry stage. An example of a plot of weight of solution remaining as a function
of time for evaporation of MIBK from a soluble vinyl chloride copolymer resin solution
is shown in Figure 18.1.
During the second stage, the diffusion rate is controlled primarily by free volume
availability. That is, the solvent molecules move through the film by jumping from free
volume hole to free volume hole. The most important factor controlling free volume availability is T 2 Tg. If solvent evaporation is occurring at a temperature well above the Tg of
the solvent-free resin, diffusion rate will not restrict the rate of solvent evaporation at any
stage in drying. If the Tg of the resin is above the temperature at which drying is occurring,
solvent evaporation will become diffusion rate controlled as T 2 Tg becomes small. No
experimental data are available as to the values of T 2 Tg when diffusion rate control
becomes important; such values are probably system dependent.
When the second stage is reached, the rate of evaporation becomes dependent on how
fast the solvent molecules get to the surface, not on their vapor pressure. As solvent loss
continues, concentration of the remaining resin solution increases, and therefore Tg
increases and the rate of diffusion slows further. If the Tg of the resin is sufficiently
higher than the temperature of the film, the rate of solvent loss will, in time, approach
zero. Years after films have been formed, there will still be residual solvent left in the
film. If solvent must be essentially completely removed in a reasonable time, the film
364
SOLVENTS
Figure 18.1. Two-stage release of solvents MIBK in Vinylite VYHH at 23 8C, initially at 20 wt% polymer.
(From Ref. [26], with permission.)
must be baked at a temperature above the Tg of the solvent-free resin. It is a common
mistake for novice polymer chemists to think that the rate of solvent loss from polymer
samples can be increased by drying under a vacuum at temperatures below Tg. Since
the rate of solvent loss has reached a stage at which it is independent of the vapor pressure
of the solvent, it is also independent of atmospheric pressure above the sample.
Solvent evaporation is affected by the addition of a plasticizer. The concentration of
solvent at which solvent loss becomes diffusion controlled decreases as plasticizer concentration increases. The amount of solvent retained after some time interval decreases as
plasticizer concentration increases. Perhaps, less obviously, in the case of mixed-solvent
systems, the ratio of slow to fast evaporating solvent retained in the film increases as
plasticizer concentration increases.
There have been many attempts to quantify the diffusion of solvents from coating films
during solvent evaporation, but success has been limited [27,28]. Some of the problems
are discussed in Ref. [29], and possible approaches to modeling the diffusion through
scaling analysis are suggested. As a broad generalization, small molecules tend to evaporate faster than large molecules in both the first and second stages because small molecules tend to be more volatile in the first stage and to diffuse faster in the second
stage. However, other factors are involved. One is molecular configuration. Since the
molecules must jump from hole to hole, the smaller the size of the solvent molecule,
the greater its chance of finding sufficiently large holes. Even though its relative evaporation rate is higher, cyclohexane is retained in films to a greater degree than toluene
because cyclohexane has a higher molar volume.
Figure 18.2 shows the results of an experiment illustrating the effect of cross-sectional
size. Nitrocellulose and acrylic lacquers were made at 20 wt% solids in 60 : 40 isobutyl
acetate (IBAc) and n-butyl acetate (BAc), and the ratios of the two solvents remaining
in the films were monitored. Like other branched-linear isomeric pairs, the relative
evaporation rate of branched IBAC is higher than that of linear BAC. Thus, the ratio of
IBAc to BAc remaining in the film decreased during the first, wet stage of solvent loss,
reaching a minimum of about 35 : 65. During the second, dry stage, the linear compound
18.3. SOLVENT EVAPORATION RATES
365
Figure 18.2. Changes in remaining solvent concentration during wet and dry stages of solvent evaporation at 238C from films of an acrylic resin (Elvacite 2013) and nitrocellulose, each initially in 60 : 40 IBAc
to Bac. (From Ref. [26], with permission.)
diffused more rapidly, so the ratio of IBAc to BAc remaining in the film increased and
approached 90 : 10 [26]. Equations have been developed that model the effect of solvent
size on diffusion based on the free volume of polymers [30].
In air dry coatings, Tg often increases to a level above ambient temperature, while significant amounts of solvent remain in the film. When this happens, solvent evaporation
becomes extremely slow, and detectable amounts of volatile solvents may remain in
the film for years. Residual solvent has been shown to affect adversely film properties
such as corrosion and moisture resistance (Section 7.3.1.1). Thus, selection of solvents
that have minimal detrimental effects may be a factor in optimizing long-term film
performance.
18.3.5. Evaporation of Solvents from High-Solids Coatings
Generally, it is more difficult to control sagging of spray applied high-solids coatings
than of their lower solids counterparts (Section 24.3). Although other factors may be
involved, there is considerable evidence that high-solids formulations lose much less
solvent than are lost by conventional coatings while atomized droplets are traveling
between a spray gun and the object being coated [31,32]. As a result, there is less increase
in viscosity and therefore a greater tendency to sag. Explanations for this decreased
solvent loss do not appear to be adequate. One factor is the colligative effect of the
lower mole fractions of solvent(s) in high-solids coating. A sample calculation illustrates
this difference. If the vehicle of a conventional coating consists of a 30 wt% solids solution
of a resin with an M̄n of 15,000 in 70% of a solvent mixture with an M̄n of 100, the
mole fraction of solvent is 0.997. However, if a high-solids coating has 70% solids of a
resin with an M̄n of 1000 and 30% of a solvent mixture with an M̄n of 100, the mole fraction of solvent is 0.811. Although this difference is in the direction of slowing solvent
evaporation from the high-solids coating, it does not seem large enough to account for
the large differences in solvent loss that have been reported.
It was suggested without experimental evidence that high-solids coatings may undergo
transition from first-stage to second-stage solvent loss with relatively little solvent loss
compared to that for conventional coatings [33]. The reasoning was that the Tg of the
366
SOLVENTS
solution in a high-solids coatings could change more rapidly with concentration, hence
reach a stage of free volume limitation of solvent loss after only a little loss of solvent.
Consistent with this hypothesis, Ellis found that high-solids polyester coatings were formulated at concentrations that were already above the transition concentration where
the solvent loss rate becomes diffusion controlled [34]. He also found that the transition
points occur at higher solids with linear molecules, such as n-octane versus isooctane,
and n-butyl acetate compared to isobutyl acetate. Further research is needed to fully understand the reasons for slower loss of solvents from high-solids coatings; meanwhile, slower
solvent loss can be a serious problem for the application of high-solids coatings.
More detailed discussion of sagging and sag control in high-solids coatings appears in
Sections 24.3 and 25.2.2.
18.3.6. Volatile Loss from Waterborne Coatings
For solventborne coatings, the formulator has available a variety of solvents having a wide
range of properties and volatility. But for waterborne coatings, the major volatile
component is water, which comes with only one vapor pressure – temperature curve,
one heat of vaporization, and so on. Furthermore, the relative humidity at the time of
application and drying of the coatings can have a major effect on rates of volatile losses
from waterborne coatings. Formulators use limited levels of organic solvents to modify
evaporation rates; however, future regulations can be expected to reduce the levels
permitted.
A major class of waterborne coatings is based on water-reducible resins, which are not
soluble in water (Section 8.3). A salt of the resin, dissolved in an alcohol or an ether
alcohol, is diluted with water. During dilution, the resin forms aggregates that are
swollen with solvent and water, dispersed in a continuous aqueous phase that also contains
some solvent. The relative evaporation rate of water is affected by the relative humidity; so
with solvents such as 2-butoxyethanol, there is a critical relative humidity (CRH) (Section
18.3.3). It has been found that the CRH is different for coatings than for water – solvent
blends without resin. For example, CRH is 65% for 10.6 wt% (based on volatile
components) 2-butoxyethanol in a coating, in contrast to 80% calculated for the same concentration of a 2-butoxyethanol –water blend without resin [35]. A possible explanation of
the difference in CRH is that the concentration of 2-butoxyethanol in the continuous water
phase is below the average and the solvent –water ratio inside the aggregates is above the
average. This type of distribution and its effect on the composition of vapor above a model
aggregate system using a more volatile solvent, t-butyl alcohol, has been demonstrated
[36]. Such effects can be critical in controlling sagging and, particularly, popping of
such waterborne coatings (Sections 23.3, 23.7, and 26.1).
The other major class of waterborne coatings uses latexes as the principal binder.
Evaporation of water from a drying latex paint film resembles first-stage drying throughout
most of the process; it is controlled by temperature, humidity, evaporative cooling, and
rate of air flow over the surface [37]. After most of the water has left, evaporation
slows as a result of coalescence of a surface layer through which water must diffuse. In
latex paints to be applied by brush or roller, it is desirable to retard the development of
a partially coalesced surface layer to permit lapping of wet paint on wet paint. This generally requires the presence of a slow evaporating solvent such as propylene glycol.
The presence of these solvents does not affect the initial rate of water loss, but does
slow the development of a surface skin [38]. The presence of such a water-soluble
18.4. VISCOSITY EFFECTS
367
solvent also facilitates the loss of coalescing solvent. Use of both types of solvents
is discussed further in Chapter 31.
18.4. VISCOSITY EFFECTS
Solvent selection can have a major effect on the viscosity of resin solutions. This results
from two factors: the viscosity of the solvent itself and the effect of solvent – resin interactions. The direct effect of solvent viscosity on solution viscosity can be seen in equations
that relate viscosity and resin concentration. Even the simplest relationship, which is valid
over only a narrow range of concentration of solutions with viscosities in the approximate
range 0.1 to 10 Pa.s, shows this dependence on solvent viscosity:
ln hsolution ¼ ln hsolvent þ K(conc)
For example, a difference of viscosity of 0.2 m Pa.s between the viscosities of two solvents (1.0 and 1.2 m Pa.s) may appear trivial but can cause a difference of 2000 m Pa.s
in the viscosity of a 50 wt% solution of a resin (10 Pa.s versus 12 Pa.s) [39]. The relationship between concentration and viscosity of resin solutions is discussed in more detail in
Section 3.4.3.
One must be careful when comparing viscosities of solutions. There are data in the
literature that make comparisons based on weight relationships, volume relationships,
weight of solvent per volume of coating, and weight of resin per volume of solution.
One must decide what type of comparison is appropriate for a particular purpose. Most
commonly, weight ratios are used when considering the viscosity of resin solutions, but
they can be misleading because air pollution regulations are based on weight of solvent
per volume of coating. If one compares the viscosity of resin solutions with different
solvents on a weight basis, one might make the wrong choice in trying to minimize the
weight of solvent per volume of coating. Since flow is generally related to volume
considerations, use of volume fraction comparisons might seem desirable, but polymer
solutions are seldom ideal, creating uncertainty about the significance of volume fraction
numbers.
An example of the effect of solvent choice on viscosity of an acrylic solution containing
400 g L21 of solution of the various solvents is given in Table 18.6. For comparison purposes, the densities and viscosities of the solvents alone are also given. The data in
Table 18.6 illustrate the substantial effect that solvent selection can have on the solution
viscosity of a given polymer. The weight concentrations of these solutions are not constant; it would have been interesting to compare also the viscosities of equal-weight-ratio
solutions. The data suggest that solvents with low density and low viscosity tend to give
low solution viscosity, as predicted by theory. However, other factors are involved.
Another factor that is sometimes important is the effect of solvent– solvent interactions
on the viscosity of solvent mixtures. In general, these effects are small except for mixtures
that include alcohols or water. When relatively small (generally, ,40%) amounts of
alcohols are mixed with other solvents, they do not increase the viscosity of the solution
proportionally to their relatively high neat viscosities [41]. This results from reduced
multiple hydrogen-bond interactions between the alcohol molecules in the presence of
other solvents. The effects are larger and less obviously predictable in the case of solutions
of solvents in water.
368
SOLVENTS
TABLE 18.6. Viscosities at 2588 C of Solutions of a High-Solids Acrylic; Solvent
Concentration 400 g L21 Solution
Solvent
Methyl propyl ketone
Methyl isobutyl ketone
Ethyl acetate
Methyl amyl ketone
n-Butyl acetate
Toluene
Isobutyl isobutyrate
Xylene
Solvent Viscosity
(m Pa.s)
Solvent Density
(g ML21)
Solution Viscosity
(m Pa.s)
0.68
0.55
0.46
0.77
0.71
0.55
0.83
0.66
0.805
0.802
0.894
0.814
0.883
0.877
0.851
0.877
80
110
121
147
202
290
367
367
Source: Ref. [40].
Large viscosity effects can result from solvent – resin interactions. To distinguish these
effects from the effects of solvent viscosity discussed previously, it is common to compare
relative viscosities (h/hs) of solutions. The effects of interactions are complex and not
fully understood in even relatively simple systems. At least two factors play a role.
First, most resins used in coatings have polar and hydrogen-bonding substituents, such
as hydroxyl groups or carboxylic acid groups, that tend to associate with polar groups
on other molecules, often substantially increasing viscosity. Solvents that can prevent
or minimize such interactions by interacting with the polar groups tend to reduce viscosity.
Polar solvents with single hydrogen-bond acceptor sites are effective in minimizing
relative viscosity; examples are ketones, ethers, and esters. Water can also serve to
reduce the viscosity of oligomers with hydrogen-bond interactions. It has been shown
that up to 20% (depending on the formulation) of water can dissolve in solvent-free coatings and that adding water, usually 5 to 15%, reduces viscosity by 40 to 60% [42].
A second factor is the effect of a given solvent on the hydrodynamic volume of the individual resin molecules with their closely associated solvent molecules. If the interaction
between resin and solvent molecules is strong, the chains of the resin molecules
become extended and their hydrodynamic volume increases. If the interaction is not so
strong, the molecules contract and the hydrodynamic volume becomes smaller. Relative
viscosity tends to be directly related to hydrodynamic volume. However, if solvent –
resin interaction is so weak that there can be resin –resin interaction rather than only
resin – solvent interaction, clusters of resin molecules form and relative viscosity and
viscosity increase. In the majority of coating applications, especially high-solids coatings,
which use low-MW resins, resin –resin interactions are the stronger of the two effects,
although hydrodynamic volume effects can be significant. Thus, most high-solids formulations include hydrogen-bond acceptor solvents. The situation with mixed solvents is
more complex. We are aware of only one publication on viscosity effects with solvent
mixtures. In a study involving a limited number of systems, Erickson and Garner [43]
found that relative viscosity was dominated by the effect of the solvent that interacted
most strongly with the resin.
A word of caution: Relative viscosities can be useful for understanding the factors
affecting viscosity, but the viscosities important in applying coatings are not relative
viscosities. A solvent giving a comparatively high viscosity solution could give a relatively low relative viscosity solution. The viscosity of resin solutions is discussed
further in Section 3.4.
18.5. FLAMMABILITY
369
Solvent selection can also affect storage stability. Inclusion of alcohols in formulations
containing melamine–formaldehyde resins can increase storage life by minimizing
cross-linking during storage (Section 11.3.3). Polyesters can transesterify with primary alcohols during storage; this can be a problem especially with water-reducible polyesters (Section
10.5). Alcohols can react slowly with epoxy resins during storage. Hydrogen-bond acceptor
solvents can extend the pot life of epoxy–amine coatings (Section 13.2.5). Reaction rates of
isocyanates with hydroxyl groups are reduced in H-bond acceptor solvents, extending pot life
(Section 12.4 and Ref. [44]). Urethane-grade esters and ketones that contain no more than
traces of alcohols or water are available.
18.5. FLAMMABILITY
Fires and explosions of flammable coating solvents can inflict terrible burns. Many tragic
accidents have occurred in which people were killed or severely burned. Sadly, these
accidents could have been prevented. Most solvents used in coatings are flammable.
Care should be exercised when working with solvents in the laboratory, in the coatings
factory, and in end use environments. Flammability depends on structure and vapor
pressure. Generally, there is both an upper and a lower level of vapor concentration that
limit flammability or explosion. If the partial pressure is low enough, not enough
energy is released during burning to maintain the vapor – air mixture above the ignition
temperature of the system. If the partial pressure of the solvent is high enough, there is
insufficient oxygen for explosion or fire. Tables of upper and lower explosive limits of
many solvents are available. A full container of solvent may present less of a fire risk
than a recently emptied container; the vapor phase in the former might have the solvent
at concentrations above the upper explosive limit, whereas the “empty” container may
have a concentration in the explosive range.
The most common cause for fires in coating factories has been static electricity. Solvent
flowing out of one tank and into another tank by gravity can pick up enough electrostatic
charge to cause a spark, which can set off a fire or explosion. To avoid such charge
accumulation, all tanks, pipes, and so on, used in handling solvents and solvent-containing
mixtures should be grounded electrically at all times. A second cause is sparking of
electric motors or faulty electrical connections. Factories and laboratories should be
equipped with explosion proof electrical installations. One should never try to bypass
these installations. Also, of course, smoking is a potential ignition source.
There are two main types of flammability tests: open cup and closed cup. Both measure
a flash point, the minimum temperature at which solvent can be ignited by a hot wire.
ASTM specifies standard conditions for both tests. Generally, open cup testers give
results more appropriate for indicating the degree of hazard of a mixture when exposed
to air, as during a spill. The closed-cup flash point more nearly describes the fire hazard
of a liquid enclosed in a container. U.S. Department of Transportation regulations for
shipment of flammable liquids are based on closed cup tests. Transportation costs can
be affected substantially by the flash points of the material being shipped. A closed cup
flash point is lower than an open cup flash point. Solvents used in architectural paints
should have closed cup flash points over 388C. Extensive tables of flash points of solvents
are available. Different sources sometimes disagree on flash point values. Variations are
not surprising for solvents such as naphthas and mineral spirits that are variable mixtures,
but variations are also found with single component solvents such as n-butyl acetate.
370
SOLVENTS
The ASTM methods are said to be accurate to +2.58C. Reproducibility is poorer at
temperatures below 08C.
Many coatings contain a mixture of solvents. It is safest to determine flash points of
mixtures experimentally. The factors affecting flash points, including molecular interactions
in blends, have been discussed by Ellis [45]. It has been reported that good predictions of
closed cup flash points for mixed solvents can be made by a UNIFAC computer program
that requires only flash points and molecular structures of the pure components [46]. The
best results were reported using UNIFAC group interaction parameters derived from flash
points of binary solvent mixtures. Calculation of useful approximations of flash points of
mixed solvents with simpler equations has been reported [47].
Risk of fire or explosion can be eliminated by meeting either of two conditions. If
solvent vapor concentrations in air are kept well outside the concentration range within
which ignition is possible or if all sources of ignition are eliminated, there can be no
fire. Unfortunately, many accidents have occurred because neither of these conditions
were satisfied. Because it is difficult to be certain that either of these conditions can be
met in all circumstances, prudent practice dictates that all possible steps should be
taken to meet both conditions, providing redundancy. The importance of good ventilation
cannot be overemphasized, especially since the vapors of solvents, especially relatively
dense ones, may stratify in stagnant air. As a result, solvent vapor concentrations can
be higher in the lower part of the workspace, perhaps high enough to explode even
when the average concentration is below the lower explosive limit.
18.6. OTHER PHYSICAL PROPERTIES
Density can be an important variable. It can have a major effect on cost. Most solvents are
sold on a weight basis, but the critical factor and, correspondingly, the critical cost in
almost all cases in the coatings field is the cost per unit volume. Most U.S. air pollution
regulations are based on weight of solvent per unit volume of coating, which also
favors the use of low density solvents in formulations (Section 18.9).
Conductivity can affect solvent choice. As discussed in Section 23.2.3, the use of
electrostatic spray guns requires control of the conductivity of the coating. In general,
formulations that have appreciable but low conductivity work best. The conductivity of
hydrocarbon solvents is too low to permit pickup of adequate electrostatic charge.
Alcohols, nitroparaffins, and small amounts of amines are common solvents or additives
to increase conductivity to a desired range. The conductivity of waterborne coatings can
pose problems for electrostatic application, such as the need to insulate the spray apparatus
and relatively fast loss of charge from spray droplets. Additives can minimize the latter
problem. For example, inclusion of glycol ethers improves the sprayability of waterborne
coatings, apparently by reducing the surface conductivity of water. Presumably, the alkyl
groups of the ethers orient quickly to the surface of the droplets.
Surface tension can be another important factor influencing solvent selection (Section
24.1). Solvent can affect the surface tension of coatings, which, in turn, can have important
effects on the flow behavior of coatings during application, as discussed at length in
Chapter 24. Solvent selection can also be a factor affecting the development of surface
tension differentials across the surface of a drying film during application and film formation; these differentials can affect flow behavior [48,49]. Since surface tensions
depend on the temperature and concentration of resins in solution, solvent volatility can
18.7. TOXIC HAZARDS
371
have a large effect on the development of surface tension differentials. For a coating to wet
a substrate, the surface tension of the coating must be lower than that of the substrate.
Solvent selection can be a factor, particularly in aqueous systems. Although surface
tension of aqueous systems can be reduced by adding surfactant, it is often more desirable
to accomplish the same purpose with a solvent such as 2-butoxyethanol. Solvent
evaporates, whereas surfactant leaves residues in the final film that may be detrimental
to properties such as adhesion and humidity resistance.
18.7. TOXIC HAZARDS
In considering the toxic risks of volatile solvents, the extent of exposure must be considered in combination with toxicity data. All solvents are toxic at some level of exposure.
Obviously, one should avoid ingesting them. The hazard from skin contact can be
controlled by wearing protective clothing. Generally, the greatest potential risk comes
from inhalation.
Three general types of toxicity data are important. Acute toxicity indicates the level of
intake in single doses that can be injurious or lethal, information that can be particularly
important in cases of accidental ingestion or spills. A second type of toxicity data
concerns the level of exposure that is safe when people are going to be exposed 8 hours
a day for long periods of time. This type of data is used, for example, to set the upper
concentration limit for solvent in a spray booth. A third type is risk of exposure over
periods of years to low levels of materials, which can increase health risks such as
cancer. When it is found, usually by animal tests, that a solvent may be carcinogenic,
very low levels of permissible exposure are set. The levels are frequently too low to be
controlled by economically feasible methods. The effect is that carcinogenic solvents
are banned. For example, benzene has not been used in coatings for many years for this
reason. Although most solvents used in large quantities in coatings have been tested
and are thought to be noncarcinogenic, it is prudent for the user to be aware of current
knowledge of the materials that he or she is using and to minimize inhalation and
contact with all solvents.
Extensive tables of all three types of data are available. In formulating coatings it is
also necessary to consider the clientele who will use them. Although coatings sold to
retail consumers are carefully labeled to include application cautions, one must assume
that many people will not read the labels. When selling to a large corporation, it is reasonable to assume that the Material Safety Data Sheets will be read and appropriate practices
will be established. But when selling to small industrial customers such as automobile
repair shops, one cannot assume that the customer will pay attention to precautions. It
may be ethical to sell a coating with a somewhat toxic solvent to one class of customer
but unethical (although perhaps legal) to sell it to another class of customer. A common
difficulty is to know what the level of exposure will be. Reference [50] describes an
approach to assessing possible exposures when retail consumers apply coatings in a room.
In 1990 the U.S. Congress listed hazardous air pollutants (HAP) for which use is to be
reduced [51]. Among those of importance in the coatings field are methyl ethyl ketone
(MEK), methyl isobutyl ketone (MIBK), n-hexane, toluene, xylenes, methyl alcohol,
ethylene glycol, and ethers of ethylene glycol. The U.S. EPA Hazardous Air Pollutants
Strategic Implementation Plan describes regulatory efforts [52]. The first step was a
voluntary program aimed at reducing emissions of 17 chemicals, including methanol,
372
SOLVENTS
MEK, MIBK, isophorone, toluene, and xylene, by 50% (of 1988 levels) by 1995. Mandatory HAP limits are included in EPA’s Unified Air Toxics Regulations, issued for all major
categories of coatings users in 1995 – 1999; for an example, see Ref. [53]. Compliance is
required within three years of the issue date. Mono- and diethers of ethylene glycol,
diethylene glycol, and triethylene glycol are listed on the HAP list, but ethers of propylene
glycol are not. t-Butyl acetate and 2-butoxyethanol (monobutyl ether of ethylene glycol)
have been removed from the HAP list. The present HAP list motivates replacement of
listed solvents with solvents not on the HAP list. Such substitutions may do little or
nothing to reduce hazards or to improve air quality if the unlisted solvent is as hazardous
as the listed one.
18.8. ATMOSPHERIC PHOTOCHEMICAL EFFECTS
Since the 1950s, it has been realized that the presence of organic compounds in the atmosphere can lead to serious air pollution problems. Terminology can be confusing; in
the older literature, such compounds are referred to as “hydrocarbons,” meaning any
organic compound, not just unsubstituted hydrocarbons. More recently, they have been
termed volatile organic compounds (VOC). The terms reactive organic gases (ROG)
and reactive organic compounds (ROC) are also used. Particularly in Europe, solvents
are classified by their photochemical ozone creation potential (POCP). Three end
effects of VOC emissions into the atmosphere are important: formation of eye irritants,
formation of particulates, and formation of toxic oxidants, especially ozone.
Although all of these factors are important, the most critical for coatings is ozone.
Ozone is a naturally occurring component of the atmosphere but is toxic to plants and
animals. When you go up in the pine forests in the Rocky Mountains and smell the
wonderful “fresh air,” the odor is ozone. Pine and oak trees emit substantial quantities
of VOC into the atmosphere, the UV level increases with altitude, and as a result,
ozone generation is high. Plants and animals, including humans, evolved in the presence
of some ozone and can tolerate its presence up to a point. However, with the rapid growth
of VOC emissions from human-made sources, ozone levels on many days of the year in
many parts of the world, especially in and around cities, have exceeded the levels that
many plants can withstand and increased risks to human health, especially of susceptible
individuals.
The EPA reports that man-made VOC emissions in the United States peaked in 1970 at
about 33.7 million short tons/year and have been declining fairly steadily ever since,
reaching about 16 million short tons/year by 2004 [54]. This .50% reduction is a
major accomplishment, especially during a period when the U.S. GDP increased 187%
and vehicle miles traveled increased 171%. Since 1970, coatings emissions decreased
roughly in proportion, accounting for 11 to 14% of total manmade emissions throughout
the period. For perspective, highway vehicles emitted about 5.9 million short tons in 1998
and 4 million short tons in 2003. In recent years, emissions from coatings have been on the
order of 2 million short tons/year, enough to be an important target for future regulation.
The coatings industry has done a lot to reduce VOC emissions, but it will be asked to do
a lot more.
There is not general agreement on the level of ozone that is considered safe. From 1978
to 1997, the U.S. National Ambient Air Quality Standard for ozone allowed no more than a
1-hour period per year (averaged over three years) when the concentration of ozone in the
18.8. ATMOSPHERIC PHOTOCHEMICAL EFFECTS
373
air exceeded 0.12 ppm. Many urban areas in the United States have had trouble meeting
this standard. In 1997, the EPA reduced the standard to 0.08 ppm. For this standard to be
attained, the three-year average of the fourth-highest daily maximum 8-hour average of
continuous ambient air monitoring data over each year must not exceed 0.08 ppm [55].
It is estimated that 47 of the 50 states have some areas that do not comply with the new
standard. (See Ref. [55] for further information on ozone regulations.)
Photochemical reactions in the atmosphere are complex and depend on many variables
in addition to the amount and structure of VOCs, especially the concentrations of various
nitrogen oxides. Full discussion of the reactions is beyond the scope of this book.
References [56 – 58] provide brief and more detailed reviews on the state of the science.
Probably, the principal pathways leading to the generation of ozone are by way of hydrogen abstraction from the VOC compounds. Some of the reactions that have been proposed
to explain ozone generation are shown in Scheme 18.1, in which RH represents a VOC
compound.
Note that nitrogen oxides (NOx) participate in the ozone-forming reaction sequence
shown in Scheme 18.1. A 1991 study by the National Research Council concluded that
in many areas of the United States, substantial reductions of ozone can be accomplished
only by reducing NOx levels in addition to or instead of reducing VOC [59]. Coatings emit
almost no NOx; the main sources are electricity generation and transportation.
An important variable in the rate of ozone generated from VOCs is the ease of abstraction of hydrogens from reactive organic gases by free radicals, such as the hydroxyl
free radicals shown in the first equation of Scheme 18.1. Data are available on the rate
constants for a wide variety of organic compounds [58]. In general terms, compounds
with hydrogens on carbon atoms alpha to amines or ethers, hydrogens on tertiary
carbon atoms, allylic hydrogens, and benzylic hydrogens are examples of easily abstractable hydrogens (Section 5.1). POCP values are determined by rates of reaction of hydroxy
free radicals with the various solvents.
Scheme 18.1
In early investigations of the effects of VOC on air pollution, organic compounds were
divided by rabbit eye irritation tests into photochemically active compounds of high and
low reactivity. It was proposed that if the emission of the highly reactive compounds could
be limited, the less reactive ones could dissipate and avoid high local concentration of pollutants. This led to the establishment of a definition of photochemically reactive solvents in
Rule 66 of the Los Angeles Air Control District. After some years of experience, it was
realized that most organic compounds are photoreactive and that the extent of dissipation
in the atmosphere after local emission had been overestimated. Changing from highly
reactive compounds to less reactive compounds might diminish the air pollution effects
374
SOLVENTS
near the scene of emission in exchange for increasing air pollution down-wind from the
emission site. Furthermore, the coatings industry objected to having to use different
solvent combinations in different parts of the country. This situation led to the conclusion
that it would be best to limit the emission of almost all organic compounds into the
atmosphere.
In establishing future regulations, there is a difference of opinion as to whether all
solvents should be considered as equally undesirable in the atmosphere, as they are
now. The present approach is simpler to enforce. However, it may well be that using
less reactive solvents to replace more reactive ones would be advantageous by allowing
at least some opportunity for dissipation in the atmosphere to minimize the probability
of local excess ozone concentrations. In Europe, some regulations are based on the photochemical ozone creation potential (POCP) of individual solvents. A list of POCP values
and examples of reformulation of solvents to minimize POCP emissions may be found
in Ref. [58]. The POCP values have been calculated by computer modeling, using reaction
kinetic data to assess changes in ozone formation and checked in smog chambers [56].
Values for about 200 VOCs are available. Xylene and related aromatic hydrocarbons
have particularly high POCP values. There are no exempt solvents.
California has established regulations covering solvents used in aerosol spray cans
based on limiting the permitted level of solvents based on the grams of ozone generated
in the atmosphere per gram of product. The levels permitted vary based on the product
in the aerosol can as well as the individual solvents contained in the formulation.
(See Ref. [60] for the California regulation and tables of products and reactivity.) In
this new system all solvents are classified as reactive organic compounds (ROC), including those that have been established as exempt from VOC regulations. A book describing
the derivation of regulations for aerosol coatings has been published by California Air
Resources Board (CARB), it is available on the Web site of the American Solvents
Council [61]. CARB states that this new approach to regulation replacing VOC limits
with reactivity-based limits can achieve equivalent air quality while providing manufacturers with more flexibility and that it may result in a more effective and cost efficient
control strategy.
It was found that the effect of hydroxy free radicals on ozone production used in
POCP depended on the ratio of NOx to ROC. Therefore, a new system of rating the effect
of solvents on ozone was developed. In a smog chamber filled with a base reactive
organic gas mixture and NOx appropriate to the area for which regulations are to be
applied, a reactive solvent was added and the incremental reactivity on ozone generation
is calculated by dividing the maximum amount of ozone generated by the addition by the
amount of compound added. Table 18.7 gives the maximum incremental reactivity
values (MIR) for some of the solvents tested. (See Section 18.9.2 for a discussion of
aerosol regulations.)
An ingenious approach to VOC reduction is use of supercritical carbon dioxide as a
component in a solvent mixture [62], detailed in Section 23.2.5. A formula used to calculate VOC is given below at the beginning of Section 18.9.1 and is discussed further there.
The units used in most U.S. regulations are weight of solvent per unit volume of coating as
applied, excluding water and exempt solvents.
VOC emissions can be affected substantially by the transfer efficiency in spray applied
coatings. When a coating is sprayed, only a part of the coating is actually applied to the
object being coated. Transfer efficiency represents the percentage of the coating used
that is actually applied to the product. As the transfer efficiency increases, the VOC
18.8. ATMOSPHERIC PHOTOCHEMICAL EFFECTS
375
TABLE 18.7. MIR Values of Selected Solvents,
m(gozone/gsolvent)
Methyl acetate
PCBTF
t-Butyl acetate
Acetone
n-Butyl acetate
Odorless mineral spirits
2-Butanol
1-Methyl-2-propyl acetate
VM&P Naphtha
Toluene
Methyl isobutyl ketone
Mixed xylenes
0.07
0.11
0.22
0.43
0.89
0.91
1.60
1.71
2.03
3.97
4.31
7.37
emissions decrease, since less coating is used. Transfer efficiency depends on many
variables, particularly the type of spray equipment utilized (Section 23.2). In some
cases, regulations have been established requiring certain kinds of spray equipment or
setting a lower limit on transfer efficiency, such as 65%.
Besides reducing the VOC content of coatings and increasing transfer efficiency, there
are two other broad approaches to minimizing VOC emissions: solvent recovery and
incineration. In some cases it is feasible to recover the solvent used in coatings.
One can pass the air flowing out of a drying chamber through activated carbon beds
where the solvent is adsorbed. After the surface of the carbon is saturated with
solvent, the material is heated to distill and recover the solvent. Alternatively, the
solvent-laden air can be passed through condensers cooled with liquid nitrogen; the nitrogen that is vaporized is fed into the drying chamber, reducing the oxygen content and,
hence, permitting higher concentrations of solvent without exceeding the lower explosive
limit. Solvent recovery is desirable when it is feasible, but feasibility is limited by low
solvent concentration in the air stream needed to stay safely below the lower explosive
limits. In the case of effluent air from spray booths, it must be below concentrations
that could have toxic hazards; the solvent concentration is generally too low to permit
economic recovery of the solvent.
Minimizing VOC emissions by incineration involves heating the effluent solvent
laden air stream in the presence of a catalyst to a temperature high enough to burn the
solvent. As with solvent recovery, this approach is economically feasible only when
solvent concentrations are relatively high. Incineration has been found to be particularly
applicable in coil coating (Section 30.4). In this case, most of the solvent is released in
the baking oven. Part of the effluent air from the baking oven is recirculated back into
the oven; the amount of air recirculation is limited so that the solvent content does not
approach the lower explosive limit. The balance of the effluent air is fed to the gas
burners that heat the oven. The solvent in the air is burned along with the gas; the fuel
value of the solvent reduces the gas requirement. Thus, VOC emission is minimized
and the fuel value of the solvent is used.
Between 1970 and 1995, VOC emissions from typical auto assembly plants
were reduced by 70 to 80% by the combination of reduced VOCs, improved transfer
efficiency, and incineration [63]. Further reductions have probably been achieved
since then.
376
SOLVENTS
18.9. REGULATION OF SOLVENT EMISSIONS FROM COATINGS
In the United States, the EPA establishes regulations regarding emission of VOC and HAP
solvents into the atmosphere. States or other subdivisions are allowed to have more restrictive regulations than the national regulations. The European Union has also established
regulations.
18.9.1. Determination of VOC
VOC values at for waterborne or solventborne coatings are calculated by the following
formula:
VOC ¼
(wt% of all volatiles wt% of water wt% of exempt solvent)(density of coating)
liters of coating liters of water liters of exempt solvents
This definition of VOC as weight/volume follows from the conventions that air pollution data are based on weights of pollutants and that coatings are sold and applied on
a volume basis. Regulatory limitations are based on VOC content of the coating “as
applied.” If, as is very common, the user is to add solvent to the coating after receiving
it from the coating manufacturer, the VOC is based on the coating after this solvent has
been added. The exclusion of water is to prevent circumventing regulations by developing
coatings with low solids that are diluted with water but still have high VOC contents
relative to the dry film thickness that has to be applied.
The amount of VOC emitted by a coating is not easily determined. In most cases, only
approximations of potential VOC emissions can be calculated, even when the exact formulation of a coating is known. Solvent can be retained in films for very long periods
of time. In latex paints, coalescing solvents are used that are only slowly released from
the coating. In cross-linking coatings, volatile by-products may be generated by the reaction. For example, MF cross-linking leads to the evolution of a molecule of volatile
alcohol for each co-condensation reaction, and in self-condensation reactions, there can
be emission of alcohol, formaldehyde, and methylal. The amount released depends on
curing conditions and the amount of catalyst used. On the other hand, when slow
evaporating glycol ether solvents are used in an MF cross-linking system, it is probable
that some of the glycol ether transetherifies with the MF resin and is not emitted from
the film. Amines used in “solubilizing” water-reducible coatings may volatilize to
different extents, depending on conditions and amine structure [64,65]. With high-solids
two package coatings, the amount of volatile material can be affected by many variables,
most obviously by the time interval between mixing the two packages and application.
Very high solids coatings use low MW oligomers; particularly when baked, some of
the oligomer may volatilize.
Experimental measurement of VOC is not straightforward, either. The amount of VOC
released depends on conditions under which the coating is used. Time, temperature, film
thickness, air flow over the surface, and in some cases, the amount of catalyst are among
the variables that affect the results obtained. Although it would seem that determination
under the conditions of actual use would be most appropriate, this is not easy to do. For
air dry coatings, the time required for the determination would be very long. For baking
coatings, there can be differences in conditions for use of the same coating. It is generally
agreed that it would be desirable to have a standard method for determining VOC
18.9. REGULATION OF SOLVENT EMISSIONS FROM COATINGS
377
emissions. However, there is little agreement as to what that standard method should be; in
view of the effect of application variables, it is doubtful if a single appropriate standard
method will ever be developed.
Methods for determination of VOC and other useful information were collected by
Brezinski in 1993 [66]. Improved methods are needed. For example, a round robin
study of ASTM D-2369-89 (the standard method for measuring volatile content) gave
results 5 to 34% higher than expected, with large standard deviations for two component
coatings; results with one component coatings correlated better [67]. An updated ASTM
method, D-2369-95, uses similar procedures; its reproducibility ranges from +3.4 to
5.3%, with little difference between one- and two-component coatings. The standard
method for determining VOCs is ASTM D-3960-96; in a round robin test of solventborne
automotive coatings, its reproducibility was 2.96%.
Measuring VOC of waterborne coatings is made more complex by the need to determine water content, and water analysis methods are surprisingly poor. A standard gas
chromatography method, ASTM D-3792-99, is only reproducible to 5%. A rather complicated Karl Fischer procedure (ASTM D-4017-02) is only reproducible to 5.5%, and this
was an improvement over earlier versions of the procedure. A round robin test of
waterborne automotive coatings by ASTM D-3960-96 had a reproducibility of only
9.75%, with most of the poor reproducibility attributable to the water analysis methods.
Note that in the formula above, as the ratio of water to regulated solvent increases,
errors in water analysis are magnified in calculation of VOC.
Current U.S. EPA regulations treat most solvents useful in coatings [except water,
acetone, t-butyl acetate, CO2, 4-chlorobenzotrifluoride (PCBTF), methyl acetate, perchloroethylene, and certain silicone fluids and fluorinated solvents] as equally undesirable.
In establishing future regulations, there is a difference of opinion as to whether all solvents
should be considered as equally undesirable in the atmosphere, as they are now. The
present approach is simpler to enforce. However, it may well be that using less reactive
solvents to replace more reactive ones would be advantageous by allowing at least
some opportunity for dissipation in the atmosphere to minimize the probability of local
excess ozone concentrations. CARB has begun to depart from the present approach. As
detailed in Section 18.9.2, new regulations for solvents used in aerosol spray cans are
based on the grams of ozone generated in the atmosphere per gram of product, not on
total solvent. The levels permitted vary based on the product in the aerosol can as well
as the individual solvents contained in the formulation. (See Ref. [68] for the California
regulation and tables of products and reactivity.)
An alternative to VOC analysis is to calculate VOC based on the formulation of the
coating. This calculation requires knowledge of the solvent content of all coating components and assumptions about what fraction of the solvents are actually emitted and
about how much additional VOC is produced by chemical reactions such as cross-linking.
Even so, calculated VOC values may be more reliable than measured values in many
cases, especially for waterborne coatings with very low VOC, in which the error of
water analysis is magnified.
18.9.2. Regulations
Regulators recognize that for adequate performance, some coatings require higher VOC
levels than others. The EPA assessed the most advanced technology for each end use
and established different maximum VOC guidelines for major applications. States and
378
SOLVENTS
localities within the United States may enforce more rigorous guidelines. California has
done so for many years, and in 2005 six northeastern states began enforcing stricter
limits. During the 1990s, the EPA guidelines ranged from 150 to 780 g L21 (1.9 to
6.5 lb/gal). Tighter EPA guidelines are expected.
Examples from the 1999 EPA VOC emission standards are 450 g L21 for architectural
enamels, 250 g L21 for exterior and interior flat paints, and 150 g L21 for traffic-marking
paint [69]. EPA regulations for automobile refinish coatings permit: wash primer, 780 g L21;
primers and primer–surfacers, 580 g L21; primer–sealer, 550 g L21; single-/two-step top
coats, 600 g L21; top coats of three or more steps, 630 g L21; and multicolor top coats,
680 g L21 [70].
In 1995 the U.S. EPA revised their wood furniture coatings regulation after extensive
discussions with the industry. It was agreed that regulations should be expressed as pounds
of VOC per pound of coating solids on the wood surfaces. The overall limits were set at
2.3 lb of VOC including a maximum of 1.0 lb of HAPs per pound of solids. In new
plants or finishing lines, HAP emissions are limited to 0.8 lb per pound of solids. For
specific coatings the following maxima were set: waterborne top coats, 0.8 lb; higher
solids solvent borne sealers, 1.9 lb; top coats, 1.8 lb; and acid-cured alkyd amino vinyl
sealers, 2.3 lb [71].
In some cases, such as automotive assembly plants, regulations are based on total
volatile emissions from a factory rather than individual coatings.
The automotive refinish coating limits apply to the entire country as the maximum
levels permitted, but states or regions are free to set more restrictive limits. In most
other applications the emission limitations apply only in regions that do not meet national
levels of ozone concentrations.
CARB frequently sets more restrictive regulations than the EPA. For example, CARB
2002 regulations [72] for architectural and maintenance coatings limit high gloss paints to
250 g L21 rather than the 450 g L21 allowed by EPA for high gloss enamels. Not only is
CARB often more restrictive than the EPA, but the South Coast Air Quality Management
District (southwestern California) often sets limits more restrictive than CARB. Effective
July 1, 2006, the VOC limits are 100 g L21 for chemical storage tank coatings, essential
public service coatings, industrial maintenance coatings, primers, sealers, and undercoats,
recycled coatings, rust preventive coatings, and specialty primers. In addition, as of July 1,
2008, the limit for flat paints will be 50 g L21. South Coast regulations can be accessed on
the Internet [73].
CARB has established regulations controlling solvents (and propellants) used in
aerosol coatings based on MIR values of incremental ozone production (gozone/gsolvent)
(Table 18.7) rather than the total solvent. MIR values can then be used to calculate the
MIR values of any mixture of solvents, which in turn permits the calculation of the
MIR per gram of coating. The emissions of reactive organic compounds (ROC) from
aerosol coatings in California was estimated to be 26.5 tons/day, with the capacity to
produce 56.3 tons of ozone per day. Table 18.8 gives the regulatory limits for grams of
ozone per gram of product for some of the 35 types of aerosol coatings.
More restrictive regulations are also being developed by the Ozone Transport
Commission (OTC) covering the 17 northeastern states, and six of these began enforcing
tighter limits in 2005. For examples of OTC regulations and comparisons with EPA,
CARB, and California South Coast regulations, see Ref. [74].
The limit on HAP emissions at any facility has been set by the EPA at 10 tons per
individual HAP per year, and a total of 25 tons per year for all HAPs. Lower limits are
REFERENCES
379
TABLE 18.8. MIR Limits (g L21) for Selected Aerosol
Coatings
Flat paints
Non-flat paints
Primers
Auto body primers
Automobile top coats
Auto bumpers and trim
Traffic paint
Marine spar varnish
1.21
1.40
1.11
1.57
1.77
1.75
1.18
0.87
set for individual chemicals thought to be more dangerous. The following solvents and
plasticizers are on the HAP list: benzene, di-2-ethyhexyl phthalate, dibutyl phthalate,
dimethylformamide (no more than 1 ton per year), dimethyl phthalate, dioxane, ethylene
glycol, mono and di alkyl and aryl ethers of ethylene glycol (except the butyl ether of
ethylene glycol), diethylene glycol, triethylene glycol, hexane, methoxyethanol, ethoxyethyl acetate, isophorone, methyl ethyl ketone, methylene chloride (no more than 4.0
tons per year), 2-nitropropane (no more than 1.0 ton per year), cresols and phenol,
toluene, triethylamine, and xylenes.
A 2004 European Directive covering decorative, functional, and protective coatings
mandates strict VOC limits to take effect in 2007, with further tightening in 2010 [74].
A paper is available that urges the establishment of more uniform international regulations of emissions of VOCs from coatings, using automotive refinish coatings as an
example [75]. The authors point out that as the globalization of industry proceeds, large
inefficiencies result from having different requirements in different countries for similar
coatings.
GENERAL REFERENCES
Ash, M.; Ash, I., Handbook of Solvents, 2nd ed., Synapse Information Resources, Endicott, NY, 2003.
Ellis, W. H., Solvents, Federation of Societies for Coatings Technology, Blue Bell, PA, 1986.
Wypych, G., Ed., Handbook of Solvents, ChemTech, William Andrew, New York, 2001.
REFERENCES
1. Ostrowski, P., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2000,
pp. 448 – 459.
2. Cooper, C.; et al., J. Coat. Technol., 2001, 73(922), 19.
3. U.S. Department of Commerce, Bureau of Census, Current Industrial Reports: Paint and
Coating Manufacturing, http://www.census.gov.
4. Burrell, H., Off. Dig., 1955, 27, 726.
5. Hildebrand, J., J. Am. Chem. Soc., 1916, 38, 1452.
6. Patton, T. C., Paint Flow and Pigment Dispersion, 2nd ed., Wiley-Interscience, New York,
1979, pp. 306 – 310.
7. Hoy, K. L., J. Paint Technol., 1970, 42(541), 76.
380
SOLVENTS
8. Lieberman, E. P., Off. Dig., 1962, 34(444), 30.
9. Ellis, W. H., Solvents, Federation of Societies for Coatings Technology, Blue Bell, PA, 1986.
10. Hansen, C. M., in Paint and Coatings Testing Manual, 14th ed., Koleske, J. V., Ed., ASTM,
Philadelphia, PA, 1995, pp. 383 – 404.
11. Grulke, E. A., in Polymer Handbook, 4th ed., Brandrup, J.; et al., Eds., Wiley, New York, 1999,
p. VII/675.
12. Hoy, K. L., Tables of Solubility Parameters, Union Carbide Corp., Chemicals and Plastics, R&D
Dept., Charleston, WV, 1969.
13. Huyskens, P. L.; Haulait-Pirson, M. C., J. Coat. Technol., 1985, 57(724), 57.
14. Rawls, R., Chem. Eng. News, 1985, March 18, 20.
15. Kamlet, M. J.; et al., Chemtech, 1986, September, 566.
16. ARCOCOMP Computer Solvent Selector Program, ARCO Chemical Co., Newton Square, PA,
1987.
17. Rocklin, A. L., J. Coat. Technol., 1976, 48(622), 45.
18. Jackson, H. L., J. Coat. Technol., 1986, 58(741), 87.
19. Skjold-Jorgenson, S.; et al., Ind. Eng. Chem. Prod. Res. Dev., 1979, 18, 714.
20. Yoshida, T., Prog. Org. Coat., 1972, 1, 72.
21. Patton, T. C., Paint Flow and Pigment Dispersion, 2nd ed., Wiley-Interscience, New York,
1979, p. 340.
22. Dillon, P. W., J. Coat. Technol., 1977, 49(634), 38.
23. Rocklin, A. L., J. Coat. Technol., 1986, 58(732), 61.
24. Watson, B. C.; Wicks, Z. W., Jr., J. Coat. Technol., 1983, 55(698), 59.
25. Hansen, C. M., Ind. Eng. Chem. Prod. Res. Dev., 1970, 9, 282.
26. Newman, D. J.; Nunn, C. J., Prog. Org. Coat., 1975, 3, 221.
27. Hansen, C. M., J. Oil Colour Chem. Assoc., 1968, 51, 27.
28. Waggoner, R. A.; Blum, F. D., J. Coat. Technol., 1989, 61(768), 51.
29. Lasky, R. C.; et al., Polymer, 1988, 29, 673.
30. Vrentas, J. S.; et al., Macromolecules, 1996, 29, 3272.
31. Wu, S. H., J. Appl. Polym. Sci., 1978, 22, 2769.
32. Bauer, D. R.; Briggs, L. M., J. Coat. Technol., 1984, 56(716), 87.
33. Hill, L. W.; Wicks, Z. W., Jr., Prog. Org. Coat., 1982, 10, 55.
34. Ellis, W. H., J. Coat. Technol., 1983, 53(696), 63.
35. Brandenburger, L. B.; Hill, L. W., J. Coat. Technol., 1979, 51(659), 57.
36. Wicks, Z. W., Jr.; et al., J. Coat. Technol., 1982, 54(688), 57.
37. Croll, S. G., J. Coat. Technol., 1987, 59(751), 81.
38. Sullivan, D. A., J. Paint Technol., 1975, 47(610), 60.
39. Patton, T. C., Paint Flow and Pigment Dispersion, 2nd ed., Wiley-Interscience, New York,
1979, p. 109.
40. Sprinkle, G. F., Jr., Mod. Paint Coat., 1983, April, 44.
41. Rocklin, A. L.; Edwards, G. D., J. Coat. Technol., 1976, 48(620), 68.
42. Jones, F. N., J. Coat. Technol., 1996, 68(852), 25.
43. Erickson, J. R.; Garner, A. W., Org. Coat. Plast. Chem. Prepr., 1977, 37(1), 447.
44. Hazel, N.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1997,
p. 237.
45. Ellis, W. H., J. Coat. Technol., 1976, 48(614), 45.
REFERENCES
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
381
Wu, D. T.; et al., FATIPEC Congress Book, 1988, Vol. IV, p. 227.
McGovern, J. L., J. Coat. Technol., 1992, 64(810), 33, 39.
Hahn, F. J., J. Paint Technol., 1971, 45(562), 58.
Overdiep, W. S., Prog. Org. Coat., 1986, 14, 159.
Smith, R. L.; et al., J. Coat. Technol., 1987, 59(747), 21.
Brezinski, J. J., in Paint and Coating Testing Manual, 14th ed., Koleske, J. V., Ed., ASTM,
Philadelphia, PA, 1995, pp. 3 – 14.
U.S. EPA, Hazardous Air Pollutants Strategic Implementation Plan, http://www.epa.gov/ttn/
atw/index.html.
U.S. EPA, Final Air Toxics Regulation for Wood Furniture Manufacturing Operations, http://
www.epa.gov/ttn/utaw/fswood.html (1995).
U. S. EPA, National Air Pollution Emission Trends, 1900– 1998. Air Emission Trends, Continued Progress Through 1994. Detailed reports are accessible at www.epa.gov, and are updated
periodically.
U.S. EPA, Final Revisions to the Ozone and Particulate Matter Air Quality Standards, http://
www.epa.gov/oar/oaqps/ozpmbro/current.htm, and http://www.epa.gov/aiA/greenbk/
ozone1hr/dec97/o3std.html.
Dobson, I. D., Prog. Org. Coat., 1994, 24, 55.
Seinfeld, J. H., Science, 1989, 243, 745.
Atkinson, R., Atmos. Environ., 1990, 24A, 1.
National Research Council, Rethinking the Ozone Problem in Urban and Regional Air
Pollution, National Academy Press, Washington, DC, 1991.
http://www.arb.ca.gov/consprod/regs/regs.htm.
http://www.americansolventscouncil.org/regulatory/HAP.asp.
Nielsen, K. A.; et al., Polym. Mater. Sci. Eng., 1990, 70, 170.
Praschan, E. A., ASTM Stand. News, 1995, October, 24.
Wicks, Z. W., Jr.; Chen, G. F., J. Coat. Technol., 1978, 50(638), 39.
Hill, L. W.; et al., in Film Formation in Waterborne Coatings, Provder, T.; et al., Eds., American
Chemical Society, Washington, DC, 1996, p. 235.
Brezinski, J. J., Manual on Determination of Volatile Compounds in Paints, Inks, and Related
Coating Products, 2nd ed., ASTM Manual Series MNI 4, ASTM, Philadelphia, PA, 1993.
Ancona, B.; et al., J. Coat. Technol., 1993, 65(820), 45.
http://www.arb.ca.gov/consprod/regs/regs.htm.
http://www.epa.gov/ttn/atw/183e/aim/fr1191.brt.
http://www.epa.gov/ttn/atw/180e/arc/fr1194.txt.
Chem Eng. News, 1995, 25.
http://www.placer.ca.gov/airpollution/regs/webrule-218-architectural-coatings.pdf.
http://www.arb.ca.gov/coatings/coatingsrules.htm.
Kapsanis, K.; Goldie, B., J. Prot. Coat Linings, 2005, 22, 36.
Drexler, H. J.; Sell, J., Eur. Coat. J., 2002, http://www.ippic.org/ecj_voc_reg.pdf.
19
Color and Appearance
Color and the interrelated topic of gloss are important to the decorative aspects of coatings
and sometimes, to their functional aspects. Although we have all experienced color since
we were babies, but most people have little understanding of it. Many technical people
think of color as an aspect of physics dealing with the distribution of visible light.
Although that is a factor, color is a psychophysical phenomenon. The difficulty of understanding color can be grasped by considering its most rigorous definition: Color is that
characteristic of light by which an observer may distinguish between two structure-free
fields of view of the same size and shape. In effect, it says that color is what is left to
distinguish between two objects when all other variables have been removed—not a
very satisfying definition.
Color has three components: an observer, a light source, and an object. (The single
exception is when the light source is the object being viewed.) There is no color on an
uninhabited island. This is not just a semantic statement; color requires an observer.
There is no color in the absence of light; in a completely darkened room there is no
color, not because you cannot see it, but because it is not there. There must be an
object; if you look out the window of a spaceship without looking at a planet or star,
there is no color—there is an observer, there is light, but there is no object.
Another major factor affecting appearance is surface roughness. If a surface is very
smooth, it has a high gloss; if it is rough on a scale below the ability of the eye to
resolve the roughness, it has a low gloss. If, however, the roughness can be resolved visually, a film may exhibit scratches, brush marks, orange peel, texture, and so on. Furthermore, there can be a combination of small scale and larger scale roughness, so films can,
for example, have low gloss and brush marks or high gloss and orange peel. The eye can
resolve irregularities in surface smoothness of approximately 25 mm, depending on the
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
382
19.1. LIGHT
383
distance from the object. Adding to the complexity, color and gloss interact—changing
either one changes the other.
19.1. LIGHT
Light is electromagnetic radiation to which our eyes are sensitive. The range of visible
wavelengths varies somewhat among individuals, but in most cases is 390 through
770 nm. The sensitivity of the eye varies as a function of wavelength, as shown in
Figure 19.1. The figure also shows the response of photomultiplier tubes and silicon photodiodes. The eye is relatively insensitive to the shorter and longer ends of the range. The
response of our eyes depends on the distribution of wavelengths of light emitted by the
light source. In the case of monochromatic light sources, the colors we see range from
violet through blue, green, yellow, and red with increasing wavelength. We see different
colors as the ratios of wavelengths in polychromatic light sources change. If we look at a
light source with nearly equal content of all wavelengths, we see white.
Sunlight is considered the standard light, but sunlight varies depending on the time of
day, latitude, season, cloudiness, and so on. When work toward understanding color
began, light from overcast north sky (in the northern hemisphere) was accepted as standard. Based on many measurements of energy distribution, a light source designated as
D65 was adopted as a standard related to average daylight. The mathematical description
of a light source is called an illuminant. A graph of standard illuminant D65 is shown in
Figure 19.1. Sensitivity of the eye, photomultiplier tube, and silicon photodiode as a function of wavelength. (From Ref. [1], with permission.)
384
COLOR AND APPEARANCE
Figure 19.2. Relative spectral power distributions of CIE standard illuminants A and D65. (From Ref. [2],
with permission.)
Figure 19.2. Tables giving energy distributions as a function of wavelength for various
bandwidths from 1 to 20 nm wide are available [3].
The energy distribution from tungsten lights is different. Another standard light source,
A, is a carefully specified tungsten light operated under specified conditions. A graph of
illuminant A is also shown in Figure 19.2, and tables of its energy distributions are available. Fluorescent lights are another type of light source, and many types are available. As
shown in Figure 19.3, they exhibit a continuous energy distribution with peaks at a series
Figure 19.3. Spectral power distribution of a cool white fluorescent lamp (IES 1981). (From Ref. [1], with
permission.)
19.2. LIGHT – OBJECT INTERACTIONS
385
of wavelengths. Even if the underlying continuous spectrum were the same as D65, the
peaks would lead to changes in colors when these lights are used as a light source in
comparison with daylight.
19.2. LIGHT – OBJECT INTERACTIONS
An important factor that affects color is the interaction between the light and the object
being viewed.
19.2.1. Surface Reflection
When a light beam is directed at a surface, some light is reflected at the surface and some
passes into the object. As shown in Figure 19.4, if the surface is optically smooth, light is
reflected at the same angle (r) as the angle of incidence (i) [4]. This kind of reflectance is
called specular (mirror-like) reflectance. By convention, an angle of incidence normal to
the surface is designated as 08 and the grazing angle is 908. The fraction of light reflected
(R) varies with angle of incidence and difference in refractive index (n) between the two
phases. If there is no difference in index of refraction, no light is reflected at the interface;
as the difference increases, the fraction of light reflected increases. For angles of incidence
near 08, the fraction reflected can be calculated using the following equations given in both
the general form, Eq. 19.1a, and where the first medium is air (n1 ¼ 1), Eq. 19.1b:
n2 n1 2
R¼
n2 þ n1
n1 2
R¼
nþ1
(19:1a)
(19:1b)
Figure 19.4. External and internal reflection and refraction of light by a nonabsorbing film (refractive
index, n1, thickness, x) with optically smooth parallel surfaces. (From Ref. [4], with permission.)
386
COLOR AND APPEARANCE
Most resins have refractive indexes of about 1.5; Eq. 19.1b shows that approximately
4% of incident light is reflected when the angle of incidence is near 08. The dependence of
reflection on angle of incidence is illustrated in Figure 19.5; reflectance approaches 100%
as the angle of incidence approaches 908. The family of curves in Figure 19.5 illustrates
the combined effects of refractive index difference and angle of incidence.
Light not reflected at the surface enters the object. When a light beam enters an object,
it is refracted; that is, it is bent, also shown in Figure 19.4. The angle of refraction varies
with the ratio of the refractive indexes of the two media:
sin r ¼
n1
sin i
n2
(19:2)
When light passes from air into a nonabsorbing plastic film with parallel, optically
smooth surfaces and an index of refraction of 1.5, the angle of refraction is smaller than
the angle of incidence (Figure 19.4). If the angle of incidence is near 08, the fraction
reflected at the first surface is 0.04, and 0.96 is transmitted into the film. If there is no
absorption, 0.96 reaches the second film/air interface. There, 0.96 0.96 of the original
light is transmitted into the air on the other side of the film, being refracted so that the angle
of refraction is equal to the original angle of incidence. There is also reflection at the
second surface, and 0.04 0.96 is reflected back toward the first surface. Again, there
is reflection of 4% and transmission of 96% at the back of the first surface. As a result,
(0.96)2 (0.04) of the original light is transmitted out of the film at an angle equal to
the original angle of incidence, and (0.96) (0.04)2 is reflected back into the film. This
bouncing back and forth continues, if there is no absorption, until the total transmission
equals 0.92 and total reflectance equals 0.08.
When a light beam passes from a medium of higher refractive index to one of lower
refractive index, the angle of the beam increases. If the angle of incidence is high
enough, all light is reflected back and none is transmitted out. If the refractive indexes
Figure 19.5. Fraction of light reflected from a smooth surface as a function of the angle of incidence i
with various differences in refractive index. (From Ref. [5], with permission.)
19.2. LIGHT – OBJECT INTERACTIONS
387
are 1.5 and 1, the critical angle (at which all the light is reflected) is calculated to be 41.88.
Total reflection is the basis for optical fibers. If light enters the end of a nonabsorbing fiber
in which there are no bends, such that the angle of incidence with the inner surface of the
fiber is less than its critical angle, all of the light will follow the fiber, even all the way
across an ocean.
If the thickness of a film is small (60 to 250 nm), interference colors will be seen. The
intensity of some wavelengths of reflected light are enhanced, and others are transmitted
selectively. With a refractive index of 1.5 in air, at a film thickness of 66 nm, bluish-white
color is reflected and yellowish-white color is transmitted; as thickness is increased, other
colors are preferentially reflected. The colors seen are affected by the angle of incidence of
the light.
19.2.2. Absorption Effects
In almost all coatings, the color we observe is affected by differential absorption of
various wavelengths of light. Colorants, dyes, pigments, and to a degree, some resins,
absorb some wavelengths of light more strongly than others. These absorptions are controlled by the chemical structures of the colorants. First, we consider the absorption
effects in transparent systems, that is, systems in which the colorant is in solution or
is so finely divided that it does not significantly reflect light at the colorant –resin
interface. The extent of absorption depends on chemical composition, wavelength,
particle size, optical path length (film thickness), concentration, and media – colorant
interactions.
Each colorant has an absorption spectrum that controls the absorption of various wavelengths of light. Chemists generally speak of molar absorptivities, 1, in units of Lmol21
path length21, corresponding to the absorbance per molar concentration of a substance.
Physicists generally speak of absorption coefficients, K, in units of path length21
mass21, corresponding to the absorbance per unit mass of the system.
The smaller the particle size, the greater the fraction of light absorbed by the same
quantity of a colorant. The highest molar absorptivities are exhibited by individual molecules in solution. In the case of pigments, the smaller the particle size of a pigment, the
greater the absorption. The longer the path followed by a beam of light through a medium
containing absorbers, the greater the degree of absorption. In transmission of a beam of
light at 08, the optical path equals the film thickness. At any other angle, the optical path
length is greater than the film thickness. If, in passing through a unit path length, one-half
of the light of a particular wavelength is absorbed and one-half is transmitted (ignoring
surface reflection), on passing through two units of path length, three-fourths is absorbed
and one-fourth is transmitted. Mathematically, this relationship is expressed by an exponential equation, as shown in Eqs. 19.3a and 19.3b, in which X is path length, I is intensity of light transmitted, and I0 is the original light intensity. Chemists commonly use
base 10 (Eq. 19.3a), and physicists base e (Eq. 19.3b):
I
¼ 101X
I0
(19:3a)
I
¼ eKX
I0
(19:3b)
388
COLOR AND APPEARANCE
Figure 19.6. Transmission spectra of idealized magentas: (a) path length ¼ x; (b) path length ¼ 2x.
Ideally, the same relationship holds when the concentration of the colorant in the
medium is changed:
I
¼ 101CX
I0
(19:4)
Equation 19.4 holds only over a limited range of concentrations; the width of the range is
system dependent. When absorbers are in solution, there are less likely to be intermolecular interactions between molecules in dilute solutions than in more concentrated
solutions. In the case of pigment dispersions, concentration effects are likely to be
larger. Another complication is possible interactions with the media. In the case of solutions, a change in solvents can sometimes lead to association of molecules effectively
increasing particle size and hence reducing absorption. There can also be effects of
hydrogen bonding between a soluble dye molecule and different solvents. This
changes the structure, leading to a change in absorption spectrum. In dispersions of
pigments, changes in the medium, such as dilution with solvent, can lead to flocculation
(agglomeration) of pigment particles; this causes a larger particle size and hence reduced
absorption.
Figure 19.6 shows transmission spectra of idealized reddish magentas. Spectra (a) and
(b) result from transmission of light through the same transparent coating of path length X
and 2X, respectively. (Surface reflection effects are ignored.) The fraction of light of all
wavelengths transmitted through the thicker coating, shown in spectrum b, is less. But
also, the relative transmissions of the blue (B), green (G), and red (R) parts of the
spectra are different. The color seen in case b would be a redder purple than in case a.
The composition is the same but the shade of the color is affected by path length. The
same type of change takes place if the concentration is doubled at the same film thickness.
This is an idealization of a real situation. One of the inks in four color process printing that
permits printing of multicolored pictures is a reddish magenta. The film thickness of
magenta ink affects the color of the magenta print and hence the combined color picture.
19.2.3. Scattering
Scattering is another phenomenon that can occur during the passage of light through a film.
If there are small particles dispersed in the film that have a different refractive index than
19.2. LIGHT – OBJECT INTERACTIONS
389
the medium, light is reflected and diffracted at the interfaces between the particles and the
medium. The physics involved in scattering is complex and beyond the scope of this brief
presentation, but the results can be stated in a simplified manner. When a beam of light
passes through a film containing nonabsorbing particles, the light is internally reflected
in all directions, so it changes from a beam to diffuse illumination inside the film. Light
that reaches the back of the top surface at angles greater than the critical angle is reflected
back into the film; part of the light reaching this surface at angles less than the critical
angle leaves the film. If the film is thick enough, no light can pass through it completely;
all the light is reflected back out of the top surface. The reflection is, however, not just at
the specular angle, but at all angles; that is, this light is diffusely reflected even though the
top surface of the film is optically smooth. The extent of light scattering by nonabsorbing
particles depends on refractive index difference between particles and medium, particle
size, film thickness, and particle concentration.
The greater the difference in refractive index, the greater the degree of light scattering.
The degree of scattering increases steeply as the difference in refractive index increases
(Figure 19.7). The degree of scattering is the same whether the particles have a higher
or lower refractive index than the medium. For example, water droplets in air (fog)
scatter light as efficiently as air droplets of the same size at the same concentration in
water (foam). An ideal white pigment would absorb no light and have a very high refractive index, so the difference in refractive index from that of binders would be large. Rutile
TiO2 comes close to meeting these specifications; it has an average refractive index of
2.73, but it absorbs some light below about 420 nm. Another crystal type of TiO2,
anatase, absorbs less light, but its refractive index is lower, 2.55. The smaller index of
refraction difference leads to less efficient scattering.
Scattering is affected by particle size (Figure 19.8). For rutile TiO2 in a resin with a
refractive index of about 1.5, the scattering coefficient for 560-nm light increases with
particle size to a maximum at 0.19-mm diameter, then drops off steeply [6]. Commercial
TiO2 has a range of particle sizes. Since efficiency drops off more rapidly on the small
diameter size side of the maximum, TiO2 pigments are produced with an average particle
diameter somewhat over 0.2 mm. See Ref. [7] for calculations showing the effect of particle size distribution and particle agglomeration on scattering. The particle diameter with
Figure 19.7. Scattering as a function of refractive index difference; particles have higher refractive
indexes than the media on the right-hand part of the curve and lower values on the left. (From
Ref. [1], with permission.)
390
COLOR AND APPEARANCE
Figure 19.8. Scattering coefficients as a function of particle size for rutile TiO2 and CaCO3. (From
Ref. [6], with permission.)
maximum scattering depends on refractive index difference. For calcium carbonate
(n ¼ 1.57), the scattering coefficient goes through a maximum at about 1.7 mm. As
expected from the small difference in refractive index, scattering by even optimum particle
size is low.
The extent of light scattering is affected by film thickness. If no absorption occurs, the
light that is not reflected back out of the top surface is transmitted unless the film thickness
is great enough so that all the light is reflected back. Scattering is also affected by concentration. At low concentrations of particles, such as rutile TiO2, transmission through a film
(neglecting surface reflection) follows a relationship similar to that for absorption:
I
¼ eScX
I0
(19:5)
As concentration increases, however, scattering efficiency reaches a maximum, then
decreases. In the case of rutile TiO2 pigmented coatings, the efficiency decreases so
much that finally, the fraction of light transmitted increases and the fraction reflected
decreases. The result can be seen in a plot of scattering coefficient S as a function of
pigment volume concentration (PVC) of rutile TiO2 in dry films [6] (Figure 19.9).
PVC is defined as the volume percent pigment in a dry film of coating. In commercial
practice, the cost effectiveness of pigmenting with TiO2 drops off sufficiently at PVCs
above about 18% that it is not generally economically sound to use a higher PVC of
TiO2. The optimum value varies somewhat from system to system; it depends on how
well the pigment is dispersed in the film and on the TiO2 content of the particular TiO2
pigment, which may vary from grade to grade by more than 10% (Section 20.1.1).
19.2.4. Multiple Interaction Effects
The effects of surface reflection, absorption, and scattering are interdependent. Light
sources generally are not narrow beams, but relatively broad beams or diffuse sources.
Surfaces are not often optically smooth; in some cases, we strive for surface roughness.
In coatings, we seldom deal with pigments or pigment combinations that only absorb or
19.2. LIGHT – OBJECT INTERACTIONS
391
Figure 19.9. Experimental plots of scattering coefficients versus PVC for selected pigments in a dry
acrylic lacquer. BCWL is basic carbonate of white lead. (From Ref. [6], with permission.)
only scatter light; commonly, both occur simultaneously. We are seldom interested in free
films, but rather in films on a substrate so that the reflectance at the bottom surface of a film
is not the same as that of the top surface of the film. The eye does not distinguish between
the light reaching it that has been reflected from the top surface, from within the film, or
from the bottom surface of a film. The combined light from all three sources is integrated
when the eye responds.
Surface reflectance increases as the angle of incidence increases (Figure 19.5). If a panel
of a high gloss (smooth surface) blue paint illuminated with a beam of white light is
observed at any angle other than the specular angle, one sees a comparatively dark blue
color, resulting from the diffuse reflectance back out of the film. If the angle of incidence
approaches 908, one sees a very light blue color at the specular angle because the observed
light contains a high fraction of surface reflected light and only a little light from within the
film. If the illumination is diffuse instead of a beam, the color is a lighter blue from most
angles of observation because one sees a higher fraction of surface reflected light. If the
blue paint is low gloss (rough surface) with exactly the same composition as in the
earlier example, a lighter blue color is seen under diffuse illumination at most viewing
angles compared with the high gloss paint. This results from a greater ratio of surface
reflected light reaching the eye. The color is darkened at most viewing angles by wetting
the surface with water, since the water fills in the surface roughness.
The effects of pigment concentration and film thickness on coatings that absorb or
weakly scatter light without rescattering the primary scattered light can be modeled by
simple equations. However, interactions resulting from both absorption and multiple scattering together are more complex, and equations that model reflection and transmission are
more complex. Figure 19.10 gives a set of Kubelka – Munk equations that model reflectance, R1 and transmittance T1 of a translucent or opaque film or slab of material of thickness X containing light absorbers and scatterers. The observer is viewing the sample in air,
while the absorbers and scatterers are embedded in a medium such as plastic or resin.
The Kubelka –Munk reflectance R and transmittance T of the film in the absence of a
refractive index boundary include the effects of absorption K, scattering S, and film thickness X. The boundary reflections from both sides of the top surface r1, and the inner and
bottom sides of the film r, as well as the corresponding boundary transmittances t1 and t
influence the reflectance R1 and transmittance T1 seen by the observer viewing the sample
392
COLOR AND APPEARANCE
Figure 19.10. Kubelka–Munk equations and assumptions.
in an air medium. Two boundary reflectances r1 and r are required because light normally
incident on the film from air has a reflectance value of 0.04, while diffuse light incident on
the boundary from inside the resin has a reflectance value of 0.596 for a material of refractive index about 1.5. The differences between R1 and R, as well as those between T1 and T,
in the case of typical plastics and resins are very significant. Other Kubelka – Munk
equations are available that permit calculations when the reflectances at the inner side
of the top and bottom sides of the film are different [8].
19.3. HIDING
Color is affected by reflectance back through the film of light reaching the substrate to
which the coating is applied. It is common to compare the colors of a coating applied
over a striped black and white substrate. If one can see the pattern of the stripes
19.4. METALLIC AND INTERFERENCE COLORS
393
through the coating, the coating is said to exhibit poor hiding. The difference results from
the reflection of light reaching the white stripes compared with absorption of light reaching
the black stripes. The effect is as if some black pigment was put in the coating above the
black stripes. If all light entering the film is absorbed or scattered back out of the film prior
to reaching the substrate, there is no effect of substrate on color; hiding is complete. Simplified Kubelka – Munk equations can be used for such opaque films (Figure 19.10).
Hiding is a complex phenomenon and is affected by many factors. Hiding increases as
film thickness increases. Low hiding coatings require thicker films; that is, coverage
decreases; the area covered (hidden) by a liter of coating is less and the cost is higher.
Hiding increases as the efficiency of light scattering increases; that is, hiding is affected
by the refractive index differences, particle sizes, and concentrations of scattering pigments present. Hiding increases as absorption increases. Carbon black pigments, which
have high absorption coefficients for all wavelengths, are particularly effective. Colorants
also increase hiding but not as much as black. Surface roughness increases hiding; a larger
part of the light is reflected at the top surface, reducing the differences of reflection resulting from differences in the substrate to which a coating is applied.
An important factor affecting hiding, sometimes forgotten in testing coatings, is uniformity of film thickness. Application of coatings commonly results in nonuniform thickness. Coatings are generally formulated to level, that is, to flow after application to make
the film thickness more uniform (Section 24.2). Leveling is often incomplete, however,
and there can be effects on hiding. Consider a poor leveling coating where an average
film thickness of 50 mm of dry coating has been applied, but there are brush marks remaining, so adjacent to each other are lines with film thicknesses of 65 and 35 mm. If the hiding
at 50 mm is just adequate, hiding will be inferior at 35 mm. The difference in color is
emphasized by the thin layers of coating being right next to the thick layers. The contrast
results in a poorer rating in hiding by the uneven film than by a uniform 35-mm film of the
same coating. Further complicating hiding is the variation of substrates over which the
coating is applied. Hiding of a white coating applied over a white surface might be
rated as excellent, but hiding of the same coating over a black surface might be rated poor.
There are quality control tests that compare hiding of batches of the same or similar
coatings, but no test is available that can provide an absolute measure of hiding [9].
The only way to establish covering power of a coating is to apply it to an appropriate
large surface with a film thickness that gives adequate hiding and then calculate the
actual coverage in m2 L21 or ft2/gal.
In some cases, hiding is not desired; an example is a coating for metallized plastic
Christmas tree ornaments. One wants transparent red, green, blue, and so on, coatings.
Preparation of a transparent coating requires that there be no light scattering within the
film; therefore, the particle size of colorant particles must be very small.
19.4. METALLIC AND INTERFERENCE COLORS
Metallic coatings are widely used on automobiles. They are made with transparent colorants together with nonleafing aluminum pigment (Section 20.2.5). These metallic coatings
exhibit unusual shifts in color as a function of viewing angle (Figure 19.11). When an
observer looks at a metallic film from an angle near the normal, the path length of light
through the film is short because it is reflected back by the aluminum. On the other
hand, from a greater viewing angle, the observer sees light that has been reflected back
394
COLOR AND APPEARANCE
Figure 19.11. Idealized diagram of the reflection of light in a metallic coating.
and forth within the film so that the path length is longer and the color is darker. Metallic
coatings are lighter in color when viewed near the normal angle (the face color) and darker
when viewed from a larger angle (the flop color). It is desirable to achieve a high degree of
color flop, that is, a large difference between face and flop colors. To do so requires a
smooth surface (high gloss), a transparent film having very little light scattering from
the resin or color pigment dispersion, and parallel alignment of the aluminum flake particles to the surface of the film. (See Section 30.1.2 for a discussion of possible mechanisms for orientation of aluminum during film formation.)
Pigments that produce colors by interference are also used in automotive coatings,
sometimes in combination with metallic flake. They are often called effect pigments.
Three types of such pigments that are used commercially are described in Section 20.2.5.
19.5. THE OBSERVER
The third critical aspect of color is the observer; both the eye and the brain are involved. The
retina of the eye has two types of photodetectors, rods and cones. These detectors are stimulated by light photons to send signals through the optic nerves to the brain. At low levels of
illumination, rods are sensitive to all wavelengths of light and become insensitive as the
level of illumination increases. Cones are insensitive at low levels of illumination but are
sensitive at higher levels. The response of cones is wavelength dependent, having three
ranges of sensitivity that overlap. One peaks in the blue region of the spectrum, another
in the green, and the third in the red region. These overlapping ranges lead to complex
signals to the brain, which integrates the signals so that we can see thousands of different
colors. At low levels of illumination, we see only with rods, and vision is limited to
shades of gray. At intermediate levels of illumination, there are responses from both rods
and cones; hence, one does not see bright colors, only grayish ones. At higher levels of illumination, the rods are inactive and we see a full range of colors from the responses of cones.
The mechanism of color vision is complex, only partly understood, and beyond the
scope of this book. There are variations from person to person in the degree of response
to the three ranges of sensitivity. As a result, colors seen by different observers are not
identical. Usually, the differences are small, but in some cases they are large. In
extreme cases, some people are color blind. There are different types of color blindness;
the most common is red –green color blindness.
19.5. THE OBSERVER
395
To specify color and predict responses to mixtures of colorants, a mathematical
model of a standard human observer was established by an international committee
of experts known by the acronym of its French name as CIE. Figure 19.12 shows a
graph of three functions that model color vision. At any given wavelength, the standard
observer has the same response to the ratio of the values of x̄, ȳ, and z̄ at that wavelength as to the monochromatic light of that wavelength. Tables of CIE Color Matching
Functions are available as a function of wavelength with bandwidths of 1, 10, and
20 nm [3].
Within the range of higher illumination, the eye adapts to changes in levels of illumination. For example, if an area of white is surrounded by black, the white looks whiter than
if the black were not present. The eye adapts to the level of light reflected by the combined
black and white and responds more to the white in the presence of the black. Similar
effects occur when two strongly colored fields of view are adjacent to each other.
Yellow surrounded by blue –green looks more orange than if the blue – green were
absent. In general, if one looks at color chips when selecting a coating, the color is different than if that coating is applied to a large surface because of the effect of the surrounding
area on the color. Many other effects result from such interactions between the eyes and
combinations of colors on a surface. (See Ref. [5] for further discussion.)
Figure 19.12. CIE color matching functions, x̄, ȳ, and z̄, for equal energy spectra. (From Ref. [1], with
permission.)
396
COLOR AND APPEARANCE
19.6. INTERACTIONS OF LIGHT SOURCE, OBJECT, AND OBSERVER
Color depends on the interaction of three factors: light source, object, and observer. If any
factor changes, the color changes. If we observe an object under a light source with the
energy distribution of illuminant A (Figure 19.4) and shift to a different illuminant, for
example, illuminant C (similar to D65), the color changes. Light source A has relatively
lower emission in the blue end of the spectrum and relatively higher emission in the red
end. The light reaching the eye when the object is illuminated with source A has more
red light and less blue light than the light reflected from the same object illuminated
Figure 19.13. Light source effects and metamerism. (From Ref. [1], with permission.)
19.7. COLOR SYSTEMS
397
with source C. The color is different. The situation is illustrated in spectral sets a and b in
Figure 19.13, which shows the different responses as product spectra of light source
object observer for an object with illuminants A and C.
The only way to be sure that a pair of coatings will match under all illuminations is to
use colorants with identical chemical composition and physical states. If the chemical
composition and physical state of the colorants in two coatings are the same, their reflectance spectra are identical, and the coatings match under any light source unless there are
differences in pigment distribution within the coatings. It is possible for two materials with
different colorant compositions, and hence different reflectance spectra, to result in the
same color under a light source. However, such a pair will not match under light
sources with different energy distributions. This phenomenon is called metamerism. In
a spectral match, the two panels change color with a new light source, but it is the
same change in both cases. In a metameric pair, the color is the same with one light
source; the colors of both panels also change when the light source is changed, but the
extent of change is different between the two panels. This situation is illustrated in
Figure 19.13, which shows responses of a second object with illuminants C (spectral set
c) and A (spectral set d). The first and second objects (a metameric pair) are identical
under illuminant C, but their colors are different under illuminant A.
19.7. COLOR SYSTEMS
The human eye can discriminate thousands of colors. However, it is difficult for a person to
tell another person what colors he or she sees. A color description such as “light grayish
blue – green” illustrates this difficulty. Many systems have been developed to permit
definitive identification of colors. There are two types of systems: one that uses color
samples in specific arrangements and one that identifies colors mathematically. All
color systems have at least three dimensions, to include all the possible colors.
The most widely used visual color system in the United States is the Munsell Color
System, in which carefully prepared and selected color chips are classified in a threedimensional system. The dimensions of the Munsell System are called hue, value, and
chroma. Hue refers to the dimension of color, described as blue, blue-green, green,
greenish-yellow, yellow, yellowish-red, red, reddish-purple, purple, purplish-blue, and
back to blue. The value dimension characterizes the lightness of a color as compared to
a series of gray samples, a gray scale. The value 0 is assigned to pure black and the
value 10 to pure white. A light blue has a high value, whereas a dark blue of the same
hue has a low value. The chroma dimension refers to the difference between a color
and a gray of the same value and hue. A bright red has a high chroma, whereas a
grayish red of the same hue and value has a low chroma.
The color chips in the Munsell System are prepared so that there are equal visual differences between all pairs of adjacent chips. The chips are labeled systematically; for
example, a chip labeled G5/6 is a green with a value of 5 and a chroma of 6. One can
look at a set of Munsell chips and know the color that someone meant when they made
such a designation. There are two limitations to this statement. The light source must
be specified. Chip G5/6 gives a different color under source A than under source D65.
Second, surface roughness affects color, so comparisons have to be made at equal gloss
levels. Two sets of Munsell chips are available: one with high gloss and the other with
low gloss. Significant errors can result in comparing either of these with materials that
398
COLOR AND APPEARANCE
are semigloss. In Europe, a different color chip system, the Natural Color System, is most
widely used [10].
The mathematical color system is the CIE Color System. It is based on mathematical
descriptions of light sources, objects, and a standard observer. Light sources are specified
by their relative energy distributions, objects are specified by their reflectance (or transmission) spectra, and the observer is specified by the CIE standard human observer
tables. For color analysis, the light reflected (or transmitted) from (or through) an
object is measured with a spectrophotometer. Since in most cases, the reflection is
diffuse, it is essential to use a spectrophotometer with an integrating sphere so that all
of the light being reflected is sampled, not just that at some narrow angle. For the
most accurate data, reflectance measurements are taken at each wavelength and the
values over the range 380 to 770 nm would be used in the summation. For most purposes,
the accuracy is sufficient using 16 measurements at 20-nm intervals from 400 through
700 nm.
To identify the color resulting from the interaction of a light source, an object, and a
standard observer, one uses the data for these three dimensions to calculate the tristimulus
values X, Y, and Z using Eqs. 19.6a, b, and c.
X¼
Y¼
Z¼
X770
(19:6a)
X770
(19:6c)
x E R
380 l l l
X770
y E R
380 l l l
z E R
380 l l l
(19:6b)
Tristimulus values are different for the same object and the same observer when the energy
distribution E of a different illuminant is used. This is as it should be, since we know that
colors change as the light source changes. The tristimulus values uniquely and unequivocally define colors. For example, X ¼ 14.13, Y ¼ 14.20, and Z ¼ 51.11 is a definitive
description of a color, but what color? Unfortunately, even experts often cannot say by
looking at the numbers. This set of tristimulus values is for a blue, but few people
could look at them and tell you it is a blue, much less whether it is a grayish-blue and
approaches being a purplish-blue, which it happens to be.
The X and Y tristimulus values can be converted to chromaticity values x and y by
normalization:
x¼
X
XþY þZ
y¼
Y
XþY þZ
(19:7)
If one has a metameric pair of coated panels, the tristimulus values and the chromaticity
values are the same with the light source under which the panels match. They are not the
same, however, if calculated with the energy distributions from another light source. When
the light source changes, the X, Y, and Z and the x and y values of both panels change, but
the changes are to different degrees.
The chromaticity values of each wavelength of the spectrum can be calculated and
plotted against each other to form the CIE spectrum locus (Figure 19.14). The ends of
the locus are connected by a straight line called the purple line. There are no purples in
19.7. COLOR SYSTEMS
399
Figure 19.14. CIE chromaticity diagram showing the location of various hues. (From Ref. [1], with
permission.)
the spectrum; in CIE color space, hues of purple lie along this line. As shown, the plot can
be divided into color areas, so one can look at the x and y values and have a reasonable idea
of the shade of a color. The third dimension rises vertically from the plot; it is the Y tristimulus axis, luminance; Y ¼ 100 (or 1, in some conventions) at the x, y values of the light
source. The Y at the spectrum locus line approaches 0. At x, y points between the spectrum
locus and the point of the source, Y is always less than 100. As the value of Y gets larger,
the gamut of possible colors narrows (Figure 19.15).
If one draws a line from the point of the source through the point of a sample extended
out to the spectrum locus, the wavelength at the intercept is called the dominant wavelength of the color. This dimension corresponds to the hue dimension in the Munsell
System, but the scale is different. If this extrapolation intercepts the purple line, the line
is extrapolated in the opposite direction, and the intercept with the spectrum locus is
called the complementary dominant wavelength. If one divides the distance from the
source point to the sample point by the total distance from the source point to the spectrum
locus (some workers express it as a percentage), one obtains the purity. Purity is the same
dimension as chroma in the Munsell System, but the scale is different. The vertical
Y dimension is a gray scale, as in the Munsell value dimension, but again, the scale is
different.
400
COLOR AND APPEARANCE
Figure 19.15. Topographical diagram of a three-dimensional CIE color space with illuminant C.
(From Ref. [1], with permission.)
Figure 19.15 shows a topographical diagram of three-dimensional CIE Color Space
with illuminant C. All real colors that can be seen with light source C fall inside this
color space. In the Munsell system, the boundaries are limited not by reality of color,
but by the color purity of the available pigments to make the reference chips. CIE
color space is not visually uniform (Figures 19.14 and 19.15). For example, a small
difference in x and y in the blue part of the color space represents a substantial difference
in color, whereas in the green part of color space, the same difference in x and y values
represents little difference in color. In this sense, Munsell color space is preferable, since
the differences are visually uniform. However, color calculations of the types needed for
instrumental color matching are feasible with CIE color space, but not with Munsell
space. If the comparisons are to be visual, one commonly uses the Munsell system. If
the comparisons are to be mathematical and include all possible color space, one uses
the CIE system.
There have been many attempts to transform CIE color space mathematically to a visually equal color space; then differences could be used as meaningful measures for specification purposes. Some progress has been made, and color differences (DE) can be
calculated using the CIE 1976 L a b equations. L represents lightness, a redness –
greenness, and b yellowness– blueness. These quantities are calculated from tristimulus
19.8. COLOR MIXING
401
values by the following equations:
Y
16
Yn
X
Y
a ¼ 500 f
f
Xn
Yn
Y
Z
b ¼ 200 f
f
Yn
Zn
L ¼ 116
In these equations, X, Y, and Z are the tristimulus values of the sample; Xn, Yn, and Zn
are the tristimulus values of the light source; f(Y/Yn) 2 (Y/Yn)1/3 for values of (Y/
Yn) . 0.008856; f(Y/Yn) ¼ 7.787(Y/Yn) þ 16/116 for values of Y/Yn equal to or less
than 0.008856; and the functions f(X/Xn) and f(Z/Zn) are similarly defined.
The classical equation for CIELAB color difference is
DE(L a b ) ¼ ½(DL )2 þ (Da )2 þ (Db )2 1=2
These equations still do not represent fully uniform color space. If specifications are
written for a line of colors specifying a fixed + range for DE for the entire series, the requirements will be more stringent for some colors than for others. Even if color difference
equations were available that were visually uniform, their use would still be limited for specifications. The color would be permitted to vary equally in any direction from the central
standard. However, it is common for people to be more concerned about deviations in one
direction in color space than in other directions. For example, there is commonly a greater tolerance for whites to be off in the blue direction than for them to be off in the yellow direction.
CIE has developed a series of improvements over the CIELAB color-difference
formula, most recently CIEDE2000 [11]. Implementation of this system requires a
22-step calculation [12]: easy with computers but beyond the space limits of this book.
A recent evaluation of 128 pairs of samples concluded that the “CIEDE2000 formula outperformed CIELAB and it is also more accurate than a panel of observers” [13].
19.8. COLOR MIXING
There are two types of color mixing: additive and subtractive. In additive mixing, the primary
colors are red, green, and blue. Additive mixing is involved in theatrical stage use of overlapping colored spot lights, as well as in color television, where three colors of dots (red, green,
and blue) are projected near each other on the screen. The lights from nearby dots are “added”
when we look at them to give colors, which depend on the ratio of the three colors in nearby
dots. In additive color mixing, equal amounts of blue and green light give blue–green (cyan);
similarly, blue and red light give purple (magenta); green and red light give yellow; and equal
addition of all colors gives white light. With appropriate light sources, lights of all colors can
be made. Note that we see as yellow the combination of all wavelengths in the green, yellow,
orange, and red segments of the spectrum.
In almost all cases, however, color mixing encountered in coatings is not additive but
subtractive mixing. We use colorants that absorb (subtract) some of the light of some
wavelengths from white light. If we add a second colored pigment to a coating containing
402
COLOR AND APPEARANCE
Figure 19.16. Transmission (or reflectance) spectra of cyan, magenta, and yellow colorants, together
with their complementary colors.
colored pigment, we subtract more; we cannot add back any intensity of wavelengths by
using an additional further colorant. The primary colors in subtractive mixing are cyan,
magenta, and yellow. If we mix equal quantities of an ideal cyan and an ideal magenta,
the result is blue. That is, the cyan absorbs red and the magenta absorbs green, so blue
remains. Similarly, cyan and yellow yield green, magenta and yellow yield red, and a
mixture of equal amounts of all three ideal absorbing colorants absorbs all light and
yields black. Idealized transmission (or reflectance) spectra of the primary colors,
which illustrate their complementary nature, are provided in Figure 19.16.
19.9. COLOR MATCHING
A large fraction of pigmented coatings is color matched. The customer originally chose
some color for a refrigerator, automobile, or other product, and a coating formulator was
given a sample of a material with this color and was asked to match the color using a
coating formulation appropriate for the end use. After the formulator establishes the appropriate ratio of the appropriate pigments and after the customer approves the color match, the
factory must color match batch after batch of the coating to meet this standard. Sometimes
the formulator develops the formula first, and the color is accepted by the customer.
19.9.1. Information Requirements
Before starting the initial laboratory color match, the color matcher needs a substantial
amount of information.
1. Metamerism. Is a spectral (nonmetameric) match possible? That is, can exactly the
same colored pigments (including white and black) be used to establish the match as were
19.9. COLOR MATCHING
403
used to make the customer’s sample? If not, the customer must realize that any match will
be metameric; that is, the colors will match under some light source but not under others.
For example, if the sample is a dyed fabric, the color of the textile cannot be matched
exactly under all lights by any coating, since the colorants cannot be identical. If, as
another example, the customer has been using a coating made with one or more pigments
containing lead compounds and wishes to have a lead-free coating, only a metameric
match is possible.
2. Light sources. If the match is to be metameric, the customer and the supplier must
agree on the light source(s) under which the color is to be evaluated. If there is to be more
than one light source, a decision should be made as to whether it is more desirable to have
a close match under one light source without regard to how far off the match might be
under other light sources, or to have a fair color match under several light sources but
not a close match under any individual light source.
3. Gloss and texture. The color of a coating depends on its gloss and texture. Some of
the light reaching the eye of an observer is reflected from the surface of the film and some
from within the film. The color seen by the observer depends on the ratio of the two types
of reflected light. At most angles of viewing, more light is reflected from the surface of a
low gloss coating than from the surface of a high gloss coating. It is impossible to match
the colors of a low gloss and a high gloss coating at all angles of viewing. There must be
agreement as to the gloss of the coating, and if the gloss of the standard is different from
the gloss desired for the new coating, the angles of illumination and viewing must be
agreed on. It is impossible to make even a metameric match of the color of a fabric
sample with a paint at all angles of viewing, since both the colorants and the surface textures have to be different. When paint manufacturers advertise on TV that their paint stores
can do this, they are engaging in misleading advertising.
4. Color properties. Colorants have to be chosen that permit formulation of a coating
that can meet the performance requirements. Need the coating have exterior durability,
resistance to solvents, resistance to chemicals such as acids and bases, resistance to
heat, meet some regulation for possible toxicity, and so on?
5. Film thickness and substrate. Since in some cases, the coating will not hide the substrate completely, the color of the substrate can affect the color of the coating applied to it,
and the extent of that effect will vary with the film thickness. This variable is particularly
important in applications such as can coatings and coil coatings, for which relatively thin
coatings are applied (,25 mm in the case of can coatings). A thin coating that was color
matched over a gray primer will not match over a red primer. A coating set up for one coat
application on aluminum will not match the color standard if it is applied over steel.
6. Baking schedule. Since the color of many resins and some pigments are affected by
heating, particularly at high temperatures, the color of a coating can be affected by the time
and temperature of baking. It is also necessary to specify what the color requirements are
for overbaking.
7. Cost. The color matcher should know the cost limitations. There is no sense in making
an excellent color match with pigments that are too expensive for a particular application.
8. Tolerance. How close a color match is needed? Coatings for exterior siding or automobile top coats require very close color matches. For many others, close matching is
unnecessary, but some customers, in their ignorance, set tight tolerance limits. Overly
tight tolerances raise cost without benefiting performance. For coatings that are going
404
COLOR AND APPEARANCE
to be produced over time with many repeat batches, the most appropriate way to set color
tolerances is to have an agreed on set of limit panels. For example, for a deep yellow
coating, these would be the limits of greenness and redness, brightness and darkness,
and high and low chroma. Since panels can change with age, spectrophotometric measurements should be made of the standard and the limit panels, and CIE tristimulus values calculated. The customer would then agree to accept any batch of coating having a color
within this volume of color space. As mentioned earlier, while DE color differences, calculated from the 1976 CIELAB Color Difference Equations, have been used to set color
tolerances, they have limitations. It is undesirable to use them as a general statement of
tolerance for a series of colors. Furthermore, they permit variation equally in any direction
from the standard color, which is frequently undesirable.
19.9.2. Color Matching Procedures
There are two approaches to matching colors: visual and instrumental. In visual color
matching, experienced color matchers examine the sample to be matched, and from
their experience select a combination of pigment dispersions that they think will permit
matching the color. A formulation is mixed and the coating applied; since gloss affects
color, the gloss must first be compared with the standard and adjusted if needed. The
color matcher looks at the resulting color in comparison to the standard and decides
what further addition of either one or more of the same pigment dispersions, or perhaps
dispersions of different pigments, are needed to adjust the color match.
Spectrophotometric curves of the original standard sample can be used in an analytical
mode to help identify component colorants in the standard, thereby simplifying pigment
selection for visual matching. Organic pigments can often be identified by examining
the absorption spectra of solutions of colorants from a sample of coating [14]. The
process continues until a satisfactory combination of pigment dispersions has been
selected. Color matching must be based on dry films. It can be useful to compare
samples of wet coatings to estimate progress toward a match, but since color changes considerably during application, film formation, and drying, decisions have to be based on dry
films prepared under conditions approximating the way in which the coating will actually
be used. When matching a new color, even a highly experienced color matcher seldom
matches a color in less than three hits, and commonly, it takes several more hits, sometimes requiring changes in one or more of the original choices of pigments.
The visual color matcher not only has the responsibility of selecting the colorants and
their ratios to make a color match, but he or she must also attempt to make the color match
so that it will be as easy as possible to produce batch after batch efficiently in the factory. It
is not economically feasible for pigment manufacturers to produce successive batches of
color pigments that are exactly the same color and color strength. Furthermore, there will
be batch to batch variation in dispersions made from these pigments. The formulation
should be set up so that it can be adjusted in the factory to permit matching the color standard despite these variations. It is desirable to use four pigments (counting black and
white, but not inerts) to make the original match. This provides the four degrees of
freedom necessary to move in any direction in three-dimensional color space. Sometimes
use of single pigments cannot be avoided, but it is undesirable especially in applications in
which close color matches are needed. Sometimes, more than four pigments are needed,
but this adds to the complications of production and should be avoided, if possible. The
19.9. COLOR MATCHING
405
factory should not make any changes in the pigments in the formulation, which would
result in a metameric match.
Visual color matching is a highly skilled craft requiring years of apprenticeship to
master. It is still practiced and still important, but the trend is toward computerized instrumental color matching. Instrumental color databases, along with computer programs, can
be used to select colorants and their ratio both for original color matches in the laboratory
and to provide information as to the amount of the various pigment dispersions to be added
in the factory to match production batches. Establishing such a program requires a major
effort. The database must be set up carefully. The pigments must be made into dispersions;
and multiple single color coatings with different concentrations of color pigment dispersion in a series of appropriate white coatings must be made and applied. The reflectance
values must be measured at 16 wavelengths; for critical matches, values at 35 wavelengths
may be needed. (See Ref. [10] for further discussion of the creation of the pigment
database.)
Analysis of computer color matching is beyond the scope of this book. References [10]
and [15] (and the general references cited) review computer color matching and provide
detailed discussions. Solid color coatings at complete hiding can be accurately color
matched by current computer color matching software. The basis for these programs is
the Kubelka –Munk theory outlined in, Figure 19.10. These programs attempt to
achieve a tristimulus match between the sample and standard or object to be color
matched. The mathematics of this operation is complicated by the nonlinear nature of
the relation between the colorant concentrations in the sample and the tristimulus
values of the sample and standard.
Typical color measurements consist of either 16 or 35 reflectance measurements at a set
of wavelengths in the visible region. To utilize these data for color matching, the reflectances must be corrected for boundary reflections by means of Saunderson’s equation to
give Kubelka – Munk reflectances, Figure 19.10. These reflectances can be used to calculate the ratio K/S for the sample to be matched at each wavelength. The K and S values are
assumed to be linear functions of the pigment concentrations chosen to match the sample.
In the case of a four pigment match:
K ¼ k 1 c1 þ k2 c2 þ k3 c3 þ k4 c4
S ¼ s1 c1 þ s2 c2 þ s3 c3 þ s4 c4 ,
((K=S )s1 k1 )c1 þ ((K=S )s2 k2 )c2 þ þ ((K=S )s4 k4 )c4 ¼ 0
for each of the 16 (35) wavelengths, since the measured K/S ratio and ki and si values for
the individual pigments are different at each wavelength. In addition, the pigment concentrations are usually expressed as percentages to satisfy the equation
c1 þ c2 þ c3 þ c4 ¼ 100
It would seem easy to select three wavelengths to give three equations that combined
with the last equation would give four linear equations with four unknowns that on solution would give the required pigment concentrations to match the sample. Unfortunately,
this is not the case in practice. Even using all 16 (35) K/S equations and using various
estimates such as tristimulus weighted averaging or least squares solutions of the determined set of 17 (36) equations in four unknowns does not in most cases give, a close
406
COLOR AND APPEARANCE
tristimulus color match. Allen devised a method that is used by most commercial software
to obtain a tristimulus match [16]. Starting from an initial estimate of pigment concentrations obtained using the K/S ratios and tristimulus weighting, these concentrations
are refined iteratively to approximate a tristimulus match.
The DX value is the difference between the measured X value of the object to be
matched and the X value of the match calculated from the initial concentrations. The
partial derivatives of X with respect to concentration are calculated from the initial
match concentrations. The values for Y and Z are calculated in the same fashion. The
details of these calculations can be found in the cited references. The resulting four
equations are solved for the four unknown Dci values, and a new set of match concentrations are calculated by adding the Dci values to the initial estimate of ci values used
to set up these equations. The process is repeated until a tristimulus match or no further
improvement in the match is obtained. Usually, the best tristimulus match with the four
chosen pigments is obtained in less than three iterations.
If the four colorants chosen are close to or identical with the colorants in the object to be
matched, a nonmetameric match results. If the degree of metamerism as judged by the
variation in color difference between the object and match for two or more illuminants
is unacceptable, different four pigment combinations can be tried until an acceptable
match is attained.
Matching of metallic colors has been more difficult to computerize because the colors
have to match at multiple aspecular angles, angles with differences between the angle of
illumination and the angle of viewing. The reflectances of the samples and all of the colorants in the database must be determined at multiple angles. Instruments are now available
that measure at a variety of aspecular angles. The angle of illumination is fixed and the
angle of viewing can be set at different angles. In the United States, the accepted aspecular
angles are 158, 458, and 1108. (See ASTM E-12.12, Standard Practice for Multiangle
Measurement of Metal Flake Pigmented Materials.)
Matching colors with pearlescent flakes, interference, and cholesteric pigments is
further complicated by changes in hue as a function of angle of illumination and angle
of viewing (Section 20.2.5). Instruments are now available that have three illuminating
angles (i.e., 258, 458, and 758) with viewing angles for each of the illuminating angles
of þ158 and 2158, but research continues, and more elaborate geometries may be
needed [17].
Other information on the colorants, such as cost, properties, regulatory restrictions, and
so on, can be built into the database. The computer can then calculate a series of alternative
colorant combinations selected for lowest cost, least degree of metamerism, excluding pigments with inadequate exterior durability, and so on. The formulator can then choose the
most appropriate combination for the particular color-matching assignment. The formulator then makes up an experimental batch of paint, applies it to the appropriate substrate,
bakes or dries it, and measures the reflectance values. The result is seldom a satisfactory
color match on the first attempt. Reflectance values of the first hit are compared instrumentally with those of the standard, and the computer calculates the amounts of the various
colorants needed so that the color will match. The number of hits required to make
color matches has been found to be significantly less using computer color matching
than the number required by even experienced visual color matchers.
The same program is used in production of factory batches of colored coatings. The
colorants to be used have already been selected and must not be changed. The laboratory
formulation is used for the initial mix, but a fraction of each of the colorants is held back.
19.10. GLOSS
407
(Otherwise, if the factory batch of some pigment dispersion were stronger than standard,
too much of that colorant would be put in.) The batch is mixed, a panel is coated, and its
reflectance is measured. The data are then used to calculate what additions have to be
made to the batch to obtain a color match. If necessary, the process is repeated. The
savings in time and hence in cost using computer color matching can be large, but the
cost of establishing and maintaining the database is substantial. Since the colors of
pigment dispersions vary from batch to batch of pigment and of pigment dispersion,
measurements of each batch usually have to be made for comparison with the standard
database values. Computer programs can be written to correct for small differences in
the colors from batch to batch.
19.10. GLOSS
Gloss is a complex phenomenon. Individuals frequently disagree on the visual assessment
of gloss and generally do not have in mind a clear definition of gloss, partly because there
are several types of gloss. Specular (mirror-like) gloss is the type most often considered in
coatings. A high gloss surface reflects a large fraction of the light that is reflected from the
surface at the specular angle, that is, the angle of reflected light equal to the angle of the
incident light beam. Lower gloss surfaces reflect a larger fraction at nonspecular angles.
When considering gloss, people visually compare the amount of light reflected at the
specular angle with the amounts reflected at other angles. If the contrast in reflection is
high, gloss is said to be high. Note that gloss is not, as seems to be commonly
assumed, directly related to the fraction of light reflected from the surface. The fraction
of light reflected at a surface increases as the angle of illumination increases
(Figure 19.5). At most angles of illumination, surface reflection from a low gloss
surface is greater than from a high gloss surface. Surface reflection at the specular
angle increases as the refractive index of an object increases (Figure 19.5).
If a surface is rough on a micro scale, the angle of incidence of a beam of light is not the
same as the geometric angle of the surface with the light beam. Light is reflected at specular angles between the light beam and individual rough facets of a surface. If a surface has
many small facets oriented at all possible angles, a beam of light is reflected in all directions. Such a surface is a diffuse reflector and has a low gloss; it is called matte or flat. Flat
is unfortunate terminology, since a perfectly smooth (flat) surface gives high gloss,
whereas a microscopically rough surface gives what we call a flat surface. At intermediate
surface roughness, the gloss is intermediate and is called semigloss, eggshell, satin, or a
variety of other terms.
A phenomenon related to specular gloss is distinctness-of-image (DOI) gloss. A perfect
specular reflector is a perfect mirror with an image that exactly mimics the original. If a
surface has perfect diffuse reflection, no mirror image can be seen. At intermediate
stages, the image is more and more blurred as the ratio of specular to diffuse reflection
decreases. Furthermore, larger surface irregularities lead to distortion of the image. Commonly, one sees some degree of both blurring and distortion.
Sheen, as the term is used in the coatings industry, refers to reflection of light when a
low gloss coating is viewed from near a grazing angle. A high gloss coating reflects a high
fraction of light whose angle is near grazing. Reflection from most low-gloss surfaces is
low at a grazing angle. A low gloss coating is said to have a high sheen if there is significant reflection at a grazing angle. The effect is easy to see but difficult to describe; there is
408
COLOR AND APPEARANCE
not a glare (as when a glossy surface is illuminated and viewed at a grazing angle) but a
“soft,” relatively high reflection.
Luster is another type of gloss effect. (To illustrate the problem of defining gloss, a dictionary definition of luster is: gloss, sheen.) Luster is directional gloss. For example, some
woven fabrics are much glossier when viewed parallel to the warp than when viewed parallel to the woof. Such fabrics exhibit variations in gloss looking at folds in a draped fabric.
If the contrasts are great, the fabric is said to have a high luster. Similar effects can sometimes be observed in textured coatings.
Haze can be said to be a form of gloss. When light enters a hazy film, it is scattered to
some degree, causing some diffuse reflection to reach the eye of the observer. The result is
similar to having some of the light reflected at nonspecular angles at the surface. The contrast between the fractions of light reflected at specular and nonspecular angles is reduced.
In pigmented coatings, it may be difficult to distinguish visually between gloss reduction
resulting from haze and from scattering of light by a pigment. It is always desirable for a
formulator to make a batch of coating with no pigment to check the clarity of a dried,
cross-linked, pigment-free film.
Bloom gives a similar effect to haze. If a liquid component of a coating film is not
soluble in the resin binder, it can separate from the body of the film in small droplets.
These can come to the surface, making it uneven, diffusing light beams striking the
surface and reducing gloss. In contrast to haze, bloom can be wiped off a surface with a
cloth damp with solvent for the blooming material; commonly, bloom reappears.
19.10.1. Variables in Specular Gloss
Some of the variables of specular gloss can be visualized by use of schematic diagrams
based on hypothetical measurements by an idealized goniophotometer. A goniophotometer is an instrument in which a beam of light can be directed at a surface at any
angle of incidence and can measure the amount of light reflected at any angle in the
same plane as the incident beam. A schematic representation of an idealized instrument
is shown schematically in Figure 19.17. Figure 19.17(a) shows a cross section through
the plane of the beam of light. Figure 19.17(b) shows a schematic view looking down
on the instrument. The drawing shows that the only light detected by the photometer is
that reflected in the plane of the beam of light. No instrument can measure light reflected
at all angles in all planes. Also, there is no way to build an instrument that would measure
light reflected at 908. In this idealized instrument, the beam of light is narrow enough so
that only a point on the surface is illuminated, and the photometer can detect a reflected
beam with a beam width approaching zero.
Figure 19.17. Simplified representation of a goniophotometer.
19.10. GLOSS
409
Figure 19.18. Schematic representation of the photocurrent reflected as a function of angle of viewing.
(From Ref. [18], with permission).
All light reflected from a perfect mirror is at the specular angle and in the same plane as
the beam of light. From a perfect diffuse reflector, light is reflected at all angles in all
planes. Line S0 in Figure 19.18 shows photocurrent as a function of angle of viewing
from a perfect diffuse reflector illuminated at 458, as measured by an ideal goniophotometer [18]. At first glance, the results do not seem rational, but it must be recalled that
the photometer can measure all the light reflected at 08 and only a fraction approaching
zero of the light reflected in all planes at 908.
Also shown in Figure 19.18 is the idealized response curve u for a semigloss white
coating. Comparison of these two curves illustrates three factors important in considering
specular gloss. First is the relative brightness A of the sample compared to that of the
diffuse reflector. Here A is the ratio of light intensities in the perpendicular direction
u0/S0. Second is height-of-gloss, h, which is calculated from the peak height, p:
h¼
P u45
p
A
¼
S45
S45 100
(19:8)
Third is the distinctness-of-image aspect of gloss d: as defined by Eq. 19.9, in which P is
the total area under the sample curve, U is the area of diffuse background reflection from
the sample, and S is the area under the standard diffuse reflector curve.
d¼
h
F
F¼
PU
S
(19:9)
If A is large, an observer will consider the gloss of a sample to be lower than that of a
sample with lower A. If h is small, the gloss of a panel will be lower than that of a sample
with higher h. It is instructive to compare a white gloss coating with a black gloss coating.
A is larger and h is smaller for the white coating because white scatters light that enters a
film, resulting in a high extent of diffuse reflection. On the other hand, a black coating
absorbs almost all the light entering its film, so there is little diffuse reflection from
within the film. If the surfaces of the two films are equally smooth, the surface reflection
410
COLOR AND APPEARANCE
is equal, but the black coating will be seen as glossier, even though it reflects less of the
incident light. It is impossible to make a white coating with as high a gloss as that of a high
gloss black coating. The possible gloss of colored coatings is intermediate; the darker the
color, the higher the potential gloss. We saw in Section 19.2.4 that gloss affects color; now
we see that color affects gloss. If the value of d is high, the surface acts as a mirror, giving a
clear image. As the value of d decreases, there is more and more blurring of the specular
image.
One must be careful when comparing these values with observer evaluation of gloss. As
has already been stated, different observers rate the gloss of pairs of panels differently.
Intensity of reflection (peak height) is not a major factor in an observer’s evaluation of
gloss; the principal factors are the contrast between reflection at the specular angle and
the distinctness of image. Gloss is also affected by the distance between the observer
and an object. An observer who is close enough to resolve irregularities on the surface
visually will say that a surface is a rough, high gloss surface. On the other hand, one
observing the same surface at a greater distance, which does not allow visual resolution
of the irregularities, will say that it is a smooth, low gloss surface.
The major factor controlling gloss of coatings is pigmentation. As a film shrinks with
volatile loss, pigment particles can cause irregularities in the surface. Roughness of the
surface varies with the ratio of pigment volume concentration (PVC) to the critical
pigment volume concentration (CPVC) in the dry film (Section 22.1). (The effect of pigmentation on gloss as solvent evaporates is discussed in Ref. [19].)
In solventborne high gloss coatings, the pigment concentration in the top micrometer or
so of a dry film contains little, if any, pigment. This pigmented-depleted layer results from
the motions within a film as solvent evaporates. Initially, convection currents are set up
within the film; and both resin solution and dispersed pigment particles move freely. As
solvent evaporates, the viscosity of the film increases and movement of pigment particles
is slowed. Movement of resin solution continues longer, so the top surface contains little
pigment. Application of a clear top coat over a pigmented base coat can increase gloss.
As PVC increases, the amount of pigment at the surface increases, decreasing gloss.
Particle size of the pigment affects gloss; if aggregates of pigment are not broken up in
the dispersion process, gloss will be lower. Flocculated pigment systems have a lower
CPVC, so at the same PVC there will usually be a greater likelihood of low gloss.
However, since large particles stop moving before small ones, flocculated particles will
stop moving sooner than well stabilized ones, which can lead to increased gloss in
low-PVC coatings. (The effects of pigment particle size and clear-layer thickness on
specular gloss are discussed in Refs. [20] and [21].)
In some coatings (e.g., furniture lacquer), it is desirable to have a low gloss but still a
high degree of transparency. This is accomplished using a small quantity of very fine particle size silicon dioxide as a pigment. The combination of small particle size and low
refractive index difference results in minimal light scattering as long as concentration is
low. When solvent evaporates from such a lacquer, the SiO2 particles keep moving
until the viscosity of the surface of the film becomes high. The result is a higher than
average concentration of pigment in the top of the film, reducing gloss at relatively
low PVC.
Poor leveling can reduce gloss. If the irregularities are large, as is commonly the case
with brush marks, the surface is seen as a glossy, but wavy surface. However, if the irregularities are small, as is commonly the case with orange peel, gloss will be rated as lower.
An irregular surface can result from poor leveling over a smooth surface but can also be
19.10. GLOSS
411
the result of applying a coating over a rough substrate. It has been shown that smoothness
and gloss are affected by the extent of roughness of a substrate, film thickness, and viscosity [22]. (See Section 24.2 for further discussion of leveling.)
Wrinkling affects gloss. If the surface of a film cross-links before the lower layers of the
film, wrinkling is likely to occur (Section 24.6). When the lower layers cross-link, shrinkage occurs, causing the top of the film to pucker up in wrinkles. If the size of the wrinkles is
small enough, gloss will be low, but looked at in a microscope, the surface can be seen to
be an irregular glossy surface.
Since the fraction of light reflected at a surface increases as refractive index
increases, gloss of high gloss coatings tends to increase as refractive index of a coating
increases, since the contrast between reflection at the specular angle and other angles is
greater. Refractive index differences between different binders are small, so although
detectable, this effect is generally small compared to surface roughness [17,18]. Refractive
indexes of pigment can also have an effect.
Latex coatings generally exhibit lower gloss than solventborne coatings at equal levels
of pigmentation in dry film (Section 32.3). There are several reasons for this:
1. Latex coatings have both resin and pigment particles as dispersed phases. During
drying of a latex paint film, there is not the same opportunity for separation to give a
pigment-depleted thin layer at the top of the film as in a solvent coating. Latexes with
smaller particle size give somewhat higher gloss films than do larger particle size
latexes. Some difference in segregation may occur, since large particles stop moving
first as a film dries.
2. Pigment-free dry films of many latexes are not transparent; they are hazy, reducing
gloss. The haze is due to the presence of dispersants and water-soluble polymers that are
not completely soluble in the latex polymer.
3. Fairly commonly, surfactants bloom to the surface of a latex coating, reducing gloss.
4. It is generally more difficult to achieve good leveling with latex coatings. Surface
roughness such as brush marks reduce gloss. The effect of brush marks depends on how
close an observer is to the surface. If one is close enough to resolve the brush marks visually, the coating can look like a wavy, gloss surface. From a greater distance, the gloss
appears to be low.
Gloss can change during the life of an applied film. In some cases the surface of the film
embrittles and then cracks as the film expands and contracts. Generally, this mechanical
failure is progressive, and after initial loss of gloss, there is film erosion. In other cases,
especially clear coats, erosion occurs first, and loss of gloss is evident only after
erosion is deep enough to cause protrusion of pigment particles of the base coat.
Erosion of binder in pigmented films can proceed to a stage in which pigment particles
are freed from binder on the surface and rub off easily; this phenomenon is called chalking.
Chalky surfaces have dramatically lower gloss, and the color therefore, changes to a
lighter color. Loss of gloss can also result from loss of volatile components after a film
is exposed; this causes film shrinkage and increases surface roughness [23]. (An excellent
review of durability and gloss appears in Ref. [23].)
In some cases, the gloss of a low gloss surface increases with use. It is fairly common
for the gloss of flat wall paints to increase if rubbed, for example, near a light switch. This
increase in gloss is called burnishing.
412
COLOR AND APPEARANCE
19.10.2. Gloss Measurement
No fully satisfactory method for measuring gloss is available, and no satisfactory rating
scale for visual observation of gloss has been developed. Although all people will agree
as to which film is glossier if the gloss difference is large, they frequently disagree in
ranking if the difference is small. Even the same observer experiences difficulty in consistently ranking a series of panels if the differences in gloss are small.
Caution must be exercised in using instruments that measure gloss. In Section 19.10.1
we discussed an idealized goniophotometer. In practice, one has to accept less than this
ideal. The beam diameter cannot actually approach zero because there must be sufficient
light intensity reflected from all angles to give a measurable response on available photodetectors. In real instruments there is a light source shining on a slit aperture at a fixed
distance from a sample surface. The reflected light goes through another slit aperture at
a fixed distance to a photodetector.
Specular gloss meters are widely used, but correspondence between meter readings and
visual comparisons is limited. Such instruments give significantly different readings with
differences in intensity of reflected light, but observers are relatively insensitive to such
differences. Furthermore, the aperture of the slit in a gloss meter is about 28, whereas
the limit of resolution of a human eye is about 0.00058 of arc [19]. A gloss meter is therefore less sensitive to distinctness-of-image than the eye (when the observer is close to the
object). The distance between the aperture and a panel is fixed in a gloss meter, whereas a
person can view a panel from any distance.
The most widely used gloss meters, also called reflectometers, are simplified goniophotometers in which one measures a response only at the specular angle. Those most
commonly used in the coatings industry can make measurements when the angles of incidence and viewing are 208, 608, and 858. A schematic drawing is shown in Figure 19.19.
The first step for using a gloss meter is to calibrate the instrument with two standards: one
with high gloss and the other with a lower gloss. If the second standard does not give the
standard reading after the instrument is set with the first standard, something is wrong;
most commonly, one or both of the standards is dirty or scratched. Other possible problems
include panel misalignment, deterioration of the light source, or a malfunction of the
photometer. One must use the standard that has been calibrated at the angle selected.
Black and white standards are available. Reflection at the specular angle is not the
Figure 19.19. Gloss meter. (From Ref. [18], with permission.)
19.10. GLOSS
413
same from a white and a black standard with equal surface roughness (Section 19.10.1). It
is not feasible to have standards of all different colors; white standards are used for light
colors and black standards for dark colors. (One should always report which set of standards is used.) Standards are standardized by NIST using a research goniophotometer,
as described in Ref. [24]. In contrast to gloss meters, such research instruments allow
both angles of illumination and viewing to be varied independently; but they are expensive
and relatively difficult to maintain.
In normal practice, one first measures at 608. If the reading obtained is over 70, readings
should be made at 208 rather than at 608 since precision is higher nearer the midpoint of the
meter reading. It is common to read low gloss panels at both 608 and 858. Readings at 858
may have a relationship to sheen. It is essential to report the angle at which readings are
made. Multiple readings should be taken and analyzed statistically. This reduces the probability that local surface irregularities or dirt particles are affecting the meter reading. It
has been said that slight brush marks do not affect readings if the direction of the brush
marks is parallel to the plane of incidence. If results are to be compared between two laboratories, it is essential to check the compatibility of their instruments. This is best done by
measuring at least three black and three white standards on each instrument. Readings are
reproducible on carefully calibrated instruments to +3%, but it is better to express error in
terms of units, such as +2 gloss units [25]. This is a high percentage of error in the low
gloss range.
Gloss meters give results in numbers between 0 and 100, but there is confusion as to
what the numbers mean. Commonly, though undesirably, they are expressed as percentages. It is better to call them gloss units or just meter readings. They are not, as some
people seem to believe, the percentage of light reflected at the surface. They are closer
to being the percentage of light reflected at that angle compared to the reading that
would be obtained if a perfectly smooth surface were measured. As noted previously,
the total reflection from a black matte surface is much higher at most angles of illumination and viewing than that from a high gloss black surface. The point can be emphasized
by comparing meter readings of the same panels at 208 and 608. Meter readings are lower
for the same panel when the setting is 208 than when it is 608 [26] (Figure 19.20).
Figure 19.20. Transition of gloss readings according to Freier [26]. (From Ref. [18], with permission.)
414
COLOR AND APPEARANCE
With sufficient care, the instruments can be used for quality control comparisons of
lots of the same or very similar coatings and for following loss of gloss on aging. For
specification purposes other than quality control, specular gloss meters are not appropriate, and one must rely on standard visual panels. The customer selects three panels of
each color, one with the desired gloss and the others representing the upper and lower
acceptable limits.
Figure 19.20 also illustrates a procedure to follow loss of gloss when a high gloss panel
is exposed to exterior weathering. Initial readings were taken at 208, then after some
exposure, at both 208 and 608, then for panels that have shown considerable loss of
gloss only at 608. Different scale units in the 208 and 608 readings (left) provide for a continuous slope at the transition point. The right scale designates gloss grade as observed
visually.
Distinctness-of-image (DOI) meters rely on using the sample as a mirror (Figure 19.21).
The reflection of a grid on the surface of the panel is compared visually to a set of photographic standards ranging from a nearly perfect mirror reflection to a blurred image in
which the grid cannot be detected. One reports a comparison of the degree of blurring
and also a qualitative statement about distortion. Correspondence with visual assessment
in the high gloss range is better than that with specular gloss meters. Instruments are
available to make comparisons based on optical density rather than relying on visual
comparisons.
Newer instruments with linear diode array detectors are being made which permit multiple measurements of light reflected at small angle increments without using an aperture
in front of the detector [27]. Computerized instruments make multiple measurements of
reflectance from small areas (approximately 100 mm in diameter) over a 10 cm2 area.
This permits separation of the reflection from micro roughness and macro roughness,
thus giving a numerical rating for gloss and a separate measurement of macro roughness,
such as orange peel or texture.
Figure 19.21. DOI meter. (From Ref. [18], with permission.)
REFERENCES
415
GENERAL REFERENCES
ASTM, ASTM Standards on Color and Appearance Measurement, 3rd ed., ASTM, Philadelphia, PA,
1991.
Burns, R. S., Billmeyer and Saltzman: Principles of Color Technology, 3rd ed., Wiley, New York,
2000.
Hammond H. K., III; Kigle-Boeckler, G., Gloss, in Paint and Coating Testing Manual, 14th ed.,
Koleske, J. V., Ed., ASTM, Philadelphia, PA, 1995, pp. 470 – 480.
Hunter, R. S.; Harold, R. W., The Measurement of Appearance, 2nd ed., Wiley, New York, 1987.
Kuehni, R. G., Color: An Introduction to Practice and Principles, 2nd ed., Wiley, Hoboken, NJ,
2004.
Pierce, P. E.; Marcus, R. T., Color and Appearance, 2nd ed., Federation of Societies for Coatings
Technology, Blue Bell, PA, 2003.
Westland, S.; Ripamonti, C., Computational Colour Science Using MATLAB, Wiley, Hoboken, NJ,
2004.
Zorll, U., Prog. Org. Coat., 1972, 1, 113.
REFERENCES
1. Billmeyer, F. W., Jr.; Saltzman, M., Principles of Color Technology, 2nd ed., Wiley, New York,
1981.
2. Billmeyer, F. W., Jr.; in Encyclopedia of Chemical Technology, 3rd ed., Wiley, New York,
1979, Vol. 6, pp. 523 – 548.
3. Wyszecki, G.; Stiles, W. S., Color Science, 2nd ed., Wiley-Interscience, New York, 1982.
4. Greenstein, L. M., in Pigment Handbook, 2nd ed., Lewis, P. A., Ed., Wiley-Interscience,
New York, 1988, Vol. 1, pp. 829 – 858.
5. Judd, D. B.; Wysecki, G., Color in Business, Science, and Industry, 3rd ed., Wiley, New York,
1975.
6. Mitton, P. B., in Pigment Handbook, Patton, T. C., Ed., Wiley, New York, 1973, Vol. 3,
pp. 289 – 339.
7. Fields, D. P.; et al., Surf. Coat. Int., 1993, 76, 87.
8. Wicks, Z. W., Jr.; Kûhhirt, W., J. Paint Technol., 1975, 47(610), 49.
9. Cremer, E., Prog. Org. Chem., 1981, 9, 241.
10. Pierce, P. E.; Marcus, R. T., Color and Appearance, Federation of Societies for Coatings Technology, Blue Bell, PA, 1994.
11. CIE, Publ. 142-2001, Central Bureau of the CIE, Vienna, Austria, 2001.
12. Sharma, G.; et al., Col. Res. Appl., 2005, 30(1), 21.
13. Luo, R.; et al., Abstr. Interim Meet. Int. Color Assoc., Porto Allegre, Brazil, 2004, http://
www.fadu.uba.ar/sicyt/color/aic2004lect.htm.
14. Kumar, R.; et al., J. Coat. Technol., 1985, 57(720), 49.
15. Rich, D. C., J. Coat. Technol., 1995, 67(840), 53.
16. Allen, E., J. Opt. Soc. Am., 1974, 64, 993.
17. Rodrigues, A., Abstr. Interim Meet. Int. Color Assoc., Porto Allegre, Brazil, 2004, http://
www.fadu.uba.ar/sicyt/color/aic2004lect.htm.
18. Zorll, U., Prog. Org. Coat., 1972, 1, 113.
19. Braun, J. H., J. Coat. Technol., 1991, 63(799), 43.
416
20.
21.
22.
23.
24.
25.
26.
27.
COLOR AND APPEARANCE
Simpson, L. A., Prog. Org. Coat., 1978, 6, 1.
Braun, J. H.; Fields, D. P., J. Coat. Technol., 1994, 66(828), 93.
Ngo, G. D.; et al., J. Coat. Technol., 1993, 65(821), 29.
Braun, J. H.; Cobranchi, D. P., J. Coat. Technol., 1997, 67(851), 573.
Nadal, M. E.; Thompson, E. A., J. Coat. Technol., 2001, 73(913), 73.
Huey, S., Off. Dig., 1964, 36, 344.
Freier, H.-J., Farbe Lack, 1967, 73, 316.
Smith, K. B., Surf. Coat. Int., 1997, 80, 573.
20
Pigments
Pigments are insoluble fine particle size materials used in coatings for one or more of five
reasons: to provide color, to hide substrates, to modify the application properties of a
coating, to modify the performance properties of films, and/or to reduce cost. Pigments
are divided into four broad classes: white, color, inert, and functional pigments. Pigments
are insoluble materials used as colloidal dispersions. Dyes are soluble colored substances;
they are used in only a few specialized coatings, such as stains for wood furniture
(Section 31.1.1). Some pigments are called lakes. The original meaning of lake was a
dye that had been converted into a pigment by irreversible adsorption on some insoluble
powder. The term lake is now sometimes used when a colored pigment is blended with
an inert pigment. When the pigment is essentially all color pigment, it is sometimes
called a toner.
Particle size affects color strength, transparency or opacity, exterior durability, solvent
resistance, and other properties of pigments. For any given pigment, the manufacturer
selects the most appropriate compromise for the particle size and designs the process to
produce that average particle size consistently. The manufacture of most pigments
involves precipitation from water; the process conditions determine the particle size.
Many pigments are surface treated during or after precipitation. The precipitated
pigment is filtered and the filter cake is dried. During drying, pigment particles become
cemented together in aggregates. The coating manufacturer generally receives pigment
from the pigment manufacturer as a dry powder of aggregates and must disperse these
aggregates to break them up to their original particle size and make a stable dispersion
(Chapter 21).
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
417
418
PIGMENTS
20.1. WHITE PIGMENTS
A large fraction of coatings contain white pigment. White pigments are used not only in
white coatings, but also in a substantial fraction of other pigmented coatings to give lighter
colors than would be obtained using color pigments alone. Furthermore, many color pigments give transparent films, and the white pigment provides a major part of the hiding
power of the coating. As discussed in Section 19.2.3, the ideal white pigment would
absorb no visible light and would have a high scattering coefficient. Since a major
factor controlling scattering efficiency is the difference in refractive index between the
pigment and the binder, refractive index is a critical property of a white pigment.
Reference [1] is a monograph on white pigments.
20.1.1. Titanium Dioxide
The most important white pigment used in coatings is TiO2. Purchases of TiO2 by U.S.
coatings companies in 2002 were $868 million [2]. The technology of TiO2 pigments is
reviewed in Refs. [1] and [3]. The factors affecting the hiding power of TiO2 are discussed
in Refs. [1] and [4]. Two different crystal types are used: rutile and anatase. Rutile is used
in larger volume primarily because it gives about 20% greater hiding power than anatase;
the average refractive index of rutile is 2.73, compared to 2.55 for anatase. On the other
hand, rutile absorbs some violet light, whereas anatase absorbs almost no visible light
(Figure 20.1). Whereas a coating pigmented with only rutile looks white, it looks yellowish alongside a coating pigmented with anatase.
The color of rutile TiO2 coatings can be adjusted by tinting. If one adds a small amount
of a dispersion of a violet pigment, such as carbazole violet, the white coating obtained is
less yellow. The violet pigment strongly absorbs wavelengths other than violet and
Figure 20.1. Reflectance spectra of rutile and anatase and anatase TiO2 coatings. (From W. A.
Kampton in Ref. [5a], Vol. I, p. 2, with permission.)
20.1. WHITE PIGMENTS
419
reduces the difference in reflection of the violet wavelengths compared to the others. This
makes the coating a light gray, but unless there is a higher reflectance white present for
comparison, it is seen as white, not gray. At the same time, there is the added advantage
that hiding power is increased due to the additional light absorption. Phthalocyanine blue
is also used as a tinting pigment. Phthalocyanine blue does not give as “white a white” as
carbazole violet, but it is less expensive and gives slightly improved hiding. The most
common tinting pigment used in whites is carbon black. Carbon black is less expensive
than either violet or blue pigments, and gives a greater increment to hiding. Since
carbon black absorbs all wavelengths of light, it does not reduce the yellowness of the
white but provides a lower cost method of increasing hiding. “Transparent TiO2” (i.e.,
very fine particle size TiO2) is used in metallic automotive coatings to give a frosted effect.
In white pigmented UV cure coatings, the lower absorption of near UV radiation by
anatase TiO2 is an advantage over rutile, since there is less interference with absorption
of UV radiation by photoinitiators. Another difference between rutile and anatase TiO2
is in their photoreactivity. Anatase pigmented coating films fail more quickly on
outdoor exposure by chalking than do rutile pigmented films. Chalking is the formation
of loose particles on the exposed coating surface resulting from the erosion of binder
by photodegradation (Section 5.2.4). Chalking is usually undesirable, although selfcleaning white house paints are deliberately formulated with anatase TiO2, so they
chalk readily. When the surface of the film erodes away, dirt accumulation is also
removed, so that the paint surface stays whiter.
In most exterior applications, chalking must be minimized, and rutile TiO2 is used.
However, although rutile is less photoreactive than anatase, it is still sufficiently reactive to reduce exterior durability. The problem is minimized by treating the surface of
the TiO2 particles during pigment manufacture [1,3]. The most common surface
treatments are silica (SiO2) and/or alumina (Al2O3); a variety of other metal oxides
are also used with silica and/or alumina. A typical surface treatment for exterior
grades of rutile is about 6 wt% silica and 4 wt% alumina. An ideal surface treatment
would be a continuous shell on all TiO2 particles to prevent contact between the TiO2
and the binder. Perfection has not been achieved, but the best chalk-resistant grades of
rutile with a shell of SiO2 show little acceleration of degradation. The best treated
anatase pigments accelerate chalking more than do treated rutiles, but much less
than does untreated anatase pigment. Since TiO2, especially rutile, absorbs UV strongly
(Figure 20.1), it competes with the binder in a coating for UV absorption. If chalkresistant grades are used, the reduced UV absorption by the binder tends to enhance
exterior durability.
Surface treatment also affects other properties of the pigments. Alumina treatments
improve dispersibility in solventborne coatings. However, alumina can partly neutralize
acid catalysts, such as those used with coatings containing Class I MF resins, reducing
cure rates (Section 11.3.1). Specially treated TiO2 pigments permit production of more
stable dispersions in waterborne coatings. TiO2 dispersions in water, called slurries, are
used in the manufacture of latex paints (Section 21.3). The slurries are manufactured by
the TiO2 producer and shipped in tank cars or trucks to replace dry pigment. The price
on a pigment basis is about the same, but the slurry is a pumpable liquid, and aggregation
of pigment particles is minimized, so there are substantial savings in material handling and
dispersion costs.
A variety of TiO2 pigments is offered by each TiO2 manufacturer. Formulators need to
be aware of the differences among the types. Some are designed for maximum exterior
420
PIGMENTS
durability, some to provide maximum gloss, some are more easily dispersed in water
systems, and others are more easily dispersed in solvent systems. The TiO2 content of
the various types of pigments varies from as high as 98% to as low as 75%. Generally,
the hiding varies with TiO2 content, so the amount of TiO2 needed for hiding varies
with pigment type. Another consideration is the range of particle sizes; the broader the
range, the larger the fraction of particles that provide less than optimum hiding.
Coating performance problems and economic waste have resulted from use of the
wrong type of TiO2 and from indiscriminate substitution of one type for another.
20.1.2. Other White Pigments
Until the late 1930s, white lead [basic lead carbonate, 2PbCO3 . Pb(OH)2] was widely
used as a white pigment. Its relatively low refractive index, 1.94, gives low hiding
power, less than one-tenth that of rutile TiO2. When TiO2 became available, the use of
white lead dropped rapidly. White lead is somewhat soluble in water and presents a
toxic hazard. Many children suffer from lead poisoning attributed to eating paint containing white lead that peels from the walls and woodwork in older buildings. As a result, the
lead content of any paint sold to retail consumers through interstate commerce in the
United States is limited to 0.06% of the dry weight. Lead poisoning should not automatically be assumed to result from lead paint. There are other possible sources of lead
exposure, including contaminated dirt.
Zinc oxide, ZnO, was formerly used as a white pigment, but because of its low refractive index, 2.02, it cannot compete for hiding power with TiO2. As discussed in Section
20.4, ZnO is used in exterior house paints as a fungicide and in some can linings as a
sulfide scavenger. About 15 million pounds of ZnO was used in coatings in 2004 [4].
Zinc oxide should not be used in primers because it is somewhat water soluble and can
therefore induce blistering by osmotic pressure when water vapor penetrates through
the top coats into the primer coat. Other white pigments, important at one time but
largely obsolete, are zinc sulfide (ZnS), with a refractive index of 2.37, and lithopone,
ZnS/BaSO4.
Small air bubbles scatter light and are another white “pigment.” In some cases, cost
reductions are possible by taking advantage of the scattering by air bubbles in coating
films. The refractive index of air, 1.00, gives a refractive index difference of about 0.50
in a typical binder. Although this difference is small compared with the difference of
1.25 between binder and rutile TiO2, it still provides some light scattering, and hence
hiding. A common way of incorporating air bubbles in a coating film is by use of such
high levels of pigmentation that after the solvent evaporates, there is not enough resin
to adsorb on all the pigment surfaces and to fill all the interstices between the pigment particles. The result is air voids, which increase hiding. Such coatings are said to be formulated at pigment volume levels above the critical pigment volume concentration (CPVC)
(Chapter 22). Since interfaces between TiO2 and air are developed, the very large refractive index difference of 1.75 is especially effective in increasing hiding. However, the
voids make the film porous and reduce the protective properties of the films. In ceiling
paints, advantage is taken of the added hiding, since requirements for other properties
are not as important.
Particles containing air bubbles provide added hiding from the air– resin interface
without imparting porosity. For example, high Tg latexes containing water trapped
within the particles are available and are finding increasing use in latex paints as a
20.2. COLOR PIGMENTS
421
partial replacement for TiO2 [7]. The latexes are prepared by sequential emulsion
polymerization [8]. Initially, a low Tg latex copolymer of acrylic acid and acrylic esters
is prepared; the particles are swollen with water, especially at alkaline pHs. These core
particles are encapsulated in shells of high Tg cross-linked polymer such as a styrene – divinylbenzene copolymer. When the paint is applied and the water evaporates, the conventional latex particles of the binder coalesce to form the film, but the high Tg latex
particles do not coalesce. Water diffuses out of the particles, leaving air bubbles within
the particles, which add to the hiding.
20.2. COLOR PIGMENTS
A wide variety of color pigments is used in coatings. References [6] and [9] provide
detailed discussions of their chemistry, properties, economics, and uses, and Ref. [10]
gives a more condensed coverage of organic pigments. First, it is appropriate to summarize
the considerations involved for selecting color pigments for a particular coating
application.
1. Color. The first criterion is the color of the pigment. Pigment suppliers provide technical bulletins that contain color chips showing the color obtainable with each pigment.
Generally, there are two or three color chips for each pigment. The mass tone is displayed
by a color chip in which the pigment is used as the sole pigment in the coating. Then there
are one or two additional color chips showing the color obtained when the pigment is used
together with TiO2 in different ratios; these are called tint colors. In some cases, color
chips showing the effect with added aluminum flake pigment (Section 20.2.5) are also
provided.
2. Color strength. Some colorants are strong and others weak, corresponding to the
absorption coefficients and particle sizes of the pigments. It may well be more economical
to use a strong, expensive (i.e., high cost per unit volume) pigment than a weak, lower cost
pigment.
3. Opacity or transparency. Depending on the end use, it may be desirable to use a
pigment that increases hiding by both scattering and absorbing radiation, or it may be
important to select pigments that scatter little, if any, light in the coating film so that a
transparent color can be obtained.
4. Ease of dispersion. Some pigments are more easily dispersed than others. If everything else were equal, one would select the most easily dispersed pigment. Many pigments
are surface treated by the manufacturers to enhance their ease of dispersion. (See Ref. [11]
for a review.) Many types of surface treatment have been used, including adsorption on the
surface of a molecule with a structure similar to that of the pigment, but with polar groups
to enhance dispersibility; monolayers of polymers; and more recently, layers of inorganic
oxides on the surface of organic pigments [12]. Pigment dispersion is a complex phenomenon involving several factors; it is discussed in detail in Chapter 21.
5. Exterior durability. Some pigments are more sensitive than others to photodegradation, which leads to loss of color or change of hue. Pigment manufacturers provide data
that can be useful for screening pigments that may be appropriate for use outdoors.
However, there can be significant variation in exterior durability, depending on the combination of pigment and resin. Therefore, for critical applications, the durability of specific
422
PIGMENTS
formulations must be determined. (See Chapter 5 for discussion of and testing for exterior
durability.) A related property is light fastness. This term is usually applied to the ability of
pigments to resist fading when used indoors. ASTM classes pigments in five light fastness
grades, I to V. Only those classed I or II are suitable for applications where long-term color
stability is critical, such as in artist paints.
6. Heat resistance. Coatings that are baked or exposed to high service temperatures
require heat resistant pigments. A few pigments (e.g., yellow iron oxides) undergo chemical changes with heat. A more common problem is that some organic pigments slowly
sublime at elevated temperatures.
7. Chemical resistance. In use, many coatings are exposed to chemicals, most often to
acids and bases, and the pigments must resist color change under those conditions. For
example, automobiles are exposed to acid rain, and home laundry machines are
exposed to alkaline detergents.
8. Water solubility. In most applications, pigments with any significant water solubility
are best avoided because they leach out of a film, resulting in loss of properties. The presence of water-soluble substances in a primer can lead to blistering of the film when water
permeates through it. In latex paints, water-soluble materials can reduce the stability of the
dispersion of the latex particles. The stability of colloidal aqueous dispersions can be particularly affected by the presence of low concentrations of polyvalent salts. On the other
hand, pigments that help protect against corrosion by passivation (Section 20.4) must be
somewhat soluble in water to be effective (Section 7.4.2).
9. Solvent solubility. Partial solubility of pigments in some solvents can be a problem.
For example, if one is making a red coating for a bicycle that is to have a white stripe on
the red coating, the red pigment must not dissolve in solvents used in the white striping
coating. If the red pigment is somewhat soluble, the stripes will become pink. Such pigments are said to bleed.
10. Moisture content. Most “dry” pigments adsorb small to moderate amounts of
water. The adsorbed water can cause serious problems with water-reactive binders such
as polyisocyanates.
11. Toxic and environmental hazards. Filter face masks should always be worn when
handling dry powder pigments; it is dangerous to inhale fine-particle dusts, even if they are
chemically inert. In most cases, there is little or no hazard once the pigment is incorporated
into a coating. However, toxicity, even in a coating film, can be important. For example, as
mentioned previously, white lead is no longer permitted in architectural coatings in the
United States. Lead-containing pigments are also prohibited in coatings for children’s
toys and furniture. Zinc chromate pigments are generally thought to be human carcinogens. In some countries, their use in coatings is prohibited; regulations can be expected
to become increasingly restrictive. Another factor that can affect pigment choice is the
increasing cost of disposal of materials containing heavy metals.
12. IR reflectance. Inorganic pigments that reflect, rather than absorb, infrared radiation are available in a variety of colors. They are known as complex inorganic color pigments and as ceramic colorants. Many have excellent outdoor durability, and when used
on building roofs and siding, their IR reflectance significantly reduces solar heating, saving
energy in hot climates. Combinations of color pigments that absorb all visible wavelengths
give black; thus, these pigments can be formulated to make gray paints with much less IR
absorption than grays made with carbon black pigments.
20.2. COLOR PIGMENTS
423
13. Cost. Cost is important, but as noted earlier, one cannot tell which pigment is the
least expensive by just looking at the price per unit weight. The critical question is: How
much will the final coating cost on a volume basis? The answer is affected by color
strength and density as well as price.
Purchases of inorganic color pigments by U.S. coating companies in 2002 totaled $327
million; those of organic pigments were $453 million [2].
20.2.1. Yellow and Orange Pigments
20.2.1.1. Inorganic Yellows and Oranges Iron oxide yellows, FeO(OH), are low
chroma brownish-yellow pigments. They give opaque films with good hiding and high
exterior durability; chemical and solvent resistances are excellent. The pigments are generally easily dispersed and are comparatively inexpensive. When heated above 1508C,
they gradually change color to a low chroma red because they dehydrate to form iron
oxide red, Fe2O3. Most iron oxide yellows are made synthetically, but some natural ore
ochre pigments are used. In some cases, particularly with the natural pigments, the presence of soluble iron and other metal salts can affect the stability of coatings, which
cure by a free radical mechanism. Extremely fine particle size iron oxide pigments are
also available; they are used when transparency is required. This effect of particle size
on opacity is an excellent example of the relationship between light scattering and particle
size (Section 19.2.3).
Chrome yellow pigments are bright, high chroma yellows. Medium chroma yellow pigments are predominantly lead chromate (PbCrO4). Greenish-yellow colorants, called
primrose yellow and lemon yellow, are co-crystals of lead chromate with lead sulfate.
Redder yellows, that is chrome oranges, are co-crystals of lead chromate with PbO.
Still redder oranges, molybdate oranges, are co-crystals of lead chromate with lead
molybdate (PbMoO4) and lead sulfate. Chrome yellows are relatively low in cost, although
higher than iron oxide, at least partly because their density is higher. Chrome yellows discolor on exterior exposure to form lower chroma yellows, but their exterior durability is
adequate for many outdoor uses. They are bleed and heat resistant. Due to their lead
content, their use is not permitted in consumer paints in the United States. Their use in
industrial applications is declining because of concern about the use of lead compounds.
Their major current use is in yellow traffic striping paint. Use of chromates for this purpose
has been banned in many European countries and in some U.S. states. Monoarylide
yellows are their predominant replacement.
Titanium yellows are made by introducing other metal ions into the lattice of anatase
TiO2 crystals, followed by calcining to convert to the rutile crystal structure.
Greenish-yellow shades are based on introducing antimony and nickel; reddish-yellow
shade grades contain antimony and chromium. Their resistance to exterior exposure,
chemicals, heat, and solvents is excellent. However, only relatively weak yellow colors
can be produced, so the cost is high, and the range of colors that can be made is restricted.
Bismuth vanadate is a yellow pigment with high brilliance and good durability that has
recently been introduced.
20.2.1.2. Organic Oranges and Yellows Chemical structures of representative
pigments are shown in Figure 20.2. The code designations are from the Color Index
424
PIGMENTS
Figure 20.2. Examples of organic yellow pigments.
system developed by the Society of Dyers and Colourists in the United Kingdom and the
Association of Textile Chemists and Colorists in the United States. The P in the code
stands for pigment, the next letter designates the hue, and the number is assigned
chronologically.
Diarylide yellows are bisazo pigments derived from 3,30 -dichlorobenzidene (e.g., PY
13), as shown in Figure 20.2. They have high color strengths and high chroma. The hue
and photostability are controlled by the number, positions, and structure of substituent
groups on the aromatic ring of the anilide portion of the molecule. Even the most photochemically stable diarylide yellow pigments fade on exterior exposure, especially when
used in tints. On the other hand, their solvent, heat, and chemical resistance are excellent.
Due to their high strength and low density, their cost is relatively low. They are used in
interior coatings, for which a bright yellow color is needed in tints, and in such applications as coatings for pencils. Grades with excellent transparency can be made. Diarylide
yellows are the major yellows used in printing inks, especially in yellow inks used in fourcolor process printing.
Monoarylide (monoazo) yellow pigments such as PY 74, also have high chromas. They
have poor bleed resistance and sublime when exposed to high temperatures. However,
their light fastness is better than that of diarylide yellows, although still inferior to that
of inorganic yellow pigments. Some opaque grades have sufficient light resistance for
use in outdoor coatings. They are replacing chrome yellows in traffic paints.
20.2. COLOR PIGMENTS
425
Nickel azo yellow is a relatively weak, very greenish yellow; in fact, as can be seen
from the designation PG 10 in Figure 20.2, it is classified as a green. Nickel azo yellow
exhibits excellent exterior durability and heat resistance. It gives transparent films and
is used predominantly in automotive metallic coatings. There have been some situations
in which bleeding into stripes has been reported, so specific applications must be
checked for bleeding.
Vat yellow pigments, as exemplified by isoindoline yellow PY 139, give transparency
with excellent exterior durability and heat and solvent resistance. They are expensive and
used only when their outstanding properties are required (e.g., in automotive metallic
coatings).
Benzimidazolone orange pigments offer excellent light fastness and resistance to heat
and solvent. They are used as replacements for molybdate orange.
20.2.2. Red Pigments
20.2.2.1. Inorganic Reds Iron oxide (Fe2O3) gives the familiar barn red color. It is a
low chroma red with excellent properties and low cost. In contrast to iron oxide yellows,
iron oxide reds are thermally stable. When their particle size is optimal for scattering, they
provide a high degree of hiding. An orange– red iron oxide is now available. There are also
very fine particle grades available that provide transparent films. The excellent exterior
durability makes the transparent grades suitable for use with aluminum in metallic automotive top coats. Transparent iron oxide is an excellent UV absorber and is used in transparent wood stains to protect the wood from photoxidation.
20.2.2.2. Organic Reds Toluidine red pigment, PR 3, is a moderate cost, bright red
azo pigment with high color strength, good exterior durability in deep colors, good chemical resistance, and adequate heat resistance to permit use in baking enamels. As shown in
Figure 20.3, it is an azo derivative of b-naphthol. Toluidine red is soluble in some solvents
and gives coatings that are likely to bleed.
Bleed resistance with azo pigments can be achieved by the presence of carboxylic acid
salts. For example, 2-hydroxy-3-naphthoic acid (BON) can be coupled with diazo compounds. Permanent Red 2B is an example of such bleed resistant high chroma red azo
pigments. It is available as the calcium, barium, or manganese salt. The somewhat
higher cost manganese salt shows better exterior durability than the calcium or barium
salts. A variety of related azo pigments that have somewhat different shades is also available. This general class is the largest volume of organic red pigments used in coatings and
inks. However, many are sensitive to bases and are not suitable for some latex paints.
Naphthol reds are a large family of azo pigments with various ring substituents (Cl,
OCH3, NO2, etc.) in the generic structure shown in Figure 20.3. They are more resistant
426
PIGMENTS
Figure 20.3. Example of organic red pigments.
to bases, soap, and acid than are the permanent reds, and have fairly good exterior durability and solvent resistance.
Quinacridone pigments are nonbleeding, heat and chemical resistant, and give outstanding exterior durability, even in light shades. However, their cost is high. Depending
on substitution and crystal form, a variety of orange, maroon, scarlet, magenta, and violet
colors are available. Large particle size grades are used when opaque pigments are needed,
and fine particle size grades are available for use in metallic automotive top coats. Surface
treatments are applied to increase the resistance of their dispersions to flocculation [13].
Many other high performance red pigments are available (see Refs. [6] and [8]).
20.2.3. Blue and Green Pigments
20.2.3.1. Inorganic Blues and Greens Iron blue, ferric ammonium ferrocyanide,
FeNH4Fe(CN)6, is an intense reddish shade blue with fairly good properties. Historically,
it was important as one of the first widely used synthetic pigments. Since the 1930s it has
been supplanted increasingly by phthalocyanine blues, which have greater color strength.
Co-crystals of various ratios of iron blue and chrome yellow are called chrome greens. The
popularity of these pigments has decreased because of the high color strength of phthalocyanine green; they can no longer be used in architectural coatings in the United States
because of their lead content.
20.2.3.2. Organic Blues and Greens The principal blue and green pigments are
copper phthalocyanine pigments, commonly called phthalo blue and phthalo green
(Figure 20.4). They exhibit outstanding exterior durability, bleed and chemical resistance,
are heat stable, and have high tinting strengths. Although their cost per pound is fairly
high, the high tinting strength and quite low density make their cost to use moderate.
Phthalo blues are available commercially in three crystal forms: alpha, beta, and the
seldom-used epsilon. The beta form has a relatively green shade of blue and is stable.
20.2. COLOR PIGMENTS
427
Figure 20.4. Representative phthalocyanine pigments.
The alpha form is redder, but not as stable; in some cases, when some grades of the alpha
form are used, there can be serious problems of change of color and strength during
storage of coatings or during baking. More stable alpha form pigments are available;
these incorporate various additives that stabilize the crystal form and minimize problems
with flocculation of dispersions. Some grades of phthalocyanine blue are slightly
chlorinated; these have greener blue shades.
Phthalo greens are made by halogenating copper phthalocyanine to produce mixtures of
isomers in which many of the 16 hydrogens of CPC have been replaced with chlorine or
mixtures of chlorine and bromine. Phthalo greens with 13 to 15 chlorines but no bromine
have blue – green shades. Partial replacement of chlorine with bromine shifts the color
toward a yellow – green shade. The yellowest shades have a high ratio of bromine to
chlorine, an example is PG 36, which has on average about 9 bromines and 3 chlorines.
20.2.4. Black Pigments
Almost all black pigments used in coatings are carbon blacks. They absorb UV radiation
as well as light, and with most binders, black colors are the most stable on exterior
exposure. Carbon blacks are made by a variety of processes of partial combustion
and/or cracking of petroleum products or natural gas. Depending on the process, the particle size and therefore the degree of jetness (intensity of blackness) varies. High color
channel blacks have the smallest particle sizes, with diameters of 5 to 15 nm, and
have the highest jetness. They are used when intensely black, glossy coatings are
desired. Furnace blacks have larger particle sizes, are lower in cost, and give less jet
black coatings. Various grades are available, with average diameters of 50 to 200 nm.
Lampblacks have a still larger particle size, on the order of 0.5 mm, and have lower
color strengths than other carbon blacks. They are primarily used in making gray coatings. They are preferable to high-color blacks for this purpose; if a small excess of
high color black is added by mistake, it may be necessary to increase the batch size
50% or more to make up for the mistake. The shade of gray obtained changes much
more slowly with the addition of lampblack, making color matching easier.
428
PIGMENTS
Approximately 30 million pounds of all types of carbon black was used in coatings in
2004, of which about 80% was furnace black [3].
Although all carbon blacks are predominantly elemental carbon arranged in polynuclear six-member rings, the chemical structure of the surfaces varies depending on
the raw material and process used in their manufacture. The surfaces are generally quite
polar and in some cases are acidic. There are also variations in the porosity. Black coatings
can present difficulties in formulation, especially when channel blacks are used. Due to the
small particle size, the ratio of surface area to volume of the pigment particles is very high.
The large surface area leads to adsorption of high ratios of resin on the pigment particles:
commonly, many times their volume of polymer, increasing their volume substantially and
hence giving high viscosity at relatively low pigment loading levels. Due to the polarity of
the surface and the large surface area, they can selectively adsorb polar additives such as
catalysts from a coating formulation. For example, in drying oil and oxidizing alkyd formulations, the metal salt driers slowly adsorb on carbon black, so that the coating dries
more and more slowly as it ages.
Acetylene black pigment increases the electrical conductivity of films. Conductivity is
required for coatings such as primers for plastic parts that are to be coated by electrostatic
spray (Section 31.3.2.2).
20.2.5. Metallic, Interference, and Cholesteric Pigments
The most important metallic pigments are aluminum flake pigments. They are produced
by milling finely divided aluminum metal suspended in mineral spirits in steel ball
mills to make thin flakes. Pigments are available in a variety of particle sizes. Two
major classes are used in coatings: leafing and nonleafing pigments. About 25 million
pounds of aluminum pigments with a value of $150 billion were used in North American
coatings in 2004 [3].
Leafing aluminum pigments are surface treated (e.g., with stearic acid) so that they have
a low surface tension. When a coating containing leafing aluminum is applied, flakes
orient at the surface, as a result of their low surface tension. This gives a bright metallic
appearance and acts as a barrier to permeation of oxygen and water vapor through the film.
As a result of barrier development, leafing aluminum pigments are used in top coats for
corrosion protection of steel structures (Section 7.3.3).
Nonleafing aluminum pigments have higher surface tensions and do not come to the
surface of the coating. They are used in metallic top coats for automobiles. As discussed
in Section 19.4, formulations using nonleafing aluminum are designed to maximize the
fraction of flakes that are oriented parallel to the film surface. Films made with transparent
color pigments and nonleafing aluminum change depth of color and shade with the angle
of viewing. Gold colored aluminum flakes can be made by vapor deposition of iron
carbonyl on the surface of aluminum flakes, followed by oxidation to generate a thin
layer of iron oxide [14]. Nonleafing aluminum pigments are surface treated to minimize
their sensitivity to acid in the environment.
The aluminum pigments used in water-reducible coatings are treated to minimize the
reaction of water with the aluminum (which generates hydrogen). In early work it was
necessary to mix in the aluminum paste just before application. Various treatments
have been developed to permit formulation of premixed coatings [15]; an example is treatment with a low acid number styrene –maleic acid copolymer.
20.2. COLOR PIGMENTS
429
Aluminum pigments are produced in solvent and sold as pastes so as to minimize problems of dusting and danger of explosion. But such pastes bring VOC into waterborne
coatings. Aluminum pigments have been developed that are pelletized with melamine –
formaldehyde resin surface coating and water stabilization treatments that can be
handled easily and stirred into water – glycol ether, which in turn can be added to waterborne coatings with minimal addition of VOC [16].
Bronze, nickel, and stainless steel pigments with platelets similar to those of aluminum
pigments are also available, but are used in lower volume. Bronze alloy flakes have “gold”
colors; depending on the alloy composition, shades from greenish yellow to reddish gold
are available. Generally, they are surface treated, so they leaf. Due to the presence of
copper, on outdoor exposure, bronze alloy pigments change color to a blotchy, muddy
green appearance.
Interference pigments are flake pigments that lead to color by interference (Sections
19.2.1 and 19.4). Pearlescent pigments are platelets of mica having thin surface treatment layers of TiO2 or Fe2O3, which serve to give interference reflection of light striking pigment surfaces [17,18]. As a result, at some spots on the surface of the platelets,
some wavelengths of light are strongly reflected and others are transmitted. At other
spots, where the film thickness of the treatment layer is different, different wavelengths
are reflected and transmitted. The result is a mother-of-pearl effect and an illusion of
optical depth. These pigments are used in automotive coatings to give an effect
related to that obtained with aluminum flake, but with the addition of hue changes
with angle.
Another type of pigment that gives colors as a result of interference effects is available [19]. The pigment particles are thin (about 1 mm in thickness) flakes composed of
a thin layer of reflective metal sandwiched between two layers of clear plastic, and with
very thin top layers of a semiopaque, very thin metal coating on each surface. As a
result of interference reflection (Section 19.4), certain wavelengths of light are reflected
and others absorbed. The color obtained depends on the film thicknesses and the angles
of illumination and viewing. As with other flake pigments, these flakes must be
oriented parallel to the surface of the coating. North American coating consumption
of pearlescent pigments in 2004 was about 4 million pounds, with a value of about
$75 million.
A newer type of interference pigment, cholesteric pigments, are based on polymeric
liquid crystal structures. One process for making these pigments starts with a combination
of an acrylated nematic liquid crystal and an acrylated chiral compound that forms helical
structures. They are dissolved in a vinyl monomer such as divinylbenzene along with a
photoinitiator. The solution is coated on a polyethylene support, linear stress is applied
to align the helixes, and the combination is UV cured. The film is removed from the
support, milled, and classified for particle size. The resulting flakes are a translucent offwhite highly cross-linked powder. The helical structure leads to interference, causing
changes in color when viewed at different angles. By changing the ratio of chiral additive
and liquid crystals, the wavelengths of light reflected and transmitted change. Five basic
colors are commercially available. They are sensitive to UV and must be used with a benzotriazole class UVA. In waterborne automotive base coats with UVA and an adequate
thickness of clear coat with UVA and HALS, they exhibit excellent durability [20].
A variety of other flake pigments for various applications are discussed in a review
paper [21].
430
PIGMENTS
20.3. INERT PIGMENTS
Inert pigments absorb little, if any, light and have refractive indexes close enough to those
of binders that they give little light scattering when used as pigments. Several synonymous
terms are used: inert pigments, inerts, fillers, and extenders. Commonly, but not always,
they are inexpensive and reduce the cost of a coating. The principal function of most
inert pigments is often to occupy volume in a film. Other functions are to adjust the rheological properties of fluid coatings and the gloss and mechanical properties of films. As
discussed in Chapter 22, many film properties are controlled by the volume of pigment
in the film. The multitude of inert pigments available is discussed in Refs. [6] and [22].
Calcium carbonate (CaCO3) pigments are widely used inerts. The lowest cost grades
are ground limestone or mixed calcium magnesium carbonate ore (dolomite). Synthetic
calcium carbonate pigments are more expensive, but they are whiter. Pigments are available with a variety of average particle sizes. In some applications, the reactivity of calcium
carbonate with acids makes carbonate pigments undesirable; especially in exterior paints,
degradation of film properties may be accelerated by acid rain. Calcium carbonate pigments should not be used in exterior latex paints. Water and carbon dioxide permeate
through the film of a latex paint; some calcium carbonate reacts to give calcium bicarbonate, which is water soluble and permeates back out of the film. On the film surface, the
water evaporates and the reaction reverses, leaving a frosting of insoluble calcium carbonate deposited on the film. Frosting is especially noticeable on dark color paints.
A wide range of clays (aluminum silicates) are used as inert pigments. They are available in various particle size ranges. Cost is frequently related to whiteness. As mentioned
in Section 20.4, bentonite and attapulgite clays are used to modify the viscosity of coatings. Mica (aluminum potassium silicate) has a platelet structure and can be useful in reducing permeability of films to oxygen and water vapor when it is oriented parallel to the
surface of the film.
Magnesium silicate minerals are also used as inert pigments. Talcs of various crystal
structures affect the film strength of coatings differently. Some talcs are platey and
reduce vapor permeability; others are fibrous and may be particularly effective in film reinforcing. Asbestos is a very fibrous magnesium silicate that is no longer used because it
causes lung cancer when fibers are inhaled.
Silicon dioxide pigments are an important group of inert pigments. Ground natural
silicon dioxide is used in a variety of particle sizes. An unusual example of an SiO2
pigment is diatomaceous earth, composed of fossil skeletons of diatoms. The large
ratio of surface area to volume affects film properties (Section 22.2). Very fine particle
size synthetic silicon dioxide pigments are used to reduce gloss of clear coatings and to
impart shear thinning flow properties to coatings, as discussed briefly in the next
section. They are expensive.
Barytes (barium sulfate) has been used as an inert pigment, especially in automotive
primer formulations. It is said to provide a “harder” primer. The density of barytes is
high (4.5), roughly twice that of most other inert pigments. In at least some cases, it is
possible that use of barytes is based on weight comparisons with other inert pigments,
rather than volume comparisons, as should be done. Although barytes is not expensive
on a weight basis, it is more expensive than most other fillers on a volume basis.
Although most inert pigments are inorganic minerals, organic materials can also be
used as inert pigments. For example, powdered polypropylene is insoluble and acts as
an inert pigment. High Tg latexes such a polystyrene latex can be used as an inert
20.4. FUNCTIONAL PIGMENTS
431
pigment in latex paints. Synthetic fibers such as aramid fibers have been shown to be effective for increasing the mechanical strength of coating films.
20.4. FUNCTIONAL PIGMENTS
Functional pigments are used to modify the application characteristics, appearance, or film
properties of coatings. An important example is corrosion-inhibiting pigments. Complex
zinc chromate pigments, strontium chromate, barium phosphosilicates, barium borosilicates, red lead, and zinc phosphate are used in primers to inhibit corrosion of steel by passivation of anodic areas. In contrast to other pigments, to function they must be somewhat
soluble in water. Because of toxic hazards, use of chromates and red lead is decreasing
rapidly. Zinc metal pigment provides corrosion protection by cathodic action and is
used in a class of primers known as zinc-rich primers. About 70 million pounds of zinc
pigment was used in North American zinc-rich primers in 2004 [3]. These pigments are
discussed in Sections 7.4.3 and 33.1.2, and considerable information is provided in
Refs. [6] and [23].
Flatting pigments are a type of functional pigment used to reduce gloss. Although low
gloss is often attained by formulating coatings with a high volume ratio of pigment in the
dry film, this is not always a desirable approach. For example, in lacquers for use on wood
furniture, it is essential to have the transparency of the dry film as complete as possible so
that the beauty of the wood grain is not concealed. Fine particle size silicon dioxide is
widely used in such coatings as a flatting pigment.
As discussed in Section 19.10.1, during solvent evaporation from the coating film, convection currents carry the fine particles to the surface, so that the pigmentation at the
surface of the dry film is high enough to reduce gloss. The combination of low total pigmentation, small particle size, and small refractive index difference gives little reduction
in transparency. Silane treatment of silica is reported to improve transparency in liquid
coatings, to provide excellent flatting with smooth surfaces, but to have little influence
on rheology [24]. Powdered polypropylene has been used in a somewhat analogous
fashion.
Some pigments are used as biocides. For example, zinc oxide is used as a fungicide.
Barnacles, algae, and other organisms can grow on the outer hulls of ships; to minimize
such growth, antifouling coatings are used. Cuprous oxide and organotin pigments have
been used in antifouling coatings; although their use is now constrained by environmental
concerns (Section 33.2). Finely divided silver is used in antimicrobial coatings.
Zinc oxide is used in linings for cans used for packing vegetables that evolve some
hydrogen sulfide during cooking, such as corn. The ZnO reacts with the H2S to form
white zinc sulfide. This prevents the deposition of black stains that result from the formation of tin sulfide by reaction between the H2S and the tin oxide layer on the tin
surface of the tin-plated cans.
Antimony oxide (Sb2O3) is a white pigment with a refractive index that is too low
(2.18) and a cost that is too high to use simply for hiding. It is used in fire-retardant coatings. When a combination of Sb2O3 and a chlorinated or brominated polymer is heated to a
high temperature, combustion by-products are generated that suppress flame propagation.
Another class of functional pigments is viscosity modifiers. In general terms, these pigments increase the low shear viscosity of coatings either to inhibit pigment settling during
storage of the coating and/or to reduce sagging after application of the coating. An
432
PIGMENTS
important example used in solventborne coatings is quaternary ammonium salt – treated
bentonite clay [25]. Another type of clay that is used in both solvent- and waterborne coatings is attapulgite clay (fuller’s earth), a fibrous clay. The fibrils of this clay associate, providing an increase in viscosity that reverses with agitation. In waterborne coatings, it also
acts by absorbing water to make swollen particles that distort in shape with stress. Fine
particle size silica has been used for many years as a viscosity modifier. The worldwide
market for flatting SiO2 pigments is reported to be 35,500 metric tons [24]. Fumed
silica (also called pyrogenic silicon dioxide) is a fine particle-size SiO2 that is useful to
increase the low shear viscosity of coatings. It is used in both solvent- and waterborne
coatings paint to give thixotropic coatings with reduced settling and with a thixotropic viscosity that allows application of thicker films without sagging. Surface treatment of SiO2
with polydimethylsiloxane gives a pigment that is in general most suitable in solventborne
coatings. In waterborne coatings it is most suitable in most cases to use the hydrophilic
untreated pigment [26]. The effect of such pigments on flow is discussed in Section 3.2.
Pigments with high infrared reflectance have been used for military camouflage.
Recently there is increased demand for infrared reflecting pigments for roofing and
siding coatings to reduce heat absorption and hence air conditioning costs. Aluminum
and white pigments have generally been used, but recently, brown, black, and brightly
colored ceramic pigments with high IR reflectance have been made available [3].
20.5. NANO-PIGMENTS
Nano-pigments are pigments of very small particle size. Some authors include pigments
with average particle diameters of less than 100 nm, but more commonly the term is
applied to diameters below 25 nm. The term is new, but nano-pigments have been used
in coatings for many years. For example, high strength channel blacks have particle diameters in the range 5 to 15 nm. Some clays can also be considered as nano-pigments, as they
can be separated (exfoliated ) into sheets on the order of 1 nm thick. A review paper on
nano-material technology applications in coatings is available [27].
Compared to the effects of conventional pigments, the use of nano-pigments, especially
nano-fillers, may improve resistance to scratch, mar, abrasion, heat, radiation, and swelling resistance, decrease water permeability, and increase hardness, weatherability,
modulus, and strain-to-failure while maintaining toughness. These properties reflect the
high surface area weight ratio of the particles. There is more surface to adsorb resin,
hence a larger fraction of immobilized resin. Based on an adsorbed layer of 10 nm, the
fraction of resin adsorbed on 300-nm pigment particles has been calculated to be 3%,
and on 50-nm particles, to be 22% [27]. In some cases, two Tgs can be detected. When
the particle size of the pigments is smaller than the wavelength of light, coatings pigmented with nano-pigments are transparent [28].
Dispersion of nano-pigments can be difficult. As particle diameters are reduced,
especially below 10 nm, the physical properties of the materials change, causing
various complications. As discussed in Chapter 21, conventional pigment dispersion
methods are often incapable of reducing particle diameters to the nano range, but some
successes have been reported [28]. In Section 21.4.4 we discuss the variables involved
in selecting mills, media, and media size; pin mills are recommended. Nano color pigments for automotive metallic coatings are prepared by grinding suspensions of the conventional color pigments in butyl acetate with a commercial dispersing agent using very
REFERENCES
433
small high-density balls in the mill. The mill base is then used to making tinting pastes
for color matching metallic colors [29]. Ball mills are used to disperse nano
(aluminum silicate) in an acrylic resin with a commercial dispersing agent. The dispersion
is used in making mar resistant clear coats [30].
It can also be difficult to make stable dispersions of nano-pigments. Due to their high
surface energy, nano-pigments have a strong tendency to agglomerate. If there is significant agglomeration, the advantages of nano-pigments will not be realized. Various
approaches have been utilized: surface treatment of the pigments, in situ polymerization
in the presence of the pigments, special pigment dispersants, and ultrasonic dispersion.
GENERAL REFERENCES
Braun, J. H., White Pigments, Federation of Societies for Coatings Technology, Blue Bell, PA, 1995.
Challener, C., Update on Inorganic Pigments, JCT Coat. Tech, 2005, 2(18), 44.
Herbst, W.; Hunger, K., Industrial Organic Pigments, 3rd ed., Wiley-Interscience, New York, 2004.
Lewis, P. A., Ed., Pigment Handbook, 2nd ed., Vol. I, Wiley-Interscience, New York, 1988.
Patton, T. C., Ed., Pigment Handbook, 3 vols., Wiley-Interscience, New York, 1973.
REFERENCES
1. Braun, J. H., White Pigments, Federation of Societies for Coatings Technology, Blue Bell, PA,
1995.
2. U.S. Department of Commerce, Bureau of Census, Current Industrial Reports: Paint and
Coating Manufacturing, http://www.census.gov.
3. Braun, J. H.; et al., Prog. Org. Coat., 1992, 20, 105.
4. Challener, C., JCT Coat. Tech., 2005, 2(18), 44.
5. Fields, D. P.; et al., Surf. Coat. Int., 1993, 76, 87.
6. (a) Patton, T. C., Ed., Pigment Handbook, 3 vols., Wiley-Interscience, New York, 1973. (b)
Lewis, P. A., Ed., Pigment Handbook, 2nd ed., Vol. I, Wiley-Interscience, New York, 1988.
7. Fasano, D. M., J. Coat. Technol., 1987, 59(752), 109.
8. Herbst, W.; Hunger, K., Industrial Organic Pigments, VCH, New York, 1997.
9. Vanderhoff, J. W.; et al., Polym. Mater. Sci. Eng., 1991, 64, 345.
10. Lewis, P. A., Organic Pigments, Federation of Societies for Coatings Technology, Blue Bell,
PA, 1995.
11. Hays, B. G., Am. Inkmaker, 1984, June, 28; 1986, October, 13; 1990, November, 28.
12. Bugnon, P., Prog. Org. Coat., 1996, 29, 39.
13. Jaffe, E. E.; et al., J. Coat. Technol., 1994, 66(832), 47.
14. Iden, R., J. Coat. Technol., 1995, 67(843), 57.
15. Muller, B., J. Coat. Technol., 1995, 67(846), 59.
16. Ferguson, R. L., JCT Coat. Tech., 2004, 1(7), 42.
17. Greenstein, L. M., in Pigment Handbook, 2nd ed., Lewis, P. A., Ed., Wiley-Interscience,
New York, 1988, Vol. I, pp. 829 – 858.
18. Chapman, D. P., J. Coat. Technol., 1996, 68(862), 19.
19. Droll, F. J., Paint Coat. Ind., 1998, 14(2), 54.
434
PIGMENTS
20. Nowak, P. J., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2004,
Paper 13.
21. Maile, F. J.; et. al., Prog. Org. Coat., 2005, 54, 150.
22. Solomon, D. H.; Hawthorne, D. G., Chemistry of Pigments and Fillers, Wiley-Interscience,
New York, 1983.
23. Smith, A., Inorganic Primer Pigments, Federation of Societies for Coatings Technology, Blue
Bell, PA, 1989.
24. Christian, H.-D., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2004,
Paper 17.
25. Memnetz, S. I.; et al., J. Coat. Technol., 1989, 61(776), 47.
26. Ettlinger, M.; et al., Prog. Org. Coat., 2000, 40, 31.
27. Fernando, R., JCT Coat. Tech., 2004, 1(5), 32.
28. Perera, D. Y., Prog. Org. Coat., 2004, 50, 247.
29. Vanier, N. R.; et al., U.S. patent 6,875,800 (2005).
30. Vanier, N. R.; et al., U.S. patent 6,916,368 (2005).
21
Pigment Dispersion
Manufacture of pigmented paints usually involves mixing liquid concentrates of dispersed
pigments with the other ingredients. In this chapter we address production of these concentrates. When they are made at the paint factory, they are commonly called mill bases, and
when they are purchased from specialist companies, they are called dispersions or
colorants.
As described in Chapter 20, pigments are manufactured with the particle size distribution that gives the best compromise of properties, but the particles often become cemented together into aggregates during processing. Breaking these aggregates and forming
stable dispersions of optimally sized pigment particles is a critical process in the manufacture of coatings. Making dispersions involves three aspects: (1) wetting, (2) separation,
and (3) stabilization. Most authors agree that there are three aspects to dispersion, but
different terms are used, sometimes with conflicting meanings. Be careful when reading
papers to determine how an author is using the terms.
21.1. DISPERSION IN ORGANIC MEDIA
In this section we discuss dispersion in organic media, usually solutions of resins in
organic solvent. Because of effects of the high surface tension and polarity of water,
pigment dispersion in aqueous media is treated separately in Section 21.3.
21.1.1. Wetting
Wetting is displacement of air, and sometimes of water or other contaminants, from the
surface of pigment particles and aggregates by the medium (i.e., the coating vehicle).
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
435
436
PIGMENT DISPERSION
Wetting is an essential requirement for pigment dispersion and requires that the surface
tension of the vehicle be lower than the surface free energy of the pigment. In organic
media this is the case for all inorganic and most organic pigments. If a pigment has an
especially low surface free energy, it is necessary to use a medium with even lower
surface tension. There can be important differences in the rate of wetting. When a dry
pigment is added to a vehicle to make a mill base, the pigment tends to clump up in clusters
of aggregates. For wetting to occur, the vehicle must penetrate through these clusters and
into the pigment aggregates. The rate of wetting is controlled predominantly by the viscosity of the vehicle; lower viscosity leads to more rapid wetting.
21.1.2. Separation
Processes are designed to separate pigment aggregates into individual crystals without
grinding crystals to smaller particle size; it is generally undesirable to decrease the
crystal size. Many different types of machinery are used to carry out the separation
stage. Some important examples are described in Section 21.4. Dispersion machines
work primarily by applying a shear stress to the aggregates suspended in the vehicle. Compressive forces resulting from impact may play a secondary role. If the aggregates are
easily separated, the machinery only needs to be able to exert a comparatively small
shear stress. If aggregates require a relatively large force for separation, machinery that
can apply a higher shear stress is required. Pigment manufacturers have been increasingly
successful in processing and surface treating pigments so that their aggregates are relatively easily separated.
Recall from Section 3.1 that shear stress is equal to shear rate times the viscosity of the
mill base. The shear rate available for any particular dispersion equipment is set by
machine design. The formulator must select appropriate dispersion machinery that can
transfer sufficient shear stress to the aggregates and formulate a mill base for the most efficient use of the selected machinery. For the fastest rate of separation of aggregates, the
mill base should have as high a viscosity as the equipment can handle efficiently. Then
the highest shear stress is exerted on the pigment aggregates, and separation is accomplished in minimum time. The engineering theory and equations modeling the forces
for separation are discussed in an excellent series of papers by Winkler, Dulog, and
co-workers [1].
21.1.3. Stabilization
Wetting and separation are important steps in making pigment dispersions, but it is seldom
a problem to carry out these two stages in organic solvent systems. On the other hand,
stabilization can frequently be a serious problem and is usually the key to making good
pigment dispersions. If the dispersion is not stabilized, the pigment particles will be
attracted to each other and undergo flocculation. Flocculation is a type of aggregation,
but the aggregates formed are not cemented together like the aggregates in the dry
pigment powder. Although substantial shear stress is required to separate the original
aggregates, flocculation can be reversed by applying relatively low levels of shear stress.
Flocculation is almost always undesirable. With light scattering pigments, the larger
particle size resulting from flocculation reduces scattering and, therefore reduces
hiding. With color pigments, the larger particle size reduces light absorption and thus
color strength. Larger floccules in the final film affect gloss. Flocculation of pigments,
21.1. DISPERSION IN ORGANIC MEDIA
437
including inert pigments, can change critical pigment volume concentration (CPVC) and
thus affect the properties of coating films (Section 22.2). Flocculated dispersions are shear
thinning and have higher viscosities at low shear rates than those of well-stabilized dispersions. Flocculated dispersions do have the advantage that any settling forms soft pigmentbearing sediments that are easily stirred back to uniformity. However, settling problems
can usually be minimized by the addition of thixotropic additives such as treated clay
or gels, without the adverse results of flocculation.
There are two mechanisms for stabilization: charge repulsion and entropic repulsion.
(As discussed in Section 9.1.1, some authors prefer to use the terms osmotic repulsion
or steric repulsion instead of entropic repulsion). In charge repulsion, particles with
like electrostatic charges repel each other. Calcium and zinc soaps can sometimes be effective pigment dispersants; by absorbing on pigment surfaces they can place ionic charges
on the surface that stabilize dispersions. However, in most cases, entropic repulsion is the
primary stabilizing mechanism in non-aqueous media. Charge repulsion is more important
in aqueous dispersions, as discussed in Section 21.3.
Entropic repulsion is a term used to describe the repelling effect of layers of adsorbed
material on the surface of the particles of a dispersion, which prevents the particles from
getting close enough together for flocculation to occur. In many dispersions of pigments in
organic media, the adsorbed layer consists of polymer molecules swollen with solvent.
The particles are in rapid random (Brownian) motion. As they approach each other,
their adsorbed layers become crowded; there is a reduction in the number of possible conformations of molecules of polymer and associated solvent in the adsorbed layers. The
resulting decrease in disorder constitutes a reduction of entropy, which corresponds to
an increase in energy and requires force; hence, resistance to the reduction of entropy
leads to repulsion, owing to the more favorable, lower energy state of the separated particles. Similarly, compression of the layers could lead to a more ordered system by squeezing out solvent. The accompanying reduction in entropy would again lead to repulsion.
Only certain parts (anchors) of the polymer are adsorbed, with other parts sticking out
into the solution (loops and tails).
Much of our understanding of entropic stabilization of pigment dispersions comes from
the seminal work of Rehacek [2]. He devised an experimental technique to determine the
thickness and composition of the adsorbed layer on the surface of a pigment dispersed in a
polymeric resin solution. First, he made a series of solutions of the resin in a solvent with
different concentrations, c1. He then dispersed a known amount of pigment in each solution. Samples of the pigment dispersions were centrifuged until a pigment-free layer
formed. The concentration of resin, c2, in this supernatant layer was determined. He
plotted (c1 2 c2)/P, where P is grams of pigment, for each sample against c2. [Unfortunately, he labeled (c1 2 c2)/P as “adsorption.” It would have made it easier to understand
his papers if he had labeled this axis “apparent adsorption.”]
A schematic representation of typical data is presented in Figure 21.1; in all cases, the
shape of the curve is the same. Rehacek extrapolated the straight line portion of the curve
to intercepts with both axes. The intercept with the c2 axis represents the concentration of
resin in the adsorbed layer on the pigment surface. The intercept with the (c1 2 c2)/P axis
represents milligrams of resin solids adsorbed per gram of pigment. It can be proven mathematically that these two values remain constant at all points on the linear portion of the
curve. At the intercept with the c2 axis, the apparent adsorption is zero; this does not mean
that no resin is adsorbed. It means that the concentration of resin in the adsorbed layer is
the same as the concentration in the initial resin solution, c1, and in the resin solution after
438
PIGMENT DISPERSION
Figure 21.1. Plot of (c1 2 c2)/P versus c2 in studies of adsorption of resin and solvent on pigment
surfaces. (Adapted from Ref. [2], with permission.)
pigment was dispersed in the resin solution, c2. Thus, the amount of resin and solvent in the
adsorbed layer is established. The surface area of a gram of pigment can be determined by
nitrogen adsorption. Using the densities of the pigment and the resin solution, one can calculate the thickness of the adsorbed layer of resin and solvent in the linear range of the plot.
In Rehacek’s work, as well as that of others [3,4], it has been found that if the
adsorbed layer thickness of resin plus solvent is less than 9 to 10 nm, the dispersion is
not stable and flocculation occurs. This adsorbed layer thickness is an average layer
thickness; some places are thinner and others are thicker. Similar studies in mixed
solvent systems with no resin present show that the adsorbed layer thickness of a combination of solvents is 0.6 to 0.8 nm and not sufficient to stabilize against flocculation [4].
With monofunctional surfactants, the adsorbed layer can be thinner and still protect
against flocculation. In one case, McKay [5] has shown that an adsorbed layer thickness
of 4.5 nm of surfactant and associated solvent was adequate. In contrast to the adsorbed
layer of resin, which is nonuniform in thickness, the surfactant layer is comparatively
uniform, so it does not have to be as thick to provide stabilization. The advantages
and disadvantages of resin – solvent versus surfactant – solvent stabilization are
discussed later.
Referring back to Figure 21.1, it is of interest to understand why the plot deviates from
linearity at lower values of c2. The values of (c1 2 c2)/P are lower at low values of c2. In
other words, there is less apparent absorption of resin per unit area of pigment surface. This
can be explained by competition between resin and solvent adsorption, which depends on
both the relative affinity of resin and solvent molecules for the pigment surface and the
concentration of resin. If the concentration of resin is high enough, the resin “wins” and
there is a complete adsorbed resin layer swollen with solvent. However, at lower concentrations, both solvent and resin are adsorbed on the particle surface, so the average layer
thickness is insufficient to prevent flocculation.
21.1. DISPERSION IN ORGANIC MEDIA
439
Rehacek observed three other differences in the behavior of dispersions with c2 values
above and below the start of the linear section of the curve. The low shear viscosities of the
dispersions below the critical concentration were higher than those above it, separation of
pigment during centrifugation was more rapid, and the bulk of the centrifugate formed was
greater. This behavior indicates that the system is flocculated below the critical concentration. Viscosity at low shear rates increases and the system becomes shear thinning
when flocculation occurs, as discussed in Section 3.5. Also, in a flocculated system
under low shear conditions, particle sizes are larger than in a nonflocculated system;
this causes more rapid settling in centrifugation. Furthermore, the floccules occupy
more volume than that of a stable dispersion of the same amount of pigment, since continuous phase is trapped inside the floccules; this leads to the bulkier layer of sediment.
What controls the thickness of the adsorbed layer? In the case of a surfactant with its
polar end adsorbed on the surface of a polar pigment, the length of its nonpolar aliphatic
chain is the primary factor. In the case of resins with several adsorption sites, the largest
single factor is probably molecular weight (MW). For example, Saarnak showed that the
adsorbed layer thickness on TiO2 dispersed in a series of BPA epoxy resins in MEK
increased from 7 nm to 25 nm as the MW of the epoxy resin increased [3]. With the
lowest MW resin, the layer thickness of 7 nm was insufficient to prevent flocculation.
Dispersions in solutions of the higher MW epoxy resins were stable.
Adsorbed layer thickness is also affected by the pigment surface. Organic pigments are
more likely to give significant differences in adsorbed layer thickness with different resin –
solvent combinations than are more polar inorganic pigments. However, even with inorganic pigments, significant differences can be encountered. For example, a TiO2
surface treated with alumina forms a more stable dispersion than that of a TiO2 surface
treated with silica in the same long oil alkyd solution [6]. The authors proposed that the
adsorbed layer is more compact on the silica-treated TiO2. Commonly, the interactions
between pigment surfaces and adsorbed molecules are considered to be hydrogenbond interactions, but some authors prefer to interpret the interactions as acid –base
interactions [7,8].
The spacing and number of functional groups along a resin chain affect layer thickness.
As an extreme example, a linear aliphatic chain resin with a polar group on every other
carbon atom would be expected to adsorb strongly on the surface of a polar pigment
such as TiO2. At equilibrium, adsorption of single molecules with interaction of successive
polar groups with the pigment surface would be favored, resulting in a thin adsorbed layer.
However, if the resin has only occasional polar groups along the chains, at equilibrium, the
longer segments between polar groups give loops and tails of resin swollen with solvent
projecting out from the pigment surface: hence, a thicker adsorbed layer.
Layer thickness is also affected by solvent –resin interaction. If the loops and tails interact more strongly with the solvent, there are more solvent molecules in the layer, the
average conformation of the resin is more extended, and the layer thickness is greater.
Resin molecules that have multiple adsorbing groups have an advantage in competition
with solvent molecules, but if the solvent interacts strongly with the pigment surface
and the resin only interacts weakly, the more numerous solvent molecules will “win”
the contest. For example, it has been found that toluene favors adsorption of macromolecules such as nitrocellulose, polyurethanes, and phenoxy resins on magnetic iron
oxide pigment particles more than does tetrahydrofuran [8]. Addition of solvent to a
stable pigment dispersion can, in some circumstances, lead to flocculation. If the ratio
of resin to solvent is just sufficient to allow adequate adsorption of resin to stabilize the
440
PIGMENT DISPERSION
dispersion, addition of more of the same solvent can shift the equilibrium, displacing part
of the resin and reducing the average adsorbed layer thickness below the critical level for
stabilization, resulting in flocculation. The dispersion is said to have been subjected to
solvent shock.
For most conventional solventborne coatings, the resin used as binder in the coating can
stabilize the pigment dispersion. Most conventional alkyds, polyesters, and thermosetting
acrylics stabilize dispersions of most pigments. In the past, most dispersions were made in
the resin to be used in the paint. If there was a problem, the two most common changes to
give adequate stability were to increase the MW or the number of hydroxyl, amide,
carboxylic acid, or other polar groups on at least that part of the resin to be used in the mill
base. It has been shown that the higher MW components of resins are selectively adsorbed [7].
It is frequently difficult to make stable pigment dispersions in the resins used in highsolids coatings. Increasing the solids of organic solution coatings requires decreasing the
MW of the resins and reducing the number of functional groups per molecule. Reduced
MW results in thinner adsorbed layers of resin and associated solvent molecules. The
reduced number of functional groups per resin molecule decreases the probability of the
adsorption of resin molecules; there is a greater probability of solvent adsorption being
favored and hence a greater likelihood of flocculation.
21.1.3.1. Pigment Dispersion Stabilizers for Non-aqueous Coatings The
trend is modern paint formulation is to add pigment dispersion stabilizers during the dispersion process. They may be used in combination with a primary resin for the coating or,
in some cases, by themselves. This practice often produces superior dispersions for
conventional-solids coatings and is essential with many high-solids coatings. Conventional monofunctional surfactants can be used for this purpose. However, if the concentration
of surfactant is low, the solvent is likely to outcompete most surfactant molecules, due to
the larger number of solvent molecules, even though the surfactant may be more strongly
adsorbed. This can be offset by increasing the concentration of surfactant molecules to
shift the equilibrium favor of surfactant, but doing so leaves excess surfactant in the
final film, which tends to reduce performance properties such as adhesion to metal.
Thus, in organic media, monofunctional surfactants are not generally a desirable choice
for stabilization. A commonly used additive is lecithin, a naturally occurring choline
ester of phosphoglycerides. It is quite strongly adsorbed on many pigments.
Surfactants have been designed that are so strongly adsorbed on a pigment that they are
preferentially adsorbed even in the presence of a large amount of solvent. For example,
phthalocyanine blue modified by covalently attaching long aliphatic side chains has
been used as a surfactant with phthalocyanine blue pigment; the phthalocyanine end of
the molecules of surfactant in effect joins the crystal structure of the surface of the
pigment particles so that little, if any, is in solution [5]. The average adsorbed layer thickness needed to protect against flocculation was shown to be about 4.5 nm. Specific proprietary surfactants have been designed for other pigments. Although effective, this approach
finds only limited use because the special surfactants are expensive and are designed to
work with specific pigments.
These difficulties have led to the design of special dispersing aids called polymeric
dispersants, A-B dispersants, hyperdispersants, or superdispersants [9– 11]. Jakubauskas
describes the design parameters of such dispersants [9]. He concluded that the most effective class of dispersant has a polar end with several functional groups (anchoring groups)
21.1. DISPERSION IN ORGANIC MEDIA
441
and a less polar tail that is soluble in the medium and has sufficient length to provide a
surface layer that is at least 10 nm thick. Tails are more effective than loops. (See
Ref. [11] for a further review of the use of hyperdispersants.) Examples of such dispersants
include polycaprolactonepolyol – polyethyleneimine block copolymers [9], polycaprolactone capped with toluene diisocyanate postreacted with triethylenetetramine [9], low MW
polyesters from polyhydroxystearic acid [10], and proprietary hyperdispersants [11].
Controlled free radical polymerization (CFRP) (Section 2.2.1.1) is used to synthesize
dispersants, as detailed in Section 21.3.1. In some cases it is desirable to treat the
surface of a pigment to provide sites on the surface to interact with the anchors of the dispersant [12,13].
A single dispersant that would anchor firmly to all pigments and be compatible with all
types of coatings would be desirable. Then a single series of dispersions could be used in
all products, substantially reducing production costs. This ideal is not attainable, but part
of the potential savings can be realized by using broadly effective dispersants to reduce the
number of series of dispersions.
Another approach to stabilization is to covalently bond long chains to the surface of
pigments. For example, the surface of SiO2 pigment particles has been reacted with trialkoxysilanes with long chain alkyl substituents [14]. Titanate orthoesters with three ethyl
groups and one alkyl group with a long alkyl chain reduce the viscosities of dispersions
of some inert pigments. It may be that the ethyl groups exchange with hydroxyl groups
on the pigment surfaces. In many cases, pigment manufacturers offer pigments that
have been surface treated to provide stable dispersions even when the pigment is dispersed
in solvent. (See Refs. [13] and [15] for reviews of the surface treatment of pigments.)
21.1.3.2. Other Considerations In high-solids coatings there is a further problem of
adsorbed layer thickness on pigment surfaces. In conventional coatings, the volume of
pigment in the wet coating of even highly pigmented coatings is quite low, so differences
in adsorbed layer thickness do not make a large difference in the viscosity of the wet
coating (provided that the dispersion is stable against flocculation). However, in highsolids coatings, especially highly pigmented high-solids coatings, the adsorbed layer
thickness can have a significant effect on the viscosity. This problem is especially limiting
in the formulation of high-solids primers, where high pigment loadings near the critical
pigment volume concentration (CPVC; Chapter 22) are often desirable. The effect of
layer thickness can be seen by considering a modification of the Mooney equation
(Section 3.5) which takes into account that the internal phase volume due to the adsorbed
layer, Va, in addition to that due to the pigment, Vp, is included in Vi: Using Eq. 21.1,
model calculations were made to illustrate the effect of adsorbed layer thickness on
viscosity [16].
ln n ¼ ln nc þ
2:5(V p þ Va )
1 (V p þ Va )=f
(21:1)
Figure 21.2 shows the calculated dependence of the viscosity of coatings with 70 NVV
as a function of the pigment volume concentration in dry films to be prepared from the
coatings. The possible level of pigmentation is more limited with the thicker adsorbed
layer than with the thinner one. The following principal assumptions were made in
making the calculations: Vi ¼ 1.2Vp and Vi ¼ 2Vp (these values are equivalent to an
8-nm and a 25-nm adsorbed layer, respectively, on a 200-nm pigment); f ¼ 0.65;
442
PIGMENT DISPERSION
Figure 21.2. Calculations of the effect of PVC on the viscosity of two formulations, both at 70 NVV but
differing in thickness of the layer of adsorbed polymer solution. See the text for assumptions. (From
Ref. [16], with permission.)
solvent: r ¼ 0.8 and h ¼ 0.4 mPa.s; and oligomer: r ¼ 1.1 and h ¼ 40 mPa.s at 70 NVV,
4 105 mPa.s at 100 NVV. The viscosity dependence was assumed to follow Eq. 21.2, in
which wr is weight fraction of resin:
ln n ¼ ln ns þ
wr
0:963 þ 0:763wr
(21:2)
In retail outlets, paints are usually inventoried as white paints and white tinting bases.
Colors are made to order by adding color pigment dispersions (colorants) to the white base
paints. Since both waterborne and solventborne paints have to be tinted, it is desirable to
have universal colorants, dispersions that can be used in both types of paint. Examples of
vehicles for such dispersions are a proprietary dispersing agent, water, propylene glycol,
modified alkylpolyglycol ether surfactant, biocide, and antifoam [17].
21.2. FORMULATION OF NON-AQUEOUS MILL BASES
The combination of resin (and/or dispersant), solvent, and pigment used to make a
pigment dispersion is called a mill base. The formulator must design a mill base for dispersing a pigment in the most appropriate dispersion equipment at optimum efficiency.
Pigment dispersion machinery is the most expensive machinery in a coatings factory
from the standpoint of both capital and operating costs. It is therefore important to maximize the amount of pigment dispersed per unit time. Higher pigment loading means more
efficient production; high loadings are possible when the viscosity of the vehicle (solvent
21.2. FORMULATION OF NON-AQUEOUS MILL BASES
443
plus resin) used in the mill base is low. Low viscosity also gives faster wetting. A properly
stabilized pigment dispersion exhibits Newtonian flow, and its viscosity follows the
Mooney equation:
ln n ¼ ln nc þ
KE Vi
1 Vi =f
(21:3)
Although mill bases usually contain a single pigment, there are circumstances in which it
is advantageous to mill two or more pigments together. Such mill bases are called composite grinds.
The volume of pigment (internal phase) can be maximized by using a vehicle of the
lowest possible viscosity (he). Solvent alone gives low viscosity, fast pigment wetting,
and high pigment content, but solvent alone cannot stabilize a dispersion against flocculation. Therefore, it is necessary to include some resin (or hyperdispersant) in the mill base.
For maximum pigment loading, it is desirable to use the minimum concentration of resin
solution that provides stability. The Rehacek procedure could be used to determine the
minimum concentration that still gives a point on the linear section of the curve.
However, the procedure is time consuming. Many years ago, Fred Daniel devised a
simpler and faster, although less accurate method [18].
21.2.1. Daniel Flow Point Method
The Daniel flow point method is a powerful tool for formulating mill bases efficiently,
especially for dispersions to be made in ball and other media mills [18]. It provides an estimate of the most appropriate resin concentration to use with a particular pigment. One
makes a series of solutions with different resin concentrations. Then one determines the
amount of each solution that must be added to a weighed amount of pigment so that
when it is dispersed by rubbing vigorously with a spatula on a flat glass plate, the resulting
dispersion has a viscosity just low enough to flow readily off the spatula. The spatula is
used as both a dispersion machine and a viscometer. Since the clearance between the
spatula and the glass plate is small, the shear rate is high; a spatula is a fairly good dispersing machine. One plots the volume of each solution needed for flow against the concentration of that solution. This plot is an isoviscosity plot; that is, the viscosity of each of the
dispersions is approximately the same. (The actual low shear rate viscosity is approximately 10 Pa.s.) An example is shown in Figure 21.3; in addition to the volume of resin
solution, the volume of solvent in the resin solution is also plotted.
Any resin – solvent – pigment combination that can give a stable dispersion shows a
minimum in the curve. Let us consider the significance of the minimum. To the right of
the minimum, the viscosity of the external phase is increasing because the concentration
of resin is increasing. As can be seen from Eq. 21.3, if he increases and the viscosity stays
constant, Vi must decrease; that is, the amount of vehicle required per unit of pigment must
increase. To the left of the minimum, the amount of vehicle per unit of pigment is also
increasing, even though he is decreasing. The concentration of resin is insufficient to
stabilize the dispersion. The pigment dispersion flocculates increasingly as the concentration of resin in the solutions decreases, and the viscosity increases more and more
steeply, so additional resin solution must be added to reach the isoviscosity level. The
minimum point corresponds to the minimum concentration of that resin in that solvent
that permits preparation of a stable dispersion of that pigment. Since the determination
444
PIGMENT DISPERSION
Figure 21.3. Daniel flow point plot: milliliters of solutions of an alkyd resin in mineral spirits per 20 g of
TiO2 as a function of NVW resin in solution. (From Ref. [18], with permission.)
is not highly accurate, it is usual to start experimental dispersions with a somewhat higher
resin concentration. If one adds more solvent to a stable dispersion at nearly the minimum
resin concentration, the dispersion will flocculate. In some cases, no minimum is found;
the amount of resin solution required per unit of pigment keeps decreasing as the concentration increases. This behavior signifies that a stable dispersion of that pigment cannot be
made with that combination of resin and solvent. Patton gives examples of data on the
Daniel flow point method [18]. He also gives information expressed in volume units of
pigment as well as the more conventional weight units.
21.2.2. Oil Absorption Values
In the early days of formulating coatings, it was observed that the weight of pigment that
could be incorporated with linseed oil in a mill base varied greatly from pigment to
pigment. To simplify formulating mill bases to approximately equal viscosity, oil absorption values were determined. Linseed oil was slowly added to a weighed amount of
pigment while rubbing with a spatula. Initially, balls of pigment wet with oil formed.
When just enough oil had been added and worked into the pigment so that one coherent
mass formed, the end point was reached. The oil absorption value is calculated as the
number of pounds (grams) of linseed oil required to reach the endpoint with 100 lb
(100 g) of pigment. At the endpoint, there is just enough linseed oil to adsorb on the
surface of all the pigment particles and to fill the interstices between close-packed
particles.
21.3. DISPERSION IN AQUEOUS MEDIA
445
Oil absorption values for different pigments vary over a wide range. The smaller the
particle size, the higher the oil absorption. A small particle size pigment has more
surface area; therefore, a larger amount of linseed oil is adsorbed on the surface. In
some cases, such as some grades of carbon black, the pigment particles are porous.
Some oil penetrates these pores, increasing the amount of oil required and hence the oil
absorption value. Diatomaceous earth (Section 20.3) has a very high surface area and
hence very high oil absorption values. Pigment density also has a major effect. High
density pigments require less weight of oil to adsorb on the surface of a unit weight of
pigment and to fill the interstices; hence, they have lower oil absorptions. Vehicles
other than linseed oil give similar oil absorption values. Therefore, data obtained with
linseed oil can be used in formulating mill bases with any vehicle that gives stable
dispersions.
The precision of oil absorption determinations by the spatula method is not high. Operators working in different laboratories commonly report values that vary by +15%. It has
been said that with experience, deviations by a single operator can be reduced to +2 to
3%. Despite these error ranges, it is common to see oil absorption values given with
three supposedly significant figures. Improved accuracy and precision can be achieved
by using a mixing rheometer (Section 3.3.2), such as a Brabender Plastometer, for carrying
out the determination [19]. The mixer is loaded with a known amount of linseed oil;
pigment is added slowly. The mixer imparts the necessary shear to separate pigment
aggregates. The power required to turn the blades is recorded. As the amount of
pigment is increased, the power requirement increases, and when the oil absorption endpoint is passed, the dispersion mass breaks up into chunks, leading to erratic readings. The
values obtained are more reproducible and are generally a few percent higher than those
obtained by the spatula method. Formulators can also take advantage of oil absorption
values provided by pigment vendors. These oil absorption values permit one to set a
starting point for pigment content when formulating a mill base. Batch-to-batch variations
can be expected. The relationship between oil absorption of pigments and the critical
pigment volume concentration of films containing those pigments is discussed in
Section 22.1.
21.3. DISPERSION IN AQUEOUS MEDIA
Dispersion of pigments in aqueous media involves the same factors as in organic media:
wetting, separation, and stabilization. However, the unique properties of water add extra
complications. First, the surface tension of water is high, so there is more likely to be a
problem in wetting the surfaces of low-polarity pigment particles. Second, in some
cases, water interacts strongly with the surface of pigments; therefore, the anchor
groups on the stabilizers have to interact more strongly with the pigment surface to
compete with water. Furthermore, many applications of aqueous dispersions are in latex
paints, so the systems have to be designed such that stabilization of the latex dispersion
and of the pigment dispersion do not affect each other adversely.
Inorganic pigments such as TiO2, iron oxides, and most inert pigments have highly
polar surfaces, so there is no problem with wetting them with water. The surfaces of inorganic pigments interact strongly with water, but the adsorbed layer of water does not
by itself stabilize against flocculation. Most organic pigments, however, require the use
of a surfactant to wet the surfaces. Some organic pigments are surface treated with
446
PIGMENT DISPERSION
adherent layers of inorganic oxides to provide a polar surface that is more easily wet by
water [13].
21.3.1. Stabilization of Aqueous Dispersions
In contrast to dispersions in organic media, stabilization by charge repulsion can be a
major mechanism in aqueous media (Section 9.1.1). The stability of the dispersions
depends on pH, since pH affects surface charges. For any combination of pigment, dispersing agent, and water, there is a pH at which the surface charge is zero; this pH is called the
isoelectric point (iep). At iep, there is no charge repulsion; above iep, the surface is negatively charged; and below iep, it is positively charged. The stability of dispersions is at a
minimum at iep + 1 pH unit [20]. The iep value for pigments varies: for example, 4.8 for
kaolin clay and 9 for CaCO3.
Since latex paints were introduced in the 1950s, continuous efforts have been made to
improve colorants. Relatively recently, a need has developed for good solvent-free colorants for use in “zero-VOC” paints. By the 1990s, introduction of waterborne automotive
base coats (Section 30.1.2) increased the need for high quality aqueous pigment dispersions. These needs motivated intensive research on pigment dispersants that continues
today. An early goal was to find polymeric dispersants that were effective in aqueous
media. A variety of acrylic copolymers made by conventional free radical polymerization
proved to be useful. For example, copolymers of (1) (meth)acrylic acid, (2) monomers
bearing one or more highly polar anchoring groups that can absorb on pigment surfaces,
and (3) a mono-(meth)acrylic ester of polyethylene oxide are effective. Such copolymers
are often used in combination with conventional surfactants.
Controlled free radical polymerization (CFRP) techniques have been studied to obtain
polymeric dispersants that are more effective and more efficient than is possible from conventional polymerization. Intensive research on dispersants made by CFRP is summarized
in a recent review [21]. These techniques can be adapted to produce AB and ABA block
copolymers, and graft, star, and hyperbranched copolymers with controlled structures.
When made with appropriate adsorbing and stabilizing groups and properly structures,
all these types of polymers can function as pigment dispersion stabilizers. (See Section
2.2.1.1 for discussion of CFRP methods.)
One type of CFRP, atom transfer radical polymerization (ATRP), permits synthesis of
polymers having predictable MWs, low MW distributions, and controllable block structures. For example, an AB block copolymer pigment dispersant is synthesized by (1)
loading a reactor with toluene, then 1-naphthalene sulfonyl chloride, copper powder,
2,20 -bipyridyl, and glycidyl methacrylate and polymerizing at 808C for 6 hours to form
the A block; (2) adding the methacrylate of the monomethyl ether of polyethylene
glycol in methyl 2-hydroxypropyl ether and polymerized for an additional 4 hours at
808C to attach a hydrophilic B block; (3) reacting the glycidyl epoxy groups with undecanoic acid to make the A block highly hydrophobic; and (4) removing the copper complex
by precipitation with magnesium silicate and filtration. The resulting AB block copolymer
solution is reduced to 70 wt% solids, dispersed in water, and the pH is adjusted to 8 with
DMAE [22].
Another type of CFRP, catalytic chain transfer polymerization (CCTP), differs from
ATRP in that a conventional initiator is used in combination with a cobalt (or other
metal) complex that mediates the reaction [23]. In one example, a graft copolymer dispersant is prepared using an azo initiator with a cobalt(II) complex as chain transfer agent.
21.3. DISPERSION IN AQUEOUS MEDIA
447
First, CCTP is used to prepare an acrylic macromonomer with 20 wt% methacrylic acid.
This macromonomer is then copolymerized with a hydrophilic nonionic macromonomer
and hydrophobic monomers. The result is a graft copolymer with a hydrophobic backbone
and both anionic and nonionic hydrophilic side chains. When neutralized with AMP, it
becomes an effective dispersant. In aqueous media, the hydrophobic backbone presumably
anchors the polymer to pigment surfaces, and the hydrophilic grafts act as tails, providing
entropic and charge stabilization. Dispersions of TiO2 in the dispersant gave base coats
with higher gloss and higher gloss after being clear coated [23].
Another type of CFRP is nitroxide-mediated polymerization (NEM) using hindered
nitroxide compounds as mediating agents to prepare narrow MW distribution block copolymers. For example, a block copolymer of butyl acrylate and dimethylaminoethyl acrylate useful as a pigment dispersion stabilizer is made by sequential bulk polymerization
using piperidine nitroxyethers as mediators [24].
The possibilities are almost limitless, but the polymer structures must be designed
correctly for good results. For example, an AB diblock copolymer of 2-(dimethyl
amino)ethyl methacrylate and t-butyl methacrylate, made by living ionic polymerization
is very effective; whereas a corresponding ABA triblock copolymer of the same components, having amine-functional blocks on both ends, leads to particle bridging and
severe flocculation [25].
A broad study of stabilization of aqueous TiO2 dispersions by anionic and nonionic surfactants concluded that a high MW nonionic copolymer provided the greatest resistance to
flocculation both in the dispersion and during drying of a gloss latex paint film [26]. Since
stabilization resulted primarily from entropic stabilization, not charge repulsion, it was not
affected by changes in pH. On the other hand, a study of stabilization of TiO2 dispersions
with sodium polyacrylate showed that both charge and entropic repulsion affected stability. The surface layer thickness did not have to be as great as in solvent based dispersions.
Even as little as an average 1-nm layer thickness combined with the charge repulsion sufficed for stabilization. Adsorption on the TiO2 surface was flat but there were loops and
tails extending into the water phase. Higher MW gave thicker films and adversely affected
viscosity [27].
Surface treatments can have important effects on the stability of aqueous pigment dispersions. Special treated TiO2 pigments have been developed for waterborne coatings.
Substantial differences in dispersion stability can result from differences in the composition and completeness of the surface treatment [28]. In some cases, the amount of
surface treatment is such that the TiO2 content is as low as 75% of the pigment weight.
Since hiding is related to the actual TiO2 content, larger amounts of such highly treated
pigments are needed to obtain equivalent hiding.
Slurries of up to 80 wt% TiO2 in water are used on a large scale with substantial cost
savings (Section 20.1.1). The dispersions are stabilized with carboxylic acid– functional
polymeric dispersants; pH is controlled by adding amine. The formulations are established
to permit minimal flocculation to retard settling. Bactericide is added to inhibit microbial
growth.
Dispersion of fine particle size pigments such as transparent iron oxide in water to give
coatings without haze can be difficult. It has been shown that good results can be obtained
using dispersants prepared by copolymerizing vinyl ether – terminated polyethers with
maleic anhydride. High color yield dispersions can be achieved that give films with
much less haze than other dispersing agents. For special frosty appearance effects, dispersions of fine particle size “transparent” TiO2 can be used [29].
448
PIGMENT DISPERSION
Most latex paint formulations contain several pigments and several surfactants. The iep
of the various pigments are different, which complicates the problem of charge stabilization. Commonly, mixtures of surfactants are used. Anionic surfactants are frequently
used as one component. Polymeric anionic surfactants (such as salts of acrylic copolymers
in which acrylic acid and hydroxyethyl acrylate are used as co-monomers) provide salt
groups for strong adsorption on the polar surface of the pigment and hydroxyl groups
for interaction with the aqueous phase; nonpolar intermediate sections add adsorbed
layer thickness. Ammonia or amine salts are recommended since the base can escape
from the film during drying giving films that are less vulnerable to water than when
sodium or potassium salts are used. Polymeric surfactants are less likely than monomeric
surfactants to lead to performance problems in the final films. Since polyacrylates are not
good wetting agents, they are frequently used with nonionic surfactants and/or an anionic
surfactant. It is also common to add potassium tripolyphosphate, the basicity of which may
assure that the pH is above the iep of all pigments. Note that potassium tripolyphosphate is
used in paints, not the sodium salt used in laundry detergents. The potassium salt is less
likely to deposit on the surface of a film as scum after being leached out of a dry paint
film by water. Large paint producers develop proprietary surfactants for their major
product lines.
Additives can sometimes affect dispersion stability. For example, a hydrocarbon
solvent used in a latex paint as a defoamer was shown to cause flocculation [30]. The
paint formula had several surfactants. It was found that the order in which the surfactants
were added controlled whether or not flocculation occurred.
Unless surface treated, organic pigments generally have surface free energies lower
than the surface tension of water. Therefore, surfactants are needed to reduce the
surface tension of the water to permit the wetting essential for pigment dispersion.
Either anionic or nonionic surfactants can be used. In the case of anionic surfactants,
the dominant mode of stabilization is probably charge repulsion. In the case of nonionic
surfactants, the relatively long polyether alcohol end is oriented out into the water and
associated with multiple water molecules; stabilization is probably predominantly by
entropic repulsion.
It has been reported that organic pigments dispersed in aqueous polyurethane dispersions (PUDs) show greater color strength than when dispersed in solventborne media.
Depending on the pigment, the increase is between 10 and 30%. The effect is attributed
to water’s ability to dissolve salts in the pigment agglomerates; these salts are not dissolved by organic solvents and increase adhesion within the agglomerates. However, it
is also reported that some pigments, particularly some quinacridone and diketopyrrololpyrrole pigments that are non-bleeding in solventborne coatings bleed from some
waterborne base coats into clear coats. This is attributed to the aggressive solvents that
are present in some aqueous vehicles, especially N-methylpyrrolidinone and butoxyethanol [31].
In many cases, paint manufacturers make only white paints and supply separate color
dispersions and formulations for the paint stores to mix with the white paints to make a
large number of colored paints for their customers. Many latex paint manufacturers buy
rather than manufacture their color pigment dispersions. The wetting and stabilization
methods used are proprietary.
There is need for more research on the factors leading to stability of dispersions in latex
paints. There is reasonable understanding of the stabilization of single pigment systems.
However, paints often contain several pigments, commonly with both polar and nonpolar
21.4. DISPERSION EQUIPMENT AND PROCESSES
449
surfaces and one or more latexes (Chapter 32). Furthermore, the formulations may contain
one or more water-soluble polymers and/or associative thickeners (Section 32.2) to adjust
the rheological properties of the paint, and these can interact with dispersants. As a result,
the selection of appropriate dispersing agent combinations is done at least partly on an
empirical basis. As in any situation in which empirical knowledge is the key to success,
it is desirable to build a database of combinations that work and do not work to facilitate
formulating with a new pigment combination or latex.
21.4. DISPERSION EQUIPMENT AND PROCESSES
A variety of equipment is used to make pigment dispersions. An important difference
between different machines is in the level of shear stress that they can exert on the
pigment aggregates. Easily separated pigment aggregates can be dispersed in low shear
stress equipment, whereas some pigment aggregates require very high shear stress for separation. Discussion of all types of such machinery and their operation is beyond the scope
of this book. Patton provides more discussion of some types [18]. Detailed and up-to-date
engineering information is often available from the machinery manufacturers. Reference
[1] deals more fundamentally with engineering aspects of some dispersion methods. We
discuss a few important types with emphasis on their advantages and disadvantages. A formulator must become familiar with dispersion equipment available in his or her company’s factories to design formulations appropriate for production with that machinery.
From a processing point of view, there are three stages in making and using pigment
dispersions: (1) premixing, that is, stirring the dry pigment into the vehicle and eliminating
lumps; (2) imparting sufficient shear stress to separate the pigment aggregates in the presence of a dispersion stabilizer; and (3) let down, that is, combining the pigment dispersion
with the balance of the ingredients to make a coating. Some machines can carry out only
the second step, some can do two of the three, and others can do all three.
Some pigment dispersions are made for specific batches of a coating; after the pigment
dispersion is made, it is let down with the other components to make the final coating.
Other pigment dispersions are made for use in several related types of coatings or as
tinting pastes for color matching a variety of coatings. To minimize the inventory of
tinting pastes, one tries to select vehicle and stabilizer combinations that are compatible
with a range of coatings made by that company.
21.4.1. High-Speed Disk Dispersers
High-speed disk (HSD) dispersers consist of a shaft with a disk that rotates at high speed in
a vertical cylindrical tank. High-speed disk dispersers are also called high-speed impellers,
dissolvers, and high-intensity mixers. The disks have flanges around the perimeter that are
bent at a sharp angle from the plane of the disk. A schematic drawing of a typical disk is
shown in Figure 21.4. Increasingly, the disks are made from engineering plastics that
provide greater abrasion resistance than steel. The shearing action takes place by the
differential laminar flow rates streaming out from the edge of the disk, which typically
rotates with a peripheral speed of 20 to 26 m s21. Sometimes, the speed of an HSD disperser is expressed in revolutions per minute (rpm), but the important speed is the peripheral speed, which depends on the radius of the disk. The shear stress developed is relatively
low, and therefore HSD machines are appropriate only for relatively easily separated
450
PIGMENT DISPERSION
Figure 21.4. High-speed impeller disk. (From Ref. [18], with permission.)
pigments. Obtaining predominantly laminar flow requires some minimum viscosity,
depending on the dimensions; it is usually somewhat over 3 Pa.s. The higher the viscosity,
the greater the shear stress exerted on the aggregates and therefore the faster the process.
The viscosity should be set such that the motor driving the shaft is run at peak power.
For maximum efficiency, pigment loading should be maximized. In solvent-based coatings, this can be done by using a vehicle solution with as low a resin concentration as provides stabilization against flocculation. The Daniel flow point method can be used to
estimate the resin –solvent ratio. Then the ratio of this vehicle to the pigment is set to
achieve a viscosity at a high shear rate such that the motor draws peak amperage. In
making latex paints, water-soluble polymers are commonly included in the pigment dispersion stage to increase viscosity during dispersion.
A diagram of flow in using an HSD disperser is shown in Figure 21.5. Centrifugal force
leads to flow up the sides of the tank. If the mill base is Newtonian and the dimensions and
Figure 21.5. High-speed impeller disperser showing the correct positioning of disk and optimum
dimension ratios. (From Ref. [18], with permission.)
21.4. DISPERSION EQUIPMENT AND PROCESSES
451
operating conditions are appropriate, the entire charge becomes intimately mixed and all
portions of the mill base pass repeatedly through the zone of highest shear, near the edges
of the disk. However, if the mill base is shear thinning, the viscosity is high at the upper
edge of the material on the sides of the tank where the shear rate is low, resulting in hang
up and incomplete mixing. Some mill bases contain a pigment designed to make the final
coating shear thinning. The resulting hang up problem can be minimized by dispersing all
of the pigments except the one that gives the shear thinning effect. After the balance of the
pigments has been separated, the pigment that causes shear thinning is added slowly. A
desirable approach to minimizing the hang up problem is to design the dispersing tank
with a slow-speed scraping blade that travels around the upper inside of the tank while
the high-speed disk is spinning in the center of the tank.
HSD machines are commonly used for premix, dispersion, and let down operations.
One initially loads the vehicle components, mixing at a low rpm, and then slowly loads
the dry pigment near the shaft. After the pigment is loaded, the speed is turned up for
the dispersion stage. In a properly formulated mill base, the dispersion stage requires
about 15 minutes. The speed is reduced and the let down is carried out with the rest of
the components in the formulation. When producing latex paints, the latex cannot be
present during the dispersion stage, since most latexes coagulate when exposed to high
shear. The latex is added in the let down stage at low speed.
Compared to other dispersion equipment, HSD machines generally have the lowest
capital and operating costs. No separate premix is required, and the let down can be
carried out in the same tank. Cleaning to change from one color to another is relatively
easy. Solvent loss can be kept to a minimum by using a covered tank. The major limitation
is that the shear stress imparted to pigment aggregates is relatively low, so the equipment
can only be used for relatively easily separated pigments. Pigment manufacturers have
made substantial progress in making pigments with easily separated aggregates. Laboratory HSDs are available, and results correlate well with production machines.
21.4.2. Rotor –Stator Mixers
Shear stress imparted to pigment aggregates can be increased by substituting rotor – stator
mixers for HSD machines. These mixers are somewhat similar to HSD machines, but the
rotating disk spins in close proximity to fixed elements. The mixers can be used for large
batches or in-line. For a general review comparing these mixers with HSD machines, see
[32]. It is claimed that, in some cases, rotor– stator mixers give superior dispersions in
aqueous media, including critical dispersions for automotive basecoats.
21.4.3. Ball Mills
A ball, or pebble, mill is a cylindrical container, mounted horizontally and partially filled
with balls or pebbles. The mill base components are added to the mill, and it is rotated at a
rate such that the balls are lifted up one side and then roll in a cascade to the lower side, as
shown schematically in Figure 21.6. Intensive shear is imparted to pigment aggregates
when the balls roll over each other with a relatively thin layer of mill base between them.
Two classes of such mills are used. In steel ball mills, the balls and lining of the cylindrical mill are steel. In pebble mills, the balls are ceramic and the mill lining is porcelain. In
early days, the balls were large smooth pebbles, hence the term pebble mills. The term
used by some to mean either class of mill and by others to mean only steel ball mills.
452
PIGMENT DISPERSION
Figure 21.6. Cascading pattern in a ball mill. (From Ref. [18], with permission.)
We use ball mill in the general sense and steel ball mill and pebble mill to specify each
particular class. Steel balls have the advantage of higher density, and therefore shearing
is greater and milling times can be shorter. But their use is limited to dark colored pigments and pigments where color is not critical (primers) because the balls wear and discolor light colors. Pebble mills are used when discoloration is a disadvantage.
Ball mills operate most efficiently when their diameter is large, so that the length of the
cascade is long. The efficiency of operation is dependent on the loading of the mill. A mill
should be loaded about half full of balls; this gives the longest cascade. The mill base
volume should be just over the volume needed to cover the balls when the mill is at
rest. If excess mill base is used, the time required for satisfactory separation increases.
In a properly loaded mill, the volume of balls is approximately 32% and the volume of
mill base is only a little over 18%, assuming that the balls are spheres of equal diameter.
If the balls or pebbles are not spheres and/or not uniform in size, there is even less space
for mill base.
Operation of a ball mill requires care in setting the rpm of the mill. If the rpm is too low,
the balls are not carried high enough. If it is too high, the balls are carried past the 608
angle shown in Figure 21.6, and some fall rather than cascade. This reduces shear, can
lead to ball breakage, and is more likely to lead to breaking the ultimate crystals of the
pigment rather than just shearing the aggregates. Experienced mill operators can judge
whether the rpm setting is proper by the sound. Milling efficiency is also affected by
the viscosity of the mill base. If the viscosity is too low, ball wear is high. If the viscosity
is too high, the balls roll more slowly and efficiency is reduced. Optimum viscosity
depends on the density of the balls and mill size. Viscosities in the neighborhood of
1 Pa.s are commonly used.
The Daniel flow point method (Section 21.2.1) was originally developed to help formulate mill bases for ball mills. It permits loading the maximum amount of pigment at the
viscosity required. Proper mill base formulating and use of proper volumes of balls and
mill bases can make large differences in the time required for satisfactory separation.
The time required depends on how easily the pigment aggregates are separated. The
minimum time is usually on the order of 6 to 8 hours. Even difficult pigments should
require no more than 24 hours. Sometimes one hears of 72 or more hours being required
for dispersion. This almost always indicates poor formulation or improper mill loading.
Ball mills can be used to disperse all but the most difficult to separate pigments.
Although the capital cost of ball mills is relatively high, operating cost is low. No
21.4. DISPERSION EQUIPMENT AND PROCESSES
453
premixing is required, and the mills can run unattended. They are difficult to clean and
therefore are most appropriate for making batch after batch of the same dispersion.
Another difficulty is that batch sizes are limited because the mills contain only about
18% by volume of mill base.
There is no directly comparable laboratory mill available. This is because the efficiency
of a ball mill is so dependent on its diameter. Production mills range from 1.25 to 2.5 m
(4 to 8 ft) in diameter. Some laboratories use jar mills, which are usually less than 30 cm
(1 ft) in diameter and rolled at a much less than ideal rpm rate. Correlation with production
operations is usually poor.
In the laboratory, quickee mills are more appropriate and are much faster for making
small batches of dispersions roughly similar to production dispersions in a ball mill. A
steel container is filled a little over half full with 30 mm steel balls and enough mill base
to somewhat more than cover the balls. The container is then shaken on a paint shaker
of the type used in paint stores. Easily separated pigments require 5 to 10 minutes. Difficult
to separate pigments may take up to an hour of shaking. Glass jars or vials partly filled with
glass beads, sand, or ceramic media can also be used as quickee mills.
21.4.4. Media Mills
Media mills (also called agitator mills) were invented to get around batch size limitations
of ball mills. In the earliest mills, the media was sand that had an average diameter of about
0.7 mm, and the mills were called sand mills. Sand is still used, but more commonly, small
ceramic balls are used; sometimes the mill is then called a bead mill. Small steel balls are
also used; then the mill is commonly called a shot mill because the steel balls look like the
shot used in shotgun shells. Shot mills are generally used to disperse carbon black or other
dark pigments.
A schematic drawing of a vertical media mill is shown in Figure 21.7. A high-speed rotor
with disks is mounted in a cylinder. The space between the rotor and the cylinder is partially
filled with sand or other media. A premixed mill base is pumped into the bottom of the mill. It
flows up through the mill, being exposed to shear as it passes between the rapidly moving
particles and out through a screen that keeps the media in the mill. A typical vertical
media mill operates with a peripheral rotor speed of 10 m s21. This impels the particles at
high speed, so that even though their size is small, the shear rate between them is high,
and, of course, in passing through the mill, a pigment aggregate passes between many
pairs of particles. Mill bases should be formulated using slightly higher resin concentrations
than the minimum in the Daniel flow point curve. The higher the viscosity of the mill base, the
longer the average residence time in the mill and hence the greater the degree of shear. For
easily separated pigments, it is desirable to use lower viscosity to achieve a greater flow rate
through the mill. Viscosities range from 0.3 to 1.5 Pa.s. For pigments that are two or three
more difficult to separate, passes through a mill (or a battery of mills in series) may be
required. Mills have been designed that automatically recycle part of the mill base.
The mill shown in Figure 21.7 has a vertical rotor, but use of horizontal media mills is
increasing. A variety of designs are available. Some have simple disks on the rotor,
whereas others have perforated, offset, or cam-shaped disks. Pin mills are also being
used increasingly. Pins are placed in opposite walls of a rotating cylinder so that the
pins on one side fit between pins on the other side. A pin mill with a motor of the some
horse power runs at higher speed, which enhances separation of beads, resulting in
higher grinding efficiency and higher energy efficiency.
454
PIGMENT DISPERSION
Figure 21.7. Media mill. (From Ref. [18], with permission.)
Nano-pigments are particularly difficult to disperse, and pin mills are recommended.
Reference [33] provides a discussion of the factors involved as the size of the media is
reduced. The ultimate size of the dispersed pigment decreases as the size of the media
decreases, so that the smallest particle size media will provide the smallest particle size
pigment dispersion in the shortest time. However, as the particle size of the media
decreases, the difficulty of separating the media balls from the pigment dispersion
increases. The factors that affect separation of the mill base from the media are:
.
Mill base viscosity. High viscosity increases the power necessary to rotate the shaft
and may surpass the design limit for rotating the shaft rapidly enough to disperse pigments. Also as the viscosity increases, the media beads are carried toward the screen,
leading to screen blockage.
21.4. DISPERSION EQUIPMENT AND PROCESSES
.
.
.
.
.
.
455
Flow rate. If the flow velocity is higher than the separation velocity, the beads are
carried to the screen.
Mill base density. It becomes harder to separate the dispersion from the beads as
density increases.
Bead size and density. As bead size decreases, separation of the beads centrifugally
becomes more difficult. As the density of the beads is increased, separation
becomes easier: from glass balls (density ¼ 2.6), to yttrium-stabilized zirconium
oxide (density ¼ 6), to steel balls (density ¼ 7.8), to tungsten carbide (density ¼ 14).
Agitator speed. Higher speed gives better separation, but there are practical limits.
Design of the separation system. The size of the screen is critical. There must be sufficient screen area in order that the velocity of the material passing through the screen is
not so high that media balls will be carried there and clog the screen.
Machine design. Horizontal pin mills are more efficient than the traditional horizontal
disk mills discussed above with small media balls.
Capital investment and operating costs of media mills are relatively low. Premixing is
required and let down must be done separately. Batch size is flexible. Cleaning is not
easy; generally, certain mills are reserved for similar colors, to minimize cleaning. Media
mills are effective for separating all but the most difficult pigment aggregates, provided
that the aggregates are small compared to the particles. Laboratory media mills are available
that correlate well with production mills. Instead of pumping the premix in from the bottom of
the mill, it is poured into the top of the mill. Horizontal laboratory mills are also available. The
use of media mills and related equipment is probably second only to that of HSD machines.
21.4.5. Three Roll and Two Roll Mills
Although formerly widely used to make dispersions for coatings, the use of three roll mills
is much more limited today. Figure 21.8 shows a schematic drawing of a three roll mill.
Aggregate separation results from the shear developed as the mill base passes through the
nip between each pair of rolls. The viscosity of the mill bases is higher than for the other
methods discussed thus far, 5 to 10 Pa.s or even higher. Since the mill base is exposed on
the rolls, solvents must have low vapor pressures to minimize evaporation.
Figure 21.8. Three-roll mill. (From Ref. [18], with permission.)
456
PIGMENT DISPERSION
Three roll mills have comparatively high capital and operating costs. Skilled operators are
required. Premixes are required and let down is a separate operation. On the positive side,
shear rates are high, so difficult to separate pigment aggregates can be processed, batch
size is versatile, and cleanup is relatively simple. Production use today is limited almost
entirely to small batches of dispersions with no solvent, or only low volatility solvent, to
make high viscosity dispersions. Three roll mills are convenient for laboratory use.
Two roll mills exert even greater shear rates than three roll mills. They are generally
used with solvent-free high MW polymer –pigment systems. They can separate even
the most difficult aggregates. They are most commonly used for dispersing very expensive
pigments. The high capital and operating costs are justified when the economic value of
achieving the last 10 to 20% of potential color yield creates a substantial difference in
product cost. Two roll mills are particularly appropriate when the dispersion is to be
used in a transparent coating. This requires that essentially all of the pigment aggregates
should be broken down to ultimate particle size to eliminate (or at least minimize) light
scattering. Another use is for dispersing certain carbon black pigments when the desired
jetness is attainable only with virtually complete separation.
Two roll mills have very high capital and operating costs. Volatile loss is complete.
Premixes are required, and further processing of the dispersion is required to make a
liquid dispersion that can be incorporated into a coating. Few coatings companies
operate two-roll mills; most purchase pigment dispersions of the type for which two
roll mills are appropriate from companies that specialize in making pigment dispersions.
21.4.6. Extruders
Extruders are used increasingly for pigment dispersion. Pigment dispersion for almost all
powder coatings is done in extruders (Section 28.4.1); also, some high viscosity liquid
dispersions are processed in extruders. Extruders have one or two screws feeding
through a cylinder and forcing the material out the end through a die. The barrel of the
extruder can be operated over a wide temperature range. Extruders can be configured so
that the shear imparted by the screw(s) is very high, and almost any pigment aggregates
can be broken down in an extruder. Difficult to disperse pigments require longer residence
times. In the case of solids, the product is usually chopped into small pieces and then pulverized either for use as a powder or for ease of dissolving.
21.4.7. Ultrasound Dispersion
Ultrasound for pigment dispersion was evaluated experimentally about 60 years ago. Lowfrequency ultrasound ( f 20 kHz) at intensities above the cavitation threshold generates
alternately growing and imploding bubbles in the liquid. The mechanical effect breaks
pigment aggregates into individual particles. In sonication times on the order of 15
seconds, tinting strength equals conventional pigment dispersions. It is reported that
scaling up of the process to industrial volumes is being studied [34].
21.5. EVALUATION OF DISPERSION
Assessment of degree of dispersion is a critical need for establishing original formulations
and optimizing processing methods as well as for quality control. Differences in degree of
21.5. EVALUATION OF DISPERSION
457
dispersion come from two factors: incompleteness of separation of the original aggregates
into individual crystals and flocculation after separation. Frequently, the coatings industry
does a poor job in this critical evaluation.
For white and colored pigments, the most effective evaluation method is by determination of tinting strength in comparison to a standard. For a white dispersion, one
weighs out a small sample of the dispersion and mixes into it a small weighed amount
of a standard color dispersion: for example, blue. One also weighs out a standard white
sample with the same ratio of the blue standard. After thorough mixing, one puts a
small amount of each tint mixture adjacent to the other on a piece of white paper and
draws down both samples with a stiff flat-ended spatula so that the edges of the two
samples touch each other. One can then compare the color of the two drawdowns. If
the batch is a darker blue than the standard, the tinting strength of the batch of white is
low, usually meaning that it is not equally dispersed. To test a color dispersion, say
blue, one carries out the same procedure but uses the standard white in both samples
and the standard blue and the batch of blue.
With dispersions of pigments that are to be used in applications such as automotive
metallic coatings, transparency is a critical requirement. Transparency can be tested by
drawing down a standard and a batch side by side on a glass plate. The degree of haze
or difference from the standard can be evaluated visually or measured instrumentally.
One can check for flocculation by pouring onto a flat surface some of the tint mix made
for the draw down. One then rubs the wet coating with a forefinger. If the color changes,
the dispersion is flocculated. For example, if a mix of blue with white becomes bluer where
it is rubbed, the blue pigment dispersion is flocculated. If the rubbed portion becomes a
lighter blue, the white is flocculated. Flocculation can also be detected by examining
the flow of the dispersion. Well-stabilized dispersions have Newtonian flow properties.
If a dispersion is shear thinning (and does not contain a component designed to make it
shear thinning), it is flocculated.
A further method of assessing pigment dispersion is by settling or centrifugation experiments. The rate of settling is governed by particle size and difference in density of the
dispersed phase from the medium. A well-separated, well-stabilized dispersion settles
or centrifuges slowly, but when settling is complete the amount of sediment is small. A
well-separated but poorly stabilized dispersion settles quickly to a bulky sediment. The
floccules settle more quickly because of their large size, and they form a bulky sediment
because continuous phase is trapped within the floccules. Flocculated sediment is readily
stirred, or shaken back to a uniform suspension, in contrast to the sediment formed from
the nonflocculated dispersion. If the pigment settles or centrifuges relatively quickly to a
compact layer, the separation step is incomplete. The larger aggregates settle more
quickly, but the volume is not large, because less internal phase is trapped in the sediment
than with a flocculated dispersion. Settling or centrifugation tests provide qualitative or
semiquantitative information sufficient for development work and quality control purposes. For research purposes, one can determine more quantitative data; the degree of flocculation can be calculated from rates of centrifugation [5].
One can also examine the dispersion with a microscope. Brownian motion is observed
in de-flocculated dispersions, but flocculation stops the motion. One must use caution in
preparing the samples for examination. In general, it is necessary to dilute the sample.
If the sample is diluted with solvent, there is a possibility of solvent-induced flocculation.
In such a case, one may report that the dispersion is flocculated when the flocculation was
actually a result of sample preparation. In assessing TiO2 dispersions with an optical
458
PIGMENT DISPERSION
microscope, for example, only the largest particles are seen. Electron microscope studies
of the surfaces of etched dry coating films can be useful for assessing variations in
dispersion [35].
Laser light scattering is probably the most precise way of measuring pigment particle
size and particle size distribution. Instruments are now available that make the measurements in liquid dispersions fast and reproducible [36]. Concentrated dispersions must be
diluted to perform the measurements, and shear forces are applied as the measurement is in
progress. These instruments are therefore better suited for measuring deagglomeration
than for measuring flocculation.
Use of infrared backscattering was pioneered by Balfour and Hird to study TiO2 dispersions [37]; they called it the flocculation gradient technique. Backscattering can be used to
measure the degree of pigment dispersion in both liquid coatings and dry films. The extent
of scattering of 2500-nm infrared radiation by a film as a function of film thickness was
measured. (Particles generally scatter longer wavelength radiation more effectively than
they do visible light.) There must be a significant difference between the pigment and
the vehicle in the refractive index at 2500 nm. A plot of backscatter against film thickness
gives a straight line whose gradient increases with increasing flocculation. Other examples
of the use of infrared backscattering measurements are given in Refs. [8] and [38].
The most widely used method of testing for fineness of grind in the coatings industry is
a drawdown gauge, commonly the Hegman gauge. Similar gauges are available, a sketch
of one such gauge being shown in Figure 21.9. A sample of the dispersion is placed on the
steel block before the zero reading and drawn down by a steel bar scraper. One then lifts
the block up and quickly looks across the drawdown sample to see at which graduation one
can start to see particles projecting or streaks caused by particles being dragged along. It is
said that the higher the scale reading, the “better” the dispersion.
Unfortunately, the device is not capable of measuring degree of dispersion. First, we
should remember that a major problem in making satisfactory dispersions is avoiding flocculation. However, the gauge cannot detect flocculation at all, since the drawing down step
breaks up any floccules. Next, the particle sizes of properly dispersed pigments are small
compared to the depth of the groove on the gauge. The depth on some gauges ranges from
0 to 10 mil (250 mm) in graduation units of 1.25 mil (approximately 30 mm). The gauge
shown in Figure 21.9 ranges from 4 mils (100 mm) to 0 in steps of 0.5 mil (12.5 mm).
Figure 21.9. Grind gauge and scraper for measuring the “fineness of grind” of pigment dispersions.
(From Ref. [18], with permission.)
REFERENCES
459
TiO2 pigment particles have an average size of about 0.23 mm, about two orders of magnitude smaller than the graduation steps of the gauge. Even aggregates of 104 to 105 particles could escape detection. Many color pigment particles are even smaller, and carbon
black particles can be as small as 5 nm. Some inert pigment particles are as large as 2 or
3 mm, still an order of magnitude smaller than the groove depths. Obviously, the gauge
cannot test whether or not all or most of the particles are less than the groove depths.
Blakely showed that in TiO2 dispersions only approximately 0.1% of the total pigmentation of a coating was responsible for an unacceptable fineness of grind rating [39].
Why are such devices used? They do give some idea of whether the big aggregates are
getting broken up, and they do give some idea of the presence of dirt particles. The determinations are fast, taking about 12 minute. A tinting strength determination by an experienced person requires 2 or 3 minutes, but after spending this time, one has an assessment
of the dispersion instead of a close to meaningless number. In the nineteenth century,
when very coarse, difficult to separate pigments were all that was available, the gauge
might have had some value. There is no excuse for using it in the twenty-first century.
GENERAL REFERENCES
Parfitt, G. D., Dispersions of Powders in Liquids, 3rd ed., Applied Science Publishers, London, 1981.
Patton, T. C., Paint Flow and Pigment Dispersion, 2nd ed., Wiley-Interscience, New York, 1979.
REFERENCES
1. Winkler, J.; et al., J. Coat. Technol., 1987, 59(754), 35, 45. Winkler, J.; Dulog, L., J. Coat.
Technol., 1987, 59(754), 55.
2. Rehacek, K., Ind. Eng. Chem. Prod. Res. Dev., 1976, 15, 75.
3. Saarnak, A., J. Oil Colour Chem. Assoc., 1979, 62, 455.
4. Dulog, L.; Schnitz, O., Proc. FATIPEC Congress, 1984, Vol. II, p. 409.
5. McKay, R. B., Proc. Int. Conf. Org. Coat. Technol., Athens, Greece, 1980, p. 499.
6. Brisson, A.; Haber, A., J. Coat. Technol., 1991, 63(794), 59.
7. Lara, J.; Schreiber, H. P., J. Coat. Technol., 1991, 63(801), 81.
8. Dasgupta, S., Prog. Org. Coat., 1991, 19, 123.
9. Jakubauskas, H. L., J. Coat. Technol., 1986, 58(736), 71.
10. Schofield, J. D.; Toole, J.; Polym. Paint Colour J., 1980, 10(24), December, 170.
11. Schofield, J. D., J. Oil Colour Chem. Assoc., 1991, 74, 204.
12. Jaffe, E. E.; et al., J. Coat. Technol., 1994, 66(832), 47.
13. Bugnon, P., Prog. Org. Coat., 1996, 29, 39.
14. Hamann, K.; Laible, R., Proc. FATIPEC Congress, 1978, p. 17.
15. Schroeder, J., Prog. Org. Coat., 1988, 16, 3.
16. Hill, L. W.; Wicks, Z. W., Jr., Prog. Org. Coat., 1982, 10, 55.
17. Bieleman, J., J. Oil Colour Chem. Assoc., 2004, 87(4), 173.
18. Patton, T. C., Paint Flow and Pigment Dispersion, 2nd ed., Wiley-Interscience, New York,
1979.
19. Hay, T. K., J. Paint Technol., 1974, 46(591), 44.
20. Morrison, W. H., Jr., J. Coat. Technol., 1985, 57(721), 55.
460
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
PIGMENT DISPERSION
Perrier, S., Surf. Coat. Int. B Coat. Trans., 2004, 87, (B4), 235.
White, D.; et al., U.S. patent 6,462,125 (2002).
Visscher, K. B.; McIntyre, P. F., U.S. patent 6,599,973 (2003).
Auschra, C.; et al., Prog. Org. Coat., 2002, 45, 83.
Creutz, S.; Jerome, R., Prog. Org. Coat., 2000, 40, 21.
Clayton, J., Surf. Coat. Int., 1997, 94, 414.
Banash, M. A.; Croll, S. G., Prog. Org. Coat., 1999, 35, 32.
Losoi, T., J. Coat. Technol., 1989, 61(776), 57.
Silber, St.; et al., Prog. Org. Coat., 2002, 45, 259.
Smith, R. E., J. Coat. Technol., 1988, 60(761), 61.
Wilker, G., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1999, pp.
360-372.
Paterson, J., Paint Coat. Ind., 2002, February 58; www.pcimag.com.
Way, H. W., JCT Tech, 2004, 1(1), 54.
Van Eecke, M. C.; Piens, M., Prog. Org. Coat., 2000, 40, 285.
Brisson, A.; et al., J. Coat. Technol., 1991, 63(801), 111.
Van, S. T.; et al., J. Coat. Technol., 2001, 73(923), 61.
Balfour, J. G.; Hird, M. J., J. Oil Colour Chem. Assoc., 1975, 58, 331.
Hall, J. E.; et al., J. Coat. Technol., 1989, 61(770), 73.
Blakely, R. R., Proc. FATIPEC Congress, 1972, p. 187.
22
Effect of Pigments on
Coating Properties
Although color and hiding are major reasons for including pigments in coatings, there are
many other effects that result from pigmentation. Pigmentation affects the viscosity of
liquid coatings as discussed in Sections 3.5 and 25.1.2. In this chapter we address the
effects of pigments on coating film properties. Most of these effects are related to the
pigment volume concentration (PVC) and the critical pigment volume concentration
(CPVC). Therefore, we discuss these concepts first. An extensive review of these
effects is available [1].
22.1. PVC AND CPVC
Traditionally, coatings formulators have worked with weight relationships, but volume
relationships are generally of more fundamental importance and practical significance.
Although there had been a few previous isolated examples of recognition of the importance of volume considerations in the performance of coatings, credit for full realization
of this importance belongs to Asbeck and Van Loo [2]. They looked at a series of performance variables as a function of PVC, that is, the volume percent of pigment in a dry film.
The term PVC should never be used to specify the volume of pigment in a wet coating.
This has been done occasionally in the literature and has been responsible for serious misinterpretations. Although PVC is usually expressed as a percentage, some authors express
PVC as pigment volume fraction in equations without bothering to tell the reader.
Asbeck and Van Loo observed that many properties of films change abruptly at some
PVC as PVC is increased in a series of formulations. They designated the PVC at which
these changes occurred as CPVC. They also defined CPVC as that PVC where there is just
sufficient binder to provide a complete adsorbed layer on the pigment surfaces and to fill
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
461
462
EFFECT OF PIGMENTS ON COATING PROPERTIES
all of the interstices between the particles in a close-packed system. Below CPVC, the
pigment particles are not close-packed and binder occupies the “excess” volume in the
film. Above CPVC, the pigment particles are close packed, and there is not enough
binder to occupy all of the volume between the particles resulting in voids in the film.
Slightly above CPVC, the voids are air bubbles in the film, and as PVC increases, the
voids interconnect and film porosity increases sharply.
22.1.1. Factors Controlling CPVC
There are large variations in CPVC, depending on the pigment or pigment combination in
a coating and the extent, if any, of pigment flocculation. With the same pigment composition, the smaller the particle size, the lower the CPVC. The surface area/volume ratio is
greater for smaller particle size pigments; hence, a higher fraction of binder is adsorbed on
the surface of the smaller pigment particles, and the volume of pigment in a close-packed
dry film is smaller. One can consider this effect from the standpoint of a modification of
the Mooney equation (Section 3.5): The volume fraction of internal phase is the sum of the
volume of pigment Vp and the volume of the adsorbed layer, Va. In a solvent-free system,
Vp equals PVC and Vp is equal to CPVC when Vp þ Va ¼ f. Thus, if Va is larger, the Vp
corresponding to CPVC is smaller. Note that at CPVC, the viscosity of the dry film
approaches infinity.
ln n ¼ ln nc þ
KE (V p þ Va )
1 (Vc þ Va )=f
(22:1)
CPVC depends on particle size distribution; the broader the distribution, the higher the
CPVC. As discussed in Section 3.5, broader particle size distribution of spherical,
dispersed-phase systems increases the packing factor. In low gloss coatings, the least
expensive component of the dry film is inert pigment; to minimize cost, it is desirable
to maximize the inert pigment content. Since properties should remain constant, one
does not want to change the PVC/CPVC ratio; the lowest cost systems are therefore,
those with the highest CPVC. Accordingly, cost is reduced by minimizing the amount
of very fine particle size pigment, since CPVC decreases with decreasing particle size,
and at the same time, by maximizing particle size distribution, which increases CPVC.
Compromises are needed, but broad particle size distribution is advantageous.
Pigment dispersion affects CPVC; CPVC of dry coatings in which the pigment is flocculated is lower than CPVC from corresponding coatings with non-flocculated pigment. Coatings with flocculated pigment clusters have non-uniform distribution of pigment and hence
are more likely to have portions where there are local high concentrations of pigment.
Solvent containing resin can be trapped inside clusters of pigment particles. When the
coating dries, the solvent diffuses out of the resin solution trapped in the floccules, leaving
behind insufficient binder to fill the spaces. In one example, it is reported that CPVC
decreased from 43 to 28 with increasing flocculation [3]. Asbeck suggests using the term
ultimate CPVC (UCPVC) to designate CPVC with a non-flocculated pigment combination
[3]. It seems to us better to recognize that flocculation is one of many variables that affect
CPVC, especially since it may be difficult to determine UCPVC experimentally.
CPVC with different pigments varies over a wide range, at least from 18 to 68%. One
might expect CPVC to be dependent on binder composition, since the composition affects
the thickness of adsorbed layers; thicker adsorbed layers might be expected to lead to
22.1. PVC AND CPVC
463
lower CPVC. Actually, CPVC with a given pigment or pigment combination seems to be
essentially independent of the binder composition (except in latex paints, which are
discussed in Section 22.2). It may be that the differences are too small to detect with the relatively imprecise methods used for determining CPVC. It may also be that the shrinkage forces
encountered when films are formed press the particles against each other to the extent that
only a minimal layer of binder is left between particles, regardless of the original adsorbed
layer thickness. No basic studies of this phenomenon have been reported.
22.1.2. Determination of CPVC
Critical pigment volume concentration has been determined by many different procedures
[4,5]. In many cases, the precision—that is, the reproducibility of the value—is relatively
poor, and the accuracy—that is, nearness to the supposed true value—is also sometimes
poor or indeterminate. A further complication is that there is no single true value
because CPVC depends on the method of measurement. Bierwagen has emphasized
that in the formation of films with PVC at or near CPVC, local fluctuations within a
film of volume fractions of binder and pigment are possible [6,7]. Thus, there could be
parts of a film with PVC . CPVC, whereas the average composition might be less than
CPVC. In view of these uncertainties, one must be careful in assessing the importance
of small differences in PVC and CPVC values.
Many changes in film properties have been used as a means of determining CPVC;
tinting strength is one of the most widely used. A series of white paints with increasing
PVC are prepared and tinted with the same ratio of color to white pigment. Above
CPVC, the white tinting strength of the coating increases due to the “white” air bubbles
above CPVC. It is suggested that the values of CPVC obtained by this and other optical
methods are somewhat low, due to nonuniform distribution of pigment, resulting in
parts of a film being above CPVC while others are below [6]. The technique is most
easily applicable to white paints, but can also be applied to colored paints.
Another method is to measure film density as a function of PVC. Since the density of
most pigments is higher than those of binders and the density of air is lower than those
of binders, density is maximized at CPVC. Due to nonuniform distribution of pigment in
the films, it is suggested that the density method tends to give high CPVC values [6]. The
CPVC can be determined by filtering a coating and measuring the volume of the pigment
filter cake. Asbeck recommends a specially designed filter that he calls a CPVC cell [3].
The CPVC for a pigment or pigment combination can be calculated from oil absorption
(OA) (Section 21.2.2), as shown in Eq. 22.2, provided that the OA value is based on either
a non-flocculated dispersion or determined at a sufficiently high shear rate that any floccules are separated. The definitions of both OA and CPVC are based on close-packed
systems with just sufficient binder to adsorb on the pigment surfaces and fill the interstices
between the pigment particles. Oil absorption is expressed as grams of linseed oil per
100 g of pigment, CPVC is expressed as milliliters of pigment per 100 mL of film, r is
the density of the pigment(s), and 93.5 is 100 times the density of linseed oil. (Both
OA and CPVC are expressed as percentages, not as fractions.)
CPVC ¼
1
1 þ (OA)(r)=93:5
(22:2)
The significance of the interrelationship depends on the observation that oil absorption
and CPVC are approximately independent of the binder, provided that the pigment
464
EFFECT OF PIGMENTS ON COATING PROPERTIES
particles are not flocculated. Since the accuracy of calculated CPVC depends on the accuracy of OA determinations, OA values determined by a mixing rheometer such as a
Brabender Plastometer (Section 3.3.2) are preferable to values determined by the
spatula rub up method [8]. Such an intensive mixer provides a greater shear rate for
separation of pigment aggregates than spatula rubbing; and also the data points are
taken while the dispersion is under high shear, so even if the binder does not stabilize
the pigment against flocculation, the volume fractions still represent a non-flocculated
system. The effects of variation in the procedure for determination of OA, including the
use of liquids other than linseed oil, are reported in Ref. [9]. The discussion of OA determination in this paper is interesting and useful; however, the conclusions drawn by the
author about CPVC are erroneous. The author failed to recognize that when the
pigment dispersions were diluted with solvent, the pigment flocculated.
Since many OA values are not accurate, Asbeck recommends against using such
calculations [3], but many workers have found the calculations useful. Although the
CPVC for individual pigments and for specific combinations of pigments can be calculated
from oil absorptions, CPVC values of pigment combinations cannot be calculated from
these values alone, since differences in particle size distribution with pigment combinations
affect the packing factor. The OA values for each combination of pigments can be determined experimentally. A variety of equations has been developed to calculate CPVC
from data on individual pigments [5]. The most successful equations use OA values, densities, and average particle sizes of the individual pigments [10,11]. The equations assume
that the particles are spheres, a fair assumption for many, but not all, pigments. Calculated
values correspond reasonably well to CPVC values determined experimentally.
22.1.3. CPVC of Latex Coatings
There is considerable controversy about the applicability of CPVC to latex paints. Some
maintain that one cannot apply the concept. Others maintain that there is no fundamental
difference in the CPVC concept between solventborne and latex paints. Still others maintain that the concept is useful in latex paints but that CPVC must be thought of differently
in latex paints than in solvent-based paints. Bierwagen emphasizes that experimental
errors in determining CPVC of latex paints are even greater than with solventborne
paints, so one must use caution in reaching conclusions [12]. Reference [13] includes a
review of the literature.
In a study of the effect of PVC on hiding of latex paints, it was concluded that CPVC
was lower in latex paint than in solvent-based paint made with the same pigment composition [14]. Patton recommended that the term latex CPVC (LCPVC) be used to distinguish this CPVC from CPVC of solventborne paints. Although CPVC is
approximately independent of the binder in solventborne paints, LCPVC varies with the
latex and some other components of latex paints. It has been found that LCPVC increases
as the particle size of the latex decreases. Also, LCPVC increases as the Tg of the latex
polymer decreases, and a coalescing agent increases LCPVC. Since LCPVC is smaller
than CPVC, the ratio of volume percent of binder in the film of a solvent based paint,
Vs, to that of a latex based paint, Vl, will always be less than 1 with the same pigment combination. This ratio has been called the binder index e [15]:
e¼
Vs
Vl
(22:3)
22.1. PVC AND CPVC
465
It has been proposed and demonstrated with a limited number of examples that this ratio
is independent of the pigment combination. If this proposal is generally true, it follows that
if one knows the binder index for a latex, one can calculate the LCPVC for paints made
with that latex and other pigments from the CPVC values calculated from oil absorptions
(or, for mixtures, by calculation from OA, density, and particle size).
The difference between CPVCs of latex- and solvent-based paints results from the
difference in film formation. To illustrate, let us consider a highly idealized comparison.
Let us assume that all the pigment particles in the solvent-based paint are spheres with the
same diameter and that the CPVC is 50%. In the latex paint, we will use the same pigment
together with spherical latex particles that have the same diameter as the pigment particles.
In the solvent-based paint, there is a layer of resin swollen with solvent on the surface of all
the pigment particles. Ideally, when a solventborne paint with PVC ¼ CPVC is applied,
the solvent evaporates and the “resin-coated” pigment particles arrange themselves in a
random close-packed order, with binder filling in the spaces between the pigment
particles. At the same ratio of pigment to binder as in the solvent-based paint, the idealized
latex paint would contain the equal numbers of latex particles and pigment particles. When
we apply a layer of the latex paint, the water evaporates and we get a closely packed
system of particles. However, some of the particles are pigment and some are latex.
There would not be a uniform arrangement of alternating latex and pigment particles in
a three-dimensional lattice; rather, there would be a statistical distribution of particles.
In some areas there would be clusters of pigment particles, and in other areas there
would be clusters of latex particles. As the film forms, the latex particles coalesce,
flowing around the pigment particles. However, the viscosity of the coalescing binder is
high, and it is difficult for the polymer to penetrate into the center of clusters of
pigment particles. As film formation proceeds, the water left inside the pigment clusters
diffuses out of the film, leaving behind voids. Although the PVC of 50% in the solventbased paint equaled CPVC, the same PVC of 50% in the latex paint results in a film
with PVC . LCPVC.
The probability of having clusters of pigment particles can be decreased by increasing
the number of latex particles until void-free films are obtained, but this necessarily reduces
PVC. If the Tg of the polymer is lower, the viscosity of the polymer at the same temperature is lower, and the distance that the latex can penetrate into clusters of pigment particles
increases. Thus, the LCPVC of paint made with lower Tg latex is higher, although still
lower than the CPVC of a solvent-based paint. Analogously, the reduction of the viscosity
of the polymer by a coalescing agent increases the LCPVC. It follows that the higher the
temperature during film formation, the higher the LCPVC should be, although no experimental test of this hypothesis has been published. It also follows, although again untested,
that LCPVC should depend on time. Although the viscosity of the polymer is high and
inhibits flow between the particles, the viscosity is not infinite, and perhaps in time,
voids would fill.
If a smaller particle size latex is used with the same ratio of pigment to binder
volume, the number of latex particles in our idealized paint would be larger than the
number of pigment particles. Now the probability of clusters of pigment particles
forming would be reduced. As a result, one would expect, as is found experimentally,
that the LCPVC increases with decreasing particle size of the latex. Particle size distribution of a latex affects its packing factor [16], which would presumably affect LCPVC.
A quantitative study of the effect of latex particle size on LCPVC using a series of
monodisperse vinyl acetate/butyl acrylate latexes with TiO2 pigment showed that
466
EFFECT OF PIGMENTS ON COATING PROPERTIES
LCPVC depended on the ratio of the number of latex and pigment particles and the ratio
of their diameters [13].
Not only does larger particle size latex decrease LCPVC, but flocculation of the latex
particles would also be expected to decrease LCPVC. If one formulates a latex paint so
that the PVC is slightly less than LCPVC with an unflocculated latex but the latex flocculates, the PVC would be greater than LCPVC. No data have been published on the
effect of latex flocculation on LCPVC. A simulation program using particle size distributions of latex and pigment and a measure of the deformability of latex particles has
been developed to predict the CPVC of simple latex paints [17].
Floyd and Holsworth proposed a different point of view about LCPVC [18]. They
suggest that a phase inversion occurs at LCPVC in which air becomes the external
phase and the pigment – latex phase becomes the internal phase. The proposal is based
on the information that latex films are somewhat porous at PVC values considerably
below LCPVC. The porosity is caused by clustering of pigment particles and of latex particles as described above. When films dry, there will be air pockets in the film; and as the
PVC increases, more and more of these air pockets join to make continuous passages of air
through the film. As PVC reaches LCPVC, the rate of increase of pore formation increases
rapidly with increased PVC. The result of this behavior is that there is not an abrupt change
of all properties at LCPVC. For example, scrub resistance shows a broad transition retaining significant scrub resistance above LCPVC. On the other hand, opacity increases quite
abruptly at LCPVC. Since porosity increases at PVC well below LCPVC (at PVCs of the
order of 20 to 30 PVC), barrier related properties such as permeability and corrosion resistance begin to change at these relatively low PVCs.
22.2. RELATIONSHIPS BETWEEN FILM PROPERTIES AND PVC
22.2.1. Mechanical Properties
Changes in volume relationships almost always control the physical properties of pigmented films. Thus, there is a major advantage to formulating coatings on the basis of volume
rather than weight relationships. For any particular application, there is a PVC/CPVC
ratio most appropriate for the combination of properties needed for that application.
Once this ratio has been established, changes in pigment combinations for that application
should be made such that this PVC/CPVC ratio is maintained. This important concept is
developed in detail by Bierwagen [4]. He also emphasizes that one must exercise particular care when formulating with PVC near CPVC, since relatively small changes in
pigment ratios or in packing, flocculation, or film formation can substantially affect film
properties [7].
Although it is desirable to have accurate determinations of CPVC available, even
reasonable estimates can be useful, since they permit one to focus a series of experiments within a range approximating the desired PVC/CPVC ratio. The decision as
to the proper pigment loading of a particular formulation should be based on actual
experimental data, not on the theoretically best PVC value. Even without CPVC data,
the concept is of value to the formulator. It is critical to recognize that performance
properties vary with volume, not with weight relationships, and that CPVC increases
with increasing particle size distribution. Qualitative use of oil absorption and density
values for individual pigments, together with recognition that CPVC increases when
22.2. RELATIONSHIPS BETWEEN FILM PROPERTIES AND PVC
467
mixtures of pigments with different particle size distributions are used, permit one to
start formulating in a reasonable range of compositions. On the other hand, using
weight relationships, one is working blindly.
In the majority of cases, Tg increases with PVC; however, there are examples in which
Tg decreases or is unchanged. It is common for there to be two Tg s, one resulting from the
continuous phase of binder and the other from the adsorbed layer of resin on the pigment
surfaces.
Tan delta is generally lower, broader, and the peak is shifted to a higher temperature for
heavily pigmented coatings than for unpigmented binder. The broadening and temperature
shift increase with increasing PVC/CPVC ratio and are more pronounced with lamellar
pigments than with spherical pigments. Figure 22.1 shows a schematic representation of
tan delta as a function of temperature for a pigmented coating compared to an unpigmented binder.
An increase in pigmentation usually increases the elastic modulus, especially in the
glass transition and rubbery regions, as shown in Figure 22.2. This results from the replacement of binder with rigid pigment. In Figure 22.2, CPVC is located between PVC 45
and 55%, as indicated by the lower E0 in the glassy region of the coatings pigmented at
PVC, of 55% and 60%.
The tensile strength of films generally increases with PVC to a maximum at CPVC but
then decreases above CPVC. Below CPVC, the pigment particles serve as reinforcing
particles and increase the strength. It can be considered that polymer molecules adsorb
on the surface of multiple pigment particles, providing the equivalent of cross-links.
Accordingly, more force is required to break this physical network as the pigment level
increases. However, above CPVC, air voids weaken the film; abrasion and scrub resistances of films drop above CPVC.
The relative tensile strength of an acrylic binder with a series of pigments is shown in
Figure 22.3. As expected, the tensile strength of films pigmented with TiO2 and barytes
rises to CPVC, then drops off. However, if the bonding between pigment particles and
the resin is poor, as with CaCO3, the tensile strength of the films drops steadily as PVC
increases. With microtalc, the structure of the pigment gives especially high reinforcement
at a low PVC.
Figure 22.1. Plot of tan delta as a function of T for a pigmented (P) and a non-pigmented (NP)
coating. (From Ref. [1], with permission.)
468
EFFECT OF PIGMENTS ON COATING PROPERTIES
Figure 22.2. Storage elastic modulus ( E0 ) as a function of temperature ( T) for a polyacrylate coating
with a different PVC (f) of TiO2. (From Ref. [1] with permission.)
The thermal expansion coefficient is affected by pigmentation. As shown in Figure 22.4,
the thermal expansion coefficient of a TiO2 pigmented epoxy coating decreases as the
PVC increases.
Internal stress in dried films of coatings whose Tg . T increases as PVC increases up to
CPVC and then decreases. Accordingly, the probability of delamination or film cracking
increases for such a system as a result of the stress.
22.2.1.1. Nano-pigment Effects Nano-pigments are those with small particle sizes;
some authors include those with average particle sizes of less than 100 nm, but more commonly the particle size is below 25 nm. The term is new, but nano-pigments have been
used in coatings for many years. For example, high strength channel blacks have particle
sizes in the range 5 to 15 nm. A review paper on nano-material technology applications in
coatings is available [19].
Compared to the effects of conventional pigments, the use of nano-pigments, especially
nano-fillers, has the potential to improve important coating properties. Improvements in
scratch, abrasion, heat, radiation, and swelling resistance, decrease in water permeability,
and increases in hardness, weatherability, modulus, and strain to failure are all possible
with nano-pigments. Of course, substantial improvements in one property may be
accompanied by undesirable changes in other properties, so formulators must be judicious
in the use of nano-particle pigments and fillers. The changes in properties reflect the high
surface area of nano-particles, which results in an increase in adsorption of resin on the
pigment surfaces and hence a larger fraction of immobilized resin. In some cases, the
films have two Tgs. When the particle size of the largest pigment particles or clusters is
22.2. RELATIONSHIPS BETWEEN FILM PROPERTIES AND PVC
469
Figure 22.3. Relative tensile strengths (ratio of pigmented film strengths to the strength of unpigmented
film) dependence on PVC for an acrylic system containing CaCO3 (†), microtalc (5), TiO2 (4), and barytes ( A). (From Ref. [1], with permission.)
smaller than the wavelength of light, coatings pigmented with nano-pigments are
transparent [1]. Based on an adsorbed layer of 10 nm, the fraction of resin adsorbed on
300-nm pigment particles has been calculated to be 3%, and on 50-nm particles to be
22% [19].
It can be difficult to make stable dispersions of nano-pigments. Due to their high surface
areas, nano-pigments have a strong tendency to agglomerate. If there is significant
agglomeration, the advantages of nano-pigments will not be realized. Various approaches
have been utilized, surface treatment of the pigments, in situ polymerization in the presence of the pigments, ultrasonic dispersion, and in some cases, conventional pigment dispersion [1]. Ball mills are used to disperse nano(aluminum silicate) in an acrylic resin with
a commercial dispersing agent. The dispersion is used in making mar resistant clear
coats [20].
470
EFFECT OF PIGMENTS ON COATING PROPERTIES
Figure 22.4. Thermal coefficient dependence on PVC at 218C and 0% RH for TiO2 pigmented
epoxy coatings. (From Ref. [1], with permission.)
22.2.2. Effects of Porosity
Stain resistance decreases above CPVC, since staining liquids can penetrate pores, leaving
color behind that is difficult to remove. Porosity also affects other properties. If one applies
a single coat of a coating with PVC above CPVC to steel and exposes the panel to humidity, rapid rusting can occur, since the pores permit water and oxygen to get to the surface
of the steel with little interference.
It is almost always desirable to make primers with a high PVC, since the rougher, low
gloss surface gives better intercoat adhesion than does a smooth, glossy surface. In most
cases, primers for metals should have a PVC that is high enough to roughen the surface and
to minimize the permeability of water and oxygen but that is lower than CPVC, where the
permeability increases. Minimum permeability often occurs at roughly PVC/
CPVC ¼ 0.9. In primers for non-metals it is sometimes desirable to formulate the
primer with PVC . CPVC. Adhesion of a top coat to such a primer is enhanced by mechanical interlocking, resulting from penetration of vehicle from the top coat into pores of
the primer. The primer PVC should be only enough higher than CPVC to assure good
adhesion. If PVC is too high, many of the pores in the primer are filled with binder
from the top coat, which decreases the PVC of the top coat, resulting in a loss of gloss.
Such a primer is said to have poor enamel hold out.
In zinc-rich primers it is important that the PVC be greater than the CPVC. The porosity
permits water to enter the film, establishing a conductive circuit with the steel surface.
Primers using 2K urethane coatings that have PVC . CPVC have the additional advantage that the porosity permits CO2 to escape without blistering.
PVC (Section 19.3) also affects hiding; as the pigmentation increases, hiding generally
increases. Initially, hiding increases rapidly but then tends to level off. In the case of rutile
TiO2, hiding goes through a maximum, gradually decreases with further increase in PVC,
REFERENCES
471
and then increases above CPVC. This increase in hiding above CPVC results from air
voids left in the film when PVC is above CPVC. The refractive index of air (1.0) is less
than that of the binder (approximately 1.5), so there is light scattering by the air interfaces
in addition to the interfaces between pigment and binder. The effect becomes large as
interfaces between air and pigment increase with increasing PVC. For example, if rutile
TiO2, with a refractive index of 2.73, is in the formulation, the difference between the
refractive index of TiO2 and air (1.73) is larger than that between TiO2 and binder
(1.23). For this reason, white ceiling paints are often formulated at PVC . LCPVC.
Tinting strengths of white coatings increase as the PVC of a series of coatings is increased
beyond CPVC. The air voids present above CPVC increase light scattering so that a
colored paint dries with a lighter color than one having the same amount of color
pigment but with PVC below CPVC.
If one applies an alkyd-based coating with PVC above CPVC to a wood substrate, one
is less likely to get blistering than with a similar coating having PVC below CPVC. When
water gets into the wood behind the coating, it can escape through pores in an alkyd
coating when PVC is above CPVC but not when PVC is below CPVC.
22.2.3. Effects on Curing and Film Formation
There are many examples of the effects of pigmentation on curing of thermosetting coatings. Titanium oxide coated with alumina inhibits the curing of coatings with MF resins as
the cross-linker. The basic alumina neutralizes the acid catalyst. Silica coated TiO2 is
acidic and does not inhibit the cure. In powder coatings, carboxy-functional resins and
alumina-coated TiO2 can interact, leading to higher yield values and higher viscosity.
Diallyl phthalate prepolymers catalyzed by dicumyl peroxide pigmented with silica
cure well. However, kaolinite virtually eliminates the curing because the acidic groups
present on the clay surface decompose the initiator [1]. On the other hand, UV-cure coatings pigmented with natural iron oxide gel in a matter of hours because of salts on the
surface of the pigment. Synthetic iron oxide pigmented coatings can be used without
difficulty.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Perera, D. Y., Prog. Org. Coat., 2004, 50, 247.
Asbeck, W. K.; Van Loo, M., Ind. Eng. Chem., 1949, 41, 1470.
Asbeck, W. K., J. Coat. Technol., 1992, 64(806), 47.
Bierwagen, G. P.; Hay, T. K., Prog. Org. Coat., 1975, 3, 281.
Braunshausen, R. W., Jr.; et al., J. Coat. Technol., 1992, 64(810), 51.
Fishman, R. S.; et al., Prog. Org. Coat., 1993, 21, 387.
Bierwagen, G. P., J. Coat. Technol., 1992, 64(806), 71.
Hay, T. K., J. Paint Technol., 1974, 46(591), 44.
Huisman, H. F., J. Coat. Technol., 1984, 56(712), 65.
Bierwagen, G. P., J. Paint Technol., 1972, 44(574), 46.
Hegedus, C. R.; Eng, A. T., J. Coat. Technol., 1988, 60(767), 77.
Bierwagen, G. P.; Rich, D. C., Prog. Org. Coat., 1983, 11, 339.
del Rio, G.; Rudin, A., Prog. Org. Coat., 1996, 28, 259.
472
EFFECT OF PIGMENTS ON COATING PROPERTIES
14. Anwari, F.; et al., J. Coat. Technol., 1990, 62(786), 43.
15. Patton, T. C., Paint Flow and Pigment Dispersion, 2nd ed., Wiley-Interscience, New York,
1979, p. 192.
16. Hoy, K. L.; Peterson, R. H., J. Coat. Technol., 1992, 64(806), 59.
17. Nolan, G. T.; Kavanaugh, P. E., J. Coat. Technol., 1995, 67(850), 37.
18. Floyd, F. L.; Holsworth, R. M., J. Coat. Technol., 1992, 64(808), 65.
19. Fernando, R., JCT Coat. Tech, 2004, 1(5), 32.
20. Vanier, N. R.; et al., U.S. patent 6,916,368 (2005).
23
Application Methods
Many factors affect the choice of method to be used for a particular application, including
capital costs, operating costs, film thickness, appearance requirements, and the structure of
the object to be coated. Reduction of VOC emissions and improved efficiency are driving
forces for improvement of application methods and equipment. Formulation of coatings
and development of application methods are strongly linked. Electrodeposition and application of powder coatings are discussed in Chapters 27 and 28, respectively. Further information and discussion of additional application methods can be found in the general
references at the end of the chapter.
23.1. BRUSHES, PADS, AND HAND ROLLERS
Brushes, pads, and hand rollers are frequently used for the application of architectural
paints. Although the same paints can usually be applied by spray gun, few do-it-yourselfers
use spray guns; on the other hand, to save time, professional painters use spray guns whenever possible.
23.1.1. Brush and Pad Application
A variety of brushes is available: narrow and wide, long handled and short handled, nylon,
polyester, and hog bristle [1]. Hog bristles are appropriate for solventborne paints but not
for waterborne paints. Nylon bristles are appropriate for waterborne paints but are swollen
by some solvents. Polyester bristle brushes can be used with either. Brushes have in
common a large number of bristles that hold paint in the spaces between the bristles.
When the paint is applied, pressure forces paint out from between the bristles. The forward
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
473
474
APPLICATION METHODS
motion of the brush splits the layer of paint so that part is applied to the surface and part
remains on the brush. Open-cell polyurethane foam “brushes” are also used.
The characteristics of paint viscosity are critical in using brushes. Pickup of paint on the
brush is controlled by paint viscosity at a relatively low shear rate, around 15 to 30 s21, the
shear rate of dipping and removing a brush from a paint can. If the viscosity is too high,
too much paint is brought out of the container with the brush; if the viscosity is too low, too
little paint is on the brush. Ease of brushing requires a low viscosity. The shear rate
between the brush and the substrate is relatively high, in the range 5000 to 20,000 s21
[2]. The viscosity at high shear rate controls the ease of brushing; high viscosity leads
to high “brush drag.” Film thickness applied is affected by the viscosity at a high shear
rate; applied film thickness increases with increasing viscosity. For most applications, a
viscosity at a high shear rate of 0.1 to 0.3 Pa.s is appropriate. (See Chapter 32 for
further discussion.) Solvents with relatively slow evaporation rates must be used to
slow the increase in viscosity of the paint on the brush.
When paint is applied by brush, the surface of the wet film has furrows called brush
marks. These furrows do not result from the individual bristles of the brush, as is
evident from comparing the size and number of brush marks with the bristles of the
brush. When polyurethane foam brushes are used, brush marks still result, even though
the brush has no bristles. Brush marks result from splitting the wet film between the
brush and the substrate as the paint is applied. Whenever layers of liquids are split, the
surfaces are initially irregular. These irregularities are made into lines by the linear movement of the brush. It is desirable to formulate a coating such that the brush marks flow out
before the film dries. Low viscosity promotes leveling but increases the probability of
sagging (Sections 24.2 and 24.3). Thixotropic flow properties are generally desirable
because they delay the increase in viscosity after brushing, permitting a compromise
between leveling and sagging. Formulation problems are particularly challenging in the
case of latex gloss paints (Section 32.3.2).
Pad applicators are also used in the do-it-yourself market [1]. The most common type of
pad consists of a sheet of nylon pile fabric attached to a foam pad that is attached to a flat
plastic plate with a handle. For low-viscosity coatings such as stains and varnishes, a
lamb’s-wool pad is used. Pads have a number of advantages over brushes. Pads hold
more paint than a brush of similar width and can apply paint up to twice as fast. In
general, pad application leaves a smoother layer than brush application. Extension
handles can be used with pads, reducing the need for moving ladders. Pads, especially
refills, are less expensive than brushes. On the other hand, pads require the use of trays,
which results in some paint loss and solvent evaporation. Cleanup of pads is more difficult
than cleanup of brushes.
23.1.2. Hand Roller Application
Hand rollers are the fastest method of hand application and are widely used in applying
architectural paints to walls and ceilings. A variety of rollers and roll coverings is available
[1]. There are rollers with built-in wells to minimize the need for dipping in the tray; there
are also power-filled rollers.
Viscosity requirements are similar to those for brush application. When paint is applied
by roller, there is also film splitting. As the roller moves, the film is stretched; flow is in an
extensional mode (Section 3.6.3) within the film. Due to imbalances of pressures, the film
breaks at different times, leaving a ribbed surface. Some of the film extends into fibers still
23.2. SPRAY APPLICATION
475
farther from the nip. As the roller moves, fibers are drawn longer and longer and eventually
break. After breaking, the ends of the fibers are pulled by surface tension to minimize the
surface area of the applied coating, but if there is not sufficient time at low viscosity for
leveling to occur, track marks are left in the coating. A further complication arises with
longer fibers that may break in two places; as a result, droplets of loose paint fly off to
land on the painter or on the floor. This is called spattering. If spattering could be eliminated, painting speed could be increased, and time spent on masking and laying drop
cloths could be reduced. However, all current formulations spatter, some much more
than others. The phenomenon is not completely understood. Glass [3– 5] has pointed
out that as the fibers are drawn out, extensional (elongational) flow rather than shear
flow is involved (Section 3.6.3). It follows that extensional viscosity rather than shear viscosity affects the development of fibers from roller application and, in turn, the roughness
of the film and the degree of spattering. Spattering can occur with any kind of paint, but is
particularly difficult to minimize in latex paints (Section 32.2).
23.2. SPRAY APPLICATION
Spraying is a common method for applying paints and coatings in architectural and
especially in industrial applications. Spraying is much faster than application by brush
or hand roller. Spraying is used on flat surfaces, but is particularly applicable to coating
irregularly shaped articles. Many different types of equipment are used for spraying; all
atomize the liquid coating into droplets. Droplet size depends on the type of spray gun
and coating; variables include air and fluid pressure, fluid flow, surface tension, viscosity,
and in the case of electrostatic application, voltage. The choice of spray system is affected
by capital cost considerations, efficiency of paint utilization, labor costs, and the size and
shape of objects to be coated, among other variables. Coating formulations must be
adjusted for the particular spray equipment and conditions.
The principal disadvantage is the inefficiency of application, since only a fraction of the
spray particles are deposited on the object being sprayed. One must mask work areas
where coating is not desired, and there is a likelihood of contaminating the area with
spray dust. Some of the droplets approach the surface and bounce back, carried by eddy
currents of air. The higher the pressure, the higher the forward velocity of the air and
as a result, the greater the percentage of bounce back. Some droplets miss the object
being coated, resulting in overspray. Some droplets may fall out of the spray pattern
under the force of gravity. Fallout is higher with large spray patterns and longer distances
between the spray gun and the surface being sprayed. The sum of all this waste determines
transfer efficiency; transfer efficiency is defined as the percentage of coating solids leaving
the gun that is actually deposited on the coated product.
Spray dust can cause many problems. If spray dust lands on the surface of a wet coating,
dirt contamination results. If the surface tension of the spray droplets is different than that
of the wet-surface, cratering will result (Section 24.4). Minimizing spray dust by choice of
application equipment, ventilation, and careful maintenance and cleaning of equipment
is vital.
Transfer efficiency is an important cost factor, since high transfer efficiency means that
less coating is used for the same surface area; VOC emissions are also reduced. Transfer
efficiency is affected by many variables. The size and shape of the product being coated is
a major variable. Transfer efficiency is low when spraying a chain link fence and high
476
APPLICATION METHODS
TABLE 23.1. Typical Baseline Transfer Efficiencies
Type of Spray Gun
Air
Airless
Air-assisted airless
High volume low pressure air
Electrostatic air
Electrostatic rotary
Transfer Efficiency (%)
25
40
50
65
60– 85
65– 94
Source: Ref. [6].
when spraying a large wall. Less obvious variables are conveyor line speed, localized air
flows, and the way that objects being coated are hung on a conveyor line. The spray
method has a major effect on transfer efficiency, whether the application is manual or automated; whether air, airless, or rotary spray guns are used; and whether an electrostatic
system is used. For manual systems, the skill of the sprayers is very important. For automated systems, system design is critical. Some factors involved are distance between the
gun and the surface, angle of the spray gun, stroke speed uniformity, extent and uniformity
of overlapping, and precision of triggering.
To compare transfer efficiency of different spray methods, the EPA, together with spray
gun and coatings manufacturers, has developed a standard procedure for comparing baseline transfer efficiency, adopted by the ASTM as Method D-5009-96. Measurement of
transfer efficiency is difficult because small changes in air flow can affect the results.
Typical percentages are given in Table 23.1; actual efficiencies are affected by many
variables besides the spray method. For example, transfer efficiency in coating a chain
link fence is low at best, and transfer efficiency in coating large wall expanses is high
in any case.
Very reactive coatings with short pot lives, such as some two-package polyurethane
coatings, require dual feed spray guns. Most commonly, the two components are
metered into a small, efficient mixing chamber just before the spray gun orifice. The
average dwell time in the mixing chamber is a fraction of a second, and the average residence time in the gun is not much longer. The equipment is designed to flush the mixing
chamber and the gun with solvent automatically, to prevent clogging when spraying is
interrupted. Frequent checking and maintenance are required to assure that the proper
ratio of components is being applied. In other types of equipment, the components are
sprayed from two orifices; mixing occurs after atomization so that the coating cannot
gel in the equipment.
23.2.1. Air Spray Guns
Air spray guns (Figure 23.1) cause atomization of a coating by fine streams of compressed
air. It is the oldest spray method and is still used. A stream of coating is driven through the
nozzle orifice at relatively low pressure, 10 to 50 kPa (1.5 to 7 psi), or by the suction
caused by rapid air flow past the outside of the orifice. The stream of coating coming
out of the orifice is atomized into small droplets by fine streams of compressed air at pressures of 250 to 500 kPa (35 to 70 psi). The degree of atomization is controlled by (1) the
viscosity of the coating (the higher the viscosity at the high shear rate encountered
going through the orifice, the larger the particle size), (2) air pressure (higher air pressure,
23.2. SPRAY APPLICATION
477
Figure 23.1. Cross section of an air spray gun and of spray gun nozzle (Delta Spray, Graco, Inc.). 1,
Wings or horns; 2, angular converging holes; 3, side-port holes; 4, annular ring around the fluid tip.
(Courtesy of Graco, Inc., Minneapolis, MN.)
smaller particle size), (3) the diameter of the orifice (smaller orifice, smaller particle size),
(4) pressure forcing or pulling the coating through the orifice (higher pressure, smaller
particle size), and (5) surface tension (lower surface tension, smaller particle size). The
outer jets of compressed air shown in the figure adjust the shape of the stream of atomized
particles coming from the gun. If these jets were not present, a cross section of the stream
would be roughly circular. Generally, an elliptical cross-sectional pattern permits more
efficient application; the pattern is a flattened cone, often called a fan. Guns can be
hand held or attached to robots.
Air spray guns are less expensive than other types of guns. Atomization can be finer
than with other spray guns. The system is versatile, and any sprayable object can be
sprayed with air spray. The level of control can be high if the operators are skilled.
However, transfer efficiency is the lowest of all the spray methods (Table 23.1).
A substantial improvement in transfer efficiency of air spraying has been made by use
of high volume low pressure (HVLP) air guns. These guns are designed to operate at lower
air pressures, 20 to 70 kPa (3 to 10 psi), but with higher air volumes and large, unrestricted
air passages to handle a large volume of air. Because of the low pressures, bounce back is
478
APPLICATION METHODS
reduced and transfer efficiencies of 65% or even higher can be achieved. HVLP guns are
used increasingly in automotive repair shops. California South Coast VOC regulations
require 65% transfer efficiency and air pressures of 70 kPa or less. An alternative approach
to HVLP guns is low volume low pressure (LVLP) guns. In LVLP guns, the air pressure
is also less than 70 kPa, but the air volume is reduced by mixing the air and coating inside
the gun.
23.2.2. Airless Spray Guns
With airless spray guns, coating is forced out of an orifice at high pressure, 5 to 35 MPa.
The coating comes out of the orifice as a “sheet.” As the sheet extends in moving away
from the orifice, flow instabilities cause the formation of ligaments, followed by further
disintegration into droplets [7]. Atomization is controlled by the relative velocity
between the sheet and the contiguous air (higher relative velocity, smaller droplets), viscosity (higher viscosity, larger particle size), pressure (higher pressure, smaller particle
size), and surface tension (lower surface tension, smaller particle size). The shape of
the fan or pattern of the spray is controlled by the orifice size and shape. Air-assisted
airless spray guns are also available; the atomization is airless but there are external
jets to help shape the fan pattern, confining the smaller droplets within the spray pattern
[8]. Both hand held and robot airless guns are available.
The size of droplets from airless spray guns is larger than those from air spray guns, 70
to 150 mm as compared with 20 to 50 mm [9]. Airless guns give a fishtail spray; that is,
there is a relatively sharp edge to the spray droplet fan, with quite uniform droplet distribution within the fan. In contrast, the fan from air guns is feathered at the edge; that is, the
number of droplets decreases at the edge of the fan, with some being quite widely spaced.
As a result of these differences, one can generally achieve more uniform film thickness
with air spray than with airless spray; air-assisted airless application gives intermediate
results.
Coatings can be applied more rapidly by airless than by air guns, permitting more rapid
production. However, as the application rate increases, the likelihood of applying excessive coating thickness also increases, particularly when coating objects with complex
shapes. Excess coating thickness is not only wasteful, but may also lead to sagging.
Since there is not a stream of compressed air accompanying the particles and because
the droplet size is generally larger, there is less solvent evaporation from the atomized particles from airless guns than from air guns. Solvents with higher relative evaporation rates
are generally required in formulating coatings for airless spray application.
The absence of the air stream in airless guns reduces the problem of spraying into
closed recesses of irregularly shaped objects. On the other hand, spraying down a recessed
section that is open on the opposite end is easier with an air gun, since the air stream helps
to carry the particles along. Airless spray equipment can cause problems with some waterborne coatings. More air is dissolved in the water under the high pressure; it comes out in
bubbles when the pressure is released as the particles leave the gun. Entrapment of bubbles
in the film can result in pinholing (Section 24.7). Also, with some latex coatings, particle
size of particles leaving the gun are larger than would be expected from the shear viscosity
of the coating. It has been shown that the extensional viscosity is relatively high in
such coatings [5].
2K coatings can be applied with dual feed airless spray equipment, which permits
combining the advantages of airless spray and dual spray [10]. Such equipment has
23.2. SPRAY APPLICATION
479
been developed to the point where it is possible to spray coatings that cure within a few
seconds, enabling use of 2K polyurea coatings (Section 12.4.1). An aerosol coating container is a type of airless spray unit. A liquefied gas, commonly propane, supplies the
pressure to force the coating out of the orifice. Since the pressure is relatively low, the viscosity of the coating must be low to achieve proper atomization.
23.2.3. Electrostatic Spraying
Transfer efficiency can be substantially higher with electrostatic spray units (Table 23.1).
In the simplest case, a wire is built into the orifice of the spray gun. An electric charge on
the order of 50 to 125 kV is impressed on the wire. At the fine end of the wire, an electric
discharge leads to ionization of the air. As atomized coating particles pass through this
zone of ionized air, they pick up a negative charge. The object to be coated is electrically
grounded. When the coating particles approach the grounded surface of the object, the
differential in charge attracts the particles to the surface. Taking the example of a chain
link fence, an increased fraction of coating deposits on the metal fence even to the
extent that particles that had passed through the holes are attracted back to the back
side of the fence. This wrap around effect permits coating both sides of the fence by spraying from only one side; there is a lower but still fairly high overspray loss even with electrostatic spraying. In large-scale production, most of the paint is applied precisely by
computer-controlled robots. With objects such as automobile bodies, overspray losses
can be reduced by over 50%, resulting in transfer efficiencies above 80%.
Different types of devices can be used for electrostatic spraying. Figures 23.2 and 23.3
show bell- and disk-type spray units called rotary atomizers. In both cases, coating is
pumped through a tube leading to the middle of the disk or bell; the unit is rotated at a
speed of about 900 rpm or higher, depending on the diameter, and the coating flows out
to the edge of the unit. As the coating is thrown off the edge of the unit, it is atomized
Figure 23.2. Bell electrostatic spray equipment. (From Ref. [1], with permission.)
480
APPLICATION METHODS
Figure 23.3. Disk electrostatic spray equipment. (From Ref. [1], with permission.)
and then passes through an electrostatic field that charges the droplets. Droplet size is controlled by viscosity and the peripheral speed of the unit. Small (2 to 6 cm in diameter) bells
that spin at speeds up to 60,000 rpm permit the application of coatings with viscosities as
high as 1.5 to 2 Pa.s, compared to 0.05 to 0.15 Pa.s used with conventional speed disks and
bells as well as spray guns. The spray pattern is adjusted with shaping air. This ability to
handle higher viscosity coatings permits use of less solvent. However, it has been reported
that application of pigmented coatings by high speed electrostatic bells tends to give films
with lower gloss [11]. It is suggested that the centrifuging effect leads to differences in
pigment content among atomized particles, leading to formation of uneven films. The uniformity of application and transfer efficiency can often be improved by compressed air
shaping of the cloud of atomized coating, analogously to the effect of the shaping air in
conventional air guns.
Low overspray losses and good wrap around with electrostatic spray depend on the
charge pickup by the atomized particles, which is controlled by the coating conductivity.
If it is too low, the particles do not pick up sufficient charge from the ionized air. This is
most likely to be a problem with coatings that have only hydrocarbon solvents, especially
aliphatic hydrocarbons. Nitroparaffin or alcohol solvents can be substituted for a portion of
the hydrocarbon solvents with good results. For coatings with free carboxylic acid groups
on the resin, addition of a small amount of a tertiary amine such as triethylamine increases
conductivity sufficiently to provide good charging. If the conductivity is too high, there is
increased danger of electrical shorting. Generally, conductivity is measured, but the results
are expressed as resistivity; optimum resistivity varies with equipment and the coating
operation, ranging from 0.05 to 20 MV. Newer gun designs permit operation with a
wide range of conductivities.
The conductivity of waterborne coatings is higher than that of solventborne coatings;
resistivities on the order of 0.01 MV have been reported [12]. The addition of slow evaporating water-miscible solvents with nonpolar ends, such as 2-butoxyethanol, reduces
surface conductivity somewhat. However, spray equipment must be specially designed
so that the coating line is isolated electrically; otherwise, the charges will be dissipated
and the atomized droplets will not be charged. Furthermore, the hazard of shocks would
23.2. SPRAY APPLICATION
481
be high. Although the entire application line could be isolated electrically, this is
expensive and there is still a hazard of electric shocks. Alternatively, just the atomizer
and a short hose connected to a voltage blocking device can be charged [13]. Another
alternative to electrical isolation is to fit the gun with an external probe that charges the
atomized paint. These expedients permit effective charging of the atomized droplets
and reduce the hazard of shocks at substantially lower cost than that to isolate the
entire application line.
Electrostatic spray application is not without limitations. The substrate must be electrically conductive so that the necessary charge differential can be set up by grounding the
object to be sprayed. It can also be difficult to get coating into recessed areas even with
airless electrostatic spray application, due to the Faraday cage effect. A Faraday cage
results from the pattern of field lines between the electrode on the gun and the grounded
object. The strong electrical field induced by the difference in voltage establishes field
lines that the atomized particles follow between the gun and the object. However, areas
surrounded by grounded metal, such as inside corners of a steel case, are shielded from the
electric fields by the metal; a Faraday cage is established and few particles enter such a
shielded area.
Electrostatic spray is the most important method for applying powder coatings
(Section 28.5.1).
23.2.4. Hot Spray
Since the viscosity of coatings must be relatively low, generally, 0.05 to 0.15 Pa.s, for
proper atomization, the solids must be quite low, especially for lacquers made with relatively high molecular weight (MW) resins. An approach for increasing solids is use of hot
spray, also called temperature conditioned spray. A hot spraying unit is designed with a
heat exchanger to heat the coating to temperatures from 38 to 658C. The unit is designed so
that the coating is recirculated when the gun is turned off, even temporarily. At the elevated temperature, the viscosity is reduced sufficiently to permit a significant increase in
solids. Although these systems were originally designed for use with quite low solids coatings such as lacquers, they are useful for high-solids coatings. For example, it is reported
that the solids of a top coat for appliances can be increased from 55% to 65% using a heater
before a disk electrostatic spray gun [2]. The viscosity of high-solids coatings generally
decreases more sharply with increasing temperature than that of conventional coatings,
which is a desirable characteristic for hot spray application. A further advantage is that
the temperature drops between leaving the spray gun orifice and arrival on the work,
leading to a correspondingly large viscosity increase. This characteristic is desirable for
reducing sagging, which can be serious with high-solids coatings (Section 24.3).
Another advantage is that temperature control eliminates the effects of variation in
ambient temperature on viscosity of the coating.
23.2.5. Supercritical Fluid Spray
Some industrial coatings are applied by supercritical fluid spray [14,15]. The supercritical
fluid of choice is carbon dioxide, which has a critical temperature of 31.38C and a critical
pressure of 7.4 MPa. In the supercritical state, carbon dioxide exhibits solvency characteristics similar to that of hydrocarbons, but it is not counted as VOC. A dual feed gun is used
with a low solvent content coating as one feed and supercritical CO2 as the other. The
482
APPLICATION METHODS
temperature must be controlled and the pressure must be above 7.4 MPa. The process has
been used with a variety of solventborne coatings for many applications [14]. VOC is
reduced by 30 to 90% without changing the MW of the resins. Waterborne coatings can
also be applied using the process [14].
The high pressure means that expensive equipment and airless guns must be used.
The supercritical fluid spray system minimizes some of the problems of utilizing
airless spray with conventional coatings [15]. When the coating leaves the orifice of
the spray gun, the very rapid vaporization of the CO2 breaks up the atomized droplets,
reducing the particle size and resulting in a narrower particle size distribution. The
droplet size is comparable to that obtained with air spray guns and significantly
smaller than that obtained with airless guns. Furthermore, the fan pattern is more
similar to the feathered pattern obtained with air guns rather than the usual pattern
for airless guns (Section 23.2.2). Loss of the CO2 is essentially complete before the droplets reach the surface, so the viscosity of the applied film is relatively high, minimizing
sagging. Transfer efficiency is also improved, reducing overspray and waste disposal
problems.
23.2.6. Formulation Considerations for Spray-Applied Coatings
The formulation of solvent mixtures for spray-applied coatings and show effective use of
spray guns requires taking into consideration the large surface area/volume ratio of the
atomized coating droplets and the flow of air over the surface of those droplets. As discussed in Section 18.3, these factors affect the rate of solvent evaporation. When lacquers
are applied by spray, over half the solvent in the coating may evaporate between the
orifice of the spray gun and the surface of the object. If the solvent mixture is balanced
properly and the spray gun is used properly, a relatively smooth, sag-free film can be
obtained. Otherwise, sagging may occur when not enough solvent is evaporated or
rough surfaces may result from dry spray. Low viscosity of the coating after it arrives
on the surface generally facilitates leveling, but it also increases sagging (Sections
24.2 and 24.3).
Proper control requires a careful balance of solvent evaporation rate with the particular
spray equipment and procedure. The greater the distance from the spray gun orifice to the
work, the greater the fraction of solvent lost. Coatings are formulated to work best at a
specific distance between the gun and the surface. This distance should be kept as constant
as possible throughout the spraying operation. The problem can be illustrated by the result
of a person spraying lacquer on a flat vertical surface by holding one’s arm in a constant
position and bending one’s wrist to spray a wider area of the surface. When the gun is
aimed perpendicular to the substrate surface, the distance is at a minimum. When the
wrist is bent to the furthest degree, the distance is at a maximum. If the solvent mixture
is properly balanced for an intermediate distance, the lacquer sprayed perpendicularly
would be likely to sag; the lacquer film at the extreme distance would tend to show
poor leveling. In extreme cases of poor spraying, both sagging and rough surfaces can
occur on the same object.
The rate of solvent loss is affected by the degree of atomization. If the average particle
size is smaller, the surface/volume ratio is higher and the extent of solvent loss is greater.
The rate of solvent evaporation is also affected by air flow over the surface of the droplets.
More solvent evaporates with air guns than with airless guns or spinning bells. The rate of
air flow through the spray booth can affect the degree of solvent loss. Temperature in the
23.2. SPRAY APPLICATION
483
spray booth can be an important factor; during hot weather, it is common to change the
solvent mixture to slow the rate of evaporation. When formulating and testing a solvent
mixture for a coating in the laboratory, the same type of spray gun at approximately the
same distance from the work that will be encountered in the customer’s factory should
be used. A final adjustment must be made in the customer’s plant under regular production
conditions. If those conditions change, the solvent mixture has to be changed. This is one
reason why industrial coatings are almost always shipped at higher concentrations than the
customer will actually use. This permits modifications of both solvent levels and solvent
composition (by changing the reducing solvent) to accommodate temperature changes in
the spray booth and other variables.
The viscosity of the coating must be adjusted to obtain appropriate atomization for the
spray gun being used. The critical viscosity for atomization is that at the high shear rates
(103 to 106 s21 [2]) encountered as the coating passes through the orifice of the gun. Coatings with high low-shear viscosity can be sprayed successfully if they are shear thinning.
The upper limitation on viscosity at a low shear rate is the need to have a satisfactory flow
rate of coating through the tubing to the spray gun, which varies from gun to gun. Architectural and maintenance paints, which must not sag significantly when thick films are
applied to large wall expanses, are shear thinning. Generally, the same coating used for
brush application can be used for spray; in the case of latex paints with some spray equipment, the viscosity at a low shear rate may have to be reduced by diluting with a small
amount of water.
Many conventional industrial coatings exhibit Newtonian flow; however, most highsolids coatings and waterborne coatings are shear thinning. Viscosity for application is
generally checked with an efflux cup (Section 3.3.5). However, efflux cups can be misleading because they do not detect shear thinning or thixotropic flow properties. They should
only be used for control purposes. The proper way to establish viscosity for production
spraying is by using the production spray gun under the conditions of use. Then, having
found the optimum formulation, one can establish an efflux cup time as the standard for
that coating for use in that gun under those circumstances.
Many water-reducible and high-solids coatings are shear thinning after reduction to
spray viscosity, which means that to give proper atomization, the efflux cup time of the
reduced waterborne coating will be longer than with most solventborne coatings. Atomization is controlled by the viscosity at high shear rate, and efflux cup time is controlled by
the viscosity at a lower shear rate. Sagging of some spray-applied high-solids coatings
cannot be controlled by adjustment of solvent evaporation rates. In such cases, the
coating has to be formulated so that it is thixotropic (Section 24.3).
In spray application of waterborne coatings, air bubbles can be entrapped in the coating
film. If they persist until the viscosity of the film surface becomes too high for the bubbles
to escape, pinholing or popping will occur. The troublesome air voids are tiny (10 mm).
The problem can be particularly severe in airless spray application, since more air dissolves in the coating due to the high pressure. When the pressure is released as the
coating leaves the spray gun, bubbles of air form in the droplets. Whereas larger air
bubbles can escape by rising to the surface and breaking before the viscosity at the
surface increases, microform bubbles dissipate only by dissolution of the air in the coating
followed by diffusion through the film [16]. In air spray, HVLP spray, and air-assisted
airless spray, the problem can be minimized by using CO2 as the driving gas instead
of air. The improvement results from the higher solubility of CO2 than of air in a waterborne coating.
484
APPLICATION METHODS
23.2.7. Overspray Disposal
In industrial production, the advantage of decreased overspray is not just savings in the
cost of the lost coating and lower VOC emissions. Any overspray has to be trapped so
that it does not contaminate the surrounding area. Water-washed spray booths (a spray
booth in which the wall behind the work being sprayed is a continuous waterfall that is
recirculated) are widely used. The overspray is collected as a sludge. Although it is sometimes possible to rework the sludge into low-grade coatings, generally the sludge must be
disposed of in approved hazardous solid waste disposal landfills. Such disposal is expensive; increased transfer efficiency can reduce waste disposal cost.
Although water-washed spray booths work efficiently with solventborne coatings, the
separation of sludge is less efficient when some waterborne coatings are sprayed. The
overspray does not coagulate completely when it strikes the water as it does with solventborne coatings. This makes separation more difficult and can limit recirculation of the
water. Froth flotation methods adapted from ore recovery processes permit relatively
rapid separation of sludge [17]. Separation of sludge from waterborne coatings can be
improved by adding to the water tank an emulsion of a melamine – formaldehyde resin
together with water-soluble cationic and/or nonionic acrylamide polymers, which
promote detackification and coating flocculation, respectively [18].
23.3. DIP AND FLOW COATING
Dip coating can be an efficient procedure for applying coatings; it offers both relatively
low capital cost equipment and low labor requirements. The principle is simple: The
object is dipped into a tank of coating and pulled out; excess coating drains back into
the dip tank. In practice, satisfactory coating application by dipping is more complex.
While excess coating is draining off the object, a gradation of film thickness develops;
the coating thickness at the top of the object is thinner than at the bottom. During draining,
solvent is evaporating. The differences in film thickness can be minimized by controlling
the rate of withdrawal of the object from the dip tank and the rate of evaporation of the
solvent. If the object is withdrawn slowly enough and the solvent evaporates rapidly
enough, film thickness approaching uniformity on vertical flat panels can be achieved.
In actual production, the rate of withdrawal is usually faster than optimum, so there is
some thickness differential between the top and bottom of the object.
Care must be exercised in selecting or changing solvents for dip coating, owing to
flammability hazards and changes in viscosity of the dip tank that result from evaporation
of solvent. Changes in viscosity result in changes in film thickness, which increases with
increasing viscosity. Achieving consistent results requires keeping the viscosity of the
coating constant, which becomes more difficult as the volatility of the solvent is increased.
Solvent is added to replace solvent lost from the tank.
Successful use of dipping in production lines requires that the coating be very stable.
Viscosity can increase not only by loss of solvent, but also by chemical reactions of
coating components. Only a small fraction of the coating in the tank is removed each
time an object is dipped. Fresh coating is added to make up for these removals, but the
coating in the tank is a mixture of old and newly added coating. Some of the original
charge of coating is present for a long time. If cross-linking reactions occur in the bath,
viscosity increases. The extent of such reactions must be minimal. While the viscosity
23.4. ROLL COATING
485
can be reduced by adding more solvent, this reduces the solids, so dry film thickness is
reduced. When oxidizing alkyds are the vehicle for a dip coating, an antioxidant must
be added to avoid oxidative cross-linking. But the antioxidant must be sufficiently volatile
that it escapes during the early stages of the baking cycle; otherwise, it inhibits crosslinking of the dry film. On the other hand, too high volatility would result in evaporation
from the dip tank. Isoeugenol is an example of a suitable antioxidant for alkyds in dip
coatings.
An advantage of dip coating is that all surfaces are coated, not just the outer surfaces
accessible to spray. However, there are difficulties in dipping irregularly shaped objects.
Coating may be held in pockets or depressions, giving pools of coating that do not
drain. To minimize this problem, the point(s) at which the objects hang on the hooks of
the conveyor line that carries it into and out of the tank must be designed and selected carefully. Objects to be dip coated must be designed with drain holes that minimize pooling but
do not interfere with the performance or appearance of the product. Lower edges,
especially lower corners, build up high film thickness. Buildup can be reduced by electrostatic de-tearing, that is, by passing the object over a highly charged electrode that causes a
charge concentration at such points, causing drops to be pulled off. Waterborne coatings
are supplanting solventborne dip coatings in many applications. They reduce flammability
hazards and VOC emissions. Electrodeposition of coatings (Chapter 27) is a dipping
process.
Flow coating and dip coating are related methods. Objects to be flow coated are carried
hung on a conveyor through an enclosure in which streams of coating are squirted on the
object from all sides. The excess coating runs off and is recirculated through the system.
The volume of coating required to fill the lines of a flow coater is less than that required for
a dip tank for objects of the same size. This reduces inventory cost and increases coating
turnover, hence reducing somewhat the problem of bath stability. Flow coaters are
designed so that the atmosphere in the enclosed area is maintained in a solvent saturated
condition. In this way, evaporation of solvents is minimized until flow and leveling have
been obtained. There is still a gradation of film thickness from the top to the bottom of the
object, but usually less than that from dip coating. Highly automated flow coating lines
have been used for applying coatings to major appliances. The design permits more
rapid line speeds than with conventional dip tanks.
23.4. ROLL COATING
Roll coating is widely used and is efficient, but is applicable only to uniform flat or cylindrical surfaces. Relatively slow evaporating solvents must be used to avoid viscosity
buildup on the rolls of the coater. The pot life of the coating must be relatively long,
since the rate of turnover of coating through the system is relatively low. There are
many types of roll coating procedures; the two most common are direct roll coating
and reverse roll coating.
In direct roll coating, the stock passes between two rollers, an applicator roller and a
backup roller, which are rotated in opposite directions. The rollers pull the material
being coated between them as shown in Figure 23.4. Direct roll coating is used for
coating sheet stock and sometimes for coil stock. In direct roll coating, the applicator
rollers are generally covered with a relatively hard polyurethane elastomer. Coating is
fed to the applicator roll by a smaller feed, or doctor, roll that is, in turn, fed by a
486
APPLICATION METHODS
Figure 23.4. Direct roll coater. (From Ref. [1], with permission.)
pickup roll. The pickup roll runs partially immersed in a tray, called a fountain, containing
the coating. Film thickness is controlled by the clearance between the feed and applicator
rolls and by the viscosity.
Several variations on direct roll coating use different types of applicator rolls. The
applicator rolls can have sections cut out so that they do not coat the entire substrate
surface. The applicator roller can be engraved with small recessed cells over its entire
surface. The cells are filled with coating, the surface is scraped clean with a doctor
blade, and the coating remaining in the cells is transferred to the substrate being coated.
Such precision coaters apply a controlled amount of coating only to the areas of the substrate that contact the roller. In another variation, the applicator roller is a brush roller for
applying thick coatings to relatively rough surfaces.
In direct roll coating, as the coated material comes out of the nip between the rollers, the
wet layer of coating is split between the roller and the substrate. As the coated stock comes
out of the nip, the release of pressure results in cavitation. As the coated stock moves away
from the nip, the walls between the cavities become filaments. Further along, filaments
may break and fall back on the fresh coating, resulting in a ribbed surface called roll tracking. The coating must be designed to level so that this tracked appearance becomes less
obvious, before the coating stops flowing. To minimize solvent loss on the rolls and to
keep the viscosity low to promote leveling, slow evaporating solvents are used, and
when possible, coatings are formulated so they exhibit Newtonian flow. The rate of
shear on the coating as it passes through the nip depends on speed and gap distance but
is on the order of 104 to 105 s21. If the coating exhibits thixotropy, it is desirable to
have the rate of recovery of the low shear viscosity be as slow as possible.
As the coating comes out of a nip, the filaments may grow quite long. When they finally
break, at least some split in more than one place, leading to drops being released, called
misting (Figure 23.5). Extended filaments result when extensional viscosity is high [5].
In reverse roll coating, the two rollers rotate in the same direction, to the material being
coated must be pulled through the nip between the two rolls, as shown in Figure 23.6. It is
generally not feasible to coat sheets by reverse roll coating, but the process is widely used
for coating coil stock. Reverse roll coating has the advantage that the coating is applied by
23.5. CURTAIN COATING
487
Figure 23.5. Cavitation and misting in roll application. (From Ref. [5], with permission.)
Figure 23.6. Reverse roll coater for coating both sides of coil stock. (From Ref. [1], with permission.)
wiping rather than by film splitting. A smoother film is formed, and the problems of
leveling are minimized.
23.5. CURTAIN COATING
Curtain coating is widely used for coating flat sheets of substrate material such as wall
panels and metal doors. Coating is pumped through a slot in the coating head so that it
flows out as a continuous curtain. The sheet to be coated is carried through the curtain
on a conveyor. The curtain must be wider than the substrate being coated, to avoid
edge effects on the film thickness. A recirculating system returns to the coating reservoir
the overflow from the sides and from between sheets. A schematic diagram is shown in
Figure 23.7.
The width of the slot, the pumping pressure, the viscosity of the coating, and the rate of
passage of the substrate being coated control film thickness. The faster the line runs, the
thinner the coating. Where applicable, curtain coating is an excellent method. No film
splitting is involved, so the film that is laid down is essentially smooth. Film thickness
can be very uniform. The coating must be very stable; solvent is added to make up for
488
APPLICATION METHODS
Figure 23.7. Curtain coater. (From Ref. [1], with permission.)
loss by evaporation. If a particle of low surface tension lands on a flowing curtain, surface
tension differential driven flow can lead to a hole in the curtain that in turn leads to a gap in
the coating being applied to the substrate (Section 24.4). As the surface tension of the
coating is increased, particles in the air are more likely to have lower surface tensions
than that of the coating, and the probability of holes is increased.
GENERAL REFERENCES
Goldschmidt, A.; Streitberger, H.-J., BASF Handbook on Basics of Coating Technology, Vincentz,
Hannover, Germany, 2003, pp. 437 – 617.
Levinson, S. B., Application of Paints and Coatings, Federation of Societies for Coatings Technology, Blue Bell, PA, 1988.
REFERENCES
1. Levinson, S. B., Application of Paints and Coatings, Federation of Societies for Coatings
Technology, Blue Bell, PA, 1988.
2. Schoff, C. K., Rheology, Federation of Societies for Coatings Technology, Blue Bell, PA, 1991.
3. Glass, J. E., J. Coat. Technol., 1978, 50(641), 72.
4. Fernando, R. H.; Glass, J. E.; J. Rheol., 1988, 32, 199.
5. Fernando, R. H.; et al., Prog. Org. Coat., 2000, 40, 35.
6. Adams, J., Spray Applications for Environmental Compliance, FSCT Symposium, Louisville,
KY, May 1990.
7. Lefebvre, A. H., Atomization and Sprays, Hemisphere Publishing Corporation, New York, 1989,
pp. 59–61.
REFERENCES
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
489
Easton, M. G., J. Oil Colour Chem. Assoc., 1983, 66, 366.
Nielsen, K. A.; et al., Polym. Mater. Sci. Eng., 1990, 63, 996.
Pucken, W., J. Prot. Coat. Linings, 2004, 21(9), 51.
Tachi, K.; et al., J. Coat. Technol., 1990, 62(791), 19.
Diana, M. J., Prod. Finish., 1992, July, 54.
Konieczynski, R. D., J. Coat. Technol., 1995, 67(847), 81.
Nielsen, K. A.; et al., Proc. Waterborne Higher-Solids Powder Coat. Symp., New Orleans, LA,
1995, p. 151.
Senser, D. W.; et al., Proc. Waterborne Higher-Solids Powder Coat. Symp., New Orleans, LA,
1995, p. 161.
Gebhard, M. S.; Scriven, L. E., J. Coat. Technol., 1994, 66(830), 27.
Fuchs, E. W.; et al., J. Coat. Technol., 1988, 60(767), 89.
Kia, S. F.; et al., J. Coat. Technol., 1991, 63(798), 55.
24
Film Defects
Obvious defects can result from incomplete coverage during application of a coating,
resulting in thin spots or holes, often called skips or holidays. Many other types of
defects and imperfections can develop in a film during or after application. In this
chapter we deal with the most important defects and to the extent possible, discuss the
causes of the defects and approaches for eliminating or minimizing their occurrence.
Unfortunately, the nomenclature for many defects is not uniform. Reference [1] provides
definitions for coating terms, including those for film defects.
24.1. SURFACE TENSION
Many defects are related to surface tension phenomena. Surface tension occurs because
the forces at an interface of a liquid differ from those within the liquid, due to the unsymmetrical force distributions on the surface molecules. The surface molecules possess
higher free energy, equivalent to the energy per unit area required to remove the
surface layer of molecules. The dimensions of surface tension are force exerted in the
surface perpendicular to a line; SI units are newtons per meter or millinewtons per
meter (mN m21). [Older units, still commonly used, are dynes per centimeter
(1 mN m21 ¼ 1 dyne cm21).] Similar surface orientation effects are present in solids,
which have surface free energies expressed in units of free energy per unit area, millijoules
per square meter (mJ m22) that are numerically and dimensionally equal to mN m21.
Frequently, people speak of the surface tension of the solids; although not formally
correct, errors do not result, because the values are identical. Reference [2] is an excellent
discussion of surface and interfacial properties, including data on many polymers.
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
490
24.1. SURFACE TENSION
491
Surface forces work to decrease the surface free energy of liquids and solids. Surface
tension works to draw liquids into spheres, since a sphere encloses a minimum ratio of
surface area to volume. In a spaceship, liquid droplets assume a spherical shape; on
Earth, the force of gravity distorts the spheres. For the same reason, surface tension
drives flow of an uneven liquid surface toward becoming a smooth surface. The smooth
surface has less interfacial area with air than the rough surface has; hence, there is a
reduction in surface free energy as a surface becomes smoother.
Segments of molecules that minimize surface tension tend to orient at the surface. The
lowest surface tension results from perfluoroalkyl groups at the surface. The next lowest is
methyl groups. Poly(dimethylsiloxane) has a low surface tension; the very flexible easily
rotatable backbone of siloxane bonds permits orientation of a large population of methyl
groups at the surface. The surface tension of linear aliphatic hydrocarbons increases as the
chain length increases, reflecting the larger ratio of methylene to methyl groups, Progressively higher surface tensions result from aliphatic chains, aromatic rings, esters and
ketones, and alcohols. Dipolar groups such as ketones, and hydrogen-bonding groups such
as alcohols, impart higher surface tension because the stabilizing intermolecular dipolar
and hydrogen-bonding interactions are absent at the surface, resulting in higher surface
energy. As the chain length of aliphatic esters, ketones, and alcohols increases, the surface
tension increases because methylene groups give higher surface tensions than do methyl
groups. Water has the highest surface tension of the volatile components used in coatings.
Addition of small amounts of surfactants to water reduces surface tension because the
hydrocarbon chains go to the surface. Addition of poly(dimethylsiloxane) to an organic
solvent solution reduces surface tension due to orientation of the methyl groups at the
surface. When a liquid is agitated, the molecules at the surface are mixed in with the
rest of the liquid. When the agitation is stopped, reorientation to give the lowest surface
tension occurs, but the equilibrium surface composition is not reestablished immediately.
When coating films are applied, they are subjected to considerable agitation. As Bierwagen
has pointed out, the surface tension of importance in governing some aspects of coating
behavior may not be the equilibrium surface tension, but rather, a dynamic surface
tension [3,4]. The time to regain equilibrium after agitation has ceased varies widely,
depending on the composition, viscosity, temperature, and probably other factors [5].
There has not been adequate quantitative study of the rates with different systems of
importance in the coatings field. Qualitatively, one can say that equilibrium is established
most rapidly when the coating contains small, flexible molecules and when there are large
differences in the polarity of components in the system as in waterborne coatings. Reaching equilibrium takes longer when the molecules with the lowest potential surface tension
groups are polymers. But if the polymers have moderate molecular weight and flexible
backbones, they can apparently reach the surface relatively rapidly. Poly(dimethylsiloxane) is an example. Low molecular weight octyl acrylate copolymers are used as additives
to reduce surface tension of coating films as they are forming. In waterborne coatings, it
has been shown that surfactants differ in their rates of reaching equilibrium surface tension
[5]. Surface tension increases with decreasing temperature, and solvents generally have
lower surface tensions than those of resins. Therefore, surface tension increases as
solvent evaporates from a film of resin solution, owing to the change in both concentration
and temperature.
If two liquids with different surface tensions are in contact with each other, the liquid of
low surface tension flows to cover the liquid with higher surface tension, since this results
in a lower overall surface free energy. Such a flow is a surface tension differential –driven
492
FILM DEFECTS
flow; some authors prefer the term surface tension gradient –driven flow. This type of flow
has been observed for millennia, but Carlo Marangoni, a nineteenth-century Italian
physicist, is credited with providing a scientific understanding of the phenomenon [6].
An example of the Marangoni effect is the flow that occurs when a clean glass containing
wine is tipped, wetting the side of the glass with the liquid, and then returned to an upright
position. Liquid then flows up the side of the glass, forming a bead of greater film thickness along the upper edge of the wetted area. When the amount of liquid collecting in the
bead becomes large, droplets flow back down the side of the glass. These “tears of wine”
have been known since biblical times, but why does the phenomenon occur? Ethyl alcohol
has a higher relative evaporation rate and lower surface tension than water, evaporation
occurs most rapidly along the edge of the layer of wine on the side of the glass. This
leads to reduction in alcohol concentration and hence higher surface tension along the
edge than in the bulk of the liquid. To minimize surface free energy, the low surface
tension (higher alcohol concentration) liquid in the bottom of the glass flows up to
cover the higher surface tension liquid at the edge. Evaporation continues at the edge,
reducing the alcohol concentration and leading to a continuation of flow of lower
surface tension liquid up the glass. The surface tension differential is also affected by
the temperature change; as alcohol evaporates, temperature decreases, further increasing
the differential in surface tension.
In summary, two types of flow result from surface tension effects. Surface tension
driven flows occur to minimize the surface area of a liquid. Surface tension differential
driven flows occur to cover a liquid or other surface of higher surface tension with a
liquid of lower surface tension.
24.2. LEVELING
Most methods of application of coatings initially lead to formation of a rough, wet film. It
is usually desirable for appearance and performance to have the irregularities level out.
The most widely studied leveling problem has been leveling of brush marks. A person
unacquainted with the field might first think that leveling results from gravitational
effects. If gravity were a significant factor, paints applied to ceilings should level more
poorly than paints applied to floors, but they do not. Based on studies of flow of
mineral oil, Orchard proposed that the driving force for leveling is surface tension and
established mathematical models for the variables that he proposed would control the
rate of leveling [7]. Orchard’s treatment has been widely applied to flow of coatings.
Patton illustrated the model for an idealized cross section of a wet film exhibiting brush
marks that follow a sine wave profile, as shown in Figure 24.1. He gives several forms
Figure 24.1. Cross section of brush marks. (From Ref. [8], with permission.)
24.2. LEVELING
493
of the Orchard equation and shows their derivations. A convenient form relating the
change in amplitude of the sine wave to time is
ln
a0 5:3g x3 dt
¼
at
l4 h
(24:1)
where a0 is the initial amplitude (cm), at is the amplitude at time t (cm), x is the average
coating thickness (cm), l is the wavelength (cm), g is the surface tension (mN m21), h is
the viscosity (Pa.s), and t is the time (s).
Leveling is fastest when the wavelength is small, the viscosity is low, the surface
tension is high, and the film is thick. The formulator has little or no control over most
of the variables. Wavelength is determined by application conditions; in brushing, it
increases as pressure on the brush increases and as the thickness of the coating increases.
High surface tension increases the rate of leveling; however, the formulator is limited in
optimizing this factor, since high surface tension can lead to other defects. Thicker films
promote leveling, however; increasing the film thickness increases the cost of coating and
increases the probability of sagging (Section 24.3) on vertical surfaces. The principal
means of control left is viscosity.
The Orchard model provides a satisfactory correlation between experimental data and
predictions when the liquid film has Newtonian flow properties and sufficiently low volatility such that viscosity does not change during the experimental observations. With coatings, however, viscosity is usually changing during the time in which leveling occurs. As
solvent evaporates, viscosity increases. Furthermore, if the system is thixotropic, the viscosity is reduced by the high shear rate exerted during application and subsequently
increases with time at the low shear rate involved in leveling. Another potential shortcoming in the Orchard treatment is the assumption of constant surface tension.
Overdiep devised methods of observation that permitted following the location of the
ridges and valleys [9,10]. He found for two alkyd coatings that brush marks leveled to an
essentially smooth film, but then ridges grew where there had been valleys, and valleys
formed where there had been ridges. Although surface tension can and does cause a
ridged film to level, it cannot cause a level film to become ridged because that creates
more surface area. Overdiep proposed that the surface tension differential is the major
driving force for emergence of the new ridges. As shown in Figure 24.1, the wet film thickness in the valleys of the brush marks is less than in the ridges. When the same amount of
solvent evaporates per unit area of surface, the fraction of solvent that evaporates in the
valleys is larger than that in the ridges. As a result, concentration of resin solution in
the valleys becomes higher than that in the ridges and the surface tension in the valleys
is higher than that on the ridges. Following the Marangoni effect, coating flows from
the ridges into the valleys. In other words, Overdiep proposed, and was able to demonstrate experimentally, that with volatile solvents, the primary driving force for leveling
is not surface tension but surface tension differential. In some cases, these differentials
lead to overshooting of the smoothest stage and cause the growth of ridges. The extent
of the flow driven by surface tension differential depends on the rate of evaporation of
the solvent.
Kojima and Moriga studied solvent evaporation and leveling of water-reducible coatings and showed that the forces driving leveling depend on the solvent in the formulation
[11]. With a fast evaporating solvent, such as isobutyl alcohol, surface tension increased
494
FILM DEFECTS
during drying, promoting leveling by surface tension differential forces, but with a slow
evaporating solvent such as ethylene glycol monohexyl ether, surface tension decreased
causing an adverse effect on leveling. Equations have been developed that model the
drying process through the changes in surface tension differentials and changes in viscosity during solvent evaporation [12].
Overdiep was particularly interested in what happens with uneven coating films over a
rough substrate [9]. He reasoned that surface tension – driven flow might give the smoothest film. However, as illustrated in Figure 24.2(a), this may be undesirable because protection in thin areas might be limited. On the other hand, as shown in Figure 24.2(b), surface
tension differential –driven flow would tend to yield equal film thickness, with the surface
of the film following the roughness of the substrate rather than being level. Overdiep
suggests that it might be best to adjust a coating so that both types of flow are significant
to achieve a compromise with reasonable film smoothness without places where film
thickness is very thin, as shown diagrammatically in Figure 24.2(c). The balance could
be controlled by the volatility of the solvent; with very low volatility, leveling would
be surface tension driven; with relatively high volatility, the major effect would be
surface tension differential; with intermediate volatility, both phenomena could be
important.
Neither the Orchard equation nor Overdiep’s work takes into account the effect of
thixotropy. Cohu and Magnin developed equations that predict the effect of thixotropy
on leveling of nonevaporative coatings [13]. Another result of poor leveling is reduction
in gloss. A study of the effect of leveling on gloss showed that three factors are important:
extent of roughness of the substrate, film thickness, and viscosity [14]. The addition of
fluorocarbon surfactant did not significantly change leveling in these experiments; presumably, surface tension differentials did not develop in the drying of the films.
In spray application, surface roughness consists of bumps surrounded by valleys rather
than ridges and valleys. Since the effect is somewhat reminiscent of the appearance of
orange skins, it is called orange peel. Figure 24.3 shows an orange peel surface. The
bumps are generally much larger than spray droplets. Most commonly, orange peel is
Figure 24.2. Alternate leveling (a –c) results after applying a coating to a rough surface. (Adapted from
Ref. [ 9], with permission.)
24.2. LEVELING
495
Figure 24.3. Typical orange peel pattern ( 15). (From Ref. [16], with permission.)
encountered when spraying coatings that have solvents with high evaporation rates. It is
common for people to conclude that due to the fast evaporation, the viscosity of the
coating on the substrate builds up so rapidly that the leveling is poor; in some situations,
that is probably the case. However, in the late 1940s it was found that leveling of sprayed
lacquer films could frequently be improved by addition of very small amounts of silicone
fluid. Contrary to the common wisdom that all leveling is surface tension driven, that is,
promoted by high surface tension, here was a case in which adding a material known to
reduce surface tension improved leveling substantially.
Hahn provided an explanation for the phenomenon [15]. When one sprays a lacquer,
initially the surface is fairly smooth, and if you watch, orange peel grows (Figure 24.3).
Hahn proposed that the growth of orange peel results from a surface tension differential – driven flow. The last atomized spray particles to arrive on the wet lacquer surface
have traveled a longer distance between the spray gun and the surface, hence have lost
more solvent, a higher resin concentration and therefore a higher surface tension than the
main bulk of the wet film. The lower surface tension wet lacquer flows up the sides of
these last particles to minimize overall surface free energy; that is, surface tension differential – driven flow grows the orange peel. If one adds a surface tension depressant, such as a
silicone fluid that orients very rapidly to the surface; the surface tension of the wet lacquer
surface and the surface tension of the last atomized particles to arrive are uniformly low, so
there is no differential to promote growth of orange peel. Octyl acrylate copolymer additives can also give an overall low surface tension and minimize orange peel growth.
It would be of interest to see what effect silicone fluid or octyl acrylate copolymers would
have on leveling of brush applied coatings. To the extent that the leveling is surface tension
differential driven, as found by Overdiep, the leveling should be made poorer, not improved.
The results of such an experiment are not given in the literature, but such additives are not
used to promote leveling in brush applied coatings, only in spray applied coatings.
Electrostatically sprayed coatings are likely to show surface roughness greater than
nonelectrostatically sprayed coatings. It has been speculated that the greater surface
roughness obtained with electrostatic spray results from arrival of the last charged
particles on a coated surface that is electrically insulated quite well from ground. These
later arrivals may retain their charges sufficiently long to repel each other and thereby
496
FILM DEFECTS
reduce the opportunity for leveling. It has also been suggested that when coatings are
applied by high-speed bell electrostatic spray guns, differentials in the pigment concentration within the spray droplets may result from the centrifugal forces [17]. These
pigment concentration differentials can lead to rougher surfaces, reduction in gloss, and
aluminum flake striping due to differential flake concentrations.
Leveling problems are particularly severe with latex paints. Latex paints, in general,
exhibit a greater degree of shear thinning and more rapid recovery of viscosity after
exposure to high shear rates than paints made with solutions of resins in organic solvents.
Due to their higher dispersed phase content, the viscosity of latex paints changes more
rapidly with loss of volatile materials than the viscosity of solventborne paints. No experimental work has been reported on the relative importance of surface tension and surface
tension differentials in leveling of latex paints; however, it seems probable that the leveling is primarily surface tension driven. The surface tension of water is high, but the presence of surfactants imparts low surface tension to latex paints. Furthermore, it is probable
that this low surface tension is established rapidly after the agitation of application stops.
Perhaps more important, the surface tension is uniformly low, since it is almost unchanged
as water evaporates. Thus, the generally poor leveling of latex paints may result in part
from the absence of surface tension differentials to promote leveling. The low surface
tension may not provide adequate driving force for leveling in a film whose viscosity
increases rapidly with time. The problems of leveling of latex paints are discussed
further in Section 32.3.2. The leveling of powder coatings is discussed in Section 28.3.
24.3. SAGGING AND DRIP MARKS
When a wet coating is applied to a vertical surface, the force of gravity causes it to flow
downward (sagging). Differences in film thickness at various places lead to differing
degrees of sagging, sometimes resulting in curtains, or drapes, of coating. The variables
that affect the volume of coating that has sagged (Vs) after time (t) are given by Patton:
Vs ¼
x3 rgt
300h
(24:2)
where x is initial film thickness (cm), r is density (g cm21), g is gravitational constant
(s cm22), t is time (s), and h is viscosity (Pa.s) [18].
Gravity is the driving force for sagging, which increases with coating density and thickness. Some latitude is available to the formulator in minimizing density and thickness, but
the major variable available for controlling sagging is viscosity. The tendency to sag can
be evaluated by observing the behavior of films applied under conditions simulating field
use, and various tests have been developed. The most common test is a sag-index blade, a
1
straight edge applicator blade with a series of 14-in. gaps of different depths at 16
-in. intervals across the blade [19,20]. A drawdown, which is a series of stripes of coating of
various thickness, is made on a chart, and immediately, the chart paper is placed in a vertical position. Sag resistance is rated by observing the thickest stripe that does not sag
down to the next stripe. For research purposes, Overdiep developed the sag balance, a
more sophisticated method, which provides a numerical basis for evaluating the extent
of sagging [10]. Overdiep also developed equations for sagging that take into consideration the changes in viscosity after application.
24.3. SAGGING AND DRIP MARKS
497
In sprayapplied solvent solution coatings, sagging can generally be minimized while
achieving adequate leveling by a combination of proper use of the spray gun (to avoid
excessively thick areas) and control of the rate of evaporation of solvents. The goal is
to manipulate viscosity so that it is initially low for leveling but builds up before noticeable
sagging occurs. In brush and hand rollerapplied coatings where slow evaporating solvents
are used, thixotropic systems permit leveling to occur before the viscosity recovers. Latex
paints, which are almost always thixotropic, are less likely to exhibit sagging than are
solvent solution paints.
Sagging can be a serious problem with high-solids coatings, especially when spray
applied. Although other factors may be involved, a major cause of sagging of high-solids
coatings is that substantially less solvent is lost during spraying (i.e., after the droplets
leave the gun and before they arrive at the substrate surface) [20,21]. This low loss of
solvent leads to less increase in viscosity of a high-solids coating than of a conventional
coating, resulting in a greater likelihood of sagging. The reasons for lower solvent loss
have not been clearly established, but are probably related to the following considerations.
First, perhaps because of their higher surface tension, high-solids coatings atomize to give
larger particle size droplets than those of conventional coatings. The lower ratio of surface
area to volume would lead to lower solvent loss. However, one should be able to adjust
spray equipment and conditions to obtain equivalent atomization. A second possibility
is a colligative effect on solvent evaporation. Resins in high-solids coatings are lower
in molecular weight, and their concentrations are higher. For both these reasons, the
ratio of the number of solvent molecules to resin molecules is lower in the case of highsolids coatings than for conventional coatings. This undoubtedly leads to a decrease in rate
of solvent loss (see Section 18.3.5 for a model calculation). However, it seems doubtful
that this difference could account entirely for the large differences in solvent loss such
as those reported by Wu [20]. A third theory is that the stage of diffusion control of the
rate of solvent loss is reached after less loss of solvent from high-solids than from conventional coatings; thus, the solvent evaporation from spray droplets of high-solids coatings is
markedly reduced [22,23]. At later stages of solvent evaporation from a film, the rate of
diffusion of solvent molecules to the surface becomes the factor limiting the rate of evaporation of solvent. Solvents with linear rather than branched backbones are desirable since
they can diffuse faster (Section 18.3.4).
Hot spraying can help control sagging (Section 23.2.4). When the coating cools on
striking the object, the viscosity increase reduces sagging. The use of carbon dioxide
under supercritical conditions is particularly helpful in controlling sagging, since the
CO2 flashes off almost instantaneously when the coating leaves the orifice of the spray
gun, thereby increasing viscosity (Section 23.2.5). High-speed electrostatic bell application permits application of coatings at higher viscosity, which also helps control
sagging (Section 23.2.3).
When sagging of high-solids coatings cannot be adequately controlled by adjustment of
solvent composition of the coating and application variables, which is not uncommon, a
useful approach is to employ additives to make the systems thixotropic. For example, dispersions of fine particle size silicon dioxide, precipitated silicon dioxide, bentonite clay
treated with a quaternary ammonium compound, or polyamide gels can be added to
impart thixotropy. One tries to formulate so that recovery to high viscosity is slow
enough to permit reasonable leveling and rapid enough to control sagging. However,
such agents increase the high shear viscosity somewhat and require higher solvent levels.
They also tend to lower gloss and their effectiveness decreases as temperature rises.
498
FILM DEFECTS
The problem of sagging in high-solids automotive metallic coatings can be particularly
severe (Section 30.1.2). Even a small degree of sagging, which might not be noticeable in
a white coating, is very evident in a metallic coating, since it affects the orientation of the
metal flakes. Use of SiO2 to impart thixotropy is undesirable, since even the low scattering
efficiency of SiO2 is enough to reduce color flop in the coatings. Acrylic microgels have
been developed that impart thixotropic flow by flocculation of the swollen gel particles
[24]. In the final film, the index of refraction of the polymer from the microgel is
nearly identical with that of the cross-linked acrylic binder polymer, so that light scattering
does not interfere with color flop. The effect of the gel particles depends on interaction
with the low molecular weight acrylic resin; Ref. [25] discusses the rheological properties
of the systems. There can also be an improvement in the strength of the final film when
microgels are incorporated [26].
Another problem that can be encountered with high-solids coatings is oven sagging,
also called hot sag [22]. The coating appears to be fine until it is put into the oven; then
sagging occurs. Oven sagging results from the strong temperature dependence of the viscosity of high-solids coatings. Compared to conventional coatings, there is a much steeper
drop in viscosity as the coated product enters the hot oven, which promotes sagging; and
the coating continues to flow even after the solvent is gone. Initial stages of cure are chain
extension of the low MW resin. The viscosity does not jump until substantial cross-linking
has occurred. Oven sagging can be controlled somewhat by zoning the oven. A lower
temperature in the initial zone allows more time for solvent loss, and perhaps some crosslinking, so viscosity increases as a result of the higher solids or higher molecular weight
before the film is subjected to high temperature.
Water-reducible coatings are less likely than high-solids coatings to give sagging problems, but there are circumstances in which they exhibit delayed sagging. The viscosity of
these coatings is very dependent on the ratio of water to solvent as well as on the solids
content (Section 8.3). As water and solvent evaporate, the residual water/solvent ratio
can sometimes decrease, leading to lower viscosity despite the higher solids content;
and sagging can result. Such behavior can depend on the relative humidity during the
flash off period after spraying. It has been found for a water-reducible acrylic enamel
that sagging occurred above, but not below, a critical relative humidity [27]. (See Sections
18.3.3 and 18.3.6 for the definition and discussion of critical relative humidity.)
A problem related to sagging is drip marks. For example, when one applies a coating to
a vertical surface in which nail heads have been counter sunk, the coating may sag, but in
addition, coating will flow out of the nail head depression where the film thickness is
greater, so that an additional flow occurs. Similar effects can occur at other places on
the paint surface, such as dripping from sharp corners, bolt holes, or cut outs, where
film thickness tends to be greater.
An investigation of sagging and drip marks in latex paints is described in Ref. [28]. A computer model was developed that was correlated with the actual performance of the paints. The
model takes into account not only the effect of gravity, thixotropy, and film thickness but also
the change in flow properties of the thixotropic system with time. It is pointed out that during
the flow of films, viscosity varies with shear stress rather than shear rate, as commonly
thought. The model also accounts for nonuniform film thickness. As an example, the
model was designed to predict the effect of a counter sunk nail head, which resulted in a
locally thicker film. The thicker film is exposed to higher gravitational force for a longer
time than the rest of the surface, resulting in greater flow out of the depression, resulting
in a dip in the surface over the nail hole and formation of a drip mark below the depression.
24.4. CRAWLING, CRATERING, AND RELATED DEFECTS
499
24.4. CRAWLING, CRATERING, AND RELATED DEFECTS
If a coating that has a relatively high surface tension is applied to a substrate having a comparatively low surface free energy, the coating will not wet the substrate. The mechanical
forces involved during application may spread the coating on the substrate surface, but
since the surface is not wetted, surface tension forces tend to draw the liquid coating
toward a spherical shape. Meanwhile, solvent is evaporating, and therefore viscosity is
increasing, so that before the coating can pull up into spheres, the viscosity is high
enough that flow essentially stops. The result is an uneven film thickness with areas
having little, if any, coating adjoining areas of excessive film thickness. This behavior
is commonly called crawling or retraction (Figure 24.4). For waterborne coatings, crawling can depend on the rate of establishment of equilibrium surface tension with different
surfactants [5].
Crawling can result from applying a coating to steel with oil contamination on the
surface. It is especially common in coating plastics. In some cases, crawling results
from failure to completely remove a mold release agent from a plastic molded part. Application of a high surface tension top coat to a low surface tension primer can also lead to
crawling. If a coating contains silicone fluids or fluorocarbon surfactants, there is likely to
be crawling when a subsequent coat is applied. If one handles a primer surface with bare
hands and then applies a relatively high surface tension top coat, it is likely that the top
coat will draw away from the oils left behind in fingerprints. This type of crawling,
which copies a pattern of low surface tension areas on the substrate, has been called telegraphing. Care is needed with this term, as this circumstance is only one of several
phenomena called telegraphing.
Crawling can also result from the presence in the coating of surfactant-type molecules
that can orient rapidly on a polar substrate surface. The polar group of the surfactant
associates with the substrate, and the long nonpolar end becomes the surface. If one
adds excess silicone fluid to a coating to correct a problem such as orange peel, small
droplets of insoluble fractions of the poly(dimethylsiloxane) can migrate to the substrate
Figure 24.4. Crawling of a top coat applied over a low surface energy primer ( 7). (From Ref. [ 16], with
permission.)
500
FILM DEFECTS
surface and spread on it, leaving a new substrate surface that the coating cannot wet,
resulting in crawling and/or cratering. A little silicone fluid can solve some defect problems, but even a small excess can cause what might be a worse problem. Higher molecular weight fractions of poly(dimethylsiloxane) are reported to be insoluble in many coating
formulations [29]. Modified silicone fluids, such as polysiloxane –polyether block copolymers, have been developed that are compatible with a wider variety of coatings and are
less likely to cause undesirable side effects. The effect of a series of additives on crawling
and other film defects has been reported [30].
High-solids coatings often have higher surface tensions than those of conventional
coatings. To achieve high-solids, lower molecular weight resins with lower equivalent
weights must be used, which requires that the concentration of polar functional groups
such as hydroxyl groups is higher, and hence surface tension is generally higher. Also, solvents that give the lowest viscosity coatings are likely to have relatively high surface
tension. Accordingly, there is a greater likelihood of crawling problems with high-solids
coatings.
Cratering is the appearance of small round depressions in a coating’s surface, generally
with a slightly raised crest. They often look somewhat like volcanic craters—hence the
name. They are also called fish eyes. Schoff calls cratering “the nastiest, most frustrating
defect of all” [31]. A schematic drawing and a photograph are shown in Figures 24.5 and
24.6, respectively. Cratering is sometimes mistaken for popping and vice versa; ways to
differentiate them are described in Section 24.7.
Cratering results from a small particle or droplet of low surface tension contaminant,
which is on the substrate, in the coating, or which lands on the wet surface of a freshly
applied film [15]. The particles or droplets can be very small. Some of the low surface
tension material dissolves in the adjacent film, creating a localized surface tension differential. As a result of the Marangoni effect, this low surface tension part of the film flows
away from the particle to try to cover the surrounding higher surface tension liquid
coating. Since, as the flow occurs, solvent evaporates, the differential in surface tension
increases, and flow continues. However, loss of solvent causes an increase in viscosity,
which impedes flow, leading to formation of a characteristic crest around the pit of the
crater.
Craters must be avoided in applying coatings. The user applying the coating should try
to minimize the probability of low surface tension contaminants arriving on the wet
coating surface. For example, the spraying of lubricating oils or silicone fluids on or
near the conveyor carrying freshly coated parts is an almost sure way to cause craters.
Paint booth air must be filtered multiple times. However, in most factories, the presence
Figure 24.5. Crater. (From Ref. [16], with permission.)
24.4. CRAWLING, CRATERING, AND RELATED DEFECTS
501
Figure 24.6. Typical crater. (From Ref. [16], with permission.)
of some contaminating particles cannot be avoided; therefore, the formulator must design
coatings that minimize the probability of cratering. Lower surface tension coatings are less
likely to form craters, since fewer contaminating particles will have even lower surface
tensions. Alkyd coatings have low surface tensions and seldom give cratering problems
(or crawling problems either). In general, polyester coatings are more likely to give cratering problems than acrylic coatings, which tend to have lower surface tensions. High-solids
coatings, because of their generally higher surface tensions and low viscosities even after
solvent loss, are more likely than conventional coatings to give cratering problems. Some
waterborne coatings are also vulnerable to cratering. In a number of cases, particles from
cosmetics worn by production workers have caused cratering of automotive coatings.
Powder coatings are quite vulnerable to cratering, as discussed in Chapter 27.
It is possible in some cases to determine the nature and source of the contaminating
particle, causing craters. Schoff and Hare advocate use of low power (2– 60 or
10280) optical microscopes for preliminary investigation of craters [31,32]. If necessary, more powerful optical microscopes and scanning electron microscopy (SEM) can be
applied to the problem. SEMs with an x-ray attachment (EDX) are especially valuable.
Two other powerful tools for such analyses are laser microprobe mass analysis
(LAMMA) and time of flight secondary ion mass spectroscopy (ToF-SIMS). LAMMA
is useful primarily in identifying inorganic materials, and ToF-SIMS is useful primarily
in identifying organic contaminants; even between different silicone oils. In the case of
a crater through a clear coat down to a base coat, ToF-SIMS showed that the contaminant
was a poly(dimethylsiloxane). The source of the siloxane was traced to a resin supplied by
a new supplier. In another case, craters in a primer coat on a car body were shown by
LAMMA to be from a skin cream used by an employee in the paint shop. In another
case, a contaminating particle in an E-coat primer was shown by LAMMA to be a
welding pearl from a copper coated steel welding wire [33]. Even with these powerful
tools, the cause cannot always be found, perhaps because the cratering agent has evaporated or dissolved in the film.
Additives can be used to minimize cratering. Small amounts of silicone fluid eliminate
cratering, but as noted before, caution is required in selecting the amount and type of
silicone fluid used to avoid crawling or recoat adhesion problems. Poly(2-ethyl hexyl
acrylate) additives such as Modaflow usually reduce cratering. These additives operate
by giving a uniform surface tension. Any low surface tension contaminant would
spread itself widely across the surface. If the entire surface has a uniform surface
502
FILM DEFECTS
tension, there will be no surface tension differential – driven flow and no cratering. A comparison of the effects of a range of additives on the control of defects such as cratering has
been reported [34,35]. Side effects such as reduction of gloss and loss of adhesion of coatings applied over the coatings are discussed. Siloxanes with a polyether terminal group or
side chain are reported to prevent film defects in waterborne coatings and to have little
effect on recoatability. They are prepared by reacting allyl-terminated polyethers with
poly(dimethylsiloxanes) with Si22H bonds [36].
There are many other examples of film defects that result from surface tension differential –driven flows. In coating tin plate sheets, the coating is applied by roller and the
coated sheets are passed on to warm wickets that carry the sheets approximately vertically
through an oven. In some cases, one can see a pattern of the wicket as a thin area on the
final coated sheet. Heat transfer to the sheet is fastest where it is leaning against the metal
wicket. The surface tension of the liquid coating on the opposite side drops locally because
of the higher temperature. This lower surface tension material flows toward the higher
surface tension surrounding coating, leaving an area of thin coating. This defect has
also been called telegraphing. Similarly, on large plastic parts with thicker reinforced
areas on the back of the part, telegraphing through to the top coat(s) can be encountered.
With the adoption of high-solids clear coats to paint plastic auto body components, there
has been an increasing problem with bond line readout. Typically, this occurs over the areas
of SMC plastic where, on the other side, there are adhesive joints to the car structure. A
series of model system studies showed that the bond line defects resulted from differential
surface tension – driven flow. The areas over the bond line heat more slowly than the adjacent areas, setting up differentials in surface tension and flow away from this bond line [37].
In spraying flat sheets, one can get an effect called picture framing, or fat edge. The
coating is thickest at the edges and thinner than average just in from the edge. The contrast
in hiding of the substrate can make the differences in film thickness very evident. Solvent
evaporates most rapidly from the coating near the edge, where the air flow is greatest. This
leads to a lower temperature and to an increase in resin concentration at the edge. Both
factors increase the surface tension there, causing the lower surface tension coating adjacent to the edge to flow out to the edge to cover the higher surface tension coating.
Surface tension differential –driven flow can also result when overspray from spraying
a coating lands on the wet surface of a different coating. If the overspray has lower surface
tension than the wet surface, cratering occurs. If the overspray has high surface tension
compared to the wet film, local orange peeling results.
In applying coatings by curtain coating (Section 23.5), the curtain of coating must
remain intact. If a particle or droplet of contaminant of lower surface tension than the
coating lands on the surface of the flowing curtain, surface tension differential – driven
flow will cause a thin area in the curtain, which can cause a hole in the curtain. When
this part of the curtain is deposited on the panel being coated, an uncoated area results.
The problem is minimized by using coatings of the lowest possible surface tension.
Since the curtain is flowing, dynamic surface tension is the important quantity. Bierwagen
has discussed this phenomenon [4].
24.5. FLOATING AND FLOODING; HAMMER FINISHES
Two related defects result from uneven distribution of pigment in a film as it is drying:
floating and flooding. Some people call both phenomena floating. We follow the more
24.5. FLOATING AND FLOODING; HAMMER FINISHES
503
common terminology by which floating describes a mottled effect and flooding is used
when the surface color is uniform but is darker or lighter in color than should result
from the pigment combination used. Floating results from horizontal separation of different pigments within the film, whereas flooding results from vertical separation.
Floating is most evident in coatings pigmented with at least two pigments. For example,
a light blue gloss enamel panel can show a mottled pattern of darker blue lines on a lighter
blue background. The pattern tends to be hexagonal, but seldom perfectly so. Alternatively,
with a different light blue coating, the color pattern might be reversed: The lines could be
light blue with the background a darker blue. These effects result from pigment segregations
that occur as a result of convection current flows driven by surface tension differentials
while a film is drying. Rapid loss of solvent from a film during drying leads to considerable
turbulence. Convection patterns are established whereby coating material flows up from
lower layers of the film and circulates back down into the film. As the fresh material
flows across the surface before it turns back down, solvent evaporates, concentration
increases, temperature drops, and surface tension increases. The resulting surface tension
differential sustains the convection current. The flow patterns are roughly circular, but as
they expand, they encounter other flow patterns and the convection currents are compressed.
If the system is quite regular, a pattern of hexagonal Bénard cells is established. The cells
are named after a seventeenth-century French scientist who pointed out the commonness of
hexagonal flow patterns in nature. As solvent evaporation continues, viscosity increases and
it becomes more difficult for the pigment particles to move. The smallest particle size,
lowest density particles continue moving longest and the largest particle size, highest
density particles stop moving sooner. The segregated pattern of floating results.
Floating is particularly likely to occur if one pigment is flocculated and the other is a
nonflocculated dispersion of fine particle size. The fine particle size pigment keeps
moving longest and is trapped where the convection current turns back into the film at
the border between adjacent cells. The border between the cells has a higher concentration
of the finer particle size material, whereas the center of the cells is more concentrated in
the coarser pigment. If in the example of the light blue coating, the white pigment is flocculated and the blue is not, one will find darker blue lines on a lighter blue background. If
on the other hand, the blue is flocculated and not the white, there will be lighter blue lines
on a darker blue background. Figure 24.7 shows convection patterns in Bénard cell
Figure 24.7. Bénard cell formation. (From Ref. [ 16], with permission.)
504
FILM DEFECTS
formation. Floating can also occur in single pigment coatings, resulting in uneven pigmentation of the surface, which reduces gloss.
Floating can be reduced by properly stabilizing pigment dispersions, so that neither
pigment is flocculated. However, one can get floating even without flocculation. This
can result from the use of pigments with very different particle sizes and densities. An
example is use of fine particle size high-color carbon black with titanium dioxide to
make a gray coating. Not only is the particle size of the TiO2 several times that of the
carbon black, but the TiO2 also has about fourfold higher density. A much larger particle
size, weaker black, such as lamp black, can be used to make a gray with a lower probability
of floating.
Floating is reduced if slower evaporating solvents are used. Surface tension differentials are then less likely to be established, and Marangoni flow and floating do not
occur. If a coating exhibits floating, the formulator should first reformulate to eliminate
the problem; most commonly, selective pigment flocculation is the cause and should be
corrected. Sometimes the flocculation is caused by the solvents used to reduce the
coating to application viscosity, in which case solvent change eliminates the problem.
In other cases, a change in the resin or dispersing agent used to stabilize the pigment dispersion is required (Section 21.1.3). The second choice is to use an additive. As with other
flow phenomena driven by surface tension differentials, floating can be prevented by
adding a minimum amount of a silicone fluid. To avoid crawling, the minimum amount
of silicone must be used, which can be facilitated by using a very dilute solution of the
silicone fluid in solvent. Since slow evaporating solvents also help reduce floating, slow
evaporating solvents are commonly used to make additive solutions.
The term flooding is applied when the color of the surface is uniform but is different
than should have been obtained from the pigment combination involved. For example,
one might have a uniform gray coating, but a darker gray than that expected from the
ratio of black to white pigments. The most troubling part of flooding is that the extent
of flooding can vary with the conditions encountered during application, leading to different colors on articles coated with the same coating. Flooding results from surface enrichment by one or more of the pigments in the coating [16,38]. The stratification is thought to
occur as a result of different rates of pigment settling within the film, which are caused by
differences in pigment density and size or flocculation of one of the pigments. Flooding is
accentuated by thick films, low vehicle viscosity, and low evaporation rate solvents—any
factor that tends to keep the film at low viscosity longer and to allow more pigment
settling. Flooding is easily detected by rubbing a section of a wet film: If the rubbed
area has a higher color strength, the coating is flooding. The remedies are to avoid flocculation and low density fine particle size pigments, and if possible, to use faster evaporating
solvents and higher viscosity vehicles. Floating, flooding, and related color defects are
discussed in Ref. [38].
Floating and flooding can also occur in latex paints. The factors leading to floating and
flooding in latex paints have been reported [39]. Latex paints include one or more surfactants in the latex, inorganic pigment dispersions, and organic pigment dispersions. In some
cases, the combinations are unstable and lead to flocculation. In other cases, there can
be incompatibility that affects only one of the pigments, which leads to floating or
flooding. When a latex paint is diluted with water, floating can result. This is attributed
to a shift in the equilibrium between the dispersed particles and water, resulting in less
stabilizing dispersant on pigment surfaces. Also, the viscosity decreases, facilitating
movement of particles; and increased surface tension may promote formation of Bénard
24.6. WRINKLING AND WRINKLE FINISHES
505
cells. Decreasing the particle size of the inert pigments (kaolin and talc) reduces floating.
Using a mixture of higher and lower molecular weight acrylic carboxylate dispersants
gives the best compromise of dispersing efficiency and stabilization against floating. Compatibility between the thickener and the dispersants in the paint can be critical. TiO2
content was found to affect floating in paints tinted with carbon black and monoazo
red; the higher the TiO2 content, the more pronounced the floating.
Although floating is usually undesirable, ingenious coatings formulators have taken
advantage of the problem by purposely inducing floating to make attractive coatings.
The coatings are called hammer finishes because they look a bit like the pattern one
would get by striking a piece of metal with a ballpeen hammer. Hammer finishes were
once used on a large scale, especially for coating cast iron components, for which it
was desirable to hide the surface roughness. Such coatings contain large particle size nonleafing aluminum pigment and dispersions of transparent fine particle size pigments, commonly phthalocyanine blue. One way of getting a hammer effect is to spray a metallic blue
coating and then spray a small amount of solvent on the wet film. Surface tension is lowest
where drops of solvent land, and surface tension differential –driven convection flow patterns are set up, leading to floating where the lines have more blue and the centers of the
patterns have more aluminum with less blue. There are self-hammer coatings, formulated
to give a hammer finish pattern without need for a spatter spray of solvent. Fast evaporating solvents are used with a resin, such as a styrenated alkyd resin, which gives fast drying.
Use of hammer finishes has decreased as smooth plastic molded parts replaced rough
metal castings in many end uses.
24.6. WRINKLING AND WRINKLE FINISHES
The term wrinkling refers to the surface of a coating that looks shriveled or wrinkled into
many small hills and valleys. In some cases, a wrinkle pattern is so fine that to the unaided
eye, the film appears to have low gloss rather than to look wrinkled. However, under magnification, the surface can be seen to be glossy but wrinkled. In other cases the wrinkle
patterns are broad or bold and are readily visible to the naked eye. Wrinkling results
when the surface of a film becomes high in viscosity while the bottom of the film is
still relatively fluid. It can result from rapid solvent loss from the surface, followed by
later solvent loss from the lower layers. It can also result from more rapid cross-linking
at the surface of the film than in the lower layers of the film. Subsequent solvent loss or
cure in the lower layers results in shrinkage, which pulls the surface layer into a wrinkled
pattern. Wrinkling is more apt to occur with thick films than with thin films because the
possibility of different reaction rates and differential solvent loss within the film increases
with thickness.
The earliest examples of wrinkling were with drying oil films, especially if all or part of
the oil was tung oil, and cobalt salts were used as the only drier. Tung oil cross-links relatively rapidly when exposed to oxygen from air, and cobalt salts are active catalysts for the
autoxidation reaction but are poor through driers (Section 14.2.2). These factors favor
differential surface cure, which results in wrinkling of the surface layer. The wrinkle
pattern can be fine or bold, depending on the ratio of tung oil to other drying oils, or in
the case of alkyd systems, the oil length of the alkyd, and the ratio of cobalt to driers
such as lead (or zirconium) salts that promote through dry. Although in many cases, wrinkling is undesirable, ingenious coating formulators turned the disadvantage into an
506
FILM DEFECTS
advantage. For many years, wrinkle finishes were sold on a large scale for applications
such as office equipment. Like hammer finishes, wrinkle finishes covered uneven cast
metal parts. Their use has dropped since plastic molded parts have supplanted many
metal castings.
Today, wrinkling is usually an undesired defect. It is most often encountered in improperly formulated or applied MF cross-linked coatings in which amines are used to
neutralize acidity in the coating formulation and/or amine blocked sulfonic acid catalysts
are used for package stability. The probability of wrinkling in such coatings increases as
the volatility of the amine increases. For example, triethylamine leads to wrinkling under
conditions in which dimethylaminoethanol does not. Increasing catalyst concentration
tends to increase wrinkling, as does greater film thickness [40].
Another situation in which wrinkling can occur is in UV curing of pigmented acrylate
coatings by free radical polymerization (Section 29.4). High concentrations of photoinitiator are required to compete with absorption by the pigment. Penetration of UV through the
film is reduced by absorption by the pigment as well as by the photoinitiator. There is rapid
cross-linking at the surface and slower cross-linking in the lower layers of the film, resulting in wrinkling. Wrinkling is likely to be more severe if the curing is done in an inert
atmosphere rather than in air. In the latter case, the cure differential is reduced by
oxygen inhibition of surface cure. UV curing by cationic polymerization, which is not
air inhibited, is even more prone to surface wrinkling.
24.7. BUBBLING AND POPPING
Bubbles can occur near the surface of a film; popping is the formation of broken bubbles
at the surface of a film that do not flow out. Both effects can occur when the surface viscosity increases to a high level while volatile material remains in the lower parts of the
film. If the viscosity is very high at the surface, some of the bubbles of solvent rise to the
surface but do not break. Popping occurs when the viscosity at the surface of a film
increases sufficiently that solvent bubbles break but the coating does not flow out
before further viscosity increase prevents leveling. If the pops are very small, the
effect is sometimes called pinholing. There often is a mixture of open pops, bubbles,
and pinholes in a surface.
Popping and bubbling result from rapid loss of solvent at the surface of a film during
initial flash off. The surface develops a high viscosity relative to that of the solvent rich
lower layers of the film. When the coated object is put into an oven, solvent volatilizes
in the lower layers of the film, creating bubbles that do not readily pass through the highviscosity surface. As the temperature increases further, the bubbles expand, finally bursting
through the top layer, resulting in popping. Meanwhile, the viscosity of the film has
increased enough so that the coating cannot flow together to heal the eruption. Popping
can also result from, or at least be made worse by, entrapment of air bubbles in a
coating. If the surface of the film has a high viscosity, the bubbles may remain in the
film until the coating goes into the oven. The air expands with higher temperature, and
the bubbles may burst through the surface. Air bubbles are especially likely to be entrapped
during spray and hand roller application of waterborne coatings. Bubbling and popping can
result from solvent that remains in primer coats or base coats when the top coat is applied.
Waterborne urethane coatings that have free isocyanate groups such as moisturecurable urethanes (Section 12.6) and 2K waterborne urethane (Section 12.7.3) coatings
24.7. BUBBLING AND POPPING
507
can show bubbling as a result of formation of CO2 from the reaction of isocyanate groups
and water.
In coating plastics, solvents can dissolve in the plastic, causing bubbling or popping
during bake. More commonly, air or moisture in voids in the plastic blow out into the
coating, causing pops. This is usually called gassing but often is indistinguishable from
solvent popping. Similarly, hydrogen and methane trapped in or under the zinc layer in electrogalvanized steel can blow out and give pops (galvanized gassing). Electrical discharge
and hydrogen evolution during cationic electrodeposition at high voltage can produce
pop-like pinholes and craters. Another potential cause of popping is evolution of volatile
by-products of cross-linking after the surface viscosity has increased to the extent that the
bubbles of volatile material cannot readily escape through the surface. Automotive waterborne base coat and solventborne clear coat combinations can experience popping if the
base coat dehydration bake does not remove enough water. The clear coat may also pop
if the dehydration bake is so hot that a porous base coat results. If solvent from the clear
is imbibed by the base coat, it may later blow out through the partially cured clear.
The probability of popping increases with film thickness, since there is a greater chance
of developing a differential in solvent content as film thickness increases. A means of
evaluating the relative likelihood that a series of coatings will show popping is to determine the maximum film thickness of each coating that can be applied without popping
when the films are prepared, flashed off, and baked under standardized conditions [41].
This thickness is called the critical film thickness for popping. Popping can be minimized
by spraying more slowly in more passes, by longer flash-off times before the object is put
into the oven, and by zoning the oven so that the first stages are relatively low in temperature. The probability of popping can also be reduced by having a slow evaporating, good
solvent in the solvent mixture. This tends to keep the surface viscosity low enough for
bubbles to pass through and heal before the viscosity at the surface becomes too high.
Popping can be particularly severe with water-reducible baking enamels, as shown in
Table 24.1. The pairs of enamels were identical except that in one set, they were reduced
for application with solvent, and in the other set, with water [41]. As can be seen, the critical film thickness for popping was consistently lower for the water-reduced coatings. The
data in Table 24.1 also illustrate another variable that affects the probability of popping.
Critical film thickness for popping in these coatings decreases as the Tg of the acrylic resin
in the coating increases. This is true in both solvent- and water-reduced coatings, but the
effect is particularly large in the water-reduced compositions.
There are probably many reasons for the greater difficulty of controlling popping in
water-reducible coatings. A variety of solvents with different evaporation rates is available
for adjusting formulations of solventborne coatings, but water has only one vapor
pressure – temperature curve, and its curve is steeper than for any organic solvent.
Water can be retained by forming relatively strong hydrogen bonds with polar groups
TABLE 24.1. Critical Film Thickness for Popping
Critical Dry Film Thickness (mm)
Copolymer Tg (8C)
228
213
28
14
32
Water
Solvent
50
30
20
10
5
120
.70, , 95
.70, , 95
55
25
508
FILM DEFECTS
on resin molecules at room temperature; these hydrogen bonds break at higher temperature, releasing the water. The heat of vaporization of water, 2260 J g21, is higher than
that of organic solvents, 373 J g21 for 2-butoxyethanol as an example. This higher heat
of vaporization slows the rate of temperature increase of films of water-reduced coatings
in an oven, further increasing the probability of popping [38] (Section 18.3.6).
In contrast to increased probability of popping with higher Tg water-reducible coatings,
popping is more likely to occur with lower Tg latex polymers. Coalescence of the surface
before the water has evaporated completely is more likely with a lower Tg latex.
Popping and cratering often appear similar. It is important to tell the difference because
the remedies are different. One clue is that popping tends to be worst where the film is
thickest, while cratering is more or less independent of thickness. Another is that pops
are often accompanied by unbroken bubbles.
24.8. FOAMING
During manufacture and application, a coating is subjected to agitation and mixing with air,
creating the opportunity for foam formation. Incorporation of foam in a coating can lead to
pinholing or popping. The problem can be particularly severe with waterborne coatings
applied by spray, especially airless spray, or by hand rollers. Formation of a foam involves
the generation of a large amount of surface area; it follows therefore that the lower the
surface tension, the less the energy required to generate a given amount of foam.
However, foam bubbles in pure low viscosity liquids are not stable and break almost instantaneously; there must be something present to stabilize the foam. Although water has a high
surface tension and therefore might not be expected to generate foam bubbles easily, bubbles
in water are more readily stabilized, since a wider variety of components in water can rapidly
migrate to the surface of a bubble to stabilize it. For example, a surfactant not only reduces the
surface tension of water, facilitating foam formation, but also migrates to the surface of the
droplets to give an oriented surface layer with a high viscosity, stabilizing the foam bubbles.
In formulating a latex paint, an important criterion in selecting surfactants or water-soluble
polymers as thickeners is their effect on foam stabilization [3,42]. Acetylene glycol surfactants such as 2,4,7,9-tetramethyl-5-decyne-4,7-diol alkoxylates are reported to be effective
surfactants that do not increase the viscosity of the surface of bubbles as much as surfactants
such as alkylphenol ethoxylates [43].
A variety of additives can be used to break foam bubbles. Most depend on creating
surface tension differential –driven flow on the surface of bubbles. If the surface tension
of a spot on the surface can be lowered, liquid from that area will flow away to try to
cover neighboring higher surface tension areas, weakening and eventually breaking the
thin wall of the bubble. For example, poly(dimethylsiloxane) fluids are effective in breaking a variety of foams, since their surface tension is low compared to almost any foam
surface. Of course, as in other uses of silicone fluids, a little may be fine, but a little
extra can cause problems. Other low surface tension additives, such as poly(octyl acrylates), act as defoamers. Small particle size hydrophobic SiO2 can also act as a defoamer
and/or a carrier for active defoaming agents [43]. Also, a small amount of immiscible
hydrocarbon solvent will often reduce foaming of an aqueous coating. However, a hydrocarbon solvent used in a latex paint as a defoamer was shown to cause flocculation [44].
In this study the paint formula had several surfactants, and the order in which the surfactants were added controlled whether or not flocculation occurred.
REFERENCES
509
Siloxanes with a polyether terminal group have been reported to prevent foaming in
waterborne coatings as well as to minimize entrapment of air bubbles [36]. Several companies sell lines of antifoam additives and offer test kits with small samples of their products.
The formulator evaluates the products in a coating to find one that overcomes, or at least
minimizes, the foaming problem. Although it is possible to predict which additive will
break a foam in a relatively simple system, such predictions are difficult for latex paints
because of the variety of components that could potentially be at the foam interface. The
combination of surfactants, wetting agents, water-soluble polymers, and antifoam can be
critical. A variety of test methods has been used to compare foaminess of coatings [42,43,45].
24.9. DIRT
Dirt has been called the most common defect of all [46]. A wide variety of solid particles
can land on the wet surface of a freshly applied film. If the particles contain low surface
tension materials, they can cause cratering. If not, they become defects if they are large
enough to be visible. Sanding dust, floor dirt, dust blown or tracked in from outdoors,
fibers from wiping cloths or clothing worn by operators, and oven dirt are examples of
such particles. Low power optical microscopy is invaluable for identifying the type of
dirt. Prevention of dirt problems requires clean raw materials, clean paint, and a clean
paint shop, preferably isolated from the rest of the factory. The air supply to spray
booths and spray guns must be clean. Sanding should be minimized and sanding dust
cleaned up before painting. Ovens should be cleaned carefully and frequently. Lint-free
protective clothing and wiping cloths reduce lint contamination. In some applications,
such as auto repair shops, such precautions cannot or probably will not be taken; then
fast drying becomes a desirable characteristic of coatings.
GENERAL REFERENCES
Pierce, P. E.; Schoff, C. K., Coating Film Defects, 2nd ed., Federation of Societies for Coatings
Technology, Blue Bell, PA, 1994.
Schoff, C. K., Coatings Clinic, monthly in JCT Coat. Tech, 1(4), 2004 on.
REFERENCES
1. LeSota, S., Ed., Coatings Encyclopedic Dictionary, Federation of Societies for Coatings
Technology, Blue Bell, PA, 1995.
2. Owen, M. J., Surface and Interfacial Properties in Physical Properties of Polymers Handbook,
American Institute of Physics, Woodbury, NY, 1996.
3. Bierwagen, G. P., Prog. Org. Coat., 1975, 3, 101.
4. Bierwagen, G. P., Prog. Org. Coat., 1991, 19, 59.
5. Smith, R. E., Ind. Eng. Chem. Prod. Res. Dev., 1983, 22, 67. Schwartz, J., J. Coat. Technol.,
1992, 64(812), 65.
6. Scriven, L. E.; Sternling, C. V., Nature, 1960, 187(4733), 186.
7. Orchard, S. E., Appl. Sci. Res., 1962, A11, 451.
8. Patton, T. C., Paint Flow and Pigment Dispersion, 2nd ed., Wiley-Interscience, New York,
1979, p. 554.
510
FILM DEFECTS
9. Overdiep, W. S., in Physicochemical Hydrodynamics, Spalding, D. B., Ed., V. G. Levich
Festschrift, Vol. II, Advance Publications Ltd., London, 1978, p. 683.
10. Overdiep, W. S., Prog. Org. Coat., 1986, 14, 159.
11. Kojima, S.; Moriga, T., Polym. Eng. Sci., 1995, 35, 1998.
12. Wilson, S. K., Surf. Coat. Int., 1997, 80, 162.
13. Cohu, O.; Magnin, A., Prog. Org. Coat., 1996, 28, 89.
14. Ngo, P.-A. P; et al., J. Coat. Technol., 1993, 65(821), 29.
15. Hahn, F. J., J. Paint Technol., 1971, 43(562), 58.
16. Pierce, P. E.; Schoff, C. K., Coating Film Defects, 2nd ed., Federation of Societies for Coatings
Technology, Blue Bell, PA, 1994.
17. Tachi, K.; et al., J. Coat. Technol., 1990, 62(791), 19.
18. Patton, T. C., Paint Flow and Pigment Dispersion, 2nd ed., Wiley-Interscience, New York,
1979, p. 572.
19. Patton, T. C., Paint Flow and Pigment Dispersion, 2nd ed., Wiley-Interscience, New York,
1979, p. 578.
20. Wu, S. H., J. Appl. Polym. Sci., 1978, 22, 2769.
21. Bauer, D. R.; Briggs, L. H., J. Coat. Technol., 1984, 56(716), 87.
22. Hill, L. W.; Wicks, Z. W., Jr., Prog. Org. Coat., 1982, 10, 55.
23. Ellis, W. H., J. Coat. Technol., 1983, 55(696), 63.
24. Maklouf, J.; Porter, S. M., U.S. patent 4,180,619 (1979). Backhouse, A. J., U.S. patent 4,268,547
(1981). Wright, H. J.; et al., U.S. patent 4,290,932 (1981).
25. Ishikura, S.; et al.; Prog. Org. Coat., 1988, 15, 373.
26. Boggs, L. J.; et al., J. Coat. Technol., 1996, 68(855), 63.
27. Brandenburger, L. B.; Hill, L. W., J. Coat. Technol., 1979, 51(659), 57.
28. Eley, R. P.; Schwartz, L. W., J. Coat. Technol., 2002, 74(932), 43.
29. Fink, F.; et al., J. Coat. Technol., 1990, 62(791), 47.
30. Berndimaier, R.; et al., J. Coat. Technol., 1990, 62(790), 37.
31. Schoff, C. K., J. Coat. Technol., 1999, 71(888), 57.
32. Hare, C. M., J. Prot. Coat. Linings, 2001, 18(12), 57.
33. Wolff, U.; et al., Prog. Org. Coat., 2004, 51, 163.
34. Schnall, M., J. Coat. Technol., 1991, 63(792), 95.
35. Waelde, L. R.; et al., J. Coat. Technol., 1994, 66(836), 107.
36. Spiegelhauer, S., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 161 –170.
37. Blunk, R. H. J.; Wilkes, J. O., J. Coat. Technol., 2001, 73(918), 63.
38. Schnall, M., J. Coat. Technol., 1989, 61(773), 33.
39. Ying, H.; et al., JCT Res., 2004, 1(3), 213.
40. Wicks, Z. W., Jr.; Chen, G. F., J. Coat. Technol., 1978, 50(638), 39.
41. Watson, B. C.; Wicks, Z. W., Jr., J. Coat. Technol., 1983, 55(698), 59.
42. Schwartz, J.; Bogar, S. V., J. Coat. Technol., 1995, 67(840), 21.
43. Heilin, W.; et al., J. Coat. Technol., 1994, 66(829), 47.
44. Smith, R. E., J. Coat. Technol., 1988, 60(761), 61.
45. Kozakiewicz, J.; et al., J. Coat. Technol., 1993, 65(824), 47.
46. Schoff, C. K., private communication, 2004.
25
Solventborne and
High-Solids Coatings
Chapters 25 through 29 present the principal classes of coatings: solventborne,
waterborne, electrodeposition, powder, and radiation cure coatings. Our intent is to
discuss the principles involved in these classes and to compare the various types of
resins applicable to each group. In Chapters 30 through 33, end use applications for
these various types of coatings are discussed.
Historically, almost all coatings were solventborne. The original motivation to reduce
solvent was to lessen fire hazards and odor and to permit cleanup with water. Since the
1960s, a major driving force has been to reduce VOC emissions. There have been significant reductions in solvent use by shifting to other classes of coatings. Some expect that in
the long run, solventborne coatings will disappear, but there are advantages to solventborne coatings over the alternatives. Capital cost for application is generally lower,
especially as compared to waterborne coatings, for which stainless steel equipment is
needed. Electrostatic spray installations for solventborne coatings are less expensive
than those for waterborne coatings. Solvent evaporation is not dependent on humidity.
There are fewer problems of air entrapment and popping. Significant reductions in
VOC emissions have been made by shifting to high-solids coatings and by refining application methods to minimize solvent requirements. Current research and development is
aimed at still higher-solids coatings and, ultimately, at solventless (solvent-free) liquid
coatings.
For some end uses, especially when cost is particularly important, a single coating
is adequate; however, for many other end uses, performance requirements can only
be met by applying at least two coats. Almost always when more than one coat is to be
applied, it is preferable to have a primer specifically designed to be the first coat, with
a different coating for the top coat. The primer is designed to adhere strongly to the
substrate and to provide a surface to which the top coat adheres well. It is not necessary
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
511
512
SOLVENTBORNE AND HIGH-SOLIDS COATINGS
for the primer to meet such other requirements as exterior durability that are critical for top
coat performance. Commonly, primers are less expensive than top coats.
25.1. PRIMERS
The first consideration in formulating a primer is to achieve adequate adhesion to the substrate (Chapter 6). The substrate should be clean and preferably have a uniformly rough
surface. The surface tension of the primer must be lower than that of the substrate. The
viscosity of the continuous phase of the primer should be as low as possible to promote
penetration of the vehicle into pores and crevices in the surface of the substrate.
Penetration is also promoted by use of slow evaporating solvents, use of slow cross-linking
systems, and whenever feasible, use of baking primers. The primer binder should have
polar groups scattered along the backbone of the resin that can interact with the substrate
surface. The presence of salts within the film should be minimal. Interaction between the
binder and the substrate should be such that the coating will resist displacement by water
when water molecules permeate through the coating to the interface (Section 7.2.3). For
primers over metal substrates, as well as over alkaline surfaces such as masonry, saponification resistance of the primer is an important criterion in binder selection.
25.1.1. Binders for Primers
The binders in many primers for metal substrates are bisphenol A (BPA) epoxy resins and
their derivatives (Sections 13.1 to 13.3 and 15.8). Although a full explanation is not available, experience has shown that in general, BPA epoxy resin –based coatings exhibit
superior adhesion to clean metal. Addition of a small amount of an epoxy phosphate
ester (Section 13.5) can further enhance adhesion, especially wet adhesion. For baked
coatings, epoxy –phenolic coatings are particularly appropriate. For air dry coatings,
epoxy – amine coatings are commonly selected. Both types have excellent wet adhesion
and saponification resistance, critical for long-term corrosion protection. In general,
baked epoxy –phenolics excel in these properties. Primers based on alkyd (Chapter 15)
and epoxy ester (Section 15.8) resins are widely used. Material and application costs
are less than with epoxy coatings, but wet adhesion is usually not as good, except possibly
to metal contaminated with oily residues. Saponification resistance of alkyd resins is
limited. The saponification resistance of epoxy esters is intermediate between those of
alkyds and epoxy – amines.
The lowest cost binders used in metal primers are styrenated alkyds (Section 15.6). They
are generally used in air dry primers. The higher Tg resulting from the large fraction of aromatic rings leads to primers that give dry-to-touch films, or even dry-to-handle films,
rapidly. Styrenation reduces the fraction of ester groups by dilution and hence may increase
saponification resistance. However, styrenated alkyds cross-link more slowly than do their
nonstyrenated counterparts. As a result, the films develop solvent resistance slowly and
have a wider time window within which they should not be top coated, owing to insufficient
cross-linking. If the primer is top coated within this time interval, lifting is likely to occur,
resulting in puckered areas in the top coat surface, due to nonuniform degrees of swelling
by the top coat solvent and to uneven film shrinkage when the film finally dries.
The surface of galvanized steel is a layer of zinc hydroxide, zinc oxide, and zinc
carbonate; all are strong, somewhat water-soluble bases. Especially if the metal is not
25.1. PRIMERS
513
phosphate treated shortly before coating, the binder in the primer must be resistant to saponification. Alkyd primers lack the necessary saponification resistance and are likely to
perform poorly, especially if the galvanized surface has been exposed to moisture. For
air dry coatings, acrylic latex vehicles are more appropriate than alkyds.
Primers are also required for surfaces other than metal substrates. Adhesion to polyolefin plastics can be difficult to achieve and can be enhanced by use of a primer based
on a chlorinated polymer (Sections 6.5 and 31.2.2). Concrete and other masonry surfaces
are alkaline and often require special surface treatments, most commonly washing with
hydrochloric or phosphoric acid. The acid wash not only neutralizes the surface alkalinity
but also etches the surface. Saponification resistant primers such as epoxy –amine or latex
primers, provide longest life. Concrete blocks have very porous surfaces, and penetration
of solventborne paints into the surface requires a relatively large volume of coating
for coverage. Substantially better coverage can be achieved by using latex paints
(Section 32.1).
25.1.2. Pigmentation of Primers
Selection of pigments and their amount is critical in primer formulation (Chapter 22).
Pigmentation level affects adhesion of top coats to the primer film. Formulations with a
high PVC/CPVC ratio are low-gloss coatings; the roughness and increased surface area
of low-gloss films give improved intercoat adhesion. In some cases it is desirable to formulate primers with PVC . CPVC. The resulting primer film is somewhat porous, permitting penetration of top coat vehicle into the pores, thereby promoting intercoat
adhesion. The PVC should be only slightly higher than CPVC or the loss of vehicle
from the top coat may be enough to raise the PVC of the top coat film sufficiently to
reduce its gloss. Primers with PVC . CPVC can be sanded more easily and are less
likely to clog sand paper than primers with PVC , CPVC. Since inert pigments are generally the least expensive components of the dry film, high PVC minimizes cost.
Since primers almost always are low gloss coatings, inexperienced formulators may not
think it is necessary to have good pigment dispersions for primers. This is not so! Since
CPVC is affected by the degree of dispersion and particularly by the extent of flocculation,
and since primers are usually formulated to be either slightly above or slightly below
CPVC, pigment dispersion can be critical. If the pigment dispersion is not stabilized
properly, CPVC decreases; so its PVC could become greater than CPVC.
If only one color top coat is to be applied over the primer, it is usually desirable for the
primer to have a similar color, since this minimizes the effect of the primer color on the
final top coat color. In many cases, top coats of several different colors are applied over
the same primer; then, it is usually desirable to use a light gray primer. Gray primers
have better hiding than white primers, when a pigment that strongly absorbs light, such
as lamp black, is present. Light gray primers have relatively little effect on top coat
color. Primers pigmented with red iron oxide provide good hiding at low cost but are
harder to hide.
In addition to adhesion, water and oxygen permeabilities are important factors in corrosion protection of metals by barrier coatings (Section 7.2.3). The pigments can strongly
affect oxygen and water permeability. Up to a PVC/CPVC ratio of about 0.9, the higher
the PVC of the film, the lower the gas and vapor permeability. To occupy the largest
possible volume of the dry film with impermeable pigment, a combination of pigments
giving a high CPVC should be used with PVC close to CPVC. Platelet-shaped pigment
514
SOLVENTBORNE AND HIGH-SOLIDS COATINGS
particles tend to provide better barriers to oxygen and water permeability. Mica and micaceous iron oxide are widely used in primers for metals because of their platelet form. One
must also use resin – pigment combinations in which the resin is strongly adsorbed on the
pigment surface. If the pigment has a polar surface with a weakly adsorbed resin, water
permeating through the film may displace the resin from the pigment surface, leading to
an increase in water permeability.
Pigments for barrier primers should be completely insoluble in water. If, for example, zinc
oxide were used as a pigment in a primer formulation, some of the zinc oxide would dissolve
in water permeating through the film, resulting in establishment of an osmotic cell that can
cause blistering. Passivating pigments can be useful for protecting steel against corrosion
when the coatings on the steel substrate have been ruptured (Section 7.4.2). However,
passivating pigments must be somewhat soluble in water in order to passivate the steel.
Furthermore, dissolving the pigments requires that the binder is able to swell with water to
a degree. These characteristics mean that blistering is more likely with a primer that contains
a passivating pigment. In most baking primers for OEM products, it is preferable not to use
passivating pigments but to rely on barrier properties to provide corrosion protection.
However, coatings for bridges, storage tanks, ships, and offshore drilling platforms cannot
be baked, and in such cases, film rupture must be anticipated. Passivating pigments or
zinc-rich primers are commonly used for such applications (Sections 33.1.2 and 33.1.3).
25.1.3. High-Solids Primers
In conventional low solids primers, the effect of pigmentation on the viscosity of the
coating is relatively small if the pigment is not flocculated. However, as solids are
Figure 25.1. Effect of pigmentation on viscosity as a function of volume solids for an unpigmented coating and for two pigmented coatings based on the same binder with pigment loadings sufficient to give
20% PVC and 45% PVC in the dry films. See Ref. [1] for the assumptions made in the calculations.
(From Ref. [1], with permission.)
25.2. TOP COATS
515
increased, the volume of the dispersed phase (including the volume of both the pigment
and of the adsorbed layer on the surfaces of the pigment particles) increases and
becomes an important factor controlling the viscosity of the coating, limiting the solids
at which a coating can be applied.
Pigmentation levels such as those encountered in low gloss coatings with PVC of 45 or
higher substantially increase the viscosity of the wet coating, requiring reduction of solids
for application. Figure 25.1, based on model calculations, shows plots of viscosity as a
function of volume solids for three sets of calculations: one for an unpigmented
coating, one for a 20 PVC (i.e., a gloss) coating, and one for a 45 PVC (i.e., low gloss)
coating. The assumptions made in carrying out the calculations are provided in Ref. [1].
With the present status of our knowledge, somewhere in the neighborhood of 60 NVV
is probably an upper limit for a primer with a PVC close to or above the CPVC. As a result,
some high-solids primers are formulated with PVC that is much lower than optimum. A
challenge to increasing solids content above 60 NVV at the optimum PVC is a means
of stabilizing the pigment dispersion in a primer with thinner adsorbed layers at reasonable
cost. However, even assuming stabilization with an adsorbed layer thickness of 5 nm, the
upper limit of volume solids is at most 75 NVV. As a result of such limitations, the major
thrust of development of reduced VOC primers is on waterborne primers.
25.2. TOP COATS
Top coats include coatings applied over primers and directly on a substrate. In the former
case, the primer provides adhesion to the substrate and a major part of the corrosion protection for coatings on metal. The top coat must adhere well to the primer and provide the
desired appearance and other properties. A single coating layer must combine both functions. In general, it is preferable to use a primer/top coat system; however, one coat application can be functional for many applications and is less expensive than multiple-coat
systems. Single coats are used on products that need little corrosion protection and for
which the need to maintain adhesion in the presence of water is not critical. When appearance and exterior durability requirements are minimal, a primer with excellent corrosion
protection properties may be used without a top coat: for example, inside the ballast tanks
of ships and for the interior of structural components of aircraft.
25.2.1. Binders for Top Coats
To an important degree, the properties of a top coat are controlled by the class of resin used
as the principal binder in the top coat. The chemistry of these various binders is discussed
in Chapters 8 through 17; in this chapter, the advantages and disadvantages of some
classes of resins used in top coats, especially for OEM product coatings on metal, are
compared.
25.2.1.1. Alkyds From the mid-1930s through the mid-1950s, alkyds (Chapter 15)
were major binders for coatings. Although being replaced increasingly by other binders,
alkyds are still used on a large scale. A major advantage of alkyds in many applications
is lower cost. A second major advantage is that application of alkyd coatings is, in
general, the most foolproof. Solventborne alkyd coatings are least subject to film
defects of all the classes of coatings. This advantage results from the low surface
516
SOLVENTBORNE AND HIGH-SOLIDS COATINGS
tension of most alkyd coatings. Therefore, there are seldom problems with crawling, cratering, and other defects that result from surface tension – driven flows or surface tension
differential –driven flows (Section 24.4). It is relatively easy to make pigment dispersions
that do not flocculate. A further major advantage is their ability to cross-link by autoxidation. This makes air dry or low temperature baking possible and avoids the need for crosslinking agents with potential toxic hazards. In essence, air is the cross-linker.
The major limitations of alkyds are comparatively poor color retention on baking and
limited exterior durability and saponification resistance. It is also difficult to achieve very
high solids in alkyd solventborne coatings (Section 15.2). A further problem can be generation of smoke in baking ovens that causes visual air pollution problems.
Oxidizing alkyd binders are used for low cost coatings such as for steel shelving,
machinery, coat hangers, and exterior coatings for drums. Individual applications may
be relatively small, but there are a host of products for which the performance of these
coatings is adequate. Oxidizing alkyds are also used for a significant fraction of architectural gloss enamels (Section 32.3.1). With metal salt driers, oxidizing alkyd coating films
air dry in a few hours, force dry at 60 to 808C in an hour, or bake at 120 to 1308C in half an
hour. Baking time is shortest with highly unsaturated drying oil– derived alkyds having an
oil length of about 60. Tung oil– modified alkyds give the fastest cross-linking, followed in
turn by mixed tung –linseed oil alkyds and then linseed oil– modified alkyds. Cured films
from these alkyds are quite yellow and turn yellow brown on overbaking. Generally, they
are used only for dark color coatings. Since the cured films still contain substantial
amounts of unsaturation and metal driers, the films embrittle with age. Better color and
color retention can be obtained by using alkyds made with fatty acids or oils such as
soy or tall oil fatty acids that are less unsaturated. These alkyds are also less expensive,
but they cure more slowly than do more highly unsaturated alkyds.
Medium oil alkyds have a substantial number of hydroxyl groups on each molecule;
these functional groups can be cross-linked with a variety of cross-linkers. Since
melamine – formaldehyde (MF) cross-linkers are generally lowest cost, they are the
most widely used. To achieve compatibility, butylated, isobutylated, or octylated MF
resins are used, not methylated ones. Generally, the more reactive Class II MF resins
with a high NH content are used (Section 11.2). The solvent system should contain
some butyl alcohol to promote package stability and reduce viscosity compared to straight
hydrocarbon solvents. Usually, no added catalyst is needed, since the residual unreacted
carboxylic acid groups on the alkyd are sufficient catalyst for the Class II MF resin. If
needed, relatively weak acid catalysts such as an alkylphosphoric can be used. Color
and speed of cure are affected by the type of fatty acid used in the alkyd. Highly unsaturated alkyds are seldom used in top coats because of their dark color, soy, tall oil, and
dehydrated castor alkyds are preferred. Some of the cross-linking results from the unsaturation and the balance from the reactions of the MF resin. The color, color retention,
and resistance to embrittlement are better than those obtained using the oxidizing alkyd
without the MF resin, since the MF cross-linking minimizes or eliminates the need for
metal driers to catalyze the oxidation reactions.
The best color, color retention, and exterior durability are obtained using nonoxidizing
saturated fatty acid –based alkyds. To minimize degradation due to hydrolysis under acidic
conditions, alkyds made with isophthalic acid (or hexahydrophthalic anhydride) are used
rather than those from phthalic anhydride. For example, solid color exterior top coats
(monocoats) based on coconut oil alkyds and MF resins were used by many automobile
manufacturers at one time. The exterior durability is good but not as good as the best
25.2. TOP COATS
517
acrylic coatings. Alkyd coatings are said to exhibit a greater appearance of depth than is
obtained with acrylic – MF coatings. The cost of these coatings is similar to polyester –
melamine coatings, and film defects during application are less likely.
The hydroxyl groups on alkyds can also be cross-linked at room temperature or under
force dry conditions with polyisocyanates. IPDI isocyanurate prepolymers are often used
for this purpose, since their color stability and exterior durability are superior to TDI
derivatives (Section 12.3.2). In many cases, use of a polyisocyanate with an oxidizing
alkyd permits sufficiently rapid development of tack free coatings so that metal driers
are not needed; this further improves exterior durability.
25.2.1.2. Polyesters Polyester resins (Chapter 10) are one of the major classes of
resins replacing alkyd resins in MF cross-linked baking enamels. Polyesters are also
widely used in urethane coatings. Their cost is generally somewhat higher than that for
oxidizing alkyds, but in some cases is less than that of nonoxidizing alkyds. Color,
color retention, exterior durability, and resistance to embrittlement are better than those
obtained with most alkyds, but exterior durability and resistance to saponification are generally not as good as with acrylics. The adhesion and impact resistance of polyester-based
coatings without primers over clean, treated steel and aluminum substrates are comparable
to those of alkyds and generally superior to those of acrylics. Polyester coatings generally
have higher surface tensions than those of alkyd coatings and hence are more subject to
crawling and surface tension differential – driven flow defects such as cratering. Most polyesters are hydroxy-terminated and are cross-linked with MF resins or with isocyanates.
The MF resins are less expensive; methylated or mixed methylated –butylated MF
resins are generally used. For lower temperature cure or air dry coatings, aliphatic polyisocyanates are used. A major advantage of polyesters over alkyds and most acrylic
resins is the relative ease of preparing polyester resins suitable for very high solids and
even solventless coatings (Section 25.2.2). Polyesters are also adaptable for use in
powder coatings (Chapter 28).
25.2.1.3. Acrylics In general terms, the major advantages of acrylic binders (Chapter
8) are their low color, excellent color retention, resistance to embrittlement, and exterior
durability at relatively modest cost. The best acrylics are quite photochemically stable
and very hydrolytically stable. In general, their surface tensions are intermediate
between alkyds and polyesters, and as a result, the susceptibility of acrylic based coatings
to film defects is intermediate. Generally, their adhesion to metal surfaces is inferior to that
of both alkyds and polyesters coatings; therefore, they are generally used over a primer.
Thermoplastic solution acrylic resins were used widely in OEM automotive coatings
but have been replaced with thermosetting acrylics (TSAs) to reduce VOC emissions
(Chapter 8). Most commonly, TSAs are hydroxy-functional, often with a minor amount
of carboxylic acid functionality. They are cross-linked with MF resins or with polyfunctional isocyanates. Either Class I or Class II MF resins can be used; the choice usually
depends on curing temperature requirements. Aliphatic isocyanate cross-linkers are
more expensive than MF resins and present greater toxic hazards but cure at lower temperatures and with HALS stabilizers usually provide somewhat greater exterior durability
and frequently have better environmental etch resistance.
25.2.1.4. Epoxies and Epoxy Esters Although the major uses for epoxy resins
(Chapter 13) are in primers, significant volumes of BPA and novolac epoxy-based top
518
SOLVENTBORNE AND HIGH-SOLIDS COATINGS
coats are used in applications where exterior durability is not an issue. They are well suited
to such purposes because adhesion of epoxy-based coatings in the presence of water or
water vapor is generally superior to other types, and they are highly resistant to saponification. Applications such as beer and soft drink can linings and the insides of aircraft
wings do not require exterior durability. Epoxy-functional acrylic resins cross-linked
with polybasic acids give films with excellent exterior durability (Section 13.1.2).
Epoxy esters (Section 15.8) offer properties intermediate between alkyds and epoxies.
An example of their use in top coats is in coating bottle caps and crowns. They exhibit the
requisite combination of hardness, formability, adhesion, and resistance to water.
25.2.1.5. Urethanes Polyisocyanate cross-linkers (Chapter 12) have been mentioned
in connection with alkyds, polyesters, and acrylics. They are widely used as cross-linkers
because of low temperature curing and abrasion resistance combined with resistance to
swelling of cured films with solvent. Urethanes can also be backbone links in resins
that have reactive groups other than isocyanates. Hydroxy-terminated urethane resins
(Section 12.8) can be cross-linked with polyisocyanates or MF resins. When MF resins
are used, the coatings have the advantage of being one package coatings without a
concern for the toxic hazards of free isocyanate. As with polyesters, there is no difficulty
making low molecular weight resins where substantially all of the molecules have a
minimum of two hydroxyl groups. Cost is higher than for polyesters, but the hydrolytic
stability and therefore the exterior durability can be superior. However, the intermolecular
hydrogen bonding of urethanes leads to higher viscosity at equal concentration solutions of
equal molecular weight compared to polyester resins. Low molecular weight, hydroxyfunctional urethanes can be blended with other generic types of hydroxyl-functional
resins to modify coating properties.
Urethanes, especially those having primary carbamate 22O22CO22NH2 groups, crosslink with MF resins to give greater hydrolytic stability under acidic conditions than that of
hydroxy-functional resins cross-linked with MF. Carbamate-functional acrylic, polyester,
and silicone resins with MF cross-linking are useful in clear automotive coatings.
25.2.1.6. Silicone and Fluorinated Resins The highest resistance to both thermal
degradation and photoxidation is obtained with silicones (Section 16.1) and fluorinated
resins (Section 17.1.4). The cost of both is high, especially that of fluorinated resins. Silicone alkyds give greater exterior durability than alkyds for air dry coatings. In a compromise to obtain superior outdoor durability at an intermediate cost, silicone-modified
polyesters and acrylics (Section 16.1.3) are relatively widely used. Fluorinated copolymer
resins cross-linked with MF resin or aliphatic polyisocyanates have outstanding exterior
durability.
25.2.2. Formulating Solventborne Coatings for Low VOC
The purpose of this section is to consider the various types of top coats, with emphasis on
the potential for reduction of solvent use while achieving the required level of performance. Although thermoplastic resins are still being used, their use will continue to fall,
owing to their high VOC. The majority of solventborne coatings currently used and
expected to be used in the future are thermosetting coatings. Conventional thermosetting
coatings have volume solids (NVV) on the order of 25 to 35%. Technical efforts on
25.2. TOP COATS
519
high-solids coatings since around 1970 have accomplished varying degrees of increases in
solids of various types of coatings for various applications.
There is no single definition of high-solids coatings. For metallic automotive top coats
(or base coats), high solids corresponds to about 45 NVV. For a highly pigmented primer,
high solids might be 50 NVV. For clear or high gloss pigmented coatings, 75 NVV or even
higher is possible. The situation is further complicated by the difficulty of exact measurement of, or even of defining, VOC of a coating (Section 18.9.1). For example, in some
cases, solvents with functional groups may partially react with the cross-linker and
hence not be evolved. On the other hand, there can be volatile by-products of cross-linking
that should be included in VOC, and low molecular weight components may volatilize
before cross-linking; the extent to which such emissions occur may vary with baking
conditions.
A limitation on solids content is the increasing difficulty of achieving desired mechanical properties as molecular weight and average functionality per molecule f̄n are
decreased and molecular weight distribution is narrowed (Section 8.2.1). In a conventional
thermosetting acrylic coating, one uses a resin with an M̄w/M̄n on the order of 35,000/
15,000, and depending on the end use, an f̄n of 10 to 20. The cross-linker might have an
M̄w/M̄n on the order of 2000/800 and an f̄n of 3 to 7. An acrylic resin that permits 45
NVV might have M̄w/M̄n on the order of 8000/3000 with f̄n of 3 to 6. As f̄n is reduced,
it becomes increasingly critical and difficult to control formulations and cure conditions
to keep all film properties within the desired ranges.
As one aims for still lower molecular weights and functionalities for still higher-solids
coatings, achieving high performance properties becomes even more difficult. For an NVV
of 70, M̄w/M̄n must be on the order of 2000/800 or less, with an average of a little over two
hydroxyl groups per molecule. It is critical that essentially all molecules have at least two
functional groups per molecule. Any molecules with only one functional group cannot
cross-link and will leave dangling ends in the network; any molecules with no functional
groups will be plasticizers, and if low enough in molecular weight, may partially volatilize
in a baking oven. It is difficult to synthesize acrylic resins for very high solids coatings
(Section 8.2.1). On the other hand, it is relatively easy to make polyurethane or polyester
resins with an M̄w/M̄n of 2000/800, or even lower, and with an average of a little over two
hydroxyl groups per molecule where essentially all of the molecules have at least two
hydroxyl groups (Sections 10.2 and 12.8).
Major progress is being made in solving the problems with conventional free radical
polymerization by using controlled free radical polymerizations (CFRPs) (Section
2.2.1.1). A resulting SuperSolids automotive acrylic clear coat is in production; solids
are reported to be 65%, with the possibility of future increases up to nearly 90%. At the
same time both the scratch and mar resistance are substantially improved [2].
In conventional coatings, the cure window is relatively large; that is, it makes little
difference if the baking temperature, baking time, or catalyst amount is off by +10%.
In high-solids coatings, the cure window is narrower [3]. If there are a large number of
hydroxyl groups on each resin molecule and 10% are not reacted, the change in properties
may be small. If, however, there are only a little over two hydroxyl groups per average
molecule and 10% are left unreacted, a significant fraction of the molecules will be tied
into the network in only one place, with detrimental effects on film properties. The
problem may be minimized by using cross-linkers with a higher f̄n. Because of their
greater average functionality, Class I MF resins generally offer broader cure windows
than those of Class II MF resins (Section 11.3). The extent of self-condensation
520
SOLVENTBORNE AND HIGH-SOLIDS COATINGS
reactions of MF resins is particularly dependent on time, temperature, and catalyst
concentration.
In using high-solids coatings, the applicator should be very careful in controlling the
time and temperature in baking ovens and in following the coating supplier’s recommendations. The formulator must be more careful in checking the film properties when the
temperature is about 108C above and below the standard temperature. In making
recommendations of cure cycles to a customer, one should use pieces of the customer’s
metal to establish the baking schedule. The critical temperature for baking is that of the
coating itself, not that of the air in the oven. Coatings applied to a heavy piece of metal
heat up more slowly than do coatings on light gauge sheet metal. On sheet metal, the
coating over a place where the sheet metal has been welded to a supporting member
heats up more slowly than it does on the rest of the surface. High-solids coatings are
more likely to be subject to variation in properties due to differences in coating temperature and time compared to conventional coatings.
To increase solids, polydispersity of resins can be decreased; however, this may narrow
the breadth of the Tg transition range (Section 4.2), which may adversely affect the mechanical properties of films [4]. An approach to overcoming this problem is to blend resins with
differences in composition, but still similar enough to each other to be compatible; for
example, polyester–acrylic blends using low molecular weight hydroxy-terminated polyesters to reduce the viscosity and thereby increase the solids [5]. Resins and cross-linkers
for high-solids coatings generally have an M̄n well below 5000. In contrast to high molecular
weight polymers, the entropy of mixing of different low molecular weight resins is large
enough to be a significant factor favoring compatibility (Section 18.2). The effect of molecular weight is illustrated by 50 : 50 blends of several acrylic and methacrylic homopolymers
that are compatible, when their M̄n values are less than 5000 but incompatible when their
M̄n values are above 10,000 [6]. The broader compatibility of low-M̄n resins permits formulation of high-solids coatings based on mixtures of resins different generic types. Not all
blends of low-M̄n resins are compatible, and there is the possibility that phase separation
could occur as molecular weight increases in the early stages of cross-linking. Unpigmented
films should be made and checked for transparency; haziness suggests phase separation may
occur on a scale large enough to detract from appearance or properties.
Isocyanurates, unsymmetrical trimer, biuret, and allophanate polyisocyanates (Section
12.3.2) with lower molecular weights and hence lower viscosities have been made available. However, the lower viscosity of these polyisocyanates is only one factor affecting the
VOC of coatings [7]. In some cases, the lower viscosity cross-linker also has a lower
equivalent weight, which requires that the weight ratio of low viscosity cross-linker to
higher viscosity polyol is decreased. For this reason, the lowest VOC could be obtained
using a somewhat higher equivalent weight polyisocyanate, even though it had a somewhat higher viscosity. Aldimines and hindered diamines permit preparation of very
high solids clear top coats (Section 12.4).
Research is now focused on ultra high-solids or solvent-free coatings. These are based
on the use of low molecular weight oligomers or reactive diluents; the distinction between
the terms is disappearing. VOC of two package polyurethane coatings can be reduced by
replacing some of the hydroxy-terminated acrylate or polyester with a reactive diluent [7].
Examples of such coreactants are aliphatic diols and low molecular weight ester or
urethane diols.
Polyesters afford the major opportunity for very low VOC coatings, since low molecular
weight oligomers can be made such that all molecules have at least two reactive groups,
25.2. TOP COATS
521
usually hydroxyl groups. For example, 1,4-butanediol esters of mixed glutaric, adipic, and
azelaic acids with an M̄n of 300 have a viscosity at 258C of 250 Pa.s [8]. Solvent-free
coatings can be formulated with polyisocyanate cross-linkers. With MF resins, film hardness of baked films is too low; however, the coatings can be thinned with water (Section
18.4), permitting formulation with a variety of other components in solvent-free coatings [8].
An example of a “zero” VOC coating that uses water as a diluent is an epoxy – hydroxyl
system [9]. Caprolactone polyols are used with 3,4-epoxycyclohexylmethyl –3,4-epoxycyclohexane carboxylate at a 2 :1 epoxy/hydroxyl ratio, with a latent triflic acid derivative
as catalyst.
VOC reduction is being aided by application methods that permit use of higherviscosity coatings. Hot spray (Section 23.2.4), high speed electrostatic disks (Section
23.2.3), and supercritical fluid spray (Section 23.2.5) are examples. Improvements in transfer efficiency are making a significant contribution to reducing VOC emissions (Sections
23.2.1 and 23.2.3). For example, high volume low pressure air spray guns have a
substantially superior transfer efficiency than that of conventional air atomization guns.
As solids increase, it becomes more difficult to avoid pigment flocculation [10]. The
primary factor controlling stabilization of pigment dispersions is the thickness of the
adsorbed layer on the surface of the pigment particles (Section 21.1.3). The very low molecular weight resins used in very high solids coatings are incapable of providing an adequate adsorbed layer. Furthermore, as the number of functional groups per molecule
decreases, the solvent can compete more effectively for adsorption sites on the pigment
surface, promoting flocculation. Hyperdispersants (Section 21.1.3) have been developed
for use in high-solids coatings.
Another limiting factor for some high-solids coatings is that of surface tension effects.
Generally speaking, as molecular weight gets lower, the equivalent weight must be
lowered even further, since the number of reactions required to achieve high molecular
weight will be greater. In most coatings, the functional groups are highly polar, such as
hydroxyl and carboxylic acid groups. Increased levels of such groups give higher
surface tensions. Furthermore, achieving high solids at a given viscosity with such
resins generally requires using hydrogen-bond acceptor solvents rather than hydrocarbon
solvents. Again, this results in higher surface tensions than that of most conventional coatings and therefore increased probability of film defects during application. When coating
metal, the importance of clean surfaces increases as solids increase. When coating plastics,
one must be careful to remove mold release agents. A broader range of plastic materials
has to be surface treated to avoid crawling and achieve adhesion (Sections 6.5 and 31.3).
Since the surface tension of a freshly applied high-solids coating is generally higher than
that of a conventional coating, a larger fraction of contaminating particles floating in the
air have lower surface tensions than that of the wet high-solids coating. Thus, cratering is
more probable (Section 24.4).
Sagging is more of a problem with high solids than with conventional coatings [11].
This results from the slower rate of evaporation of solvents from high-solids coatings
than from conventional coatings (Section 24.3). Although the reasons for this difference
have not been elucidated completely, the consequences pose significant problems.
Sagging of spray applied high-solids coatings cannot be controlled easily by adjusting
the evaporation rate of the solvent in the coating or by changing the distance between
the spray gun and the substrate. Sagging can be minimized by using hot spray (Section
23.2.4) or supercritical fluid spray (Section 23.2.5). For many applications, thixotropic
flow properties must be built into a high-solids coating for spray application. Fine
522
SOLVENTBORNE AND HIGH-SOLIDS COATINGS
Figure 25.2. Viscosity as a function of temperature for a conventional and a high-solids resin solution.
The high-solids solution is of a 1500 molecular weight polyester at 90% solids in ethylene glycol monoethylether acetate. The conventional solution is of a 20,000-molecular weight polyester at 25% solids in
methyl ethyl ketone and ethylene glycol monoethyl ether acetate. (Adapted from Ref. [1], with
permission.)
particle-size SiO2, bentonite clay pigments, zinc stearate, and polyamide gel thixotropes
are additives that can be useful. Another alternative is to include a modest fraction of
highly reactive cross-linker that will cause a quick increase in viscosity; mixtures of isocyanates and MF resins can be used [12]. The problem of sagging in metallic coatings used
on automobiles is particularly serious; microgel particles have been developed that are
effective (Sections 24.3 and 30.1.2).
Even when sagging is not encountered during application, it can sometimes occur with
high-solids coatings during baking. The temperature dependence of viscosity of highsolids coatings is greater than for conventional coatings, as illustrated in Figure 25.2.
Before the cross-linking reactions have proceeded far enough to increase the viscosity,
the increasing temperature during baking has already decreased the viscosity sufficiently
to lead to sagging [1,11]. The phenomenon is called oven sagging.
GENERAL REFERENCE
Hill, L. W.; Wicks, Z. W., Jr., Prog. Org. Coat., 1982, 10, 55.
REFERENCES
1.
2.
3.
4.
5.
6.
Hill, L. W.; Wicks, Z. W., Jr., Prog. Org. Coat., 1982, 10, 55.
http://automotive.dupont.com/en/newsEvents/article20030305b.html.
Bauer, D. R.; Dickie, R. A., J. Coat. Technol., 1979, 54(685), 101.
Kangas, S. L.; Jones, F. N., J. Coat. Technol., 1987, 59(744), 99.
Hill, L. W.; Kozlowski, K., J. Coat. Technol., 1987, 59(751), 63.
Cowie, J. M. G.; et al., Macromolecules, 1992, 25, 3170.
REFERENCES
523
7. Jorissen, S. A.; et al., Proc. Waterborne Higher-Solids Powder Coat. Symp., New Orleans,
LA, 1992, p. 182.
8. Jones, F. N., J. Coat. Technol., 1996, 68(852), 25.
9. Eaton, R. F.; Lamb, K. T., J. Coat. Technol., 1996, 68(860), 49.
10. Hochberg, S., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA, 1982, p. 143.
11. Bauer, D. R.; Briggs, L. M., J. Coat. Technol., 1984, 56(716), 87.
12. Teng, G.; et al., Polym. Mater. Sci. Eng., 1997, 76, 117.
26
Waterborne Coatings
Before about 1950, almost all coatings were solventborne. Introduction of latex architectural paints was the first major step away from solventborne coatings; the driving forces
were better performance, easier clean up, and reduced fire hazards. Since 1970, there
has been a further trend away from solventborne coatings to reduce VOC emissions,
and use of waterborne coatings in the United States now exceeds the volume of solventborne coatings. The trend is expected to continue as restrictions on VOC become more
stringent. Most waterborne coatings contain some organic solvent. The solvents play a
variety of important roles in resin manufacture, coating production and application, and
film formation. Research efforts are under way to reduce and in some cases, eliminate
the need for solvents.
The two largest classes of waterborne coatings are water-reducible coatings (Section
26.1) and latex coatings (Section 26.2); smaller amounts of emulsion coatings (Section
26.3) are also used. The terminology is not uniform [1]. We use the term water-reducible
for resins made in solvent and reduced with water to form a dispersion of resin in water.
Other authors refer to these resins as aqueous dispersion resins or water-soluble resins, but
they are not really soluble in water. Some authors call latex coatings emulsion coatings; we
recommend against this practice to avoid confusion of latexes (dispersions of solid
polymer particles in water) with coatings that really are emulsions (dispersions of
liquids in liquids). Latex resins are prepared by emulsion polymerization. Polyurethane
latexes are called aqueous dispersion resins. There is growing use of blends of
different types, for example, combinations of latex and water-reducible resins. Water is
also used as a cosolvent in solventborne coatings, which we call water thinning
(Section 25.2.2).
In relation to organic solvents, water has both advantages and disadvantages. Water
presents no toxic hazard, is odor free, and is not flammable. Nonflammability reduces
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
524
26.1. WATER-REDUCIBLE COATINGS
525
risks and insurance costs and permits use of less makeup air in baking ovens, reducing
energy consumption in some cases. There are no emission or disposal problems directly
attributable to use of water. With some formulations, cleanup of personnel and equipment
is easier with waterborne coatings; however, in other cases, cleanup is more difficult. The
cost of water is low; but it does not necessarily follow that the cost of waterborne coatings
is low.
A wide range of solvents with different evaporation characteristics enables the formulator to fine-tune evaporation rates from solventborne coatings, but there is only one type
of water. The heat capacity and heat of vaporization of water are high, resulting in high
energy requirements for evaporation. With a given amount of energy available for evaporation, water evaporates more slowly than do solvents with similar vapor pressures,
although the vapor pressure of water increases relatively rapidly with increasing temperature. Evaporation of water is affected by relative humidity (RH) and variations in RH when
coatings are applied can lead to problems (Section 26.1).
The surface tension of water is higher than that of any organic solvent. In waterreducible coatings, pigments can be dispersed in solvent solutions of the resins before
addition of water, and the surface tension can be reduced with solvents such as butyl
alcohol or a butyl glycol ether. The surfactants in latex paints reduce surface tension,
improving pigment wetting and enabling the coating to wet many types of surfaces.
The presence of surfactants tends to give films having poorer water resistance; various
approaches are being used to reduce or even eliminate the need for free surfactants.
Water increases corrosion of storage tanks, coating lines, ovens, and so on. This
requires that corrosion resistant equipment be used in applying waterborne coatings,
increasing capital cost. For example, mild steel lines may have to be replaced with stainless steel. Since it is electrically conductive, water requires special adaptations of electrostatic spray equipment, which increases cost (Section 23.2.3).
The largest fraction of water-reducible resins is used in OEM product coatings. Waterreducible electrodeposition coatings have such different formulation and application
procedures that they are discussed separately in Chapter 27.
Latex systems are used in a majority of architectural coatings. The general principles of
latex coatings are covered in Section 26.2, detailed discussion is postponed to Chapter 32.
Small but increasing fractions of both OEM product coatings and special-purpose coatings
are latex based.
26.1. WATER-REDUCIBLE COATINGS
Most water-reducible coatings are produced as high-solids solutions in water-miscible
solvent(s). The resins have either carboxylic acid or amine groups that are at least partially
neutralized with low molecular weight amines or acids, respectively. Pigment is usually
dispersed in the partly neutralized resin solution, cross-linker is added along with additives
such as catalysts, and the coating is diluted with water to application viscosity. The resin is
not soluble in water or in the aqueous solution of the solvent that results. Instead, aggregates form with salt groups oriented at the water –particle interfaces and with low polarity
parts of the resin molecules in the interior of the aggregates. The solvent partitions
between the water phase and the aggregates. The aggregates are swollen by solvent as
well as by water that associates with the salt groups and dissolves in the solvent. The crosslinker is dissolved in the resin – solvent aggregates, and the pigment also resides inside the
526
WATERBORNE COATINGS
aggregates. The aggregates exist in a dynamic equilibrium, with solvent moving freely
from one to another and probably with aggregates fusing and separating.
When solutions of these neutralized resins are diluted with water, the change of viscosity with concentration is abnormal. Figures 8.1 to 8.3 show typical relationships
between resin concentration and log viscosity when water-reducible acrylic resin solutions
are diluted with water. As water is added, there is an initial rapid drop in viscosity. As
more water is added, the viscosity plateaus and, frequently, with further addition of
water, the viscosity increases. Still further addition of water leads to a rapid drop in viscosity. The solids of the coatings at application viscosity are usually lower than the
solids of solventborne coatings. While the viscosity of the organic solvent solution of
resin is Newtonian, the viscosity of the water-diluted system in the region of the peak
or plateau is highly shear thinning. When diluted to application viscosity, flow is
usually only slightly shear thinning. As explained in Section 8.3, this behavior is consistent
with the formation of swollen aggregates.
As explained in Section 8.3, the pH of these systems is also abnormal. In the case of
carboxylic acid– functional resins neutralized with a low-molecular weight amine, the
pH is basic even though less than stoichiometric amounts of amine are frequently used.
As application solids are approached on dilution, the viscosity, decreases rapidly as
water is added and is very dependent on the ratio of amine to carboxylic acid. If a
small excess of water is added inadvertently, the viscosity may be too low. However,
this problem can usually be remedied by adding a small further increment of amine,
which increases the viscosity.
Selection of the amine for neutralization is an important formulation consideration
(Section 8.3). Amines are expensive, and they add to VOC emissions. Volatility and
base strength are important. If there is insufficient amine, there will be macrophase separation of the coating. Therefore, it is desirable to select amines that stabilize the aggregate
dispersion efficiently at low concentration. The principal factor controlling the required
amine level appears to be water solubility of the amine. For example, the amount of tripropylamine required is substantially larger (on an equivalent weight basis as well as a weight
basis) than the amount of triethylamine required. Still more efficient are aminoalcohols;
N, N-dimethylaminoethanol (DMAE) is probably the most widely used amine. Somewhat
less efficient but still more effective than trialkylamines are morpholine derivatives.
Commonly only about 20% of the volatiles are organic solvents and the VOC emitted
per unit volume of coating, excluding water, are fairly low. In a water-reducible gloss
acrylic coating, VOC is equivalent to the amount of solvent emitted by a solution
acrylic with approximately 60 NVV. Air quality goals in the future will require still
lower VOC emissions; research is under way.
Most types of resins can be chemically modified to make them water reducible. The
most widely used resins are acrylics with both carboxylic acid and hydroxy functionality
(Section 8.3). Water-reducible polyesters are used, but their hydrolytic stability is limited
and low molecular weight cyclic oligomers can volatilize in baking ovens (Section 10.4).
Since polyesters can be synthesized in the absence of solvent, it is possible to make
solvent-free dispersions [2]. The resin is cast as a solid, is pulverized, and is stored as a
solid until needed, delaying possible hydrolysis. When a coating is to be made, the
powdered resin is stirred into a hot aqueous solution of dimethylaminoethanol to make
a dispersion. Water-reducible alkyds (Section 15.6) are fairly popular in Europe, but
their use is limited because of the difficulty in achieving adequate saponification
resistance for package stability. Water-reducible epoxy esters (Section 15.8) and uralkyds
26.1. WATER-REDUCIBLE COATINGS
527
(Section 15.7) have better hydrolytic stability. Water-reducible urethanes have superior
saponification resistance and a minimum of cyclic oligomers (Section 12.7). Solvent-free
and low solvent water-reducible urethanes have also been reported [3]. Urethane resins
can give films with excellent properties, but are generally more expensive than acrylic
resins.
Coatings made with water-reducible resins have advantages and limitations. An
important advantage follows from the high molecular weight that can be used, as high
as that of resins used in conventional solution thermosetting coatings. This is possible
because the viscosity at application dilution is almost independent of molecular weight.
For example, one can use water-reducible acrylics with an M̄w/M̄n on the order of
35,000/15,000 and an average of around 10 hydroxyl groups and 5 carboxylic acid
groups per molecule along with a class I melamine – formaldehyde (MF) cross-linker
and achieve properties essentially equal to those of a conventional solution acrylic
enamel. The window of cure is comparable to that of the conventional thermoset
acrylic. The problems related to molecular weight and functionality of high-solids coatings are not encountered. A disadvantage is that the application solids are low. Typically,
such coatings are applied by spraying, roll coating, or curtain coating at around 20 to 30
NVV. The low solids means more wet film thickness has to be applied to achieve the same
dry film thickness. On the other hand, in automotive metallic coatings, the low solids is an
advantage because it is associated with better orientation of the aluminum pigment in
the film.
Water-reducible resins, cross-linked with blocked isocyanates, have been reported for
clear coats [4,5]. The combination of blocked isocyanate and MF resin cross-linking gives
coatings having both mar and etch resistance.
The viscosity of water-reducible resin coatings depends strongly on the ratio of water to
solvent. Relative humidity and temperature during application affect the ratio of the
evaporation rates of water and solvent. If the humidity is above a critical level, water evaporates more slowly than even slow evaporating solvents such as 2-butoxyethanol (Section
18.3.6). In extreme cases, the viscosity of the coating during flash off after application can
decrease, rather than increase, as evaporation continues. The result is delayed sagging [6].
Variations in humidity during application and drying can lead to problems. If the RH is
over 70%, the rate of evaporation of water is very slow, and at 100% there is no net evaporation of water. In the mid-RH range, the effect can be reduced by a relatively modest
increase in temperature, since RH decreases with temperature. However, if the humidity
is very high, the only recourses have been to cool the air to condense out some of the
water and then rewarm it, an expensive expedient, or to wait until the RH is lower. For
factory applied coatings, it is generally desirable to formulate for best application at a relatively high RH, say 60%, since it is less expensive to increase RH than to decrease it.
Popping (Section 24.7) during baking of water-reducible coatings is more difficult to
control than with solventborne coatings [7]. The probability of popping increases as
film thickness increases. Since the solids are low and the rate of evaporation of water
during spraying and flash off is low, the wet film thickness required to apply the same
solids is higher than with solventborne coatings. The high heat of vaporization of water
compared to solvents leads to a slower rate of heating of the coating and hence slower
evaporation. Also, until the amine evaporates, there are polar salt groups that tend to
retain water. When the coated article enters an oven, water evaporates fastest from the
surface, so the viscosity of the surface layer increases. Subsequently, when the water
remaining in the lower layers of the film volatilizes, some of the bubbles of water
528
WATERBORNE COATINGS
vapor either cannot break through the surface layer, leaving a bubble, or break through at a
stage when the viscosity of the surface layer is so high that a crater cannot flow out.
Popping is affected by the base strength and volatility of the amine. Popping can be minimized by spraying the coating in more, thinner coats (often called passes) so that there is a
greater chance for evaporation as the film thickness is built up; by having longer flash off
times before entering the baking oven; and by zoning the oven, such that the first part of
the oven is at lower temperature, permitting water to diffuse out of the film before the
viscosity at the surface increases unduly. In some installations, infrared ovens can be
used to drive off most of the water before entering the baking oven. All these alternatives
have relatively high capital or operating costs.
The greatest difficulties occur when spray applying thick coats. One cannot apply completely uniform films by spray. To assure that sufficient coating is applied on all areas,
there will be some areas with substantially more than the average film thickness. Also,
there is a greater chance of air entrapment with spray applied coatings. If the air
bubbles do not break, the air in the bubbles expands when the coating is baked, ultimately
leading to popping. Air entrapment is particularly severe with airless spray, since at the
high pressure, more air dissolves in the coating and then comes out as pressure decreases
on leaving the gun (Section 23.2.6). Siloxanes with a polyether terminal group have been
reported to minimize air entrapment in waterborne coatings [8].
It has been shown that the problem of air entrapment can be reduced by using carbon
dioxide rather than air for spraying [9]. The carbon dioxide largely evaporates during the
time of flight between the spray gun and the substrate. Popping can be minimized by reducing Tg of the resins to the lowest level compatible with the film properties required. Use of
some slow evaporating solvent such as 1-propoxy-2-propanol or the monobutyl ether of
diethylene glycol in the formulation assists in reducing the probability of popping.
Thin coatings applied by roll coating or curtain coating seldom have a popping
problem. The films are more uniform than spray applied films are and there is no air
entrapment problem. With reasonable care about flash off times, curtain coating of relatively thick films can be accomplished without major problems. Water-reducible coatings
are used on a large scale for such applications as can coatings and panel coatings.
One might think that the high surface tension of water would cause serious problems
with crawling and cratering. This is usually not the case, presumably because orientation
of nonpolar segments of solvent to the surface is rapid, reducing surface tension.
With air dry water-reducible alkyd coatings, equivalency of properties with solvent
alkyd coatings has been difficult to achieve. Water-reducible alkyds can be made, as
can modified alkyds such as silicone-modified alkyds, but their hydrolytic instability
leads to limited storage life once the paint has been made and water added. With
careful inventory control, they can be used for some industrial applications, but they
are generally unsatisfactory for trade sales paints, for which a shelf life of two or more
years is needed. Also, until the neutralizing amine required for water dilution evaporates
from the film, the films are water sensitive. The time required to lose amine is related to
both the volatility and the base strength of the amine. Ammonia is the most volatile amine
and is widely used. However, the last of the amine is lost at a stage when volatile loss is
controlled by diffusion rate through free volume holes, not by volatility. The time for
amine loss may be reduced by using a combination of ammonia and a less volatile but relatively weak basic amine such as N-methylmorpholine. Even after the amine has evaporated, the films contain carboxylic acid residues and are sensitive to water and,
particularly, to bases.
26.2. LATEX-BASED COATINGS
529
At one time it was widely assumed that waterborne urethane coatings that have free
isocyanate groups would not be feasible, due to the reactivity of isocyanates with
water. However, two package (2K) urethane coatings with adequate pot life for commercial use have been widely commercialized [10]. Reference [11] reviews 2K urethanes
extensively. A water-dispersible, aliphatic, “hydrophilically-modified” polyisocyanate is
used in one package, and the second package contains a hydroxy-terminated, waterreducible polyurethane with carboxylic acid groups from 2,2-dimethylolpropionic acid
(Section 12.8). A 2 : 1 ratio of 22N55C55O to 22OH is used to offset possible reaction
with water. Films cross-linked within a week at 258C when the RH was 55% or lower.
At high humidity, solvent resistance did not develop at 258C, but warming to 318C permitted cross-linking even at 80% RH. Acrylic polyols can also be used [12]. Pot life problems can be severe since water can diffuse into the dispersed particles of the isocyanate
component, leading to cross-linking, which can preclude good coalescence. Such problems can be minimized by use of appropriate mixing and application equipment [12].
Alternatively, low viscosity polyisocyanates such as blends of HDI dimer and trimer
can be used with aqueous polyurethane dispersions [13]. Very low viscosity HDI iminooxadiazenedione (unsymmetrical trimer) is reported to incorporate easily into 2K waterborne urethane coatings [14]. We discuss waterborne polyurethanes further in Section
12.7. Graft copolymer and semiblock copolymer aqueous dispersions prepared by catalytic chain transfer polymerization have the advantage of narrower molecular weight
distributions compared with water-reducible acrylic resins made by free radical
polymerization [15].
Waterborne primers are particularly important, since reducing VOC levels in solventborne primer formulations is particularly difficult (Section 25.1.3). Waterborne coatings
are especially suited for dip coating applications (Section 22.3) because they eliminate
fire hazards of solventborne dip coatings. For example, water-reducible maleated epoxy
esters (Section 15.8) are used for spray or dip applied primers for steel. All electrodeposition coatings are water-reducible coatings (Chapter 27). Interior linings for beverage cans
are water-reducible coatings, based on grafting acrylates, styrene, and acrylic acid onto
BPA epoxy (Section 13.4).
26.2. LATEX-BASED COATINGS
Latexes have been used for many years in architectural coatings and are the major type of
vehicle for these coatings. For certain household applications, such as flat wall paint, the
advantages of latex paints over any solventborne paint are so large that solventborne paints
are seldom marketed. Important advantages of interior latex paints include rapid drying,
low solvent odor, absence of odor of oxidation by-products of drying oils and alkyds,
easy cleanup, reduced fire hazard, and better long term retention of mechanical properties.
For exterior paints, a major advantage is superior exterior durability of high-performance
latex paints relative to drying oil or alkyd paints. On wood siding, blistering is reduced,
since the latex films are more permeable to water vapor. On the other hand, adhesion of
latex paints to chalky surfaces is inferior to that of solventborne paints (Section 32.1).
A further advantage of acrylic, styrene –acrylic, and styrene – butadiene latexes over
alkyds is their superior resistance to saponification. Paints with shelf lives of many
years are attainable. Latex paints generally show better adhesion to galvanized metal surfaces than do alkyd paints. They also generally show better performance over cement and
530
WATERBORNE COATINGS
concrete surfaces than alkyds, which tend to be hydrolyzed by the alkaline surfaces. Latex
paints also give better coverage over porous cement surfaces (Section 32.1).
The polymers in latex coatings exist as discrete particles and do not move from one
latex particle to another, although solvent and water molecules may do so. Latex coatings
form films by coalescence of the polymer particles (Section 2.3.3). Coalescence can occur
only if the film formation temperature is higher than the Tg of the polymer particles.
Whereas initial coalescence proceeds rapidly at temperatures just a few degrees above
Tg, completion of coalescence is relatively slow unless the temperature is significantly
higher than Tg. For most architectural paints, slow final coalescence is not a real
problem, so the Tg need be only a little below the film formation temperature. In baked
industrial coatings, film formation should be complete by the time the coated article
comes out of the baking oven; therefore, baking temperatures have to be significantly
above the Tg.
There are limitations to latex paints, particularly in regard to the minimum film forming
temperature (MFFT), which allows proper coalescence for film formation. To have a final
film with a high enough Tg to resist blocking, it is common to use coalescing solvents in
the formulation. The coalescing solvent dissolves in the polymer particles, reducing the Tg,
permitting film formation at a lower temperature. After film formation, the coalescing
solvent slowly diffuses out of the film and evaporates. Even with the use of coalescing solvents, however, there are limitations on the temperatures required for good film formation
(Section 32.1).
Latex paints with less organic solvent or, in some cases, no VOC have been developed
and marketed. Blends of latexes of different Tg and particle sizes [16,17] and latexes with
particles having gradient Tg made by sequential polymerization (Sections 2.3.3 and 9.2)
[18] have been used. Thermosetting latexes permit use of a low Tg polymer, facilitating
film formation at relatively low temperatures; subsequent cross-linking gives the
needed block resistance and other properties (Section 9.1.4). Some thermosetting
latexes are used in two-package coatings for industrial use; pot life has to be long
enough to permit coalescence before a significant degree of cross-linking occurs. For
one package latexes, polymers with allylic substitutions that permit autoxidative crosslinking and polymers with trialkoxysilyl groups are being developed. Also, hybrid
alkyd –acrylic latexes have been prepared by dissolving an oxidizing alkyd resin in the
monomers used in the emulsion polymerization.
Polyurethane latexes are usually called aqueous polyurethane dispersions (PUDs) and
are generally cross-linkable (Section 12.8). An important factor promoting their increasing
use is that films of higher Tg polymers can be formed at ambient temperatures than with
acrylic and other conventional latexes, hence, reducing or eliminating the need for coalescing solvents. The water acts as a plasticizer for the polyurethane, reducing the Tg of
the polymer permitting film formation [19].
Latexes, water-reducible acrylic resins, and acrylated water-reducible polyesters can
be blended with PUDs. Acrylic monomers can also be polymerized in PUDs to make
hybrid acrylic–urethane resins. There have been three major approaches to preparing
aqueous acrylic/urethane systems: (1) blending of PUDs and latexes (or other waterreducible acrylic resins) (Section 12.7.2), (2) polymerizing acrylic monomers in PUDs
(Section 12.7.2), and (3) and cross-linking urethane–vinyl systems (Section 12.7.2.1).
In broad terms, acrylate/PUD blends combine the advantages of both acrylic latexes
and PUDs. Acrylic latexes are lower in cost and have excellent exterior durability. The
PUDs give films with higher abrasion resistance. The MFFTs of PUDs are lower than
26.2. LATEX-BASED COATINGS
531
those of acrylic latexes having the same Tg, which permits reduction or even elimination of
coalescing agents and hence reduction of VOC. Hybrid resins give films with higher
tensile strength than acrylic films. Acrylated PUDs with functional groups can be crosslinked by a variety of cross-linkers. Such cross-linked films are of particular interest in
automotive base coats.
Another limitation in the formulation of latex coatings is the difficulty of formulating
high gloss latex paints (Sections 19.10.1 and 32.3). The main problem of formulating gloss
coatings results from the random distribution of pigment and latex particles as the volatiles
evaporate, so there is not the same chance of obtaining a pigment free or low pigment
content upper surface of the film as with solventborne coatings. This problem can be minimized but not eliminated by using a small particle size latex. Gloss is also limited by the
presence of surfactants in the dry film, which can lead to haze and blooming. Latexes with
very low surfactant content or with surfactants that can polymerize are being developed to
minimize or eliminate this problem (Section 9.1.1).
Latex coatings tend to be excessively shear thinning. When the viscosity at high shear
rate is set appropriately for application, the viscosity at the low shear rates encountered in
leveling tends to be high. This is one of several reasons that leveling of latex-based coatings tends to be poorer than that of solventborne coatings. Flat coatings with only fair
leveling have acceptable appearance, but as gloss is increased, the unevenness of the
films becomes undesirable. Substantial progress has been made by use of associative
thickeners (Section 32.3).
Use of latexes has been more limited in OEM coatings, for many reasons. The problems
of evaporation of water on a conveyor line and in ovens mentioned in Section 26.1 are part
of the problem. There can also be a problem of popping with latex coatings. It is desirable
to use a latex polymer with as high a Tg as possible to minimize the chances of coalescence
of the latex particles at the surface of the film before the water has evaporated completely.
Probably, the major limitation for industrial applications has been the flow properties of
latex-based coatings. In addition, uncross-linked latexes lack the robust resistance to
solvents, water, and mechanical damage that is required for many OEM applications.
Leveling requirements for many industrial coatings are more rigorous than for architectural coatings. In general, latex coatings have exhibited relatively high degrees of shear
thinning and, in many cases, thixotropy. In some cases, these flow characteristics result
from flocculation of latex particles. Flocculation of latex increases low shear viscosity
to a major degree; furthermore, it exacerbates the gloss problem. Use of associative thickeners minimizes this problem. Since most of the published work with associative thickeners has been done with latex paints for architectural end uses, discussion of them is
deferred to Section 32.3. Combinations of latexes with water-reducible resins are used
increasingly. Frequently, these combinations have better flow properties than latex
coatings.
An advantage of latexes is that their high molecular weights provide mechanical properties that are fully satisfactory for many applications without need for cross-linking.
Where good solvent resistance is needed, cross-linking is generally necessary. The viscosity of latexes is independent of molecular weight, so they can be applied at relatively
high solids even though their molecular weight is high. In highly pigmented coatings, the
fraction of internal phase volume becomes so large that the solids level has to be reduced;
but the reduction is done primarily with water so that VOC emissions are minimal.
As pressure to reduce VOC emission increases still further, it is to be anticipated that
there will be substantial increases in the use of latexes, particularly thermosetting latexes,
532
WATERBORNE COATINGS
in industrial applications of both OEM product coatings and special purpose coatings. For
example, an emulsion of a 3,5-dimethylpyrazole blocked HDI isocyanurate modified so
that it readily emulsifies in a hydroxy-functional acrylic latex has been suggested as a
binder for clear coats [5].
Another method for cross-linking latexes is to copolymerize a few weight percent of
diacetone acrylamide into the latex and then formulate the coating with adipic dihydrazide
cross-linker; the cross-links are hydrazones formed from the carbonyl group on diacetone
acrylamide and the dihydrazide [20]. Yet another method is to carry out the emulsion
polymerization with hydroxyl-functional monomers in the presence of mixed-ether MF
resins [21]. Both approaches are reported to give films with substantially enhanced hardness and solvent resistance.
26.3. EMULSION COATINGS
Although most waterborne coatings are based on water-reducible or latex binders, emulsion coatings have found some applications and may have potential for wider use.
Two package coatings in which one package is a BPA (or novolac) epoxy resin solution
and the second is an amine-terminated cross-linking agent (Section 13.2.2) which contains
a nonionic surfactant have had significant commercial applications [22 – 24]. The amine
cross-linker package is diluted with water, and the epoxy resin solution package is
added with vigorous stirring. The pot life is limited to a few hours, since the epoxy
resin can react not only with the amine groups, but also, slowly, with water. Although
pot life of solventborne epoxy –amine coatings is limited by viscosity increase, emulsion
systems show little, if any, change in viscosity with age, since the viscosity is controlled by
internal phase concentration, not by molecular weight, and because reaction with water
does not lead to cross-linking. Instead, pot life is limited by reduced gloss of applied coatings or by inferior film properties. Such emulsion epoxy paints are used where hard, easily
cleaned wall coatings are needed, for example, in hospitals and food processing plants.
The residual surfactant reduces the corrosion protective properties for application to
metal surfaces. Emulsion coatings based on aliphatic epoxy resins and aliphatic carboxylic
acid-functional resins are reported to give superior pot life and improved properties [25].
An approach to eliminating surfactants is to use nitroethane as one of the solvents [26,27].
As discussed in Section 13.2.6, nitroethane forms a salt with an amine group of an amineterminated polyamide, which then acts as a surfactant. When the film dries, the nitroethane
evaporates, leaving a less water sensitive film. Such coatings are used in high performance
applications such as aircraft primers (Section 33.4).
Amine salts of acrylic copolymers such as methacrylic acid/methyl methacrylate/ethyl
acrylate/styrene (40 : 20 : 20 : 20) can be used as emulsifying agents for epoxy – phenolic
coatings [28]. During baking, the carboxylic acid groups of the acrylic react with epoxy
groups, incorporating the surfactant into the polymer structure, avoiding the ill effects
of nonreactive surfactants on film properties. Another example is nitrocellulose lacquers
emulsified into water for use as top coats for wood furniture (Section 31.1) [29]. The emulsions have a significantly lower VOC than solventborne lacquers, but longer times are
required to achieve print resistance.
Alkyd emulsion paints have been formulated [30] and are used commercially,
especially in Europe. Higher gloss can be obtained than with latex paints and there is
not the problem of minimum temperature required for film formation versus blocking
REFERENCES
533
resistance experienced with latex paints. However, durability is poorer due to yellowing
and embrittlement of alkyds. Long oil alkyds with anionic surfactants provide the most
stable emulsions. Drying generally is slower than with solventborne alkyds. A contributing factor may be the partitioning of driers between the water and alkyd phases. Alkyds are
emulsified into some latex paints to improve adhesion to chalky surfaces (Section 32.1).
Alkyd emulsions, formed initially as a water-in-oil emulsion, which subsequently
inverts to an oil-in-water emulsion, are reported to have smaller as well as narrower distribution particle size and to be more easily formulated into paint. The alkyd is heated
before addition of water; particle size varies with temperature. In general terms, the
shorter the oil length of the alkyd, the higher the temperature for optimum emulsification.
Superior film properties are obtained using an emulsifier with a polyunsaturated hydrocarbon chain to cross-link with the alkyd after application [31].
GENERAL REFERENCES
Glass, J. E., Ed., Technology for Waterborne Coatings, American Chemical Society, Washington,
DC, 1997.
Karsa, D. R.; Davies, W. D., Eds., Waterborne Coatings and Additives, Royal Society of Chemistry,
Cambridge, 1995.
Pruskowski, S. J., Jr., Ed., Waterborne Coatings Technology, Federation of Societies for Coatings
Technology, Blue Bell, PA, 2005.
REFERENCES
1. Padget, J. C., J. Coat. Technol., 1994, 66(839), 89.
2. Engelhardt, R., Proc. Water-Borne Higher-Solids Powder Coat. Symp., New Orleans, LA, 1996,
p. 408.
3. Rosthauser J. W.; Nachtkamp, K., in Advances in Urethane Science and Technology, Frisch,
K. C.; Klempner, D., Eds., Technomic Publishers, Westport, CT, 1987, Vol. 10, p. 121.
4. Briggs R. L.; et al., U.S. patent 6,099,899 (2000).
5. Rimmer, I.; et al., Eur. Coat. J., 1998, 98, 722.
6. Brandenburger, L. B.; Hill, L.W., J. Coat. Technol., 1979, 51(659), 57.
7. Watson, B. C.; Wicks, Z. W., Jr., J. Coat. Technol., 1983, 55(732), 61.
8. Spiegelhauer, S., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 161 – 170.
9. Gebhard M. S.; Scriven, L. E., J. Coat. Technol., 1994, 66(830), 27.
10. Jacobs, P. B.; Yu, P. C., J. Coat. Technol., 1993, 65(622), 45.
11. Wicks, Z. W., Jr.; et al., Prog. Org. Coat., 2002, 44, 161.
12. Dvorchak, M.; Hunter, J., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 1997, p. 515.
13. O’Connor, J. M.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
1997, p. 458.
14. Richter F.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2003,
pp. 71– 85.
15. Huybrechts, J.; et al., Prog. Org. Coat., 2000, 38, 67.
534
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
WATERBORNE COATINGS
Winnik, M. A.; Feng, J., J. Coat. Technol., 1996, 68(852), 39.
Eckersley, S. A.; Helmer, B. J., J. Coat. Technol., 1997, 69(864), 97.
Hoy, K. L., J. Coat. Technol., 1979, 51(651), 27.
Satguru, R.; et al., J. Coat. Technol., 1994, 66(830), 47.
Mestach, D. E. P.; et al., U.S. patent 6,730,740 (2004).
Huang, Y.; Jones, F. N., Prog. Org. Coat., 1996, 28, 133.
Albers, R., Proc. Water-Borne Higher-Solids Coat. Symp., New Orleans, LA, 1983, p. 130; U.S.
patent 4,352,898 (1982).
Wegman, A., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1997,
p. 92.
Galgoci, E.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1997,
p. 106.
Eslinger, D. R., J. Coat. Technol., 1995, 67(850), 45.
Wegmann, A., J. Coat. Technol., 1993, 65(827), 27.
Lopez, J. A., U.S. patent 4,816,502 (1989).
Kojima, S.; et al., J. Coat. Technol., 1993, 65(818), 25.
Winchester, C. M., J. Coat. Technol., 1991, 63(803), 47.
Hofland, A., in Technology for Waterborne Coatings, Glass, J. E., Ed., American Chemical
Society, Washington, DC, 1997, p. 183.
McNamee, W.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2004, Paper 27.
27
Electrodeposition Coatings
Electrodeposition is an efficient method of applying high performance coatings. Electrodeposition coatings are a major class of waterborne coatings, which are used in largest
volume as primers, but also in single and two coat systems. The general principle is
relatively simple; development and initial commercialization were complex and difficult. Electrodeposition coating systems are anionic or cationic, alternatively anodic or
cathodic, corresponding to deposition at the anode or cathode, respectively. E-coat,
electrocoat, electropaint, ED, and ELPO are fairly common synonyms for electrodeposition coatings. Cationic E-coat is used worldwide for priming auto bodies, and its
adoption in the 1970s and 1980s led to a major improvement in the corrosion resistance
of motor vehicles. Reference [1] provides a review of electrodeposition coating.
In anionic coatings, negatively charged particles of coating in an aqueous dispersion are
electrophoretically attracted to a substrate, which is the anode of an electrochemical cell. The
coating particles are precipitated by hydrogen ions generated there by electrolysis of water.
For cationic coating, the object is made the cathode, and positively charged particles of
coating are attracted to the cathode and precipitated on its surface by hydroxide ions. In
both types, thermosetting binders are used, and coatings are almost always baked. The coatings must be designed so that all coating components are attracted to the electrode at the same
rate; otherwise, the composition will change with time. A vehicle in which pigments can be
dispersed and cross-linkers dissolved must form a stable, electrically charged dispersion of
aggregate particles when diluted with water. The pigment must be preferentially wet by
the resin so that it does not migrate out of the resin aggregates. It is possible to electrodeposit
polymer films from solutions of the salt of a polymer in water, but one cannot use a dissolved
polymer as a binder for pigmented coatings; the resin, pigment, cross-linker, and other
components would not be deposited in the same ratio over time. One can electrodeposit
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
535
536
ELECTRODEPOSITION COATINGS
polymer films from a latex. But latexes cannot be used in pigmented coatings, since pigment
particles and latex particles would be attracted to the substrate at different rates.
The coating is diluted to 10 to 20% solids with water. This relatively low solids is used
for three reasons. When the coated object is brought out of the tank, it carries a layer of the
bath liquid with it that must be rinsed off; losses are less, and rinsing is easier, if the solids
are low. Furthermore, ultrafiltration (Section 27.4) is easier with lower solids. For coating
products such as automobile bodies and appliances, the electrodeposition tanks are very
large—up to 500,000 L. The tank is kept full at all times; therefore, part of the capital
investment is the cost of a tank full of coating—the investment is lower with 10 to 20%
solids than it would be with 50% solids.
A critical requirement for electrodeposition coatings is indefinite stability after dilution.
As the coating is applied, coating solids are removed from the bath, and they have to be
replaced continually to maintain the same composition in the tank. Ideally, the tank would
never be emptied. Some material from the original loading will be in the tank for long
times and must therefore be very stable to hydrolysis and mechanical agitation. The crosslinker must be stable in the diluted coating at a pH over 7 for anionic coatings and under 7
for cationic coatings. Stability to oxidation is critical, since air is mixed into the bath continually by the agitation. If an oxidizing type vehicle is used, an antioxidant that volatilizes
in the baking oven is an essential additive.
27.1. ANIONIC ELECTRODEPOSITION COATINGS
Resins used in anionic systems are substituted with carboxylic acid groups that have an
acid number in the range 50 to 80 mg of KOH per gram of resin. The pigments and
other components are dispersed in the resin, and the carboxylic acid groups are partially
neutralized with an amine such as 2-(N,N-dimethylamino)ethanol. To load the tank, the
coating is diluted to about 10% solids with water. The degree of substitution with salt
groups is designed to be such that the resin is not soluble in water but rather, forms aggregates on dilution. The aggregates are stabilized as a dispersion in water with salt groups on
the outer surface of the particles. Even with less than the theoretical amount of amine to
neutralize the carboxylic acid groups, the pH is above 7, due to entrapment of unneutralized COOH groups in the center of aggregate particles (Section 8.3).
In early work on electrodeposition primers for automobiles, maleated linseed oil (Section
14.3.5) was used as a vehicle. The anhydride moiety was bonded to the linseed oil molecules
by carbon–carbon bonds that cannot be hydrolyzed, and hence the coating showed reasonable stability. However, adhesion to steel was relatively poor, and maleated linseed oil was
soon replaced with maleated epoxy esters (Section 15.8), which had even better hydrolytic
stability and provided superior adhesion to steel. The cross-linking obtained through
the drying oil fatty acid esters in the epoxy ester was supplemented by using some
melamine–formaldehyde (MF) resin as a cross-linker. Mixed methyl ethyl ether Class I
MF resins (Section 11.2) are most appropriate for electrodeposition. Such resins have
sufficient solubility in water to permit easy incorporation but are more soluble in the resin
aggregates than in water, so it deposits in constant ratio as the bath is used over long times.
Carboxylic acid – substituted resins made by reacting moderate molecular weight (MW)
polybutadiene with maleic anhydride have also been used as vehicles for anionic primers.
Since the backbone linkages are all carbon– carbon bonds, there is no problem of hydrolysis in the bath.
27.1. ANIONIC ELECTRODEPOSITION COATINGS
537
Maleated epoxy esters are not appropriate for many top coat applications because their
color stability and chalking resistance are poor. The most widely used resins for anionic electrodeposition top coats are acrylic copolymers made using acrylic (or methacrylic) acid and
2-hydroxyethyl methacrylate as co-monomers. The carboxylic acid groups permit formation
of stable aqueous dispersions with electronegatively charged particles; the hydroxyl groups,
as well as the carboxylic acid groups, serve as sites for cross-linking with MF resin.
The primary reaction occurring at the anode in anionic E-coats is electrolysis of water
to yield hydrogen ions. The hydrogen ions neutralize carboxylate ions on the resin at the
surface of the anode. Neutralization removes the charge that stabilizes the aggregates
against coalescence. When the salt groups are neutralized, the surface is less polar;
there is less swelling of the surface with water, so stabilization is eliminated and the particles coalesce on the metal surface.
Not all the salt groups have to be neutralized for precipitation to occur. When the film is
formed, some salt groups, and hence ammonium ions, are trapped within the film. Side
reactions can also occur at the anode. Iron can dissolve to form ferrous ions that are oxidized to ferric ions. The ferric ions can form insoluble salts with the carboxylic acids on
the resin, leading to reddish-brown discoloration. Discoloration is not a problem for
primers but is a limitation for making light colored or white electrodeposition top coats.
Zinc phosphate conversion coatings generally reduce discoloration to acceptable levels.
Anionic top coats can be used on aluminum, since aluminum salts do not affect the
color of a coating. In fact, the protective oxide layer on aluminum is enhanced when it
is the anode, so anionic E-coat is preferable to cationic for coating aluminum.
An important side reaction in electrocoating of phosphate conversion– treated steel is
partial dissolution of the iron– zinc phosphate layer by hydrogen ions generated at the
anode surface:
This partial removal of conversion coating has two potentially serious effects. Damage
to the phosphate coating can lead to poorer adhesion to the steel surface and less corrosion
protection. Also, the soluble ion concentration in the bath is increased, which leads to
higher conductivity of the water phase. As discussed later, maintaining constant relatively
low conductivity of the water phase is critical.
Electrolysis of water at the anode also generates oxygen. The oxygen can be generated
at the metal surface after it has been coated, leading to film rupture as the bubbles of
oxygen escape through the film. At the cathode, hydrogen gas and hydroxide ions are generated. The hydroxide ions neutralize the ammonium counterions:
The amines formed are water soluble, and only a fraction of the amine is removed when
the coated article is taken out of the bath. This leads to accumulation of amine in the bath.
Methods of control of amine concentration are discussed in Section 27.4. An advantage of
electrodeposition compared to other waterborne systems is that only a small amount of
538
ELECTRODEPOSITION COATINGS
counterion is deposited on the substrate with the rest of the coating. Some anionic electrodeposition tanks have been run using potassium hydroxide as the neutralizing base.
27.2. CATIONIC ELECTRODEPOSITION COATINGS
Cationic coatings have positively charged aggregates that are attracted electrophoretically to
the cathode. The resins have amine groups neutralized by a low MW, water soluble acid
such as formic, acetic, or lactic acid. Coatings stable at a pH a little below 7 are preferred;
otherwise stainless steel or other expensive corrosion resistant piping and handling facilities
should be used. Some suppliers recommend corrosion resistant equipment in all cases. Commercial cationic electrodeposition tanks are operated within a narrow pH range, 5.8 to 6.2.
The resins used in the formulation of cationic E-coat automotive primers are based on
BPA epoxy resins.
.
.
.
.
.
The epoxy resin is reacted with an aminoalcohol, such as diethanolamine (DEA) to
yield a resin with amine and hydroxyl groups.
The resin product is reacted with an isocyanate half-blocked with an alcohol (e.g.,
2-ethylhexyl alcohol (2-EHOH) or 2-butoxyethanol) to make a self cross-linkable resin.
Catalyst and pigment dispersions are added.
The amine groups are neutralized with a volatile carboxylic acid such as acetic or
formic acid.
The coating is dispersed in water.
The blocked isocyanate cross-linking agent is stable in the slightly acidic water system,
whereas MF resins are not. During baking, the blocked isocyanate reacts with a hydroxyl
group to form a urethane cross-link. Older E-coat primers contained basic lead silicate as a
catalyst, which interacted with the phosphate coating to enhance adhesion. Lead-free
E-coats have replaced the older formulations.
27.2. CATIONIC ELECTRODEPOSITION COATINGS
539
The vast majority of cationic E-coats are cross-linked with blocked isocyanates.
Upwards of 400 references, mostly patents, between 1980 and 1999 are given in a
review of blocked isocyanates [2]. Many variations of epoxy derivatives are described
in the literature.
2,4-TDI has the advantage of a large difference in reactivity between the para and
ortho isocyanate groups allowing for the formation of predominantly monoblocked
intermediate However, some deblocking of both isocyanate groups occurs during cure,
releasing some free TDI, which can migrate into base coats causing discoloration.
Non-yellowing primers can be made with aromatic isocyanates by reacting TDI halfblocked by the butoxyethoxyethoxy alcohol with BPA epoxy to make an oxazolidone,
which is then used with a conventional modified epoxy resin [3]. In some, delamination
occurs on exterior exposure between the E-coat and top coats; incorporation of HALS
and a UV absorber in a 2-butoxyethoxyethyl alcohol-blocked TDI primer has been
reported to overcome this problem [4].
MDI and polymeric MDI have the advantage over TDI in that they are less mobile than
TDI, thus less likely to diffuse into base coats. E-coat dispersions are reported to be more
stable using a polymeric MDI (12% 2,40 , 35% 4,40 , and 53% poly) blocked with methyl
alcohol (or other alcohols) [5]. Partially reacting the MDI with polyether polyols before
blocking gives coatings with greater flexibility. A cationic polyurethane dispersion
(PUD), prepared from polymeric MDI partially blocked with 2-butoxyethoxyethyl
alcohol, TMP, and N,N-dimethylaminopropylamine is neutralized with formic acid and
dispersed in water to give a solvent-free cross-linker for E-coats [6].
In recent years primary focus has been on reducing the cure temperature by using
other blocking agents and/or catalysts while maintaining the long term hydrolytic stability of the diluted coating. Since 2-EHOH was the standard for many years, it is used as
the base for comparison. Based on patent citations, it appears that 2-butoxyethanol is the
most widely used blocking agent today. Such glycol ether-blocked isocyanates crosslink at lower temperatures than those blocked with non-ether alkyl alcohols, since the
reverse reaction of the glycol ethers to re-form the blocked isocyanates is slower. In
addition, owing to lower volatility, glycol ethers remain in the film longer, promoting
better flow and leveling. Glycol amides such as N,N-dibutylglycolamide have been
disclosed as a blocking agent permitting lower temperature cure in E-coats than with
2-EHOH [7].
Methylethylketone oxime (MEKO)-blocked isocyanates are reported to give primers
that cure in 20 minutes at 1208C [8]. Such oxime systems were widely used in Japan;
but it has been shown that in the E-coat process, HCN is generated leading to dangerous
waste discharges [9].
Considerable effort has been devoted to development of new catalysts for E-coats. There
have been three driving forces: preparation of lead-free primers, reduction of curing temperatures, and increased hydrolytic stability. Dibutyltin oxide has been the most commonly
used catalyst; it gives good stability, but it is insoluble and must be dispersed with the other
pigments; it is not highly efficient, and it generally requires a lead compound as a cocatalyst. Trialkyltin compounds such as bis(trioctyltin) oxide (TOTO) are claimed to be
effective catalysts especially when used with ZnO [10]. A patent discusses various possible
catalysts and particularly recommends TOTO, which permits curing at 1508C [11]. It is
claimed that with TOTO one can obtain thicker coatings, greater throw power, and better
edge coverage. TOTO is low in toxicity and is a liquid that is easily incorporated into the
coating, and can be used to make clear coatings.
540
ELECTRODEPOSITION COATINGS
Bismuth salts have been reported to be effective catalysts. Bismuth salts of lactic acid and
dimethylolpropionic acid (DMPA) offer the advantage that the phosphate conversion coating
does not need a chromate wash when the salts are used as catalysts for the E-coat primer [12].
The effectiveness of a series of catalysts on the cure of a cationic E-coat in which a glycol
ether-blocked polymeric MDI was the cross-linker concluded that bismuth tris(isostearate),
plus isostearic acid, gave the best cure response as well as superior resistance to hydrolysis
[13]. See Ref. [2] for an extensive review of catalysts in cationic E-coats.
While aromatic isocyanates have been used in most cationic E-coat primers, problems
arise if UV radiation reaches the primer surface as a result of inadequate coverage by the
base coat. Photodegradation of the primer can lead to discoloration and possibly to delamination after outdoor exposure. Therefore, primers have been developed using aliphatic
isocyanates that are more stable to photodegration. For example, caprolactam blocked
IPDI/TMP prepolymer can be used as the cross-linker [14].
In cationic E-coating there is not the problem of acid dissolving the phosphate
conversion coating that exists in the case of anionic deposition. Depending on zinc
concentration in the treatment bath, different crystals can be deposited (Section 6.4.1).
At relatively high zinc concentrations, the crystals are predominantly hydrated zinc
phosphate, Zn3(PO4)2 . 4H2O, called hopeite. Under zinc-starved conditions, the crystals
have been identified as phosphophyllite, Zn2Fe(PO4)2 . 4H2O [16]. For E-coating steel the
preferred treatment is phosphophyllite, not hopeite. On zinc-coated steels for E-coats, Zn–
Ni, Sn– Mn, and Zn– Mn –Ni phosphate conversion treatments are used since straight zinc
phosphate would give only hopeite [17].
A recent innovation has been the introduction of two-coat electrodeposition combinations for priming of automobiles. The first E-coat is electroconductive, which allows
a second E-coat to be applied. The second E-coat is formulated to replace a spray-applied
coating of a primer-surfacer, reducing application cost. A conductive E-coat primer, formulated with 2-EHOH-blocked TDI, an epoxy –amine resin, and electroconductive black
pigment, has been disclosed [18].
E-coats based on BPA epoxies and TDI or MDI are used only as primers. If used as top
coats, they would exhibit poor color retention and exterior durability. For top coats, acrylic
resins with blocked aliphatic diisocyanates give better color retention and exterior durability. 2-(N,N-dimethylamino)ethyl methacrylate and hydroxyethyl methacrylate as
co-monomers provide the amine groups for salt formation and the hydroxyl groups for
cross-linking. The curing temperature required for alcohol-blocked aliphatic isocyanates
to react with hydroxyl groups is high. One can use oxime-blocked isocyanates that crosslink at lower temperatures, but then the bath stability may be limited. Alternatively, one
can make an acrylic resin using glycidyl methacrylate as a co-monomer; the pendant
epoxy groups can be reacted with amines to yield secondary amines or mixed secondary
and tertiary amines. Alcohol-blocked aliphatic isocyanates can be used to cross-link the
secondary amine groups; high temperature bakes are required, but chemical stability in
the tank is excellent.
Cationic E-coats have replaced anionic types as automotive primers owing primarily to
superior corrosion protection. Amine-substituted cationic resin binders provide greater
corrosion protection for steel, resulting, perhaps, from strong interaction between the
amine groups and the substrate surface, thereby enhancing wet adhesion, the most critical
factor in corrosion protection (Section 7.3.2). Wet adhesion of lead-free coatings is
especially critical for corrosion protection, since lead-containing compounds, which inhibited corrosion in earlier E-coats, can no longer be used.
27.3. EFFECT OF VARIABLES ON ELECTRODEPOSITION
541
27.3. EFFECT OF VARIABLES ON ELECTRODEPOSITION
Deposition of coating is not instantaneous with application of current. Some time elapses
before enough hydrogen ions (anionic E-coat) or hydroxide ions (cationic E-coat) are
formed to neutralize enough charges on aggregate particles to cause precipitation. After
this initial time interval, the rate of deposition is affected by the rate of electrophoresis
of the aggregates. This rate is affected significantly by the impressed voltage; the
higher the voltage, the faster the film deposition. Coatings are designed to coat in 2 to
3 minutes with 225 to 400 V; the high voltage is not needed to electrolyze the water,
but rather, to increase the driving force for electrophoretic attraction of particles to the
electrode and for adequate coating of recessed areas.
The first areas covered are the edges of the metal, since current density is highest there.
As film thickness increases, electrical resistance increases, reducing the rate of deposition
at those sections coated first. There is a limiting film thickness beyond which deposition of
further coating stops or, at least, becomes very slow. After edges are coated, outer flat surfaces of the object are coated, followed by recessed and enclosed areas. The further back in
a recess, the later the area is coated. Particularly for corrosion protective primers, it is
desirable to have the entire surface of the steel coated, so it becomes important to coat
the furthest recessed areas in the 2 to 3 minute dwell time in the tank.
The rate of deposition is also affected by the equivalent weight of the coating. The
higher the equivalent weight, the greater the amount of coating precipitated by each
hydrogen (or hydroxide) ion and therefore the faster the buildup of film thickness. On
the other hand, it is critical to have a low enough equivalent weight so that there are sufficient polar salt groups to maintain stability of the dispersion of aggregates in the coating
bath. The rate is also affected by the amount of soluble low MW ions present in the bath.
These ions are attracted to the electrodes in competition with the aggregate particles. Since
they are small, they move more rapidly in the electrophoretic field. Thus, concentration of
soluble salts must be low and maintained close to constant.
A coating that deposits in recessed areas quickly is said to have a high throw power
(called throwing power by some authors). Standard tests for throw power determine
how far up into a pipe or an open-ended box the coating is applied at a standard
voltage in a standard time. Throw power increases with higher impressed voltage and
longer dwell times. However, if the voltage is increased to too high a level, there will
be film rupture of the coating applied to the outer surfaces. At a sufficiently high
voltage, the current breaks through the film, leading to local generation of gas under the
film (oxygen in anionic electrodeposition and hydrogen in cationic electrodeposition),
and bubbles of gas can blow out through the film, leaving film defects. It has been
shown that, at higher voltages, electric discharges occur through the film during electrodeposition causing visible sparks [19]. These sparks may also be responsible for film ruptures. Sparking is reported to occur at lower voltages (240 V) when the substrate is
galvanized steel than when the substrate is steel, about 300 V or higher. As conductivity
of the deposited film increases, film rupture tends to occur at lower voltages. Direct current
electricity obtained by rectifying AC electricity has a relatively wide variation in voltage;
the current is said to ripple. The effect of rippling can be to break through at lower
(average) voltage [20]. Model equations have been developed to permit predicting
throw power and film build on complex shaped surfaces such as an automobile [21].
Throw power is affected by the conductivity of the bath; the higher the conductivity, the
greater the throw power. However, there is a limitation; as the conductivity due to the
542
ELECTRODEPOSITION COATINGS
presence of soluble salts increases, the rate of electrophoresis of coating aggregates
decreases. Increased numbers of salt groups on the resin increase conductivity and
hence throw power, but equivalent weight drops, which decreases the rate of deposition.
Entrapment of conductive material in the film increases with a corresponding increase in
the likelihood of film rupture. A compromise on conductivity must be reached. Conductivities used are in the range 1000 to 1800 microsiemens (mS); the older unit, mho, is
still used; 1 mS ¼ 1 mmho.
Film rupture and hence throw power are affected by variations in coating composition.
If the viscosity of the aggregates precipitated on the surface is high, full coalescence is not
achieved and a porous film results, giving high film conductivity and poor throw power.
On the other hand, if the viscosity of the precipitated aggregates is very low, the films
will be soft; then, if any electrolysis takes place below the surface, bubbles will break
through easily and film rupture will be severe. A compromise must be made between
the extremes. The glass transition temperature of the resin is a controlling factor, and correspondingly, the temperature at which the electrodeposition is carried out is important.
Temperature of the bath must be controlled within a fairly narrow range, typically 32 to
358C. Higher Tg resins can be used only if solvent is included in the formula. Many
E-coats now use a small amount of solvent, but eliminating it is a desirable environmental
goal. Solvents can affect the electrical conductivity of the deposited film, so care must be
taken in their selection. Excess solvent leads to film rupture at lower voltage, which
reduces throw power. Also, the partition coefficient must be such that almost all the
solvent is dissolved in the aggregates, with very little in the water, or the solvent concentration in the bath will build up over time.
Pigment concentration also affects coalescence. If the PVC is near or above CPVC and the
amount of solvent, if any, is small, the deposited film will not coalesce. Even if low enough to
permit coalescence, the amount of pigment has a significant effect on leveling of the film.
Since the film as applied has a low solvent concentration, its viscosity is very dependent
on pigment concentration. Unless the PVC is low, in most coatings less than half of
CPVC, leveling will be poor because of the high viscosity. Because of the low level of pigmentation compared to that of conventional primers, the gloss of electrodeposited primers is
higher, especially if the pigment content is reduced to low enough levels to permit good leveling. The PVC also affects flow away from edges during baking; higher PVC reduces this
problem, so a compromise between leveling and edge flow must be made.
27.4. APPLICATION OF ELECTRODEPOSITION COATINGS
A schematic diagram of an E-coat system is shown in Figure 27.1. The object to be coated
is hung from a conveyer and carried into a dip tank. Both to return coating carried out of
the tank by the object and to avoid local accumulation of excess coating on the object, the
object is rinsed with ultrafiltrate as it comes out of the tank. At the rinsing stage, the
coating film has not been cross-linked but is of sufficiently high viscosity to remain
intact; only the bath liquid is washed from the surface of the coating. Concentration of
coating in the bath must be kept constant by adding makeup coating continuously to
replace coating solids removed by deposition. The heat exchanger maintains temperature
within a narrow range.
The bath liquid is recirculated continuously through an ultrafiltration unit. Adoption of
ultrafiltration was an essential factor in allowing large-scale cationic electrodeposition.
27.4. APPLICATION OF ELECTRODEPOSITION COATINGS
543
Figure 27.1. Cationic automotive primer electrodeposition installation. (Adapted from Ref. [22], with
permission.)
The ultrafiltration membrane permits removal of excess water and water-soluble materials
while not removing the aggregates containing the resin, pigment, and cross-linker. Ultrafiltration permits maintaining the concentration of soluble salts essentially constant so that
there is constant conductivity.
Figure 27.1 also shows an electrolyte tank. This represents a system for controlling
the concentration of solubilizing agent (acid for cationic coatings and amine for
anionic ones). When coating is precipitated on the substrate surface, a corresponding
amount of solubilizing agent is released. The concentration of solubilizing agent must
be maintained at a constant level. A small amount of solubilizing agent is carried out
of the bath with the coated product, and some is removed by ultrafiltration, but these
losses of solubilizing agent are less than the amount being released by electrocoating.
There are two ways of maintaining the balance. A sufficiently low level of solubilizing
agent content is used in the makeup coating, and the remainder of the required level of
solubilizing agent is the excess left in the bath. The other, more effective approach is to
have the counter electrode in a microporous polypropylene box. The pore size of the
membrane must be such that the aggregate particles cannot pass through the membrane,
but water and carboxylate (or ammonium) ions can easily go through. In some cases it is
necessary to use ion-selective membranes. The clear liquid in the box is recirculated to
the electrolyte tank, where the concentration is monitored and corrected automatically.
Maintenance of the proper level of solubilizing agent is critical. It is common to hear
that the pH of the bath must be kept constant. Although this is true, the pH of these
coatings is insensitive to the ratio of the weak acid and weak base (Section 8.3).
Conductivity is a more important control criterion. The application lines can be
highly automated, with feedback control of rate of addition of make-up coating, solubilizing agent, and water.
544
ELECTRODEPOSITION COATINGS
27.5. ADVANTAGES AND LIMITATIONS OF ELECTRODEPOSITION
Electrodeposition is used for applying coatings to a variety of products. Almost all new
automobiles have electrodeposited primer coats. Many appliances are primed by electrodeposition. Aluminum extrusions, drapery fixtures, metal toy trucks, and steel furniture are
a few of many examples of single coat electrodeposition applications.
Electrodeposition can be a highly automated system with low labor requirements,
especially as compared to spray application. A startling example of a savings of labor
has been reported in the use of cationic electrodeposition to apply an epoxy coating to
air conditioners [23]. The former coating system of a flow coated primer and a spray
applied acrylic top coat required 50 people, including those who did the required
touchup and repair. The E-coat operation requires only one operator. Further savings
result from elimination of coating losses from overspray. Coating utilization in excess of
95% is reported. The economic advantage of the combination of these factors is large in
assembly line operations. However, the capital cost of the automated line is high, limiting
applicability of highly automated lines to large production operations. The electrodeposition unit is the most expensive equipment in an auto assembly plant. Simpler installations
are used for applications such as coating metal toys. Development of electrodeposition in
the 1960s and 1970s was a long and expensive process, and even today it is expensive to
start a new system and get it running smoothly.
Solvent content of E-coats is relatively low, so VOC emissions are low and fire hazard
is reduced. Another environmental advantage over spray applied coatings is that there is
no overspray sludge disposal (unless a poor job of formulating, coating production, bath
maintenance, or control makes it necessary to dispose of 500,000 L of bad coating!). Since
the solids of the deposited film are high, only 3 to 5 minute flash off time is required before
entering the oven, an advantage relative to spray applied coatings.
Another advantage, assuming adequate throw power, is that complete coverage of surfaces is obtained. There can be differences in film thickness; for example, the recessed
areas will generally have thinner deposits than the exposed face areas, but all of the
surface will have some coating. Recessed and enclosed areas that cannot be coated by
spray application can be coated by electrodeposition. Objects with many edges, such as
drapery fixtures, can be coated better by electrodeposition than any other way. The
edges are uniformly coated, in contrast to excess coating at the bottom edges of objects
coated by dipping.
Sagging can be experienced when the electrocoated part is heated in an oven. However,
due to the relatively high viscosity of the coating immediately on application, severe
sagging is less likely than during spray or conventional dip application methods. Also,
one is less likely to encounter the relatively large differences in film thickness between
the top and bottom of the dipped object commonly experienced with conventional dip
coating.
Uniform film thickness can lead to a problem, especially with relatively highly pigmented primers: the applied coating follows the surface contours of the metal closely, so a
rough metal gives a rough primer surface. Some authors refer to this phenomenon of replication of the substrate surface profile as telegraphing. If there is minimal replication of
metal scratches or a rough metal profile in the surface of the primer coat, some authors
say the primer shows good filling. E-coat films are relatively thin, varying from 15 to
30 mm, depending on coating composition and application variables. When E-coat
primers are sanded to smooth the surface and improve adhesion, only light scuff
27.6. AUTODEPOSITION COATINGS
545
sanding is done. If necessary, heavier spot sanding will smooth out sags, runout, seeds, and
so on, and remove dirt, weld balls, and weld spatter. However, spot sanding will often
result in sand-through to expose metal. If so, the area must be patched with a spot
primer, a catalyzed spray primer that contains corrosion inhibitors.
Substantial variations in smoothness of the primer surface can result from changes in
E-coat primer composition [24]. Leveling can be improved by reducing the pigment
content of the primer. With lower PVC, the viscosity of the coating after application
and before cross-linking is lower. During baking, some leveling occurs before crosslinking increases the viscosity such that no further flow can occur. The lower viscosity
can also permit undesirable flow; in some cases, edge coverage is reduced by flow
away from the edges of the coated product. The edges heat up first, reducing the
surface tension of the coating in that area, leading to flow over the adjoining higher
surface tension area. Also, when the pigment content is decreased, gloss of the films is
increased.
There can be a problem achieving adequate adhesion of top coats to the surface of electrodeposited primers. Adhesion is affected by the cross-link density of the E-coat; overbaking can lead to poorer intercoat adhesion. The gloss of electrodeposited primers is
relatively high, and the smooth surface makes achieving intercoat adhesion more difficult.
When the PVC is further reduced to promote leveling, the surface is still smoother to the
further detriment of intercoat adhesion. For auto bodies, a primer– surfacer is usually
applied over the E-coat primer. If the solvents are selected carefully, penetration into
the electrodeposited film promotes intercoat adhesion. The primer –surfacer can have
PVC . CPVC to enhance adhesion of the top coat, and it can be sanded to smooth the
surface without reducing adhesion. The primer– surfacer also improves adhesion of
sealants and adhesives and protects the E-coat from UV degradation.
The substrate must be conductive; most applications have been for metal primers or for
one-coat metal coatings. Two coat applications require that the first coat be conductive and
are used in production in only a few instances.
Another limitation that electrodeposition shares with any other dipping system is the
difficulty of formulation changeovers. If it is decided to change the color of an automotive
primer, what do you do with 500,000 L of coating? At least once when this problem was
faced, the coating supplier developed the new color primer to be compatible with the old
color primer, so that the new coating could be introduced into the existing tank. Of course,
for a period of time there was a slow change of primer color from the old color to the new
color. This was not serious for a primer, but would not be acceptable for a top coat. Use of
electrodeposition for top coats is limited to applications for which long runs of the same
color are made (such as agricultural equipment, garden tractors, and lawn mowers) so that
a line can be dedicated to a particular color. In small installations, such as for coating
toys, one coating bath can be pumped out of the electrodeposition tank into a holding
tank, where it must be kept continuously agitated, and another color is pumped into the
tank. But these are relatively small tanks; the procedure would not be economically
feasible with large tanks.
27.6. AUTODEPOSITION COATINGS
Autodeposition coating (also called autophoretic coating) is, as the name implies, a process
in which coatings are applied without external electricity, in contrast to anionic and
546
ELECTRODEPOSITION COATINGS
cationic electrodeposition. A paper has been published comparing the properties of steel
coated by autophoresis with a cationic E-coat [25]. The coating is a stabilized poly(vinylidene chloride) latex bath containing a weak acid (hydrofluoric acid), an oxidant (hydrogen peroxide), deionized water, an accelerator (iron salt), surfactants, and additives. When
a steel object to be painted is immersed in the coating bath, two reactions take place. The
principal reaction is
ð27:1Þ
A minor reaction is
ð27:2Þ
The ferrous ions precipitate the latex and pigment dispersion in the coating as an iron
complex. The ferrous ions in the complex are oxidized to ferric ions by the hydrogen
peroxide and hydrofluoric acid:
ð27:3Þ
The FeF3 is then available for a chain reaction via reaction (27.1).
Initially, the deposited coating is adherent and porous, allowing the acid to continue to
penetrate to the substrate as the reaction continues. The first areas of the steel coated are
the anodic areas of the steel, such as edges. As the initial anodic areas are coated, other
areas of the steel become anodic and coating continues. After application, the coating is
baked at 1058C to remove water and become cohesive.
The films from the autophoretic coating are rougher than those of cationic coating.
Physical properties of the films are similar. Corrosion protection by intact autophoretic
films is excellent, but when gouged through, the protection is inferior to the cationic
primer. The resistance of the autophoretic coatings to brake fluid is inferior. Thus,
cationic E-coat is preferable for coating car bodies, and autophoretic coating uses are
limited to applications where the more limited corrosion protection is adequate: for
example, seat structures, fans, head light casings, and the like. Only black coatings have
been available.
Where the film properties are adequate, autophoresis has several advantages over
E-coat: Bake temperatures are lower. Only thorough cleaning of the steel is required,
not phosphate coating. Capital cost of application tanks and equipment is lower. Any
active metal can be coated. The coatings have zero VOC.
In addition to poly(vinylidene chloride), a variety of other binders have been patented:
vinylidene chloride, acrylic acid, methyl acrylate copolymer latex [26], an acrylic latex
[27], a BPA epoxy resin with an IPDI uretdione prepolymer as a blocked cross-linking
agent, and a resole phenolic resin with an acrylonitrile –butadiene latex [28]. Flexibilized
BPA epoxy resins with caprolactam-blocked HDI isocyanurate are reported to give coatings with excellent adhesion and corrosion protection without the need for a phosphate
conversion coating [29].
REFERENCES
547
GENERAL REFERENCE
Dini, J. W., Electrodeposition: The Materials Science of Coatings and Substrates, Noyes
Publications, Norwich, NY, 1993.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Krylova, L., Prog. Org. Coat., 2001, 42, 119.
Wicks, D. A.; Wicks, Z. W., Jr., Prog. Org. Coat., 2001, 41, 1.
Kaufman, M. L., U.S. patent 5,272,189 (1993).
Zwack, R. R.; Eswarakrishnan, V., U.S. patent 5,389,219 (1995).
McCollum, G. J.; Mauer, G. W., Eur. patent EP 236,050B1 (1990).
Schafheutle, M. A.; et al., U.S. patent 5,977,247 (1999).
Moriarity, T. C., U.S. patent 4,452,963 (1984).
Yasuoka, Y.; et al., Jpn. Kokai Tokkyo Koho JP 02,067,382 (1990).
Nonomura, M.; et al., Toxicol. Environ. Chem., 1993, 39, 65. Modler, H.; Nonomura, M.,
Toxicol. Environ. Chem., 1995, 48, 155.
Bossert, E. C.; et al., U.S. patent 5,880,178 (1999).
Bossert, E. C.; et al., U.S. patent 6,042,893 (2000).
Kerlin, K. G.; Hamacher, P., U.S. patent 5,702,581 (1997).
He, Z. A.; et al., PCT int. patent appl. WO/0047642 (2000).
Debroy, T. K.; Chung, D.-Y., U.S. patent 4,755,418 (1988).
Dyett, M. J., J. Oil Colour Chem. Assoc., 1989, 72, 132.
Schoff, C. K., J. Coat. Technol., 1990, 62(789), 115.
Batzill, W.; et al., U.S. patent 4,882,090 (1989).
Smith, R. E.; Boyd, D. W., J. Coat. Technol., 1988, 60(756), 77.
Vincent, J. J., J. Coat. Technol., 1990, 62(785), 51.
Boyd, D. W.; Zwack, R. R., Prog. Org. Coat., 1996, 27, 25.
Wismer, M.; et al., J. Coat. Technol., 1982, 64(688), 35.
Miranda, T. J., J. Coat. Technol., 1988, 60(760), 47.
Gilbert, J. A., J. Coat. Technol., 1990, 62(782), 29.
Almeida, E.; et al., Prog. Org. Coat., 2003, 46, 8.
Hall, W. E., U.S. patent 6,312,820 (2001).
Roberto, O. E.; Maxim, M. A., U.S. patent 5,486,414 (1996).
Kucera, H. W.; Cowles, R. S., U.S. patent 6,476,119 (2002).
Bammel, B. D.; Fristad, W. E., Proc. Waterborne High-Solids Powder Coat. Symp.,
New Orleans, LA, 2003, pp. 491 – 500.
28
Powder Coatings
Worldwide production of powder coatings in 2001 was estimated to be about 900,000
metric tons [1]. Powder coatings are used most extensively in Europe (43%) and Asia
(26%); followed by North America (23%) and the rest of the world (8%). On a percentage
basis, use of powder coatings grew rapidly from the 1970s through 2000; the growth rate
slowed in the early 2000s [2].
The general principle is to formulate a coating from solid resins, pigments, and additives, to melt mix them under high shear to disperse pigments in a matrix of the major
binder components, and to pulverize the resulting solid. The powder is applied to the
substrate, usually metal, and fused to a continuous film by baking. Thermosetting and
thermoplastic powders are available; the major portion of the market (.90%) is for
thermosetting types.
28.1. BINDERS FOR THERMOSETTING POWDER COATINGS
Binders for thermosetting powder coatings consist of a mixture of a primary resin and a
cross-linker, often called a hardener. The major types of binders are grouped somewhat
arbitrarily into several classes, as shown in Table 28.1. The terminology has grown historically and has become confusing. Epoxy powder coatings include only coatings based on
BPA and novolac epoxy resins with amine, anhydride, or phenolic hardeners. Hybrid coatings also contain BPA epoxy resins but are cross-linked with carboxy-functional polyester
resins. Polyester coatings contain polyesters with various cross-linkers other than BPA and
novolac epoxies; the term is used only for coatings that exhibit good to excellent exterior
durability. Acrylic coatings contain acrylic resins with various cross-linkers. In addition,
various blends, sometimes called alloys, of these classes are used. Blends are gaining
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
548
28.1. BINDERS FOR THERMOSETTING POWDER COATINGS
549
TABLE 28.1. Classes of Thermosetting Powder Coatings
Common Name
Epoxy BPA (or novolac) epoxy
Hybrid COOH-functional polyester
Polyester
COOH-functional polyester
OH-functional polyester
Acrylic
Epoxy-functional acrylic
OH-functional acrylic
UV cure
Acrylate-functional resin
Epoxy-functional resin
Primary Resin Cross-Linker
Polyamines, anhydrides, or phenolics
BPA epoxy
Triglycidylisocyanurate or hydroxyalkylamide
Blocked-isocyanate or amino resin
Dibasic acid
Blocked-isocyanate or amino resin
Free radical
Cationic
increasing importance, for example, in primers for auto bodies [3]. Among the factors in
choosing a class of powder coating for an application are protective properties, exterior
durability, and cost. Differences in exterior durability are indicated in Figure 28.1.
In powder coatings, it is necessary to control the balance of binder Tg, M̄n, f̄n, and reactivity [5]. It must be possible to process the material without significant cross-linking, and
the resulting powder must not sinter (start coalescing) or start to cross-link during storage,
but it must fuse during baking, level to form a desirable film, and cross-link. In general, the
primary resins are amorphous polymers with a Tg high enough to avoid sintering of the
powder and with an M̄n of a few thousand. Recommended minimum binder Tg values
are 408C in Europe and 45 to 508C in North America, reflecting the higher temperatures
during shipment and storage in parts of North America. A typical powder coating having a
Tg of 508C could be melt processed at about 808C and could be handled and stored at temperatures up to about 408C. When heated in a baking oven, its viscosity would drop briefly
Figure 28.1. Florida outdoor exposure data on various types of powder coatings. (From Ref. [4], with
permission.)
550
POWDER COATINGS
to about 10 Pa.s, allowing coalescence, flow, and leveling, and with continued heating for
15 minutes at a temperature in the range 130 to 2008C, it would cross-link. The Tg values
referred to are those of the combination of the primary resin plus the cross-linker. The Tg
required for the primary resin alone varies, depending on the cross-linker used with it. UV
cure powder coatings can be cured at temperatures as low as 1008C.
Because Tg is higher, powder coatings can attain good hardness at somewhat lower
cross-link density than most liquid coatings designed for similar applications. This attribute tends to work in the favor of powder coatings, which often combine excellent hardness with excellent impact resistance.
28.1.1. Epoxy Binders
Epoxy powder coatings are the oldest and still one of the largest classes of thermosetting
powder coatings. Decorative coatings have been based on BPA epoxy resins with n values
of 3 to 5, usually made by the advancement process (Section 13.1.1). The trend, however,
is to lower molecular weights (MW) with some epoxy resins having n as low as 2.5, to
provide better flow in thin film applications. For protective coatings, n values range up to
8. The most commonly used cross-linkers are dicyandiamide (dicy) (Section 13.2.1) or a
modified dicyandiamide. The curing reactions are complex [6]. Modified dicys are more
soluble in epoxy resins and tend to form uniform films more readily. 2-Methylimidazole is
a widely used catalyst.
Epoxy powder coatings have good mechanical properties, adhesion, and corrosion protection; however, their exterior durability is poor. Applications for decorative types include institutional furniture, shelving, and tools. Applications of protective epoxy coatings include
pipes, rebars, electrical equipment, primers, and automotive underbody parts. Where
enhanced chemical and corrosion resistance is needed, phenolic resins (Section 13.3.1) are
used to cross-link epoxy resins, with 2-methylimidazole as catalyst. Novolac epoxy resins
(Section 13.1.2) or blends of novolac and BPA epoxies give higher cross-link densities
than BPA epoxies alone. All these coatings discolor and chalk on exterior exposure. Polycarboxylic acid anhydrides (Section 13.3.2), such as trimellitic anhydride, are sometimes used
with BPA epoxy resins in applications for which greater resistance to yellowing and to
acids and solvents is needed. The latter coatings are generally being replaced with hybrid coatings that have somewhat better exterior durability and less question of toxic hazard.
28.1.2. Hybrid Binders
BPA epoxy resins are cross-linked with carboxylic acid – terminated polyester resins with
M̄n of a few thousand. Such formulations are called hybrid powder coatings, implying that
they are intermediate between epoxy and polyester coatings. Hybrid coatings have better
28.1. BINDERS FOR THERMOSETTING POWDER COATINGS
551
color retention and UV resistance than epoxy powder coatings, but still do not have good
exterior durability. Examples of end uses include water heaters, fire extinguishers, radiators, and transformer covers.
A variety of polyesters have been described. Most are derived from neopentyl glycol
and terephthalic acid (TPA) with smaller amounts of other monomers to adjust the Tg
to the desired level and give branching to increase the f̄n above 2 (Section 10.6). An
example is a polyester from neopentyl glycol (NPG) (364 parts by weight, 3.5 mol),
TPA (423 parts, 2.55 mol), adipic acid (AA) (41 parts, 0.24 mol), and trimellitic anhydride
(TMA) (141 parts, 0.74 mol) [7]. The acid number of the resin is 80 mg KOH per gram of
resin. The relatively high trimellitic anhydride content increases f̄n, compensating for the
low (about 1.9) f̄n of the BPA epoxy. The primary cross-linking reaction is ring opening of
the oxirane groups by carboxylic acids (Section 13.3.2). Esterification and transesterification reactions involving hydroxyl groups of the epoxy resin and homopolymerization reactions of oxirane groups may also play a role. A catalyst such as an ammonium or
phosphonium salt, for example tetrabutylammonium bromide or choline chloride,
permits baking temperatures in the range 160 to 2008C. Often, the polyester resins are
supplied with the catalyst blended in.
A proprietary, modified BPA epoxy that exhibits greater flow with comparable sintering resistance compared to conventional BPA epoxies has been reported [8].
28.1.3. Polyester Binders
Depending on the cross-linker, carboxylic acid– functional or hydroxy-functional
polyester resins are used. Flow properties of powder coatings containing carboxylic
acid –terminated polyesters tend to be poorer than those made with hydroxy-terminated
polyesters. Triglycidylisocyanurate (TGIC) (Section 13.1.2) has been widely used as a
cross-linker for carboxylic acid –terminated polyesters with basic catalysts. TGIC based
powder coatings have good exterior durability and mechanical properties. Examples of
end uses are outdoor furniture, farm equipment, fence poles, and air conditioners.
Although TGIC is expensive, the amounts required are relatively small because of its
low equivalent weight. Typical binders contain 4 to 10 wt% TGIC and 90 to 96 wt% of
carboxylic acid– terminated polyester. The polyesters used are generally less branched
than those used in hybrid coatings because of the higher functionality of TGIC compared
to BPA epoxies. One polyester, for example, is made from NPG (530 parts by weight,
5 mol), TPA (711 parts, 4.3 mol), isophthalic acid (IPA) (88 parts, 0.47 mol), pelargonic
acid (58 parts, 0.37 mol), and TMA (43 parts, 0.22 mol); the acid number is 35 [7]. Such
resins are prepared in a two-stage process to minimize the problems caused by the high
melting point and low solubility of TPA (Section 10.6). Alternatively, dimethyl terephthalate is used in place of TPA and the polymer is formed by transesterification
(Section 10.6). High equivalent weight (low acid number) is desirable, since this
reduces the required amount of TGIC, but cross-link density decreases with the higher
equivalent weight. There is an optimum for each application.
A study of a variety of catalysts for TGIC – polyester powder coatings was made to find a
catalyst that will cure at a temperature of 1208C while still having adequate storage stability
and flow. The work was aimed at a powder coated primer for use in coating aluminum alloy
for aircraft. The properties of certain aluminum alloys change if exposed to temperature
over 1208C. The most appropriate catalyst was found to be benzyltrimethylammonium
552
POWDER COATINGS
chloride [9]. A combination of a fumaric unsaturated polyester and an allyl-functional
polyester with TGIC and benzoyl peroxide as a free radical initiator is reported to cure at
a temperature of 1258C [10].
A problem with NPG – TPA polyesters is that they can contain cyclic esters without
functional groups; such compounds have been shown to bloom to the surface of polyester – TGIC coatings. A combination of NPG and 2-butyl-2-ethyl-1,3-propanediol with
TPA and IPA gives a semicrystalline polyester with much less cyclic ester and no blooming
when cross-linked with TGIC to make gloss coatings [11].
There has been increasing concern about toxic hazards of TGIC. Partly as a result, tetra(2hydroxyalkyl)bisamides (Section 17.4) are being used as cross-linking agents for carboxylic
acid–functional polyesters in exterior durable coatings [12]. These coatings also have good
mechanical properties and flow. Water is evolved from the cross-linking reaction, which
may limit the film thickness that can be applied without popping. Polyesters with terminal hindered carboxylic acid groups, cross-linked with tetra(2-hydroxyalkyl)bisamide, provided
powder coatings with improved flow, less limitation on film thickness to avoid popping,
and excellent acid rain resistance. A hindered COOH group increases flow due to the
reduction of hydrogen bonding between COOH groups. The hindered COOH groups also
reduce the cure rate, providing for improved leveling [13]. An example of such a polyester
is made in three stages. First, NPG is esterified with TPA to give a hydroxy-functional polyester. Second, that polyester is reacted with IPA to give a carboxylic acid–terminated
polyester. Then that polyester is reacted with hydroxypivalic acid to give a polyester
terminated with tertiary carboxylic acid groups [14].
In North America, polyester coatings with blocked aliphatic isocyanates (Section 12.5)
as cross-linkers for hydroxy-functional polyesters are widely used. The coatings have
exterior durabilities equal to, or somewhat better than, TGIC cross-linked polyesters,
and the superior mechanical properties and abrasion resistance typical of polyurethane
coatings. Blocked isocyanate –polyester powder coatings generally show better flow
than most powder coatings, perhaps because the unreacted cross-linkers or blocking
agents released by unblocking are good plasticizers. Examples of their end uses are automobile wheels, lighting fixtures, garden tractors, fence fittings, and playground equipment.
Derivatives of isophorone diisocyanate (IPDI), bis(4-isocyanatocyclohexyl)methane
(H12MDI) isocyanurate, and tetramethylxylidene diisocyanate (TMXDI) low MW prepolymers (Section 12.3.2) are examples of blocked isocyanates that are solids. Blocked
isocyanates from sterically crowded isocyanates such as TMXDI have the potential advantage of unblocking at a somewhat lower temperature [15,16]. Probably, the most widely
used blocking agent is e -caprolactam. Caprolactam-blocked H12MDI –TMP prepolymer
is reported to cure in 20 minutes at 1608C (208C lower than with caprolactam-blocked
IPDI or HDI), and give excellent exterior durability, flexibility, and impact resistance
[17]. A disadvantage of caprolactam is build up of volatilized caprolactam in baking
ovens. Also, there is pressure to reduce curing temperatures. Oxime blocking agents are
being used, since they react at lower temperatures [16]. The sterically crowded blocked
isocyanates from diisobutyl and diisopropyl oximes give the lowest cure temperatures
[16]. Oxime-blocked isocyanate coatings tend to yellow during cure, especially if overbaked. Also, there is concern about toxic hazards with oximes. 3,5-Dimethylpyrazole and
1,2,4-triazole (and mixtures of the two) give blocked isocyanates that combine lower cure
temperatures and no yellowing [18]. 3,5-Dimethylpyrazole-blocked H12MDI is reported
to cure coatings at 1508C with excellent flow and leveling [17].
28.1. BINDERS FOR THERMOSETTING POWDER COATINGS
553
As cross-linkers, uretdiones (isocyanate dimers) have the distinct advantage that they
cleave to give isocyanates, so that no blocking agent is evolved (Section 12.5). The
cure temperatures without catalyst are high, and considerable work has been done on catalysis. Most recently it has been recommended that tetraalkylammonium carboxylates be
used as catalyst along with TGIC to react with free carboxylic acid groups on the polyester
that inhibit the cross-linking. The combination permits curing at 1508C [19]. A patent has
been applied for in which an IPDI uretdione with the NCO groups partially reacted with
TMP, benzyltrimethylammonium chloride as catalyst, and TGIC as an acid scavenger are
used with a hydroxy-functional polyester in formulating a powder coating [20].
An amino resin, tetra(methoxymethylol)glycouril (TMMGU), is also used as a crosslinker for hydroxy-functional resins [21]. Since methyl alcohol is generated as a
by-product, film thickness may be limited by popping or retention of bubbles in the
cured film. Release of the methyl alcohol from the film can be aided by a variety of
approaches that slow cross-linking, such as use of methyltoluenesulfonimide as a catalyst
with a solid amine such as tetramethylpiperidinol as an inhibitor. Toluene sulfonamide –
modified melamine –formaldehyde resins can also serve as cross-linkers for hydroxyfunctional polyesters [22]. The reaction with these resins releases volatile by-products
also; it is reported that buffering by a cure rate regulator, 2-methylimidazole, minimizes
popping [23].
28.1.4. Acrylic Binders
A variety of acrylic resins can be used in powder coatings: Hydroxy-functional acrylics
can be cross-linked with blocked isocyanates or glycolurils, and carboxylic acid –
functional acrylics can be cross-linked with epoxy resins or with carbodiimides [24].
Hydroxyalkylamides have also been used to cross-link COOH-functional acrylic
powders [25]. Controlled radical polymerization by ATRP of COOH-functional acrylic
resins results in narrower MW distribution than by conventional free radical polymerization. (See Section 2.2.1.1 for a discussion of the polymerization process.) Such a resin
had a M̄w/M̄n of 1.25 compared with 1.95 for a similar resin made by conventional free
radical polymerization. The melt viscosity at 1808C of the ATRP polymer was
11.2 Pa.s compared with 57.2 for the conventional resin. A powder coating using the
ATRP polymer and bis(N, N-di-2-hydroxyethyl)adipamide as a cross-linker was compared with a similar powder coating made with the conventional free radical copolymer.
The ATRP powder coating gave cured films with better leveling than the conventional
polymer and had superior storage stability. The lower melt viscosity and hence
narrow MW distribution of the ATRP polymer resulted in better leveling. The reduced
amount of very low MW molecules reduced the fusion of the powder particles during
storage [26].
The greatest interest has been in epoxy-functional acrylics made with glycidyl methacrylate (GMA) (Section 13.1.2) as a co-monomer and cross-linked with dicarboxylic acids
such as dodecanedioic acid [HOOC(CH2)10COOH] [27] or carboxylic acid– functional
resins [28]. An epoxy-functional acrylic for automotive primer –surfacer is said to
require an M̄n below 2500, a calculated Tg above 808C, and a monomer composition
such that melt viscosity is less than 40 Pa.s at 1508C [29]. Such a resin can be made
with 15 to 35% GMA, 5 to 15% butyl methacrylate (BMA), with the balance being
methyl methacrylate (MMA) and styrene. Such an acrylic evaluated for automotive
clear coats had an M̄n of 3000, an M̄w/M̄n of 1.8, and a Tg of 608C [28]. Epoxy-functional
554
POWDER COATINGS
acrylic polymers can be synthesized by emulsion polymerization using
cyclodextrin-assisted polymerization. Cyclodextrin is dissolved in water, the acrylic
monomers are emulsified into the solution, and a redox initiator system is added. The
product is a dispersion of the epoxy-functional acrylic, which is filtered and dried [30].
Acrylic powder coatings generally have superior detergent resistance and are used for
applications such as washing machines. Acrylics tend to be incompatible with other
powder coatings, requiring caution when changing coating types to avoid contamination,
which can result in cratering. As with liquid coatings, acrylic powder coatings tend to have
poorer impact resistance than polyesters [31].
28.1.5. Silicone Binders
The use of silicone and silicone/polyester resins in high heat resistance powder coatings
has been reported. Silicone powder coatings require lower processing temperatures than
other powder coatings. To avoid bubbling due to the escape of water during cross-linking,
film thickness is limited to 2 mil or less. The coatings were cured at 2328C [32].
28.1.6. UV Cure Powder Coatings
Powder coatings have been developed that are cured by UV. (See Chapter 29 for discussion of UV curing.) This process permits rapid cure at lower temperatures [33,34]. Since
the powder is stable in the dark, premature reaction during powder production is minimized. Both free radical – and cationic-cure coatings have been made. Free radical –
cure coatings use acrylated epoxy resins (Section 29.2.4) and/or acrylated polyesters or
unsaturated maleic polyesters as binders. The effects of annealing before UV cure of
methacrylated BPA epoxy resin, acrylated BPA epoxy resin, silylated acrylated BPA
epoxy resin, and an acrylated polyester have been investigated. The methacrylated BPA
resin films annealed at 1108C were very inflexible. The acrylated BPA epoxy resin
provided high MEK rub resistance and good smoothness when annealed at 1708C.
The silylated acrylated BPA epoxy resin showed excellent smoothness and impact resistance when annealed at 908C; however, the MEK rub resistance was lower than that of the
other films. The acrylated polyester showed good smoothness when annealed at 1908C and
was intermediate in other properties [35].
A UV cure powder coating for fiberboard formulated with an acrylated polyester and an
acrylated BPA – polyphenoxy resin has been reported. The coating is sprayed on the fiberboard, which is backed up with a grounded copper plate. After fusing under IR the coating
is UV cured [36].
Crystalline unsaturated polyesters with difunctional vinyl ethers have been
recommended as a vehicle for UV cure powder coatings that are fused with IR at temperatures as low as 1208C followed by UV [37]. For example, a crystalline polyester from
1,6-hexanediol, fumaric acid, and terephthalic acid is used with the reaction product of
HDI with 4-hydroxybutyl vinyl ether [38]. Semi-crystalline low MW polycarbonate
diols esterified with methacrylic acid permit storing at 408C and curing at 100 to
1158C. Synthesis of the resin is discussed in Section 17.9 [39].
Cationic UV cure coatings use BPA epoxy resins as binders. Photoinitiators must be
incorporated in the formulation. After application, the powders are fused by passing
under infrared lamps and then are cured by passing under UV lamps. Film formation with
infrared lamps can be carried out at film temperatures below 1208C, and films UV cure in
28.3. FORMULATION OF THERMOSETTING POWDER COATINGS
555
1 second or less while still hot. This permits use on heat sensitive substrates such as wood
and some plastics. Good leveling is possible because viscosity does not begin to increase
until cure is initiated by UV. As with any other UV cure system, pigments can interfere
with curing since the pigments may absorb UV, limiting the film thickness that can be
cured. Although some pigmented UV cure powder coatings can be cured, the main interest
has been for clear coatings.
28.2. BINDERS FOR THERMOPLASTIC POWDER COATINGS
The first powder coatings were thermoplastic coatings, but they now account for less than
10% of the North American market. Thermoplastic powder coatings have several disadvantages compared to thermosetting coatings. They are difficult to pulverize to small particle sizes; thus, they can be applied only in relatively thick films. To have good film
properties the binders must have high MWs and/or strong intramolecular hydrogen
bonding. As a result, when they are baked even at high temperatures they are viscous
and give poor flow and leveling.
Vinyl chloride copolymers (PVC) and, to a more limited extent, polyolefins, polyamides (nylons), fluoropolymers, and thermoplastic polyesters are used as binders. High
vinyl chloride – content copolymers (Section 17.1.1) are formulated with stabilizers and
a limited amount of plasticizer, often a phthalate ester, so that the Tg is above ambient
temperature. The partial crystallinity of PVC may help stabilize the powder against sintering. Vinyl powders are generally applied as quite thick films, 0.2 mm and higher, by fluidized bed application, discussed in Section 28.5.2. Dishwasher racks, handrails, and metal
furniture are examples of end uses.
Polyolefin powders give coatings with low water absorption and excellent chemical
resistance. They are used to coat lab and food-handling equipment. Nylon 11 –nylon
12 –based powder coatings exhibit exceptional abrasion and detergent resistance. They
are used as antifriction coatings and as coatings for hospital beds, clothes washer
drums, and other applications that must withstand frequent cleaning or sterilization and
have good toughness and wear resistance. Fluoropolymers such as poly(vinylidene fluoride) and ethylene– chlorotrifluoroethylene copolymers are used for coatings that require
exceptional exterior durability, such as aluminum roofing and window frames, and also
for resistance to corrosive environments, such as equipment for chemical plants. The
application of poly(vinylidene fluoride), nylon 11, and nylon 12 powders by high velocity
thermal spray (Section 28.5.2) to aluminum and steel has been studied [40].
Thermoplastic polyester coatings are sometimes made using scrap or recycled poly (ethylene terephthalate). Polyolefin-based powder coatings are used, for example, in carpet backing,
but the volume used in metal applications has been limited by generally inferior adhesion.
Ethylene–acrylic acid (EAA) and ethylene–methacrylic acid copolymer resins that give
coatings with substantially better adhesion [41,42] are available (Section 28.5.2).
28.3. FORMULATION OF THERMOSETTING POWDER COATINGS
The challenge facing formulators is satisfying a combination of conflicting needs: (1)
minimization of premature cross-linking during production; (2) stability against sintering
during storage; (3) coalescence, degassing, and leveling at the lowest possible baking
556
POWDER COATINGS
temperature; and (4) cross-linking at the lowest possible temperature in the least possible
time. Furthermore, flow and leveling must be balanced to achieve acceptable appearance
and protective properties over the range of expected film thicknesses. Coatings that flow
readily before cross-linking can form smooth films, but they may flow away from edges
and corners because of the surface tension differential – driven flows that result from the
faster heating of the edges (Section 24.3).
If the Tg of a coating is high enough, sintering can be avoided. However, coalescence
and leveling at the lowest possible temperature are promoted by low Tg. Short baking
times at low temperatures are possible if the resins are highly reactive and if the baking
temperature is well above the Tg of the final cross-linked film. However, such compositions may cross-link prematurely during extrusion, and the rapid viscosity increase as
the particles fuse in the oven limits the ability of the coating to coalesce and level. Compromises are needed. With current technology, a crude rule of thumb is that the lowest
feasible baking temperature is 508C above the melt extrusion temperature and 70 to
808C above the Tg of the uncured powder. Thus, minimum baking temperatures are
about 125 to 1358C for a powder with a Tg of 558C.
Several studies have addressed the changes in viscosity during film formation
[21,25,43– 45]. Nakamichi used a rolling ball viscometer to measure viscosities of
powder coatings on a panel during heating [43]. Results for three types of coatings are
shown in Figure 28.2. In each case, viscosity is high immediately after fusion of the
Figure 28.2. Non-isothermal viscosity behavior of powder coatings during film formation as functions
of time and panel temperature. 1, Acrylic– dibasic acid type; 2, polyester-blocked isocyanate type;
3, epoxy– dicy type. (From Ref. [43], with permission.)
28.3. FORMULATION OF THERMOSETTING POWDER COATINGS
557
powder but drops off sharply with increasing temperature. Viscosity levels off as crosslinking reactions begin to increase the MW and then increases rapidly as the coating
approaches gelation. Flow is governed by the lowest viscosity attained and by the
length of time that the coating stays near that viscosity, called the flow window. In
Figure 28.2, coating 2 will flow more than coating 1 even though the lowest viscosity is
about the same because the reaction of coating 2 is slower and the flow window is longer.
Some authors discuss the temperature dependence of viscosity in terms of Arrhenius
type relationships and discuss the activation energies for viscous flow. As indicated in
Section 3.4.1, the dependency of viscosity on temperature does not actually follow Arrhenius relationships, but rather is dependent on free volume availability. The most important
factor controlling free volume availability is T 2 Tg.
An area that needs more study is the important question of the driving forces for
coalescence and leveling. Coalescence of latex particles has been studied extensively
(Section 2.3.3), but not coalescence of powder coatings. The main driving force for
coalescence may be the reduction of surface area driven by surface tension; in keeping
with this hypothesis, it has been suggested that high surface tension promotes coalescence
[44]. Low melt viscosity promotes leveling, but the mechanism involved is not well understood. It has been proposed that the driving force for leveling is surface tension as reflected
by the Orchard equation (Section 24.2) [44]. The data obtained fit the Orchard equation
reasonably well for relatively thin and/or fluid films, but thicker and/or somewhat
more viscous films leveled better than predicted. It has been reported that surface
tension differential flow can cause cratering of powder coatings and that addition of
small amounts of additives such as poly(octyl acrylate) derivatives may overcome this
problem [44]. In the same paper it was reported that epoxidized soy fatty acids and hydroabietyl alcohol reduce orange peel but do not affect cratering.
Reference [46] reviews the many factors affecting leveling of powder coatings. Liquid
coatings that exhibit Newtonian viscosity level much better than thixotropic coatings.
Blocked isocyanate coatings generally exhibit better leveling than other types since the
blocking agent (caprolactam or oxime) plasticizes the film, lowering viscosity, until it
evaporates. Particle size distribution of the powder affects surface smoothness; the
largest particles should be no larger than two-thirds of the proposed film thickness.
Film thickness has a major effect; if the film thickness is less than about 2 mil, particle
size affects smoothness; above about 5 mil sagging is likely to occur. Surface tension, particularly surface tension differentials across the surface, can have major effects on leveling
and cratering. Additives, particularly low-surface-tension polyacrylates, are widely used
to promote leveling and reduce cratering. Commonly, these polyacrylates are adsorbed
on the surface of silica to ease incorporation into the powder coating. The effectiveness
of the additives depends not only on their chemical structure but also on MW and concentration. With one additive, a MW of about 6000 is more effective than a MW approaching
100,000. Concentrations must generally be kept below 1.5% since excess additive gives a
tacky surface to the cured film. Higher concentrations of additive are effective if they are
also hydroxy-functional so that they cross-link into the binder. The additives also reduce
or eliminate cratering.
Severe cratering can be caused by cross contamination of one type of powder with
another. Contamination can occur during powder production or application, for
example, if a spray booth is inadequately cleaned when switching from one type of
powder to another. A medium MW acrylic additive with a broad MW distribution is
reported to solve this problem in some cases. In this case, the additive must be added in
558
POWDER COATINGS
a premix, not adsorbed on silica, since the high MW fraction tends to adhere tenaciously to
the surface of the silica.
Surface tension also affects wetting of the substrate; if it is too high, poor wetting
occurs, which leads to defects such as cratering. On the other hand, low surface tension
affects leveling adversely. Furthermore, uneven surface tension on the surface of coatings
leads to film defects. The problem is further complicated by the fact that surface tensions
decrease with increasing temperature, and the extent of this decrease varies with structure.
Surface tensions over a range of temperatures from 138 to 1848C have been studied by
asymmetric drop shape analysis (ADSA) of epoxy resin and epoxy resin compounded
with five acrylate additives and two polyester-modified methylalkylsiloxane additives.
The acrylate additives reduced the high temperature surface tension of the epoxy resin
with little difference between the acrylate additives. The siloxane additives decreased
the viscosity further than the acrylates, especially siloxanes, in which the alkyl groups
were short chain groups compared to those with long chain groups [47].
Dynamic mechanical analysis (DMA) (Section 4.5) is useful for characterizing cured
films [22,41]. A DMA study showed that the Tg values of cured films of a series of decorative powder coatings, including a hybrid coating, a TGIC-polyester coating, and a
blocked isocyanate – polyester coating, were all in the range 89 to 928C [22]. The
average MWs between cross-links, M̄c (Section 4.2), for the cured films were in the
narrow range 2500 to 3000. It is noteworthy that years of trial and error formulation in
different laboratories with such disparate binders led to similar Tg and M̄c values. On
the other hand, a protective epoxy powder coating with a modified dicy cross-linker
gave cured films with a Tg of 1178C and an M̄c of 2200. These studies show that
powder coatings are formulated with higher Tg and lower cross-link densities than are
liquid coatings for the same end uses. Similar but not identical mechanical properties
are attainable.
DMA can be a powerful tool for helping suggest starting points for development work
on new binders for powder coatings for similar applications. Two different factors control
the Tg of resins: chemical composition and MW. It has been reported that higher MW,
more flexible resins are advantageous in that they can have adequate package stability
and also flow more easily during baking than a lower MW resin of similar Tg that has
more rigid chains [4].
Differential scanning calorimetry (DSC) is a very useful tool for Tg determination and
measurement of cure response; its use for studying powder coatings is reviewed in
Ref. [48]. It has also been recommended that a combination of modified dynamic load –
thermal mechanical analysis, DSC, and TGA be used to determine melting point, onset
flow, flow, and gel point [49]. DSC studies showed that pigmentation with TiO2 had
only a weak reinforcing effect and had essentially no effect on Tg, suggesting that
pigment –binder interaction is weaker in powder coatings than in liquid coatings [22].
Very commonly, on the order of 0.1 to 1% benzoin (melting point 133 to 1348C) is
included in the formulations of many types of powder coatings. It is said to improve
the appearance of films and to act as an antipinholing agent and a degassing aid. Other
additives with similar melting points are ineffective. The mechanism of benzoin’s
action has been the subject of much conjecture over the years. One study showed
benzoin plasticizes the melt and increases the flow window of polyester – glycoluril formulations, indicating it can improve leveling; high levels (1.4 to 2.4%) were used [21].
Another study has shown that benzoin evaporates from the applied film during the
first 6 to 8 minutes of cure and suggests to the authors voids left in the film facilitate
28.3. FORMULATION OF THERMOSETTING POWDER COATINGS
559
degassing [50]. In another study it was proposed that the effect of benzoin is related to its
air oxidation to benzil in the gas phase Tg. The study showed further that derivatives of
benzoin such as p-chlorobenzoin are effective in degassing and lead to less yellowing
of films than benzoin [51].
Wrinkle powder coatings have also been formulated with tetramethoxymethylglycoluril (TMMGU) (Section 11.4.3) cross-linkers using amine-blocked catalysts such as
2-dimethylamino-2-methylpropanol –blocked p-toluenesulfonic acid [21].
28.3.1. Low Gloss Powder Coatings
Since there is no volatile solvent, the volume of pigment in powders approaches the PVC
of the final film. At a PVC near CPVC, the viscosity of the fused powder would be far too
high for acceptable leveling. It has been shown that as pigmentation increases above a
PVC of about 20, the problems of leveling increases, due to the increase in melt viscosity
[31]. Thus, a common method for making low gloss liquid coatings by using a PVC that
approaches CPVC (Section 22.1) can be used to only a limited degree in powder coatings,
due to the increase in melt viscosity.
Low gloss and semigloss powder coatings have been prepared using approaches other
than high PVC [52]. Some reduction in gloss can be achieved by incorporating polyethylene micronized wax with as much inert pigment as flow permits. In hybrid polyester
powders, addition of organo-metallic catalysts along with wax is said to reduce gloss
somewhat further. Low gloss hybrid polyester coatings are made by using a large
excess over the stoichiometric ratio of epoxy resin with a high acid number polyester
and curing at high temperatures. By selection of catalysts such as cyclohexylsulfamic
acid (cyclamic acid) and stannous methane sulfonate, it has been found that smooth
matte finishes can be obtained with tetramethoxymethyl glycoluril (TMMGU) crosslinked polyester coatings [21].
Another approach is to blend two different primary resins or two different cross-linkers
with substantially different reactivities or with poor compatibility. For example, combination of a BPA epoxy resin, a carboxy-functional polyester resin, and a carboxyfunctional acrylic were used with a BPA – amine adduct or a blend of imidazole and
dicyandiamide as curing agent. The epoxy resin and polyester resins were compatible,
the polyester and acrylic resins were reported to be relatively compatible, but the epoxy
resin and acrylic resin were not compatible [53]. A prepolymer of dimethylolpropionic
acid (DMPA), a diisocyanate blocked with caprolactam, and a hydroxy-functional polyester are used in low gloss powder coatings; the COOH groups from the prepolymer are
cross-linked with a hydroxyalkylamide, TGIC, or GMA acrylics [17].
Addition of styrene –maleic anhydride copolymer to hybrid powder coatings together
with tetrabutylphosphonium bromide catalyst gives a low gloss coating. The reduction
in gloss is attributed to the dual cure system of anhydride – epoxy and COOH –epoxy
reactions [54].
560
POWDER COATINGS
Low gloss (matte) powder coatings with excellent properties have been made with a
bifunctional polyester having both OH and COOH functional groups, a GMA acrylic
copolymer, caprolactam-blocked IPDI, and dodecanedioic acid [55]. The gloss can be
controlled by the mole fraction of GMA in the acrylic resin. Gloss can be further
reduced by mixing a small amount of small particle size fumed SiO2 with the extruded
powder formulation.
Generally speaking, the gloss reduction achieved by the methods described above may
vary with film thickness and baking conditions. Thus, care during application is needed to
achieve consistent gloss.
28.4. MANUFACTURE OF POWDER COATINGS
Manufacture of powder coatings poses production and quality control issues quite different from those of liquid coating manufacture.
28.4.1. Production
Most powder coatings are manufactured by the same process: premixing, melt extrusion
(Figure 28.3), and pulverization (Figure 28.4). All major ingredients must be solids;
some liquid additives are used, but they must first be melted into one of the solid components to make a master batch which is then granulated. The granulated ingredients,
resins, cross-linkers, pigments, and additives are premixed in a batch process. A variety
of premixers is used; it is essential that they provide a uniform, intimate mixture of the
ingredients. The premix is fed through a hopper to an extruder in a continuous process.
The barrel of the extruder is maintained at a temperature moderately above the Tg of
the binder. In passing through the extruder, the primary resin and other low melting or
low Tg materials are fused, and the other components dispersed in the melt. The extruder
operates at a high rate of shear so that it can effectively separate pigment aggregates. The
melt can be extruded through a die—either a slot to produce a flat sheet or a series of round
orifices to produce spaghetti. More commonly, to reduce heat exposure, the melt is
extruded through a die with a larger bore to produce sausage that is fed between chilled
rollers to flatten it into a sheet and cool it.
Extruders have developed into sophisticated and rather expensive pieces of equipment.
Two types are commonly used: single screw and twin screw; in both types, a powerful
motor turns screws to drive the material through a barrel. The screws and barrel are
configured to mix the material thoroughly and apply a high rate of shear. A popular
single screw extruder uses a reciprocating action in addition to radial turning of the
screw to effect mixing and dispersion. Twin screw extruders use a combination of
screw segments and kneading segments. Both types of extruders are capable of excellent
dispersion of most pigments. They operate with relatively high viscosity formulations at
high shear rates and hence efficiently separate pigment aggregates (Section 21.4.6).
However, there is a trade-off between separating pigment aggregates and production
rate. Production capacity can be increased by pushing more material through the extruder
per unit time. Residence time in the extruder is sometimes reduced to 10 seconds or less,
but at some point, pigment dispersion, especially with some organic pigments, begins to
suffer. Poor color development and color variability may result (Section 28.4.2).
Polymeric dispersants used in pigmented powder coatings can substantially improve
Figure 28.3. Line for premixing, melt extrusion, and granulation. (Adapted from Ref. [7], p. 242, with permission.)
561
562
POWDER COATINGS
Figure 28.4. Line for pulverization and classification. (Adapted from Ref. [7], p. 252, with permission.)
pigment dispersion. For example, carbon black polymeric dispersants give increased
jetness and pigment loading can be increased with lower melt viscosity. Similarly, TiO2
pigmented powder coatings can be made with higher hiding at lower film thickness [56].
The granules are then pulverized. A variety of pulverizers is used. Some, such as pin
disk mills and hammer mills, work on the principle of striking airborne granules with
metal dowels or hammers mounted on a rapidly spinning disk. The newer opposed jet
mills work by causing high velocity collisions of granules with one another. Opposed
jet mills perform well for small (,12 mm) diameter powders needed for thin film application. Thermosetting extrudates are brittle and relatively easy to pulverize, but thermoplastics are generally quite tough and can be difficult to pulverize. For thermoplastics it
is often necessary either to cool the mill with liquid nitrogen or to grind dry ice along
with the granules so that the temperature is kept well below the Tg of the binder to
offset the heating effect of the milling process. Even so, thermoplastic powders are
generally available only in large particle sizes.
Some mills partially classify (i.e., fractionate) the powder, automatically returning
oversize particles for further pulverization. Further size classification, by sieving and/or
by an air classifier, is usually needed. The coarse particle fraction is sent back into the
mill for further reduction in size. The fines are collected in a bag filter and recycled in
the next batch of the same coating to be processed through the extruder. Finally, the classified powder is bulk blended for uniformity, packaged, and shipped to the customer.
28.4. MANUFACTURE OF POWDER COATINGS
563
A relatively new process uses supercritical CO2 as a solvent during extrusion [57]. (See
Section 23.2.5 for discussion of the use of supercritical CO2 in other applications.) The
powder coating materials and CO2 are premixed under pressure and then fed to an extruder
still under pressure. The extruder can be operated at lower temperatures than conventional
processing permits, reducing the risk of premature cross-linking in the extruder, hence use
of more reactive formulations. When the compound comes out of the extruder, the CO2
flashes off, reducing the temperature and pulverizing a major part of the product. The
product is classified; only the larger particle size fraction needs to be pulverized in the
conventional manner. It is said that the process affords excellent pigment dispersion.
A suspension method of preparing powders has been recommended for making narrow
size distribution GMA epoxy-functional acrylic or dicarboxylic acid powders for automotive clear coat powders [58]. An aqueous dispersion of a solution of the resins and crosslinkers is prepared, and the solvent is distilled off. The resulting suspension is centrifuged,
washed with water, and dried. The particle size distribution is said to be much narrower
than that of powders made by the conventional pulverizing process.
28.4.2. Quality Control
Close quality control of all components is required. In solvent-based coatings, the effect of
small differences in MW or MW distribution can be readily adjusted by small variations in
the solids of the coating. In powder coatings, there is no solvent in the formula with which
to make such adjustments. The only way of maintaining consistent quality of powder
coatings is to assure that the raw materials have no significant variation in MW and
MW distribution as well as monomer composition.
In processing thermosetting compositions, care must be exercised so that no more than
a minimal amount of cross-linking occurs at the elevated temperatures in the extruder. The
rate of travel through the extruder should be as rapid as possible consistent with achieving
the necessary mixing and dispersion of pigment aggregates. Reprocessing should be
minimized. In an extreme case, reprocessed material could gel in the extruder. More commonly, reprocessing can lead to increases in MW, due to some cross-linking reactions,
giving powders that show incomplete coalescence or poor leveling after application.
Only a limited amount of fines from the micropulverizer should be put in any one batch
of coating. If reprocessing is required, it is best to use a limited amount of the batch to
be reprocessed in each of several new batches rather than reprocessing the batch alone.
Color matching is often the most troublesome aspect of powder coating production. It is
more difficult than with liquid coatings, particularly with reprocessed powders. Generally,
one cannot blend batches of powder to achieve satisfactory color matching; it must be
done using the appropriate ratio of pigments in the extruder mix. In laboratory processing,
it is feasible to make an extruder mix with an estimate of the ratio of colorants needed to
match a color. The color of the coating is then checked against a standard. Color can be
adjusted to match the standard by mixing into the initial batch the estimated additional
quantities of pigments and running the coating through the extruder and pulverizer
again. In the laboratory, a third hit might be possible. The number of hits required is
kept to a minimum by using computer color matching programs (Section 19.9). In
production, however, it is almost essential to have the mix right the first time. Potential
problems can be minimized by extruding a small fraction of the production batch and
then checking its color. If the color match is satisfactory, processing is continued. If the
color needs adjusting, the initial fraction is granulated and returned to the hopper along
564
POWDER COATINGS
with the additional amounts of pigments required. Some batch to batch variation of color is
inevitable in pigment manufacture. For closely color matched powder coatings, the
pigment manufacturer is requested to supply selected batches with narrow tolerance
limits. In the case of large production runs, several extruder batches can be blended
before pulverizing to average out batch to batch differences. With sufficient care, color
reproducibility can be satisfactory for all but the most demanding end uses.
Other important quality control issues are particle size and particle size distribution.
These important variables become increasingly critical as particle size is reduced for
thin film (,50 mm) applications. Particle size distributions can be measured by passing
the powder through a stack of graduated sieves and weighing the fraction retained on
each sieve or by a variety of instrumental methods, such as laser diffraction particle
sizing [59]. The important effects of particle size and its distribution on the handling
and application characteristics of the powder are discussed in Section 28.5.1.
28.5. APPLICATION METHODS
Almost all thin-film powder coatings are applied by electrostatic spray. Other application
methods, important for protective and thick-film powder coatings, include fluidized bed,
electrostatic fluidized bed, and flame spray. See the general references listed at the end
of the chapter for more extensive discussions.
28.5.1. Electrostatic Spray Application
Electrostatic spray (Section 23.2.3) is the major process for applying powder coatings. The
powder is fluidized in a hopper and conveyed by a stream of air to an electrostatic spray
gun. This gun consists of a tube to carry the airborne powder to an orifice at which an electrode is located. The electrode is connected to a high voltage (40 to 100 kV) low amperage
power supply. Electrons emitted by the electrode react with molecules in the air, generating a cloud of ions, called a corona, around the orifice. The corona probably consists
predominantly of HO2 and, perhaps, O2
2 ions. Powder particles come out of the orifice,
pass through the corona, and pick up anions. The object to be coated is electrically
grounded. The difference in potential attracts the powder particles to the surface of the
part. They are attracted most strongly to areas not already covered, forming a fairly
uniform layer of powder even on irregularly shaped objects. During application, size segregation of the particles has been observed, with the finer particles more prevalent near the
substrate, presumably due to a combination of the higher charge on small particles and
their fitting between larger particles [60]. The particles cling to the surface strongly
enough for the object to be conveyed to a baking oven, where the powder particles fuse
to a continuous film, flow, and cross-link. Heat can be furnished by either conventional
convection ovens or by infrared (IR) lamps; use of induction heating has also been proposed. One comparison of alternative heat sources concluded that IR is least expensive
and can be adjusted to provide more rapid cure [61].
The powder particles that do not adhere to the substrate (overspray) are recovered as a
dry powder. This powder is usually recycled by blending with virgin powder. A schematic
diagram of electrostatic spray application of powder coatings, including recovery of overspray, is given in Figure 28.5. Production spray guns are usually mounted on automated
28.5. APPLICATION METHODS
565
Figure 28.5. Production equipment for electrostatic spray application of powder coating, showing the
collection of overspray powder. (From Ref. [4], p. 18, with permission.)
reciprocators; such equipment functions smoothly with minimal worker attention as long
as a single color and type of powder are being applied.
A limitation of powder coatings is the difficulty of changing colors. When spraying
liquid coatings, one can flush the gun with solvent and shift the feed lines to the guns
from those feeding one color to those feeding another. In this way, successive objects
on a conveyor line can easily be painted different colors. However, if changeover of
powder coatings was attempted in this way, there would be sufficient dust in the air in
the spray booth for color contamination both on the product and in the overspray collecting
units. In production, the operation must be closed down, the booth cleaned, and the overspray collecting units changed to collect the next color. Although spray booths and overspray units have been designed to minimize cleanup time, it is still only economically
feasible to make a reasonably long run of one color before shifting over to another
color. Many important applications of powder coatings, such as fire extinguisher cases,
are single color end uses, or are applications, such as metal furniture, for which runs of
single colors are long enough to dedicate a spray booth to each of the limited number
of colors involved.
Water-washed spray booths have been suggested to avoid the cost of installing
special spray booths for shorter runs [62]. The overspray collected in the water is
dispersed by a low foaming nonionic surfactant and flocculated with a melamine
resin (Section 23.2.7).
566
POWDER COATINGS
In electrostatic spray, the charged powder particles wrap around the grounded object to a
degree and coat exposed surfaces not in direct line with the spray gun. Nevertheless, in
production, it is often desirable to use spray guns on both sides of the object, making it possible
to coat uniformly such intricately shaped objects as automobile wheels and tube-and-wire
metal furniture. The process is strongly affected by the Faraday cage effect (Section
23.2.3). As a result, it is difficult to get full coverage of areas such as interior corners of
steel cabinets; the interior of pipes can only be coated with a spray gun inside the pipe.
Film thickness increases with increasing voltage and decreasing distance between the
spray gun and the product being coated. Larger particle size powders tend to give
increased film thickness. One can apply thicker coatings by heating the object to
be coated before applying the powder. However, film thickness is limited by the fact
that once a certain film thickness has been reached, the powder coating acts as an insulator
and does not attract further particles. The insulating properties of the powder coating mean
that defective coated parts generally cannot be recoated and must be stripped to be
recoated. Coating over other coatings and over plastics frequently requires application
of a conductive primer coating.
There is considerable room for improvement of the process and need for a better understanding of how it works. For example, there is no satisfactory physical explanation for
why the powder clings to the object as well as it does [63]. One opportunity for improvement is to increase efficiency of charging by the corona of the electrostatic gun. It is estimated that only about 0.5% of the anions in the corona become attached to powder particles
[64]. The rest are attracted to the nearest grounded object, where they at best do no good
and at worst may reduce deposition efficiency and increase the Faraday cage effect [65].
Research on ways to improve this situation is under way; some experts think that if charging efficiency could be raised to 10%, deposition efficiency might become high enough to
reduce overspray sufficiently that it would no longer be necessary to collect it.
Particle size and distribution have a critical effect on powder coating. The range of particle sizes must be limited; as a rule of thumb, the predominant particle diameter should be
somewhat less than the intended film thickness, and the largest particles should be no more
than twice the film thickness. Figure 28.6 shows a typical distribution, presumably for a
powder coating to be applied at a thickness of around 50 mm [66].
Figure 28.6. Particle size distribution of a typical white epoxy powder coating. (From Ref. [66], with
permission.)
28.5. APPLICATION METHODS
567
Very fine particles do not flow properly in hoppers and feed lines. In general, only 6 to
8 wt% of the particles are smaller than 10 mm in diameter. Small particles have a higher
surface area/volume ratio and hence acquire a higher charge/mass ratio as they pass
through the corona of an electrostatic gun. After charging, the particles are affected by
three forces: the electrostatic field, air movements, and gravity. Theoretical calculations
predict that gravity should be the predominant effect on very large particles and that air
flow should predominate for the very small ones [67]. These theoretical predictions
were confirmed by particle size analysis of overspray; with a typical powder, the overspray
was rich in particle sizes below 20 and above 60 mm, indicating that the intermediate sizes
were transferred most effectively. Small particles probably penetrate best into Faraday
cages. Large particles flow better through the application system and retain their charge
longer and therefore cling to the object better between application and baking. The
dusts of very small (,1 mm) particles may present a toxic hazard if they are inhaled.
On balance, it appears that particle sizes concentrated in the range 20 to 60 mm are best
for powder coatings that are to be applied at film thicknesses of 30 to 60 mm.
Automotive clear coats with maximum smoothness are reported to result with fine particle size powders having a narrower size distribution [58,68]. However, application of
such powders (averaging 10 mm) presents problems. Modification of application equipment has been reported to solve many of the problems [68]. Constant feeding of
powder was accomplished by controlling temperature and humidity in the powder feed
system and incorporating an agitator in the system to aid fluidization. Spitting of aggregates of particles was minimized by redesigning the gun nozzle. Transfer efficiency was
improved by increasing charging efficiency, minimizing free ions, and adding an external
electrode of intermediate charge to increase the electric field intensity near the part being
coated.
Another means of applying electrostatic powder employs triboelectric charging of the
powder particles. The particles are charged by the friction generated from streaming
through a poly(tetrafluoroethylene) spray tube in the gun instead of a high voltage
source generating a corona at the gun orifice. The mechanism is analogous to the
buildup of a static charge on a comb when combing your hair. Since there is not the
large differential in charge between the gun and the grounded work being sprayed, no significant magnetic field lines are established and Faraday cage buildup is minimal, facilitating coating hollows in irregularly shaped objects. Smoother coatings are obtained. On the
other hand, throughput is slower and stray air currents can more easily deflect the particles
between the gun and the object being coated. Triboelectric charging is widely used in
Europe and is gaining popularity in North America. Guns are available that combine triboelectric and corona charging.
28.5.2. Other Application Methods
Fluidized beds are the oldest method of applying powder coatings. The equipment consists
of a dip tank, the bottom of which is a porous plate. Air is forced through the porous plate
and acts to suspend the powder in the dip tank in air. The flow behavior of air suspensions
of powders resembles that of fluids, hence the term fluidized bed. The object to be coated is
hung from a conveyor and heated in an oven to a temperature well above the Tg of the
powder. The conveyor then carries the part into the fluidized bed tank. Powder particles
fuse onto the surface of the object. As the thickness of fused particles builds up, the
coating becomes a thermal insulating layer, so that the temperature at the surface of the
568
POWDER COATINGS
coating becomes lower, finally reaching the stage at which further particles do not stick to
the surface. The last particles that attach to the coated surface are not fused completely, so
the conveyor must then carry the object into another oven, where the fusion is completed.
Film thickness depends on the temperature to which the part is preheated and on the Tg of
the powder. Thin films cannot be applied in this fashion. Most commonly, the method is
used for applying thermoplastic coating materials.
Electrostatic fluidized beds are similar, but electrodes are added to generate ions in the air
before the object passes through the powder. The object to be coated is grounded. The
process is illustrated schematically in Figure 28.7. Powder is attracted to the object by electrostatic force, as in electrostatic spray. The object can be heated when thick films are desired,
but heating is not necessary. This method is used to apply thermoplastic and some thermosetting powders: for example, electrical insulation coatings. There is no overspray and
powder losses are minimal; color change-over is also easier. Thinner films can be applied
compared to conventional fluidized bed coating. However, very thin films are hard to
apply, there is a strong Faraday cage effect, and it is difficult to coat large objects uniformly.
Powder coil coating is also practiced. In some cases, the powder is applied to the coil
strip by electrostatic spray [34]. In another process, an electrostatically charged strip is
passed through a cloud of powder, then to the oven for fusing [70]. The method makes
possible coating of perforated or preembossed metal and has the advantage of no VOC
emissions. Capital cost is said to be lower, but line speeds are somewhat slower than
with conventional coil coating.
Flame spray is another technology for applying thermoplastic powder coatings
[41,42,71]. In a flame spray gun, the powder is propelled through a flame, remaining
there just long enough to melt. The molten powder particles are then directed at the
Figure 28.7. Electrostatic fluidized bed coating apparatus. 1, Air inlet; 2, air regulator; 3, porous membrane; 4, object to be coated; 5, electrodes; 6, fluidized powder; 7, cloud of charged powder; 8, ground.
(From Ref. [69], with permission.)
28.6. ADVANTAGES AND LIMITATIONS OF POWDER COATINGS
569
object to be coated. The flame heats and melts the polymer and heats the substrate above
the melting temperature of the polymer so that the coating can flow into irregularities in
the surface to provide an anchor for adhesion. The combination of flame temperature (on
the order of 8008C), residence time in the flame (small fractions of a second), Tg of the
coating, particle size distribution, and substrate temperature must be carefully balanced.
Particle size distribution must be fairly narrow; very small particles pyrolyze at 8008C
before larger particles can melt.
Poly(vinylidene fluoride), ethylene –chlorotrifluoroethylene, perfluoroalkoxyalkane,
and fluorinated perfluoroethylenepropylene have been applied by flame spray to steel
and showed excellent salt spray corrosion resistance. The chemical resistance of perfluoroalkoxy alkane polymer was particularly outstanding [72].
In contrast to other methods of application, flame spray permits application in the field,
not just in a factory. Examples of commercial or experimental applications include drum
linings, metal light poles, bridge rails, concrete slabs, and grain railcars, among others.
Since the application is not electrostatic, nonconductive substrates such as concrete,
wood, and plastics can be coated. The coating need not be baked, so substrates that will
withstand just the temperature from the impinging spray can be coated by this method.
Since the coatings are thermoplastic and are not applied electrostatically, it is possible
to use powder to repair damaged areas of coating, which is generally not possible with
other powder coating methods.
Disadvantages of flame spray application include limitations on the service temperatures of objects coated with thermoplastic coatings (758C in the case of EAA-based coatings) (Section 28.2) and the need for careful control of application variables [41].
Overheating the polymer can lead to thermal degradation and hence to poor coating
performance, with no visual indication of degradation until the polymer begins to
pyrolyze. Adhesion to steel is also affected by application variables [42]. The carboxylic
acid groups promote adhesion, but coatings are subject to cathodic delamination (disbonding) because of the acid groups.
28.6. ADVANTAGES AND LIMITATIONS OF POWDER COATINGS
In this section, advantages and limitations of powder coatings are compared to solventborne or waterborne coatings. The primary advantages may be summarized as follows:
VOC emissions are low; flammability and toxicity hazards are substantially reduced;
thick films of 100 to 500 mm can be applied in a single operation; and energy consumption
can be reduced. At first blush, the low energy requirements of powder coatings are surprising, since their baking temperatures are generally higher than those of most solventborne
baked coatings. However, with little or no volatiles being emitted into the oven, the air in
the oven can be recirculated with almost no make-up air. In contrast, with solventborne
coatings the solvent concentration in the air in the oven must always be kept well
below the lower explosive limit, resulting in the need to heat a substantial volume of
make-up air. Also, air flow through the spray booth can be lower with powder coatings
because it is no longer necessary to keep the solvent concentration in the air below a
safe concentration for people to be in the spray booth. In winter, the cost of heating air
flowing at a high rate through a spray booth can be high. Since there is no drying or
flash off time, parts can be hung closer to one another on the conveyor. There is no
sagging or dripping.
570
POWDER COATINGS
An important further factor reducing the overall cost of powder coatings is that the
overspray powder can be collected in filter bags from the spray booth and reused. This
not only increases paint utilization, but also eliminates the cost and difficulty of disposing
of the sludge obtained from the water wash spray booths used with liquid paints.
Important limitations of powder coatings are:
1. Explosive hazards. Whereas the absence of solvents eliminates the flammability
problem, suspensions of powder in air can explode. Consequently, manufacturing and
application facilities must be designed to avoid powder explosions. With good engineering
and good housekeeping, the processes can be operated safely. Triboelectric charge
systems are less likely to initiate explosions than corona charge spray guns, which can
spark if brought near a grounded conductor.
2. Inability to coat large or heat sensitive substrates. With electrostatic spray and fluidized bed methods, only baked applications can be considered, and since the baking temperature must be fairly high, only substrates that can withstand the baking cycle can be
used. This limits applications of conventional thermosetting powder coatings almost
entirely to metal substrates. UV cure powder coatings can be used on some heat sensitive
substrates, and as noted in Section 28.5.2, flame spray can be used for some other
applications.
3. Appearance limitations. Broadly speaking, powder coatings can have good appearance, but some of the appearance effects attainable with liquid coatings are difficult or
impossible to match with powders. Color matching is more difficult than with liquid coatings, and color reproducibility can also be difficult. Since there is no solvent loss, there is
not the degree of shrinking during film formation required for metallic coatings to show
the color flop typical of automotive coatings. One can make powder coatings that
contain aluminum flake pigments; they sparkle but display little color flop. As discussed
in Section 28.3, a compromise between leveling and edge coverage is inevitable. The usual
result is that the formulator must settle for a film with some orange peel to get acceptable
edge coverage. The thinner the film, the more difficult the problem becomes. As also discussed in Section 28.3, there are problems making and controlling gloss of low and
medium gloss powder coatings.
Achieving satisfactory appearance is a major obstacle to development of powder clear
coats for autos and trucks [73]. Uniform film thickness must be applied to the entire car
body because leveling of powder coatings is quite sensitive to variations in film thickness.
This requires close control of powder flow rates; which is an engineering challenge. Films
thinner than 1.0 mil are difficult to achieve because of pinholing. The entire operation,
including reclamation of overspray, must be scrupulously clean to avoid dirt particles,
which are especially noticeable in clear coatings. It is difficult to coat objects with
outside sharp corners and inside corners tend to get low film thickness, due to the
Faraday cage effect.
4. Materials limitations. Since all major components must be solids, a smaller range of
raw materials is available to the formulator. Furthermore, it is not possible to make thermosetting powder coatings for which the Tg of the final film is low. This limits the range of
mechanical properties that can be formulated into a powder coating.
5. Limits of production flexibility. Economics of production and application suffer
badly whenever frequent color changes are needed. Cleanup between color changes is
time consuming. Powder coatings are best suited to reasonably long production runs of
REFERENCES
571
the same type and color of powder. Water-washed spray booths permit shorter runs, but the
advantage of recycling overspray is lost. New engineering designs have reduced the
problem of color changes.
Some of the limitations of powder coatings can be overcome by making aqueous dispersions (slurries) of powders. This eliminates the potential for powder explosions, broadens the range of application methods, and reduces storage stability problems. The Tg of the
powder no longer needs to be high to avoid sintering, and therefore more flexible coatings
can be formulated and lower baking temperatures can be utilized. Settling problems can be
minimized and appearance improved by using smaller particle size, less than 10 mm, in the
slurries. However, new challenges are added. Aqueous dispersions must be stabilized
without undue sacrifice of properties due to the presence of surfactant, and spray
rheology must be controlled. Aqueous powder dispersions are being used in automotive
clear coats.
GENERAL REFERENCES
Bate, D. A., The Science of Powder Coatings, 2 vols., Selective Industrial Training Associates, Ltd.,
London, 1990.
Jilek, J. H., Powder Coatings, Federation of Societies for Coatings Technology, Blue Bell, PA, 1991.
Misev, T. E., Powder Coatings Chemistry and Technology, Wiley, New York, 1991.
Misev, T. A., Proc. Int. Conf. Org. Coat., Athens, Greece, 1997, pp. 321 – 335.
REFERENCES
1. Powder coatings, http://specialchem4coatings.com.
2. Phillips, P., Coat. World, 2004, 9(10), 24.
3. Gribble, P. R., Proc. Int. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
1998, p. 218.
4. Richert, D. S., in Kirk– Othmer Encyclopedia of Chemical Science and Technology, 3rd. ed.,
Wiley, New York, 1982, Vol. 19, p. 11.
5. Kapilow, L.; Samuel, R., J. Coat. Technol., 1987, 59(750), 39.
6. Fedtke, M.; et al., Polym. Bull., 1993, 31, 429.
7. Misev, T. A., Powder Coatings Chemistry and Technology, Wiley, New York, 1991, p. 48.
8. Fischer, G. C.; McKinney, L. M., J. Coat. Technol., 1988, 60(762), 39.
9. Merfold, G.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2004,
Paper 38.
10. Subramanian, R.; Sullivan, C. J., Proc. Waterborne High-Solids Powder Coat. Symp., New
Orleans, LA, 2001, pp. 381 – 389.
11. Shah, N. B.; Nicholl, E. G., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 2003, pp. 217 – 230.
12. Kronberger, K.; et al., J. Oil Colour Chem. Assoc., 1991, 74, 405.
13. Boogaerts, L.; et al., Proc. Waterborne High-Solids Coat. Symp., New Orleans, LA, 2000,
pp. 504 – 515.
14. Moens, L.; et al., U.S. patent 6,720,079 (2004).
572
POWDER COATINGS
15. Pappas, S. P.; Urruti, E. H., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA,
1986, p. 146.
16. Witzeman, J. S., Prog. Org. Coat., 1996, 27, 269.
17. Rawlins J. W.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2002, pp. 357 – 369.
18. Engbert, Th.; et al., Farbe Lack, 1996, 102, 51.
19. Grenda, W.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2005,
pp. 270 –281.
20. Spyron, E.; et al., U.S. patent appl., 20040219367.
21. Jacobs, W.; et al., Prog. Org. Coat., 1996, 29, 127.
22. Higginbottom, H. P.; et al., Prog. Org. Coat., 1992, 20, 301.
23. Higginbottom, H. P.; et al., Polym. Mater. Sci. Eng., 1992, 67, 175.
24. Taylor, J. W.; et al., J. Coat. Technol., 1995, 67(846), 43.
25. Yeates, S. G.; et al., J. Coat. Technol., 1996, 68(861), 107.
26. Barkac, K. A.; et al., U.S. patent 6,670,043 (2003).
27. Kenny, J. C.; et al., J. Coat. Technol., 1996, 68(855), 35.
28. Agawa, T.; Dumain, E. D., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 1997, p. 342.
29. Green, C. D., Paint Coat. Ind., 1995, September, 45.
30. Flosbach, C.; et al., Prog. Org. Coat., 2003, 48, 177.
31. van der Linde, R.; Scholtens, B. J. R., Proc. 6th Annu. Int. Conf. Cross-Linked Polym.,
Noordwijk, The Netherlands, June 1992, p. 131.
32. Popa, P. J.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1999,
pp. 383 –398.
33. Thames, S. F.; Rawlins, J. W., Powder Coat., 1996, 7(6), 19.
34. Misev, T. A.; van der Linde, R., Proc. Int. Conf. Org. Coat., Athens, Greece, 1997.
35. Rawlins, J. W.; Thames, S. F., Proc. Waterborne High-Solids Powder Coat. Symp., New
Orleans, LA, 2000, pp. 185 – 200.
36. Moens, L.; et al., U.S. patent 6,790,876 (2004).
37. Udding-Louwrier, S.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 1999, pp. 420 –430.
38. Twigt, F.; Van Der Linde, R., U.S. patent 5,703,198 (1997).
39. Löwenhielm, P.; et al., Prog. Org. Coat., 2005, 54, 269.
40. Blanton, M.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1999,
pp. 399 –412.
41. Glass, T.; Depoy, J., Soc. Manuf. Eng. Finish. Conf., FC91– 384, Cincinnati, OH, September
1991.
42. Sugama, T.; et al., Prog. Org. Coat., 1995, 25, 205.
43. Nakamichi, T.; Mashita, M., Powder Coat., 1984, 6(2), 2.
44. de Lange, P. G., J. Coat. Technol., 1984, 56(717), 23.
45. Hannon, M. J.; et al., J. Coat. Technol., 1976, 48(621), 42.
46. Hajas, J.; Juckel, H., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
1999, pp. 273 – 283.
47. Wulf, M.; et al., Prog. Org. Coat., 2000, 38, 59.
48. Gherlone, L.; et al., Proc. Int. Conf. Org. Coat., 1997, p. 128.
49. Belder, E. G.; et al., Prog. Org. Coat., 2001, 42, 142.
REFERENCES
573
50. Maxwell, B. E.; et al., Prog. Org. Coat., 2001, 43, 158.
51. Jahromi, S.; et al., Prog. Org. Coat., 2003, 48, 183.
52. Richart, D. S., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1995,
p. 1.
53. Lee, S. S.; et al., Prog. Org. Coat., 2003, 45, 266.
54. Schmidhauser, J.; Havard, J., Proc. Int. Waterborne High-Solids Powder Coat. Symp.,
New Orleans, LA, 1998, pp. 391 – 404.
55. Dumain, E.; et al., U.S. patent 6,310,139 (2001).
56. Maxwell, L. D.; et al., FSCT Annual Meeting, 2004.
57. Koop, P. M., Powder Coat., 1995, 6(2), 58.
58. Satoh, H.; et al., Proc. Int. Conf. Org. Coat., Athens, Greece, 1997, p. 381.
59. Barth, G. H.; Sun, S. T., Anal. Chem., 1989, 143.
60. Huang, Z.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1997,
p. 328.
61. Dick, R. J.; et al., J. Coat. Technol., 1994, 66(831), 23.
62. Kia, S. F.; et al., J. Coat. Technol., 1997, 69(875), 23.
63. Meyer, E. F., III, Polym. Mater. Sci. Eng., 1992, 67, 220.
64. Hughes, J. F., Electrostatic Spraying of Powder Coatings, Research Studies Press, Letchworth,
England, 1984.
65. Hughes, J. F., Ytbehandlingdagar, Stockholm, Sweden, May 1992, p. A-1.
66. Tech. Bull. SC:586-82, Shell Chemical Co., Houston, TX, 1989.
67. Bauch, H., Polym. Mater. Sci. Eng., 1992, 67, 344.
68. Yanagida, K.; et al., J. Coat. Technol., 1996, 68(859), 47.
69. Jilek, J. A., Powder Coatings, Federation of Societies for Coatings Technology, Blue Bell,
PA, 1991, p. 15.
70. Graziano, F. D., Proc. Int. Conf. Org. Coat., Athens, Greece, 1997, p. 139.
71. Misev, T. A., Powder Coatings Chemistry and Technology, Wiley, New York, 1991,
pp. 347 – 349.
72. Leivo, E.; et al., Prog. Org. Coat., 2004, 49, 69.
73. Maty, J., Paint Coat. Ind., 1988, January, 24.
29
Radiation Cure Coatings
Radiation cure coatings cross-link by reactions initiated by radiation, rather than heat.
Such coatings have the potential advantage of being indefinitely stable when stored in
the absence of radiation; following application, cross-linking occurs rapidly at ambient
temperature on exposure to radiation. Rapid cure at ambient temperature is particularly
significant for heat sensitive substrates, including paper, some plastics, and wood. See
Section 28.1.6 for a discussion of UV cure powder coatings.
Two classes of radiation cure coatings are (1) UV cure coatings, in which the initial step
is excitation of a photoinitiator (or photosensitizer) by absorption of photons of UV –
visible electromagnetic radiation, and (2) EB (electron beam) cure coatings, in which
the initial step is ionization and excitation of the coating resins by high energy electrons.
Cross-linking is initiated by reactive intermediates that are generated from the photoexcited photoinitiator in UV curing and from excited and ionized resins in EB curing.
Infrared and microwave radiation are also used to cure coatings, but these systems are
not included here, since the radiation is converted to heat, which initiates thermal curing.
While the physical volume of radiation cure materials used in the United States is relatively small, perhaps 50,000 metric tons per year, their economic importance is disproportionate to their volume. They are essential to the production of computer chips, optical
fibers, printed circuit boards, and printing plates, and they are used in a variety of other
economically important applications, including dental and medical applications.
29.1. UV CURING
Two classes of polymerization reactions are used in UV curing: free radical and
cation-initiated chain-growth polymerization. Although there have been attempts to use
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
574
29.1. UV CURING
575
photoreactions in which the radiation leads to generation of a reactive functional group,
this approach is not useful for coatings. In such reactions, each photon absorbed can
effect only one cross-linking reaction. In chain reactions, absorption of a single photon
can lead to the formation of many cross-links.
A key requirement for UV curing is a UV source that produces high intensity UV radiation at low cost without generating excessive infrared radiation. Major radiation sources
in commercial use are medium pressure mercury vapor lamps. Such electrode lamps are
tubes up to 2 m long; power outputs of 80 W cm21 are in wide use; and lamps with outputs
up to 325 W cm21 are available. The radiation has continuous wavelength distribution
with major peaks at 254, 313, 366, and 405 nm, among others. Visible radiation and a
minor but not insignificant amount of infrared radiation that causes some heating is also
emitted. The wavelength distribution of radiation from a medium pressure mercury
vapor electrode lamp is shown in Figure 29.1.
Radiation is emitted in all directions around tubular lamps, and its intensity drops off with
the square of the distance from the source. To increase the efficiency of absorption, lamps
are mounted in an elliptical reflector with a focal length such that the maximum intensity of
radiation is focused at the distance between the lamp and the coated surface being cured. A
limitation of UV curing is that the distance between the lamp and the coating on various
parts of the object being coated must be fairly uniform. Hence, UV curing is most easily
applicable to coating flat sheets or webs that can be moved under the UV lamps or cylindrical objects that can be rotated under or in front of the lamps. Since thermal energy is also
produced, the lamp housing must be water and air cooled. UV radiation is hazardous and can
lead to severe burns. It is essential to avoid exposure of eyes to the radiation. The lamps
are housed in enclosures; when the enclosure is opened, the current to the lamps is turned off
automatically. Depending on the radiation source, a greater or lesser amount of ozone is
generated. Since ozone is toxic, the UV unit must be ventilated. (These and other safety
considerations in radiation curing are reviewed in Ref. [1].)
Figure 29.1. Energy distribution of radiation from a medium pressure mercury vapor lamp. (From U.S.
patent 3,650,699, as cited in Ref. [2], with permission.)
576
RADIATION CURE COATINGS
In some cases, there are advantages of using other wavelength distributions, especially
those with increased fractions of radiation in the very near UV – visible region. Changes in
distribution can be made by doping the lamp with traces of other elements besides mercury
or by having fluorescent coatings on the lamp tube that absorb short UV and emit longer
UV radiation.
Electrodeless lamps, powered by microwaves, also enjoy substantial commercial use.
Electrodeless lamps are more suitable for doping, since the lifetime of lamp electrodes is
generally reduced by dopants. Electrodeless lamps have the further advantage of essentially instantaneous start-up and restart. On the other hand, electrodeless lamps are
more expensive. (For further discussion of UV sources, see Ref. [2].) The utilization of
DC power supplies for both electrode and electrodeless lamps has been shown to result
in constant radiation, which is reported to enhance the efficiency of polymerization relative to pulsed radiation [3].
Excimer lamps emit very narrow bands of UV [4,5]. They use silent electrical discharge through two quartz tubes and an enclosed gas volume. Electronically activated
molecules are produced in the gas phase and decompose within nanoseconds to
produce photons of high selectivity. The narrowness of distribution is comparable to
that of lasers, but in contrast to laser sources, the radiation is incoherent and can therefore
be used for large area applications. Different excimer lamps are available or in commercial development that have principal emissions at 172, 222, 308, and 351 nm. The high
intensity of very narrow wavelength distribution can provide rapid cures. The 172 nm
excimer lamps are reported to activate acrylates directly [5]. Xenon chloride excimer
lamps that produce energy at 308 nm have been shown to effect UV cure in shorter
times than do mercury vapor lamps at equal photoinitiator concentration or at equal
speed with lower initiator concentration [6]. The 308 nm excimer lamps are reported to
be particularly effective for cationic curing using triarylsulfonium salt photoinitiators
that absorb in that region [7].
For initiation via any lamp source, there must be absorption of radiation by the photoinitiator (or, as discussed later, by some substance that leads to the generation of an
initiator). The fraction of radiation absorbed within a coating film, IA/I0, is related to
the molar absorptivity 1, concentration of photoinitiator C, and optical path length of radiation in the film X. Assuming that no other absorber is present and neglecting surface
reflection, the fraction of radiation of a given wavelength absorbed is expressed as
IA
¼ 1 101CX
I0
(29:1)
The molar absorptivity 1 can be concentration dependent, so the range of concentration
over which the equation is valid may be limited. Molar absorptivity varies with wavelength, so the fraction of radiation absorbed also varies with wavelength. Also, the intensity of radiation from the source varies as a function of wavelength. The total number of
photons absorbed per unit time depends on the combination of these factors. When a
photoinitiator absorbs a photon, it is raised to an excited state, which leads to generation
of an initiating species. But there are also other possible fates of the excited state photoinitiator. It may emit energy of a longer wavelength; that is, it may fluoresce or phosphoresce. It may be quenched by some component of the coating or by oxygen. It may
undergo other reactions besides those that lead to initiator generation. The efficiency of
generation of initiating species is an important factor in the selection of a photoinitiator.
29.1. UV CURING
577
Whereas most photoinitiators absorb a single photon, efficient two-photon absorbing
photoinitiators have been developed which can generate free radical [8] or cationic
[9] initiator species. Such photoinitiators are particularly useful for three-dimensional
applications, including three-dimensional microfabrication and optical data storage
systems.
The rate of polymerization reactions is related to the concentration of initiating radicals
or ions. It would therefore seem that the higher the initiator concentration, the faster the
curing. As one increases from very low concentrations to somewhat higher concentrations,
the rate of cure increases. However, there is an optimum concentration. The rate of cure in
the lower part of the film decreases above this concentration. If the concentration is high
enough, such a large fraction of the radiation is absorbed in the upper few micrometers of
the film that little radiation reaches the lower layers. Since the half-life of free radicals is
short, they must be generated within a few nanometers of the depth in the film where they
are to initiate polymerization. Although the half-life of the acid species in cationic cure
coatings is substantially longer, migration may be limited by diffusional constraints as
polymerization and cross-linking proceed.
Optimum photoinitiator concentration is dependent on film thickness: The greater the
film thickness, the lower the optimum concentration. Lower concentration is favorable
from the standpoint of cost, since the photoinitiator is generally an expensive component.
However, comparing the cure speeds of films of different thicknesses, each containing the
optimum concentration of photoinitiator for its thickness, the time required to cure a thick
film is longer than to cure a thin film. This follows since less radiation is absorbed in any
volume element with increasing film thickness. The problem is further exacerbated when
surface cure is oxygen inhibited, as is the case with free radical polymerization. (Oxygen
inhibition is discussed in Section 29.2.3.) In general, one should determine the concentration of photoinitiator just sufficient to give the required extent of cure at the surface;
this concentration gives the maximum rate of cure for that film thickness of that system
at the lowest possible cost.
The problem of achieving both surface and through cure can be ameliorated by using a
photoinitiator or a mixture of photoinitiators having two absorption maximums with distinctly different molar absorptivities near different emission bands of the UV source. The
emission band that is highly absorbed by the photoinitiator(s) is absorbed more strongly
near the surface, and less UV is available for absorption in the lower layers. This band
is most important for counteracting oxygen inhibition (Section 29.2.3) but does not contribute substantially to through cure. The weaker absorption of a second emission band is
more uniform throughout the film to provide through cure.
Another factor that affects absorption of UV by a photoinitiator is the presence of competitive absorbers or materials that scatter UV radiation. In designing vehicles for UV cure
systems, it is desirable to minimize their absorption of UV in the range needed for excitation of the photoinitiator. Pigments can cause problems by absorbing and/or scattering
UV radiation; the effects of pigmentation are discussed in Section 29.4.
Although film thickness is one variable that affects path length, it is not the only one. If
the same coating is applied over a black substrate and over a highly reflective metal substrate, the rate of cure over the metal substrate is close to twice as fast as over the black
substrate. UV radiation that passes through the film to the black substrate is absorbed,
but that reaching a smooth metal substrate is reflected and passes through the film
twice; the path length is twice the film thickness, so there is almost twice the opportunity
for absorption.
578
RADIATION CURE COATINGS
29.2. FREE RADICAL INITIATED UV CURE
In coatings cured by free radical initiated UV, free radicals are photogenerated and initiate
polymerization by adding to vinyl double bonds, primarily acrylates. Two classes of
photoinitiators are used: those that undergo unimolecular bond cleavage and those that
undergo bimolecular hydrogen abstraction from some other molecule.
29.2.1. Unimolecular Photoinitiators
A range of unimolecular photoinitiators has been studied [10]. The first ones used on a
large scale commercially were ethers of benzoin. Benzoin ethers undergo cleavage to
form benzoyl and benzyl ether radicals:
Both these radicals can initiate polymerization of acrylate monomers [11]. There is
doubt whether the benzyl ether radical can initiate the polymerization of styrene [10].
The package stability of UV cure coatings containing such benzoin ethers tends to be
limited. Apparently, this is due to the ease of abstraction of the hydrogen on the benzyl
ether carbon. Any organic material contains traces of hydroperoxides that slowly decompose. The resulting radicals can abstract the benzylic hydrogen, leading to initiation of
polymerization and hence poor package stability. Package stability is improved if the
benzylic carbon is fully substituted. Accordingly, the ketal, 2,2-dimethoxy-2-phenylacetophenone, is an effective photoinitiator with good package stability. Photocleavage
produces benzoyl and dimethoxybenzyl radicals. The latter is a sluggish initiator; however,
it can undergo further cleavage to the highly reactive methyl radical; the extent of this
cleavage increases with increasing temperature.
2,2-Dialkyl-2-hydroxyacetophenones (a-hydroxyacetophenones) are also commercially
important photoinitiators with good package stability, including 2-hydroxy-2-methyl1-propan-1-one (HMMP) shown below where R55R155CH3, and 1 benzoylcyclohexanol,
where R and R1 form a cyclohexane ring. These photoinitiators tend to give less yellowing
than do phenyl-substituted acetophenones, including benzoin ethers as well as the ketal
above, probably because benzylic radicals are not generated by photocleavage.
29.2. FREE RADICAL INITIATED UV CURE
579
All these photoinitiators have an acetophenone chromophore, the absorptivity of which is
enhanced by electron donating substituents on the benzoyl ring. This is exemplified by the
morpholino-substituted photoinitiator 2-dimethylamino-2-benzyl-1-(4-morpholinophenyl)butan-1-one, which is recommended for pigmented coatings [12,13]. Yellowing can be a
problem in clear coatings.
Acylphosphine oxides are another class of unimolecular photoinitiators. Several acylphosphine oxides are available: for example, diphenyl-2,4,5-trimethylbenzoylphosphine oxide.
Irradiation generates the corresponding benzoyl and phosphinyl radicals. Acylphosphine
oxides tend to be nonyellowing and give good package stability.
Bis(2,6-dimethoxybenzoyl)-2,4,4-trimethylpentylphosphine oxide (BAPO) on irradiation
is reported to generate four radicals that can initiate polymerization [14]. The strong absorption of near UV–visible radiation makes possible curing of relatively thick films of white
pigmented coatings. Although BAPO is yellow, photocleavage results in reduced absorptivity
(bleaching), so BAPO can be used in white coatings. The absorption spectrum can be
broadened by using a blend with HMMP, providing combined surface and through cure at
lower cost.
29.2.2. Bimolecular Photoinitiators
Photoexcited benzophenone and related diarylketones, such as xanthone and thioxanthones, do not cleave to give free radicals, but can abstract hydrogens from a hydrogen
donor to yield free radicals that initiate polymerization. Thioxanthones such as 2isopropylthioxanthone are used when their high absorption in the very near UV is desirable to permit absorption in the presence of pigments or dyes that absorb UV strongly
in the longer wavelength UV range.
Widely used hydrogen donors are tertiary amines with hydrogens on a carbon
atoms such as 2-(dimethylamino)ethanol (DMAE) and in printing inks, methyl
p-(dimethylamino)benzoate. It has been shown that the accompanying ketyl free radical
does not initiate polymerization. Use of N-phenylphthalimide as a photosensitizer with
isopropylthioxanthone and N-methyl-N,N-diethanolamine accelerates photocuring of
580
RADIATION CURE COATINGS
acrylates at least twofold [15]:
An advantage of bimolecular initiators with amine co-initiators is reduced oxygen inhibition, discussed in the next section. A disadvantage is that excited states of these initiators
are generally longer lived than those of unimolecular photoinitiators and therefore more
readily quenched by oxygen as well as by vinyl monomers with lower triplet energies.
N-Alkylmaleimides have been shown to be photoinitiators for the polymerization of
acrylates. The effectiveness is enhanced by using benzophenone as a photosensitizer
and an amine. For example, a combination of N-methylmaleimide, benzophenone, and
methyldiethanolamine in 1,6-hexanediol diacrylate gave essentially the same exotherm
and peak time as 2,2-dimethoxy-2-phenylacetophenone, and substantially higher than
benzophenone and amine. However, the maleimides are toxic [16].
29.2.3. Oxygen Inhibition
Oxygen inhibits free radical polymerizations. In coatings, this inhibition is particularly
troublesome, since coating films have such a high ratio of surface area, where oxygen
exposure is high, to total volume. Oxygen reacts with the terminal free radical on a propagating molecule to form a peroxy free radical. The peroxy free radical does not
readily add to another monomer molecule; thus, the growth of the chain is terminated.
The terminal radical from methyl methacrylate has been shown to have a rate constant
for reaction with oxygen that is 106 times that for reaction with another monomer
molecule. Furthermore, the excited states of certain photoinitiators are quenched by
oxygen, thereby reducing the efficiency of generation of free radicals.
Several approaches are available to minimize this problem. Curing can be done in an
inert atmosphere, but this is relatively expensive. The most economical inert atmosphere
approach is to use a CO2 atmosphere. CO2 is relatively inexpensive and its high density
makes it easier to contain. It permits reduction in photoinitiator concentration and use
of slower curing reactants [17].
High intensity UV sources minimize but do not eliminate the problem with fast curing
systems. In effect, free radicals can be generated so rapidly that their high concentration
can deplete the oxygen at the film surface, permitting other radicals to carry on the
polymerization before more oxygen diffuses to the surface.
A formulating solution is to incorporate paraffin wax in the coating. As the coating is
applied and cured, a layer of wax comes to the surface of the coating, shielding the surface
from oxygen. Although effective, the residual wax detracts from the appearance of the film
and makes recoating difficult.
Another common approach is to add a chain transfer agent, including tertiary amines,
having a-CH groups; alkyl ethers, having a-CH groups; and thiols. Although peroxy free
radicals are poor initiators, they readily abstract H atoms from such chain transfer agents,
and the resulting a-amino, a-ether, or thiyl radical can initiate a new growing polymer
chain.
29.2. FREE RADICAL INITIATED UV CURE
581
Free radicals on carbon atoms alpha to amines and ethers also react rapidly with
oxygen. Thus, amines and ethers not only act as chain transfer agents but also serve to
deplete oxygen levels. Accordingly, benzophenone – amine initiating systems substantially reduce oxygen inhibition. Oxygen inhibition can also be reduced using unimolecular
photoinitiators by adding small amounts of an amine. This process is useful when lowintensity lamps or short exposure times are employed.
29.2.4. Vehicles for Free Radical Initiated UV Cure
In the first UV cure coatings, styrene solutions of unsaturated polyesters (Section 17.3)
were used along with benzoin ether photoinitiators. However, styrene is sufficiently volatile that a significant amount evaporates between application and curing. Furthermore, the
rate of polymerization is relatively slow compared to that of acrylate systems. However,
the cost is low, and they continue to find some applications.
Most current coatings use acrylated reactants. Acrylate rather than methacrylate esters
are used, since acrylates cure more rapidly at room temperature; they are also less oxygen
inhibited. Furthermore, polymerization of acrylates tends to terminate by combination,
whereas methacrylate polymerization terminates largely by disproportionation. The
extent of cross-linking and higher molecular weights (MW) are favored when termination
of growing radicals occurs by combination.
In general, the vehicle consists of three types of acrylate esters: multifunctional
acrylate-terminated oligomers, multifunctional acrylate monomers, and acrylate monomers. The monomers range from mono- to hexafunctional. Many formulas include mixtures of mono-, di-, and trifunctional acrylates. The monomers are also called reactive
diluents. Multifunctional oligomers contribute to high rates of cross-linking, owing to
their polyfunctionality, and in large measure control the properties of the final coating,
due to the effect of the backbone structure on such properties as abrasion resistance, flexibility, and adhesion. Their viscosity is too high alone and monomers are required to reduce
viscosity for application. Multifunctional acrylate monomers also give fast cross-linking,
owing to polyfunctionality, but have lower viscosities than oligomers. Viscosity is
lowered further by most monofunctional acrylates, but the rate of cross-linking is reduced.
Quantitative kinetic studies of photopolymerization can be carried out by methods such
as photo-differential scanning calorimetry [4] and real-time infrared spectroscopy. Conventional infrared and Raman spectroscopy are also valuable. Kinetic studies furnish
useful insights, but because of the large number of variables involved, a complete
quantitative description of UV curing processes is difficult to construct, especially with
practical formulations containing three types of acrylates.
It would be desirable to react all of the acrylate groups during photopolymerization to
eliminate residual double bonds that might cause undesirable long-term changes in film
properties. In practice, this is usually not possible, but it is often possible to polymerize
a large fraction of the acrylate groups. The extent of conversion of the double bonds is
affected by the choice of components; in general terms, extent of conversion increases
with an increase in monofunctional monomer, presumably because the small monomer
molecules can diffuse through the film during the reaction. Extent of conversion is also
affected by free volume availability; in general, conversion is increased by use of components that give low Tg films. As polymerization proceeds, Tg of the binder increases.
If, as is commonly the case, Tg approaches the temperature at which curing is being
carried out, the rate of reactions slows. Reactions become very slow at temperatures
582
RADIATION CURE COATINGS
only a little above Tg ¼ Tcure. Since there is heat from the radiation source and a reaction
exotherm, Tcure is somewhat above ambient temperature. (See Ref. [4] for a review of
cross-linking rates and extents of conversion of various systems.)
Acrylated oligomers are prepared from a variety of starting oligomers. The properties
of films made from the acrylated oligomers are affected by the average number of acrylic
double bonds per molecule, MW, and by the structure of the “core” oligomer. For example,
acrylated urethane oligomers tend to give coatings with a good combination of hardness
and elasticity. Any polyol or hydroxy-terminated oligomer can be reacted with excess
diisocyanate (OCN22R022NCO) to yield an isocyanate-terminated oligomer. This oligomer can then be reacted with hydroxyethyl acrylate at ambient or moderately elevated
temperature to yield an acrylated urethane oligomer. It has been shown that the gloss
retention on exposure to UV of UV cure coating films decreases as the MW of the diol
from which the urethane diacrylate is prepared increases. It is proposed that this reflects
a higher cross-link density with the lower MW oligomers [18].
Another route is to react oxirane groups of epoxy resins with acrylic acid. The
ring-opening reaction yields the acrylic ester and a hydroxyl group as shown. Like conventional epoxy resins, the acrylated epoxies tend to give coatings with good toughness,
chemical resistance, and adhesion.
Various catalysts (e.g., triphenyl phosphine) are used, so the reaction is carried out at as
low a temperature as possible. Care is required to avoid polymerization of the acrylic acid
or esters during the process. Inhibitors are added to trap free radicals. Some inhibitors,
notably phenolic antioxidants, are effective only in the presence of oxygen, so the reaction
is commonly carried out under an atmosphere of air mixed with inert gas. Variations in
reaction conditions and catalyst composition can result in significant differences in the
product. The most widely used epoxy resin is the standard liquid BPA epoxy resin
(n ¼ 0.13), yielding predominantly the acrylated diglycidyl ether of BPA. Epoxidized
soybean or linseed oil also react with acrylic acid to give lower Tg oligomers with
higher functionality.
Acrylated melamine – formaldehyde (MF) resins have been prepared by reacting etherified MF resins with acrylamide. These resins have the potential to cure in two ways: by
UV curing through the acrylate double bonds, and thermally by residual alkoxymethylol
groups on the MF resin. It is reported that UV curing followed by thermal curing provides
films with increased hardness, improved stain resistance, and improved durability [19].
Many multifunctional acrylate monomers have been used; examples are
trimethylolpropane triacrylate, pentaerythritol triacrylate, 1,6-hexanediol diacrylate, and
29.2. FREE RADICAL INITIATED UV CURE
583
tripropyleneglycol diacrylate. Care must be used in handling them because many are skin
irritants and some are sensitizers [1]. A range of monofunctional acrylates has been used.
Those with lowest MW tend to reduce viscosity most effectively, but they may be too volatile. Ethylhexyl acrylate has sufficiently low volatility. Ethoxyethoxyethyl acrylate, isobornyl acrylate, 2-carboxyethyl acrylate, and others are also used. Small amounts of
acrylic acid as a co-monomer promote adhesion. 2-Hydroxyethyl acrylate has low volatility, high reactivity, and imparts low viscosities, but its toxic hazard is too great in
many applications. N-Vinylpyrrolidone (NVP) is an example of a non-acrylate monomer
that copolymerizes with acrylates at speeds comparable to that of acrylate polymerization;
NVP is particularly useful because the amide structure promotes adhesion to metal and
reduces oxygen inhibition; but it also introduces a possible toxic hazard.
Decker found that monomers with carbamate, oxazolidine, or carbonate ester groups in
addition to the acrylate functionality give faster curing (by as much as 100-fold) and more
complete conversion than did simple acrylates [12]. Addition of a reactive acrylated
carbamate monomer to more slowly polymerizing traditional acrylate monomers dramatically increased overall reactivity. Factors affecting the inherent reactivity of such
monomers have been investigated intensively [13,20,21]. Three theories have been
advanced: (1) hydrogen abstraction/chain transfer that would essentially increase the
monomers’ functionalities, (2) hydrogen bonding, and (3) electronic and resonance
effects. A recent study of monomers with methyl groups blocking potential sites for hydrogen abstraction [20] provided evidence in favor of theory 1 as the dominant factor,
although theories 2 and 3 are not excluded.
C6F13(CH2)2OCOCH55CH2 as a monomer in an epoxy –acrylate oligomer as a UV
cure coating for wood paneling provides high hydro- and oleo-phobicity, improved
chemical resistance, and better scratch resistance than do coatings without the fluorine
monomer [22].
The combination of unsaturated polyesters with vinyl ether resins has been reported to
UV cure at rates comparable with those of acrylates [23]. It is also reported that 1-propenyl
ethers can be UV cured at rapid rates. They are made by isomerizing the corresponding
allyl ethers [24]. Maleated resins [25] as well as maleimides [4] can also be used for copolymerization with vinyl ethers. Maleimide – vinyl ether compositions have also been
reported to cure at rates comparable to those of acrylates without the use of photoinitiators.
This is possible because maleimides absorb in the 300 to 310 nm region, so the monomer
serves as its own photoinitiator. Initiating free radicals are reported to arise from
hydrogen-abstraction by the photoexcited maleimide. These compositions are particularly
efficient with high intensity lamps that emit primarily at 308 nm, where the maleimides
have peak absorbance [4].
Other UV cure copolymer compositions are amine – enes [26] and thiol – enes.
However, applications have been limited: in the former case by relatively slow cure
rates, and in the latter by the odor of thiols. Recent years have seen a resurgence of interest
in thiol –enes because “they exhibit all of the advantages of typical photopolymerizations . . . and have excellent mechanical properties. In addition, thiol –ene polymerizations
have the advantages of being relatively uninhibited by oxygen, exhibit delayed gelation,
and enable radical polymerization of a wide range of thiol and vinyl functional group
chemistries” [27]. Typical thiol –ene formulas contain a tri- or tetra-functional thiol, an
ene (a compound with two or more C55C double bonds), and a photoinitiator. Vinyl
ethers and highly reactive norbornene derivatives are especially useful ene components
[28]. For example, a mixture of trimethylolpropane tri-3-mercaptopropionate with
584
RADIATION CURE COATINGS
pentaerythritol trivinyl ether can be UV cured in air with very little oxygen inhibition.
Other vinyl ether –thiol systems can be cured in air by sunlight [29].
Propagation of thiol – ene cure involves addition of a RS. radical to the ene to form a
carbon centered radical followed by chain transfer to RSH to form another RS. radical.
The process follows step-growth kinetics rather than the chain growth kinetics of conventional photopolymerization.
29.2.5. Waterborne UV Cure Coatings
Waterborne UV cure coatings are reported to have several advantages over conventional
UV coatings. Since the resin is dispersed in water, no reactive diluent is needed. There is
no emission of volatile, possibly toxic monomer. The viscosity of the coating is independent of MW, so that low viscosity for spraying is obtained by adjusting the solids with
water. The high MW means that fewer reactions of double bonds are needed, hence the
film shrinkage is much lower, which in turn improves adhesion to substrates such as
metal. Reaction of fewer double bonds is also expected to result in reduced restricted
shrinkage and therefore less internal stress in the cured coatings. The main drawback is
that the water must be flashed off before exposure to the UV source.
The water is evaporated in an oven at 808C, followed by UV curing; the higher temperature leads to faster, more complete cure. Since the dispersions are stable at high
dilution, overspray caught in a water wash spray booth can be recovered by ultrafiltration.
It is reported that it is not even necessary to add further photoinitiator to the recovered
coating. Furthermore, the UV cure dispersion can be blended with other waterborne coatings, thus permitting enhanced exterior durability [30,31]. Acrylated PUDs have been used
as vehicles for waterborne UV cure coating on wood paneling. Extractables after curing
are very low, due to the absence of unreacted low MW reactants. Adhesion is reported
to be excellent due to reduced shrinkage during cure due to the lower ratio of double
bonds compared to that of conventional UV cure coatings. Low gloss coatings can be
made by addition of silica pigment to the formula since the evaporation of water carries
pigment to the film surface, and shrinkage due to water evaporation provides a low
gloss surface. After application, but before cure, the water is removed by force drying,
which means that nonvolatile photoinitiators must be used. 1-Benzoyl hydroxycyclohexyl
phenyl ketone was reported to give excellent results. Water soluble photoinitiators are also
available. The heating required to evaporate the water has the desirable side effect of
increasing conversion, reducing the residual double bonds [32].
Decker and co-workers studied ways to speed up UV curing of water-based acrylic –
PUD coatings [33]. They recommend curing the coating while it is still hot following
water removal; otherwise, mobility restrictions will retard cure. Another important
factor is selection of the base needed to stabilize the PUD; volatile tertiary amines are
recommended; most of these amines presumably evaporate during water flash-off.
Nonionically stabilized acrylated PUDs do not require an amine to stabilize the
dispersion [34].
Waterborne UV cure coatings based on acrylated polyesters have been disclosed. An
emulsifying polyester and a water insoluble polyester, both acrylated, are mixed, neutralized with aqueous ammonia, and dispersed in water. A photoinitiator (HMMP) is added.
The coating is applied to wood, flashed off, and UV cured. The films obtained have excellent properties, and the grain of the wood is not raised [35].
29.3. CATIONIC UV CURE
585
29.3. CATIONIC UV CURE
Photoinitiators for cationic polymerization are typically onium salts of very strong acids
[4,36,37]. Iodonium and sulfonium salts of hexafluoroantimonic and hexafluorophosphoric acids are examples. Irradiation of diaryliodonium and triarylsulfonium salts yields
strong protic acids of the corresponding counter anions, as well as radical cations; both
initiate cationic polymerization. Biscumyliodonium tetrakis(pentafluorophenyl)borate is
reported to give substantially faster cure rates than obtained with other onium salts [38].
Onium salts can also be utilized as photoinitiators for free radical polymerization as
well as for concurrent cationic/free radical polymerization, since free radical species
are also formed in their photolysis. These reactions are shown for a triphenylsulfonium
salt (anion omitted) following the primary unimolecular bond cleavage.
Diaryliodonium and triarylsulfonium salts absorb radiation only weakly above 350 nm;
however, their spectral response can be extended into the near UV – visible as well as into
the mid-visible range by use of photosensitizers [17,37]. Anthracene-bound sulfonium
salts exhibit substantially higher reactivity than bimolecular anthracene-sensititized sulfonium salts in epoxy-based UV cure coatings and photoresists, which has been attributed to
greater efficiency of intramolecular sensitization as mobility diminishes in such compositions [39].
Reported to be particularly effective are 308-nm excimer lamps using triarylsulfonium
salts as photoinitiators [7]. Diphenyliodonium salts have limited solubility and are toxic.
Substituted derivatives such as bisdodecylphenyliodonium hexafluoroantimonate have
both higher solubility and lower toxicity [38].
29.3.1. Vehicles for Cationic UV Cure
Homopolymerization of oxirane groups (equations in Section 13.3.6) is the major type of
cationic polymerization used commercially. The counterions must be very weak nucleophiles: in other words, anions of very strong acids. Hexafluoroantimonates, hexafluorophosphates, and the recently introduced borate derivatives are particularly effective. In
contrast to free radicals, cations do not react with each other. Consequently, in the
absence of nucleophilic anions, cation initiated cross-linking can continue after exposure
to the radiation source until the reactive cations become immobilized. Reactions with
water and alcohols terminate polymer growth. Nevertheless, cross-linking occurs even
if only one pair of epoxy groups reacts before reaction with water; furthermore, the termination reaction is accompanied by regeneration of a proton, corresponding to chain transfer.
The absence of oxygen inhibition further distinguishes cationic from free radical
polymerization.
586
RADIATION CURE COATINGS
Bisphenol A (BPA) epoxy resins react slowly at ambient temperatures. Higher
temperatures promote the reaction; optimum cure temperatures are about 70 to 808C.
Rapid curing of coatings based on BPA epoxy resins can be effected by using
a combination of UV and infrared sources. Cycloaliphatic epoxides, such as
3,4-epoxycyclohexylmethyl-30 ,40 -epoxycyclohexane carboxylate show higher reactivity,
which results from the added ring strain due to the fused epoxy ring system that promotes
ring opening of the oxonium ion during the propagation step. Their low viscosity makes
such epoxides useful as reactive diluents. Epoxy-functional silicone polymers have been
recommended for use in UV cure release coatings [38].
Vinyl ethers and styrenes, particularly p-alkoxystyrenes, rapidly polymerize cationically. Cationic photopolymerization of vinyl ethers is more rapid than polymerization
of epoxy-functional reactants [4,36]. Less photoinitiator is required; moreover, multifunctional vinyl ether monomers are said to show a very low order of toxicity. Vinyl ether
monomers are interesting in their own right and are also useful as a highly reactive component in epoxy resin coatings. An example of a divinylether monomer, reported to have a
very high rate of polymerization, is the divinylether derived from the reaction product of
chloroethyl vinyl ether with bisphenol A [36]:
Cationic UV cure coatings have been reported using vinyl ether oligomers, prepared
from toluene diisocyanate/polyethylene glycol prepolymers terminated with 4-hydroxybutylvinyl ether together with ethylene glycol divinyl ether as reactive diluent, a diaryliodonium tetrakis(pentafluorophenyl)borate salt as initiator and isopropylthioxanthone as
photosensitizer [40].
Tensile strength of films from a vinyl ether –terminated poly(dimethyl)siloxane/TDI
prepolymer were higher than with propylene glycol and the cure rate was more rapid. 1Propenyl ether compounds are reported to cure rapidly, cationically, using triarylsulfonium hexafluoroantimonate salts [24].
29.4. EFFECTS OF PIGMENTATION
Since many pigments absorb and/or scatter UV radiation, they generally inhibit UV curing
to some degree [41]. Scattering leads to reflection back out of the film, and absorption also
reduces UV availability to the photoinitiator. The effect becomes more serious as film
thickness increases, since the pigment plus the photoinitiator reduce the amount of UV
that can reach the lower layers of the film. With a strongly absorbing pigment such as
carbon black, the films that can be UV cured are limited to 1 to 2 mm in thickness.
Hence, printing inks pigmented with carbon black can be UV cured, but not black coatings.
The most widely used pigment in all other types of coatings is rutile TiO2. However, rutile
TiO2 absorbs some violet light and essentially all but the very nearest UV, even at quite thin
film thickness. Rutile white coatings with good hiding can be cured using a blend of 25%
BAPO and 75% HMMP (Section 29.2.1) [14]. To permit flow, low levels of pigmentation
are used (PVC of 6), requiring film thicknesses of 50 to 100 mm for hiding. Anatase TiO2
does not absorb as strongly in the near UV, so that thicker films can be UV cured than
with rutile, but hiding by anatase is not as efficient.
29.4. EFFECTS OF PIGMENTATION
587
Equations have been developed that permit calculation of the fraction of UV of each
wavelength that will be absorbed by photoinitiator in the presence of pigments that
absorb and scatter UV [42]. The equations permit calculation of the total fraction of radiation absorbed as a function of wavelength, and perhaps more important, permit calculation of absorption in the bottom layer of films of different film thicknesses. The
database needed and the calculations would be very extensive, so it appears more practical
to make model calculations using the equations to illustrate the effect of variables and to
guide formulators considering selection of pigments, photoinitiators, and their concentrations. Figure 29.2 illustrates the effect of concentration of photoinitiator on absorption
of UV by photoinitiator in the bottom 0.1 mm of 15 mm films. The calculations were based
on assumptions intended to illustrate the effect of 20 PVC of rutile TiO2 in the film. The
photosensitizer absorption was based on 2-chlorothioxanthone (CTX). With the assumptions used, the appropriate concentration of CTX for this system would be about 0.33%.
Curing is favored by using UV sources with higher contents of very near UV radiation
and by using photoinitiators such as CTX with higher absorption coefficients in the very
near UV wavelengths. As discussed previously, there is an optimum photoinitiator concentration that decreases as film thickness increases. Pigmented coatings are more sensitive than unpigmented coatings to this effect. The effect of reflection by the substrate can
be critical in determining the film thickness that can be cured.
A further problem of pigmentation in curing arises from the large differential of absorption of UV at the surface and at the bottom of the films, which can result in wrinkling
Figure 29.2. Absorption by CTX of 375 nm radiation in the bottom 0.1 mm of 15 mm films with 20 PVC
TiO2 over an 85% reflectance substrate as a function of CTX concentration. (From Ref. [41], with
permission.)
588
RADIATION CURE COATINGS
(Section 24.6). If the surface layer of a film cures while the bottom layer is still fluid, then
when the bottom layer cures it shrinks, causing the top layer to wrinkle. This effect is
particularly likely to be seen with free radical cure done in an inert atmosphere or with
cationic cure. In both cases, there is no inhibition of the surface cure by oxygen.
Pigmentation also affects flow properties. UV cure coatings are generally applied
without solvent, which means that the volume of pigment in the wet coating is almost
as high as the PVC of the final film. Without solvent, viscosity tends to be higher than
desirable for good flow in the short time between application and cure. Even at 20
PVC, there is a significant increase in viscosity due to the presence of the pigment. By
using sufficient monomer in place of oligomer, the viscosity can be reduced to application
viscosity, but this reduces the curing speed. As one tries to make higher PVC coatings, the
problem becomes more difficult. As PVC approaches CPVC, viscosity of a wet coating
approaches infinity. The problem is compounded by the difficulty of stabilizing the
pigment dispersion against flocculation with the relatively low MW vehicle (Section
21.1.3). Flow problems are just as serious for cationic systems as for free radical
systems. However, pigment flow problems are expected to be alleviated in waterborne
UV cure compositions (Section 29.2.5), which are applied as dispersions.
For many wood finish applications, transparent low gloss coatings are desired, but they
are difficult to achieve with radiation cure coatings. Low gloss is attained in transparent lacquers for finishing furniture and paneling by adding low concentrations of small particle
size SiO2. When the lacquer is applied, the solvent evaporates, setting up convection currents in the drying lacquer film. The SiO2 particles are carried up to the surface of the film
by these convection currents. As the viscosity of the surface increases, the SiO2 particles are
trapped at the surface. The resulting high PVC at the surface gives low gloss. In UV cure
coatings, there is no solvent to evaporate and therefore no mechanism to concentrate the
pigment at the surface. If the level of pigmentation is increased to the extent necessary
to provide low gloss, the viscosity is too high for application. As discussed in Section
29.2.5, low gloss transparent wood coatings are possible with waterborne UV cure.
Some progress has been made in reducing gloss by use of a dual UV cure method. A
UV cure coating containing as much small particle size SiO2 as viscosity permits and
designed so that its cure is strongly oxygen inhibited is applied and cured in air. The
lower layer of the coating cures preferentially. As the lower layer cures, it shrinks, exerting
an unbalanced stress on the pigment particles in the upper parts of the film, forcing them
toward the surface. An alternative explanation is that reactive monomer or oligomer
undergoes net migration into the lower polymerizing region, thereby concentrating
pigment in the upper nonpolymerizing region. Then the coating is UV cured again, this
time under an inert atmosphere. Now, the surface layer cures, resulting in further shrinkage and further increasing PVC near the surface. In this way, medium gloss transparent
coatings can be made. Several production lines have operated using this dual cure method.
29.5. ELECTRON BEAM CURE COATINGS
High-energy electron beams (EB) can be used to polymerize acrylate coatings. In EB cure
compositions, the high energy electrons cause direct excitation of the coatings resins (P) as
well as their ionization into radical cations (Pþ.) and secondary electrons (e2).
29.6. DUAL UV/THERMAL CURE AND SELECTED APPLICATIONS
589
The major fate of the radical cations is recombination with the relatively low energy
secondary electrons, which can yield additional excited state resins. The excited state
resins (P ) primarily undergo homolytic bond cleavage to free radicals that initiate
polymerization of the acrylated resins.
The high energy electrons are generated by charging a tungsten filament at a high negative potential, 150 to 300 keV. The electrons are directed by magnets in a curtain through a
metal “window” to the coatings to be cured. (The term electron curtain is also used.) The
types of vehicles used are the same as those used in free radical –initiated UV curing.
Since the polymerization reactions are oxygen inhibited and there is little differential in
absorption between the top and the bottom of the film, it is essential to do EB curing in
inert atmospheres.
Cationic polymerization of vinyl ethers [4,43], as well as highly reactive epoxides [44]
can be induced by EB radiation in the presence of diaryliodonium and triarylsulfonium
salts. The readily reducible onium salts have been implicated in the generation of cationic
initiators by oxidation of free radical species as well as by capture of secondary electrons.
Scavenging of secondary electrons by onium salts tends to lengthen the lifetime of cations
and to improve the prospects for cationic polymerization [43].
The principal advantages of EB curing over UV curing are that no photoinitiator is
needed and that pigments do not interfere with the curing. These advantages are often
offset by the higher capital cost of the electron beam or curtain generating equipment,
the need to use inert atmospheres, and the need for shielding to protect workers from
the electron beam. The advantage of being able to cure pigmented systems is real but is
of limited importance in coatings, since the flow problem is not alleviated by EB curing.
29.6. DUAL UV/THERMAL CURE AND SELECTED APPLICATIONS
UV curing is generally limited to flat sheets or moving webs or cylindrical objects, which
can be rotated in front of the radiation source, to achieve uniform cure. However, dual
UV/thermal cure compositions have been developed for multi-shaped substrates, such
as an auto body, to provide thermal cure in areas, called shadow areas, that are not
exposed directly to the radiation. An investigation of various approaches to UV curing
on automobiles has been described [45]. A dual UV/thermal cure coating, exposed
with a combination of two fluorescent lamps (one with maximum intensity at 313 nm,
the other at 351 nm) and a xenon flash lamp, followed by thermal curing for 30 minutes
at 708C, gave excellent film properties. The rapid UV cure minimizes sagging, and the
subsequent thermal cure provides cure in the shadow areas. A related UV/thermal dual
cure study of auto clear coats utilized a free radical photoinitiator/acrylated resin composition blended with an alkyd– MF resin composition and reported good gloss and mar
resistance of the cured coating [46].
Dual UV/thermal cure compositions have been reported, based on thermal redox generation of radicals, so that the same functionality, such as acrylate groups, can be used for
both UV and thermal cure [47]. Examples include through cure of black coatings. Dual
UV/thermal cure waterborne coatings have also been reported. Compositions containing
carbamate-functional acrylated PUDs together with MF resins can be UV cured followed
by thermal cross-linking of the carbamate groups with the MF resins [48].
Researchers are actively seeking photolatent catalysts that generate acids or tertiary
amines when exposed to UV [49]. Such catalysts are useful in dual UV/thermal cure
590
RADIATION CURE COATINGS
compositions. For example, a photolatent catalyst that releases strong acid can promote
acid catalyzed thermal cure, such as with MF resins, avoiding package stability problems.
In combination with conventional photoinitiators, the UV exposure step not only effects
UV curing, but also generates acid for the subsequent thermal step. Such acid precursors,
known as photoacid generators (PAGs), are commonly precursors of sulfonic acids such
as oxime sulfonates [50]. PAGs are used widely in the electronics industry for photoimaging. Photolatent catalysts that release amines can extend the pot life of 2K polyurethanes to a full working day, of substantial value in auto refinish shops and other
applications [49].
UV cure printing inks, particularly lithographic inks, are widely used. The viscosity of
lithographic inks is relatively high, on the order of 5 to 10 Pa.s, so there is not a flow
problem. The films are thin, 2 mm or less, so interference of pigment with curing can
be offset by use of highly absorbing photoinitiators and/or high photoinitiator concentrations. Since the films are not continuous, adhesion or curling difficulties resulting
from shrinkage are not a major problem. The advantage of fast curing is large, as it
allows the webs to be wound up quickly without offsetting on the back of the substrate.
In printing of tin plate and aluminum for cans, there is the advantage that the four
colors can be printed in line simply by having a UV lamp after each printing station; operating costs are lower than with heat cure inks. Ink jet printing is another rapidly growing
market for UV inks [44].
Clear coatings for paper, plastic film, and foil are probably the largest volume market
for radiation cured coatings [44]. The unpigmented coatings cure very fast at low temperatures. The important uses of UV curing for top coating wood furniture and filling of particle board are discussed in Chapter 31. Another large scale application that benefits from
the fast curing at low temperature possible with UV cure coatings is top coating of vinyl
flooring. It is especially applicable to flooring with patterns made by expanded foams,
since the foam pattern is heat sensitive. Acrylated urethane oligomer coatings have the
high abrasion resistance needed in a floor covering; acrylated polyester coatings have
superior stain resistance.
Abrasion resistant clear coatings for plastics have been developed. The poor abrasion
resistance of clear plastics is a distinct disadvantage in applications such as replacing glass
in uses ranging from glazing to eyeglasses to automotive head lamps. Silicone-based coatings have been used for many years to improve the abrasion resistance (Section 4.4.1).
These coatings require long thermal cure schedules and have high VOC levels. Radiation
cure coatings based on trialkoxysilyl-functional acrylate monomers and colloidal
silica cure in seconds and have low VOC emission [51]. They have excellent adhesion
to polycarbonates and other plastics and have abrasion resistance equal or superior to
solvent-based coatings [51,52].
An application for cationic UV cure epoxy coatings is for coil coating on tin plate for
the exterior side of can ends. The superior adhesion of epoxy coatings makes them preferable to acrylate coatings for this application. In some cases, a UV cure coating is applied
on one side of the strip and UV cured. Then a thermally cured epoxy coating is applied on
the other side. When the strip is put through the oven to cure the thermally cured coating,
the cross-linking of the UV cure coating is also advanced. More flexible films are obtained
using epoxidized polybutadiene [53].
A large use of UV curing is for coating optical fibers for wave guides used in telecommunication cables. The process is carried out at high speeds in towers up to 25 m tall. The
glass fiber is pulled from a heat softened blank of glass and then coated through a die
29.7. ADVANTAGES AND LIMITATIONS
591
coater, that is, a reservoir with holes on each end through which the glass is pulled. The
UV lamps are aligned parallel to the fiber as it comes out of the coater; elliptical reflectors
are placed both behind the UV source and on the opposite side of the fiber from the source.
The fiber is at the focal point of the pair of reflectors. In this way, the coating is given very
high dosages of UV, permitting cure speeds on the order of 20 m s21 without an excessive
bank of UV lamps. Acrylated urethane oligomers have been used since abrasion resistance
is a critical requirement. Two UV cure coatings, a soft first coat with special optical properties and a harder outer coat, are applied and cured by the time the fast-moving coated
fiber reaches the bottom of the tower [54].
A UV curable optical fiber coating transparent in deep UV has been developed to allow
a Bragg grating to be engraved on the fiber directly by laser exposure through the coating.
The vehicle selected was an acrylate-terminated dimethylsilicone resin, and the initiator
was 2,4,6-trimethylbenzoyldiphenylphosphineoxide.
UV curing is also the basis for important photoimaging processes used in the printing
and electronic industries [55,56]. In these applications, UV curable coatings on selected
substrates are exposed through an image bearing transparency such as a film negative.
Cross-linking occurs selectively under the transparent regions. The unexposed regions
are subsequently removed, commonly by washing out with aqueous or organic solvent.
The resulting image transfer into the coated substrate is a critical step in making printing
plates as well as printed circuit boards and integrated microcircuits (e.g., computer chips).
29.7. ADVANTAGES AND LIMITATIONS
A survey of 130 people involved in radiation curing assessed the advantages, disadvantages, and prospects for the technology [44]. Eighty-nine percent of the respondents
reported that reduced solvent emission is an important motivation for use of radiation
cure. Generally, the formulations contain no solvent and VOC emissions are negligible.
Very short cure times are possible at ambient temperature with package stable coatings.
In the case of clear acrylate compositions, curing times required are fractions of a second.
Since the curing can be done at ambient temperatures, the coatings are applicable to heat
sensitive substrates such as paper, some plastics, and wood. The energy requirement is
minimal. Major losses in energy encountered in thermal curing systems are not experienced. The cure is done at or near to ambient temperature, so no energy is required to
heat the object being coated except in the case of waterborne UV coatings. (The IR radiation from the UV lamps does cause a modest temperature increase.) In addition, heat loss
from air flow through a baking oven, required to keep the solvent concentration below the
lower explosive limit, is eliminated. The extent of this advantage is, of course, dependent
on the cost of energy. The use of radiation curing in Europe and Japan has been greater
than in the United States, at least in part because the cost of energy is higher in Europe
and Japan than in the United States.
Capital cost for UV curing is low. This results primarily from the small size of the
curing units as compared with ovens for baking. A UV cure coating needs only about
0.5 second exposure for curing. A line being run at a speed of 60 m min21 need be
exposed to UV over a distance of only 0.5 m. A UV curing unit with four lamps can do
this in a total length of about 2 m. In contrast, a thermal cure oven has to be 100 or more
meters long to provide time for curing at this speed. In addition, there has to be space for
flash off of at least some of the solvent from conventional coatings before the coated object
592
RADIATION CURE COATINGS
enters the oven and space for it to cool back down to handling temperature. The savings in
building space can thus be large. Half of the end users in the survey report “improved
physical properties/product performance” as a major motivation for using radiation
cure [44].
There are, however, limitations and disadvantages to radiation curing. It is most applicable to flat sheets or webs for which the distance to the UV source or the window of the
EB unit can be approximately constant. Cylindrical or nearly cylindrical objects can be
rotated in front of the radiation source and fibers can be pulled through a circle of
sources, but irregularly shaped objects are not easy to expose uniformly.
In the case of UV curing, pigmentation can limit the thickness of films that can be
cured. In the extreme case of carbon black pigmented films, the limit is about 2 mm. Electron beam cure coatings do not have this limitation because there is no gradient of
initiation from top to bottom of the film. Owing to the effects of pigmentation on flow,
the level of pigmentation is limited in both UV cure and EB cure coatings, excepting
waterborne formulations. Furthermore, low gloss coatings cannot be applied. These limitations can be overcome by using solvent containing coatings, but this eliminates many of
the advantages.
Shrinkage during curing can lead to adhesion problems on metal and some plastic substrates (Section 6.2). When polymerization results from chain addition of free radicals to
double bonds, there is a substantial decrease in volume, since the length of the carbon–
carbon bonds formed is shorter than the intermolecular distance of monomer units. The
degree of shrinkage is related directly to the number of double bonds reacted. In acrylic
UV or EB cure coatings, the potential shrinkage is 5 to 10%. Since cure occurs in less
than a second, there is not time for much of the volume shrinkage to take place before
the free volume in the coating is so limited that movement becomes restricted. The
stress from the restricted shrinkage applies a force to offset the adhesion forces
holding the coating to a substrate so that less external force is required to remove the
coating. Shrinkage can be minimized by using higher MW oligomers or higher ratios
of oligomer to low MW reactive monomers, but these expedients increase viscosity
and tend to increase cure time. Curing of epoxy systems by cationic photoinitiators
leads to less shrinkage. The decrease in volume resulting from the formation of polymers
from monomers is partially offset by an increase in volume from epoxy ring opening.
Typical epoxy UV cure compositions shrink less than 3%. Adhesion of cationic cure coatings can be further enhanced by surface treatment of the substrate [57]. Some of the problems with UV curing can be reduced by using waterborne UV coatings as discussed in
Section 29.2.5.
Porous substrates such as paper and wood give good adhesion, since the mechanical
effect of penetrating into the surface of the substrate holds the coating on the substrate.
However, there can be a curling problem with thin substrates as a result of the tension
from shrinkage on one side of the substrate.
The photoinitiator in free radical initiated coatings is only partly consumed in curing, so
a substantial fraction remains in the films. The photoinitiator can accelerate photodegradation of films exposed outdoors. Conventional UV stabilizers and antioxidants tend to
reduce UV cure rates, which limits their use. These constraints limit the use of UV cure
coatings in the range of exterior applications but do not exclude it. Studies on automotive
clear coats have shown that a UV curable formulation of specified acrylated resins, photoinitiators, UV absorber, and hindered amine light stabilizer (HALS) showed exterior
durability after five years in Florida comparable to that of conventional thermal cure
clear coats [58].
REFERENCES
593
Cationic coatings have the advantage that they do not generate free radicals when films
are exposed outdoors. Electron beam cure coatings have the distinct advantage that there is
no photoinitiator in the system. However, even with EB curing, the coatings used must
have excess unsaturation to cure on rapid schedules, and residual unsaturation can
adversely affect exterior durability. Also, there are trapped free radicals that can react
with oxygen after the cure and generate peroxides. Radiation-cure coatings tend to be
more expensive than conventional coatings, so that commercial applications are primarily
those where UVs cure permit unique advantages.
GENERAL REFERENCES
Belfield, K. D.; Crivello, J. V., Eds., Photoinitiated Polymerization, ACS Symp. Ser. 847, American
Chemical Society, Washington, DC, 2003.
Koleske, J. V., Radiation Curing of Coatings, ASTM International, West Conshohocken, PA, 2002.
Oldring, P. K. T., Ed., Chemistry and Technology of Formulating UV Cure Coatings, Inks, and
Paints, Vols. 1 – 5, SITA Technology, London, 1991– 1994.
Scranton, A. B.; et al., Eds., Photopolymerization Fundamentals and Applications, ACS Symp.
Ser. 673, American Chemical Society, Washington, DC, 1997.
REFERENCES
1. Golden, R., J. Coat. Technol., 1997, 69(871), 83.
2. Bean, A. J., in Radiation Curing: Science and Technology, Pappas, S. P., Ed., Plenum Press,
New York, 1992, p. 308.
3. Okamitsu, J. K.; et al., U.S. patent 6,908,586 (2005).
4. Jonsson, S.; et al., Prog. Org. Coat., 1996, 27, 107.
5. Roth, A., RadTech Rep., 1996, September – October, 21.
6. Sharp, K.; et al., J. Coat. Technol., 1997, 69(865), 77.
7. Braddock, J. K., RadTech ’96 Proc., Vol. 1, p. 478.
8. Cumston, B. H.; Perry, J. W., Nature, 1999, 398, 51.
9. Zhou, W.; et al., Science, 2002, 296, 1102.
10. Hageman, H. J., Prog. Org. Coat., 1985, 13, 123.
11. Carlblom, L. H.; Pappas, S. P., J. Polym. Sci. A Polym. Chem., 1977, 15, 1381.
12. Decker, C.; Moussa, K., J. Coat. Technol., 1993, 65(819), 49.
13. Beckel, E. R.; et al., Macromolecules, 2004, 37, 4062.
14. Rutsch, W.; et al., Prog. Org. Coat., 1996, 27, 227.
15. Cavitt, T. B.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2002, pp. 431 – 446.
16. Hoyle, C. E.; Jonsson, S., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 2000, pp. 16– 29.
17. Studer, K.; et al., Prog. Org. Coat., 2003, 48, 92, 101.
18. http://www.sartomer.com/wpapers/5060.pdf.
19. Gummeson, J. J., J. Coat. Technol., 1990, 62(785), 43.
20. Beckel, E. R.; et al., Macromolecules, 2005, 38, 3093.
21. Kalambi, H.; et al., Polymer, 2005, 46, 4735.
22. Bongiovanni, R.; et al., Prog. Org. Coat., 2002, 45, 359.
594
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
RADIATION CURE COATINGS
Friedlander, C. B.; Diehl, D. A., U.S. patent 5,536,760 (1996).
Artysz, D. M.; et al., Prog. Org. Coat., 2003, 46, 302.
Noren, G. K., Polym. Mater. Sci. Eng., 1996, 74, 321.
Noren, G. K., J. Coat. Technol., 1993, 65(818), 59.
Cramer, N. B.; et al., Macromolecules, 2003, 36, 7964.
Jacobine, A. F.; et al., J. Appl. Polym. Sci., 1992, 45, 471.
Jonsson, S.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2003,
pp. 295 –310.
Gerlitz, M.; Awad, R., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2001, pp. 187 – 194.
Gerlitz, M.; et al., U.S. patent 5,990,192 (1999).
Reich, W.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1999,
pp. 448 –457.
Decker, C.; et al., JCT Res., 2004, 1(2), 127.
Van Den Berg, K. J.; et al., U.S. patent 6,987,135 (2006).
Urbano, E.; et al., U.S. patent 6,265,461 (2001).
Crivello, J. V., J. Coat. Technol., 1991, 63(793), 35.
Pappas, S. P., in Radiation Curing: Science and Technology, Pappas, S. P., Ed., Plenum Press,
New York, 1992, p. 1.
Priou, C.; et al., J. Coat. Technol., 1995, 67(851), 71.
Pappas, S. P.; et al., J. Photochem. Photobiol. A, 2003, 159, 161.
Kayaman-Apohan, N.; et al., Prog. Org. Coat., 2004, 49, 23.
Pappas, S. P.; Wicks, Z. W., Jr., in UV Curing: Science and Technology, Pappas, S. P., Ed.,
Technology Marketing Corp., Norwalk, CT, 1978, Vol. I, pp. 78– 95.
Wicks, Z. W., Jr.; Kühhirt, W., J. Paint Technol., 1975, 47(610), 49.
Lapin, S. C., in Radiation Curing: Science and Technology, Pappas, S. P., Ed., Plenum,
New York, 1992, p. 241.
Lawson, K., RadTech Rep., 1998, 12(1), 25.
Maag, K.; et al., Prog. Org. Coat., 2000, 40, 93.
Decker, C., Polym. Prepr., 2004, 45(2), 190.
Studer, K.; et al., Prog. Org. Coat., 2005, 53, 126, 134.
Schwalm, R.; Meisenberger, S., U.S. patent 6,916,854 (2005).
Dietliker, K., Spectrum, 2004, 17(3), 12.
Photoacid Generators for Microlithography, tech. bull. e-9/2003 Ciba Specialty Chemicals,
Tarrytown, NY, January 2003.
Blizzard, J. D.; et al., Proc. Waterborne Higher-Solids Powder Coat. Symp., New Orleans, LA,
1992, p. 171.
Lewis, L. N.; Katsamberis, D., J. Appl. Polym. Sci., 1991, 42, 1551.
Cazaux, F.; et al., J. Coat. Technol., 1994, 66(838), 27.
Masson, F.; et al., Prog. Org. Coat., 2004, 49, 1.
Monroe, B. M., in Radiation Curing: Science and Technology, Pappas, S. P., Ed., Plenum Press,
New York, 1992, p. 399.
Funhoff, D.; et al., Prog. Org. Coat., 1992, 20, 289.
Molenaar, F.; et al., Prog. Org. Coat., 1993, 22, 393.
Valet, A., Prog. Org. Coat., 1999, 35, 215.
30
Product Coatings for
Metal Substrates
In 2002, U.S. shipments of product coatings for OEM (original equipment manufacture)
application had a value of $5.26 billion [1]. This was approximately 33% of the value
of all U.S. shipments of coatings. In this and the following chapter on coatings for nonmetallic substrates, we discuss some of the major OEM product coating end uses to illustrate
the factors involved in selecting coatings for particular applications. A large fraction of
industrial coatings is used on metals. Multitudes of products are coated; space permits discussion of only a few. We have selected four of the larger end uses: automotive, appliance,
container, and coil coatings.
30.1. OEM AUTOMOTIVE COATINGS
OEM automotive coatings for cars and trucks are the largest single market for industrial
coatings. The global market for automotive coatings in 2001 was reported to be $6.6
billion; the U.S. market accounted for $2.5 billion [2]. Coatings for repair and refinishing
of cars are discussed in Section 33.3. The discussion here is further limited to coatings
applied to the exterior bodies of automobiles. Coatings for plastics in cars are discussed
in Section 31.2.2. Many other coatings are used on car interiors, wheels, trunk linings,
air filters, and so on. The coating operation is a major part of auto and truck assembly,
occupying about half the space and using more than half the capital investment in the
entire assembly plant.
There are two major reasons for coating an automobile: appearance and corrosion
protection. Initial appearance can be a critical factor in the sale of a car; the coating is
the first thing a potential buyer sees. The purchaser can be attracted by the good appearance
of the coating; furthermore, if the coating is poor, the purchaser may assume that the
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
595
596
PRODUCT COATINGS FOR METAL SUBSTRATES
manufacturer is careless not only in applying the coating, but also in other aspects of
manufacturing the automobile. Maintaining the excellent appearance of the coating over
many years of service is also critical. If the coating does not continue to look good, not
only may the owner of the car be unhappy but others may decide not to buy that make of
car because they have seen that the coating does not stand up well. High-gloss coatings
are used, and gloss retention is a critical requirement. Corrosion protection is another critical
performance requirement. The use of salt on icy highways provides an environment that can
lead to rapid corrosion of steel. In the 1970s and 1980s, coatings were developed that
together with improved steel, provide protection against corrosion for the lifetime of the car.
The steel used in car bodies may be cold-rolled steel; it may have been electrogalvanized or hot-dip galvanized; or it may be a special alloy that includes zinc. Before
coating the car, body parts are fabricated and assembled, including the doors, hood, and
trunk lid. After careful cleaning to remove dirt and oil, the body is chemically conversion
coated (pretreated) as detailed in Section 6.4.1. For steel, the conversion coating process
deposits a mesh 1 to 5 mm thick, of zinc phosphate and zinc– iron phosphate crystals on
the surface. Modern conversion coatings are complex formulas, often containing manganese in addition to zinc. Different conversion coatings may be needed for zinc coated steel
and alloys. The car body is then carefully cleaned and rinsed; the last rinse water contains a
dilute solution of chromic acid having the primary purpose of preventing the onset of
rusting during the short time before painting begins.
30.1.1. Electrodeposition Coatings and Primer– Surfacers
Virtually all of the steel vehicles in the world are primed with cationic E-coat as discussed
in Section 27.2. E-coat primers provide excellent adhesion to metal and resist displacement by water. The strong driving force of the electrophoretic application is thought to
aid penetration into the phosphate crystal mesh on the metal surface. The amine groups
on the binder resin provide strong interaction with the phosphate coating and the metal
surface. The binders are designed to be nonsaponifiable; even if the coating is scraped
through, undercutting will be very slow, and the coating will continue to protect the
metal next to the gouge against corrosion. The E-coat covers all metal surfaces, including
recessed areas, such as the interior of rocker panels below the doors that cannot be coated
by spray application. The process provides a relatively uniform film thickness on all metal
surfaces of about 25 mm of dry film.
Although E-coat primers provide excellent corrosion protection, there are three problems
involved in their use. (1) They lack good UV resistance. If the top coat allows UV to reach the
E-coat, degradation occurs at the E-coat surface, destroying intercoat adhesion and leading to
delamination. Subsequent coats must prevent UV from reaching the E-coat surface. (2) The
film thickness can be so uniform that surface irregularities in the metal and conversion
coating are copied in the surface of the primer film. (3) It can be difficult to achieve the
required adhesion of top coat to the E-coat surface. E-coat primers have been formulated
that provide greater smoothness (the term filling is often applied) by reducing the pigment
content of the primer. The viscosity of the uncured primer film is lower, and when the
primed car body is baked, more leveling can occur than if the PVC were higher. Although
this aids in obtaining surface smoothness, adhesion of top coat to the smooth, glossy
E-coat surface is more difficult, making achieving chip resistance more difficult.
There have been many efforts to develop E-coats that can be coated directly with top
coats, but most auto assembly plants use an intermediate coating variously called a
30.1. OEM AUTOMOTIVE COATINGS
597
primer surfacer, a tie coat, or just a primer. In some cases, there is an additional antichip
coating in areas such as the rocker panels and nose of the hood.
Three classes of primer – surfacers are currently used. Worldwide, in 2004, these
coatings were 73.9% solventborne, 19.5% waterborne, and 6.6% powder [3].
An example of a binder in a chip resistant solventborne primer –surfacer is a one
package urethane polyester coating, utilizing a blocked isocyanate [4]. In some cases, a
two package (2K) polyurethane primer –surfacer is used. The primer –surfacer has a
lower cross-link density than the E-coat primer, providing greater opportunity for the
solvent in the top coat to penetrate into this surface than into the surface of the E-coat
primer. The primer – surfacer is formulated with a higher PVC, so that surface roughness
provides a greater opportunity for top coat adhesion. Also, a highly pigmented primer –
surfacer can be relatively easily sanded to flatten defects such as dirt. The primer film
is usually thick enough to avoid sanding through to bare metal, but if this happens, the
spot is repaired with catalyzed coatings containing passivating pigment such as strontium
chromate. An example of a waterborne chip resistant primer is a urethane acrylic, based on
blocked isocyanates and water-reducible acrylic resins [5]. The mechanical properties
involved in stone chip resistance and approaches to testing are discussed in Ref. [6].
Powder primer– surfacers are being used over E-coat on some cars to reduce VOC
emissions. They also provide improved chip resistance. Adhesion to the E-coat surface
is presumably obtained by the high baking temperature, 1508C, used with the powder
primer– surfacer. The greatest interest has been in epoxy-functional acrylics made with
glycidyl methacrylate (GMA) (Section 13.1.2) as a co-monomer and cross-linked with
dicarboxylic acids such as dodecanedioic acid [HOOC(CH2)10COOH] [7] or a carboxylic
acid –functional resin [8]. An epoxy-functional acrylic for automotive primer– surfacer is
n below 2500, a calculated Tg above 808C, and a monomer composaid to require an M
sition such that melt viscosity is less than 40 Pa.s at 1508C [9]. Such a resin can be
made with 15 to 35% GMA and 5 to 15% butyl methacrylate (BMA), with the balance
being methyl methacrylate (MMA) and styrene.
A current trend is toward use of color key primers; the colors are picked for use under a
group of top coats with related colors. Use of several different color primers is a significant
expense but can improve the appearance when top coats have thin areas with less than 100%
hiding. Color key primers also make stone chip damage less noticeable since chipping to the
primer results in little or no color change. A further advantage is that colored primers can be
sprayed under the hood, in the trunk, and on the inside of doors. Then a relatively thin layer
of the more expensive top coat gives a satisfactory appearance. The primer must have
good exterior durability to compensate for thin spots in the top coat or excessive loss of
UV absorbers from the top coat on baking. Otherwise, UV radiation will degrade the
surface of the primer, causing loss of adhesion of top coat to primer and, ultimately, to
delamination. UV absorbers and HALS stabilizers (Section 5.2.) can be added to the
primer– surfacer to improve its UV resistance. A UV absorber in the base coat and top
coat that strongly absorbs UV in the wavelength range 290 to 310 nm also helps protect
the primer from degradation. Partial migration of UVA and HALS among the coats is to
be expected.
30.1.2. Top Coats
The primary purpose of top coats is appearance. Top coats are high gloss and must
maintain their appearance for long periods. While the primary factors in maintaining
598
PRODUCT COATINGS FOR METAL SUBSTRATES
appearance are resistance of the resin binder to photoxidation and hydrolysis, durability of
the pigment in the film is also a critical factor (Chapter 5). There are many other requirements, including resistance to scratching, to acid precipitation (rain, fog, and dew), to bird
droppings, to sudden thunder showers on a car that has been sitting in the hot sun, to the
impact of pieces of gravel striking the car, to gasoline spillage, and so on. Until the early
1980s almost all top coats were monocoats, a single coating composition applied in
several coats. Monocoats have been largely supplanted by base coat –clear coat
systems; a base coat containing the color pigments covered by a transparent coating.
Base coat – clear coat systems provide better gloss and gloss retention than monocoats.
In both monocoat and base coat– clear coat systems, a majority of top coats are metallic
or other polychromatic colors. As discussed in Sections 19.4 and 20.2.5, metallic coatings
give an attractive appearance, due to the change in color with the angle of viewing. When
viewed at an angle near the perpendicular, the color is light, and when viewed at larger
angles, the color is darker. The extent of this phenomenon is called the degree of color
flop, or just flop. High color flop depends on three factors: minimal light scattering by
the coating matrix between the aluminum flakes; a smooth, high gloss surface; and
orientation of the aluminum flakes parallel to the surface. Other special effect top coats
are pigmented with interference pigments such as pearlescent and cholesteric pigments.
Minimal light scattering by the coating matrix requires that pigment selection and
dispersion provide transparent films when prepared in the absence of the effect
pigment. Most dispersions are made with media mills. Nano color pigments for automotive metallic coatings are prepared by grinding suspensions of the conventional color
pigments in butyl acetate with a commercial dispersing agent using very small balls in
the mill. The mill base is then used to make tinting pastes with excellent transparency
for color matching metallic colors [10].
Also, all resins and additives in the coating must be compatible so that there is no
haziness in the unpigmented film. The final surface should be as smooth as possible.
Surface unevenness can be affected by many factors. Surface roughness can result from
roughness of the initial metal substrate and phosphate layer [11]. Rough primer– surfacers
can result from dry spray on difficult to paint areas such as the verticals on or below the
deck lid. Leveling of orange peel after spraying can be difficult to avoid in vertical areas,
where relatively dry spray is used to avoid sagging. Door jambs can be particularly
difficult and are the part most noticed when you get into a car. Overspray can result in
poor leveling.
Orientation of aluminum flakes parallel to the surface has been achieved in everyday
production for decades, but it has never been entirely clear why orientation is as good
as it is [12,13]. Most workers have concluded that an important factor is shrinkage of
the film after application; shrinkage is accompanied by increasing viscosity of the film
most rapidly near the surface as the solvent evaporates, leading to a viscosity gradient.
This viscosity gradient causes the upper edge of an aluminum platelet to be immobilized
before the lower edge, and the platelet swivels toward parallel orientation as the film
shrinks. In general, the lower the solids of the coating as it arrives at the substrate, the
better the alignment will be. Some workers think that spray droplets spread out when
they strike the surface, and the resulting flow forces tend to align the flake particles parallel
to the surface. It has also been suggested that the spray droplets must penetrate the surface
of the wet film and crash on the substrate [13]. The rapid drop in viscosity as a high-solids
clear coat goes into the oven can cause flow, resulting in some loss of orientation. These
and other factors may influence orientation. Changes in atomization, air flow, solvent
30.1. OEM AUTOMOTIVE COATINGS
599
evaporation rates, and gun to surface distance during spraying can cause substantial
variations in flake alignment [14]. Transfer efficiency is greatly improved by electrostatic
spraying; however, it is more difficult to obtain good surface smoothness and metal orientation using electrostatic spray. It has been suggested that there may be some alignment of
the metal flakes parallel to the lines of force in the electrostatic field that are perpendicular
to the substrate. Use of automatic high-speed rotary bell spray guns further complicates the
problem of obtaining a uniform appearance.
Most monocoats are formulated with thermosetting acrylic and melamine–formaldehyde
(MF) resins. To achieve high gloss, pigmentation levels are low. In solid colors, PVCs of 8 to
9% are used; in the case of metallic colors, PVCs are much lower—on the order of 2 to 4%.
Owing to the low pigment content, relatively thick films (on the order of 50 mm) are required
for hiding. Since the coatings are spray applied, some areas have significantly over 50 mm of
dry coating: for example, around the A post, between the windshield and the front door
openings. Higher PVCs are possible with base coats because there is no requirement for
gloss. Relatively thin (12 to 20 mm) base coat films can provide adequate hiding [15].
Until recently, the highest solids at which it has been possible to formulate high
performance solventborne monocoats is about 45 NVV. This has also been near the
upper limit of solids that provides outstanding exterior durability. If the molecular weight
(MW) of the acrylic resin is reduced further to increase the solids above 45 NVV, the
average number of functional groups per molecule and MW get so low that the fraction
of molecules with single functional groups becomes too large to permit good properties
with thermosetting acrylic resins synthesized by free radical polymerization (Section 8.2.1).
Clear coats can be sprayed at somewhat higher solids than monocoats. As discussed
in Section 24.3, as resin solids increase, the problems of controlling sagging increase.
Automotive coatings are particularly vulnerable to sagging, since the film thickness
applied is large and variable. As solids have been increased, it has become necessary to
add thixotropic agents to increase the viscosity at low shear rates to minimize sagging
[16,17]. Conventional thixotropic agents make a coating hazy due to light scattering,
thereby reducing gloss. Therefore, viscosity modifiers have been designed with refractive
indexes similar to those of the acrylic binder. Those most widely used are acrylic microgels, which are highly swollen gel particles that are lightly cross-linked so that they can
swell, but not dissolve, in the liquid coating (Section 24.3). The microgels make the
coating thixotropic (i.e., the viscosity decreases with the high shear going through the
spray gun, stays low long enough for leveling, but then increases, reducing sagging).
The mechanism of their action is not fully understood, but swelling of the particles and
flocculation are thought to play a role [15,16].
Application of a clear coat over a pigmented base coat gives higher gloss. Historically,
clear coats were not used in automotive finishes because of the cost of an extra coating step
and because the exterior durability of available clear coatings was inadequate. However,
with development of better binders and light stabilizers, especially combinations of HALS
and UV absorbers (Section 5.2), clear coats now have long-term exterior durability. Total
combined film thickness is only slightly greater than that of a single coat system. The base
coat contains roughly twice the PVC of a monocoat. Thus, a base coat with a dry film about
12 to 20 mm thick, depending on the color, has about the same hiding as a 50 mm monocoat. The clear coat thickness should be about 40 to 50 mm, but considerable variation is
experienced in production. Thin (less than 30 to 35 mm) areas appear satisfactory but have
inferior durability. Properly formulated and applied base coat – clear coat systems have
high gloss and long-term gloss retention. The advantage is particularly marked for metallic
600
PRODUCT COATINGS FOR METAL SUBSTRATES
coatings. The same overall VOC emission can be achieved using a lower solids base coat
and a higher solids clear coat. The lower solids in the base coat permits better orientation
of the aluminum in the coating and hence superior color flop. The higher level of pigmentation leads to a somewhat rougher surface of the base coat than of a single coat, but this is
covered by the clear coating. Waterborne base coats generally give superior metal
orientation, since they are applied at 20 to 25 NVV.
Although it would be possible to bake the car body after application of the base coat,
then apply the clear coat, and bake the car again, it is more economical to apply the clear
over a wet base coat and bake the overall system just once. Part of the solvent must evaporate from the base coat film before clear coat is sprayed over it; otherwise, the force of the
spray of the clear coat would distort the base coat. On the other hand, if too much solvent
has evaporated, intercoat adhesion may be adversely affected. The solvent must be
permitted to flash off until distortion will not occur, and then the top coat is applied.
The flash-off time required is about 2 minutes but varies with spray booth ventilation
conditions as well as coating formulations. In some systems, to permit application of
the clear coat in 2 minutes, it is necessary to incorporate a small amount of wax, zinc
stearate, or cellulose acetobutyrate in the base coat formulation [18]. The additive apparently orients at the surface and minimizes distortion of the layer. Care must be exercised,
since excess additive could interfere with intercoat adhesion.
Worldwide, in 2004, 57.7% of base coats were solventborne and 42.3% were waterborne. In the United States about 35% were waterborne; in Germany almost 90% were
waterborne [3].
Solventborne base coat binders are often thermosetting acrylic resins with MF crosslinkers. It has been reported that hydroxy-functional urethane modified polyesters are
also useful in base coats [18].
Published papers on waterborne base coats mention MF cross-linkers with waterreducible acrylics, water-reducible polyester – polyurethanes, and acrylic latexes
[19 – 21]. Use of acrylic latexes appears to be gaining favor as experience is gained in
designing the complex combinations of rheological characteristics needed for application
and flake orientation [12,20]. Waterborne base coats require a dehydration bake in which
most of the water and some of the solvent is driven off before application of the top coat.
This process must be carefully controlled. If too much water remains in the film, it will
blow out during the clear coat bake resulting in popping. If the base coat is overbaked,
it may become porous and adsorb clear coat solvent, affecting flow of the clear coat
and again, leading to popping.
The lower film thickness of the base coat and the flash-off time required before applying
the clear top coat reduce the popping problem (Sections 24.7 and 26.1) that tends to occur
with waterborne single coats. Some plants apply two thin base coats, which further minimizes the problems of popping and sagging. Control of sagging during application requires
that the waterborne base coat be shear thinning, which also reduces surface distortion during
subsequent application of the clear coat. The low solids (15 to 20 NVV) of waterborne base
coats facilitates aluminum orientation. Opaque colors can be applied at 25 to 35 NVV.
Waterborne base coats are formulated for application at relative humidity (RH) of about
60%, since it is less expensive to increase RH than to decrease it. Acrylic–polyurethane
dispersions (PUDs) are in wide use for formulating base coats. For example, base coats
are formulated with an acrylic–urethane PUD and an MF resin [22] (Section 12.7.2).
Water-dispersible carboxymethylcellulose acetobutyrate has been shown to provide
reduced dry-to-touch times, improved metal orientation, reduced sagging due to faster
30.1. OEM AUTOMOTIVE COATINGS
601
viscosity build up, and improved leveling in waterborne base coats [23,24]. Addition of
alkyl esters of 2,4,7,9-tetramethyl-5-decyne-4,7-diol to a waterborne base coat gives
film excellent flow and leveling without orange peel. There is virtually no foaming
[25]. A highly cross-linked copolymer of MMA, BA, ethylene glycol dimethacrylate,
and acrylamide dispersed in water is used as an additive to waterborne base coats. The
acrylic dispersion minimizes strike in by solvent from solventborne clear coat, which
results in nonuniformity; it also provides excellent aluminum orientation [26].
Worldwide, in 2004, clear coats were 97.9% high-solids solvent-borne (of which 82.7%
were 1K coatings and 15.2% were 2K), 1.2% waterborne, and 0.9% powder [3].
Many types of clear coats are being used and others evaluated. In addition to transparency and high gloss, major considerations include acceptable VOC, resistance to environmental etching, mar resistance, and cost. Most clear coats are acrylic resins with various
cross-linkers. The acrylic resin is the major factor controlling VOC content. Considerable
research has been aimed at blending low MW polyester and polyurethane polyol resins
with acrylic resins to increase solids without reducing the MW of acrylic resins below
the level that provides adequate film properties.
Soon after introduction of base coat –clear coats, a serious problem was encountered.
After a few days or weeks in some locations, shallow, unsightly pits developed in the
clear coat. The suspected cause was acid hydrolysis of cross-links in the top coat, so
the coatings were said to have poor acid etch resistance. Since other factors are now
known to contribute to the problem (Section 5.5), the terminology now used is environmental etch resistance [27]. Outdoor tests of environmental etch resistance, performed
near Jacksonville, Florida, give useful but still not very reproducible results. No other
exposure site has been found that gives better reproducibility. A laboratory test has
been developed with cyclic exposure to acidic water, darkness, UV exposure, and humidity exposure [28].
To have good environmental etch resistance, a coating must have good hydrolytic
stability under acidic conditions. The activated ether groups in hydroxy-functional
acrylic – MF cross-linked films are subject to hydrolysis (Section 11.3). Environmental
etch resistance of MF cross-linked coatings can be improved by increasing Tg, which
minimizes penetration of water into the film. Since urethanes are more resistant to hydrolysis under acidic conditions, isocyanate cross-linked acrylics generally have superior
environmental etch resistance. Many urethane clear coats have been two package (2K)
coatings (Section 12.4); however, weighing against the use of 2K coatings is the need
for dual spray systems, concern about toxicity, and cost. Another approach is to use
hydroxy-functional urethanes. For example, a low MW hydroxy-functional urethane
derived from the reaction of a triisocyanate and a diol such as neopentyl glycol has been
reported to give good clear coat properties when cross-linked with MF resin [29].
Moisture-curable trialkoxysilylacrylic resins provide good etch resistance [30,31]. Clear
coats based on a combination of MF resin and trialkoxysilyl cross-linking tend to
have better environmental etch resistance than those based on MF cross-linking alone.
Epoxy-functional acrylics cross-linked with dicarboxylic acids or anhydrides also
provide good environmental etch resistance. Since reduced solubility of water in films
increases environmental etch resistance, highly fluorinated resins (Section 17.1.4)
improve environmental etch resistance but are expensive.
Another approach to improved environmental etch resistance is to use MF resin to
cross-link carbamate-functional resins (Section 11.3.4) [32]. Such clear coats have been
shown to have high mar and acid resistance, exterior durability, and hydrolytic resistance.
602
PRODUCT COATINGS FOR METAL SUBSTRATES
Higher solids are also possible. For example, a coating made with the reaction product of
hydroxypropyl carbamate with an isocyanate prepolymer derived from IPDI, together with
an MF resin and dodecylbenzenesulfonic acid catalyst, has 85 wt% weight solids.
In another study, a carbamate-functional siloxane resin, carbamate-functional acrylic
resin, and MF resin was used in formulating clear coats. The carbamate-functional siloxane
is prepared by reacting a hydroxy-functional siloxane with 1-methoxy-2-propylcarbamate
(see Section 16.1.1.1). The acrylic resin is made by reacting a hydroxypropyl acrylate/
butyl methacrylate copolymer with 1-methoxy-2-propylcarbamate. A dispersion of
colloidal silica in a solution of the carbamate-functional siloxane is added. The resin
composition with nano-pigmentation gives a clear coat with outstanding gloss retention
after weathering and mar/scratch resistance testing [33].
Mar resistance is discussed in Section 4.4.2; Ref. [34] provides a review of mechanical
properties required for mar resistance. Marring can result from automatic car washes,
keys, dry wiping, brushing against tree branches or shrubs, and so on. Mar resistance
requires a hard film that is highly elastic; this combination requires compromises. Films
with low coefficients of friction improve mar resistance.
Various cross-linking reactions are being worked on. Hydroxy-functional acrylic
polymers can be cross-linked with MF resins (Section 11.3), polyisocyanates (Section
12.4), or combinations of the two. Epoxy-functional acrylic resins can be cross-linked
with dicarboxylic acids (Section 13.3.2). Trialkoxysilyl-functional acrylic resins can be
cross-linked by reaction with atmospheric water [35]. In baking automotive clear coats,
methoxysilyl groups of trimethoxysilyl-functional acrylic resins hydrolyze at room temperature to silanol groups, which co-react during baking to cross-link the coating. The
scratch resistance before and after weathering is superior to MF cross-linked coatings.
Such coatings are reported to be widely used [36].
Use of auxiliary cross-linkers such as blocked polyisocyanates and/or MF
resins further enhances the properties obtained with silyl-functional acrylics [31].
3,5-Dimethylpyrazole- or 1,2,4-triazole-blocked polyisocyanates are preferred, since
they permit curing at somewhat lower temperature and do not have the yellowing
problem of methyl ethyl ketone – blocked isocyanates (Section 12.5).
MF cross-linked coatings generally provide excellent mar resistance, but as noted
previously, generally have inadequate environmental etch resistance. Urethane crosslinked coatings have good environmental etch resistance, but poor mar resistance.
Also as noted previously, cross-linking with a combination of urethane and MF and
cross-linking silyl-functional acrylics with auxiliary cross-linkers can provide both
mar and environmental etch resistance. Mar resistance can be further improved by incorporating some fluoroalkyl acrylate in the acrylic resins to decrease the coefficient of
friction. Carboxylic acid cross-linked epoxy-functional acrylics suffer from inferior
mar resistance.
Acrylic resins prepared by controlled free radical polymerization (CFRP) are being
used in automotive clear coats. The solids of the SuperSolids clear coats are reported to
be 65% with the possibility of future coatings with up to 90%. At the same time the
scratch and mar resistance are substantially higher. Du Pont has been awarded the EPA
clean air achievement award for their work [37]. In contrast to conventional free radical
polymerization, CFRP permits preparation of block copolymers so that all molecules of
the polymer have at least two functional groups at low MWs. Also, the MW distribution
is narrower, which permits higher solids at the same number average MW. (See Section
2.2.1.1 for discussion of CFRP.)
30.1. OEM AUTOMOTIVE COATINGS
603
Use of 2K waterborne urethane coatings is growing (Section 12.7.3). In a review of the
status of 2K waterborne urethane clear coats, the advantages of using unmodified polyisocyanates and in line mixing of the two components are emphasized [38].
Compared to the effects of conventional pigments, the use of nano-pigments, especially
nano-fillers, is said to result in improved scratch, abrasion, heat, radiation, and swelling
resistance, decrease in water permeability, and increase in hardness, weatherability,
modulus, and strain to failure while maintaining toughness. These properties reflect the
high surface area of nano-pigments so that there is an increase in adsorption of resin on
the pigment surfaces, hence a larger fraction of immobilized resin. In some cases, the
films have two Tgs. When the particle size of the pigments is smaller than the wavelength
of light, coatings pigmented with nano-pigments are transparent. Nano-sized ceramic
spherical pigments are used in clear coats to reduce marring; the coatings are sometimes
called ceramic paints.
Nano-sized zinc zirconium silicate has been patented for use as a UV absorber in
automotive clear coats. The small particle size does not reduce transparency of the
clear coat. Other advantages are insolubility and nonvolatility [39].
A very desirable research goal is development of UV curable automotive clear coats.
They would have the advantages of low bake temperature and very low VOC emissions.
The rapid low temperature cure can overcome the problem of sagging encountered with
very high solids coatings. UV cure coatings on flat panels can give film properties suitable
for automotive clear coats. A study of the effect of various photoinitiators, UV absorbers,
and hindered amine light stabilizers (HALS) on the curing and exterior durability of
UV-cure automotive clear coats over a white base coat and a blue metallic base coat
has been published. A clear coat formulated with a specified combination of acrylated
resins, photoinitiators, and UV stabilizers, including a HALS derivative, applied over a
blue metallic base coat, showed exterior durability after 5 years in Florida comparable
to conventional thermal cure clear coats, a surprising technical achievement [40].
However, the UV sources available for flat sheet curing cannot provide uniform cure on a
three dimensional object such as an automobile. An investigation of various approaches to
UV curing on automobiles has been described [41]. A dual UV/thermal cure coating
exposed with a combination of two fluorescent lamps (one with maximum intensity
output at 313 nm and the other at 351 nm) and a xenon flash lamp, followed by thermal
curing for 30 minutes at 708C is reported to give excellent film properties. The rapid UV
cure minimizes sagging, and the subsequent thermal cure provides cure in the shadow areas.
Waterborne dual UV/thermal cure clear coats are based on a waterborne coating made
from a resin prepared by reacting a HEA – HDI allophanate, IPDI isocyanurate,
1,6-hexanediol, thioglycolic acid, and TMP urethane with the residual isocyanate
groups blocked with 1,2,4-triazole and neutralized with TEA [41]. A dual cure coating
is also formulated with a PUD, an acrylic latex, and adipic dihydrazide. The PUD is
made from IPDI, HEA, monomethyl ether of a polyethylene glycol, a sulfonate functional
polyester diol, and ethoxylated TMP triacrylate, dispersed in water, and chain extended
with hydrazine. The latex is a copolymer of MMA, MAA, EA, and diacetoneacrylamide.
See Section 17.6 for discussion of cross-linking via diacetoneacrylamide [42].
Very low emissions could also be achieved using powder coatings for exterior automotive coatings (Chapter 28). Powder automobile top coats have been in a development stage
since the late 1960s. In Japan in the early 1980s, powder coatings were used as nonmetallic
monocoats. Major limitations are the need for a separate spray facility for each color (or
very expensive cleanup between color changes) and the very poor color flop of metallic
604
PRODUCT COATINGS FOR METAL SUBSTRATES
powder coatings. These limitations do not apply to clear coatings. Major development
effort is being applied to powder clear coats, and they are used in at least two European
car assembly plants. Fine particle size powder is required to achieve the smooth films
needed for clear coats. For example, a powder coating made with an epoxy-functional
acrylic resin with average particle size of 10 mm with a narrow distribution of particle
size formulated with dodecanedioic acid having a particle size of about 3 mm, gave
good surface smoothness [7]. The small particle size of the dodecanedioic acid is required
for the extruder mix to achieve the desired properties. Special application equipment is
required for application over nonconductive surfaces and to handle the small particle
size [43].
Aqueous dispersions of powder clear coats have also been adopted in one auto plant. An
aqueous powder coating is prepared by dispersing powder obtained by spray drying a
combination of a water-reducible hydroxy- and carboxy-functional acrylic resin, a PUD
made with IPDI having terminal NCO groups blocked with a combination of malonic
ester and methyl ethyl ketoxime, and tris(alkoxycarbonylamino)triazine [44].
Generally, plastic components are coated separately and then installed on the car body.
There would be substantial advantages to a “universal coating” that could be applied to
flexible and rigid plastics and steel in one operation. (The design criteria for such a
coating are discussed in Ref. [45].) However, formidable problems must be overcome
to make universal coatings practical. As discussed in Chapter 31, plastic parts usually
require different coatings than metal parts.
30.1.3. Factory Repair Procedures
During assembly of a car, it is common for the coating to be damaged or to have blemishes
caused by dirt, resulting in need for repair. The wide range of defects that occur during
manufacture and use of automobiles are discussed in an excellent pair of papers [46].
Reducing the frequency and cost of repairs is a major goal in auto production. Some factories have achieved “first-run OK” rates of 90%, but others fall well short of this level.
Repairs are made at several stages in the process: after application of the primer–surfacer
to the car body, after the base coat and clear coat have been applied, after assembly, and
after shipment to the dealer. Once the glass, upholstery, tires, and the like have been
installed, the car can no longer be baked at the temperatures for which regular production
coatings are designed. Either the entire car can be baked to about 808C or the area repaired
can be heated somewhat above 808C with infrared lamps.
When cars were coated with acrylic lacquer, repair was relatively simple, since the thermoplastic systems stay soluble in the solvent in the repair lacquer. However, thermosetting
enamels are more difficult to repair. Achieving adhesion to the surface of the cross-linked
coating is more difficult. Minor blemishes in the clear coat can be corrected by light
sanding and buffing; spot repair of small defects going deeper into the paint is sometimes
possible. Larger damage, however, may require refinishing of the entire panel to which the
damage occurred or even repainting the entire car body. In panel refinishing, the top coat is
removed, any bare metal is primed, and special repair base coat and clear coats are applied
to the entire panel. Since the coating cannot be baked at a high temperature, additional
strong acid catalyst must be added to coatings, which contain MF cross-linkers, to
allow curing at lower temperature. The excess catalyst remains in the film and can lead
to more rapid hydrolysis. Durability of such repairs is good, but not as good as the original
coating. Urethane 2K repair coatings are being used increasingly, since they cure at
30.2. APPLIANCE COATINGS
605
relatively low temperatures without loss in long term durability. It is unlikely that
corrosion resistance of repaired areas is equal to that of unrepaired areas, because the
E-coat is often sanded through during the repair process. As with OEM auto finishes,
dual UV/thermal cure compositions, which have the potential of minimizing sagging
and ensuring cure in shadow areas, are being studied for repair clear coats (Section 29.6).
30.2. APPLIANCE COATINGS
Major markets for OEM appliance coatings are for washing machines, dryers, refrigerators, air conditioners, and ranges. In some cases, single coats are used, but a major
part of the market is for primer – top coat systems. For applications that require corrosion
protection, such as washing machines and air conditioners, cationic E-coat primers
(Section 27.2) are used on the highest quality products. Nonelectrodeposition primers
are frequently applied by flow coating (Section 23.3). To minimize VOC, water-reducible
epoxy ester – based primers are appropriate.
Cationic E-coat epoxy coatings are used for some applications as single coats.
Although epoxy coatings chalk badly on exterior exposure, the drum of a dryer or the
interior of an air conditioner does not get exposed outdoors, but a high degree of corrosion
protection and impact resistance are needed. The uniform coverage of edges attainable
with E-coat permits use of 12 mm E-coats to replace 50 mm solution epoxy coatings on
air conditioners, with a substantial reduction in manpower required for application
while maintaining the necessary performance [47]. White E-coats have been applied to
appliances. Anionic acrylic E-coats can be used on aluminum, but for steel, as discussed
in Section 27.1, discoloration results from iron salt formation. Single cationic E-coats
based on acrylic resins (Section 27.2) avoid this problem on steel.
Thermosetting acrylic coatings are generally used over primers. For single coat
compositions, polyesters are more commonly used because they tend to exhibit better flexibility and adhesion to treated steel or aluminum than do acrylics. The most commonly
used cross-linkers are amino resins. For washers and dishwashers, benzoguanamine–
formaldehyde resins (Section 11.4.1) are used because they impart greater resistance to
alkaline detergents. For other applications, conventional MF resins are used because of
lower cost. In end uses for which performance requirements are not severe, such as
hot water heaters, lower cost semioxidizing alkyd/MF based coatings may be most
appropriate.
Use of powder coatings as top coats for appliance applications has been growing
rapidly and is reported to be the largest single market for thermosetting powder coatings
[48] (Chapter 28). The long runs of single colors make powder coatings a natural choice
for appliances. A limitation is the greater difficulty of achieving good leveling with
powder coatings then with liquid coatings. Whereas some orange peel can be desirable
to conceal metal irregularities, powder coatings may have too much orange peel.
However, in some cases, particularly in Europe, consumers are accustomed to the
orange peel typical of porcelain enamels, and the finish from powder coatings is readily
accepted. Low VOC emissions, low fire risk (with proper precautions in handling
powders), low energy requirements, and reuse of overspray powder are strong economic
and environmental reasons for using powders. Hydroxy-functional polyester or acrylic
resins are used as binders with a blocked isocyanate or tetramethoxymethylglycoluril
(Section 11.4.3) as cross-linker. Alternatively, carboxylic acid – functional resins are used
606
PRODUCT COATINGS FOR METAL SUBSTRATES
together with triglycidyl isocyanurate (Section 13.1.2) or tetra(hydroxyethyl)adipamide
(Section 17.5) as cross-linker.
Another approach to reducing VOC emissions at the appliance manufacturer’s factory
is to use coil coated metal (Section 30.4). The solvent emissions occur at the coil coating
factory, where they can be burned to provide part of the fuel for curing the coatings. A
potential problem with coil coated metal is that the edges are bare. The appliance must
be designed so that cut edges are turned under and protected by a sealant. In some
cases, it is possible to weld through the coating in areas that are not visible.
30.3. CONTAINER COATINGS
Container coatings were historically called metal decorating coatings, since a major
portion of the business was in coating flat sheets, followed by lithographic printing and
a finishing varnish (clear top coat) to protect the ink. There were many uses for such
sheets: metal boxes, trays, wastepaper baskets, bottle caps and crowns, and most important, cans. Plastics have replaced coated metal in most of these applications other than
bottle caps, crowns, and cans. The field is now usually referred to as container or can
coatings. In the United States alone, about 135 billion cans are produced annually, of
which about 100 billion are beverage cans.
Most cans are food or beverage containers, and one of the key requirements is that
there be no possibility of introducing toxic compounds into the foods or beverages. In
the United States, all can linings must be acceptable to the Food and Drug Administration
(FDA) and in the case of meat products, the Department of Agriculture. Contrary to what
many people seem to believe, the FDA does not approve coatings; it lists acceptable ingredients. Rules are published in the Code of Federal Regulations (CFR), Title 21, Part 175.
The most important is 21CFR175.300, which deals with resinous and polymeric coatings.
In most cases if all of the components of a new coating have already been used in can
coatings, a new coating will be acceptable. In some cases it is necessary to prove that
no material is extracted into any food or beverage that will be packed in the can.
However, new raw materials must pass extensive tests. Toxicity considerations affect predominantly interior can coatings, but in some cases there are also restrictions for possible
contamination from exterior can coatings. When metal sheets coated on both sides are
stacked, the exterior coating on one sheet is in direct contact with the interior coating
on the next sheet in the stack. In this configuration, migration of low MW components
between coatings is possible.
In the European Union, food contact legislation for coatings is formally different from
country to country. But European food contact legislation applicable to plastics is used
more and more for can coatings, especially with respect to specific migration limits for
chemical starting compounds.
The effect of coatings on the flavor of the food or beverage packed in the container is
critical. Although flavor requirements are particularly important for interior coatings, care
must also be exercised with coatings for exterior application. Flavor changes can result
from extraction of some contaminant from the coating; absorption of flavor agents from
the packed food into the coating, reducing the flavor of the food; or failure to isolate
the food or beverage from the metal of the can. Flavor can be affected by minute
amounts of substances. To assure that all residual solvent and other volatile flavor detractors are driven out of the coatings, high baking temperatures are used. The only way to
30.3. CONTAINER COATINGS
607
evaluate the effects of coatings on flavor is by making test packs of the food or beverage in
the container and tasting the food or beverage. As a result, major suppliers of can coatings
maintain flavor panels of people trained to taste and, particularly, to use consistent words
to describe flavors.
For beer cans, a major aspect of the flavor problem is to prevent contact between the
beer and the can because metals catalyze flavor changes in beer. For this reason, the
final interior coating is spray applied after formation of the can to avoid potential problems
from breaks in the can lining resulting from stresses during can forming operations.
Linings are spray applied to the interior of soft drink cans not only to protect flavor, but
also to protect the can; the acid present in most soft drinks could eat through the metal
without a coating barrier. An interesting sidelight on history is that in the early days of
packing pineapple products and grapefruit juice in cans, coatings for lining these cans
that would resist their high acidity were not available. They were packed in cans with
heavy-tin-weight linings to protect the steel bodies. Tin affects the flavor of the pineapple
products and grapefruit juice and acts as a bleaching agent, allowing light color fruits and
juices to retain their light color. Even though organic coatings that could be used with
these products are now available, most are still packed in heavily tinned cans. Apparently,
consumers are used to and prefer canned pineapple and grapefruit products that have a
tinny flavor.
There are two major classes of cans: three-piece and two-piece cans. In a three-piece
can, one piece is the body and the other two are the ends. Metal sheets are coated, or
continuous strips of metal are coil coated. Blanks for the can bodies are stamped out of
the coated metal, formed, and sealed into the cylindrical body. The body is formed,
usually into a cylinder, and sealed by soldering, by welding, or with an organic adhesive.
Soldering is now restricted to nonfood cans because of concern about lead toxicity. The
side seam is sprayed with a fast drying coating called a side striper to cover exposed
metal resulting from soldering or welding. Solventborne side stripers predominate;
some waterborne stripers, and for containers for aggressive products, powder coatings,
are used. The coatings are applied while the metal is still hot, assisting in curing of the
coating. The ends are made separately; they are stamped out of coated sheets or coil
coated metal, formed, and equipped with a formed-in-place rubber gasket. One end is
put on by the can maker; the other end is put on after the can has been filled.
Coatings on metal sheets are generally applied by direct roll coating (Section 23.4). In
some cases, transfer rollers with sections cut out are used so that the coating is applied only
in selected areas. For example, the edges that will be soldered or welded at the sides of the
body of the can are commonly not coated. On coming out of the coater, the sheets are fed
onto wickets attached to a conveyor. This permits the sheets to go through the baking oven
in an almost vertical position, reducing the necessary length of the oven. Baking schedules
vary with the application. Interior food can coatings may be cured in less than a minute
at 200 to 2108C, and exterior white and varnish coatings may receive only 10 minutes
at 1508C. Inks are applied by offset lithography; in which the ink is transferred from a
lithographic plate to a rubber blanket and is offset to the sheet.
There are two processes for making two-piece cans. Drawn and wall-ironed (DWI or
D&I) cans are formed by drawing a cup from a flat blank and then ironing the walls to
thinner thickness and greater depth. In draw – redraw (DRD) cans, a coated blank is
formed into a shallow cup and then drawn one or two more times to achieve the desired
height and configuration and shape of the bottom of the can. The other piece of a two-piece
can is an end, similar to the ends of a three-piece can.
608
PRODUCT COATINGS FOR METAL SUBSTRATES
Examples of DRD cans are shallow cans for tuna fish and taller cans for vegetables and
pet foods. Flat sheets are coated, and then the can is drawn and formed. In some cases, such
as shoe polish and auto wax cans, printed sheets are formed into two-piece cans. The design
must be distortion printed; that is, the print must be designed so that it looks correct after
the distortion that results from forming. In general terms it costs less to coat and print flat
sheets before forming than to coat and print a formed can. The ability of the coating to
withstand this degree of formation depends on the depth of draw. It is not just a matter
of how deep the can is, but also how wide. Forming a large diameter can involves less
distortion than forming a narrow can; a deep can involves more distortion than a
shallow can. Bottle caps and crowns are made from coated and distortion printed sheets
and then are punched out and formed.
Beverage cans and some food cans are DWI cans; uncoated metal is drawn and formed
before coatings and inks are applied to the exterior, and a lining is sprayed into the interior
of the can. The exterior coating is applied by rotating the can against the transfer roll of a
small coater, and inks are transferred from a litho plate to a soft rubber roller that in turn
transfers the inks to the can surface. Baking is done in short time cycles at high temperature. Some lines produce over 2000 cans per minute. It is estimated that the coating
reaches a peak temperature of about 2058C for only about 1 second. Only partial crosslinking occurs in the short time. The cross-linking is completed when the can is baked
again after the interior lining is applied. Interior coatings are applied by spray. A small
spray gun is inserted automatically into a spinning can; the gun sprays and then is
pulled out of the can. To assure removal of essentially all solvent, the final part of the
curing is done with air directed into the can bodies. Typically, the cure schedule is
about 2 minutes at 2008C. Final forming of the top of the can to fit the end is done
after the coating and printing are completed.
There are three major types of metal used in cans: tin plate, steel, and aluminum. The
choice of metal depends on the end use. Beverage cans are two-piece aluminum cans.
Pressure from the carbonation of either soft drinks or beer keeps the thin walled aluminum
cans sufficiently rigid. In Asia and Europe some tinplate cans are used. Most food cans are
three-piece cans made from tin-plated steel. In the case of food cans, which do not have
significant interior pressure, aluminum does not compete with steel because thicker
walls of aluminum would be required to achieve adequate rigidity. However, many
shallow fish cans are being made with aluminum. There has been some discussion of
using aluminum for food packing by putting a small piece of dry ice into the pack
before sealing. Vaporization of the CO2 provides the pressure necessary for rigidity.
However, people associate slightly bulging food cans with spoilage, so consumer acceptance has been slow. A large fraction of pet food cans are made from treated steel—called
black plate—since it is less expensive and the highly reflective tin coating is not needed.
30.3.1. Interior Can Linings
The composition of interior can coatings depends on the food or beverage to be packed in
the can. In most cases, if the food is to be cooked, it is cooked in the can. A common
cooking cycle is 60 minutes at 1218C. Most beer is pasteurized in the can at a lower
temperature. In both cases, the interior and exterior coatings must maintain their adhesion
and integrity through the cooking or pasteurizing process. A paper has been published
reporting studies of tests for dry and wet adhesion of several coatings on various tin
plate substrates. It was found that electrochemical impedance spectroscopy (EIS) was
30.3. CONTAINER COATINGS
609
the most reliable test procedure. Organosol (Section 17.1.2) and epoxy –phenolic
(Section 13.3.1) coatings showed the best dry and wet adhesion, both before and after
sterilization [49].
Most vegetables and fruits are packed in cans with an interior coating called an R
enamel. Historically, R enamel was a phenolic varnish. Now, it is more common for R
enamel to be a phenolic resin modified with some drying oil derivatives during its synthesis or an epoxy – phenolic coating—a resole phenolic (Section 13.6.1), with a BPA
epoxy and phosphoric acid catalyst. To achieve adequate flexibility, high MW epoxy
1007 or 1009 resins (Section 13.1.1) were formerly used. New resole phenolics have
been developed that can be used with lower MW 1001 or 1004 BPA resins, thus decreasing
VOC emissions [50]. For packing vegetables that give off hydrogen sulfide during
cooking, such as corn, fine particle size ZnO pigment is dispersed in the coating, which
is called a C enamel. The ZnO reacts with H2S to form white ZnS. This prevents or conceals formation of unsightly black tin sulfide by reaction of tin oxides with H2S.
Waterborne coatings have also been developed to reduce VOC emissions. Epoxy–
phenolic emulsions with nonionic surfactants have been used but tend to have marginal
properties. Waterborne coatings based on amine salts of a carboxylic acid– functional
epoxy phenolic are said to have properties superior or equal to those of solventborne
coatings [51].
Cans for fish and other oily foods are generally lined with an epoxy-phenolic or resole
phenolic resin (Section 13.6.1). To obtain sufficient formability to permit drawing of
two-piece fish cans, cross-link density is reduced by using a mixture of p-cresol and
phenol in making the resole phenolic resin. Poly(vinyl butyral) is commonly incorporated
in the formulation to promote adhesion and act as a plasticizer. The extent of cross-linking
is adjusted for flexibility, together with resistance to swelling and softening by oils from
the fish (and in some cases, additional oil in which the fish is packed), during processing
and storage. Thin film thickness is used; this helps permit forming without film rupture.
Aluminum flake pigment is sometimes incorporated in the coating to minimize
permeability. In linings for meats such as ham, one of the key requirements is that the
coating permit the meat to slide out of the can easily after the top is removed. This requires
incorporation of a release agent, such as petroleum wax, in the can lining.
An increasing fraction of food cans are DRD or DWI two-piece cans. The draw – redraw
process requires that coatings have a greater degree of ductility than the coatings used for
interiors of three-piece food cans. Most commonly, vinyl organosol (Section 17.1.2) coatings lightly cross-linked with phenolic resins or MF resins have been used, since the vinyl
chloride copolymer binder shows ductility below Tg [52] (Section 4.2). In some European
countries, there is concern about possible toxic reaction products from HCl scavengers and
toxic emissions from factories recycling vinyl chloride copolymer – coated cans [53]. This
has led to work on other more distensible coatings. For example, epoxy resins have been
designed that will take considerably deeper draws than those required for conventional
BPA epoxy coatings [54].
Large volumes of coatings are used for lining beverage cans. Historically, most spray
applied interior coatings were solventborne vinyl chloride copolymers or high MW
epoxy/amino resin/phenolic coatings. The volume solids of these coatings at application
viscosity were very low, 12 to 15% NVV. The high cost of solvent and the high level of
VOC emissions forced a change. Waterborne coatings based on acrylic – epoxy graft copolymers are now widely used in lining both beer and soft drink cans [55,56]. Styrene/ethyl
acrylate/acrylic acid side chains are grafted onto a BPA epoxy resin. The resin is
610
PRODUCT COATINGS FOR METAL SUBSTRATES
“solubilized” with dimethylaminoethanol (DMAE) in glycol ether solvent and reduced
with water (Section 13.4). To lower the cost of some soft drink can linings, it has been
possible to use a special latex binder blended with small amounts of the graft copolymer
and epoxy phosphate as adhesion promoters.
Concern has been expressed about use of BPA epoxy resins in interior coatings, since
BPA is an estrogen mimic. Procedures for more accurate analysis of BPA extractables
have been published [57]. Studies of extractables from beverage can linings have
shown no detectable BPA using analysis methods sensitive to 5 ppb. Average migration
from food can linings was 37 ppb, a level that is estimated to lead to a maximum potential
dietary exposure of approximately 2.2 ppb, which is below the levels considered to be of
concern [58].
30.3.2. Exterior Can Coatings
Most food cans are not coated on the outside; rather, paper labels are used. The appearance
of coated and printed cans is more attractive than that of paper labels. Some food and all
beverage cans are coated and/or printed. Printed metal cans are not affected by water
such as from ice chests or condensation when cans are brought out of a refrigerator into
humid air.
The general procedure used on the exterior of sheets for the bodies of three-piece cans
is to apply a base coat, often called an enamel, print on the base coat up to four colors by
offset lithography, and finally, top coat with a finishing varnish. The most common color
for base coats is white, but a variety of colors are used for specific products. Color stability
and lack of color change on baking and can processing (sterilization) are critical requirements. To minimize VOC emissions, waterborne acrylic coatings cross-linked with MF
resins are widely used vehicles. The lithographic printing inks are either baking or UV
cure inks. The vehicles for the baking inks are long oil alkyd resins with some MF
resin cross-linker, while the vehicles for the UV cure inks are blends of acrylated epoxidized soy (or linseed) oil with acrylated epoxy resin and acrylate reactive diluents (Section
29.2.4).
The finishing varnishes are also acrylic –MF or sometimes polyester – MF binder compositions. To decrease friction in forming machines and conveyors, the finishing varnish
commonly contains a small amount of petroleum wax or fluorinated surfactant to reduce
surface tension so that the cans will have a low coefficient of friction. Two inks are
applied, the sheets are baked, the other two inks are applied, the finishing varnish is
applied, and the sheets are baked again. This is called wet-ink varnishing, since the
varnish is coated on the printed surface before the last two inks are cured. With UV
cure inks, each ink is partially cured by passing under a UV lamp before the next ink is
applied. Commonly, finishing varnishes are not used over UV cure inks.
For 2-piece cans, white waterborne acrylic base coats are used. The film thickness varies
from 8 to 15 mm; even at 15 mm, hiding is less than 100%, but the coating is definitely white.
An increasing proportion of two-piece cans are simply printed with inks 2 to 3 mm in thickness with no base coat. Appearance of these cans is inferior, but they cost less than cans that
are coated with a base coat before printing. A transparent yellow–brown base coat is used on
some cans to give a bronze or gold color to the relatively shiny metal surface.
UV cure coatings are also used on some beverage cans (Chapter 29). Exterior finishing
varnishes are acrylated resin coatings cross-linked by photogeneration of free radicals.
Coatings for the exterior side of can ends are epoxy resins cross-linked by photogeneration
30.4. COIL COATING
611
of acid. The UV cure end coating is applied to the side of the sheets that will be on the
outside of the cans and passed under the UV lamps, giving a partially cured film; the
other side, which will be on the inside of the can, is coated with an FDA-listed epoxy –
phenolic thermal cure coating. When the epoxy –phenolic is cured in an oven, the cure
of the UV cure epoxy coating is advanced to completion by the photogenerated acid
still present in the films. This thermal complement to the UV cure also enhances adhesion
of the coating.
30.4. COIL COATING
Steel and aluminum are manufactured in long, continuous strips that are rolled into coils.
In many cases, the coils are cut, formed, and fabricated before they are coated, as in
production of automobiles, for example. In other cases, it is possible to coat metal coils
and later fabricate the final product from the precoated metal. When precoated metal
can be used, there can be substantial advantages (Section 30.4.1), and coil coating has
grown into a major industrial coating process [59] whose growth continues.
Coil coating started in 1935 as a process for coating venetian blind slats. The strips of
metal were about 5 cm wide, and the line was run at a rate of about 10 m min21. Modern
coil lines can coat metal up to 1.8 m wide at rates as high as 275 m min21. Most lines run
at a rate of around 100 to 200 m min21. A schematic drawing of a coil line is shown in
Figure 30.1.
The metal is shipped from the steel or aluminum mill in coils weighing up to 25,000 kg
that are 0.6 to 1.8 m wide and 600 to 1800 m long. In some lines, as shown in Figure 30.1,
the first step is precleaning; brushes remove any physical contaminants. The strip then
goes to the entry accumulators. The rollers of the accumulators move apart to accumulate
a significant length of coil, so that when one coil is about to run out, the next coil can be
stitched (welded) on while the accumulator rolls move together, supplying strip to the line
without interruption of the process. After the stitching is done, the new coil is fed to the
coating line. As the process continues, the accumulator rolls gradually separate to store
strip for the next change of coils.
Next, the strip is carried through the metal treatment area. Detergent washing and rinsing
are followed by application of various conversion coatings, including phosphates, complex
oxides and chromates, water rinse, and finally, chromic acid or other rust-inhibiting rinse.
Because of concerns about toxicity of chromates, proprietary chromate-free conversion
coatings are being adopted. All the cleaning and conversion coating procedures must be
designed to work at high speeds. Since the metal is moving at 100 to 200 m min21 or
more, the total time for cleaning and treatment is on the order of a minute or less. Next,
the strip is carried through a dry-off oven and finally, to coaters and a baking oven. A
laminator is also shown in Figure 30.1, although laminating of film in line with coating is
not common. It is more usual to have two coating stations, each followed by a baking
oven. Often, a primer and a top coat are applied on one side of the strip and another coat
on the back side. It is fairly common to coat, cure, and then print one or more colors.
Most coatings are applied by reverse roll coating, but sometimes direct roll coating is
used for thinner films (Section 23.4). Reverse roll coating is used for thicker films for
which close control of thickness is critical and for which the flow advantages that result
from the wiping action, in contrast to the film splitting action in direct roll coating, is
important. To achieve even film thickness across a strip 2 m wide, the rolls have to be
612
Figure 30.1. Coil-coating line. (From Ref. [59], with permission.)
30.4. COIL COATING
613
crowned, that is, designed so that their diameter is greater in the middle than on the edges,
since the pressure involved tends to bow the rolls to a small degree. To avoid damage to
the rollers, the line is programmed so that very shortly before a stitched section joining two
coils passes through the coating nip between the rollers, the rollers are automatically
separated slightly and then almost immediately returned to normal operating pressure.
Coil coating of embossed metal tends to hide the embossing. To retain the pattern, after
coating, the stock moves under a high-velocity air knife, which removes excess coating
from the embossed areas. As a result, the coil is reported to have a uniform thickness of
coating on the entire surface, which maintains the embossed design [60].
At the high line speeds, the dwell time in the ovens is less than a minute, even with
long ovens. In some cases, the dwell time is as low as 10 seconds, but more commonly
is in the range 12 to 45 seconds. After a short initial period, hot air is directed over the
surface of the coatings at high velocity. The air temperature can be as high as 4008C.
The temperature reached by the coating on the metal is the critical temperature for
curing the coating. This temperature cannot be measured directly, but is closely related
to the metal temperature, which can be measured. The temperature considered most
important is the peak metal temperature (PMT); this can be as high as 2708C. After the
coating is baked, the strip passes through the exit accumulator to the rewind. The exit
accumulator stores coated strip during removal of a coated coil. In some lines the strip
passes over chilling rolls, or through water quenching, to reduce the temperature before
the strip is rewound into a coil. The pressure in the center of the rewound roll is very
high; consequently, the Tg of the coating on the metal must be very high, to avoid blocking.
For most applications, however, the cross-link density must not be so high that the film
loses elasticity, which is needed to allow subsequent fabrication of the steel into its
final form.
Near infrared curing of coil coatings has been started on at least one coater. Cure
is reported to be complete in 1 to 4 seconds compared to 12 to 45 seconds on a
conventional line.
The exhaust air from the hoods over the coaters, and particularly from the oven,
contains solvents. On most lines, the exhaust air streams are used as part of the air used
to burn the gas to heat the ovens or are fed through an oxidizer. In this way, part of the
residual heat from the oven exhaust is recycled and the solvent is burned. Burning the
solvent essentially eliminates VOC emissions, and the fuel value of the solvent is
recovered. As a result, there has been less incentive to change to waterborne or high-solids
coatings in coil coating applications than in other applications. There are still pressures to
reduce solvent content because the fuel value of solvents is low compared to the cost of the
solvents and because some lines are not equipped to burn solvent. New regulations
controlling emission of hazardous air pollutants (HAPs) that will go into effect in 2006
require that stack emissions not exceed 0.38 lb of HAP per solid gallon of coating
applied. Most coil coaters can now meet this requirement, but some will have to
upgrade their lines. This addition to the VOC requirements will give a further advantage
for coil coating over many other application methods.
Coatings on aluminum are frequently single coats; but on steel, primer– top coat
systems are most widely used. Binders for primers were traditionally based on BPA
epoxy resins; epoxy esters and epoxy – MF resins are examples. However, polyurethane,
polyester, and waterborne latex primers are being used increasingly.
Many types of coating binders are used for top coats. Oxidizing alkyds with MF resin
are the lowest cost and are sometimes used on the reverse side of the coated strip as a
614
PRODUCT COATINGS FOR METAL SUBSTRATES
backer; polyesters are supplanting alkyds in this application. Backer coatings may be
pigmented or unpigmented, and they contain a small amount of incompatible wax. The
purpose of the backer is to avoid metal marking of the top surface coating by rubbing
against a bare metal reverse side of the coil. Alkyd –MF coatings are also used as top
surface coatings for which corrosion resistance requirements and/or exterior durability
requirements are modest. Polyester – MF binders are widely used, especially as single
coats; exterior durability and corrosion protection are generally superior to those obtained
with alkyd coatings. Over the years, incremental improvements in the design of polyester
resins and in formulations have led to polyesters with substantially improved outdoor
durability. Polyester – blocked isocyanate coatings have been used to a degree for
applications for which abrasion resistance and flexibility are particularly important.
Close temperature control in the ovens is important with urethane coatings, since
urethanes may discolor and decompose relatively rapidly at the baking temperatures
involved in coil coating. Thermosetting acrylic –MF coatings are used, generally over
primers.
For greater exterior durability, one can use silicone-modified polyesters and
silicone-modified acrylic resins (Section 16.1.2). For example, one might use 30%
silicone-modified polyester resin with a small amount of MF resin as a supplemental
cross-linker as the binder for color top coats for high performance residential or industrial
siding. In the same quality line, the white top coat might well be just a polyester –MF
coating. Although the white might start to chalk slightly after many years of outdoor
exposure, this would not affect the appearance adversely. However, even a small
amount of chalking of a color, coating makes an easily seen change of color, due to the
change in surface reflection. Such changes are particularly serious in exterior siding,
since the exposure varies depending on the location on the building. The resulting nonuniform chalking of the color coating is very evident. For the highest exterior durability,
polyvinylidene fluoride resin coatings are used (Section 17.1.4). In some cases, such
coatings show only slight indication of change after exposure outdoors for more than
25 years.
Organosol and plastisol coatings are used for some coil coating. The relatively low
viscosity organosols are applied at about 25 mm, and the higher viscosity plastisols are
used for film thicknesses of 100 mm or more. As discussed in Section 17.1.2, the
vehicle in these coatings is a dispersion of vinyl chloride copolymer in plasticizer and
solvent. Such coatings provide reasonable exterior durability with excellent fabrication
properties. They have the advantage that they need not be cross-linked and hence can
be run with as little as a 15 second dwell time in the oven. Solution vinyl resins are
used for coil coating metal for can ends for beverage containers.
Increasing use is being made of latex vehicles for coil coating. They have the advantage
of high MW, so mechanical properties can be achieved without need for much, if any,
cross-linking. High-gloss coatings cannot be made, and there can be greater problems
of leveling than with solventborne coatings. Leveling problems are minimized by using
reverse roll coating and by using associative thickeners to control viscosity. Associative
thickeners minimize latex particle flocculation relative to conventional water-soluble
polymeric thickeners such as hydroxyethyl cellulose (Sections 24.2 and 32.3).
An extensive study of the corrosion resistance of 12 coil coated sheets at three South
American sea shore locations has been reported. Location 1 was marine-polar, with low
temperature, high humidity, very low SO2 and Cl2 exposure, and little sun. Location 2
was marine with moderate temperature, high humidity, moderate SO2 and Cl2 exposure,
30.4. COIL COATING
615
with high sun exposure. Location 3 was marine-desert with high temperature and
humidity, very high SO2 and Cl2 exposure, and high sun exposure. At location 1, the
best performance was shown by an alkyd –MF coating on galvanized steel. At location
2, the best performance at incision marks was a polyester primer with chromium
pigment on Galvalume, on the unruptured surface the best performance was with three
samples: an epoxy –acrylic primer with a silicone-modified polyester top coat on
galvanized metal, a polyester primer with a silicone-modified polyester top coat on
Galvalume, and a polyester primer and polyester top coat on Galvalume. At location 3,
the best performance was a polyester primer and polyester top coat on Galvalume. At
all three locations, the poorest performance was an epoxy primer with a polyester top
coat. No details were given of the binders in the various coatings [61].
UV cure coatings are potentially of great interest in coil coating because of rapid cure
and limited emissions. Active research is under way, but actual production is limited so
far. Cationic UV cure epoxy coil coatings on tin plate are used for the exterior side of
can ends. The superior adhesion of epoxy coatings makes them preferable to acrylate coatings for this application. In some cases a UV cure coating is applied on one side of the strip
and UV cured. Then a thermally cured epoxy coating is applied on the other side. When
the strip is put through the oven to cure the thermally cured coating, the cross-linking of
the UV cure coating is also advanced.
Powder coatings are also being applied to coil stock. One approach, which is in limited
commercial use, is to electrostatically spray the strip with automatic spray guns [62]. A
disadvantage of this process is that line speeds are slow. Another process being introduced
is to run the coil strip through a “cloud” of electrically charged powder particles; the strip
then passes into an induction heating oven for fusing and curing [63]. Since there is no
contact of rollers with the coil, embossed or perforated metal can be coated. It is projected
that line speeds can be higher than for conventional coil coating.
30.4.1. Advantages and Limitations of Coil Coating
Important advantages of coil coating have promoted the growth of the business to a major
component of industrial coatings. For long production runs, the cost is low compared to
that of coating preformed metal. The rate of application of coating is faster and the
labor cost is lower. Coating utilization is essentially 100%. Oven designs are such that
energy use in curing is much more efficient. In general, floor space requirement is
much less, so capital cost for buildings is lower. Since solvent is restricted to the immediate area of a roll coater, fire hazards and toxicity hazards are reduced over those of spray
application. Since in most cases, solvent is incinerated, VOC and HAP emissions are generally very low. Film thickness of the applied coatings is more uniform than is generally
obtained in coating preformed products. Since the coatings are applied on uniform thickness metal, curing of all parts of the coating tends to be more uniform than in curing coatings on fabricated products. For many applications, the performance of coil coatings is
superior. This difference is particularly evident when comparing high quality precoated
exterior siding to house paint. A large part of this difference results from the superior performance of baked coatings to that of air dry coatings. The manufacturer using precoated
metal gains some substantial advantages. The VOC and HAP emissions and fire hazards
associated with coatings application are eliminated, and insurance costs drop. There is no
waste disposal problem with sludge from spray booths. There may be a substantial saving
in floor space.
616
PRODUCT COATINGS FOR METAL SUBSTRATES
There are limitations to coil coating. The capital cost of a line is very high; therefore,
lost production time in shutdowns is expensive. It is economical only for fairly long runs
of the same color and quality of coating. The cost of changing color is high because the
coater must be shut down for cleanup. However, many modern lines have multiple
coating heads, which substantially reduce downtime due to color changes. In a line of
coated metal with several colors, the inventory cost can be high, since an inventory of
several colors must be kept. If a stylist changes colors, the obsolete inventory cost can
be high or, said another way, the flexibility of changing colors is more limited than
when the assembled product is coated.
Color matching is a challenge to a coil coating supplier. Generally, very close color
matches are needed. Hiding is less than complete, and the color of the metal or primer
affects the color of the coating. Color can also be affected by the high temperature
baking schedule. It is not possible to duplicate in the laboratory a curing schedule of 30
seconds with high-velocity 4008C air. A color matcher must learn to compare color
differences that can be expected in the laboratory with what will happen on a particular
coil coating line and then do the color matching, taking the difference into consideration.
The coated metal must be able to withstand fabrication into the final product without
film rupture. This may require acceptance of somewhat softer, more flexible films than
could be specified if the product were coated after fabrication. When the coated metal
is die cut to make the eventual product, bare edges of metal are exposed. Welding of
coated metal can be a problem. Coatings that do not interfere with the welding process
are available; however, the appearance of the coated surface is destroyed around the
welded area. In some applications, bare edges are not a problem and the product is not
welded. In other cases, it is critical for the designer of the finished product to ensure
that the cut edges do not show and are not in locations where they will be subject to
corrosion, and that any welding will be done in areas that will not show.
Examples of large applications for coil-coated metal are siding for residential use,
original siding for mobile homes, venetian blinds, rain gutters and downspouts, fluorescent
light reflectors, appliance cabinets, can ends, and can bodies for fruits and vegetables.
Newer examples include antimicrobial coil coatings for making stock for air conditioning
duct work, metal kitchen counters, refrigerator boxes, walk-in coolers, and the like. The
antimicrobial agent is called AgION, a silver-ion-on-zeolite product. Legislation is
increasingly requiring many roofing materials to have high reflectance and high emittance,
White and light colored coil coated metal has high reflectance, thus reducing heat absorption. Furthermore, infrared reflective pigments (Section 20.2) enable formulation of “cool
building” coatings in a broad range of colors. However, some radiation is absorbed, so
there is still heating. This heat is lost by thermal emittance into the atmosphere and into
the building [64].
Preprimed weldable coil stock is used to a limited extent in some automobile plants to
replace E-coat primer on fabricated cars. There is also the possibility of E-coating the car
bodies assembled with precoat stock, presumably with a conductive coil-coated primer.
GENERAL REFERENCES
Fettis, G., Ed., Automotive Paints and Coatings, VCH Verlagsgesellschaft, Weinheim, Germany,
1995.
Gaske, J. E., Coil Coating, Federation of Societies for Coatings Technology, Blue Bell, PA, 1987.
REFERENCES
617
Goldschmidt, A.; Streitberger, H.-J., BASF Handbook on Basics of Coatings Technology, Vincentz,
Hannover, Germany, 2003, pp. 688 – 727.
McBane, B. N., Automotive Coatings, Federation of Societies for Coatings Technology, Blue Bell,
PA, 1987.
REFERENCES
1. U.S. Department of Commerce, Bureau of Census, Current Industrial Reports: Paint and
Coating Manufacturing, http://www.census.gov.
2. Chem. Engi. News, 2002, 80(42), 27.
3. Crawley, L. C., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2005,
pp. 3 – 10.
4. Koenig, E.; et al., U.S. patent 5,532,313 (1996).
5. Wamprecht, C.; et al., U.S. patent appl. 20030114578.
6. Ladstadter, E.; Gessner, W., Proc. Int. Conf. Org. Coat. Sci. Technol., Athens, Greece, 1986,
p. 203.
7. Kenny, J. C.; et al., J. Coat. Technol., 1996, 68(855), 35.
8. Agawa, T.; Dumain, E. D., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 1997, p. 342.
9. Green, C. D., Paint Coat. Ind., 1995, September, 45.
10. Vanier, N. R.; et al., U.S. patent 6,875,800 (2005).
11. Cheever, C. D.; Ngo, P.-A. P., J. Coat. Technol., 1989, 61(770), 65.
12. van Beelen, D. C.; et al., Proc. Int. Conf. Org. Coat. Sci. Technol., Athens, Greece, 1989,
p. 39.
13. Tachi, K.; et al., J. Coat. Technol., 1990, 62(782), 43.
14. Weaks, G. T., Proc. ESD/ASM Adv. Coat. Technol. Conf., 1991, p. 201.
15. Mezger, F., Proc. 29th Int. Conf. Coat. Sci. Technol., Athens, Greece, 2003, p. 251.
16. Bauer, D. R.; et al., Ind. Eng. Chem. Prod. Res. Dev., 1985, 21, 686.
17. Boggs, L. G.; et al., J. Coat. Technol., 1996, 68(855), 63.
18. Broder, M.; et al., J. Coat. Technol., 1988, 60(766), 27.
19. Vachlas, J. Oil Colour Chem. Assoc., 1989, 72, 139.
20. Wagstaff, I., Proc. ESD/ASM Adv. Coat. Technol. Conf. 1991, p. 43.
21. Backhouse, A. J., J. Coat. Technol., 1982, 54(693), 83.
22. Vogt-Birnbrich, B.; et al., U.S. patent 6,069,218 (2000).
23. de Wit, J. S.; et al., FSCT Annual Meeting, 2004.
24. Seo, K. S.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2004,
Paper 22.
25. Galgoci, E. C.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2004, Paper 29.
26. Swarup, S.; et al., U.S. patent 6,762,240 (2004).
27. Gregorovich, B. V.; Hazan, I., Prog. Org. Coat., 1994, 24, 131.
28. Quill, J.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2005, pp.
308 – 320.
29. Gardon, J. L., J. Coat. Technol., 1993, 65(819), 24.
30. Chen, M. J.; et al., Surf. Coat. Int., 1996, 79, 539.
618
PRODUCT COATINGS FOR METAL SUBSTRATES
31. Nordstrom, J. D., Proc. Waterborne Higher Solids Powder Coat. Symp., New Orleans, LA,
1995, p. 492.
32. Rehfuss, J. W.; St. Aubin, D. L., U.S. patent 5,356,669 (1994). Rehfuss, J. W.; Ohrbom, W. H.,
U.S. patent 5,373,069 (1994).
33. Anderson, L. G.; et al., U.S. patent 6,803,498 (2004).
34. Courter, J. L., J. Coat. Technol., 1997, 69(866), 57.
35. Furukawa, H.; et al., Prog. Org. Coat., 1994, 24, 81.
36. Yaneff, P. V.; et al., Proc. Waterborne Higher-Solids Powder Coat. Symp., New Orleans, LA,
2001, pp. 109 – 125.
37. htpp://automotive.dupont.com/en/newsEvents/article20030305b.html.
38. Hovestadt, W.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2001, p. 421.
39. Vanier, N. R.; Munro, E., U.S. patent appl. 20030158316.
40. Valet, A., Prog. Org. Coat., 1999, 35, 215.
41. Schwalm, R.; et al., U.S. patent 6,747,088 (2004).
42. Overbeek, G. C.; et al., U.S. patent 6,730,733 (2004).
43. Yanagida, K.; et al., J. Coat. Technol., 1996, 68(859), 47.
44. Sapper, E.; Woltering, J., U.S. patent 6,296,903 (2001).
45. Ryntz, R. A., SAE Tech. Paper Ser. 880597, Society of Automotive Engineers, Warrendale, PA,
1988.
46. Schoff, C. K., JCT Coat. Tech., 2004, 1(2), 34; 1(3), 22.
47. Miranda, T. J., J. Coat. Technol., 1988, 60(760), 47.
48. Powder Coating, http://www.specialchem4coatings.com.
49. Beiro, M., Prog. Org. Coat., 2003, 46, 97.
50. Nootens, C.; et al., Surf. Coat. Int., 1997, 80, 50.
51. Oberressl, P.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
1997, p. 296.
52. Palackdharry, P. J., Polym. Mater. Sci. Eng., 1991, 65, 277.
53. Hickling, M., Polym. Mater. Sci. Eng., 1991, 65, 285.
54. Dubois, R. A.; Sheih, P. S., J. Coat. Technol., 1992, 64(808), 51.
55. Woo, J. T. K.; et al., J. Coat. Technol., 1982, 54(689), 41.
56. Woo, J. T. K.; Eley, R. R., Proc. Waterborne Higher-Solids Powder Coat Symp., New Orleans,
LA, 1986, p. 432.
57. Wingender, R. J.; et al., J. Coat. Technol., 1998, 70(877), 75.
58. Howe, S. R.; et al., J. Coat. Technol., 1998, 70(877), 69.
59. Gaske, J. E., Coil Coatings, Federation of Societies for Coatings Technology, Blue Bell, PA,
1987.
60. Challener, C., JCT Coat. Tech, 2004, 1(2), 50.
61. Rosales, B. M.; et al., Prog. Org. Coat., 2004, 50, 105.
62. Blatter, K.; et al., Coat. World, 1998, 3(2), 26.
63. Graziano, F. D., Proc. Int. Conf. Org. Coat., Athens, Greece, 1997, p. 139.
64. Cocuzzi, D. A.; Pilcher, G. F., JCT Coat. Tech, 2004, 1(2), 22.
31
Product Coatings for
Nonmetallic Substrates
Many products made from wood and plastics are coated in factories; examples are wood
furniture and paneling, hardboard paneling and siding, plastic auto body parts, and computer cases. Do-it-yourselfers paint wood products at home, but the products they use are
generally different and are discussed in Chapter 32. Other nonmetallic substrates are also
coated in factories (e.g., paper, textiles, and films). Because of space limitations, we will
focus on wood, plastics, and glass.
31.1. COATINGS FOR WOOD
The world market for resins for industrial wood coatings in 2004 (Table 31.1), as reported
in a Bayer study, had a total value of 1.9 million euros [1]. The world market for formulated
coatings for industrial wood finishing in 2004 was 2.2 million metric tons. The percentage
of each type is shown in Table 31.2. The U.S. market for wood finishes dropped from
303 million liters in 1995 to 202 million liters in 2003. The primary cause was the
marked increase in furniture imports from Asia. This is consistent with the large remaining
market for nitrocellulose lacquers, much of which is used in the Chinese furniture
industry [2].
31.1.1. Coatings for Wood Furniture
There are many styles of furniture, and furniture coatings are affected by styling. Furniture
styling and manufacture are different in North America than in Europe; most of our
discussion deals with the U.S. market. A large fraction of wood furniture made in or
imported into the United States is styled to resemble antique furniture. Styling is generally
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
619
620
PRODUCT COATINGS FOR NONMETALLIC SUBSTRATES
TABLE 31.1. Worldwide Market for Resins and
Intermediates for Industrial Wood Coatings, 2004
Resin Type
Percent
Nitrocellulose
Alkyds
Polyols
Polyisocyanates
UV cure (urethane acrylates)
Acrylic
Unsaturated polyesters
Polyurethane dispersions
7
6
18
19
15
10
10
15
Source: Ref. [1].
initiated in the high end of the market: expensive wood furniture that involves a great
deal of hand labor and artistry in its manufacture. A representative finishing process is
described in some detail since a major part of furniture is finished to resemble high
style furniture but uses manufacturing and finishing techniques that permit lower costs.
Fine quality furniture is made from a combination of woods. The tops and sides are
made from five-ply plywood, the legs and rails are made from solid wood, and carved-wood
decorative pieces are commonly attached to the furniture. The center ply of the plywood
can be wood, but other cores, such as chipboard, are used. Chipboard is made by pressing
wood chips and an adhesive binder into sheets and curing the sheets. Chipboard is a
desirable core material because it is more dimensionally stable than wood. A ply of
wood veneer is laid up with glue on each side of the core. Another layer of veneer is
laid up on each side, cross-grain to the first layer. The top, or face, veneer, the side seen
on the furniture, is usually from a hardwood selected for the beauty of its grain pattern.
Grain patterns are affected by the kind of tree from which the wood is obtained and by
the way the veneer is cut from a log. The top of a table is generally too large for the face
veneer to be made from only one piece; it is laid up from several pieces. They are carefully
selected and put together to give a particularly beautiful pattern. Adhesive is needed not
only between the layers in the plywood, but also to connect the edges of the individual
pieces of veneer in the top face. Each tabletop is unique because the grain patterns in
the separate pieces of veneer are different. In many cases, elaborate patterns are laid up.
Many kinds of wood are used: pecan, walnut, and oak are among the most commonly
used. All these hardwoods have open pore structures and prominent grain patterns. In
some styles, mahogany, which has straighter grains and shallower pores, has been
TABLE 31.2. Worldwide Market for Industrial Wood
Coatings, 2004
Coating Type
Nitrocellulose lacquers
Polyurethanes
UV curing
Acrylic
Unsaturated polyester
Waterborne
Source: Ref. [1].
Percent
31
41
3
12
6
7
31.1. COATINGS FOR WOOD
621
popular. Woods that have little pore structure, such as maple and birch, have a more
limited market than those with bold patterns.
In U.S. furniture manufacturing, the various components are cut for a run of the same
style and sets of furniture. The furniture is assembled before it is finished. The various
components are of different colors; if the final color of the furniture is to be lighter
than any part of any of the component pieces, the wood is bleached with a solution of
30% hydrogen peroxide; the bleaching solution is activated with sodium or potassium
hydroxide. Bleaching is now seldom done in the United States.
In earlier times the next step was to spray a dilute solution of hide glue on the surface.
Now, low solids poly(vinyl alcohol) size solutions are used. The process is still called glue
sizing. After the size has dried, the wood fibrils that stick up from the surface are stiffened
so that the surface can be smoothed by sanding to minimize concentration of stain at the
fibrils, which would give an unattractive appearance. With some woods, sanding without
glue sizing gives satisfactory results.
Next, the wood is colored with a solution of acid dyestuffs in solvent or with dispersions
of transparent pigments. The stain is selected and applied to give the overall base color to
the furniture, Spraying stain requires skill to obtain even coloration.
A wash coat is a low solids [less than 12 wt% (NVW)] low viscosity lacquer. A thin
layer is applied to minimize displacement of the stain, to stiffen wood fibers for sanding,
and to prepare the surface for the next step, filling.
The purpose of filling is to color the pores of the wood to emphasize the grain pattern
and to fill the pores to near the same level as the rest of the wood. The pores must be filled
with the colored filler without leaving filler on the surfaces between the pores. Usually, the
color of the filler is a dark brown and the stains are lighter yellowish or reddish browns, but
for special effects, one can use white filler with black stain or other color combinations.
The vehicle is linseed oil and/or a linseed long oil alkyd with a hard resin such as
limed rosin, and driers with aliphatic hydrocarbon solvents. The color pigments used
for dark colors are natural earth pigments with a high loading of inert pigments at a
PVC over 50%. The entire piece of furniture is sprayed with a liberal coating of filler,
the filler is “padded” into the pores by rubbing vigorously with a pad of cloth, and then
excess filler is wiped off. If the wood has been properly wash coated and the filler is
wiped evenly when the degree of solvent flash off is right, the pores of the wood are
filled. Adequate drying of the filler is critical because residual solvents can result in shrinkage, graying, and loss of adhesion of succeeding coats.
A sanding sealer is then applied. Its purpose is to prepare the surface for application of top
coat. The sealer must be easy to sand smooth without “gumming” the sandpaper. A formulation for a typical sanding sealer includes nitrocellulose; a hard resin, such as a maleated
rosin esterified with a polyol; and a plasticizer such as blown soy oil (Section 14.3.1).
Sanding sealers contain 3 to 7% of zinc stearate, based on the lacquer solids, to aid
sanding. After diluting with lacquer thinner to spray viscosity (about 20% volume solids),
the sealer is applied and after drying is sanded smooth. UV cure sealers with very low
VOC are being used by some manufacturers, especially under waterborne coatings.
Wood is not naturally as uniform in color as the overall staining makes it. By spraying
different shades of stains to selected areas, the color can be varied and the grain highlighted. Shading is a skilled art; in recent years shading stains have been replaced in
large measure by padding stains. Padding stains are made with similar dyes but have
some binder and somewhat slower evaporating solvents; they are applied by hand with
a rag moistened with the stain, and less application skill is required.
622
PRODUCT COATINGS FOR NONMETALLIC SUBSTRATES
The furniture is then sprayed with a top coat. Nitrocellulose is used as the primary
binder in high-style furniture lacquers for three major reasons. First, the appearance of
furniture finished with nitrocellulose lacquers is outstanding; the lacquers provide an
appearance of depth, fullness, and clarity of the grain pattern that has not been matched
by any other coating. Second, lacquers dry quickly, so they can be rubbed a short time
after application and then packed and loaded into a truck for shipment without printing.
The term printing in this context means surface imperfections resulting from wrapping
material denting the surface of a coating that is too soft. Third, lacquers are thermoplastic
and permanently soluble, so they are easily repaired in case of damage during shipment.
The mechanical properties of the final film are better with higher molecular weight (MW)
nitrocellulose, but solids are reduced. A hard resin such as an esterified maleated rosin is
blended with the nitrocellulose. Plasticizers, usually including a short oil coconut alkyd
(Section 15.4), provide the required flexibility. The balance of nitrocellulose, plasticizer,
and hard resin is critical. If the coating is too soft, it will be difficult to rub; if it is too hard,
the lacquer will be subject to cracking as the wood expands and contracts with changes in
moisture content and/or exposure to rapid temperature changes. If the lacquer is applied
over a pale colored wood, UV absorber (Section 5.2.1) is added to reduce yellowing of the
wood. Citric acid is also commonly added to chelate iron salts that produce reddish colors
with phenolic compounds naturally present in wood.
After the initial coat of top coat is applied, the next step is distressing. Distressing uses a
variety of techniques to give an antique appearance to furniture. Stylists are not trying to
make fake antiques; rather, the furniture is styled to have an antique appearance. In the
old days, quill pens with black india ink were used, resulting in drops of ink falling on tabletops. Now little drops of black pigment glazing stain are applied. Sometimes, a colonial forefather was a little tipsy when he sat down to write and put the feather end of the quill pen into
the inkwell and then set it down on the tabletop. Sometimes, colonials walked on their tables
with hobnailed boots or banged the tabletop with their pewter beer mugs. It is startling to
walk by the finishing line in a fine furniture factory and see someone swinging a chain at
the tops of those carefully prepared pieces of furniture. Another common thing to see is a
little black pigmented stain rubbed into inside corners. Over the years in antique furniture,
dirt accumulated in corners, so the new furniture is made permanently dirty in these areas.
Proper distressing requires artistic skill.
Finally, another coat of top coat is applied and the lacquer is dried, usually in a force dry
oven at 40 to 608C. Then the lacquer surface is rubbed, first by sanding with fine sandpaper
and a lubricant and then rubbing with cloth and rubbing compound. The result is a soft
appearing low gloss finish.
While the solubility of nitrocellulose lacquer finishes is an advantage in that it permits
easy repair, it can also be a disadvantage. If solvents such as nail polish thinners are spilled
on a lacquer surface, the finish will be marred. A way of solving this problem and still
obtaining the beauty of a nitrocellulose finish is to add a small amount of polyisocyanate
cross-linker to the lacquer just before application. The isocyanate cross-links with the
hydroxyl groups on the nitrocellulose. If this step is to be done, all solvents must be
urethane grade, that is, contain no more than traces of alcohols and water, and the nitrocellulose must be wet with plasticizer rather than with isopropyl alcohol.
Wood furniture produced by this process is beautiful but expensive. A much larger part of
the market is for furniture designed to look as much as possible like this high-style furniture
but which can be manufactured and finished at lower cost. There are many approaches with
varying degrees of cost savings; an intermediate cost approach is described next.
31.1. COATINGS FOR WOOD
623
Instead of using plywood tops and sides made with expensive hardwood veneers that
must be finished painstakingly, one can use printed tops and sides with solid wood legs
and frames. Particle board, that is, pressed wood chips and particles with a resin binder,
is first filled with a UV cure filler. Note that this is a very different filler than referred to
above for filling the pores in wood. Particle board has a rough surface, and before use
in furniture tops, it must be coated with a filler and sanded smooth to give an even
surface for the application of a base coat and printing. This was originally done with
highly pigmented solventborne coatings applied at about 30 vol% solids. However, the
low solids requires two or three coats to give sufficient film thickness to permit sanding
to a smooth surface without cutting down into the substrate. With UV cure, the volume
loss is only a few percent, so an adequate layer can be applied in one coat. This together
with the fast cure means that UV cure fillers are less expensive to apply. The pigments
used are clays and silica that absorb or scatter only minor amounts of UV. Aluminum trihydroxide absorbs still less UV and may be a particularly effective inert pigment [3]. The
inert pigment minimizes the problem of oxygen inhibition. The high level of pigmentation
leads to poor leveling, but it does not matter, since the surface is sanded. Furthermore, the
pigmentation results in less resin being at the surface, so air inhibition is less of a problem.
Originally, the fillers were styrene – polyester systems, but the higher speed of acrylics has
led to their adoption despite their higher cost. The filler is applied to the particle board with
a roll coater that has a brush roller.
The filled and sanded panel is next coated with a base coat, a highly pigmented
nitrocellulose lacquer whose color becomes the overall underlying color of the piece of
furniture, corresponding to the color of the stain in high-style furniture. The stain used
on the solid wood parts approximately matches the color of the base coat. The base
coat is then ready for offset gravure printing. The gravure printing cylinders are made
from photographs taken of carefully selected hardwood top veneer plywood that was
stained, filled, shaded, and distressed by highly skilled finishers. One cylinder prints the
darkest tones of the original finished wood, a second prints the medium depth tones,
and the third prints the lightest tones. The printing ink colors for the three cylinders are
selected to approximate the colors of the three depths of color on the original. The inks
are pigment dispersions in plasticizer, so they adhere well to the base coat and to subsequent lacquer coats. A light coat of lacquer is applied to protect the print.
The frame and legs of the furniture are assembled and finished up to the stage of applying
top coat, followed by assembly of the printed tops and sides onto the furniture. Then the
entire piece of furniture is coated with semigloss nitrocellulose lacquer top coat. Semigloss
lacquer, not gloss lacquer, is used to give a finish with a hand-rubbed effect without extensive hand rubbing. The semigloss is obtained by pigmenting with a low level of fine particle
size SiO2, such that the PVC of the final film averages 2 to 4%. When the solvent evaporates
after application, convection currents resulting from solvent loss carry the pigment to the
surface of the film, where it is “trapped” in the viscous surface layer. As a result, the
PVC of the top layer of the film is high enough that low gloss can be obtained, while
the overall PVC is low enough that only a minor degree of light scattering occurs.
A disadvantage of furniture finishing systems discussed so far is that they require use
of low solids coatings with very high VOC emissions. The low solids requires multiple
coats, or at least more passes with a spray gun, so the application cost is high. Hot
spray (Section 23.2.4) is frequently used. A temperature of 658C permits increasing
solids from 20 – 24 NVW to 28– 34 NVW. This results in a significant reduction in
VOC emissions, but still far short of probable future requirements.
624
PRODUCT COATINGS FOR NONMETALLIC SUBSTRATES
In 1995 the U.S. EPA revised their wood furniture coatings regulation after extensive
discussions with the industry. It was agreed that regulations should be expressed as pounds
of VOC per pound of coating solids on the wood surfaces. The overall limits were set at
2.3 lb of VOC, including a maximum of 1.0 lb of HAP per pound of solids. In new plants
or finishing lines, HAP emissions are limited to 0.8 lb per pound of solids. For specific
coatings the following maxima were set: waterborne top coats, 0.8 lb; higher-solids
solvent-borne sealers, 1.9 lb; top coats, 1.8 lb; and acid-cured alkyd amino sealers, 2.3 lb.
Limits of 275 g L21 of VOC for coatings for wood furniture took effect in California in
1994. As VOC regulations become more stringent, it seems probable that a larger fraction
of furniture tops will be high pressure laminates with wood grain prints. Laminates have
been used for many years on commercial and institutional furniture; now their use in
household furniture is increasing.
There have been efforts for many years to replace nitrocellulose lacquers to achieve
higher solids. Alkyd – urea –formaldehyde (UF) top coats, frequently called conversion
varnishes or catalyzed finishes, have been used for many years. Tall oil alkyds with a
butylated –UF resin are typical. Just before use, one adds about 5% of p-toluenesulfonic
acid catalyst, based on UF resin solids. A flatting pigment dispersion and a small
amount of poly(dimethylsiloxane) to minimize orange peel (Section 24.2) are usually
added. The top coated furniture is force dried at 65 to 708C for 20 to 25 minutes. The
NVW is typically about 38%, about twice that of lacquers. Hot spray can further increase
the solids, but pot life must be watched carefully, and as the solids increase, low gloss
becomes more difficult to achieve. With conversion coatings the appearance of depth is
not as good as that obtained with a nitrocellulose lacquer, but is still presentable. The
coating is solvent resistant and more heat and gouge resistant than are lacquer coatings.
Repair is more difficult but is needed less frequently. Conversion varnishes are most
widely used on furniture for commercial use and for kitchen cabinets.
High-solids coatings based on isocyanate chemistry are a more recent developments.
Aldimine–isocyanate two package (2K) 90% solids gloss clear coats (Section 12.4) have
been formulated that have a pot life of about 3 hours and a dry time of about 1/2 hour
[4]. Moisture-curable urethane coatings are used in coating wood in applications where
high abrasion resistance is needed. Isocyanate prepolymer mixtures such as propoxylated
ethylenediamine/TDI and 2,2,4-trimethylpentane-1,3-diol/TDI prepolymers are used as a
vehicle for moisture-curable wood coatings [5]. A moisture-curable wood finish based on
a TMP/1,4-butanediol/IPDI prepolymer with 3-glycidyloxypropyltrimethoxysilane is
reported to cure by hydrolysis of the alkoxysilane groups and subsequent reactions not
only with other alkoxysilane groups but also between the silanol groups and isocyanate
groups [6].
Considerable effort has been expended on waterborne finishes for wood furniture.
Application of a waterborne coating directly on wood leads to excessive grain raising,
which may limit their use to applications where there is already a solventborne sealer
on the wood. Substantial reductions of solvent use are possible by emulsifying nitrocellulose (NC) lacquers into water [7]. Lacquers with VOC of 300 to 420 g L21, excluding
water, can be made compared with VOC on the order of 750 g L21 for conventional
NC lacquers. Solids of the NC solution internal phase can be maximized by using only
true solvents, esters and ketones. Short oil alkyds using ester solvents (other than the
small amount of xylene needed for azeotroping in producing the alkyd) are selected as
plasticizers. Water wet NC is used rather than conventional isopropyl alcohol wet NC.
A surfactant is incorporated in the lacquer as an emulsifying agent; sodium alkylphosphate
31.1. COATINGS FOR WOOD
625
surfactants are recommended. The advantage of easy strippability for repair is retained.
However, a long dry time and slow development of print resistance have limited their
use. Combined nitrocellulose/acrylic latex – based lacquers have been recommended.
Furniture sealer and top coat formulations with VOC in the range 240 to 400 g L21
have been described [8]. Film physical and appearance properties are reported to be
superior to those of acrylic latex counterparts.
Thermosetting latex top coats are also used. For two package coatings, a variety of
cross-linkers for hydroxy-functional and carboxylic acid –functional latexes can be used.
Hydroxy-functional latexes are readily prepared by using hydroxyethyl (meth)acrylate
as a co-monomer. The resulting latexes can be formulated with urea – formaldehyde
(UF) or melamine– formaldehyde (MF) resins as cross-linkers. UF resins tend to evolve
high levels of formaldehyde from hydrolysis; MF resins evolve lesser amounts of formaldehyde. MF resins are soluble in BA/HEMA/MAA latex polymers so that they act as
plasticizers to promote interdiffusion. As temperature increases, the interdiffusion rate
increases faster than the rate of cross-linking, until many of the potential reactions have
occurred, but at some point the XLD reaches a point at which that interdiffusion ceases.
The rate of cross-linking can be increased by the addition of a catalyst such as pTSA.
Although this permits shorter cure times, the pot life of the coating decreases and the
residual acid in the final film can lead to reduced stability of the films to hydrolysis
since the catalyst also catalyzes hydrolysis.
Water-reducible acrylic resin –based coatings (Section 8.3) have been used. At a
sufficiently high MW, they can be used as thermoplastic coatings. Thermosetting
resins can be cross-linked with polyaziridine cross-linkers (Section 17.7). There is
concern about the toxic hazard of aziridine cross-linkers in spray application. Methylated
urea – formaldehyde resins can be used as cross-linkers under force dry conditions.
Acetoacetoxy-functional water-reducible acrylic resins made with acetoacetoxyethyl
methacrylate as a co-monomer are cross-linked with diamines [9] (Section 17.6).
Two package waterborne polyurethane top coats (see Section 12.7.3 for a discussion of
various types) are being used increasingly, since they provide low VOC emissions with a
minimum of HAP solvents. The films have excellent abrasion resistance. The cost is relatively high, and special application equipment is required [10]. A statistical study of the
effect of several variables on the performance of hydroxy-functional acrylic latexes crosslinked with a nonionic hydrophilically modified polyisocyanate finishes for wood kitchen
cabinets has been published [11]. It was found that a high hydroxy content of the latex
(hydroxyl number 52), small particle size of the latex, core-shell preparation of the
latex, and low Tg (obtained by increasing level of coalescing solvent) enhanced performance. Pigmented waterborne 2K coatings formulated with an acrylic latex and HDI isocyanurate that has been hydrophilically modified by reacting with a sterically hindered amino
sulfonic acid salt is reported to disperse easily and give superior stain resistance compared
with other 2K waterborne urethane coatings and 2K solventborne urethane coatings [1].
Solvent-free, waterborne 2K epoxy – amine furniture coatings were compared to
solventborne coatings in a study sponsored by the U.S. EPA [12]. Many of the film properties were comparable, but appearance would limit the types of application for which this
waterborne coating would be satisfactory.
In Europe and to a more limited degree, in the United States, UV cure acrylic top
coats are used. A large part of European furniture is manufactured in flat sheets and
assembled after coating. Furthermore, the finish on much European furniture is glossy.
UV curing is ideal for such an application. The fast cure at low temperature is appropriate
626
PRODUCT COATINGS FOR NONMETALLIC SUBSTRATES
for wood. Early coatings were based on styrene – polyester resins, but acrylic coatings
were later adopted because they cure more rapidly and emission of styrene is avoided
(Section 29.2.4). They are applied before assembly of the furniture by roll coating on
tops already finished up to the top coat. The coatings are gloss or high semigloss, are
solvent resistant, and have excellent mechanical properties. They have the advantage of
little or no VOC emission. Their use in the United States has been limited, where most
furniture uses low semigloss coatings and is finished after assembly. Furthermore, as discussed previously, most U.S. furniture has a low gloss top coat, which is difficult with
solvent-free UV cure coatings. Waterborne UV cure coatings have recently been developed that permit formulation of low gloss coatings (Section 29.2.5).
31.1.2. Panel, Siding, and Flooring Finishes
Large quantities of coatings are applied to plywood and hardboard sheets to make
precoated interior paneling and exterior siding. A small part of the market is for finishing
expensive, high quality hardwood plywood used for paneling executive offices. This
paneling is finished in essentially the same way as that used for the highest end of the
wood furniture market. A variety of other products constitutes the bulk of the market.
One class of commodity wall paneling is made using three-ply plywood whose top
veneer is a relatively featureless, low cost luan. The panel is dried to remove surface
water, grooved to make a plank effect, and then sanded. A dark colored lacquer is
applied to the grooves by pinpoint automatic spray guns and the panel is dried. A
lacquer sealer colored with dyes is applied. The panel is sanded again and then printed
with two or three prints to give the appearance of walnut, rosewood, or some other attractive wood with prominent grain patterns. Finally, a low gloss top coat is applied. The
coatings have generally been nitrocellulose lacquers. Although this business was developed in the United States, most production has shifted overseas, initially to South
Korea and more recently to the Philippines. The predominant source of the wood is the
Philippines, so finished product rather than raw material can be shipped; also, air pollution
regulations in South Korea and the Philippines are less stringent.
A related but quite different application is for “door skins.” In the manufacture of
interior doors, a common approach is to assemble the door and then laminate a top
veneer (the skin) to the particle board surface. The skins are veneer, laid up and adhered
to kraft paper. The veneer is commonly birch or luan, sometimes printed with a hardwood
grain pattern. Commonly, the coating on door skins is gloss or high semigloss. Whereas
low gloss is preferred on wall paneling to avoid glare from reflection of lights, higher
gloss is preferred on doors because it is easier to clean. Since the surface area of doors
is relatively small, glare is not a serious problem. This combination of factors has led to
the widespread use of UV cure coatings for door skins.
A variety of UV compositions has been studied for wood coatings. For example,
C6F13(CH2)2OCOCH55CH2 as a monomer in an epoxy – acrylate oligomer as a UV cure
coating for wood paneling provides high hydro- and oleo-phobicity, improved chemical
resistance, and better scratch resistance then those of coatings without the fluorine
monomer [13].
Part of the interior paneling market is precoated hardboard. Hardboard is made by
mixing wood fibers and shreds, then curing in hydraulic presses. The lignin in the wood
acts as a binder to hold the wood fibers together; sometimes the lignin is supplemented
with phenolic resins. Depending on the pressure applied, the density of hardboard can
31.1. COATINGS FOR WOOD
627
be varied. Furthermore, patterns can be embossed into the surface of the hardboard. A
variety of hardboards are made. Using smooth-surface hardboard, one can apply a base
coat, prints, and a semigloss top coat to make 4 by 8 foot panels that look like any kind
of wood. Smooth surface hardboard is often coated by curtain coating (Section 23.5),
since this gives level films.
Tongue and groove paneling can be simulated by routing out grooves and painting them
a dark brown color. The paneling is finished by roll coating a base coat that does not flow
into the grooves appreciably, followed by printing and top coating. A pattern of ticks
(short line dents scattered over the surface, as in hardwoods) can be embossed. When
the board is finished, it not only has the grain of the wood but also has the little dips of
a tick pattern as well as the grooves to give the effect of planked paneling.
One can simulate wood with holes, such as pecky cypress. The hardboard is embossed
with the hole pattern copied from real pecky cypress paneling. The hole coat, generally a
dark brown, is applied using roller brushes to assure that the coating covers surfaces inside
the holes completely. Then a base coat is applied by precision coating. A precision coater
is a roll coater in which the application roll is a gravure printing roll covered uniformly
with cells so that the upper surface of the panel is coated but no coating is applied in
the holes. Generally, the color is a relatively light brown to provide a contrasting background color. Then grain pattern prints are applied, followed by a transparent semigloss
top coat. One is not restricted to replicating pecky cypress or wormy chestnut; brick,
stone, travertine marble, and other materials can be simulated. Other interior hardboard
paneling is designed primarily for bathrooms. The coatings are high gloss pigmented coatings. Sometimes, joints are embossed in to make the panels look like tiles. Sometimes,
prints such as marble are used with a clear gloss top coat.
Since, in contrast to plywood, hardboard can withstand high temperatures, baking
coatings, with all their advantages in performance properties, are used. Since unlike
wood, hardboard does not undergo grain raising, water-reducible acrylic – MF finishes
are used. Although the average film thickness of a base coat is relatively high, popping
is not a major problem, since the coating thicknesses applied by roll coating and curtain
coating are more uniform than when coatings are applied by spray (Sections 23.4 and 23.5).
High-density hardboard is also used for exterior siding. The largest volume of siding is
factory primed with a primer designed to have at least 6 months of exterior durability
before exterior house paint is applied (Section 32.1). Acrylic – MF binders are used in
the primers, generally waterborne to reduce VOC. Challenges in formulating these
primers are to be sure that the coating cross-links sufficiently that the coated boards
coming out of the oven can be stacked without blocking and also that the paint to be
applied to them in the field will have good adhesion. The combination requires careful
control of cross-link density. The primers are low gloss because high loading with inert
pigment reduces cost and enhances adhesion of paint to the surface. The largest volume
is used in siding for tract houses. The primed siding is erected onto the house, but painting
is postponed until the house is sold so that the buyer can pick the paint color. An extensive
study of exterior durability of different hardboards with different combinations of coatings
has been published [14]. The best results were obtained when preprimed board was field
coated with an alkyd primer and acrylic latex top coats. Petrolatum and other oily
substances are present in some hardboards. If the paint applied on the siding is porous,
over time some substances can migrate to the upper surface of the paint, leading to
discoloration or sticky areas that will pick up dirt to give a blotchy appearance. The
problem can be minimized by using only paints with PVC , CPVC.
628
PRODUCT COATINGS FOR NONMETALLIC SUBSTRATES
Fully prefinished exterior siding is manufactured on a smaller scale than preprimed siding.
The primed board is top coated in the factory with a baking acrylic enamel or, for greater
exterior durability, a silicone-modified acrylic resin–based coating (Section 16.1.2). The
durability is superior to field applied paint, but flexibility in color selection for tract homes
is sacrificed. Fully prefinished siding is more commonly used on commercial buildings.
Kitchen cabinet waterborne coatings are formulated with hydroxy- and carboxyfunctional acrylic latexes with a urea –formaldehyde cross-linker. While some coalescing
solvent is needed, VOC emissions are much lower than with solventborne coatings. The
variables in formulating such coatings are discussed in Ref. [15].
Waterborne UV cure clear coats for application to wood articles such as window frames
having good exterior durability have been developed. A polyurethane dispersion (PUD)
terminated with acrylate groups made with aliphatic isocyanates is used with low MW
acrylate coreactants [1]. Waterborne UV cure acrylate-functional PUD give coatings
with satisfactory adhesion [17].
Nano-aluminum oxide (average diameter 13 nm) markedly improves the scratch resistance of UV cure wood flooring coatings, The Al2O3 is dispersed in the UV curing coating,
which also includes 3-trimethoxysilylpropyl methacrylate as a coupling agent between the
pigment and binder [18].
Moisture-curable urethane coatings are used for applications such as floor coatings,
for which abrasion resistance and hydrolytic stability are important. They are called
urethane coatings even though these cross-links are urea rather than urethane groups.
Since urea groups also form intermolecular hydrogen bonds, presumably they can affect
resistance to mechanical stress in a manner similar to urethane groups (Section 12.6).
Another wood flooring coating is based on a urethane – acrylic PUD and pentaerythritol
tris(b-(N-aziridinylpropionate) [16]. Waterborne wood floor finishes prepared using
drying oil– modified PUDs show a combination of abrasion resistance and resistance to
heel marking. 2K waterborne urethane coatings in which the urethane cross-linker is a
hydrophilically-modified allophanate polyisocyanate are useful in wood finishing.
Uralkyds have been widely used for finishing wood flooring and other wood products
because of their abrasion resistance and ease of application. However, they have a VOC of
over 500 g L21. New regulations in southern California and the northeastern states require
275 g L21. This VOC can only be attained with waterborne resins. (See Section 15.7 and
Ref. [19] for a discussion of such coatings.)
31.2. COATING OF PLASTICS
Plastics have become a major substrate for coatings. Design of the coatings is complicated
by the variety of polymers used in plastics and the range of approaches that are possible for
decorating the surfaces. Reference [20] has several useful chapters in which various
approaches are discussed; coatings for plastics are reviewed in Ref. [21]. For molded plastics there are two alternatives: (1) in-mold coating, in which the inside of a mold is coated,
the plastic material is put into the mold, and the outer surface is the coating when the part is
taken out of the mold; or (2) post-mold coating, in which the molded part is coated.
31.2.1. In-mold Coating
Polar thermosetting plastics can be in-mold coated; a variety of coating materials is used
for in-mold coating, depending on the plastic. To assure good adhesion to the plastic, it is
31.2. COATING OF PLASTICS
629
desirable to use compositions related to the composition of the plastic: for example,
the use of gel coats with glass-reinforced plastic parts. Gel coats are pigmented
styrene – unsaturated polyester resins (Section 17.3) that are cured along with the resin
in the molding compound so that the coating is chemically bound to the main body of
the plastic. The molding resin is usually a styrene solution of a phthalic anhydride,
maleic anhydride, propylene glycol polyester. For some products, similar polyesters are
used in the gel coats. More expensive isophthalic, maleic anhydride, neopentyl glycol
resins are used in gel coats when superior hydrolytic and exterior durability are needed.
In-mold coating with gel coats is used in making glass reinforced styrene – polyester
molded products such as boats and shower stalls.
With polyurethane RIM (reactive injection molding) parts, in-mold coatings with free
hydroxyl groups permit reaction with isocyanate groups in the molding compound. Rigid
urethane steering wheels are made by coating the inside of the mold with the appropriate
color enamel. Rigid urethane foam parts to replace wood carvings for furniture are in-mold
coated. The inside of the mold is sprayed with a lacquer which is color-matched to the base
coat of the furniture. In the past many in-mold coatings were solventborne coatings,
frequently low solids lacquers. These high VOC coatings are being replaced with
higher solids coatings, waterborne coatings, and powder coatings.
Coating of interior automotive plastic components has for many years been done by
post-mold coating. Waterborne in-mold coating is now beginning to be introduced. It is
reported that the new technology achieves rapid part throughput, high transfer efficiency,
greater gloss control, and superior mar and chemical resistance [22].
For some plastic automobile parts that are to be assembled onto the steel body of the car
and painted with the same top coat as the rest of the body, the interior of the mold can be
coated with a primer to provide adhesion to the plastic and intercoat adhesion with the top
coat. Sometimes, such a primer is made with conductive pigments, such as acetylene
carbon black, so that electrostatic spray will be effective over the plastic parts.
A further alternative approach is to use coated films as a laminate for in-mold application of coatings [23]. The coating is reverse roll coated on a smooth, heat resistant,
high gloss polyester film. A size coat is applied to the surface of the coating, and then a
film of plastic of the same polymer (or a compatible polymer) as the plastic part to be
coated is laminated. The resulting coated film is vacuum formed into a mold, coated
side out, followed by a sheet of the plastic. The part is removed from the mold and the
polyester film is stripped off, leaving the coating on the surface of the plastic substrate.
Multiple layers of coating can be applied. For example, for finishing plastic parts for
use with base coat –clear coat systems, the polyester film would be coated first with the
clear coat, followed by the base coat, and then laminated to the plastic film.
31.2.2. Post-mold Coating
Coatings are also applied to plastic articles after they are fabricated. Many of the coatings
used for metals can be adapted for certain plastics, and many of the selection criteria are
the same as for metals. However, there are important differences. Adhesion (discussed in
Section 6.5) is usually the central problem in coating plastics. The coatings must be more
flexible than those for metal. This follows from the easier deformability of many plastics
and all elastomers; the coatings must be at least as easily deformed as the substrate. As a
rough guideline, elongation-at-break (Section 4.2) of the coating should be greater than
that of the plastic substrate.
630
PRODUCT COATINGS FOR NONMETALLIC SUBSTRATES
To obtain good adhesion, the surface of the plastic must be clean. Machining oils, sanding
dust, finger prints, and so on, must be removed. Many plastic parts have residual mold
release agents on the surface. If a mold release must be used in making the plastic article,
it is desirable to use a water-soluble soap such as zinc stearate as a mold release, since it
is relatively easy to remove. Wax mold release agents are more difficult to remove, and
silicone or fluorocarbon mold releases should not be used for plastic parts that are to be
painted. Surface contaminants and mold release are removed by spray washing in three
stages. First, a detergent wash is used followed by a water rinse, and finally, by a deionized
water rinse. At the end of the washer, droplets of water are removed by compressed air jets,
and the part goes into a drying oven before coating.
After the surface is clean and dry, the surface tension varies widely, depending on the
type of plastic. In general, thermosetting plastics and thermoplastics with polar structural
groups, such as nylons, have relatively high surface tensions, although lower than those of
metals. Often, they can be coated without further surface treatment. However, the surface
tension of some less polar plastics, particularly polyolefins, is lower than that of most coatings, and the polarity of the surfaces is very low. Primers are required that have very low
surface tensions and that can penetrate the surface of the plastic. Alternatively, the surface
of the plastic can be treated to increase its surface tension and provide polar groups that
can hydrogen bond with components of the coating to promote adhesion (Section 6.5).
A very challenging problem in coating plastics is the finishing of plastic bumpers,
grills, and other exterior plastic components of automobiles. Papers of a symposium
on painted plastic components published in JCT Research provide an excellent
summary of the present status of the work [24]. The components are molded from a
polypropylene – ethylene/propylene/butadiene rubber blend (called TPO in the United
States and PP-EPDM in Europe and Asia). Due to the low surface tension of the crystalline
polypropylene surface, coatings will not adhere well, so a surface treatment is required.
Adhesion can be achieved by applying a thin coat of a chlorinated polyolefin (CPO)
primer. CPOs are prepared by grafting maleic anhydride to an isotactic polypropylene
and chlorinating the resulting graft copolymer. A wide variety of CPOs is available
with different MW, maleic content, and degree of chlorination. Most primers have been
solventborne, incorporating acetylene black to make the coating conductive to permit
electrostatic spraying of a base coat on the primed part.
One of the papers in the symposium gives an excellent presentation of efforts to understand the mechanism by which the CPO primers promote adhesion [25]. The chlorinated
polymer is thought to migrate into the surface of the polyolefin, resulting in a surface to
which a broader range of coatings can adhere. A large number of variables can affect
the performance. Injection molded plastics tend to provide better adhesion than compression molded plastics. The structure of the CPO, MW, degree of chlorination, maleic
content, and film thickness are important variables. The solvent used can make a major
effect on the extent of penetration; hydrocarbons give greater penetration than do polar
solvents. Baking generally promotes penetration. Care must be exercised to assure that
the solvent is all evaporated before the base coat is applied, to prevent it blowing out
through the film when the top coats are baked. Although solvents can improve adhesion,
care must be exercised in the selection of solvents for plastics such as TPO (thermoplastic
polyolefin blended with rubber) so that the crystallinity of the plastic is not disturbed, since
that would affect the cohesive strength of the plastic [26]. The TPO primer films must be
on the order of 3 to 5 mm thick; the primer is thermoplastic, and thicker films may lead to
cohesive failure. The need for thin films dictates that the primer must be applied in high
31.2. COATING OF PLASTICS
631
VOC formulations. The crystallinity and cohesive strength of TPO is also affected by the
baking temperature used for curing the coatings [27]. It is also reported that a solution of
chlorinated polyolefin can be added to the top coat to obtain adhesion to TPO without the
need for a separate primer [28]. (See Section 6.6 for a discussion of testing methods.)
Adhesion of top coat to the primer is sometimes inadequate. Addition of ethylene/vinyl
acetate copolymer to the primer is reported to improve intercoat adhesion [29]. Another
approach is to graft acrylic side chains by esterifying some of the maleic groups with
HEA and then making a graft copolymer with acrylate monomers. 2K urethane base
coats are formulated with the resin and are reported to resist delamination without the
use of a separate primer [30].
Automobile base coat and a clear coat are applied over the primed TPO. For example,
an acrylic – MF base coat with a trimethoxysilyl-functional acrylic clear coat can be
applied. A UV absorber and HALS compound are included in a 2K urethane clear coat
to assure exterior durability. However, durability is not as good as that obtained with
the same system over steel. It has been found that the UVA and hindered amine light
stabilizer (HALS) can migrate through the coatings and down into a TPO plastic, resulting
in poorer durability. To minimize this problem, the use of UVAs and HALS with hydroxyl
groups that will react with an isocyanate in the clear coat to bind the stabilizers to the
polymer chains leads to much higher retention of stabilizers in the clear coat and improved
durability [31].
In baking automotive clear coats for plastics, methoxysilyl groups of the trimethoxysilylfunctional acrylic resins hydrolyze at room temperature to silanol groups, which react during
baking to cross-link the coating. The scratch resistance before and after weathering is
superior to that of MF cross-linked coatings [32].
A waterborne coating for poly(phenylene oxide) plastic interior auto parts made with a
urethane – acrylic hybrid PUD and two acrylic latexes has been disclosed [33,34]. A waterborne primer, formulated with graft copolymers of acrylic monomers on polypropylene,
enhances adhesion to polypropylene (or blends of polypropylene/ethylene –propylene)
rubber, as well as to acrylic top coats [35].
Instead of applying a special primer, the surfaces can be oxidized to yield polar groups;
there are several approaches to oxidation treatment [36]. Oxidizing agents such as chromic
acid/sulfuric acid baths are effective and have been used for many years. However, disposal of chromate containing wastes is now tightly controlled. To avoid Cr(VI), sodium
hypochlorite with a detergent has been recommended for treating the surface of polyolefin
polymers to improve adhesion [33]. Alternatively, a surface can be oxidized by directing
an oxidizing flame from propane or butane onto the surface. Care is required to ensure that
all surfaces are adequately treated. Flame treatment is widely used in Europe for the
surface treatment of plastic parts for automobiles. Yet another approach to oxidation of
the surface is by corona discharge. The plastic parts are passed through a cloud of
ionized air generated, as described in Section 6.5, by electrodes with many wire ends at
high voltage. The ions oxidize the surface of the polyolefin. For irregular shaped parts,
the corona discharge can be carried out in a vacuum chamber; this method is frequently
called plasma discharge. Still another approach is to spray a solution of benzophenone
on the surface of a polyolefin and then expose the part to UV radiation. This leads to
the generation of free radicals that initiate autoxidation of the surface. This process is particularly effective with rubber-modified polyolefins. A drawback of these four approaches
is that they do not make the surface conductive enough for efficient electrostatic spraying.
In such cases, TPO primers with conductive pigments are preferred.
632
PRODUCT COATINGS FOR NONMETALLIC SUBSTRATES
From the standpoint of adhesion, it is preferable for the plastic part to have a somewhat
roughened surface. Plastics that are highly filled with pigments have rougher surfaces to
which it is easier to adhere. Alternatively, the surface of the mold can be designed to
impart some roughness to the molded pieces. Adhesion can be promoted by baking at
temperatures above the Tg of the plastic. However, in many cases, heating above Tg
distorts the molded part. This is one of the advantages of in-mold coating: Adhesion is
promoted by the heat involved in molding the plastic, avoiding the problem of heat distortion when baking post-coated plastic moldings.
Adhesion can also be promoted by using solvents in the coating that are soluble in the
plastic. This reduces the Tg of the plastic at the surface, permitting penetration of binder
from the primer into the surface of the substrate. However, very volatile strong solvents
must be avoided on high Tg thermoplastics such as polystyrene and poly(methyl methacrylate); such solvents cause crazing, that is, the development of a network of microcracks in
the surface of the plastic. A possible explanation for crazing is that the solvent penetrates
into the plastic; then, if it is very volatile, it evaporates rapidly from the surface while there
is still solvent left just below the surface. When the solvent diffuses out from below the
surface, there is a reduction in volume, leading to stresses sufficient to crack the surface
layer of the plastic. Another potential problem with using penetrating solvents is the
possibility of solvent popping caused by the release of solvents from the plastic after crosslinking of the coating is well advanced. In a study of factors that affect popping on sheet
molding compound parts (molded fiberglass reinforced polyester plastics), it was found
that the problem can be minimized by using primers with the lowest possible solvent
permeability [37]. Permeability can be reduced by using highly cross-linked or partially
crystalline binders in the primer.
Moisture-curable urethane coatings have been used on plastic flooring. For example, a
coating with high abrasion resistance, particularly for urethane elastomer flooring, is
formulated with a moisture-curable urethane made from polytetramethylene glycol,
1,4-butanediol, H12MDI, and MDI [38]. Just before application, DBTDL and DABCO
solutions are added as catalysts.
Electrostatic spraying of plastics is difficult, since the plastics are usually insulators and
therefore cannot be adequately grounded to give the charge differential needed to attract
charged spray droplets to the surface. Surface resistivity of plastics and time for charges
on their surfaces to dissipate can be measured and correlated with the feasibility of electrostatic spraying [39]. The charge dissipation time is affected not only by the composition of
the plastic, but also by the humidity in the spray booth. Conductive tie coats or primers can
be applied to the plastic; the effect of a range of conductive primers on the charge dissipation time has been studied [26]. It has also been shown that the effect of surface charges on
the plastic substrate in the electrostatic spraying can be reduced by placing a continuous
grounded metal backing in contact with the plastic substrate [40]. When the object
being sprayed has both plastic and metal components, there can be nonuniform deposition
of coatings even when a conductive primer has been applied to the plastic, since the metal
can distort the electrostatic field near the plastic – metal interfaces.
The lower curing temperatures possible with two package polyurethane coatings make
them particularly attractive for many applications, to minimize the possibility of heat
distortion of the plastic product [41]. The higher abrasion resistance and flexibility available with urethane coatings are frequently also advantageous. While acrylic resins can be
cross-linked with polyisocyanates, higher solids are possible with hydroxy-functional
polyesters. For example, a polyester resin prepared from adipic acid, isophthalic acid,
31.3. COATINGS FOR GLASS
633
neopentyl glycol, and trimethylolpropane in a mole ratio of 1 : 1 : 2.53 : 0.19 with an M̄n of
730 and an f̄n of 2.09 hydroxyl groups per molecule was cross-linked with various trifunctional aliphatic isocyanates at a 1.1 : 1 NCO/OH ratio [41]. The coating provided a good
balance of properties on a variety of plastics, including good impact resistance at 2298C.
Another approach to coating plastics subject to heat distortion is radiation curing
(Chapter 29). The temperatures involved in radiation curing are generally not substantially
above ambient and can be further reduced by filtering out IR radiation. For example,
UV-cure gloss top coats are applied to plastic flooring. The low temperature curing is particularly important in coating flooring such as vinyl tile. Plastic flooring often has an
embossed foam decorative layer as the upper surface; elevated temperatures could
collapse the foam.
Dual UV/thermal cure primer-sealer coatings for unsaturated polyester – styrene glass
fiber sheet molding compound (SMC) automotive components have been developed to
solve the serious popping problem that occurs when solventborne coatings are applied
and baked. As the coated part is baked, gas from pores in the plastic and solvent from
the coating lead to popping. The problem has been solved by applying a primer–sealer
to the SMC components based on a 2K urethane coating with acrylate functionalities.
The coatings are exposed in a UV chamber to a maximum temperature of 658C and
then baked at 1218C for 20 minutes. In the UV chamber, the components are hung on a
conveyor with UV lamps at several angles so as to minimize shadow areas [42]. Adhesion
is facilitated by the much lower shrinkage than in traditional UV curing since the fraction
of unsaturation is low due to the relatively high MW resin.
Use of some solvent in UV cure coatings permits using less low viscosity monomeric
acrylates, thereby reducing shrinkage during cure and minimizing the adhesion problem.
Proprietary UV cure base coats and clear coats for flexible plastic automotive components
reduce the time for application and cure to 7.5 minutes, compared to 35 minutes for
thermal cure coatings [43].
In many applications, the function of the coating is decorative, such as for color; but
there are many examples of functional coatings for plastics. The UV cure top coat
applied to plastic flooring mentioned earlier has the functions of increasing wear life,
stain resistance, and gloss retention. Coatings are applied to polyethylene tanks to
reduce permeability. Magnetic coatings are applied to tapes and sheets to make recording
tapes.
Transparent plastics can serve as replacements for glass for applications ranging
from window glazing to eyeglasses, but they are less abrasion resistant than glass
(Section 4.4.1). Surface coatings to improve abrasion resistance have been developed
based on alkoxysilanes and colloidal silica [44]. The coatings provide excellent abrasion
resistance to plastics such as polycarbonate, but have relatively high VOC levels and long
curing times. Radiation cure coatings have been developed with low VOC that cure in a
few seconds (Section 29.6).
31.3. COATINGS FOR GLASS
In general, it is difficult to achieve adhesion to untreated glass such that the adhesion
resists soaking in water or strong bases. Such adhesion is possible with the use of reactive
silanes as discussed in Sections 6.4.3 and 16.2 [45]. For example, an aqueous polyurethane
dispersion (PUD) having blocked NCO groups together with a hydroxy-functional PUD
634
PRODUCT COATINGS FOR NONMETALLIC SUBSTRATES
and silane additives is used for coating glass bottles. The coating is baked at 1908C for
20 minutes and resists loss of adhesion when the coated bottle is soaked in water or in
a solution of NaOH [46]. Because NaOH solutions are used to clean recycled bottles,
this resistance is critical.
A commercially important application for coatings on glass is coating optical fibers for
wave guides used in telecommunication cables (Section 29.6).
GENERAL REFERENCE
Ryntz, R. A., Adhesion to Plastics: Molding and Paintability, Global Press, Moorhead, MN, 1998.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Irle, C., private communication, Bayer Material Science, June 2005.
Morrison, S., http:/specialchem4coatings (2005).
Parker, A. A.; Martin, E. S., J. Coat. Technol., 1994, 66(829), 39.
Dvorchak, M. J.; et al., Coat. World, 1997, November – December, 28.
Pedain, J.; et al., U.S. patent 4,727,128 (1989).
Fong, J. J. J., U.S. patent 5,124,210 (1992).
Winchester, C. M., J. Coat. Technol., 1991, 63(803), 47.
Haag, H. F., J. Coat. Technol., 1992, 64(814), 19.
Esser, R. J.; et al., Prog. Org. Coat., 1999, 36, 45.
Dvorchak, M. J., J. Coat. Technol., 1997, 69(866), 47.
Feng, S. X.; et al., J. Coat. Technol., 1999, 71(899), 51.
Huang, E. W.; et al., Coat. World, 1997, November– December, 21.
Bongiovanni, R.; et al., Prog. Org. Coat., 2002, 45, 359.
Bailey, W.; et al., J. Coat. Technol., 1990, 62(789), 133.
Howard, C.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2001,
pp. 490 –503.
Kohman, R. G.; et al., U.S. patent 6,444,134 (2002).
Irle, C.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2001,
pp. 455 –463.
Landry, V.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2006,
pp. 225 –235.
Caldwell, R. A., JCT Coat. Tech, 2005, 2(3), 30.
Satas, D., Ed. Plastic Finishing and Decoration, Van Nostrand Reinhold, New York, 1986.
Ryntz, R. A., Painting of Plastics, Federation of Societies for Coatings Technology, Blue Bell,
PA, 1994.
Cannon, K.; Young, D. L., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 2005, pp. 134 –143.
Fridley, C. H., Proc. Soc. Manuf. Eng. Finish.’91 Conf., FC91-374, 1991.
JCT Res., 2005, 2, 351 – 416.
Ryntz, R. A., JCT Res., 2005, 2, 351.
Ryntz, R. A.; et al., J. Coat. Technol., 1995, 67(843), 45.
REFERENCES
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
635
Ryntz, R. A., Prog. Org. Coat., 1996, 27, 241.
Tech. Bull. GN-4-88, Eastman Chemical, Kingsport, TN, 1998.
Tech. Bull. GN-4248, Eastman Chemical, Kingsport, TN, 2004.
Masuda, T.; et al., U.S. patent 6,861,471 (2005).
Yaneff, P. V.; et al., JCT Res., 2004, 1, 201.
Yaneff, P. V.; et al., Proc. Waterborne Higher-Solids Powder Coat. Symp., New Orleans, LA,
2001, pp. 109 – 125.
Mori, N.; et al., U.S. patent 6,465,563 (2002).
Haag, H. F., U.S. patent 5,053,256 (1991).
Hintze-Bruening, H.; Borgholte, H., Prog. Org. Coat., 2000, 40, 49.
Lane, J. M.; Hourston, D. J., Prog. Org. Coat., 1993, 21, 269.
Ryntz, R. A.; et al., J. Coat. Technol., 1992, 64(807), 29.
Harmer, W. L., U.S. patent 4,273,912 (1981).
Garner, D. P.; Elmoursi, A. A., J. Coat. Technol., 1991, 63(803), 33.
Elmoursi, A. A.; Garner, D. P., J. Coat. Technol., 1992, 64(805), 39.
Shoemaker, S. H., J. Coat. Technol., 1990, 62(787), 49.
Joesel, K. H., 7th Int. Coat. Plast. Symp., Troy, MI, 2004.
Cannon, K.; et al., Proc. Waterborne Higher-Solids Powder Coat. Symp., New Orleans, LA,
2005, pp. 34– 42.
Blizzard, J. D.; et al., Proc. Waterborne Higher-Solids Powder Coat. Symp., New Orleans, LA,
1992, p. 171.
Pluddemann, E. P., Prog. Org. Coat., 1983, 11, 297.
Martin, L.; et al., Proc. Waterborne Higher-Solids Powder Coat. Symp., New Orleans, LA,
2000, pp. 373 – 382.
32
Architectural Coatings
In 2002, U.S. shipments of architectural coatings had a value of $7.211 billion [1]. This
was about 45% of the value of all U.S. coatings shipments. Architectural coatings are
also called trade sales paints. There are two overlapping markets for these paints: contractors and do-it-yourselfers, who paint their own homes and furniture. The contractor is
particularly concerned with the cost of application, which is greater than the cost of the
paint. For example, contractors want paints that cover in one coat and touch up readily.
The do-it-yourselfer is more likely to be concerned about ease of cleanup, odor, and the
range of colors available. Concern with the cost of paint itself varies significantly
within both groups. In the United States, paint is sold through several distribution
systems: large merchandisers, hardware and lumber stores, paint stores owned or franchised by paint companies, and independent dealers. Some paint is sold directly to
large contractors by paint manufacturers. There are three, sometimes overlapping
classes of paint companies: large companies that do extensive national advertising to
promote trade name paints; smaller regional companies that sell only locally, primarily
to local contractors but commonly with regional stores for sales to contractors and individuals; and companies that sell private-label paints to large merchandisers or hardware
chains, primarily for do-it-yourself use. Some paint companies design their paints to be
most suitable for the climate and style trends in specific regions of the United States.
Large companies generally make several lines of the major product types: for example,
good, better, and best. The good grade is designed for the person whose principal criterion
for selecting paint is price per unit volume. Such paint is usually adequate for the application with fair coverage. But buying the cheapest paint is usually a false economy.
The best paint is usually designed to give the longest life and to have as good coverage
as can be designed into that class of paint; it may also be easier to apply and have
better appearance. The “better” paint is a compromise between the two.
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
636
32.1. EXTERIOR HOUSE PAINT
637
Although some colored paints are produced at the factory, the majority are manufactured as tintable paints that are custom-tinted at the point of sale. The manufacturing
company supplies an extensive line of color chips or standards, with liquid tinting
colorants and formulations to use with paints called bases, in order to match the color
chips. This makes it possible for the paint store to provide thousands of colors with
only a limited inventory.
A multitude of products is made for the architectural market. We restrict discussion to
the three largest classes of paints: exterior house paint, interior flat wall paint, and gloss
enamels. Space limitations do not allow inclusion of smaller but still significant product
lines, such as stains, varnishes (Sections 14.3.2 and 15.7), floor paints, and many specialty
products.
32.1. EXTERIOR HOUSE PAINT
An extensive discussion about finishing wood with many types of coatings is provided in
Ref. [2]. This section is limited to exterior house paint, that is, paint applied to exterior
wall surfaces. Paint for trim around windows, doors, and on shutters is generally semigloss
paint or gloss enamel, discussed in Section 32.3. The great majority of exterior house paint
sold in the United States is latex paint. The exterior durability of latex paints is superior to
that of air dry solventborne house paints. Latexes are made by emulsion polymerization
(Chapter 9) and undergo film formation by coalescence (Section 2.3.3). Solventborne
house paints are used if outside painting must be done when the temperature is below 1
to 78C (the minimum application temperature is given on the label), since latex paints
do not coalesce properly at such low temperatures. Latex paints do not adhere well to a
chalky surface, that is, a surface with a layer of poorly adherent pigment and eroded
old paint; careful washing to remove chalk is required. As of 1999, EPA national
maximum VOC emission for exterior flat paint was 250 g L21 [3]. (See Section 18.9.2
for further information about EPA and California Air Resources Board regulations.)
Latex paints perform better than oil or alkyd paints on wood siding. When oil and alkyd
paints were widely used on wood house siding, paint failures by blistering were common.
Water can get into siding from the back side. Since films from oil and alkyd paints have
low water vapor permeability, when water reaches the back side of a film and starts volatilizing in the heat from the sun, blisters are formed. Latex paints are not cross-linked and
have Tg values below summer temperatures and therefore have relatively high moisture
vapor permeabilities. When water gets to the back of a latex paint film, the water vapor
can pass through the film; blistering is unusual.
Latex paints are more resistant than oil or alkyd paints to grain cracking. Wood expands
and contracts as its moisture content increases and decreases. A coating applied with an
acrylic latex binder maintains its extensibility after many years of exterior exposure and
can expand and contract with the wood. Highly unsaturated oil and alkyd coating films
continue to cross-link with exterior aging, becoming less extensible and more likely
to crack as the wood expands and contracts. Oil and alkyd paints also commonly fail
by chalking after exposure. High quality latex paints are less likely to fail by chalking
or, at least, have longer lifetimes before chalking occurs to a serious extent.
Leveling of latex paints is generally not as good as that of solventborne paints. This is
not viewed as too serious a problem in exterior house paints, but it is with gloss paints, as
638
ARCHITECTURAL COATINGS
discussed in Section 32.3. This disadvantage is offset in exterior house paints by the
accompanying superior sag resistance of latex paints compared to solventborne paints.
Since they remain thermoplastic with aging, latex paints are more prone to retain dirt
on their surface than are oil or alkyd paints. Dirt pickup is undesirable anywhere, but is a
particular problem in places where lignite and soft coal are used for heating and cooking,
where white latex paint turns an unsightly blotchy gray after only a few months exposure.
As discussed in Section 2.3.3, film formation of latex paints occurs by coalescence of the
polymer particles. Coalescence requires that the temperature during film formation be
above the Tg of the particles. If the paint is to be applied at low temperature, the Tg
must be correspondingly lower. However, this means that when the temperature is high
during the summer, the film is soft and susceptible to dirt pickup, even when the dirt
content in the air is not excessive. Dirt pickup decreases as Tg is increased from 08 to 108C
and styrene–acrylic latex paints show less dirt pickup than straight acrylic latex paints [4].
However, they tend to chalk in sunny exposures of the southern latitudes of North
America, and thus have not been employed extensively in the United States. Lower-gloss
paints generally have lower dirt pickup than that of higher-gloss paints.
The condition of the wood surface to be painted can be a significant factor in the performance of paint. Contamination with dirt and oil can interfere with adhesion. When bare
wood is exposed outdoors, the surface degrades quickly. Painting degraded surfaces can
result in what looks like adhesive failure but is actually cohesive failure of the wood. It
is reported that even a three to four week exposure of freshly cut wood to weather can
affect adhesion adversely [2]. Many woods discolor when exposed to UV; this is particularly noticeable in outdoor exposure but happens more slowly indoors. Two coat finishes
have been developed that markedly decrease the discoloration of wood. The first coat next
to the wood contains a nitroxy HALS (Section 5.2.3), and the second coat contains a combination of bis-, tris, and tris-para-substituted/tris-resorcinol-triazines as UV screeners
(Section 5.2.1) [5].
It is recommended that before painting new wood, joints and cracks be caulked and that
the wood be treated first with a paintable water-repellent preservative and then primed
with an alkyd or latex primer. On many kinds of wood, a latex paint can be used for
both the primer and top coat. However, woods such as redwood, cedar, and some pines
contain water-soluble materials that extract into a latex paint, leading to reddish-brown
stains. The extractives are naturally occurring phenolic compounds. Special stain-blocking
latex primers are made for use over redwood and cedar. Formerly, these primers were formulated with a somewhat soluble lead pigment, which formed insoluble salts with the
phenolic compounds. To avoid lead compounds, other ways to insolubilize the phenolics
have been found. One approach is to include in the formulation a cationic ion-exchange
latex. Many houses are sided with preprimed hardboard siding (see Section 31.1.2 for a
discussion of painting this siding).
As mentioned earlier, another potential problem in applying latex house paints can be
poor adhesion to chalky surfaces. The surface of chalky paint is covered with a layer of
loosely held pigment and paint particles. When latex paint is applied over such a
surface, the continuous phase of the paint penetrates among the particles down to the substrate, but latex particles are large compared to the interstices among the chalk particles
and do not penetrate significantly. When the water evaporates and coalescence occurs,
the paint film penetrates the chalk particles only partially, with limited binding of the
chalk particles together and to the substrate. As a result, adhesion can be poor. Latex
chalk sealers made with very small particle size latexes are available.
32.1. EXTERIOR HOUSE PAINT
639
Alternatively, one can use an oil- or alkyd-based primer. With oil or alkyd paints, the
continuous phase penetrates among the chalk particles; the binder from the paint surrounds
the chalk particles and penetrates through them down to the substrate surface, minimizing
the adhesion problem. Adhesion of latex paints can be improved by careful cleaning of the
surface to be painted to remove all chalk. For large areas such as entire wall sections of a
house, power washing is effective for removing chalk as well as dirt and loose paint.
Another formulating method is to replace about 15% of the latex polymer binder with
long oil alkyd resin or synthetic drying oil. The alkyd or oil is emulsified into the latex
paint. After application, as the water evaporates, some of the alkyd or oil can penetrate
between the chalk particles, providing improved adhesion. For the paint to have a good
storage life, the alkyd or oil should be hydrolytically stable. The drying oil or alkyd
reduces the exterior durability of the paint compared to pure latex paints, especially in
resistance to chalking, fading, and mildew growth. The problem of adhesion over
chalky surfaces is becoming less common as the incidence of chalky weathered surfaces
declines, owing to the decreasing use of oil-based paints and “cement paints.”
Table 32.1 gives a formulation recommended by a latex manufacturer for a high quality
exterior white house paint [6]. This formulation is not given to recommend it or these raw
materials over others; rather, it provides a framework for discussing the reasons for the
TABLE 32.1. Exterior White House Paint
Material
Water
Triton CF-10
Colloid 643
Tamol 731A
Potassium tripolyphosphate
Ti-Pure R-706
Minex 4
Attagel 50
Natrosol 250 MHR (2.5%)
Kathon LX (1.5%)
Rozone 2000
Disperse at high speed for 13 to 20 minutes,
then let down with Rhoplex Multilobe 200
Ropaque Ultra
Colloid 643
Texanol
Triton CF-10
Propylene glycol
Acrysol RM-2020 NPR
Natrosol 250 MHR (2.5%)
Water
Weight (lb)
Volume (gal)
65.42
1.00
2.00
14.18
1.00
225.00
150.00
5.00
125.00
1.70
6.00
10.25
0.11
0.28
1.50
0.05
6.70
6.85
0.25
15.00
0.20
0.66
320.00
45.00
2.00
11.10
0.50
16.35
11.00
14.40
85.55
36.00
5.25
0.28
1.40
0.07
1.90
1.25
1.25
10.25
1122.20
100.00
Properties
PVC (%) (TiO2 19.18; Minex, 19.60; Attagel, 0.72; Ropaque, 7.85), 47.4
Volume solids (%), 34.8
Stormer viscosity, equilibrated (KU), 96
ICI viscosity, equilibrated (m Pa.s), 16
VOC (g L21, excluding water), 100 g L21
PVC
19.18
19.60
0.72
7.85
640
ARCHITECTURAL COATINGS
many components of a latex paint and for the indicated order of addition during making of
the paint. Note that the formulation is given in terms of both volume and weight of the
ingredients. In general, the weight figures are used for production purposes, although
fluid metering devices are sometimes utilized. The volume figure indicates the volume
contribution to the wet paint.
Triton CF-10 is a low foam nonionic surfactant to aid in wetting the pigment. (See
Section 9.1.1 for a discussion of surfactants.) Colloids 643 is an antifoam agent (Section
24.8). The minimum level of antifoam necessary for controlling foam should be used;
excess antifoam can result in crawling when paint is applied and can interfere with color
acceptance (Section 32.2). Selection of antifoams is empirical; the manufacturers of proprietary antifoams offer test kits with samples of various products. The formulator tries
the samples to find which is most effective with the particular formulation. Tamol 731A
is a sodium salt of a carboxylated polyelectrolyte used as a pigment dispersant. Potassium
tripolyphosphate (KTPP) is also a pigment dispersant.
Ti-Pure R-706 is a rutile TiO2 pigment, surface treated for ease of dispersion in waterborne coatings. The PVC of TiO2 in the formulation is 19.18%. The Ropaque Ultra added
later in the formulation provides additional light scattering. In this formulation, dry
powder TiO2 is called for. In many cases, large manufacturers of latex paints use TiO2
in liquid form (slurries), that is, dispersions of TiO2 in water supplied by the pigment
manufacturer for more economical material handling (Section 20.1.1).
Minex 4 (nepheline, an alkaline aluminum silicate) is an inert, pigment having lower
cost than an equal PVC of TiO2. Its refractive index is sufficiently similar to that of
the binder in the paint that it contributes little directly to hiding. The fine particle size
allows it to serve as a spacer pigment to increase the hiding efficiency of the TiO2.
Use of spacing pigments is discussed further in Section 32.2. Attagel 50 is an attapulgite
clay to provide thixotropy to control settling and syneresis.
A variety of types of inert pigments can be used. Particle size, cost per unit volume, and
color are major selection criteria. Calcium carbonate is inexpensive and is sometimes used
as an inert pigment in exterior latex paints, but there is a potential problem. Latex paint
films have quite high water permeability; water and carbon dioxide, which are in equilibrium with carbonic acid, can permeate the film. Calcium carbonate dissolves in carbonic
acid to give a calcium bicarbonate solution that can diffuse out of the film. On the surface,
the water evaporates, leaving a deposit of calcium bicarbonate in areas protected from
rainfall, such as porch ceilings and under eaves, which then reverts to insoluble calcium
carbonate. The white deposit, called frosting, on the surface of the film is undesirable,
especially with colored paints. Thus, calcium carbonate should not be employed in
latex paints that may be tinted to a dark color. Furthermore, calcium carbonate– pigmented
paints are subject to degradation by acid rain; not only can frosting result, but also
film properties deteriorate and the films become more subject to mildew, due to porosity
developed in the films by loss of calcium carbonate [7].
Natrosol 250 MHR is a hydroxyethylcellulose thickener. Kathon LX (chloromethyl/
methylisothiazolone) is a biocide that serves as a preservative. Bactericides are needed
to control bacterial growth in the can of paint. There are three adverse effects of bacterial
growth: A putrid odor can develop; the metabolic processes release gases that can build up
enough pressure to blow off the can lid; and since bacteria produce enzymes that can
degrade cellulose derivatives such as HEC, the viscosity of the paint may drop.
Many other compounds are used as fungicides and bactericides [8]. Consumption of
biocides for use in paints in 2002 is reported to be $60 million [9]. In many exterior
32.1. EXTERIOR HOUSE PAINT
641
paints, zinc oxide is used as one of the white pigments. It contributes somewhat to hiding,
but its scattering efficiency is poor because the refractive index difference between it and
the binder is relatively small. Zinc oxide is used primarily as a fungicide. It is also helpful
in latex primers for tannin-stain blocking and for rust inhibition. Zinc oxide can cause
large viscosity increases during storage of some latex paints. The effect of ZnO is
system dependent; the variables and possible formulation approaches to minimize the
instability problems are discussed in Ref. [10].
Testing of biocides is difficult because fungal and bacterial growth are dependent on
ambient conditions and because many fungicides and bactericides are effective against
only a limited number of organisms. ASTM D-2574-96 is a standard test method for resistance of latex paints in the container to attack by microorganisms. Diluted paints are more
likely than undiluted paints to support bacterial growth; a test method based on inoculating
diluted latex paint with several bacteria and fungi species has been reported to be reliable
[11]. A rapid impedimetric test procedure has been reported [12]. A preliminary high
throughput study of factors affecting the effectiveness of bactericides in latex paints has
shown that the presence of phosphate groups accelerates microbial growth [13].
The best control of bacterial growth is to avoid contamination of the paint with bacteria.
Housekeeping in a latex paint factory should be as good as it is in a food-processing facility. Furthermore, bactericides kill bacteria, but they do not deactivate enzymes that have
been produced by bacteria. If the factory has places where bacteria can grow, enzymes can
get into paint. The enzymes catalyze splitting of the HEC molecules and the viscosity
drops, even though there is adequate protection against bacteria. Some thickeners are
not subject to attack by bacteria.
Rozone 2000 (1,2-dichloro-2-n-octyl-4-isothazoline-3-one) is a mildewcide to minimize mildew growth on the applied paint film. Mildew is perhaps the most common
cause of deterioration of the appearance of latex house paints. To be effective, an
organic mildewcide must migrate to the surface of the film, where it is subject to leaching
by water. When painting over mildewed paint, the old mildew should be killed with a mildewcide or with bleach before applying the new paint. If the biocides are adsorbed on the
surfaces of zeolite, they are released more slowly from the film and thus the paint films
maintain their antibacterial activity for longer times [14]. The search for more effective
mildewcides is an active research area.
Latex cannot be added until after pigment dispersion is complete because some latexes
would coagulate with the high shear exerted in the dispersion step. Therefore, after the
pigment is dispersed, the high-speed impeller is slowed to avoid coagulation and excessive
foaming.
Multilobe 200 is an acrylic latex in which groups of a few particles are polymerized
together into non-spherical lobed particles [15]. At equal concentration, the viscosity of
this latex is higher than the viscosity of comparable latexes with spherical particles,
since the packing factor is smaller than for a spherical particle latex. Paints formulated
with a lobed latex are shear thinning and give a higher high shear viscosity at the same
low shear viscosity than do conventional latexes. The higher viscosity at shear rates
experienced during application leads to thicker films with low levels of thickener and
hence lower cost. Adhesion to chalky surfaces is said to be improved as compared to
paints formulated with conventional latexes.
Latexes with part of the MMA replaced by styrene have lower cost, lower dirt pickup,
and reasonably satisfactory exterior durability, although in sunnier latitudes they tend
to chalk and fade more. Vinyl acetate – acrylate copolymers are even less expensive.
642
ARCHITECTURAL COATINGS
While their hydrolytic stability is inadequate for exterior use in climates with high humidities and rainfall, they are used outside in arid climates. Copolymers of vinyl versatate (a
C10 branched carboxylic acid vinyl ester) (Section 9.3) have better exterior durability than
that of vinyl acetate copolymers and are widely used in Europe as vehicles for exterior
paint [16,17]. A VAc/BA/vinyl ester of a C12 branched acid copolymer latex in exterior
paints is reported to have performance equal to that obtained with all acrylic latexes at a
substantially lower cost [18].
Ropaque Ultra is a high Tg latex that contains nano-bubbles of water within the particles
(Section 20.1.2) [19]. When the paint film dries, the water diffuses out of the particles,
leaving nano-voids that scatter light, reducing the TiO2 requirement. In calculating
PVC, one includes the volume of these particles, since they do not coalesce with the
latex binder particles during film formation.
Texanol is a coalescing agent; it is a mixture containing primarily the monoisobutyrate
ester of 2,2,4-trimethylpentane-1,3-diol (TMB). It reduces Tg during film formation and
then slowly evaporates. The high degree of steric hindrance of the ester group of
Texanol provides hydrolytic stability.
Propylene glycol is in the formulation for two main reasons. First, it acts as antifreeze to
stabilize the paint against coagulation during freezing and thawing. The expansion of
water as it freezes exerts substantial pressure on the latex particles and can push them
together with enough force to overcome the repulsion by the stabilizing layer, resulting
in coagulation. The glycol lowers the freezing point; even if the temperature gets low
enough to freeze the mixture, it freezes to slush, so that less pressure is exerted on the
latex particles. The second function of propylene glycol is to control the rate of drying
of the paint to permit wet lapping without disruption of the edge of the film. As water
evaporates from a latex paint film, viscosity increases rapidly as a result of the increase
in volume fraction of the internal phase. When paint is applied with a brush or roller,
the edge of the wet paint film is painted over (lapped) so that no substrate is left unpainted
or with only a thin layer of paint. With oil paint, there is no problem with wet lapping.
With a latex paint, by the time lapping occurs, the viscosity of the applied paint may
have increased enough so that the film is semisolid, but since only limited coalescence
has occurred, the film is weak. The pressure of a brush or a roller can break up the
film, resulting in irregular chunks of paint film along the lapped edge. This can be minimized by slowing the rate of evaporation of the continuous phase by incorporating slowly
evaporating propylene glycol. In addition, propylene glycol aids coalescence, although it
is less effective than other coalescents.
Acrysol RM-2020 is a nonionic associative thickener. It helps control viscosity of the
paint as a function of shear rate, improving application characteristics. Associative thickeners permit formulation of latex paints that exhibit less shear thinning, so that viscosity at
high shear rates can be higher; as a result, thicker wet films can be applied. Formulation
with associative thickeners also reduces viscosity at low shear rates, so leveling is also
improved without causing undue sagging or settling. (The thicker wet film in itself
helps promote leveling, too, since the rate of leveling depends on wet film thickness, as
discussed in Section 24.2.)
The viscosity of the external phase controls its rate of penetration into a porous substrate such as wood. If penetration is rapid, the viscosity of the paint above the porous
surface increases rapidly, resulting in poorer leveling. The final water addition permits
adjustment of the viscosity and solids content of the final paint. There is variation from
batch to batch of the viscosity of any paint formulation.
32.2. FLAT AND SEMIGLOSS PAINTS
643
The PVC of the paint is 47.4%; it is a relatively low gloss paint, as are most exterior latex
paints. The gloss of this latex exterior house paints is quite low. The popularity of exterior
satin finish paints has increased in recent years. Gloss is controlled by the PVC/CPVC ratio
(Chapter 22). Exterior durability, as well as many other properties, is also affected by PVC.
It is desirable to maximize CPVC, since the PVC of the paint at the same PVC/CPVC ratio
increases with the CPVC. With a given latex, a major factor controlling the CPVC of latex
paints is the particle size distribution of the pigments. Multiple inert pigments maximize the
breadth of particle size distribution, which increases CPVC and permits higher pigment
loading at the same PVC/CPVC, decreasing cost (Sections 22.1.1 and 22.1.3).
Commonly, the gloss of latex paints is lower than that of oil paints. The volume solids of
the paint is 34.8%, substantially lower than that of exterior oil- or alkyd-based paints. As a
result, coverage per unit volume of latex paints is lower than that of oil-based paints, but the
cost per unit volume of latex paints is lower. The grams of VOC per liter of paint, excluding
water, is 100, lower than that of alkyd- and oil-based house paints and well under current
California air pollution regulations. However, these regulations are expected to tighten.
Note that despite the widely recognized shortcomings of the Stormer viscometer discussed in Section 3.3.6, in this case (and in most paint labs in the United States), it is
the viscometer used. The viscosity of the unsheared paint is measured before shearing
reduces the viscosity related to the thixotropy of the paint. The ICI viscosity is determined
on a special cone and plate viscometer that measures viscosity at a high shear rate similar
to that exerted by application of the paint by brush or roller. It is important to have the
proper high shear viscosity because this is a major factor controlling film thickness.
Emulsions of methylsiloxane resins, an amino-functional polysiloxane emulsion, and a
styrene – acrylic latex are used to formulate exterior house paints. Such coatings have been
used on a large scale in Europe for many years but are just being introduced into the United
States. They have outstanding exterior durability [20].
There is a considerable market for coatings that increase the exterior durability of wood
but are not pigmented, or are lightly pigmented, as transparent stains. Results of a variety
of treatments of coatings on fir and oak have been reported. The best results were obtained
with waterborne transparent stain with an alkyd– acrylic dispersion binder pigmented with
iron oxide. The performance was better than that of another system without the iron oxide
in the stain and with a top coat containing a UV absorber (UVA) and hindered amine light
stabilizer (HALS) [21].
A study of the performance of exterior latex paints on masonry has been published [6].
Key properties are water resistance, moisture permeability, exterior durability, dirt pickup
resistance, color retention, low chalking, and good elasticity. A series of all acrylic and
styrene – acrylic latexes were evaluated with a range of pigments. While styrene containing latexes in clear films had poorer exterior durability and color retention than all-acrylic
latexes, this difference became minimal in pigmented paints. The overall best performance
was shown using BA/MMA/EHA/MMA, and EHA/styrene latexes. EHA-containing
latexes were noted for soiling resistance. EA/MMA showed poor performance. The
inert pigments also affected performance; for example, SiO2 increased chalking.
32.2. FLAT AND SEMIGLOSS PAINTS
The largest volume of architectural paint is interior flat wall paint; almost all is latex paint.
Although flat wall paints are most common, low semigloss (“satin” or “eggshell”) wall
644
ARCHITECTURAL COATINGS
paints are becoming more popular for their “soft” appearance. As of 1999, the EPA
permitted level of VOC for interior flat paints was 250 g L21 [3]. However, lower
VOCs will undoubtedly soon be required.
The major advantages of latex paints over the older oil-based paints are:
1. Fast drying and less sagging. If desired, two coats can be applied to the walls of a
room during a day, the furniture moved back, and the room used that night. The rapid
increase in low-shear viscosity in the early stages of film formation reduces vulnerability
to sagging, even when relatively thick films are applied.
2. Low odor. The odor of mineral spirits and by-products from oxidation of drying oils
in solventborne paints is unpleasant for days after walls are painted. Although the odors
are less offensive, some odor from the slow evaporation of coalescing solvents from
latex paints persists for up to a week. A reason for developing low VOC paints is to
eliminate this odor problem.
3. Ease of cleanup. Spills, dripped spots, brushes, and rollers are easily cleaned with
soapy water in the case of latex paints; solvent is required with oil paints. However, cleaning up must be done promptly because once latex has coalesced, cleaning it up is more
difficult than with oil paints.
4. Low VOC emissions. Latex paints were widely adopted before 1967, when the first
VOC regulations were adopted in the United States. Progress is being made in reducing the
already low VOC, as discussed in Section 32.3. There is also the advantage with waterborne paints of substantially reduced fire hazard. Not only is solventborne paint flammable, but also rags wet with oil-based paints in a confined space can undergo
spontaneous combustion, that is, ignite as a result of heat generated by autoxidation.
5. Less yellowing and embrittlement. White and light color acrylic and vinyl – acrylic
latex paints retain their color better than do alkyd paints, which gradually turn yellow with
age. Alkyds made with oils that contain very little linolenic acid, such as sunflower or safflower oils (Section 15.1.1), yellow more slowly than alkyds based on soybean oil or
tall oil fatty acids, but they still yellow with time. Alkyd films also eventually become
brittle, as autoxidative cross-linking continues far beyond the point of optimum film
properties.
Wall and trim paints are generally stocked as white paint, and companion tintable
paints are called tint bases. Liquid (or sometimes powdered) colorants are added to the
tint base to make a color chosen by the customer from a large array of color cards. This
requires that equal white tinting strength be maintained with the tint bases through
quality control, or the colors obtained will differ. Furthermore, any new formulation
must have the same white tinting strength as that of the formulation being replaced, or
else the color cards and formulations in dealer stores will have to be replaced. In each
quality line, two, or sometimes three or four, tinting bases are included. Often,
the white paint itself in a can accept up to a few fluid ounces of colorant per gallon, to
create very light or pastel colors. Bases to make deeper colors have lower levels
of TiO2, depending on how much colorant is to be added. Very deep bases may have
no TiO2 at all. Intermediate depths of shade could be matched using a lighter base, but
the cost would be excessive because more tinting color would be needed to match
the colors; hiding would be greater than needed. Tint bases are made with surfactants
that help ensure compatibility between colorant and tint base. The tinting colorants are
commonly “universal colorants” that can be used to tint either waterborne or solventborne
paints.
32.2. FLAT AND SEMIGLOSS PAINTS
645
Users are sometimes confused by the change in color of latex paint as it dries; the color of
a dry film is darker than the color of the wet paint. In the wet paint, the interfaces between the
water (n ¼ 1.33), polymer particles (n ¼ approximately 1.5), TiO2 (n ¼ 2.73), and inert
pigments (n ¼ approximately 1.6) scatter light to a greater degree than when the paint is
dry. The dry paint has fewer interfaces as a result of coalescence of the latex particles,
and there are smaller refractive index differences since the pigment particles are in a
polymer matrix instead of water. Since light scattering decreases as water evaporates
and the latex particles coalesce, the color gets darker; hiding also decreases as the paint
dries.
Since exterior durability is not needed, lower cost vinyl acetate copolymer latexes
(Section 9.3) are the principal binders for interior products. The Tg of VAc homopolymer
is about 328C, so that a softer comonomer such as butyl acrylate (BA) is needed to reduce
Tg. Ethylene –VAc latexes have been introduced to replace VAc/BA latexes. Ethylene is
not only a more effective co-monomer than BA for reducing the Tg of VAc polymers but
appears to provide more effective plasticization as well. Paints made with VAc/ethylene
latexes require less coalescing solvent, show better touch-up properties at low and room
temperatures, and have better scrub resistance [22].
Table 32.2 is a supplier suggested formulation for an ultra low VOC interior eggshell
paint [23]. The VOC for the paint is 6 g L21.
Natrosol Plus 330 MBR is a modified hydroxyethylcellulose associative thickener. Tamol
1124 is a dispersant that is the ammonium salt of a hydrophilic acid copolymer. AMP-95 is
2-amino-2-methyl-1-propanol. Strodex PK-90 is a dispersant that is said to promote
color acceptance. Triton CF-10 is a low foam nonionic surfactant. Drewplus L-745 is an antifoam. Nuosept 95 is a biocide, a mixture of 5-substituted 1-aza-3,7-dioxabicyclo[3.3.0]
octanes. Ti-Pure R-706 is a surface-treated rutile TiO2 designed for use in waterborne coatings. Mattex is a surface-treated kaolin clay pigment. Attagel 50 is an attapulgite clay to
TABLE 32.2. Ultra Low VOC Interior Eggshell Paint
Component
Water
Natrosol Plus 220 MBR (3%)
Tamol 1124
AMP-95
Strodex PK-90
Triton CF-10
Drewplus L-145
Nuosept 95
Disperse the addition under agitation
Ti-Pure R-706
Mattex
Attagel 50
Let down with
Airflex EF811 (58.0%)
Water
Drewplus L-145
Acrysol RM-2020 NPR
Acrysol RM-8W
Benzoflex 9-88
Total
Weight
Volume
150.0
50.0
6.0
2.0
2.0
2.0
1.5
1.5
18.0
6.0
0.6
0.2
0.2
0.2
0.2
0.2
230.0
110.0
5/0
6.9
5.0
0/3
418.9
83.0
3.0
20.0
8.0
6.0
1098
47.0
10.0
0.4
2.3
0.9
0.6
99.1
646
ARCHITECTURAL COATINGS
control settling and syneresis. Airflex EF 811 is a VA/ethylene latex. Acrysol RM 2020 NPR
and Acrysol RM-8W are hydrophobically modified ethoxylated polyurethanes (HEUR)
associative thickeners. Benzoflex 9-88 (dipropylene glycol dibenzoate) is a coalescent.
For interior applications for which greater water resistance is needed, acrylic latexes or
vinyl versatate copolymer latexes are used. A copolymer latex of BA, styrene, and 0.5%
acrylamide shows superior scrub resistance [24]. Copolymer latexes of VAc and vinyl
versatate are used in both interior and exterior paints. These copolymers have higher
hydrolytic stability than conventional vinyl acetate copolymers. A combination of ethoxylated undecyl alcohol, a cellulose ether, and sodium vinyl sulfonate can be used as a surfactant/protective colloid in preparing the latexes. The effect of changes in chain length of
the ethylene oxide units on the properties of latexes and paints has been reported. Increasing degree of ethoxylation gave smaller particle size latex particles and higher viscosity. In
flat paints, optimum properties are obtained with 2 to 3% emulsifier concentration and 17
to 28 degree of ethoxylation. In gloss and semigloss paints, gloss improves with an
increase of emulsifier content up to about 4%, where it levels off. The gloss improves
as the degree of ethoxylation increases to 17. Above 4% concentration and 17 ethoxyl
units, blocking increases dramatically. Similar results were obtained with VAc/ethylene
latexes [25]. VAc/vinyl ester of a C12 branched acid latexes in interior flat wall
paints have better scrub and stain resistance than VAc/BA latexes and styrene/acrylic
latexes [26].
Acrylated castor oil has been used as a co-monomer, resulting in latexes that are formulated into semigloss paints with better block resistance while still permitting low temperature film formation. The latexes have the further advantage that coalescing solvents are
not needed [27].
When painting ceilings, one is particularly anxious to get hiding in one coat, since
painting over one’s head and moving the ladder requires more effort than painting
walls. The problem is particularly challenging because ceiling paints are commonly
plain white, to reflect light diffusely. Since there are no color pigments in a white paint
to absorb light, the hiding by white paints is poorer than any color paint made from it.
The problem is compounded by the decrease in hiding when a latex paint film dries.
The user thinks he or she has applied enough paint to hide marks on the ceiling, but
comes back an hour or so later and finds that marks show through the dry paint. Special
ceiling paints minimize this problem by formulating with PVC above CPVC. Dry
paint films with PVC above CPVC have voids of air with n ¼ 1 that add additional
light scattering by the new interfaces between air and polymer as well as air and
pigment. Formulations can be adjusted so that wet hiding and dry hiding are approximately equal. The films do not have as high mechanical strength as films of paint with
PVC , CPVC, and resistance to staining is poorer, but neither property is important
for ceiling paints.
Owing to the high cost of TiO2, coatings are not generally formulated with a PVC of
TiO2 more than 18% since incremental hiding at a higher PVC is not cost efficient.
(This value is dependent on the actual TiO2 content of the pigment and the stability of
the dispersion.) Rather than using large quantities of TiO2, lower cost inert pigments
are used to occupy additional volume. Although there is some controversy, most
workers accept that inert pigments with particle sizes smaller than that of the TiO2,
called spacer inerts, increase the efficiency of the TiO2 [28]. A mathematical model has
been developed that can be used to improve the spacing of TiO2 by predicting the
optimum inert pigment size distribution and concentration for a given formulation [29].
32.2. FLAT AND SEMIGLOSS PAINTS
647
Aluminum silicate pigments are reported to permit replacing 50% of TiO2 while maintaining good whiteness and hiding power. The effectiveness is reported to result not only from
its action as a spacer, but hiding also results from the fact that the pigment is porous, so that
it has better hiding power than most inert pigments [30].
Another approach to greater hiding at lower cost is to use a high Tg latex, such as polystyrene, as a pigment [31]. When the latex binder coalesces, the high Tg latex does not
coalesce, but the particles remain separate, as with any other pigment. Including the dry
volume of the high Tg latex particles as part of the pigment volume, paints can be formulated with a PVC greater than CPVC, resulting in air bubbles that increase hiding but
without making the surface of the film porous. The paints provide equal hiding at lower
TiO2 content while retaining good enamel hold out and stain resistance. No convincing
explanation of why an intact surface film forms in such paints has been published, but
the method is said to be used on a large commercial scale.
Yet another approach to minimizing the TiO2 requirement is to use as pigments special
high Tg latexes, such as Ropaque Ultra [19], as shown in the formula for an exterior house
paint given in Table 32.1 and explained in the text following the table.
A large fraction of latex flat wall paint is applied by roller. During roller application, latex
paints spatter, some to a major degree. Paints with high extensional viscosity are likely to
spatter severely [32] (Section 23.1.2). Extensional viscosity increases when high molecular
weight (MW) water-soluble polymers with very flexible backbones are used as thickeners in
a latex paint [33]. Spattering can be minimized by using low MW water-soluble thickeners
with rigid segments in the polymer backbone, such as low MW HEC.
Whereas in many applications, the increase in external phase viscosity resulting from
use of water-soluble polymeric thickeners is desirable, an application for which it is undesirable is on concrete block walls. Concrete block surfaces contain holes that are large
compared to the pigment and latex particles, as well as pores that are small compared
to them. When a solventborne paint is applied to concrete block, the coverage is low,
since so much paint penetrates into the smaller holes. When regular latex paint is
applied, the coverage is better. The lower viscosity of the external (continuous) phase relative to that of solventborne paint leads to more penetration of continuous phase into the
small pores of the block so that the viscosity of the remaining paint increases rapidly
and there is less penetration into the large holes. The coverage with latex paint can be
improved further by omitting water-soluble thickener from the formulation. This further
reduces the viscosity of the external phase, which can then penetrate more rapidly into
the small pores and hence give even more rapid buildup of viscosity of the remaining
paint. The effect is enhanced by the requirement that to have equal starting viscosity,
the volume fraction of internal phase in the modified paint must be significantly higher
than when water-soluble polymer is present. With increasing volume fraction internal
phase, the viscosity increases more rapidly as the continuous phase is drained off in the
small pores, so penetration into the large holes is reduced. A further advantage of latex
paints, especially acrylic, styrene/acrylic, or styrene/butadiene latexes, relative to oil or
alkyd paints is that they are not subject to saponification by the basicity of the concrete
block and the mortar joints.
In a study of changes in composition and properties of latex paints applied to porous
inorganic substrates, it was found that a significant amount of coalescing agent penetrated
into the substrate, hence increasing the Tg of the film. There were also indications that
binder penetrated. The authors recommend that it is particularly important to formulate
paints with PVC close to CPVC for application to porous substrates [34].
648
ARCHITECTURAL COATINGS
When latex paint films form, surfactants sometimes bloom out of the film and collect on
the surface. Usually, this is not noticeable; however, if water condenses on the paint
surface, surfactant on the surface can dissolve. Then, when the water evaporates, surfactant concentrates in the last remaining water droplets, leaving brown spots of surfactant
on the surface. The problem can be minimized by avoiding dark colored surfactants.
However, surfactant can leave whitish spots on dark color paints. It has been shown
that nonionic surfactants are more compatible with latex polymers and less likely
to bloom than anionic surfactants [35]. Use of alkenyl-functional surfactants in latex
preparation has been reported. The surfactant copolymerizes with the monomers
during polymerization, giving latexes that are reported to have superior shear stability
and freeze – thaw resistance. Paints prepared with the surfactant-bound latexes overcome
the problems associated with the use of conventional surfactants such as surfactant
blooming [36].
Although all latex paints have a low VOC compared to solventborne paints, there is still
regulatory pressure to reduce or eliminate the VOC. To accomplish this, coalescing solvents and propylene glycol in the formula must be reduced to low levels or eliminated.
Several approaches have been followed. Sequencing the composition of the monomer
feed during polymerization so that the last part of the monomer feed has a higher fraction
of monomers, giving a low Tg as in the power feed approach discussed in Section 9.1.3,
would permit reduction of coalescing solvent. Addition of acrylic acid in the later part
of the co-monomer feed has been suggested; water associates with a salt of the carboxylic
acid, plasticizing the surface of the particles, promoting coalescence, and reducing the
need for coalescing solvent. Another method is to use cross-linkable latexes with a
lower Tg; cross-linking after film formation increases block and dirt resistance to offset
the effect of the lower Tg (Section 9.4).
It is estimated that 120 million pounds of volatile coalescing agents are emitted in one
year in the United States. A cross-linkable coalescing agent, the propylene glycol monoester of sunflower fatty acids, has been developed that essentially eliminates evaporation
of the coalescent, thus reducing VOC. It is also reported to have less odor, increased scrub
resistance, and produce higher gloss [37]. ADM received a 2005 presidential green award
for the development.
An approach to eliminating the need for coalescing solvent is to use blends of high and
low Tg latexes [38,39]. When a film of such a blend dries, the high Tg latex does not
coalesce, but is dispersed in the continuous phase from the low Tg polymer. The hard particles act to reinforce the low Tg polymer film, increasing the storage modulus (E0 ) of the
film so that its block resistance is superior to that of a film from only the low Tg latex. The
properties depend on the ratio of hard and soft latexes and on the particle sizes of
the latexes. For example, excellent block resistance was obtained with a combination of
60% of a 37/61/2 styrene/BA/MAA latex with a particle size of 475 nm and Tg ¼ 98C
and 40% of a 70/28/2 styrene/BA/MAA latex with a particle size of 118 nm and
Tg ¼ 628C. No coalescing agent was required for film formation at ambient temperature.
The superior block resistance is said to result from the fact that close packing leads to a
high concentration of the small-particle size latex at the surface of the film [39]
(Section 32.3). Another approach to eliminating VOC is to make nano polymer/clay composite latexes. The latexes are made by polymerizing relatively low Tg acrylic monomers
with clay dispersed in the water in which the emulsion polymerization is carried out. Films
prepared from the latex have unusual toughness without compromising film formation.
The latexes are used in zero VOC paints with low tack and dirt pickup [40] (Section 9.3).
32.3. GLOSS ENAMELS
649
Semigloss paints are made with lower pigment contents than flat paints. They
are frequently used to paint doors and other woodwork. They are usually prepared using
acrylic latexes to give better wash resistance, compared to other latex types.
Combined aqueous polyurethane dispersions (PUDs) and aqueous vinyl resins offer
several potential advantages over each of the two separate types of polymers. (See
Section 12.7.2 for a discussion of acrylated PUDs.) In general terms, polyurethane
based polymers offer superior abrasion resistance, due to intermolecular hydrogen
bonding. Vinyl resins have lower raw material costs and lower processing costs. PUDs
have low minimum film formation (MFFT) temperatures in relation to dry film Tg due
to the plasticizing effect of hydrogen bonding with water, reducing the Tg of the
polymer particles. Used with high Tg acrylics, the MFFT is reduced without reducing hardness. For example, the open time of latex paints can be extended by adding a PUD
with a Tg lower than that of the latex in the paint. A paint made with a latex composed
of n-butyl acrylate (BA), 2-ethylhexyl acrylate (2-EHA), methyl methacrylate (MMA),
and methacrylic acid (MAA), and modified with a PUD with a Tg of 2408C has a
longer open time, permitting re-brushing to ensure good hiding and lapping [41]. Coalescing solvents are required to permit film formation with latexes at low temperatures,
whereas they are not required with PUDs.
Further improvement in properties can be obtained by polymerizing the acrylic monomers in a PUD or in a prepolymer precursor for a PUD. Semigloss wood coatings are formulated with a carboxy-functional PUD to which acrylic monometers, including
N-(2-methacryloxyethyl)ethylene urea, are added and polymerized [42]. Latexes contain
surfactants that can bloom to the surface of films, reducing gloss, whereas many PUDs
can be made without surfactants.
32.3. GLOSS ENAMELS
The term enamel connotes a hard, glossy surface analogous to porcelain enamel. In the
United States, gloss enamels are used both indoors and outdoors for trim around
windows and doors, for shutters, for wood furniture, and on kitchen and bathroom
walls. In some other countries, they are also used for walls. For gloss enamels, the
merits and drawbacks of alkyd paints and latex paints are about evenly balanced, and
latexes have partly, but not completely, displaced alkyds. Mounting pressure to reduce
VOC may eventually tilt the balance toward latexes unless there is a technical breakthrough with alkyds. The 1999 EPA VOC emission standard for gloss paints is
380 g L21 and 450 g L21 for quick-dry enamels [3]. These levels are lower in California
and certain eastern states.
32.3.1. Alkyd Gloss Enamels
As described in Chapter 15, alkyd enamels cross-link by autoxidation to form tough films
that adhere well to a variety of surfaces and are block resistant and water resistant. An
advantage of gloss alkyd enamels is that they have higher gloss than latex enamels. As
discussed in Section 19.10.1, a polymer layer with very low pigment content forms at
the upper surface of the coating during film formation from solution vehicle paints; this
does not happen with latex paints.
650
ARCHITECTURAL COATINGS
The principal advantage of gloss alkyd paints is that they afford good hiding with one
coat when used over a surface without sharp color contrasts. One-coat hiding is especially
important to painting contractors, for whom the cost of application is higher than the cost
of paint. Several factors account for the difference in hiding between alkyd and latex gloss
paints. The NVV of a typical alkyd gloss paint can be 66% or even higher, whereas the
NVV of a latex gloss paint is limited to about 33%. To obtain the same dry film thickness,
one has to apply twice as thick a wet film of the latex paint.
Another factor that affects hiding is leveling. Assume that a uniform dry film thickness
of, say, 50 mm of a paint provides just satisfactory hiding. If the paint levels poorly, there
will be streaks of thinner film, say, 35 mm, and thicker film, say, 65 mm, and the hiding of
the uneven film will be poor. Hiding is likely to be inferior to that of a uniform 35 mm dry
film of the same paint. Since the 35 and 65 mm areas are immediately next to each other,
the contrast emphasizes the poor hiding. Furthermore, not only is the hiding affected
adversely by poor leveling, but also the contrast in colors resulting from the poor
hiding in the valleys compared to the ridges emphasizes the poor leveling. Alkyd paints
for brush application generally level better than latex paints because the solvent is slow
evaporating mineral spirits. The difference in leveling is amplified when the paint is
applied under warm, dry conditions. Yet another factor is that the volume fraction of
internal phase in latex paints is higher than in alkyd paints since both the latex and
pigment particles are dispersed phases. Therefore, even if the volatile material evaporates
at equal rates, the viscosity of a latex paint increases faster.
As described in Section 24.2, Overdiep demonstrated that leveling of solventborne,
brush-applied paints is promoted by surface tension differential –driven flow in the wet
paint film [43]. When solvent evaporation begins, the fraction of solvent lost from
the valleys of the brush marks is greater than that lost from the thicker ridges. Since the
surface tension of the solvent is lower than that of the alkyd, the surface tension of the
more concentrated alkyd solution in the valleys becomes higher than that in the ridges.
The resulting differential in surface tension causes the paint to flow from the ridges into
the valleys to minimize overall surface tension, thus promoting leveling. This driving
force may be absent in the leveling of latex enamels (Section 32.3.2).
When selecting paints for repainting, it should be remembered that latex paints can be
used over alkyds or latexes as long as the surface is properly prepared, but use of alkyds
over latexes is risky. The solvent in alkyds may penetrate uncross-linked latex films and
cause lifting.
The drawbacks of alkyd gloss enamels are largely the same as those of alkyd flat wall
paints (Section 32.2). They include slow drying, odor, yellowing and embrittlement with
age, and the need to use solvent for cleanup. In the future, perhaps the most important consideration will be VOC content. There is disagreement as to what level of VOC can be
achieved with alkyd paints while retaining reasonable application and film properties.
Many in the coatings field doubt that VOC can be reduced much below 250 g L21
without a technical breakthrough. The problems of low-VOC alkyds include application
characteristics, through dry, color change, and durability.
High-solids alkyds are discussed in Section 15.2. Solids can be increased to an extent
by solvent selection, especially the use of hydrogen-bond acceptor solvents to reduce
intermolecular hydrogen bonding. Although solids can be increased by reducing MW
and by using narrower MW distributions, both of these approaches lead to inferior film
properties and durability if carried very far.
32.3. GLOSS ENAMELS
651
A promising approach to increasing solids is the use of reactive diluents (Section 15.2).
Such additives are low MW materials designed to reduce viscosity almost as efficiently as
solvent, but also to co-react with the oxidizing alkyd during film formation. This permits
reduction in VOC while maintaining film properties. An example is dicyclopentenyloxyethyl methacrylate. This reactive diluent has both an acrylate double bond and an activated
allylic position. In the presence of driers, it can co-react with an alkyd such as a long oil
linseed alkyd. Using this reactive diluent with specially designed alkyds is reported to
permit formulation of gloss alkyd paints with a VOC of 155 g L21. The properties of
the films are said to approach those of an alkyd paint with a VOC of 350 g L21 [44].
Other types of reactive diluents are condensation products of mixtures of drying oil
acid amides and acrylamide with (hexaalkoxymethyl)melamine [45] and oligomeric
polyallyl ethers [46]. Although the use of reactive diluents is attractive, their cost and performance characteristics require further improvement. Use of polyallyl ethers has been
questioned because of the potential for formation of toxic, volatile acrolein during
cross-linking.
When paints containing oxidizing alkyds and driers are exposed to air, premature
cross-linking will occur at the wet paint surface. This skinning is controlled by additives,
as reviewed in Ref. [47]. The most effective was found to be methyl ethyl ketone oxime
(MEKO). MEKO forms a complex with cobalt ions, [Co(MEKO)1 – 8]3þ. The complex is
in equilibrium with the components, so that while it is present it reduces the effectiveness
of the cobalt. After the paint is applied, MEKO evaporates relatively rapidly, freeing the
cobalt drier. The MEKO is added to the coating just before filling into containers. In some
cases a further small amount of a solution of MEKO is added after the containers are filled
and just before they are sealed, to prevent skinning during storage. Addition of 0.2% of
MEKO prevents skinning for more than 250 days. It increases the dust-free time of an
alkyd coating to 2 hours from 1 hour 45 minutes but the fully dried time is reduced
from 4 hours 45 minutes to 4 hour. The initial effect results from the time for the
MEKO to evaporate. The shorter drying time is thought to result from faster permeation
of oxygen in the earlier stages after application while MEKO is inhibiting surface cure.
Two approaches to waterborne alkyds for trade sales paint have been investigated
extensively. Efforts to develop gloss enamels based on water-reducible alkyd resins
(Section 15.3) have not been very successful. A key obstacle has been the difficulty of
making alkyds with sufficient hydrolytic stability for the required two-year shelf life.
Water-reducible alkyds generally contain significant levels of solvent and may not
reduce VOC much below 250 g L21. Emulsions of alkyds in water can also be used to
prepare gloss enamels [48] (Section 26.3). They are used in Europe but have not been
widely adopted in the United States. The VOC content can be lower, but the surfactants
required for emulsification cause foaming on application and adversely affect film properties;
and the shortcomings of odor and poor resistance to yellowing and embrittlement remain.
32.3.2. Latex Gloss Enamels
Development of latex gloss enamels has proven to be one of the most challenging problems of coatings technology. Key problems and potential solutions are discussed in
this section. One problem is that of obtaining high gloss. As noted in Section 32.3.1,
the clear polymer layer associated with the surface of alkyd enamels does not readily
form as latex coalesces. The ratio of pigment to binder at the surface of a latex paint
652
ARCHITECTURAL COATINGS
film can be reduced somewhat by using a finer particle size latex, but there is still a difference from alkyd paint films. Gloss is affected by flocculation of either pigment or latex.
The choice of dispersing surfactants and thickeners, and their order of addition, can
affect gloss. In one study it was shown that when a mixture of polyacrylic acid (PAA)
and HEC was used as a thickener, gloss was higher if the HEC was added first than
when PAA was added first [49]. Nonionic surfactants resulted in higher gloss than
anionic surfactants. Other factors tending to reduce the gloss of latex paint films are the
haze that results from incompatibility of surfactants (and possibly other components)
and the blooming that results when surfactants migrate to the surface of the film. These
problems can be minimized by making latexes with the lowest possible surfactant
content and by selecting surfactants for pigment dispersion that are as compatible as
possible.
Considerable work has been done using blends of compatible water-soluble resins with
latexes to achieve high gloss. In floor wax applications, morpholine salts of styrene/
acrylic acid copolymers have been used. When the film forms, the solution resin concentrates in the surface layer. When the film dries, the morpholine evaporates, leaving the free
carboxylic acid groups. The film has sufficient water resistance for normal use, but it can
be removed by mopping with ammonia water. This level of resistance is an advantage
for floor wax but is not adequate for paint films. There have been some proprietary
resins marketed that give better gloss with less loss of resistance properties.
In exterior and to a lesser degree, interior applications, the advantage of high initial
gloss alkyd enamels is more than offset by the better gloss retention and resistance to
cracking exhibited by latex gloss paints. Depending on location, alkyd enamels exposed
outdoors lose so much gloss in a year or two that the coating may become nearly flat.
Although a latex enamel starts out with a lower gloss, it retains most of its initial gloss
for several years. In interior applications, the low odor, superior color retention, and
greater resistance to cracking are important advantages of latex gloss paints.
The principal limitation of gloss latex paints is not their lower gloss; the major problem
is achieving adequate hiding in one coat. The factors involved are discussed in Section
32.3.1. Not much can be done about the relatively low NVV of latex paints. The formulating challenge is to make latex enamels that can be easily applied in films thick enough for
one-coat hiding. For maximum hiding, a latex enamel should be formulated to have a
higher viscosity at high shear rate than that of an alkyd enamel. In practice, as is discussed
later, the high shear rate viscosity of latex paints has traditionally been formulated to be
lower than that of alkyd paints, further compounding the problem. Wet film thickness
can be controlled to a degree by the extent to which the painter brushes out the paint.
The painter tries to judge how much he or she should brush out by how well the wet
paint is covering. In the case of alkyd paints, there is relatively little difference between
the wet and dry hiding power of the paint. However, the wet hiding power of latex
paints is greater than their dry hiding power. This increases the difficulty of judging
how far to brush out a latex paint.
The need for good leveling is also described in Section 32.3.1. The driving forces for
leveling of latex paints do not seem to have been considered in the literature. It is not
obvious how surface tension differentials could develop during film formation of latex
enamels. The surface tension of the water phase of latex paints is controlled primarily
by the surfactants in the paint, which suggests that there probably is little change in
surface tension as water evaporates. If this hypothesis is correct, the driving force for
leveling of latex paints is the relatively small force of surface tension – driven leveling
32.3. GLOSS ENAMELS
653
rather than the larger force of surface tension differential – driven leveling thought to
operate with alkyd enamels. Another factor that may be important is dynamic surface
tension. It has been shown that some surfactants reach equilibrium surface tension more
rapidly than others [50]. The need for research is evident.
However, the major factor affecting the leveling and therefore the hiding of gloss latex
paints is probably their rheological properties. As they have traditionally been formulated
with water-soluble thickener polymers such as HEC, latex paints have exhibited a much
greater degree of shear thinning than alkyd paints. This has led to latex paints having
too low a viscosity at high shear rate, so the applied film thickness tends to be too thin,
and too high a viscosity at low shear rate to permit adequate leveling. The problem is
especially severe because the rate of recovery of viscosity after exposure to high shear
rates is generally rapid with latex paints. The use of Stormer viscometers (Section
3.3.6) has been at least partly responsible for the prolonged time before the problem
was well defined. This viscometer measures something related to viscosity in a midrange
of shear rates but gives no information about viscosity in either of the critical regions—at
high and low shear rates.
The reasons for this greater dependency of viscosity on shear rate in latex coatings have
not been fully elucidated: at least two factors may be involved, flocculation of latex particles and/or pigment particles in the presence of HEC, and possible entanglement of
chains of swollen high MW HEC [51]. Progress in minimizing the problem has been
made by using associative thickeners, of which many kinds are available. (See Section
3.5.1 for discussion.) They all are moderately low MW hydrophilic polymers with two
or more long chain nonpolar hydrocarbon groups spaced along the backbone. Use of
such thickeners permits formulation of latex paints that exhibit less shear thinning, so
that viscosity at high shear rates can be higher; as a result, thicker wet films can be
applied [51,52]. Formulation with associative thickeners also reduces viscosity at low
shear rates, so leveling is also improved. (The thicker wet film in itself helps promote
leveling too, since the rate of leveling depends on wet film thickness, as discussed in
Section 24.2.)
It has been shown that good formulations with associative thickeners not only level
better but also give somewhat higher gloss [53]. Higher gloss, as well as better rheological
properties, have been reported by use of a combination of associative thickeners [54],
which may reduce flocculation of the TiO2. Another factor in the improved gloss may
be that associative thickeners are effective with small particle size latexes, which give
the highest gloss [51]. It has also been shown that paints thickened with associative thickeners spatter less when applied by roller than do those thickened with HEC, and that most
of the thickeners resist viscosity loss by bacterial action [51]. Sag control of latex paints is
more difficult with associative thickeners but still easier than with alkyd paints; leveling
and gloss are still not equal to those of most alkyd paints [53].
Another shortcoming of latex paints, particularly evident in gloss formulations, is the
time required to develop final film properties. Part of the problem is that users are, in a
sense, fooled by the drying properties of latex paints. They dry to touch more rapidly
than do alkyd paints and can be handled sooner. However, they require longer to reach
their ultimate properties. For example, even though latex paints dry more rapidly than
alkyd paints, a longer time is necessary to develop the block resistance required to
prevent windows and doors from sticking or to permit putting heavy objects on a newly
painted shelf. The initial film formation of latex particles is rapid, but full coalescence
is limited by the availability of free volume. Since the T 2 Tg must be small, free
654
ARCHITECTURAL COATINGS
volume is small. The situation is helped by using coalescing solvents. However, loss of
these solvents is controlled by diffusion rate, which is also limited by T 2 Tg. It has
been reported that latex particles having inner layers of relatively high Tg with a gradient
to low Tg on the outer shell of the particles can provide film formation at low temperatures,
yet achieve block resistance relatively quickly [55]. It has been recommended that a high
Tg latex be used with larger amounts of carefully selected coalescing agents [56]. The
dimethylether of dipropylene glycol is reported to be an effective coalescing agent with
a relatively rapid rate of evaporation from films [57]. Whereas the initial loss of coalescing
agent is controlled by volatility, losses in later stages are controlled by the rate of diffusion
through the film, and the loss rate becomes very low. There is need for further progress,
but as this problem is unlikely ever to be solved completely, there is a need to educate
consumers better on the limitations of latex paints.
Progress has been made in developing gloss latex paints with adequate hiding by one
coat, but further efforts lie ahead. The problem is made more difficult by the lack of adequate laboratory tests to measure gloss (Section 19.10.2) or absolute hiding (Section 19.3).
The lack of adequate test procedures is particularly troublesome because many of the companies that supply raw materials to the paint industry, many of the people who establish
specifications and regulations, and sadly, even people working as paint formulators are
not aware that the laboratory tests are inadequate.
Another problem of some gloss latex paints is poor adhesion to old gloss paint surfaces
when the new dry paint film gets wet. After wetting with water, some latex paint films can
be peeled off the old paint surface in sheets. Such a film is said to exhibit poor wet
adhesion. There is always a problem of achieving adequate adhesion when repainting old
gloss paint surfaces, even with alkyd paints, but the problem with latex paints is more
severe. It is essential to wash any grease off the surface and to roughen the surface by
sanding, but even with such surface preparation, many latex paints do not show good wet
adhesion. Wet adhesion improves with age, but it can remain seriously deficient for
several weeks or even months. Latexes that minimize the problem of wet adhesion have
been developed by several manufacturers. Incorporation of small amounts of hydrogenbonding polarizable co-monomers such as methacrylamidoethyleneurea (Section 9.1)
improves wet adhesion and wet scrub resistance.
Although VOC of latex paints is lower than that of alkyd paints, latex enamels are
usually formulated with higher VOCs than other latex paints. Extra coalescing solvent
is used to permit use of latexes with the relatively high Tg needed for block, scrub, and
stain resistance. Their VOC is often only slightly less than the maximum (currently,
250 g L21) allowed by the strictest regulations. There is mounting pressure to reduce
VOC emissions further. One approach is to select the most efficient coalescing solvents
to minimize the amount required. For example, the acetate of propylene glycol n-butyl
ether has been recommended as an efficient coalescent [57]. 2-Ethylhexyl benzoate has
been recommended as a replacement for Texanol as coalescent. It is reported that 25%
less 2-ethylhexayl benzoate is required, thus reducing the VOCs; the odor is lower; and
the evaporation rate is slower [58]. Another approach is to use blends of high and low
Tg latexes [38]. The binder must be transparent, which requires that the difference in
refractive index be small and that the particle size of the high Tg latex be small.
Core – shell copolymers of BA, MMA, and AA can be used to formulate gloss latex
paints having high block resistance with no coalescent. The final films were dispersions
of hard particles in a continuous softer phase. The best compromise of properties was
with 25 to 35% of hard phase (MMA) [59].
REFERENCES
655
Still another approach is to use thermosetting latexes (Section 9.4); this permits use of
low Tg latexes that form films without the need for coalescing solvents but that reach adequate hardness through cross-linking. For example, an undisclosed latex that cross-links
by autoxidation is reported to permit formulation of zero VOC gloss latex paints [60].
An acrylic latex that oxidatively cross-links, Rhoplex HG-700, is recommended for
gloss paints. It is used with an associative thickener and a surfactant that promotes leveling. The latex has a MFFT of 168C. Films of the paint gain block resistance rapidly with
full development of cross-linked properties over a period of a month. Paint films show a
208 gloss of 50 to 60 and a 608 gloss of 80 to 90. VOC is reported to be 150 g L21.
Adhesion to old alkyd paint surfaces and exterior durability are reported to be excellent
[61].
Satisfactory later gloss enamels having low VOC have been reported by combining
associative thickeners and a cross-linking later [60]. An associative thicker (hydrophobically modified polyacrylic acid ammonium salt) increases open time to permit wet lapping
and improves freeze-thaw resistance. This permits elimination of propylene glycol thus
reducing VOC [60].
Acrylate – urethane dispersions have also been reported to give, exterior latex paints
with good gloss and exterior durability [62].
GENERAL REFERENCE
Feist, W. C., Finishing Exterior Wood, Federation of Societies for Coatings Technology, Blue Bell,
PA, 1996.
REFERENCES
1. U.S. Department of Commerce, Bureau of Census, Current Industrial Reports: Paint and
Coating Manufacturing, http://www.census.gov.
2. Feist, W. C., Finishing Exterior Wood, Federation of Societies for Coatings Technology, Blue
Bell, PA, 1996.
3. http://www.epa.gov/ttn/atw/183e/aim/fr1191.brt.
4. Smith, A.; Wagner, O., J. Coat. Technol., 1996, 68(862), 37.
5. Hayoz, P.; et al., Prog. Org. Coat., 2003, 48, 297.
6. Tech. bull. 78739, Rohm & Haas, Philadelphia, PA, 2004.
7. Hook, J. W., III; et al., Prog. Org. Coat., 1994, 24, 175.
8. Wachtler, P.; Kunisch, F., Polym. Paint Colour J., 1997, October, 2.
9. Challener, C., JCT Coat. Tech, 2004, 1(5), 54.
10. Mattei, I. V.; et al., J. Coat. Technol., 1991, 63(803), 39.
11. Jaquess, P. A.; McLaurin, M. C., J. Coat. Technol., 1993, 65(823), 77.
12. Cooke, P. K.; et al., J. Coat. Technol., 1991, 63(796), 33.
13. Rhoades, A. M.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2004, Paper 22.
14. Edge, M.; et al., Prog. Org. Coat., 2001, 43, 10.
15. Chou, C.-S.; et al., J. Coat. Technol., 1987, 59(755), 93. Rhoplex Multilobe 200, tech. bull.,
Rohm & Haas, Philadelphia, PA, 1992.
656
ARCHITECTURAL COATINGS
16. Prior, R. A.; et al., Prog. Org. Coat., 1996, 29, 209.
17. Decocq, F.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1997,
p. 168.
18. Wang, H. W.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2001, pp. 61– 76.
19. Fasano, D. M., J. Coat. Technol., 1987, 59(752), 109.
20. Mangi, R.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 41–56.
21. Jirous-Rajkovic, V.; et al., Surf. Coat. Int. B Coat. Trans., 2004, 87(1), 70.
22. Gilicinski, A. G.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2000, pp. 516 – 526.
23. Publ. 151-0290.6, Air Products and Chemicals, Allentown, PA, 2002.
24. Porzio, R. S.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2001, pp. 245 – 260.
25. Heldmann, C.; et al., Prog. Org. Coat., 1999, 35, 69.
26. Yang, H. W.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2000, pp. 308 – 321.
27. Zhu, P.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 139 –144.
28. Braun, J. H., J. Coat. Technol., 1988, 60(758), 67.
29. Temperley, J.; et al., J. Coat. Technol., 1992, 64(809), 33.
30. Christian, H.-D.; Feller, A., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 2003, pp. 115 –128.
31. Ramig, A., Jr.; Floyd, F. L., J. Coat. Technol., 1979, 51(658), 63, 75.
32. Massouda, D. B., J. Coat. Technol., 1985, 57(722), 27.
33. Glass, J. E., J. Coat. Technol., 1978, 50(641), 56.
34. Perara, D. Y.; Eynde, D. V., J. Coat. Technol., 2001, 73(919), 89.
35. Evanson, K. W.; Urban, M. W., J. Appl. Polym. Sci., 1991, 42, 2309.
36. Grade, J.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 145 –157.
37. http://admworld.com 2005.
38. Winnik, M. A.; Feng, J., J. Coat. Technol., 1996, 68(852), 39.
39. Eckersley, S. A.; Helmer, B. J., J. Coat. Technol., 1997, 69(864), 97.
40. Lorah, D. P.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2004, Paper 19.
41. Gray, R. T.; Lee, J., U.S. patent 6,303,189 (2001).
42. Chung, J. K.; et al., U.S. patent 6,031,041 (2000).
43. Overdiep, W. S., Prog. Org. Coat., 1986, 14, 159.
44. Larson, D. B.; Emmons, W. D., J. Coat. Technol., 1983, 55(702), 49.
45. Resimene AM-300 and AM-325, Tech. Bull., Monsanto Chemical Co. (now Solutia, Inc.),
St. Louis, MO, 1986.
46. Badou, I.; Dirlikov, S., Polym. Mater. Sci. Eng., 1994, 70, 334.
47. Bieleman, J., Surf. Coat. Int. A, 2003, (10) 411.
48. Hofland, A., in Technology for Waterborne Paints, Glass, J. E., Ed., American Chemical
Society, Washington, DC, 1997, p. 183.
49. Hulden, M.; et al., J. Coat. Technol., 1994, 66(836), 99.
REFERENCES
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
657
Schwarz, J., J. Coat. Technol, 1992, 64(812), 65.
Reynolds, P. A., Prog. Org. Coat., 1992, 20, 393.
Fernando, R. H.; et al., J. Oil Colour Chem. Assoc., 1986, 69, 263.
Hall, J. E.; et al., J. Coat. Technol., 1986, 58(738), 65.
Lundberg D. J.; Glass, J. E., J. Coat. Technol., 1992, 64(807), 53.
Hoy, K. L., J. Coat. Technol., 1979, 51(651), 27.
Mercurio, A.; et al., J. Oil Colour Chem. Assoc., 1982, 65, 227.
Geel, C., J. Oil Colour Chem. Assoc., 1993, 76, 76.
Arendt, W. D.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2001, pp. 497 – 509.
Schuler, B.; et al., Prog. Org. Coat., 2000, 40, 139.
Monaghan, G., Formulating Techniques for a Low VOC High Gloss Interior Latex Paint, FSCT
Symposium, Louisville, KY, May 1996.
Tech. bull. 77265, Rohm & Haas, Philadelphia, PA, 2004.
Aznar, A. C., Prog. Org. Coat., 2006, 55, 43.
33
Special Purpose Coatings
The term special purpose coatings designates industrial coatings that are applied outside a
factory. In 2002, they accounted for about 20% ($3.15 billion) of the value of U.S. coatings shipments [1]. Although the volume is the smallest of the three classes of coatings, it
is substantial and the value per unit volume is the highest. Separate profit figures are not
available, but historically, the operating profit percent for entrenched suppliers is also the
highest. Many different end uses are involved; our discussion centers on four of the larger
markets: maintenance, marine, automobile refinish, and aircraft coatings.
33.1. MAINTENANCE PAINTS
The term maintenance paints is generally taken to mean paints for field application,
including paints for highway bridges, refineries, factories, power plants, and tank farms;
not usually included are paints for office buildings or retail stores, which are classified
as architectural coatings. For many maintenance paints, a major requirement is corrosion
protection (Chapter 7). Sometimes, the terms heavy duty maintenance paints, anticorrosive paints, protective coatings, or industrial paints are used, all implying that the
paints must perform more effectively in aggressive environments than trade sales paints.
Although selling prices are important, the major economic demands of customers
emphasize proven performance, the time interval to be expected between repainting, and
service, rather than the cost of the paint. The frequency of repainting in a factory is
especially critical, since production may have to be shut down for repainting.
Since no laboratory tests adequately predict field performance of maintenance coatings,
potential customers want to inspect actual field use examples of a coating system being
recommended to them. State highway departments and the larger oil, chemical, and
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
658
33.1. MAINTENANCE PAINTS
659
construction companies employ specialized engineering groups responsible for working
both with coating suppliers to select appropriate coatings for their company’s needs and
with applicators to specify the application parameters. These groups also serve as inspectors to see that the coatings are applied properly. Records are kept of surface preparation,
application conditions, coating composition, and coating supplier; and the performance of
coatings on the various installations is monitored. The composition of the coatings is a
major variable, but surface preparation and application procedures are also critical to performance. The Steel Structures Painting Council is a valuable source of information about
the effects of such variables in its Steel Structures Painting Manual [2]. The problems of
predicting the service life of coatings are discussed in the introductory paragraphs of
Chapter 4, and the status of corrosion testing and performance prediction is described
in Section 7.5. As discussed in those sections, reliability theory is a potentially powerful
tool [3].
The most commonly used method for surface preparation has been sandblasting. It can
be very effective for removing surface rust; it also roughens the surface, increasing the
contact area with the coating, which in turn promotes adhesion. However, sandblasting
creates hazardous dust: small particles of silica and paint debris, often containing toxic
pigments. Elaborate and expensive containment systems are necessary to prevent
worker exposure and contamination of the environment. As discussed in Section 6.4,
many modifications and alternatives to sandblasting are being studied. They include dry
abrasive blasting with a variety of other media, and ultrahigh pressure hydroblasting at
pressures above 25,000 psi. Whatever the method, only a relatively limited area should
be cleaned at a time, and primer should be applied as soon as possible. One of the
authors was asked to explain why there had been a massive failure of coatings on a
Taiwan oil tank farm located only a few hundred meters from the sea. The tanks had
been carefully sandblasted and a high performance coating system had been used.
However, there was a delay of some days between sandblasting and coating. In the meantime, spray from the sea deposited salt on the surface; the primer was sprayed over
salt-contaminated steel. The salt crystals dissolved in water that permeated though the
coating, leading in a few weeks to blistering and massive failure.
The coating system must be carefully selected for the particular installation. Most
systems include at least two types of coatings: a primer and a top coat. Frequently,
more than one layer of primer and/or top coat is applied, and in some cases, a combination
of primer, an intermediate coat, and top coats is used. Primers provide the primary corrosion control, but intermediate and top coats also have significant effects on corrosion
protection by reducing oxygen and water permeability of the combined films. Top coats
also serve to protect the primer and provide other properties, such as gloss, exterior durability, and abrasion resistance. Three classes of primers are used: barrier primers, zinc-rich
primers, and primers containing passivating pigments.
33.1.1. Barrier Coating Systems
As discussed in Section 7.3, a requirement for a barrier system primer is that it have excellent wet adhesion. The primer should have a low viscosity continuous phase with slow
evaporating solvents to permit as rapid and complete penetration as possible into microscopic cracks and crevices in the metal surface. Amine-substituted binders are particularly
resistant to water displacement. Phosphate esters such as epoxy phosphates have also been
shown to enhance wet adhesion (Section 13.5). The Tg of the fully reacted binder should
660
SPECIAL PURPOSE COATINGS
generally be only a little above the ambient temperature at which the curing is to be
carried out. If the Tg of the fully reacted cross-linked film is too high, the rate of the crosslinking reaction becomes mobility controlled, and the reaction may stop prior to complete
conversion (Section 2.3.2).
Two package (2K) primers formulated with bisphenol A (BPA) and/or novolac epoxy
resins in one package and a polyfunctional amine in the other are widely used, since they
provide good adhesion combined with excellent saponification resistance (Section 13.2).
The novolac epoxy resins are prepared from novolac resins made with phenol, cresol,
bisphenol A, and/or tetraphenolethane. To reduce viscosity, the triglycidyl ether of trimethylol propane (TMP) can be used as a reactive diluent. A coating formulated with
the triglycidyl ether of TMP and a phenol –formaldehyde novolac resin showed better
resistance to methylene chloride, acetic acid, and sulfuric acid than did one formulated
with a bisphenol F resin [4].
Pot life is increased by using relatively high molecular weight (MW) epoxy resins and
amine cross-linkers at relatively low solids, since this minimizes the concentrations of
reactive groups. Especially when the solvents are slow evaporating, the vehicle can penetrate well into surface irregularities. Pigmentation should be high but somewhat below the
critical pigment volume concentration (CPVC), to reduce permeability to oxygen
and water and give a low gloss surface that enhances adhesion of top coat to the
primer. Intercoat adhesion is further enhanced by applying top coat before cure of
the primer is complete. Only a thin coat of primer is needed, but it is generally desirable
to apply multiple coats of primer to assure that the entire surface of the metal has been
coated. Sometimes, two different epoxy formulations are used: one as primer and the
other for intermediate coats. To reduce VOC emissions, higher solids coatings
are being adopted. As solids increase, the challenges of formulating coatings with adequate pot life becomes more difficult. The lower MW reactants in high-solids coatings
increase the possibilities for toxic hazards.
Waterborne epoxy–amine primers are being used increasingly (Section 13.2.6).
Novolac epoxy resins are frequently used in waterborne coatings since the higher MW
does not affect viscosity. Novolac epoxies dispersed with a proprietary surfactant combined
with a dispersion of a proprietary amine are reported to give excellent performance [5].
Alkyd formulations are used as primers, although their use is declining. 2K urethane
coatings are also used as primers. Aromatic polyisocyanate prepolymers are chosen
over aliphatic polyisocyanates due to their lower cost. They show strong adhesion to
metals, including galvanized steel, concrete, and wood. Their resistance to saponification
is a major advantage compared to alkyd resins in primers; this is especially important in
coatings for galvanized steel and concrete, which have highly alkaline surfaces. Moisturecurable urethane coatings are also used as primers. Commonly primers are formulated
with PVC . CPVC, resulting in somewhat porous films. This permits water penetration
through the film, minimizing the CO2 bubbling from reaction with water.
The poor exterior durability of BPA and novolac epoxy –amine coatings limits
their use as top coats to interior applications, for which their generally high chemical
resistance makes them useful. An exception is that most epoxy – amine coatings do not
resist acetic acid (or similar organic acids) very well. In contrast to inorganic acids,
acetic acid dissolves in the films; this effect is promoted by the presence of amines,
especially if cross-link density is inadequate. Novolac epoxy resins have a higher
average functionality than BPA epoxies and generally provide greater organic acid
resistance.
33.1. MAINTENANCE PAINTS
661
Generally, top coats have a different composition than primers. Chlorinated polymers
such as vinyl chloride copolymer resins and chlorinated rubbers are used as top coat
vehicles, since they have low moisture vapor and oxygen permeabilities (Section
17.1). Chlorinated resins require stabilization against photodegradation (Section 5.4).
Polyvinylidene fluoride/acrylate copolymer latexes offer outstanding exterior durability
[6] (Section 17.1.4). Since they are not cross-linked, they remain solvent sensitive and
are therefore not appropriate for applications such as petroleum refineries and chemical
plants. Use of chlorinated resins is limited by the high VOC required due to their high
MW. Polypropylene has low water permeability and is being used for top coats on
pipelines [7]. A system of an epoxy primer, an intermediate coat of a polar-modified
polypropylene copolymer, and a polypropylene top coat is used. Since polypropylene
is subject to photoxidation, UV stabilization is required.
Two package urethane coatings are used increasingly as top coats, due to lower VOC
and high solvent resistance of the cured films; urethanes are particularly useful when
abrasion resistance is an important requirement (Section 12.4). An important variable in
2K polyisocyanate –polyol coatings is the NCO/OH ratio. In ambient-cure systems, it
is often found that a ratio of the order of 1.1 : 1 gives better film performance than a
1 : 1 ratio. A probable reason is that part of the NCO reacts with water from solvent,
pigment, or air to give urea cross-links. To the extent this happens, two hydroxyls are
unreacted for each water molecule; excess NCO minimizes residual unreacted hydroxyl
groups. Solvent resistance is also improved. Since polyisocyanates are typically lower
in viscosity than the polyol, excess NCO gives a lower VOC content [8].
2K urea coatings made from polyamines and polyisocyanates have the advantages that
they cure rapidly even at freezing temperatures, have almost zero VOC, and have good
low temperature flexibility. Aromatic amines react slowly enough that they can be used
in 2K coatings. For example, MDI – polyoxypropylene prepolymers with aromatic
amines are used in maintenance coatings for which gloss and color retention are not critical [9]. Roofs and floors can be walked on within 30 seconds after they are coated with
such polyureas.
For 2K ureas that require exterior durability, hindered aliphatic amines are used. In one
example, one package was a mixture of hindered and less hindered di- and tri-amines, and
the other was HDI isocyanurate. The coatings had adequate pot life, tack free time of 2.3
minutes at 258C, strong adhesion to a variety of substrates, chemical resistance, and
exterior durability. No primer is required [10].
The chemistry of moisture-curable urethanes is described in Section 12.6, where it is
explained that they are not actually urethanes, but ureas. These are one-package (1K) formulations that can cure at temperatures as low as 08C, generally by reaction with moisture
in the air. They have certain advantages over two-package (2K) maintenance coatings. The
possibility of incorrect mixing ratios is eliminated, as are potential problems of mixing
more 2K coating than can be used before the viscosity has increased too much for application or of not enough and running out before a work shift ends. A further advantage is
that 1K formulations can be applied to damp surfaces. As compared to 1K systems that
cure by autoxidation, moisture-curable coatings can have superior resistance to oxidation
and exterior exposure (if aliphatic diisocyanates are used).
Compared with alkyd maintenance paints, moisture-curable urethanes generally exhibit
superior abrasion resistance, chemical and solvent resistance, hydrolytic stability, superior
adhesion retention over alkaline surfaces such as cement-based substrates or metal
substrates, and with aliphatic isocyanates, superior exterior durability. Reference [11]
662
SPECIAL PURPOSE COATINGS
provides a useful discussion of moisture-curable urethane coatings, especially for maintenance coatings. They are used for applications such as floor coatings, for which abrasion
resistance and hydrolytic stability are especially important. They can also be applied at
low temperatures and relatively high humidity. The coatings can be applied over metal
when the temperature is below the dew point since the isocyanate groups react with
the water condensed on the surfaces [12]. Reference [13] describes several examples of
application to various bridges and other applications.
On the other hand, moisture-curable urethanes are more expensive than alkyds. A
further drawback is that cure rates depend on the water content of the air; at low temperatures, higher relative humidity is required than at higher temperatures, since water content
decreases with decreasing temperature for the same relative humidity. At high humidity
and temperature, cure is rapid, but the carbon dioxide released by the reaction of isocyanate with water can be trapped as bubbles, especially in thick films. See Ref. [14] for a
discussion of the effects of temperature and humidity and other application considerations.
A further consideration is that solvents, pigments, and other coating components must be
essentially water free. Moisture-curable coatings are often used as clear, glossy coatings
because it is expensive to remove adsorbed water from pigments. Molecular sieves can be
used to adsorb water, but they reduce gloss. Pigmented moisture-curable coatings can
be made using a water scavenger such as an alkyl orthoformate or p-toluene sulfonylisocyanate. Such water scavengers, especially sulfonylisocyanates, must be handled with care due to
toxicity. Isophorone diisocyanate (IPDI) can be used as a moisture scavenger [15]; moisturecurable urethane coatings using IPDI are used on military vehicles. Oxazolidines are reported
to be effective moisture scavengers for aliphatic isocyanate-based coatings, presumably with
less toxic hazard and lower cost [16]. 4-Ethyl-2-methyl-2-(3-methylbutyl)-1,3-oxazolidine
provides rapid water removal while not adversely affecting the stability or performance of
a coating.
In most cases, atmospheric water is the curing agent, but in some cases spray application of liquid water or steam on the uncured film is recommended. A patent describes
the use of steam fed spray application of moisture-curable coatings; cure is very rapid,
and film thickness as high as 200 mm are feasible [17]. Through cure can also be
accelerated by putting a hygroscopic liquid in a moisture-curable formulation. For
example, N-methylpyrrolidinone [18] and g-butyrolactone [19] have been used for this
purpose [20].
2K waterborne urethane coatings have very low VOC emissions. A comparison of their
performance with that of solventborne high solids 2K urethane coatings showed that when
applied at 258C (758F) or below, the performance was similar. Under high humidity/high
temperature conditions the films from the solventborne coating showed poor surface
smoothness and appearance, whereas the waterborne coating had superior appearance.
Five-mil films of the waterborne coating could be applied satisfactorily, but solventborne
coatings are only recommended up to 2 mils. The waterborne coating was formulated with
a water-reducible acrylic resin and a hydrophilically modified polyisocyanate [21].
Alkyd resins are still used, although their saponification resistance and exterior durability are inferior. Coatings based on alkyds are generally lower in cost and have intermediate VOC emissions. As a result of their low surface tension, alkyd-based coatings
are less likely to develop film defects during application.
As described in Section 15.7, uralkyds have properties superior to those of alkyds.
Studies indicate that the rate of water permeability in alkyd films is higher than in the
uralkyd films [22].
33.1. MAINTENANCE PAINTS
663
Waterborne autoxidizable urethane coatings are being used (Section 15.7.2). A partial
ester of linseed oil fatty acid with PE is reacted with MDI and then with maleic anhydride
to form a maleated uralkyd [23]. The resin is diluted in 2-butoxyethanol, neutralized with
TEA, and diluted with water. Polyunsaturated acid– substituted aqueous polyurethane dispersions (PUDs) are also being used. For example, a PUD made by reacting a diisocyanate
with a diol, a fatty acid monoglyceride, and dimethylolpropionic acid (DMPA) is neutralized with an amine and then dispersed in water. Water acts as a chain extender for terminal
isocyanate groups [24]. Hydrolytic stability of the aqueous dispersions is excellent. Films
also have excellent abrasion resistance. Drying speed, color, and cost are dependent on
the drying oil used. Aliphatic diisocyanates provide superior color retention and exterior
durability at a higher cost. Cost can be reduced by blending in 10 to 20% of acrylic latex.
Corrosion protection generally increases as water vapor and oxygen permeability of the
coatings are reduced (Section 7.3.3). High levels of pigmentation reduce permeability and
cost, but also reduce gloss. Sometimes, low gloss intermediate coats are applied over the
primer, followed by a high gloss top coat. Platelet pigments, such as mica and micaceous
iron oxide, that can orient parallel to the surface of the coating as the solvent in the coating
evaporates are particularly effective in reducing permeability. Pigmentation of the final
top coat with leafing aluminum flake pigment is desirable because an almost continuous
layer of aluminum forms at the surface.
Since zinc rapidly develops strongly basic corrosion products, alkyds are not suitable
vehicles for galvanized steel. An extensive study showed that epoxy – polyamide primer
with a urethane top coat gave the best performance on galvanized steel when exposed
for 4 12 years at Cape Kennedy, Florida [25].
33.1.2. Systems with Zinc-Rich Primers
A typical modern coating for steel bridges consists of three layers: a zinc-rich primer, an
epoxy intermediate coat, and a polyurethane top coat. The primer provides galvanic protection, the intermediate coat is a barrier, and the top coat is an additional barrier for
enhanced weather resistance.
As discussed in Section 7.4.3, zinc-rich primers provide high performance either when
the surface cannot be completely cleaned of rust or when complete penetration into surface
irregularities cannot be achieved. The zinc acts as a sacrificial metal, protecting the
steel from corrosion. To be effective, the level of zinc pigmentation must be such that
PVC . CPVC. The high level of pigmentation assures electrical contact between the
zinc particles and porosity that permits water to enter the film, establishing a conductive
circuit with the steel surface. Even after a significant fraction of the zinc is consumed, the
primer continues to provide protection, perhaps because the pores fill with Zn(OH)2 and
ZnCO3, whose alkalinity may provide passivation [26] or perhaps because they form a
barrier to water and oxygen. In most applications, top coats are required to protect the
zinc against corrosion, reduce probability of mechanical damage, and provide a desired
appearance. See Ref. [27] for a review of the types of zinc-rich primers together with
their application and performance.
There are two classes of zinc-rich primers: inorganic and organic, each available in
solvent-borne and waterborne formulations. The binder of inorganic zinc-rich primers is
a prepolymer derived by reacting tetraethyl orthosilicate with a limited amount of
water. The chemistry of cross-linking is described in Section 16.4. Cross-linking
depends on moisture. When the coating is applied at low humidities, especially if
664
SPECIAL PURPOSE COATINGS
the temperature is high, film properties such as abrasion resistance may be adversely
affected [28]. If it is necessary to paint under hot and dry conditions, the coating should
be mist-sprayed with water immediately after application. A review of the effects of
zinc particle size and shape, extender pigments, and application techniques of ethyl
silicate –based primers is available [29].
Organic primers command a share of the market by their advantages of greater tolerance to incomplete removal of oils from the substrate, easier spray application, and
better compatibility with some top coats [30]. The binders in organic zinc-rich primers
are often based on 2K epoxy – polyamine (polyamide) resins. Polyurethane binders are
used increasingly. For example, the Michigan Department of Transportation recently
adopted a urethane zinc-rich primer as its standard [31], replacing an epoxy zinc-rich
primer [32].
In the older literature there is published evidence that in some circumstances, inorganic
zinc-rich primers can provide better protection than organic primers. For example, in a
seacoast environment, a six-year service life was estimated for inorganic primers compared to three years for the organic type [30]. More recently, however, organic zinc-rich
primers have been adopted for many demanding applications, indicating that knowledgeable people expect them to be long lived. For example, in the early 2000s, zinc – epoxy
primers were supplanting zinc – silicate primers in North Sea oil well applications [33];
the change is said to be driven by health, safety, and cost considerations as well as unspecified improvements in technology. Organic zinc – epoxy – urethane systems are favored
increasingly and sometimes required in many U.S. states [34] and are typically used for
onshore and offshore steel windmills [35].
Zinc-rich primers formulated with moisture-curable urethanes are used for new construction when the metal surface is blast cleaned to white metal conditions [36]. Reference
[37] provides a discussion of variables in formulating such zinc-rich primers. Resistance to
saponification is critical for a binder for zinc-rich coatings since the zinc reacts with
oxygen and water, producing zinc hydroxide and zinc carbonate. Thus, urethanes are
much more suitable than alkyd resins. Epoxy – amine 2K zinc-rich primers have been
used widely, but the urethane primers have the advantage of being 1K coatings.
In response to the need to reduce VOC, waterborne zinc-rich primers have been developed. As of 2005 they had captured only a small fraction (,3%) of the heavy-duty maintenance market in Europe, and early products admittedly performed poorly [38]. Both
inorganic and organic waterborne zinc-rich primers are now available that are reported
to rival solventborne counterparts in protective properties and have much lower VOC;
different application methods are required [38]. For inorganics, a typical binder is a combination of potassium, sodium, and/or lithium silicates with a dispersion of colloidal silica
[39]. Excellent performance on oil and gas production facilities in marine environments
has been reported [40].
A challenge in using zinc-rich primers is proper selection and application of the top
coat. The activity of the primer depends on maintaining its porosity by having PVC .
CPVC. If the top coat vehicle penetrates through pores in the primer, PVC of the
primer is reduced to a level approximately equal to CPVC, reducing the effectiveness
of the primer. Penetration of the continuous phase of a top coat into pores is controlled
primarily by the viscosity of the external phase. Penetration can be minimized by applying
a mist coat (i.e., a very thin coating) of top coat first. The solvent evaporates rapidly from a
thin coating, so the viscosity of the continuous phase increases rapidly, minimizing penetration and sealing the pores. Proper application requires considerable skill, since excess
33.1. MAINTENANCE PAINTS
665
mist coat in an area leads to increased penetration into the pores, but lack of coverage of an
area results in insufficient sealing in that area, so the vehicle of the next coat of top coat
will penetrate into the pores. Sealing the pores minimizes pinholing or bubbling during
application of thick layers of top coat. It is desirable for the color of the mist coat to
contrast with that of the zinc-rich primer to aid the sprayer in applying complete but
not excessive coverage. After the mist coat is applied, application of further coats requires
no special spraying skill. Since alkaline zinc oxide, hydroxide, and carbonate are present
on the zinc surface, top coats in contact with the primer coat must be stable against saponification. Two package urethane, vinyl, or chlorinated rubber coatings are used. Often, an
intermediate epoxy coat is applied to the primer, followed by a urethane top coat.
33.1.3. Systems with Passivating Pigment Containing Primers
When extensive film damage must be anticipated and when the substrate cannot be completely cleaned (especially when oily rust will remain), or when the coating cannot penetrate into the surface irregularities, passivating pigment primers are the primers of choice
(Section 7.4.2). In contrast to the situation with baked OEM product coatings, for which
passivating pigments are seldom appropriate, there are many cases in field applications for
which they are preferred despite the fact that blistering is more likely to occur.
A range of primer vehicles is used. Alkyds have the advantage that they are relatively low
cost and wet oily surfaces, but they are deficient in saponification resistance. Epoxy –amine
primers provide greater saponification resistance and good wet adhesion. Epoxy ester
binders are intermediate in cost and performance.
For many years, oil and alkyd paints with lead anticorrosive pigments were commonly
used on bridges and other steel structures. They performed fairly well, especially since
their relatively low cost made it economical to apply thick films. But repainting is a
major problem—safe removal of the old lead paint and replacement with a lead-free
coating is extremely expensive. Hare estimated a cost of $54 to $217 per square meter
in 1997, which is probably much higher today [41]. Today, the emphasis is on coatings
that will afford the longest possible intervals between repainting.
Zinc yellow was, for many years, the passivating pigment of choice (Section 7.4.2).
However, zinc yellow is carcinogenic, and care must be taken to avoid inhaling or ingesting spray dust, sanding dust, or welding fumes; in some countries, its use has been prohibited. Zinc yellow is being replaced by other passivating pigments; possible replacement
pigments are discussed in Section 7.4.2.
In recent years, the use of latex paint systems for maintenance applications has increased.
When water is applied to a freshly abrasive blasted steel surface, there is almost instantaneous rusting, called flash rusting. Flash rusting occurs with latex paints. To avoid
it, the formulation should include an amine such as 2-amino-2-methyl-propan-1-ol
(AMP). Use of mercaptan-substituted compounds as additives to prevent flash rusting
has also been recommended [42].
A paper has been published reporting on studies of factors influencing flash rusting and
tests for evaluating it. Relative humidity during application is a major factor, especially
when it is over 80%. Use of a “complex zinc compound in a mixture of anionic ionogenic
solvents” gave better protection than a 10% aqueous solution of a 9 : 1 ratio of sodium
benzoate and sodium nitrite solution. The composition of the latex affects the extent of
flash rusting. A copolymer of acrylic acid, methacrylic acid, 2-ethylhexyl acrylate, and
styrene gave good results. The easiest way to evaluate flash rusting is to apply
666
SPECIAL PURPOSE COATINGS
pigment-free films to the steel since then there is nothing to hide the development of the
red rust layer as it forms. A variety of other test methods are reported [43].
Since latex particles are large compared to the size of many of the crevices in the
surface of steel, and since the viscosity of the coalesced latex polymer is extremely
high, complete penetration into the crevices cannot be expected. Therefore, it is essential
to use a passivating pigment. The pigment must be chosen such that the concentration of
polyvalent ions is low enough so that package stability of the latex is not affected
adversely, but is still high enough that it can serve its passivating function. Strontium
chromate is less soluble than zinc yellow and is preferred. Zinc phosphate has been
used. Among the newer passivating pigments, zinc –calcium molybdate and calcium
borosilicates have been recommended.
Acrylic, styrene – acrylic, and vinylidene chloride/acrylic latex polymers are completely resistant to saponification. Special proprietary latexes are sold that provide enhanced
adhesion to metal in the presence of water. It has been suggested that amine-substituted
latexes are especially appropriate for wet adhesion. Use of 2-(dimethylamino)ethyl acrylate or methacrylate as a co-monomer is a way to incorporate amine groups on the latex
polymer. Methacrylamidoethylethyleneurea has also found use as a wet adhesion promoting monomer. Another approach for enhancing wet adhesion is to use an alkyd or
a modified drying oil to replace part of the latex polymer. The alkyd is emulsified
into the coating. After application, as the water evaporates, the emulsion breaks and
some of the alkyd can penetrate the fissures in the steel surface. Epoxy esters are
more hydrolytically stable than alkyds and would be expected to provide better corrosion
protection.
Owing to the high moisture vapor and oxygen permeability of most latex paint films,
it is desirable to use some platelet pigment in the formulation. Mica is used in both
primers and top coats. Final top coats using leafing aluminum (Section 20.2.5) are
particularly appropriate. To avoid problems of the reactivity of aluminum with water,
a special grade of leafing aluminum is mixed into a latex base paint just before
application.
The need for reducing VOC emissions has been particularly critical in California. Since
commercial paint suppliers did not have latex paints available for maintenance coatings
for bridges, the California State Office of Transportation Laboratory undertook the formulation and application of latex paints to several highway bridges in order to evaluate their
potential utility [44]. Latex primers were used in some cases and inorganic zinc-rich
primers in others; in all cases, latex top coats were used. It is reported that some
systems were still performing well after over five years field exposure. Based on the
experience obtained, it was recommended that application should only be done when
the temperature is above 108C and when the relative humidity is less than 75%. Use of
latex systems can be expected to expand as requirements for reduced VOC emissions
become more stringent.
Traffic striping paints are a large-volume part of the special purpose coatings market. In
the past, solventborne combinations of alkyds and chlorinated rubber formulated in
fast-evaporating solvents were used. Glass bead pigments are used to increase reflectance
of headlight beams back to the driver. Alkyds were displaced by preformed tape, epoxies,
and polyureas
Latex traffic paints are used widely in Europe, and several U.S. states are putting them
through final use tests. They have been compared to pre-formed tape, polyurea paint, and
epoxy paint. Their useful life was found to be three years, not as long as the other paints but
33.2. MARINE COATINGS
667
lower in total cost, including the earlier replacement. The epoxy paint was next lowest in
cost but had the added disadvantage that recoat adhesion was poor and the old traffic paint
had to be removed [45]. Thermosetting latexes are used exemplified by an ammonia neutralized binder consisting of a BA/MMA/MAA/acetoacetoxyethyl methacrylate latex made
using dodecyl mercaptan for MW control, aminoethylaminopropyltrimethoxysilane, and
a dimethylaminoethyl methacrylate/AA latex [46].
33.2. MARINE COATINGS
The marine coatings market includes coatings for pleasure craft, yachts, as well as for
naval and commercial ships. Coatings for pleasure craft are commonly sold on a retail
basis to the individual owner. A range of products is available; one for wood is spar
varnish. The original spar varnish (Section 14.3.2) was a phenolic – tung oil varnish; the
tung oil provides high cross-linking functionality, and the phenolic resin imparts hardness,
increased moisture resistance, and exterior durability. Although some phenolic –tung oil is
still used by traditionalists, the bulk of this market has shifted to uralkyds, which provide
greater abrasion and water resistance (Section 15.7).
The larger part of the market is for commercial and naval shipping and some shore
installations that are exposed to marine environments. A variety of substrate surfaces
are coated, and many require special coatings. Reference [47] provides a useful overview
of the market and products. A paper written in 2004 reports that the annual U.S. market for
marine coatings by market is: U.S. Navy, $4.6 million; commercial vessels, $14 million;
and the offshore drilling platform and related applications, $48 million [48]. The global
market is reported to be 500 million liters, of which 55 million liters are antifouling coatings [49]. The U.S. market is relatively small because most ships are painted in ports
where regulations are less strict and labor costs are lower.
Coatings for exterior surfaces above the waterline are generally analogous to those supplied for heavy duty maintenance, as discussed in Section 33.1. Ultra high pressure hydroblasting is used to clean the surface; salt is a common contaminant in shipyards, and
hydroblasting removes all salt from the surface. A primer must be applied soon after blasting to avoid recontamination. Inorganic zinc-rich or epoxy primers are used. In a study of
the performance of a range of coatings over zinc-rich primers on panels exposed for 10
years in Jacksonville, Florida, inorganic zinc-rich primer with an intermediate epoxy
coating and an aliphatic polyurethane top coat gave the best overall performance [50].
Inorganic zinc-rich primer with an epoxy second coat and a vinyl top coat received the
next highest rating for overall exterior performance. Where the requirements are somewhat less demanding, a two component epoxy primer with a silicone alkyd top coat
may perform adequately.
2K coatings formulated with an epoxy-functional silicone resin (Section 16.1.1) and a
carboxylic acid– functional acrylic resin have been evaluated as marine coatings for
topside application. They are reported to give superior exterior durability compared to
urethane coatings, including less dirt pick up and easier cleaning [48].
The exterior durability requirements of coatings for ship superstructures are severe,
since in addition to the direct UV, further UV is reflected off the water surface, and the
humidity is also generally high. Alkyd coatings have been the standard; however, urethane
coatings are being used increasingly because of their greater durability. Alkyds remain
popular because they are relatively easy for crew members who are not trained painters
668
SPECIAL PURPOSE COATINGS
to apply. Especially for interior paints, fire retardancy is important; chlorinated alkyd coatings have been recommended [51]. Latex paints are used to a minor extent; although initial
gloss is lower with latex than with alkyd coatings, gloss retention and resistance to cracking are superior. The low solvent content of latex paints reduces the fire hazard of paint in
storage on a ship. Deck paints have the further requirement of being skid resistant. 2K
aliphatic urethane coatings satisfy the exterior durability and abrasion resistance requirements. Skid resistance is obtained by mixing coarse sand into the paint just before
application.
For many ship bottom areas, 2K coal tar– epoxy coatings have been found to be particularly effective; they have also been used in ballast areas. The coatings must be handled
with care in view of the potential carcinogenicity of components of coal tar. Coal tar is
being replaced by petroleum derived hydrocarbon resins; in addition to lower toxicity
hazard, the lighter color of hydrocarbon resins facilitates inspection in ballast areas.
One package contains hydrocarbon resin, amine-terminated polyamide resin, and
pigment; the other contains the epoxy resin. Multiple coats are applied to give a total
film thickness on the order of 400 mm. The coatings have high chemical resistance and
high dielectric strength. The latter property is especially important for ships that use
zinc or magnesium sacrificial metal anodes to provide cathodic protection against corrosion (Section 7.2.2). Over properly prepared surfaces, the expected lifetime in seawater
immersion is on the order of seven years [47]. Pigmentation with leafing aluminum
enhances performance. To protect against environmental disasters, most oil tankers entering U.S. coastal waters are required to have a double-hull construction. This has increased
the surface area of ballast spaces, which is as much as six times the outer hull area. The
difficulty and high cost of repair work in double-hull construction dictates that the
coating systems offer very long term protection against corrosion.
Fluorinated urethane coatings are used in tanks such as fuel tanks, septic tanks, and
sometimes bilge tanks [52]. Fuel tanks must be protected against corrosion, since they
are frequently filled with seawater ballast after the fuel is consumed. The lower free
energy surface of the coatings simplifies cleaning.
The most challenging class of marine coatings is antifouling coatings. Spores and
larvae of a variety of plants and animals, ranging from algae to barnacles, can attach to
the underwater hull of a ship. The growth of plants and animals on ship bottoms increases
the roughness of the hulls, which in turn increases the turbulence of water flow over the
surfaces and drag; hence, speed decreases and fuel consumption increases. Removal of
such growths requires putting a ship into dry dock. The economic penalty of fouling is
very large. Review papers detail the variables involved, the history of antifouling coatings,
and the status of the coatings available and the various approaches being worked on
[49,53].
The major approach has been to apply coatings that contain biocides that leach out of
the coating. The biocide must be a very general toxic agent. The leaching rate must be
above a critical level required to kill all the organisms settling on the surface over a
long period of time. Leaching rates decay exponentially in accordance with a first order
release of biocide. The first biocide used on a large scale was cuprous oxide. Typical
binders are solution vinyl resins containing rosin salts. The binder is water sensitive, so
loss of cuprous oxide continues after the surface layer is depleted of cuprous oxide. To
have sufficient cuprous oxide at the end of the life of the coating, the rate of loss of
cuprous oxide in the early stages of use has to be excessive. The antifouling lifetimes
of such coatings in actual use are on the order of 7 to 24 months.
33.2. MARINE COATINGS
669
The economic need was for a coating with a lifetime of 48 months or longer. Coatings
based on copolymers of tributyltin (TBT) esters of methacrylic acid with conventional
acrylates were developed. They are hydrophobic and fairly low in Tg. The TBT near
the surface slowly hydrolyzes, liberating active biocide. As this happens, Tg increases
and the surface becomes more hydrophilic. As a result, surface layers separate from the
paint after they have released their biocide, exposing a fresh layer with more biocide.
The rate of hydrolysis is controlled by the design of the copolymer and also by the incorporation of a slightly water-soluble pigment. Leaching of the pigment from near the
surface of the film increases the contact area between polymer and water, increasing the
hydrolysis rate. Part of the pigment can be cuprous oxide, to provide further biocidal
activity. Since the rate of loss of biocide is controlled by the rate of hydrolysis at the
surface, the rate of leaching is approximately linear with time rather than exponential.
Furthermore, slow dissolution of polymer keeps the surface of the film smooth, which
favorably influences speed and fuel consumption. These coatings are called self-polishing
coatings. Depending on conditions, service lifetimes are up to five years.
When ships are in port, leaching of toxicants continues. In some harbors, concentrations
of toxicant can build up sufficiently to affect marine life. There is concern about potential
heavy metal entrance into the food chain, affecting fish that might be caught for human
consumption. Therefore, regulations controlling biocide release are getting more stringent.
In the mid-1980s, use of tributyltin compounds was banned on vessels less than 25 m in
length, affecting most pleasure craft. The International Maritime Organization has
drawn up legislation to restrict its use on all commercial vessels. If ratified, TBT will
be banned as of the beginning of 2008 [53].
Use of copper based antifouling paints has again increased, and self-polishing copper
acrylate coatings have been made available. However, their service lives are shorter
than that of tin containing self-polishing coatings. Copper release by hydrolysis of the
polymer is not sufficient to provide antifouling on its own; therefore, copper oxide and
a biocide are used with it. An example of an organic toxicant is 4,5-dichloro-2-n-octyl4-isothiazolin-3-one. It is reported to be an effective toxicant, and it degrades rapidly in
seawater, so that there is no bioaccumulation [54]. Although copper is much less toxic
than tin, some countries are banning its use as well.
Another approach to self-polishing antifouling coatings is the use of the zinc salt of
hydrogenated rosin with a plasticizer such as oleic acid, Cu2O, and a biocide. A study
of laboratory evaluations of the rates of leaching and dissolution of the upper layers of
the coating is available [55].
Intensive research in recent years has been aimed at attempts to find means of controlling fouling other than by use of toxicants [53]. It is a difficult problem. A promising
approach is the development of coatings to which adhesion of fouling organisms is so
limited that fast ship movement through the water or underwater brushes or hoses are sufficient to remove any fouling without dry docking. They are used on fast ferries, naval
ships, and some motor boats. The coating is applied over an epoxy primer for corrosion
protection. Service life is over three years, and the excellent leveling properties give a
smooth surface, which reduces drag. Research is directed at improving abrasion resistance
and tear strength of silicone elastomer coatings. Use of volatile methylsiloxanes, which are
exempt under VOC regulations, as a solvent permits use of higher MW elastomeric
silicones in the binder, thereby improving the physical properties [56]. Unfortunately,
speeds of 18 to 22 knots are needed to remove most of the growths [53]. Ships such as
freighters and tankers do not go that fast.
670
SPECIAL PURPOSE COATINGS
Coatings formulated with poly(4-hydroxystyrene) applied to the bottom of motor
boats reduce drag and significantly decrease fuel requirements. The same coatings
were applied to sea plane floats and permit take off with much shorter taxi distances
before lift off [57].
A paper is available that discusses the requirements for a coating from which
biogrowths are most easily removed. The polymer in the coating should have a flexible,
linear backbone which introduces no undesirable interactions and a sufficient number of
surface-active side-chain groups that are free to move to the surface to give the desired
low surface free energy. The polymer should have a low elastic modulus and give a
surface that is smooth at a molecular level, to avoid infiltration by a biological adhesive,
leading to mechanical interlocking. The backbone and the surface active side chains
should exhibit high molecular mobility. The thickness should be such as to control fracture
mechanics at the interface. Furthermore, the polymer should be chemically stable for long
periods in a marine environment. Polysiloxanes with side chains of nonafluorohexylmethylsiloxane or trifluoropropylmethylsiloxane are possible examples of polymers that
could meet the requirements [58].
Self-polishing coatings that contain no biocides are being developed. One example is
based on copolymers of a fluorinated monomer (e.g., 2,2,2-trifluoroethyl methacrylate),
an organosilyl monomer (e.g., trimethylsilyl methacrylate), and other (meth)acrylic
monomers. The erosion rate is adjusted by adjusting the proportions of monomers [59].
33.3. AUTOMOBILE REFINISH PAINTS
The value of coatings for cars and trucks outside assembly plants is almost as large as that
for OEM automotive coatings, discussed in Section 30.1. Some refinish paints are used for
overall finishing of cars and new trucks in special colors, but the largest segment of the
market is for repair. When a car has been in an accident and a fender is straightened or
a new door is installed, these parts must be coated so that the color matches the color
of the original coating. There are major technical problems meeting the demanding application and performance requirements, and also major marketing and distribution problems. Hundreds of types of cars are made each year, each in as many as a dozen
different colors. Although most cars on the road are five years old or less, many are
well over 10 years old. If there is an accident involving a 15-year-old Jaguar, the owner
expects to be able to bring it to a repair shop and have that section of the car painted to
match the rest of the car. Furthermore, the body shop expects to be able to call the
paint distributor and have the necessary liter of paint delivered by the next day.
The refinish industry is under increasing pressure to reduce emissions, complicating
the technical challenge. The national limits for VOC emissions are much more liberal
than those for many states, such as California. The present national EPA regulations
permit: wash primer, 780 g L21; primers and primer–surfacers 580 g L21; primer–sealer,
550 g L21; single-/two-step top coats, 600 g L21; top coats/three or more steps,
630 g L21; multicolor top coats, 680 g L21 [60].
In some cases, coating manufacturers make and stock small containers to match the
colors. These are ready to be shipped to dealers and repair shops before the first new
car comes off the assembly line. However, these “factory packaged” colors are limited,
due to inventory costs and the variations in color shade at the OEM manufacturing site.
In most cases, the manufacturers provide formulations and tinting color bases to
33.3. AUTOMOBILE REFINISH PAINTS
671
distributors (or larger repair shops) to permit matching any original color. Establishing the
formulations requires great color matching skill, especially for metallic colors.
Both air dry and force dry (65 to 808C) coatings are used. The heat in force dry coatings
makes a large difference in enhancing the quality of the repair. In Europe and Japan, most
shops are equipped with force dry ovens. In the United States, use of force dry ovens is
increasing, but there is still a significant fraction of repair shops that use air dry coatings.
Repair of damage after an accident is usually accomplished with replacement parts, as
labor costs to bump out dents have become too high. Replacement parts are usually
received from the manufacturer coated with electrodeposition primer. In some cases,
the damaged panel is repaired, then surface preparation is critical. The old coating
surface must be cleaned to remove dirt, tar, and wax by scrubbing with detergent,
rinsing thoroughly, and drying; in some cases solvent cleaners are required. Scuff
sanding may be required to remove chalky pigment and degraded polymer. If the old
paint is cracked, it must be sanded down to bare metal. Any breaks through to metal
require that the edge of the area be feathered out, that is, sanded with a bevel so that
there is a smooth change in film thickness. When bare metal is exposed, it must be
washed free of grease with solvent.
A primer –surfacer is applied after checking whether the coating on the car can withstand the primer – surfacer solvent. Current OEM coatings give no problem, but if the
car has been refinished in the relatively recent past, the coating might be only partially
cross-linked and lift when solvent is applied to it. In the past in the United States and
still in South America and Asia, the most common primer– surfacers are made with
nitrocellulose – alkyd binders. To achieve fast dry, a medium oil length, rosin-modified,
tall oil alkyd is used [61]. The rosin modification increases the Tg of the alkyd. The
primers are highly pigmented, for example, with a PVC of 38% and sometimes even
higher. The use of nitrocellulose with rosin-modified alkyd and high levels of pigmentation permit sanding the primer –surfacer within 30 minutes after application. The primer is
sanded smooth with fine grit paper, and the spray dust is removed. Application solids of
such primers is low. Stricter VOC controls and OEM warranty repairs are forcing the
use of epoxy and urethane primers, which also give superior performance. In areas with
the most restrictive VOC regulations, waterborne undercoats are being used. The initial
products have been formulated with acrylic latexes or polyurethane dispersions. Drying
of these coatings is dependent on humidity, and the time required for drying in high
humidity is excessive. They are being replaced by thermosetting 2K waterborne epoxy
(Section 13.2.6) or urethane (Section 12.8) primers.
The eco-efficiency of a UV cure primer compared with 2K urethane and epoxy primers
has been reported [62]. The VOC of the UV primer was 1.7 lb/gal compared with
4.7 lb/gal for the 2K urethane and 4.4 lb/gal for the epoxy primer. Only one coat was
needed for the UV primer, in contrast to two for the others. Curing time was 3 minutes
compared with 25 and 20 minutes for the urethane and epoxy primers, respectively.
The material cost for the UV primer was substantially higher, but the higher cost is
largely offset by lower labor, energy, cleanup, and rework costs. Furthermore, the capacity
of the shop was increased by 25%. The primer is in commercial use. BASF was awarded a
presidential green chemistry award for the development in 2005.
Waterborne 2K primer –filler coatings based on acetoacetate – amine systems are being
investigated. One package contains a polyacetoacetate such as trimethylolpropane triacetoacetate; the other contains an emulsion of a polyamine compound such as the reaction
product of a low MW BPA epoxy resin and a polyamine such as diethylenetriamine
672
SPECIAL PURPOSE COATINGS
neutralized with acetic acid. The primers show good stability, good pot life, excellent
adhesion, fast curing, and sandability [63].
Two broad classes of refinish top coats are used: lacquer (thermoplastic) and enamel
(thermosetting). A major application advantage of lacquer is fast dry; they have short
out-of-dust times (i.e., the time required for the film to become sufficiently dry that dust
particles do not adhere to the surface). There tends to be a great deal of dust in the air in
repair shops, and fast dry is a major advantage, reducing contamination of the freshly
applied coating. A disadvantage of acrylic lacquers for refinish is that when dried at room
temperature, their gloss is not high enough. To match the gloss of the OEM applied
coating, refinish lacquers must be polished with rubbing compound, adding to the cost.
In acrylic refinish lacquers, thermoplastic acrylic polymers with high methyl methacrylate content (Section 2.2.1) are blended with cellulose acetobutyrate (CAB) and a
plasticizer such as butyl benzyl phthalate. Gloss retention is good but inferior to
current OEM coatings. Application solids are low (10 to 12 NVV) and VOC emissions
are very high. Lacquers are now used in North America only for antique car restoration
and by the do-it-yourself market. They are still being used in countries, where there
are few environmental regulations, such as in parts of Central and South American
countries.
The other broad class of refinish top coats is enamels; there are several types. The oldest
and lowest in cost are alkyd enamels; the alkyd is a medium oil, isophthalic, soy, or tall oil
alkyd (Section 15.1). These enamels have several advantages in addition to low cost.
Fewer film defects such as crawling or cratering occur during application, due to their
low surface tension. Their gloss is high enough to match OEM monocoats without polishing. VOC emissions are substantially less than with lacquer. On the other hand, out-of-dust
time approaches half an hour, and if recoating is necessary, it must be done within four
hours after application or after 24 hours. At intermediate times, when the film is partially
cross-linked, a second coat of enamel can lead to lifting. Gloss retention of alkyd coatings
is poorer than that of OEM coatings or refinish lacquer.
The properties of alkyd enamels have been improved in several ways. Triisocyanates
such as isophorone isocyanurate (Section 12.3.2) can be added just before spraying.
They serve as auxiliary cross-linkers, reacting with free hydroxyl groups of the alkyds.
This supplemental cross-linking reduces out-of-dust time and recoating problems.
Another approach is to use methacrylated alkyds (Section 15.4), which give shorter
out-of-dust times but cross-link less rapidly, so the films remain sensitive to gasoline
for a protracted time. Another type resin is made by reacting an acrylic resin made with
glycidyl methacrylate as a co-monomer with drying oil fatty acids. Films from such a
vehicle show out-of-dust times and durability approaching that of acrylic lacquers and
superior to that of alkyds. Since the relatively high Tg acrylic backbone provides the
necessary initial dry, driers are not needed to accelerate the oxidative cross-linking.
The absence of driers improves exterior durability.
With the increasing use of base coat –clear coat finishes (Section 30.1.2), corresponding
coatings for use in refinishing were required. The base coat must be able to hide with two
coats. Relatively low solids are required to permit alignment of aluminum or other flake
pigments. Solvent evaporation must be relatively fast to allow clear coat to be applied
within an hour without disturbing the flake orientation. Acrylic lacquer was used initially
for refinish base coats but to accelerate drying is being replaced with a binder based on a
polyester with cellulose acetobutyrate, and to assist in fixing the flake, a polyethylene –
polyvinyl acetate wax suspension as a rheology control. Such base coats have a relatively
33.3. AUTOMOBILE REFINISH PAINTS
673
high VOC and are used with very high solids clear coats (up to 85% weight solids), so that
the combined VOC is within regulations.
Waterborne base coats based on either acrylic latexes or polyurethane dispersions are
being made available. These waterborne coatings have a volume solids of 15%, which
permits flake orientation comparable to OEM coatings. Treated aluminum flake must be
used to avoid hydrogen evolution (Section 20.2.5). Rheology control additives are also
required. Relative humidity conditions affect the drying speed; the coatings can be
force dried or air dried with high volume dry air flow to remove most of the water
before applying the clear coat.
Two package urethane enamels have become the most important for clear coats.
Thermosetting hydroxy-functional acrylic resins similar to those used in OEM coatings
can be used together with a polyfunctional isocyanate cross-linker. Polyfunctional
aldimines and hindered amines are being used to further increase solids (Sections 12.4
and 25.2.2). A copolymer of an adduct of glycidyl versatate and methacrylic acid, butyl
acrylate, styrene, and methyl methacrylate with low viscosity HDI isocyanurate is used
as a vehicle for a high-solids clear coat [64]. Unsymmetrical HDI trimer (HDI iminooxadiazenedione) has been recommended for use in 2K acrylic clear coats for refinishing
since 1 to 3% higher solids can be applied as compared with HDI isocyanurate coatings
[65]. Use of t-butyl methacrylate as a comonomer in acrylic polyols permit somewhat
higher solids, a 1-hour pot life, a tack free time of 35 minutes, and a higher gloss than
that of other acrylic polyols. Use of EPA exempt acetone and 4-chlorobenzotrifluoride
as part of the solvents permits making a 3.5 lb/gallon VOC [66]. For applications
where the volatility of acetone is acceptable but the ketone structure is not, methyl
acetate is an alternate exempt solvent. Recently, the EPA ruled that t-butyl acetate is
accepted as an exempt solvent; its use awaits state inclusion in implementation plans.
Bicyclic orthoesters are used commercially with polyisocyanates in very high gloss 3K
clear coats. The third component is the catalyst to promote hydrolysis of the orthoester
and reaction of the resulting hydroxyl groups with the isocyanate component. Clear
coats with a pot life of 4 hours are reported with 230 g L21 VOC that flash off in 1 to
2 minutes and dry in 5 minutes at 608C [67]. A patent application discloses the
use of 2-butyl-2-ethyl-1,3-propanediol-spiroorthoester prepared by reacting 2-butyl2-ethyl-1,3-propanediol with tetraethylorthosilicate, together with HDI isocyanurate as
cross-linker and DBTDL as catalyst. The DBTDL acts as a catalyst for both the hydrolysis
of the orthoester and the reaction of the resulting hydroxyl groups with the isocyanate. Pot
life is reported to be 7 hours [68].
Silylated acrylic resins have been developed that are reported to combine good appearance, mar resistance, acid resistance, and exterior durability [69]. The silyl groups are introduced using 3-trimethoxysilylpropyl methacrylate as a co-monomer. After application, the
resins are moisture-curable through the silyl group (Section 16.4). Very fast cure refinish
clear coats can be formulated using pentaerythritol tetraskis-3-meraptopropionate, HDI
isocyanurate, and a Zr(AcAc)2 catalyst. Ambient temperature cure time is reported to be
10 minutes, and at 608C, 4 minutes [70].
Dual thermal and UV cure clear coats are being investigated for refinishing. 2K coatings with a polyisocyanate in one package and a hydroxy-functional urethane/acrylate oligomer were used. The coatings were first cured at temperatures in the range 40 to 608C
followed by UV curing. The isocyanate cross-linking served to give acceptable properties
even in shadow areas when the films were UV cured. The mar and chemical resistance are
reported to be superior to a conventional 2K urethane clear coat [71]. A UV cure coating
674
SPECIAL PURPOSE COATINGS
system for spot repairing has been patented. A pigmented primer –surfacer with a BPA
epoxy acrylate and hexanediol diacrylate vehicle and a mixture of acylphosphine oxide
and hydroxyacetophenone type photoinitiators, is applied, UV cured, and sanded. A
pigmented base coat formulated with an aliphatic polyurethane acrylate and a mixture
of acylphosphine oxide and hydroxyacetophenone type photoinitiators is applied and
UV cured. A clear coat with the same vehicle as the base coat, to which a benztriazole
UV absorber and a HALS compound are added, is applied and UV cured [72].
Monocarboxylic acid –modified polyesters cross-linked with HDI or IPDI isocyanurate
have also been recommended [73]. Durability of these coatings is in the same range as
OEM coatings, and they give high gloss without polishing. Out-of-dust times are intermediate between lacquers and alkyd enamels. The out-of-dust times can be reduced by
increasing the amount of catalyst (often dibutyltin dilaurate), but pot life is also
reduced. Refinish shops cannot generally afford dual-mixing guns, so the pot life has to
be at least long enough to permit complete spraying of a car—over an hour. (See
Section 25.2.2 for discussion of approaches to formulating coatings with reduced VOC
and increased pot life.)
VOC emissions have been reduced by changing from conventional air spray guns to
high volume low pressure air guns (HVLP). HVLP guns permit better transfer efficiency
and hence lower paint use and lower VOC emissions (Section 23.2.1).
Although paint suppliers emphasize the need to wear masks and to mix and apply the
paint in a well-ventilated spray booth, some refinishers do not follow recommended safety
precautions. Especially if cross-linkers with some volatile diisocyanate components are
supplied, sprayers can develop respiratory problems. Some shops will not use 2K urethane
coatings. Refinish coating suppliers are pursuing non-isocyanate cross-linking systems.
It is important to bear in mind that any reactant that can cross-link hydroxyl, carboxyl,
or amine groups on a synthetic polymer can also cross-link proteins. Any such cross-linker
will be toxic. A new system must be designed in which the reactive component is not
volatile and has a high enough MW to minimize permeation through the skin and
membranes.
A significant aftermarket product is truck bed lining. The principal coating is a black
pigmented 2K urethane applied with dual feed spray guns. Fast gel coatings are needed
to avoid sagging owing to the film thickness of more than 100 mils. The isocyanate is
MDI/polyethylene glycol diol prepolymer. The resin package is a combination of polyether diol and triol with an aromatic diamine. The black pigment is to protect the
binder against UV degradation. An aliphatic 2K coating has also been developed that
permits colors that match the color of the truck. The isocyanate is HDI isocyanurate,
and the resin package binder is a combination of polyester polyols, a polyaspartate, and
aliphatic diamine. The amines are necessary to have a short gel time, The aliphatic
coating can be applied as a sole coating, or to reduce the cost the aromatic product
can be applied at a film thickness of about 120 mils followed by 40 mil of the aliphatic
coatings as a top coat [74].
33.4. AIRCRAFT COATINGS
The U.S. market for aircraft and aerospace coatings in 1990 was estimated to be $60
million [75]. The majority of the business is for exterior primer and top coats for
aircraft—roughly 50 : 50 for commercial and military aircraft.
33.4. AIRCRAFT COATINGS
675
Excellent adhesion and corrosion protection are required for a range of substrates:
aluminum, aluminum alloys, and composite plastics. The high strength aluminum alloys
used for aircraft structures are more vulnerable to corrosion than is ordinary aluminum.
The coatings must resist swelling by phosphate ester hydraulic fluids, lubricating
oils, and fuel. Swelling resistance is usually obtained by relatively high cross-link
density. However, the coatings must also have good flexibility at low temperatures and
excellent abrasion resistance while the aircraft is flying through dust, rain, or sleet at
high speed. Military aircraft in the Middle East must be repainted very frequently due
to abrasion. Flexibility and abrasion resistance tend to decrease with increasing cross-link
density.
Achieving strong adhesion to aluminum requires surface treatment. Chromate based
treatments have been used, but the toxicity of chromates has required the development
of chromate free treatments. See Section 7.3.2 for a discussion of various treatments.
Epoxy – amine primers are generally used (Section 13.2). For aircraft exteriors they are
generally topcoated with urethanes. For interior areas, such as in hold or cargo areas,
internal structural members, and wheel wells, the primers are generally not top coated
except for special applications. The interior coatings are expected to last for the life of
the aircraft and are not stripped and recoated as exterior coatings are. As aircraft are
used beyond their original design lifetimes, there can be problems. For example, in military fuel tankers, which are often older than their pilots, fragments of primer come loose
from the inside the wings and clog fuel filters. These primers are generally formulated with
polyfunctional novolac epoxies (Section 13.1.2) with amido-amines or amine adduct
curing agents to give relatively high cross-link density and maximum resistance to lubricating and hydraulic fluids. Passivating pigments are used to control corrosion in the case
of breaks through the film. Strontium chromate is preferred over zinc yellow, since its
solubility in water is somewhat lower, and therefore the rate of leaching of the pigment
from the film by water is lower (Section 7.4.2). Strontium chromate also has greater
heat resistance than zinc yellow.
Primers for exterior surfaces are designed with lower cross-link density by using only,
or primarily, BPA epoxy resins. Amine-terminated polyamides (Section 13.2.2) are
usually used as the cross-linking agents. The long chain fatty acid parts of the polyamides
provide good wetting to the metal surface and flexibility to the final film. Cross-linking is
catalyzed with 2,4,6-tris(dimethylamino)phenol. The lower cross-link density provides
greater flexibility and permits easier stripping of the coatings for repainting. Strontium
chromate is used as a pigment, and again owing to toxic hazard, chrome free primers
are being evaluated.
Top coats are almost always 2K urethane coatings. Most commonly, hydroxyterminated polyesters are used with isocyanurate trimers from hexamethylene diisocyanate or isophorone diisocyanate; commonly, a high isocyanate/hydroxyl ratio is used so
that the coatings are partially moisture cured (Sections 12.4 and 12.6). Cross-linking is
catalyzed by organotin compounds.
Aircraft coatings have had relatively high solvent contents, but there is pressure to
reduce VOC emissions. Waterborne epoxy –amine primers are being used increasingly
[30] (Sections 13.2.6 and 26.3). Some reduction of solvent in top coats has been achieved
by using lower MW resins and polyisocyanates, but progress has been limited by the
requirements for rapid dry at ambient temperatures, which is harder to achieve as the
MW is reduced. Very fast reacting coatings requiring dual spray show promise of very
low or even zero VOC. Aldimine based coatings are being evaluated (Section 12.4).
676
SPECIAL PURPOSE COATINGS
The exteriors of airplanes are frequently repainted. The biggest problem is not
application of the new paint, but removal of the old paint. The paint strippers that have
been used for many years contain substantial amounts of methylene chloride, which is
toxic and environmentally undesirable. Studies are under way to eliminate the use of strippers by using mechanical methods for paint removal. The removal method must not
weaken the metal or composite plastic substrate. For example, sandblasting will remove
coatings, but also erodes the metal surface. Alternative blasting media such as plastic
beads, dry ice pellets, and crystalline starch are being investigated (Section 6.4.1).
REFERENCES
1. U.S. Department of Commerce, Bureau of Census, Current Industrial Reports—Paint and
Coating Manufacturing, http://www.census.gov.
2. Kearne, J. D., Ed., Steel Structures Painting Manual, Vol. I, Good Painting Practices, 3rd ed.,
1993; Vol. II, Systems and Specifications, 7th ed., Steel Structures Painting Council, Pittsburgh,
PA, 1995. Hare, C. H., Protective Coatings, Steel Structures Painting Council, Pittsburgh, PA,
1995.
3. Martin, J. W.; et al., Methodologies for Predicting Service Lives of Coating Systems, Federation
of Societies for Coatings Technology, Blue Bell, PA, 1996.
4. Kincaid, D. S.; Schulte, J. A., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA,
2001, pp. 127 – 141.
5. Kincaid, D. S.; et al., Proc. Waterborne Higher-Solids Coat. Symp., New Orleans, LA, 2002,
pp. 303 –317.
6. Wood, K. A.; et al., http://specialchem4coatings (2005).
7. Guidetti, G. P.; et al., Prog. Org. Coat., 1996, 27, 79.
8. Jorissen, S. A.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
1992, p. 182.
9. Takas, T. P., JCT Coat. Tech, 2004, 1(5), 40.
10. Giles, D.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2003,
pp. 167 –182.
11. Gardner, G., J. Prot. Coat. Linings, 1996, 13(2), 34.
12. Schutz, D., Mater. Perform., 1998, 37(2), 32.
13. Bell, Q., Surf. Coat. Aust., 1982, 30(9), 50.
14. Gardner, G., J. Prot. Coat. Linings, 1996, 13(2), 81.
15. Urs, B. H., U.S. patent 4,304,706 (1981).
16. Robinson, G. N.; et al., J. Coat. Technol., 1993, 65(820), 51.
17. Liebl, H.; et al., U.S. patent 4,816,542 (1989).
18. Zhou, L. I., U.S. patent 6,709,539 (2004).
19. Brinkman, R. A.; Richards, J. I., U.S. patent 5,705,547 (1998).
20. Chandalia, K. B.; et al., U.S. patent 5,391,614 (1995).
21. Giles, D.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2001,
pp. 337 –348.
22. Xu, Y.; et al., Prog. Org. Coat., 2002, 45, 331.
23. Reichhold Albert Chemie AG, Br. patent GB 1,199,810 (1970).
24. Liu, W.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 1999,
p. 202.
REFERENCES
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
677
Drisko, R. W., J. Prot. Coat. Linings, 1995, 12(9), 27.
Feliu, S., Jr.; et al., J. Coat. Technol., 1993, 65(826), 43.
Kline, H. H., J. Prot. Coat. Linings, 1996, 13(11), 73.
Eccleston, G., J. Prot. Coat. Linings, 1998, 15(1), 36.
Parashar, G.; et al., 2001, Prog. Org. Coat., 42, 1.
Paert, J., J. Prot. Coat. Linings, 1992, 9(2), 50.
Anon., J. Prot. Coat. Linings, 2004, 21(7), 37.
Phifer, E., J. Prot. Coat. Linings, 1992, 9(2), 48.
Doble, O., J. Prot. Coat. Linings, 2004, 21(4), 22.
Kowalski, G.; et al., J. Prot. Coat. Linings, 2003, 20(1), 73.
Muhlberg, K., J. Prot. Coat. Linings, 2004, 21(4), 30.
Schwindt, J., J. Mater. Performance, 1996, 35(12), 25.
Sonntag, M., Polym. Colour J., 1982, 172(4081), 50.
Aamodt, M., Coat. World, 2005, 2(10), 30.
Montes, E., J. Coat. Technol., 1993, 65(821), 79.
A. Szokolik, J. Prot. Coat. Linings, 1995, 12(5), 56.
Hare, C. H., J. Prot. Coat. Linings, 1997, 14(11), 50.
Reinhard, G.; et al., Prog. Org. Coat., 1992, 20, 383.
Kalendova, A., Prog. Org. Coat., 2002, 44, 201.
Warness, R., Low-Solvent Primer and Finish Coats for Use on Steel Structures, tech. rep.
FHWA/CA/TL-85/02, 1985.
Acrylic Lines, Rohm & Haas, Philadelphia, PA, summer 2004.
Hermes, A. H.; et al., U.S. patent appl. 20030224184.
Bleile, H. R.; Rodgers, S., Marine Coatings, Federation of Societies for Coatings Technology,
Blue Bell, PA, 1989.
Fransehn, P.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2004, Paper 7.
Morrison, S., SpecialChem4Coatings and Inks, 2005, (1).
Sghibartz, C. M., FATIPEC Congress Book, 1982, Vol. IV, p. 145.
Wake, L. V.; et al., J. Coat. Technol., 1995, 67(844), 29.
Brady, R. F., Jr., Polym. Mater. Sci. Eng., 1998, 78, 249.
Yebra, D. M.; et al., Prog. Org. Coat., 2004, 50, 75.
Raber, L., Chem. Eng. News., 1996, 74(29), 9.
Yerba, D. M.; et al., Prog. Org. Coat., 2005, 53, 256.
Bausch, G. G.; Tonge, J. S., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans,
LA, 1996, p. 340.
Gesser, H. D.; Lafreniere, D. R. T., Prog. Org. Coat., 2004, 49, 372.
Brady, R. F., Jr., Prog. Org. Coat., 1999, 35, 31.
Aubers, M. A.; et al., U.S. patent 6,767,978 (2004).
http://www.epa.gov/ttn/atw/180e/arc/fr1194.txt.
McBane, B. N., Automotive Coatings, Federation of Societies for Coatings Technology, Blue
Bell, PA, 1987.
Wall, C. A., et al., FSCT Annual Meeting, 2004.
Geurink, P.; et al., Prog. Org. Coat., 2003, 48, 153.
Petit, H.; et al., Prog. Org. Coat., 2001, 43, 41.
678
SPECIAL PURPOSE COATINGS
65. Richter, F.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2003,
pp. 71–85.
66. Ball, P. A.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA, 2002,
pp. 119 –124.
67. van Beelen, D. C., FSCT Annual Meeting, 2004.
68. Van Den Berg, K. J.; et al., U.S. patent appl. 20040122203.
69. Chen, M. J.; et al., Surf. Coat. Int., 1996, 79, 539.
70. Klinenberg, H.; van Beelen, J. C., U.S. patent appl. 20030181625.
71. Maag, K.; et al, Prog. Org Coat., 2000, 40, 93.
72. Maag, K.; et al., U.S. patent 6,472,026 (2002).
73. Wilt, T. F.; et al., U.S. patent 5,468,802 (1995).
74. Zielinski, D. P.; et al., Proc. Waterborne High-Solids Powder Coat. Symp., New Orleans, LA,
2004, Paper 35.
75. Chattopadhyay, A. K.; Zentner, M. R., Aerospace and Aircraft Coatings, Federation of Societies
for Coatings Technology, Blue Bell, PA, 1990.
34
Perspectives on
Coating Design
Formulation of coatings is a challenging assignment. Sometimes, scientists look down
their noses at the coatings formulator without realizing that the task of formulating a
new coating can be more technically challenging than much so-called “pure” research.
Any coating must meet a multitude of requirements. There are innumerable possible
raw materials, combinations, and proportions. Test methods are generally subject to
large ranges of error and frequently do not give results that predict use performance
well. The formulator is faced with variable substrates and application methods. There
are commonly severe cost constraints. Frequently, the volume of any one coating is too
limited to justify expenditure of large amounts of time; furthermore, time available to
solve the problem is usually limited. In fact, as one looks at the complexities of the
field, one sometimes wonders how a useful coating ever gets formulated.
Historically, the problems of formulation were somewhat eased by following a
procedure of making small modifications of coatings known to be satisfactory. Over
time, excellent coatings were formulated in this way. This approach of continuous
improvement, with a focus on the customer’s needs and close contact between the formulator and the customer’s engineering group, continues to be important. However, it has
become necessary to make major changes in formulations in less time than was formerly
used to make even small changes. This need results from a variety of factors, but particularly from the introduction of VOC emission controls and the increasing number of raw
materials identified as having potentially serious toxic hazards.
Most of this chapter is written from the perspective of the United States, but our experience teaching courses in many countries around the world indicates that the differences in
the challenges to coatings design in different countries are small compared with the
common denominators.
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
679
680
PERSPECTIVES ON COATING DESIGN
The challenge of formulation is intensified by the need for increasing productivity and
creativity. We all want more of the good things of life, but the only way this desire is made
possible for the population as a whole is to increase the aggregate productivity. Some politicians and some uninformed segments of the population tend to associate productivity
problems with production line workers. Clearly, there is need for increased productivity
of labor, but overall productivity depends on the efficiency with which management, sales
personnel, accountants, clerical personnel, and laboratory personnel, including formulators, work. In the inimitable words of Pogo: “We have met the enemy and they is us.”
At least as important, and perhaps more important, is the need for increased creativity
in formulation. Some of the problems facing the coatings formulator seem impossible.
Some of them are, in fact, impossible, but many “impossible” problems aren’t really
impossible but require creative visualization of ways around the supposedly insuperable
hurdles. Enhanced creativity should be a critical component of most aspects of people’s
lives, and the need for creativity is especially important in technical people, including
formulators.
The coatings industry has been known as an industry with a relatively low profit
margin. It is commonly argued that more money cannot be invested in research and development because of low profit margins. Actually, the inverse is probably truer. Profit
margins are low because a large fraction of the technical effort is spent trying to copy competitors’ products and/or by applying the same old ideas as those being used by every
other formulator rather than focusing on innovative research and development.
There is no magical single route to increasing creativity and productivity. This chapter
provides some ideas based on the accumulated experience of four technical people
who have worked in different aspects of the coatings field. Other people will have other
suggestions; the particular approach is not important, but the end result is critical.
There are several important aspects involved in working on a formulating problem.
Although, to a degree, these are sequential steps, continuous review and reevaluation of
these aspects is essential as the work progresses.
1. Define the problem. The first stage in working on a formulating problem is to define
it. This seems so obvious that one would think that it need only be mentioned in passing.
However, experience shows that inadequate definition of the problem is a large factor in
unsuccessful technical projects. Appropriately defining a formulating problem is in itself a
difficult and relatively time consuming effort, but effort invested at this stage often saves
time overall. It is particularly critical to define a coating problem in terms of performance
requirements. The statement “Formulate a coating to meet Specification Number XXXX
with a 30% reduction in VOC emissions” does not define the performance requirements in
most cases. Very often, the specification is based on quality control tests that a satisfactory
previous coating happened to pass, but quality control tests are not designed to predict performance of a new product. Consequently, an effective formulator must know or find out
what the actual end use requirements are.
One should also examine closely the statement of a problem that he or she receives. On
one occasion, a laboratory was asked to formulate a “harder” coating; this was done and
the customer accepted the new formulation, but after some months of use, the customer
complained loudly that the new coating had a poor wear lifetime. The customer really
wanted improved abrasion resistance and assumed that greater hardness would provide
it. The customer was mistaken, but the fault was not really his; the fault was that of the
coating formulator, who didn’t ask: “Why does the customer think that he wants a
PERSPECTIVES ON COATING DESIGN
681
harder coating?” There is great need for direct contact and interaction between the
customer and the formulator who works on a problem.
It is not adequate to define a problem as: “Match competitor A’s Coating No. YYY at a
lower cost.” First, this definition propagates a self-fulfilling prophecy—that profits in the
paint industry will continue to be low. Second, by the time the work has been finished, the
competitor may have produced a better product than YYY, and thus a match for YYY is no
longer adequate to get the business. Third, the difficulties of analysis, especially for additives, are such that one can seldom analyze a competitor’s formulation precisely enough to
duplicate it. Matching the laboratory test results of the competitor’s product does not
assure that the product performance will be equal. The only satisfactory definition of
the technical aspects of a project is a listing of detailed performance requirements. In
any list of this sort there is a range of degrees of importance of the various needs.
Some are essential requirements, others are important, still others would be “nice if we
could do it,” and still others are in the category “As long as we are pipe dreaming, why
not put this goal down, too?” The laboratory worker should clearly understand how
each of the performance needs fit into such a scale.
Applicable regulations must be known—not just current regulations, but best estimates
of what they will be over the lifetime of the potential project. No one can predict future
regulations accurately, but a choice as to goals from this point of view must be made.
After all, assuming that current regulations are going to continue unchanged is also a
prediction of the future. Toxicity problems should be assessed, taking into consideration
possible future developments. Starting a new study of a corrosion protective primer based
only on zinc yellow pigment, with a target date for significant sales five years or so ahead,
is risky, since it is known that zinc yellow is a human carcinogen.
Test methods must be agreed on. Few laboratory tests are adequate predictors of use
performance. Since coatings are complex compositions and end use requirements are variable, there is a real danger in relying on any one test. Some widely used tests, such as salt
spray tests for corrosion protection, have repeatedly been shown not to correlate with end
use results (Section 7.5). It is absurd to base decisions for a major research project on
whether or not the product will or will not pass a salt spray test. The problem of evaluation
should be faced in advance. A variety of tests relating to mechanical, spectroscopic, and
thermal properties should also be carried out using coatings with known field performance
as standards for comparison. Dickie has published a methodology for integrating the
results of laboratory performance tests, field history, environmental factors, design
parameters, and the fundamentals of degradation to predict service performance [1].
The problems involved with predicting the service life of coatings have generally been
underestimated. See Ref. [2] for a review of the problems of prediction and the proposed
use of reliability theory for developing better methodologies.
Cost requirements are an integral part of the definition of a project. The formulator must
know what the real permissible upper cost is, not just the amount to which some salesperson hopes the cost can be kept down. Similarly, the timing requirement must be known.
Some projects have to be finished by a certain date. Some do not have a specific deadline,
although obviously, the sooner they are finished, the better. Unrealistic cost and time goals
can lead to wrong decisions in project planning. The technical person has the obligation to
let any possible time slippage or cost increase be known to others as soon as he or she is
aware of it.
The potential value of a project should be compared with the estimated total cost of the
project. Some companies use elaborate discounted cash flow methods of analyzing the
682
PERSPECTIVES ON COATING DESIGN
potential value of a project. Our experience has been that such analyses have killed many
good projects along with bad ones. It should be recognized that estimates can be used to
“prove” that any project will or will not be economically sound. On the other hand, there
have been frequent occasions when a little thought would have shown that there was no
sensible possibility of a return approaching the potential cost of the project.
Unfortunately, misunderstandings about the definitions of projects are all too common.
A way of minimizing such misunderstandings is to have the person who will be doing the
work write out his or her understanding of the project. It is far better to take the time
required to reach consensus on a project definition at the outset than to risk disagreements
midway or even near the project’s end.
2. Do a background search. Too often, laboratory workers jump on their horses
and ride off in all directions. First, one should assess the available knowledge. Review
the pertinent scientific literature, review (with appropriate concern for bias) suppliers’
technical data bulletins; and review any pertinent background with fellow workers,
in the company’s files, and on the Internet. Discussions with technical employees
of raw material suppliers are often very helpful. Discussions with the customers’ technical
and engineering personnel can also be useful. Increasingly, companies are accumulating
computer data banks that compare actual field performance with composition variables;
such data can provide useful ideas.
It is particularly critical to be sure that a problem is not impossible. Some problems are
impossible, and there is no point in working on them. No amount of wishful thinking will
ever permit the development of a flat jet black paint or a white paint with gloss equal to
that of a high gloss jet black paint (Section 19.10.1). No one can match the color of a gloss
coating at all angles of illumination and viewing with that of a low gloss coating (Section
19.9.1). No one will ever make a kinetically controlled one package coating with six
months of package stability at 308C that will cure in 30 minutes at 808C (Section
2.3.2). Stop such projects before they start.
3. Identify the approaches. Commonly, a technical person starts working on a project
based on the most obvious approach to the problem. This approach is probably the same
one that his or her counterpart in a competitor’s laboratory will select. The greatest opportunity to apply creativity comes early in a project. Devise all of the approaches you can to
solving a problem; solicit ideas from fellow workers. Don’t be dissuaded by veterans who
say: “We tried that 20 years ago and it didn’t work.” Assess the merits of their comments,
but in 20 years, many other things have changed; what didn’t work then might work now.
Set the problem aside for a few days and try again to come up with really different ways by
which the problem might be solved. Open up your mind; get out of ruts. Having accumulated a variety of possible approaches on which to work, try to assess their merits and pick
the one or two most promising. Sometimes, the one picked will be the first approach identified, but commonly, it is not.
As the basic understanding of the factors controlling the performance of coatings
increases, the opportunity for basing experimental approaches on sound scientific understanding increases. It is sad to see the failures to apply known understanding. To take an
obvious example, it is well established that in room temperature – cure coatings, crosslinking may become limited by the availability of free volume. If the Tg of the fully
reacted system is significantly above room temperature, the cross-linking reaction will
slow and probably stop before completion. Why bother to prove this again? Why not
PERSPECTIVES ON COATING DESIGN
683
start out with combinations of raw material that can react fully? Yet it is common for paint
formulators to be given a sample of a new raw material—say a cross-linker—and substitute it in a formulation they are using, only to find out that it doesn’t work. In many cases,
just looking at the formulation will tell you not to bother even running the experiment, but
rather, to change the resin being cross-linked to fit with the characteristics of the new
cross-linker.
Understanding the principles that control exterior durability (Chapter 5) permits a
better estimate of the durability of a new coating than any laboratory tests available.
Using this knowledge permits one to concentrate efforts on compositions that have a
reasonable chance of being appropriate. A person who understands recent work on corrosion protection can make a better prediction regarding corrosion protection by some
new formulation by looking at the formulation in the light of corrosion principles
(Chapter 7) than can be predicted by running salt fog chamber tests. Understanding the
scientific principles gives a head start on the formulation of a coating with improved
performance.
It is appropriate to repeat a theme that recurs throughout the book: Volume relationships, rather than weight relationships, are almost always the critical values. A good
motto for a coatings formulator is: “Think volume.”
Understanding relationships between properties and compositions is approaching the
stage at which one will be able to design the binder for a formulation from first principles.
We know what factors control Tg; we know what factors control cross-link density; we
have some good leads as to the factors that control the breadth of the glass transition
region; and we have a fairly good understanding of the relationships of these three characteristics with the performance of films. The day is not far off when it will be possible to
design a resin and cross-linker for an end use without having to rely as much as is necessary today on trial and error. But this will only be true for the formulator who studies the
advances in basic understanding and then tries to apply the principles.
4. Consider experimental approaches. The complexity of modern coatings and the
increasingly stringent performance requirements being demanded by customers coupled
with an industry wide reduction in technical support staff requires that efficiency of laboratory activities needs to improve continually to have a company remain competitive.
Except in the simplest projects, two major factors face the experimental worker. First,
most test methods used in the coatings industry are subject to considerable variation in
results, and there are many possible variables and different possible responses in many
different end uses. Second, there are almost always conflicting requirements based on
performance, process, and economic considerations. To address these challenges, the coatings developer needs to choose carefully which experiments to run, use the most efficient
laboratory approach to get the data needed, and make efficient use of all the data that are
available.
Statistical experimental design and statistical analysis of data are particularly applicable to coatings formulation, since productivity can be increased by the use of statistics.
Proper experimental design permits acquiring more information, with a higher degree of
confidence in the results, with fewer experiments. Applying design concepts before starting at the bench can facilitate choosing which experiments are the most important.
Students are still being taught in school that “thou shall never change more than one
variable at a time.” This commandment should have just one word changed, so that it
reads “thou shall always change more than one variable at a time.” There are two major
684
PERSPECTIVES ON COATING DESIGN
problems with changing only one variable at a time. First, there are so many potential
variables and levels of those variables that one could still be changing single variables
years later. Second, changing one variable at a time does not permit identification of interactions among variables. For example, the best catalyst at one temperature is not necessarily the best catalyst at another temperature. The best pigment for a coating that must
have excellent exterior durability made with one class of binder is not necessarily the
best pigment to use with a different class of binder.
The strength of experimental design is that it permits changing multiple variables simultaneously in ways that permit separation of the effects of the different variables and
identification of interactions among the variables. It can also identify variables that
many of us tend to ignore related to operator biases and those that come from running
an experiment over and over, resulting in either an improvement or a reduction in skill
at doing that experiment. Many books are available to provide background in experimental
design. Although relatively old, Ref. [3] is a useful introduction because it uses examples
from the coatings industry to illustrate the advantages and limitations of various types of
experimental designs. More extensive coverage may be found in Ref. [4]. Courses in
experimental design are available; every coatings formulator should take such a course.
Any plan for a project of significant size should include statistical experimental design.
Perhaps the most important application of experimental design is to evaluate the robustness of a formulation after the basic components are settled upon. Many formulators do all
of the development work using one batch of the raw materials or raw materials from only
one source. The production environment will always provide many sources of variability;
raw materials come with producer specifications that show ranges of product properties
that will be acceptable, production processes at both manufacturer and end user have predictable variations, and purchasing departments are known to change raw materials based
on favorable pricing. Each of these variations, either alone or synergistically, has the
ability to affect product performance negatively. Failure of the formulator to anticipate
changes that could happen after he or she finalizes the formulation can and indeed does
lead to major failures in the field, with a corresponding financial impact. Proper application of experimental design will allow for evaluation of the interplay between variables
by testing raw materials and processes at the extremes of their specifications.
Data analysis is also of critical importance. How many replicates are needed to obtain a
test result with a 90 or a 95% confidence limit? Chemists, particularly, are used to standard
deviations and do not seem to realize that 10 + 2 means that there is about a 1 : 3 chance
that the “real value” is greater than 12 or less than 8. Commonly test results subject to wide
variations are used to decide between two formulations, the differences between which are
small compared to the errors in the test results. The worrisome thing is not that poorer
materials may be accepted for further development, but that good ideas may be
discarded because of erroneous test results. It is desirable to analyze data obtained from
statistical experimental designs, since this gives the opportunity of allocating the differences in results between the different variables, the different interactions, and error.
Error is the unexplained remaining difference. If it is large, there can be two possible
explanations: The test methods may have error ranges larger than the differences being
investigated, or there may be one or more important uncontrolled variables. In either
case, before proceeding further, one should do something about the test results either by
increasing the number of replications and/or by identifying the other variable(s).
Laboratory methods and practices in many coating laboratories have not changed
substantially over the last 50 years. Of course, there have been improvements in
PERSPECTIVES ON COATING DESIGN
685
measuring methods, product specifications, and some productivity tools, but life for
a modern formulator is still a sequential process of manually preparing one or two
formulations at a time, manually preparing samples, and manually measuring properties,
all followed by manual entry of results into a lab journal or spreadsheet. Even with an
intelligent research plan, coupled with selection of reliable nonsubjective testing
methods, this process leaves a lot to be desired in terms of efficiency, reproducibility,
and accuracy.
Although there are current restraints, notably cost, combinatorial or high throughput
methods are considered because of their potential importance in the coatings industry.
These methods essentially make use of the advancements in computers, automation and
analytical methods to address the laboratory experimentation with an eye on efficiency,
reproducibility, and accuracy. In a fully integrated high throughput laboratory, the
entire process is controlled by a central computer system that tracks composition,
samples, and test results through four major steps:
1. Advanced software populates a database space (known as a library) based on formulation parameters selected by the researcher. This includes calculating weights and
volumes of the various ingredients, specifying the process conditions and times, and
determining in what order the experimental steps are conducted.
2. The information within the library is converted into instructions for a sample preparation robot that will combine all ingredients, mix, heat, and carry out the reactions
as specified. Any deviations and the actual weight volume used are stored in the
database and connected to the library description.
3. From the mixing and reacting steps, samples are transferred automatically to containers used for analytical measurements (multiple copies can prepared) and the
analysis is performed with results sent back to the central database and connected
to the library description.
4. At the end of the experiment, the database program automatically generates a
custom report that is sent to the researcher requesting the study.
Aside from the opportunity to prepare large numbers of samples, several other key
benefits arise out of the automation process described above. First, connectivity of
results from experiment to formulation composition is automatic and in a standardized
format. Second, record keeping accuracy increases dramatically as all information in
the integrated computer system is obtained and maintained digitally. Third, the reproducibility and accuracy of measurements is improved as human error and bias is taken
out. This is especially true in time dependent experiments; robotic systems can be
programmed to do a routine in exactly the same amount of time, every time.
The stages usually found in a high-throughput process are:
.
Primary screening. Large numbers of experiments are run and tested for only one or
two properties that are considered knockouts if they fail to reach a given level. For
many applications in catalyst or drug discovery fields, this primary screen may be
used on tens of thousands of samples for a given application in a single day. For a
coatings application, such a knockout might be: Does the sample thermally yellow
more than a certain amount on storage at 1008C? Samples that do not meet the criteria
686
.
.
PERSPECTIVES ON COATING DESIGN
are tagged as rejects, whereas those that exceed the criteria are tagged as “hits” and
are selected for further testing.
Secondary screening. The most promising hits from the primary screening are prepared in larger sample sizes and tested for a larger number of properties using
more advanced techniques that will allow a clear ranking of a sample potential
versus others and versus a known standard. Such tests might be run on hundreds of
samples over a period of weeks, and again, hits are selected for further evaluation.
Tertiary screening. The select number of hits from the secondary screening are passed
through traditional laboratory experimentation and testing to validate earlier results
and assign a final rank to those that go into full blown development.
The major benefit of this process of sequentially screening out guaranteed failures,
followed by screening out by key performance characteristics, is the increased probability
of choosing the best candidate to undergo full blown development.
Application of combinatorial and high throughput methodologies is rapidly being
implemented in the materials sciences and formulation science. A number of international
conferences regularly feature these methods for polymeric materials, and university
researchers are also setting up national competency centers for this purpose. An excellent
overview for material science has been prepared by Potyrailo and Amis [5]. In 2002 the
U.S. National Institute of Standards and Technology (NIST) initiated a Combinatorial
Methods Center to develop metrological standards for formulation science fields [6].
The involvement of national laboratories such as NIST is very important in terms of establishing traceable standards around which new high throughput analytical methods will be
developed.
High throughput developments for preparing coatings raw materials and specifically
coatings formulation and testing are rapidly increasing. Successful application to formulation of UV curable hard coats for plastics [7], blocked isocyanate systems [8], antifouling coatings [9], and thermostabilizers, among others, has been shown. However, reading
the reports of successful applications reveals that there is as yet no standardization in the
coatings industry and that major hurdles are yet to be overcome before this becomes a
common tool. Key concerns related to samples and to implementation are:
.
.
.
.
.
Automated sample preparation of pigmented systems has not been reported.
Edge effects in small films can have a major impact on measured properties.
Off the shelf testing solutions are not available.
Clear cost –benefit benchmarks of the technology are not documented.
The cost of implementation is far beyond that of all but the largest companies’
resources.
Universities and contract laboratories can be retained to perform high throughput
research for corporations.
The latter two points should raise a level of concern as one contemplates getting
involved. As discussed earlier, experimental design and data analysis, have clear applicability to coatings with immediate benefits that greatly exceed any investment, but that
relationship is not so clear for high throughput methods. Within the area of pharmaceuticals with billion-dollar annual profit potentials or in catalyst development for olefin polymerizations, where a few percent improvement in performance is worth tens, if not
GENERAL REFERENCES
687
hundreds, of millions of dollars in profit, it is easy to justify an expenditure of $10 million
to improve success. Very few opportunities of this size are available in the coatings
marketplace, so care must be taken not to jump in too fast and to do so only after
careful consideration.
5. Perform the work. One should think first, but one must also work, and work efficiently. One can think through what has to be done one or more days before working,
so as to mesh together as many different tasks as possible. In general, it is desirable to
identify the most difficult goal of the project and concentrate initial efforts on this
aspect of the problem. If the most difficult problem cannot be solved, there is no need
to solve the relatively easy parts of the task. Time management is one of our most critical
needs. It is fashionable to attend seminars on time management, and indeed, they may
help, but primarily, each individual needs to think through how he or she can spend his
or her time most effectively. Plans and the problem definition should be reviewed frequently; the situation may have changed, and you may end up solving a problem that is
no longer there. In complex problems involving several people from different disciplines,
planning procedures that permit monitoring progress in all aspects of the project should be
used.
6. Write a report. Most technical people detest writing reports, but in a discipline such as
coatings, there is need for continuous accumulation of data. Writing a report forces one to
review the work so far and plan the next steps. Valuable parts of reports are the experiments that did not work. Unsuccessful experiments represent a wealth of information
that can be used to solve production problems with a current product, to minimize
future work, or to meet the requirements of a coating for a different application. Unsuccessful experiments are also useful for preparing patent applications in activeness. It is
particularly critical to get the results of actual field uses and the performance obtained
into the database.
Standardization of the reporting of trials and results is very important, especially given
the increase in electronic data storage. The report on a new formulation should spell out
the reasons for inclusion of each of the components in the formulation. Consistent and
accepted nomenclature have to be adopted, and analytical results have to be standardized
and be nonsubjective. Otherwise, when others look at, or query, the formulation results
years later, they may not have any idea why a certain component is there, what it really
is, or what the results mean. This situation will complicate the reformulation that often
becomes necessary, for example, by a change in the customer’s process, and may
render the data unusable for future comparative research.
The coatings field is frustrating because there are so many variables to deal with, but
this is also what makes it fun and challenging. The primary factor controlling success is
enthusiasm to tackle and solve complex problems.
GENERAL REFERENCES
Leedy, P. D.; Ormond, J. E., Practical Research Planning and Design, 8th ed., Pearson/Merrill
Prentice Hall, Upper Saddle River, NJ, 2005.
Rekab, K.; Shaikh, M., Statistical Design of Experiments with Engineering Applications, Chapman
& Hall/CRC Press, Boca Raton, FL, 2005.
688
PERSPECTIVES ON COATING DESIGN
REFERENCES
1. Dickie, R. A., J. Coat Technol., 1992, 64(809), 61.
2. Martin, J. W.; et al., Methodologies for Predicting the Service Lives of Coatings Systems,
Federation of Societies for Coatings Technology, Blue Bell, PA, 1996.
3. Grinsfelder, H., Resin Rev., 1969, 19(4), 20; 1970, 20(1), 20; 20(2), 25; 20(3), 16.
4. Hill, H. E.; Prane, J. W., Applied Techniques in Statistics for Selected Industries, Coatings,
Paints, and Pigments, Wiley, New York, 1984.
5. Potyrailo, R. A.; Amis, E. J., Eds., High-Throughput Analysis: A Tool for Combinatorial
Materials Science, Plenum Press, New York, 2003.
6. http://polymers.msel.nist.gov/combi/index.html, accessed June 11, 2005.
7. Chisholm, B. J.; et al., in High-Throughput Analysis, Potyrailo, R. A.; Amis, E. J., Eds., Plenum
Press, New York, 2003, pp. 611 – 630.
8. Bach, H.; et al., Farbe Lack, 2002, 108(4), 30.
9. Chisholm, B. J.; et al., PMSE Prepr., 2005, 93, 906 – 907.
Index
All acronyms used in the text are included in alphabetical order in the index
AA, Acrylic acid; Adipic acid
AAEM, see Acetoacetoxyethyl methacrylate
A-B dispersants, 440, 447
ABMBN, see Azobis(2-methylbutyronitrile)
Abrasion resistance, 68, 83–85, 90, 93–94, 581, 591,
603, 628, 664, 658–659, 669, 675
of coatings for plastics, 590, 632 –633
effect of CPVC, 467–468
effect of friction, 84, 86, 602, 610
effect of nano-pigments, 432, 458, 603
of polyurethanes, 84, 88, 232, 261– 262,
624 –625, 632 –633, 661 –663, 667 –668
effect of rubber particles, 84
testing, 90, 93–94
Abrasive particle blasting, 125– 126, 659,
664– 665, 676
Absorption of IR radiation, 422
Absorption of light, 387–388, 391, 418– 419,
422, 427
Absorption of UV radiation, 97, 99– 102, 105–107,
114, 418–419, 425, 427, 575–577, 586–587
Acetaldehyde, 192
Acetoacetoxyethyl methacrylate (AAEM), 343 –345,
625, 667
Acetoacetylated resins, 131, 186, 343–345, 625
Acetylacetone (2,4-pentanedione), 241 –242, 258
Acetylene black, 428, 623, 630
Acetylene glycol alkoxylate surfactants, 508
Acid-base interactions in adsorption, 128, 439
Acid etching, see Environmental etching
Acid precursors, 223
Acid rain, 97, 154, 640. See also Environmental
etching
Acoustic testing of adhesion, 135
Acrolein, 319, 651
Acrylamide copolymers, 165–166, 229, 651
Acrylated carbamates, 583
Acrylated castor oil, 190, 646
Acrylated epoxidized linseed oil, 57, 582
Acrylated melamine-formaldehyde resins, 582
Acrylated oligomers, 57, 581–583, 588, 590–592, 626
Acrylated epoxidized soybean oil, 582
Acrylic based plastisols, 166
Acrylic esters, Tg of homopolymers, 16
Acrylic lacquers, see Lacquers, acrylic
Acrylic latexes, 178, 186–191, 529–530, 600, 638,
643, 647, 649, 654
autoxidizable, 655
hybrid alkyd/acrylic latex, 195–196, 313,
530, 643
hybrid acrylic/urethane PUDs, see Polyurethane
dispersions, hybrid acrylic/urethane
thermosetting, 193–196, 532, 603, 625,
628, 655
trialkoxysilylated, 332
Acrylic microgels, 498, 522, 599
Organic Coatings: Science and Technology, Third Edition, by Zeno W. Wicks, Jr., Frank N. Jones,
S. Peter Pappas, and Douglas A. Wicks
Copyright # 2007 John Wiley & Sons, Inc.
689
690
INDEX
Acrylic powder coatings, see Powder coatings;
Thermosetting, acrylic
Acrylic resins, 4, 16, 19–20, 23, 159– 174, 517
allyl copolymers, 164
amide-functional, 165 –166, 229
amine-substituted, 132, 540
blends with polyesters, see Polyesters, acrylic
blends
carbamate-functional, 165, 227, 329, 518,
589, 602
carboxylic acid-functional, 58, 161, 165, 167–174,
287, 290, 344, 553, 559
controlled free radical polymerization (CFRP), 23,
164– 165, 446–447, 519, 602
for electrodeposition, 537, 540, 605
epoxy-functional, 166, 277, 287, 518, 553,
559– 560, 563, 601–602, 604, 615, 672
drying oil esters from, 672
high solids, 162– 165, 519–520, 599, 602
hydroxy-functional, 160– 165, 247–248, 252,
257– 258, 261–264, 266, 330, 322, 517,
553, 601–602, 673
isocyanate-functional, 165– 166, 246
latexes, see Acrylic latexes
organocopper substituted, 669
organotin substituted, 669
phosphate-functional, 130, 146, 161, 165
silicone-modified, 86, 110, 330– 331, 518, 614
Tgs, 16
thermoplastic (TPA), 65, 159– 160, 517. See also
Lacquers, acrylic
thermosetting (TSA), 160–174, 517
functionality of, 161– 164, 519
trialkoxysilyl-functional, 8– 9, 165, 590, 601–602,
628, 673
water-reducible, 166–174, 262 –264, 266, 280,
344, 526–527, 530, 537, 597, 600, 610, 625,
627, 662
amine selection for, see Amines for waterreducible coatings
carbonate-modified, 170
pH on dilution, 167, 169, 526
silicone-modified, 331
viscosity changes on dilution, 167– 172, 526
Activation energy, 29–32
Activation energy for viscous flow, 53
Activity coefficient in solvent mixtures, 381
Acylphosphine oxides, 579, 674
Acylurea blocked isocyanates, 255
N-Acylurea cross-links, 346
Addition polymerization, see Chain-growth
polymerization
Additive color mixing, 401
Adhesion, 121 –135, 512 –515
to chalky surfaces, 185, 313, 529, 637–639
cooperative, 129
covalent bonding effects, 130 –131
effect of amine groups on, 129 –130, 132, 146,
596, 659, 665–666
effect on corrosion, see Corrosion, effect of wet
adhesion
effect of dodecylbenzenesulfonic acid, 125, 224,
602
effect of epoxy phosphates, see Epoxy phosphates
effect on formability, 80
effect of internal stress,123–124, 584–585,
588, 592
effect of pigmentation, 124, 133, 470, 514,
544–545
effect of surface tension, 122, 124–125, 131, 630
effect of viscosity, 122 –123, 146, 512
effect of wetting, 124– 125, 132, 630
fracture mechanics, 123 –124
to galvanized steel, 127, 512–513, 660, 663
to glass, 130–131, 331, 633 –634
hydrogen-bond effects,127– 128, 132, 630
intercoat, 132 –133, 513, 545, 631, 654, 660
mechanical effects, 121–123, 127, 133
to metal, 125 –131, 144, 512–515, 521, 583,
592, 660
to plastics, see Plastics, adhesion to
testing for, 133–135
of UV cure coatings, 123, 207, 590, 633
wet, 129–130, 143–146, 148–149, 153–154, 512,
654. See also Corrosion, effect of wet adhesion
Adipic dihydrazide, 196
ADSA, see Asymmetric drop shape analysis
Adsorption:
effect on adhesion, 128
on pigment surfaces, 301, 437– 444, 461, 463,
467–469, 521
Aerosol spraying, 374, 377 –378
After tack, 303
AFM, see Atomic force microscopy
AgION, 616
AIBN, see Azobisisobutyronitrile
Air-assisted airless spraying, 476, 478, 483
Air bubbles as pigments, 420, 463, 471, 642,
646 –647
Air conditioners, coatings for, 544, 605
Aircraft coatings, 94, 247, 674–676
Air entrapment, 478, 483, 509, 528
Air flow effect on solvent evaporation, 357–359,
361, 363, 366, 376, 478, 482
Air inhibition, see Oxygen, inhibition
Airless spraying, see Spraying, airless
Air pollution regulations, see Volatile organic
compounds, regulations; Hazardous air pollutants
Air spraying, 358, 476 –478, 482 –483
Aldimines, 249, 260, 345, 520, 673. See also
Isocyanates, reactions with aldimines
Alkyd coatings, 336, 471, 493, 501, 505, 512–513,
515 –517, 605, 613–615, 637–639, 643, 647,
649 –654, 660 –662, 671 –673
INDEX
Alkyd resins, 4, 306 –320, 512 –513, 515– 517
acrylate-functional, 320
amide-modified, 45, 315
chlorinated, 667
cross-link density, 77
cross-linking with isocyanates, see Isocyanates,
reaction with alkyds
cross-linking with melamine resins, see
Melamine-formaldehyde-resins,
cross-linking of alkyds
emulsions, 313, 532 –533, 651
high solids, 312 –313, 650 –651
hybrid alkyd/acrylic latexes, 195 –196, 313,
530, 643
hydrolytic stability, 311, 313, 516
long oil, 306, 308, 313, 533, 621, 651
maleated, 314
medium oil, 306, 308, 516, 672
methacrylated, 672
molecular weight distribution, 311 –312,
319 –320
nonoxidizing, 306, 314–315, 516–517, 622
oil length, 306 –307
oxidizing, 307–313, 428, 516, 605, 613,
671 –672
average functionality, 307–310
plasticizers for nitrocellulose lacquers, 314–315,
622
rosin-modified, 309, 671
saponification, 127, 148 –149, 152, 320,
512 –513, 663
short oil, 307, 315, 622
silicone-modified, 329–330, 528, 667
styrenated, 320, 512
synthesis, 315 –319
thixotropic, 45, 315
viscosity control during synthesis, 316, 318–319
water-reducible, 313–314, 526, 528, 651
Alkylethoxylates, 180
Allophanates, 233–234, 238–239, 245,
252, 258
Allophanate based polyisocyanates, 245,
252, 520
Allyl alcohol, 164, 181–182
Allyl monoethoxylate, 164
Allyl monopropoxylate, 154
Allyl-functional latexes, 195
Allyl-functional polyesters, 552
Alternating copolymers, 2
Aluminum, 137, 141–142, 608, 611, 613, 675
corrosion of, 141, 345, 675
electrodeposition on, 537, 605
passivation of, 141, 149, 151
powder coating of, 551, 555
surface treatment, 127, 144–145, 675
Aluminum trihydroxide, 623
Aluminum oxide (nano-size), 628
691
Aluminum pigments:
leafing, 147–148, 428, 609, 663, 666, 668
nonleafing, 393, 428, 505, 598 –600, 672
in metallic colors, see Metallic color coatings
Aluminum silicate pigments, 640, 647
nano-sized, 433
Amic acid catalysts, 287
Amidoamines, 282. See also Polyamides,
amine-terminated
Amine adducts of epoxy resins, 281, 283, 675
Amines as adhesion promoters, see Adhesion, effect of
amine groups on
Amines as corrosion inhibitors, 142, 152
Amine cross-linkers for epoxy resins,
see Epoxy–amine coatings
Amine reactions with isocyanates, see Isocyanates,
reaction with amines
Amine-terminated polyamides, see Polyamides,
amine-terminated
Amines in UV curing, 579– 581
Amines for water-reducible coatings, 167, 169– 173,
314, 526, 529
N-(Aminoethyl)-3-aminopropyltrimethoxysilane, 145,
331, 667
Aminolysis of silicone resins, 328
Aminolysis by solvents, 284, 350
2-Amino-2-methyl-1-propanol (AMP), 172–173, 224,
263, 645, 665
3-Aminopropyltriethoxysilane, 260, 331–332
3-Aminopropyltrimethoxysilane, 130, 264, 332
Amino resins, 4, 211 –229
Ammonium fluoride, 245
Ammonium persulfate, 179, 188
Amorphous state, 12–13
AMP, see 2-Amino-2-methyl-1-propanol
t-Amyl peracetate, 19
t-Amyl peroxide, 164
AN, Acrylonitrile
Anatase titanium dioxide, see Titanium dioxide,
anatase
Anchimeric effect, 191, 199, 204–205, 314
Anhydrides as epoxy cross-linkers, 286–287, 548,
550, 559
Anhydroglucose unit, 338, 340
Anionic electrodeposition, see Electrodeposition,
anionic
Anionic surfactants, see Surfactants, anionic
Antifoams, 508–509, 640, 645
Antifouling coatings, 431, 668–670
Antigraffiti coatings, 338
Antimony oxide, 431
Antioxidants, 103–105, 298, 485, 536. See also
Hindered amine light stabilizers
Antiskinning agents, 651
MEKO, 301
MEKO (cobalt ion complex), 301, 651
Appliance coatings, 228, 544, 554, 605–606
692
INDEX
Application defects, 315, 490–509
Application methods, 473–488
brush, see Brush application
curtain coating, 487 –488, 502, 528, 627
dipping, see Dip coating
electrodeposition, see Electrodeposition
flow coating, 488, 605
pad, 474
of powder coatings, 564–569, 604, 615
roll, see Roll coating
spray, see Spraying
AQUEVAP, 362
Aramid fiber pigments, 431
Architectural coatings, 2, 366, 369, 378 –379, 529,
636–655
Arizona exposure, 112 –113
Arrhenius equation, 16, 29–30, 280, 287, 557
Ascorbic acid, 179
Aspartate ester coreactants with isocyanates,
249–250, 674
Aspecular angles, 406
Associative thickeners, 63, 531, 642, 645, 653, 655
ASTM, 91, 152, 154, 369, 377, 641
Asymmetric drop shape analysis (ADSA), 558
Atmospheric photochemistry, 373–375
Atomic force microscopy (AFM), 109, 129
Atom transfer radical polymerization (ATRP), 23,
446, 553
ATR, see Attenuated total reflectance spectroscopy
ATRP, see Atom transfer radical polymerization
Attapulgite clay pigments, 45, 430, 432, 640, 645
Attenuated total reflectance spectroscopy (ATR), 129
Auger analysis, 129
Autocatalysis, 98, 180, 192, 298
Autodeposition coating; Autophoreticw coating,
545–546
Automotive coatings, 44–48, 451, 457, 538–542, 550,
552–553, 567, 570, 595–605, 670–674
Autoxidation, 103–104, 108, 195, 297–301, 309, 314,
320, 322, 505, 516, 530, 644, 649, 645
Autoxidizable polyurethane dispersions, see
Polyurethane dispersions, autoxidizable
Azeotropic effect, 316, 319, 362
Aziridines, 195, 345–346, 628
polymerization, 345
reaction with carboxylic acids, 345–346, 625
2-Aziridinylethyl methacrylate, 132
4,40 -Azobis(cyanovaleric) acid, 191, 194
Azobisisobutyronitrile (AIBN), 19, 21, 164, 192
Azobis(2-methylbutyronitrile (ABMBN), 164
Azo initiators, 19–21, 164, 191–192, 194
Azo pigments, 424 –426
BA, n-Butyl acrylate
BAc, n-Butyl acetate
Backscattering of infrared, 458
Bactericides, 447, 541, 640 –641, 645
Ball mills, 443, 417– 453, 469
BAPO, see Bis(2,6-dimethoxybenzoyl)-2,4,4trimethylpentylphosphine oxide
Barium borosilicate, 150, 431
Barium metaborate, 150
Barium phosphosilicate, 150, 431
Barium soap stabilizers, 109, 336
Barium sulfate (Barytes), 430, 467
Base coats, 186, 447 –448, 501, 506–507, 623, 627,
672 –673
Batch emulsion polymerization, 177, 185
Bead mills, 453–455
Beer cans, see Beverage cans
Bénard cells, 503– 504
Bentonite clay pigment, 430, 432, 497, 522
Benzil, 559
Benzimidazolone orange, 425
Benzoguanamine, 213, 225
Benzoguanamine-formaldehyde resins, 225, 605
Benzoic acid, 309 –310
Benzoin, 558 –559
Benzoin ethers, 578, 581
Benzophenone, 579–581, 631
Benzothiazoylthiosuccinic acid, 150
tridecylammonium salt, 150
zinc salt, 151
1-Benzoyl hydroxycyclohexyl phenyl
ketone, 584
Benzoyl peroxide (BPO), 19, 21, 164, 320, 340, 552
Benzyldimethylamine, 280
n-Benzyl-N,N-dimethylaniliniumsulfonates, 223
Benzylidene malonates, 101
Benzyltrimethylammonium chloride, 287, 551–552
b-Transitions, 79
Beverage cans, 84, 90, 286, 290 –292, 529, 606–610
Bicyclic orthoesters, 673
Binder, 4
Binder index, 464
Bingham bodies, 8
Biocides, 431, 541, 640–641, 645, 669– 670
2,20 -Bipyridyl, 23, 301, 313, 446
Bis-(4-aminocyclohexyl)methane, 260
2,20 -Bis-[4-(3-n-butoxy-2-glycidyloxy-1-propoxy)
phenyl]propane, 275
Bis-cumyliodonium-tetrakis(pentafluorophenyl)
borate, 585
Bis-(N,N-di-2-hydroxyethyl)adipamide, 553. See also
Hydroxyalkylamides;
Tetra(2-hydroxyalkyl)bisamide
Bis(2,6-dimethoxybenzoyl)-2,4,4-trimethylpentylphosphine oxide (BAPO), 579, 586
Bisdodecylphenylphenyliodonium
hexafluoroantimonate, 585
Bis(2,2,6,6)-tetramethylpiperidinyl)amine, 256
Bis(2-hydroxyethylvinyl) ether, 586
Bis-(methyl isobutyl ketone) ketimine of
ethylenediamine, 249
INDEX
Bismuth 2-ethylhexanoate, 258
Bismuth lactate, 540
Bismuth tri(isostearate), 540
Bisphenol A (BPA), 271–274, 286, 288, 322. See also
Epoxy resins, bisphenol A
butoxymethylolated, 286
Bisphenol F, 275
Bis(trialkoxysilyl)alkanes, 331
Bis(trimethoxysilyl)ethane (BTSE), 131, 145, 331
Bis(trimethoxysilylpropyl)amine, 145, 332
Bis-trioctyltin oxide (TOTO), 539
Bis, tris, tetra-p-substituted-tris-resorcinol-triazene,
108
Biurets, 233, 244–245, 252, 265, 520
Black box exposure, 103, 105, 113
Bleaching wood, 621
Bleeding, 424 –425
Blistering, 111, 143, 145– 146, 149– 150, 155, 264,
420, 422, 471, 514, 529, 637, 659
Block copolymers, 8, 441, 446–447, 500
Blocked amines, 260, 279 –280, 633
Blocked catalysts, 31, 242, 280, 287, 506, 559
Blocked isocyanates, 250 –259
alcohol blocked, 250, 538– 540
amine blocked, 250, 256
1-caprolactam blocked, 254–256, 259, 540, 552,
557, 559–560
catalysis of reactions, 258 –259
diethyl malonate blocked, 252, 256–257
3,5-dimethylpyrazole blocked, 255–256, 259,
532, 552
encapsulated, 258
kinetics of reactions, 251–252
organotin compound catalysis, 252, 255, 258– 259
inhibition by carboxylic acids, 259
oxime blocked, 254–255 258–259, 539–540,
552, 557
phenol blocked, 254 –255
reactions with amines, 253
reactions with polyols, 252, 257 –259
steric crowding effects on unblocking, 254
1,2,4-triazole blocked, 255 –256, 552, 603
uretdiones, 234, 250, 257–258, 553, 603
in waterborne coatings, 259
Blocked reactants, 31, 260. See also Blocked
isocyanates
Blocking, 37, 189– 190, 193, 530, 648, 654
Blooming, 182, 262, 408, 411, 648 –649, 652
Blown oils, 302
Blue pigments, 419, 426– 427
Blushing, 282–283
BMA, n-Butyl methacrylate
Bodied oils, 302, 304
Bond line readout, 502
Boron trifluoride etherate, 289
Bounce back, 473, 475
BPA, see Bisphenol A
693
BPA epoxy resins, see Epoxy resins, bisphenol A
BPO, see Benzoyl peroxide
Brabender plastometer, 445, 464
Branched polymers, see Polymers, branched
Brittle–ductile transition temperature, 76, 86, 187
Brittle failure, 74–75, 86
Bronze pigments, 429
Brookfield viscometer, 49–50
Brush application, 473–474, 492–494, 642–644,
649–650, 652
Brush drag, 474
Brush marks, 382, 393, 410 –411, 413, 474,
492–494, 650
BTSE, see Bis(trialkoxysilyl)ethane
Bubbling, 478, 483, 506–507, 527 –528, 553, 662
Buffers, 179, 182, 192, 196
Burnishing, 84, 411
1,4-Butanediol, 207, 521, 624, 632
2-Butoxyethyl acetate, 350
t-Butyl acetate, 284, 313, 350, 372, 377, 673
t-Butyl acrylate, 15
Butylated phenolic resins, 293
t-Butylbenzylamine, 256
Butyl benzyl phthalate, 672
t-Butylcarbodiimidoethyl methacrylate, 195
2-Butyl-2-ethyl-1,3-propanediol, 210, 552, 673
2-Butyl-2-ethyl-1,3-propanediol-spiroorthoester, 673
t-Butyl hydroperoxide, 179
n-Butyl isocyanate, 241
t-Butyl isocyanate, 241
n-Butyl methacrylate (BMA) oligomeric, Tg of
solutions, 17
t-Butyl peracetate, 164
t-Butyl peroctanoate, 164
t-Butyl peroxybenzoate, 195
p-t-Butylphenol, 292
Butylphosphoric acid, 223
Butyl stannoic acid, 329
t-Butyl urethanes, 280
CAB, see Cellulose acetobutyrate
Cadmium soap stabilizers, 109, 336
Calcium bicarbonate, 111, 430, 640
Calcium borosilicates, 150, 666
Calcium carbonate pigments, 111, 430, 467, 640
light scattering by, 390
Calcium driers, 301
Calcium phosphosilicates, 150
Calcium tripolyphosphate, 150
Calcium zinc phosphate, 151
California Air Resources Board (CARB), 374,
377–378
Cans:
drawn and wall ironed (DWI), 607 –609
draw-redraw (DRD), 607–609
three-piece, 607
two-piece, 607–609
694
INDEX
Can coatings, see Container coatings
Capacitance, 153
Capillary effect on coalescence, 35 –36
1-Caprolactam blocked isocyanates, see Blocked
isocyanates, 1-caprolactam
Caprolactone polyols, 207, 260, 264, 288, 521
CARB, see California Air Resources Board
Carbamate-functional acrylic resins, see Acrylic
resins, carbamate-functional
Carbamate-functional polyester resins, 209
Carbamate-functional siloxanes, 329
Carbamate salts, 282
Carbazole violet, 418
Carbonate methacrylate, 347
Carbon black, 106, 393, 419, 427 –428, 432, 445, 453,
456, 459, 504–505, 562, 586, 592
Carbon dioxide:
blushing of epoxy–amine coatings, 282
in processing powder coatings, 563
supercritical, 374, 563. See also Spraying, with
supercritical fluids
2-Carboxyethyl acrylate, 265, 583
Carboxymethylcellulose acetobutyrate, 600
Carcinogenicity, 30
Car wash tests, 85–86, 93
Casson equation, 43, 45–46
Castor oil, 303, 315
Catalysis. See also Driers; Sulfonic acid catalysis;
Tin catalysts
general acid, 217
specific acid, 217
weak acid –weak base, 279
Catalytic chain transfer polymerization
(CCTP), 23, 529
Cathodic delamination, 149, 155, 569
Cationic electrodeposition, 507, 538–540,
596–597, 605
Cationic polymerization, 288 –289, 506, 555,
585–586, 589–590
CCTP, see Catalytic chain transfer polymerization
Cedar, 638
Ceiling paints, 420, 471, 646
Cellulose acetobutyrate (CAB), 339 –340, 600, 672
Cellulose derivatives, 4, 338–340
C enamel, 609
Centrifuging pigment dispersions, 437, 455, 457
Ceramic clear coats, 603
Cetyl alcohol, 245
CFRP, see Controlled free radical polymerization
Chain-growth polymerization, 17–23, 159, 177, 299
Chain-stopped alkyds, 310
Chain transfer, 18 –21, 100, 108, 160, 162, 184,
192, 580
Chain transfer agents, 21, 164, 580
Chain transfer polymerization (CCTP), 446
Chalking, 91, 100, 112–113, 107, 185, 313, 411, 419,
529, 637 –639
Charge repulsion, 180, 437, 446 –447
CHDA, see 1,4-Cyclohexanedicarboxylic acid
CHDM, see Cyclohexanedimethanol
Chipboard, 620
Chip resistant coatings, 263, 597
Chlorinated ethylene/vinyl acetate
copolymers, 337
Chlorinated polyolefins (CPO), 132, 337, 630
Chlorinated resins, 132, 147, 335– 338, 513,
630, 666
degradation, 100, 109–110
Chlorinated rubber, 100, 132, 147, 336– 337, 666
p-Chlorobenzotrifluoride (PCBTF), 321, 350,
377, 673
p-Chlorobenzoin, 559
Chloromethyl/methylisothiazolone, 640
Chloromethyl oxirane, see Epichlorohydrin
m-Chlorophenyl isocyanate, 239
N-3-(p-Chlorophenyl)-N 0 ,N 0 -dimethylurea, 280
Chloroplatinic acid, 329
2-Chlorothioxanrhone (CTX), 587
Cholesteric pigments, 406, 429
Choline chloride, 551
Chroma, 397, 399, 704, 424–425
Chromate ion effect on corrosion, 150
Chromate pigments, 150– 151, 423–424, 426, 431,
665, 675
Chromate treatments of aluminum, 127
Chromaticity values, 398– 399
Chromatography:
gas (GC), 225, 296, 377
gel permeation, (GPC), 12, 208, 210
high performance liquid (HPLC), 219, 274, 296
hydrodynamic, 184
size exclusion (SEC), 12
Chrome green pigments, 426
Chrome yellow pigments, 150, 423– 424, 426
Chromic acid, 127, 144, 611, 631
Chromium(III) sulfate, 144
CIE color system, 398 –401
chromaticity values, 398–399
CIEDE 2000 color difference, 401
CIELAB color difference, 401, 404
CIE 1976 L a b equations, 401
color matching functions, 395
color space, 400
complementary dominant wave length, 399
dominant wave length, 399
luminance, 399
purple line, 398
purity, 399
spectrum locus, 398–399
standard human observer, 395
tristimulus values, 398, 401, 404
Citric acid, 209, 622
Clay pigments, 45, 62, 190–191, 430, 432, 437, 446,
471, 497, 522, 640, 645
INDEX
Clear coats, 80–86, 93, 101–102, 103–105, 109,
113– 114, 116– 117, 161, 164–166, 249–250,
252, 257, 329, 332–333, 501–502, 507, 553,
589– 592, 598– 604, 673– 674
Closed cup flammability testing, 369–370
CMC, see Critical micelle concentration
Coagulation of latexes, 181, 422, 451, 642
Coalescence:
of latex particles, 34 –37, 176, 189–190, 193, 262,
366 –367, 376, 465, 530 –531, 638, 642, 645,
648, 653
of powder coatings, 550, 555 –557
Coalescing agents, 36–38, 190, 465, 530, 642–649,
654– 655
cross-linkable, 648
Coal tar-epoxy coatings, 668
Cobalt driers, 77, 300–301, 313, 320, 505
cobalt bis(2-ethyl hexanoate), 258
Cobalt naphthenate, 34
Cobwebbing, 66, 159
Coconut oil, 315, 516, 622
Coefficient of friction, 84, 86, 610
Cohesive energy density, 352
Cohesive failure, 90, 134, 321
Cohesive strength, 134, 321, 630
Coil coating, 78, 93, 161, 205, 222, 229, 330, 375, 403,
485– 486, 568, 590, 606– 607, 611– 616
Cold rolled steel, 128, 138
Colligative properties, 12
Color, 382–407
change on drying, 645
face and flop, 394. See also Flop of metallic colors
interrelation with gloss, 107, 112–113, 403, 411
matching, 402–407, 563–564, 570, 598, 616
metallic, 393–394, 406, 570. See also Metallic
color coatings
metamerism, 397, 402 –403, 405 –406
pigments, 417, 421 –428
strength, 417, 421, 424–427
systems, 397 –401
Color blindness, 394
Color difference equations, 400– 401, 404
Color matching, see Color, matching
Color matching functions, 395
Color mixing, 401–402
Combinatorial experimentation, 252, 685–687
Compatibility, 355, 516
Complementary dominant wavelength, 399
Concrete, coatings for, 513, 530, 660
Concrete block coatings, 513, 647
Condensation polymerization, see Step growth
polymerization
Conductive coatings, 540, 623, 630
Conductivity of solvents. 350, 370, 480
Cone and plate viscometers, 48, 52, 318
Conical mandrel test, 92
Contact angle, 122, 124, 131
695
Container coatings, 161 –162, 323, 590, 611. See also
Beverage cans
Controlled free radical polymerization (CFRP),
16–18, 22–23, 164 –165, 446 –447, 519
Controlled shear (CSR) DMA rheometer, 88
Controlled stress (CSS) DMA rheometer, 88
Conversion coatings, 126–127, 144, 537, 540
Copolymerization, 21–22
monomer-starved conditions, 22, 160, 178, 186,
191–192
Copper, 23, 446
Copper (II) bromide, 2
Copper phthalocyanine, see Phthalocyanine blue;
Phthalocyanine green
Corex D filter, 113
Corona discharge, 132, 631. See also Spraying,
electrostatic
Core-shell latex, see Latex, core-shell
Corrosion, 137–155
barrier protection against, 142, 514, 659– 663
cathodic protection against, 141, 151–153, 668
effect of coating Tg, 146 –148, 153 –154
effect of oxygen, 138–140, 142, 146– 148
effect of pH, 139 –140, 146
effect of salts, 138– 141, 154–155, 659
effect of temperature, 140, 146–148,
153–154
effect of wet adhesion, 143–146, 148–149,
153–154, 540, 659, 666
filiform, 149
inhibition, 142
passivation, 140 –141, 149 –151, 431, 514,
667–668, 675
protection by intact coatings, 142 –148
protection by non-intact coatings, 148
protection by sacrificial metal, 142,
151–152, 668
testing, 153 –155
Corrosion inhibiting pigments, 141, 149–151, 422,
431, 514, 667–668, 675
Coverage, 636, 643, 647
CPO, see Chlorinated polyolefins
CPVC, see Critical pigment volume concentration
Cracking, 68, 80, 82–83, 92–93, 111, 113–114, 468,
637, 652
force to crack, 82– 83
Cratering, 500 –502, 516 –517, 521, 528, 557–558,
672
Crawling, 87, 499–501, 516, 521, 528, 640, 672
Crazing, 132, 632
Creep experiments, 72–73
Cresols, 379
p-Cresol, 291– 292
CRH, see Relative humidity, critical
Critical film thickness for popping, 507
Critical micelle concentration (CMC), 180,
183–184
696
INDEX
Critical pigment volume concentration (CPVC), 437,
441, 461 –468, 470 –471, 588, 627
calculation from oil absorption, 463–464
cell, 463
determination, 463–464
effect on color, 463
effect on film density, 463
effect of flocculation, 410, 437, 462, 464, 466
effect on gloss, 410, 513, 643
effect on hiding 420, 470–471, 646 –647
effect on intercoat adhesion, 133, 470, 513, 545
effect on mechanical properties, 466–468
effect of particle size distribution, 462, 464 –467
effect on permeability, 147 –148, 513 –514
effect on porosity, 470–471, 513
effect on scrub resistance, 467
effect on stain resistance, 470, 647
effect on tensile strength, 80–81, 467
factors controlling, 462–463
of latex paints, (LCPVC), 464–467
ultimate (UCPVC), 462
of zinc-rich primers, 151, 470, 664
Critical pressure, 481–482
Critical relative humidity (CRH), 362, 366, 498
Critical temperature, 481
Crockmeter, 85
Cross hatch adhesion testing, 135
Cross-link density (XLD), 161 –164, 174, 194, 200,
211, 234, 241, 260, 309, 627
calculation from stoichiometric data, 77
determination by MRI, 77
effect on brittle–ductile transition temperature, 76
effect on intercoat adhesion, 133
effect on permeability, 147, 625
effect on Tg, 16–17, 283
relation to storage modulus, 76– 78
relation to solvent swelling, 77, 94
Cross-linking, 8
effect of on coalescence of latexes, 38, 194
effect of mobility on, 34, 247, 283, 581–582
effect on Tg, 16–17, 76, 247
reaction rates of, 29– 32
Crystallinity, 13, 336– 337, 347, 552, 554
CSR, see Controlled shear DMA rheometer
CSS, see Controlled stress DMA rheometer
CTX, see 2-Chlorthioxanthone
Cupric bromide, 23
Cuprous oxide, 431, 668–669
Cure window, 82, 162, 222, 519
Curling, 592
Curtain coating, 487–488, 502, 528, 627
Cyclamic acid, 559
Cyclic esters, 200, 207 –208, 210, 308, 319, 526
Cyclic carbonates, 347
Cyclodextrin, 554
Cyclohexylcarbodiimidoethyl methacrylate, 195
1,4-Cyclohexanedicarboxylic acid (CHDA), 205, 211
1,3-Cyclohexanedimethanol, 204
1,4-Cyclohexanedimethanol (CHDM), 203–204,
210 –211, 221, 322
Cyclohexyl isocyanate, 255 –256
Cyclohexylsulfamic acid, 559
Cyclopentadiene-modified drying oils, 304
DABCO, see Diazabicyclo[2.2.2]octane
Dangling ends, 16, 76, 163
Daniel flow point method, 443–444, 450, 452– 453
DBTDL, see Dibutyltin dilaurate
DBU, see 1,8-Diazabicyclo[5.4.0]undec-7-ene
DDBSA, see Dodecylbenzenesulfonic acid
Deblocking temperature, 31, 251–255, 258–259
Defects, see Application defects
Deformation, 35, 70 –71, 76, 78–79, 85 –86, 88,
91 –93
Degassing, 558–559
Degree of polymerization, 10, 26–27, 163, 218– 220
number average, 10, 26–27
number fraction of degree of polymerization
(np), 26
weight average, 10
Dehydrated castor oil, 303, 309, 323, 516
Delamination, 123, 131, 134, 153, 540
Dendritic polymers, 8
Densification, 81–82, 87
Density:
effect on floating, 503
effect on flooding, 504
effect on sagging, 496
effect on oil absorption, 445
of solvents, 352, 356, 367–368, 370
DETA, see Diethylenetriamine
Dewetting, 125, 224
DGEBA, see Diglycidyl ether of BPA
Diacetoneacrylamide, 196
2,2-Dialkyl-2-hydroxyacetobenzophenones, 578
Diallyl phthalate, 471
Diarylide yellow pigments, 424
Diaryliodonium salts, 585
Diatomaceous earth, 430, 445
Diazabicyclo[2.2.2]octane (DABCO), 237–238,
241 –242, 246, 255, 632
1,8-Diazabicyclo[5.4.0]undec-7-ene
(DBU), 239, 344
N,N-Dibutylaminoglycolamide, 539
Dibutyl maleate, 185
Dibutyl phosphate, 242
Dibutyl phthalate, 315, 379
3,3-Di-n-butyl-3-stanna-2,4-dioxa-8-thio-spiro[5.3]
nonane, 241
Dibutyltin diacetate, 239
Dibutyltin di(benzene sulfonate), 240
Dibutyltin dichloride, 241
Dibutyltin di(2-ethylhexanoate), 240
Dibutyltin dihexafluoroarsenate, 240–241
INDEX
Dibutyltin dilaurate (DBTDL), 237, 239–243, 246,
249, 252, 258– 259, 265, 336, 632
hydrolysis, 259
Dibutyltin diperfluorobutyrate, 240–241
Dibutyltin di(2-ethylhexanoate), 240
Dibutyltin maleate, 109
Dibutyltin methoxide, 240–241
Dibutyltin oxide, 312, 539
3,30 -Dichlorobenzidene, 424
1,2-Dichloro-2-n-octyl-4-isothiazolin-3-one, 641, 669
Dicumyl peroxide, 471
DICY, see Dicyandiamide
Dicyandiamide (DICY), 280, 550, 559
Dicyclcohexylcarbodiimde, 312
Dicyclopentadiene, 304
Dicyclopentadienyloxyethyl methacrylate, 651
Diels– Alder reaction, 109, 304
Diethanolamine, 262, 322
Diethylammonium triflate, 288
Diethyl 2-bromo-2-methylmalonate, 23
Diethylene glycol, 26, 372, 379
Diethylene glycol monobutyl ether, 528, 539
Diethylene glycol monomethyl ether, 195
Diethylenetriamine (DETA), 280–282, 671
Diethyl malonate blocked isocyanate, 252, 256–257
Differential scanning calorimetry (DSC), 15, 80
Diffusion control of solvent evaporation, see
Evaporation of solvents, diffusion control
Diglycidyl 1,2-cyclohexanedicarboxylate, 277, 287
Diglycidyl ether of bisphenol A (DGEBPA), 34, 272,
274– 275
Diisobutyl ketoxime, 552
Diisopropanolamine, 342
Diisopropylethylamine, 263
Diisopropyl ketoxime, 552
Diketene, 343
Diketopyrrololpyrrole, 448
Dilatant flow, 43
Dilatometry, 15
Dilauryl thiodipropionate (LTDP), 103
Diluents, 351. See also Reactive diluents
Dimer acids, 204–205, 281, 302
2,5-Dimercapto-1,3,4-mercaptobenzimidazole, 151
N,N-Dimethoxymethyl-N 0 ,N 0 ,N 0 ,N 00 -tetramethylmelamine, 218
2,2-Dimethoxy-2-phenylacetophenone, 578, 580
Dimethyl adipate, 207, 342, 360
2-Dimethylamino-2-benzyl-1-(4-morpholinophenyl)
butan-1-one, 579
2-(N,N-Dimethylamino)ethanol (DMAE), 167 –169,
171– 173, 225, 315, 446, 506, 526, 536, 579
2-(N,N-Dimethylamino)ethyl acrylate, 447, 666
2-(N,N-Dimethylamino)ethyl methacrylate, 447, 540,
666
2-N,N-Dimethylamino-2-methylpropanol, 559
N,N-Dimethylaminopropylamine, 287, 539
Dimethylaniline, 341
697
Dimethyl azaleate, 207
Dimethylbenzylamine, 241
Dimethylbenzylpyridinium hexafluoroantimoniate,
289
Dimethyldimethoxysilane, 329
5,5-Dimethyl-1,3-dioxane-2-one, 170
Dimethyl dimethylene carbonate, 347
Dimethylformamide, 379
Dimethyl glutarate, 207, 342
2,5-Dimethyl-2,5-hexanediamine, 258
Dimethylisopropylamine, 315
1,4-Dimethylolcyclohexane, see
Cyclohexanedimethanol
2,4-Dimethylolphenol, 291
Dimethyl phthalate, 379
2,2-Dimethylolpropionic acid (DMPA), 210, 233,
261–263, 314, 322, 529, 559, 663
3,5-Dimethylpyrazole (DMP) blocked isocyanates,
255–256, 259, 532, 552
Dimethyl succinate, 201, 342, 360
Dimethyl terephthalate, 211, 312, 551
Dimethyltin diacetate (DMTDA), 239
Dinonylnaphthalene disulfonic acid (DNNDSA), 224
Dioctyl phthalate, 168, 258, 315
Dip coating, 323, 484 –485, 529, 542, 544–545
Dipentaerythritol, 303, 319
Diphenyldichlorosilane, 326
Diphenyl-2,4,6-trimethylbenzoylphosphine oxide, 579
Dipropylene glycol, 275
Dipropylene glycol dibenzoate, 646
Dipropylene glycol dimethyl ether, 37, 654
Dipropylene glycol monomethyl ether, 322
Direct impact testing, 93
Direct pull adhesion tests, 133 –134
Direct roll coating, 65, 485–486, 607, 611
Dirt, 509
Dirt retention, 108, 111, 191, 638, 641, 643, 648
Dispersing agents, see Surfactants, for stabilization of
pigment dispersions
Dispersion equipment (machinery), 443, 449– 456
Dispersion of pigments, 388, 407, 410 417, 419–421,
432–433, 435–459, 469
in aqueous media, 445 –449, 640, 652
effect on CPVC, 462
evaluation of degree, 456 –459
for high solids coatings, 441–442, 521
in organic media, 439 –445
in powder coatings, 560– 561
Dispersion polymerization, 336
Distinctness of image gloss (DOI), 407, 414
Distressing, 622
Divinylbenzene, 191, 421, 429
Divinyl ethers, 586
DMA, see Dynamic mechanical testing
DMAE, see 2-(N,N-Dimethylamino)ethanol
DMTDA, see Dimethyltin diacetate
DMP, see 3,5-Dimethylpyrazole
698
INDEX
DMPA, see Dimethylolpropionic acid
DNNDSA, see Dinonylnaphthalene disulfonic acid
Dodecanedioic acid, 166, 553, 560, 597, 604
Dodecylbenzenesulfonic acid (DDBSA), 125, 224,
602
Dodecyl mercaptan, 194
DOI, see Distinctness of image gloss
Dominant wavelength, 399
Door skins, 626
DRD, see Draw–redraw cans
Drawn and wall-ironed (DWI) cans, 607–609
Draw–redraw (DRD) cans, 607–609
Driers, 77, 103, 258, 298, 300 –301, 313, 320, 505,
651, 672
Drip marks, 498
Dry abrasive blasting, 125–126, 659,
664–665, 676
Dry hiding, 652, 670
Drying index, 297
Drying oils, 103, 298–304
in alkyd resins, 306–307, 309, 312–313, 316
blown, 302, 621
bodied oils, 302, 304
conjugated oils, 301, 303
driers for, see Driers
drying index, 297
in epoxy esters, 322–323
functionality, 297–298, 303
maleated, 303–304, 536
mechanism of drying, 297–299
nonconjugated, 297– 301
odor during drying, 300, 304, 529, 644
synthetic, 295, 302–304
in uralkyds, 321– 322
vinyl-modified, 304
Dry-to-touch, 27, 320, 512, 653
DSC, see Differential scanning calorimetry
Ductility, 76, 609
DWI, see Drawn and wall-ironed cans
Dynamic mechanical analysis (DMA), 15, 74, 76,
80, 87 –89, 94, 114, 224–225, 558
Dynamic surface tension, 491, 502, 653
Dyes, 417, 621
EA, Ethyl acrylate
Early water resistance, 314
EB curing, see Electron beam curing
ECH, see Epichlorohydrin
E-coat, see Electrodeposition coating
Edge coverage, 541, 545, 556, 570
EEW, see Epoxy equivalent weight
Efflux cup viscometers, 50 –51, 483
EGDMA, see Ethylene glycol dimethacrylate
2-EHA, 2-Ethylhexyl acrylate
2-EHMA, 2-Ethylhexyl methacrylate
2-EHOH, 2-Ethylhexyl alcohol
EIS, see Electrochemical impedance spectrometry
Elastic deformation, 70 –72
Elastically effective net work chain, 76
Elastic modulus, see Storage modulus
Electrochemical impedance spectrometry
(EIS), 151, 153
Electrodeless UV sources, 576
Electrodeposition coating, 535 –545
adhesion to, 545, 596 –597
anionic, 536 –538, 605
application, 541– 543
blocked isocyanate cross-linkers, 251, 254,
538–540
cationic, 507, 538–540, 596–597, 605
catalysis, 539 –540
conductive, 540
edge coverage, 541, 545
filling, 544
film rupture during, 507, 541 –542
pH of baths, 536, 538, 543
rate of, 541
throw power, 541 –542, 544
ultrafiltration, 542–543
variables in, 541–542
Electromotive series, 138, 140–141
Electron beam (EB) curing, 588–589, 592
Electron curtain, 589
Electron microscopy, 458. See also Scanning
electron microscopy
Electron spin resonance spectroscopy (ESR), 114–115
Electrophoresis, 535, 541
Electropolymerization, 141, 151
Electrostatic fluidized beds, 568
Electrostatic spraying, see Spraying electrostatic
a-Eleostearic acid, 296– 297, 301, 312
Elongation-at-break, 71 –72, 74–76, 83, 87, 92, 111
Elongation-at-yield, 71, 75–76
Embrittlement, 298, 301, 312 315, 516, 650
EMMAQUA, 113, 116
Empirical tests, 90–94
Emulsions in coatings, 285, 313, 329, 532–533,
624 –625, 639, 643, 651
Emulsion polymerization, 176 –186
batch process, 177, 184, 191
core-shell, see Latex, core-shell
linear power feed, 189– 190, 648
molecular weight control, 184–185, 194
particle size control, 183– 185
semicontinuous batch process, 177, 184, 192
sequential polymerization, 184, 186, 189,
192, 421, 530
surfactants in, 173 –185
Emulsions, viscosity of, 60
EN, see Extent of neutralization
Enamels, 605, 610
alkyd, 516, 649–651, 672–673
latex, 651 –655
Enamel hold out, 47, 647
INDEX
Enamines, 342, 345
Encapsulation, 33, 258
Endocrine disruption, 286, 610
Energy of vaporization, 352
Enthalpy of vaporization, 351
Entropic repulsion, 181–182, 437, 447
Entropy of mixing, 351–352, 520
Environmental etching, 109 –110, 165, 227, 232, 332,
517, 601–602
testing for, 109 –110, 601
Environmental Protection Agency (EPA), 349,
371– 373, 377– 378
Epichlorohydrin (ECH), 271–272
Epoxidized soybean oil, 276, 313, 557, 582, 610
Epoxidized linseed oil, 276, 290, 582, 610
Epoxy–acrylic graft copolymers, 290, 529, 609–610
Epoxy–amine coatings, 278 –285, 512
aminolysis by solvents, 284, 350
blushing, 282–283
catalysis, 278– 279, 282, 675
coal tar modified, 668
corrosion protection primers, 146, 150, 513–514,
660, 664–665, 668, 675
effect of hydrogen-bond acceptor solvents on
reactivity, 278, 284
for glass, 130 –131
graininess, 282
ketimine systems, 279–280
mobility effect on cure, 283, 660
package stability, 280, 284–287
pot life, 279–280, 660
scumming, 282
siloxane modified, 284
waterborne, 150, 285, 532, 660, 671, 675
pot life, 285, 532
3,4-Epoxycyclohexylmethyl-30 40 -epoxycyclohexanecarboxylate, 277, 287, 521, 586
b-(3,4-epoxycyclohexyl)ethyltriethoxysilane,
195, 332
Epoxy equivalent weight (EEW), 272
Epoxy esters, 146, 322 –324, 512, 518, 665
maleated, 323–324, 529, 536
water-reducible, 323, 526
Epoxy–phenolic coatings, 609
Epoxy phosphates, 130, 144, 146, 290, 512, 610, 659
Epoxy powder coatings, see Powder coatings, epoxy
Epoxy resins, 4, 271 –290, 512, 517–518. See also
Acrylic resins, epoxy-functional
acrylated, 554, 582, 674
advancement process, 274 –275, 322, 550
aliphatic, 277, 287, 289–290
bisphenol A (BPA), 53, 271– 275, 279, 281–286,
288, 290, 538 –539, 550 –551, 559, 609
adhesion effects, 128 –129, 275, 281, 283, 286,
323, 512
amine adducts, 281, 283, 675
brominated, 276
699
butyl-modified, 275, 283
effect on pigment dispersion stability, 439
exterior durability of, 100, 275 –276, 550
functionality, 272–273, 284
hydrogenated, 53, 276
molecular weight, 272–275, 279
siloxane modified, 284
silicone-modified, 331
silylated, acrylated, 554
viscosity, 54, 272–275, 277
bisphenol F, 275
cross-linking with amines, see Epoxy–amine
coatings
cross-linking with anhydrides, 287 –288, 550, 559
cross-linking with carboxylic acids, 116, 166, 199,
209, 211, 286–288, 518, 532, 550–551, 553,
559, 561, 567, 604
cross-linking with mercaptans, 288– 289
cross-linking with isocyanates, 247, 288
homopolymerization, 289 –290, 586, 615
hydrolysis, 272– 273, 284
nonhydrolyzable chlorine content, 273
novolac, 276, 284–286, 292
in powder coatings, see Powder coatings, epoxy
reaction with fatty acids, 322–324
reaction with hydroxy groups, 287–288
reaction with 2-hydroxyalkylamides, 343
reaction with phenols, 146, 274 –275, 286, 290,
512, 532, 609, 611
storage stability, effects of water and alcohols,
284, 369
taffy process, 273 –274
UV cure coatings, 554, 615
waterborne, see Epoxy–amine coatings, waterborne
ESR, see Electron spin resonance spectroscopy
Esters, hydrolytic stability of, 109, 200, 202–205,
209–210, 311–314, 315, 323– 324
Esterification. 24 –27, 199– 211, 306– 324, 343
Estrogen mimic, 286, 610
Ethanolamine, 262
Ethoxyethoxyethyl acrylate, 583
2-Ethoxyethyl acetate, 357, 379
Ethoxylated C12/14 alcohols, 180
Ethoxylated trimethylolpropane triacrylate, 603
Ethoxylated undecyl alcohol, 192, 646
Ethyl aspartate derivatives, 249–250, 674
Ethyl 3,3-di(t-amylperoxy)butyrate, 164
Ethylene/acrylic acid copolymer, 555
Ethylene/chlorotrifluoroethylene copolymers,
555, 559
Ethylenediamine, 261–264, 322
Ethylene glycol dimethacrylate (EGDMA), 160
Ethyleneimine, see Aziridine
Ethylene/(meth)acrylic acid copolymers, 555
Ethylene/propylene rubber, primer for, 630–631
Ethylene/vinyl acetate latex, see Vinyl
acetate/ethylene latex
700
INDEX
Ethylene/vinyl acetate copolymer, 631
2-Ethylhexyl benzoate, 654
Ethyl linoleate, 299, 301
N-Ethylmorpholine (NEM), 173 –174
Ethylsilane, 333
Ethyltriphenylphosphonium acetate, 275
Ethyltriphenylphosphonium hydroxide, 274
Evaporation of solvents, 28–29, 351, 356– 357, 474,
478, 480 –488
diffusion control, 28, 363–365, 497, 654
during spraying, see Spraying, solvent evaporation
during
effect of air flow over surface, see Air flow effect on
solvent evaporation
effect on leveling, 360, 493 –494, 496, 650
effect of ratio surface area/volume, 358 –359,
360– 361, 482, 650
effect on surface tension, 491–494, 500, 502, 505
effect of temperature, 352, 357, 359–360,
363– 365, 491–492, 502, 507
effect of vapor pressure, 350, 357 –359, 363
of mixed solvents, 361– 366,
relative, 356, 358– 361, 362, 366
of water, 358–363, 366, 525
Evaporometer, shell thin film, 359
EW, equivalent weight
Excimer lamps, 576
Excited state quenchers, 100–101, 576, 580
Experimental design, 683–687, 686
Extenders, see Inert pigments
Extensional flow, see Flow, extensional
Extensional viscosity, see Flow, extensional
Extent of neutralization (EN), 167, 173
Exterior durability, 97–118, 315, 529, 592, 643,
645, 667
chalking, see Chalking
cracking, fracture mechanical studies, 83–84
effect of pigments, 106–108, 115, 417, 419,
421– 427, 431
epoxy esters, 323
epoxy resins, 100, 275–277, 286
gloss retention, see Gloss, retention
silicone resins, 99, 110, 328 –330, 332
testing for, 111– 118
accelerated outdoor, 112 –113
accelerated weathering devices, 115–118
analysis of changes in mechanical properties,
83– 84, 114–115
black box, 103, 105, 113
Fresnel reflector, 113
Exterior house paints, 109, 111, 181, 187, 190, 329,
529, 637 –643
Extruders, 456, 560–561
Fabrication resistance, see Formability
Falling sand abrasion test, 94
Fall out from spraying, 475
Faraday cage effect, 481, 566–568, 570
FDA, see Food and Drug Administration
Ferric hydroxide, 140, 149
Ferric phosphate, 126
Ferrous hydroxide, 139, 140
Ferrous sulfate, 183
Filiform corrosion, 149
Fillers, see Inert pigments
Fillers (for wood), 588, 621
Filling, 544, 621
Film defects, see Application defects
Film formation, 27 –38. See also Latex, coalescence
by coalescence, 34– 38, 530, 542, 550, 555–557,
638, 648– 649, 653
by cross-linking, 28–34
by solvent evaporation, 28
Film splitting, 65, 474, 486–487
Film thickness:
effect on color, 387–388, 403
effect on corrosion protection, 148, 153
effect on hiding, 393, 650, 652
effect on leveling, 493–494, 557
effect on light scattering, 389–390, 586–587
effect on mechanical properties, 80, 82,
86, 91–92
effect on popping, 507 –508, 527 –528, 552–553,
600
effect on sagging, 478, 483, 496– 498
effect on solvent evaporation, 361, 363, 376
effect on UV curing, 577, 586–587
effect of viscosity on, 474, 484, 642, 652–653
effect on wrinkling, 173, 505
Fineness of grind, 458
Finishing varnish, 610
Fire retardant coatings, 276, 431, 668
Fish oil, 296
Flame spraying, 568–569
Flame treatment of plastics, 132, 631
Flammability of solvents, see Solvents, flammability
Flash points, 369–370
Flash rusting, 665 –666
Flatting pigments, 431, 624
Flavor, 606– 607
Flexibility, 74–82, 92 –93, 329
Floating, 502 –505
Flocculation, 436 –440, 443 –444, 448, 457–458,
521, 531, 653
effect on color, 388, 436, 457
effect on CPVC, see Critical pigment volume
concentration, effect of flocculation
effect on floating, 503–504
effect on flooding, 504
effect on gloss, 410, 436, 653
effect on hiding, 436
of latexes, 180
effect on settling, 437, 447, 457
effect on viscosity, 62, 437, 588
INDEX
Flocculation gradient technique, 458
Flooding, 504
Floor coatings, 83–85, 89, 93, 321 –322, 590,
626, 632–633
Flop of metallic colors, 340, 394, 570, 598, 600, 603
Florida exposure, 103, 105, 112–113, 115–116, 663
Flow. See also Viscosity
dilatant, 43
effect of pigments, see Viscosity, effect of
pigments
extensional, 64 –66, 475, 486, 647
Newtonian, see Newtonian Flow
non-Newtonian, 43, 49, 51, 56
normal force, 64
plastic, 43
shear, 41–63, 185
shear thickening, 43, 49
shear thinning, see Shear thinning
thixotropic, see Thixotropy
turbulent, 63–64
of UV cure coatings, 588
Flow coating, 485, 605
Flow window, 557
Fluidized bed application of powder coatings,
567– 568
Fluorescent light, 384 –385
Fluorinated polymers, 147, 518, 555, 569, 602,
614, 661
Fluorocarbon surfactants, 84, 494, 499, 610
Fluoroolefin vinyl ether copolymers, 337–338
Foaming, 181, 508– 509, 640, 645
Food and Drug Administration (FDA), 606, 611
Formability, 74–82, 92 –93
effect of adhesion on, 80
effect of film thickness, 80, 609
Fourier transform IR spectroscopy (FTIR), 109, 115,
129, 225, 266, 298–299
photoacoustic FTIR, 115
Fox equation, 16
Fracture mechanics, 82 –83, 123 –124
Fractional extent of reaction ( p), 26–27
Freckle exposure testing, 113
Free energy of mixing, 351 –352
Free radical polymerization, 17–23, 160, 163 –165,
168, 552, 563, 577–584, 592
oxygen inhibition, see Oxygen, inhibition
Free volume, 14–15, 28, 33 –34, 36
effect on adhesion to plastics, 131
effect on film formation, 14, 33, 283
effect on film permeability, 146
effect on latex coalescence, 36, 654
effect on powder coalescence, 557
effect on reaction rates, 33, 283
effect on solvent evaporation, 28, 363, 365, 654
effect on viscosity, 28, 52
Freeze-thaw stability, 182, 642, 655
Freezing point depression, 12
701
Fresnel reflectors, 113
Friction, effect on abrasion resistance, 84, 86,
602, 610
Frosting, 111, 419, 430, 640
FTIR, see Fourier transform spectroscopy
FT Raman analysis, 299
Fuller’s earth, see Attapulgite clay
Fumaric acid, 340, 347, 552
Fumaric unsaturated polyester, 340, 552
Functionality, 24 –25
average ( f¯n), 25, 162 –163, 170, 200–202,
206–211, 519
of drying oils, 298, 303
effect on cross-link density, 77
of epoxy esters, 323
of epoxy resins, 273, 276, 279, 284–285
of hydroxy-functional polyesters, see Polyesters,
functionality
of melamine-formaldehyde resins, 78
of monomers (F), 24, 77, 200
average (F ), 24, 200
of oxidizing alkyds, 307– 312
of polyisocyanates, 244, 258, 265
of resins, 24
of thermosetting acrylic resins, 117, 163–164, 519
Functional pigments, 431– 432
Fungicides, 431, 640 –641
Furniture coatings, 378, 410, 619 –626
UV cure, 588, 590, 625–626
Galvalume, 615
Galvanized steel, 127 –128, 142, 145, 151, 155, 186,
512–513, 615, 660, 663
Gas chromatography, 225, 296, 377
Gassing, 507
GC, see Gas chromatography
Gelation, 50, 62, 200, 259, 302, 308, 319
Gel coats, 341, 629
Gel particle effect on viscosity, 61, 498, 599
Gel permeation chromatography (GPC), 12, 208, 210
Genotoxicity, 301
Gibbs free energy equation, 351
Glass, 14
adhesion to, 130– 131, 321, 633
coatings for, 130 –131, 633 –634
microspheres, 85
reaction of surface with trialkoxysilanes, 321
Glass transition temperature ( Tg), 14–17
of acrylic polymers, 14–17, 38
of alkyds, 307
breadth of transition, 79
calculation of for copolymers, 16
changes during reaction, 33–34, 12 –19, 283,
659–660
of cross-linked polymers, 16– 17, 76, 283
determination of, 15, 75–76
effect on adhesion to plastics, 132, 632
702
INDEX
Glass transition temperature (Tg) (Continued )
effect on corrosion protection, 146– 148, 152–154
effect on CPVC of latex paints, 465
effect on cross-linking of epoxies, 36, 283,
659– 660
effect on cross-linking of urethane coatings, 247
effect on electrodeposition, 542
effect on environmental etching, 601
effect on fabrication, 78– 80
effect on film formation, 33– 38
effect on heat distortion of plastics,
132, 632–633
effect on impact resistance, 79
effect on internal stresses, 123
effect on latex coalescence, see Coalescence of
latex particles
effect of molecular weight, 15
effect on permeability, 146
effect of pigmentation, 467–468
effect on solvent loss from films, 28, 363–365
effect on UV curing, 581–582
effect on viscosity, 53, 58, 520
of epoxy esters, 323
of epoxy resins, 272, 275, 283
of latex polymers, 35–38, 185, 191, 638, 645,
647– 648, 653–654
of plastisol resins, 336
of polyester resins, 200, 202, 205–206, 208,
211, 340
of polyethylene, 15
of poly(methyl methacrylate), 15, 35
of polypropylene, 15
of powder coatings, 549, 558
of siloxanes, 15, 327
of solutions, 17
of solvents, 17
of uralkyds, 321
Gloss, 382, 407–414
bloom, 408, 411, 552
distinctness of image (DOI), 407, 413
effect of clear coats, 598 –599
effect of electrostatic spraying, 480
effect of flocculation, 410, 436, 651
effect on intercoat adhesion, 133, 513, 545,
654, 660
effect of leveling, 410–411, 494, 542, 653
effect of pigmentation, 410, 430 –431, 496, 502,
542, 588, 590
effect of wrinkling, 173, 411, 505
of electrodeposition primers, 542, 596
enamels, 504, 516, 649–655
of furniture lacquers, 410, 584, 588, 623, 626
haze, see Haze
interrelation with color, 107, 112–113, 403, 411
of latex paints, 190, 411, 531, 651– 655
luster, 408
measurement, 408–409, 412–414, 654
of powder coatings, 559– 560, 570
of primers, 133, 470, 660
rate of loss (GLR), 114– 115
retention, 115, 250, 264, 414, 596, 598–599, 609,
652, 661, 668, 672
sheen, 407 –408, 413
of UV cure coatings, see UV coatings, gloss
Gloss meters, 412 –414
GLR, see Gloss, rate of loss
Glue sizing, 621
Glutaric acid, 205
Glycerol, 24–25, 307 –308, 309, 313,
315, 320–321
decomposition during alkyd preparation, 321
Glycerol carbonate, 347
Glycidyl esters, 271
of versatic acid (neodecanoic acid), 161, 277, 673
Glycidyl ethers, 271
of butanediol, 277
of n-butyl alcohol, 277
of o-cresol, 277
of cyclohexanedimethanol, 277
of neopentyl glycol, 277
of trimethylolpropane, 277
Glycidyl methacrylate (GMA), 156, 277, 287, 289
3-Glycidyloxypropyltriethoxysilane, 329
3-Glycidyloxypropyltrimethoxysilane, 145, 331,
333, 624
Glycidyl versatate, 161, 673
Glycoluril, 213, 228 –229
Glycoluril-formaldehyde resins, 211, 228–229,
553, 558
GMA, Glycidyl methacrylate, see Glycidyl
methacrylate
Goniophotometers, 408– 409, 412–413
GPC, see Gel permeation chromatography
Graft copolymers, 8, 182, 290, 314, 320, 336, 340,
529, 609
Grain cracking, 82, 111, 637
Graininess in films of epoxy–amine coatings, 282–283
Grain raising, 624
Grain of wood, 620 –621
Gravelometer, 90
Gravity:
effect on drip marks, 498
effect on sagging, 496, 498
Gravure printing, 293, 623
Green pigments, 426 –427
Halogenated thermoplastic polymers, 335–338. See
also Chlorinated resins; Fluorinated polymers
HALS, see Hindered amine light stabilizers
Hammer finishes, 505
Hammer mills, 562
Hand-rubbed effect, 623
HAP, see Hazardous air pollutant
Hardboard, 626–629
INDEX
Hardness, 91, 83, 86, 91–92
HASE, see Hydrophobically-modified alkaliswellable emulsions
Hazardous air pollutants (HAP), 284, 15 –7, 349–350,
372, 378–379, 624
Hazardous waste disposal, 484
Haze, 408, 411, 531, 598–599, 652
HBA, see 4-Hydroxybutyl acrylate
HDI, Hexamethylene diisocyanate
HEA, Hydroxyethyl acrylate
Head-to-head polymerization, 19, 108
Head-to-tail polymerization, 19
HEAT, see Hydrophilically-modified ethoxylated
aminoplast
Heat distortion of plastics, 132, 632–633
Heat resistance of silicone resins, 328
Heat of vaporization, 508
HEC, see Hydroxyethylcellulose
Heel marking, 322, 628
Hegman gauge, 458
HEMA, Hydroxyethyl methacrylate
HEUR, see Hydrophobically-modified ethoxylated
polyurethanes
Hexa(aminomethoxymethyl)melamine
amides, 313, 651
Hexafluoroantimonic acid, 289, 585
Hexafluoroarsenic acid, 289
Hexafluorophosphoric acid, 223, 289, 585
Hexafluorozirconic acid, 127, 144
Hexahydrophthalic anhydride (HHPA), 301 –302,
311, 516
Hexamethoxymethylmelamine, (HMMM), 219, 221.
See also Melamine-formaldehyde resins, Class I
1,6-Hexamethylenediamine, 345
1,6-Hexanediol/hexamethylene diisocyanate (HDI),
allophanate derivatives. 245
1,6-Hexamethylene diisocyanate (HDI) biuret,
244– 245, 265
1,6-Hexamethylene diisocyanate (HDI) uretdione,
245, 257–258, 265, 529
Hexamethylolmelamine, 214
1,6-Hexanediol, 203, 205 221, 554
1,6-Hexanediol diacrylate, 167, 580,
582, 674
HHPA, see Hexahydrophthalic anhydride
Hiding, 80, 392 –393, 403, 470–471, 597, 599, 610,
616, 646–647, 650, 652– 654
High performance liquid chromatography (HPLC),
219, 274, 296
High volume, low pressure air (HVLP), 476 –478,
483, 674
High solids coatings, 219, 222, 518–522
acrylic, 162, 164–265, 519–520
alkyd, 312–313, 650–651
application by hot spray, 521, 623
pigment dispersions for, 441–442, 521
polyesters, 205–208, 366, 520 –521
703
primers, 514 –515
sagging, 366, 481– 482, 497–498, 521–522,
584, 599 –600
solvent effect on viscosity, 308, 312, 521
solvent evaporation from, 365, 521
surface tension, 521
urethanes, see Urethane coatings, high solids
High-speed disk (impeller) dispersers (HSD), 449–451
High throughput experimentation, 252, 685–687
Hindered amine light stabilizers (HALS), 104–107,
110, 113 –116, 232, 256, 429, 517, 539, 592,
631, 638
H12MDI, Bis(isocyanatocyclohexyl)methane
HMMM, see Hexamethoxymethylmelamine
HMMP, see
2-Hydroxy-2-methyl-1-phenylpropan-1-one
Hole coats, 627
Hookean deformation, 79
Hooke’s law, 79
Homogeneous nucleation, 183 –184
Hopeite, 126, 540
Hot spray, 481, 521, 623
House paint, see Exterior house paints
HPLC, see High performance chromatography
HPMA, Hydroxypropyl methacrylate
HSD, see High-speed (impeller) dispersers
Hue, 397, 399, 406
Huggins equation, 54
Humidity, see Relative humidity
Humidity resistance, 97, 112– 114, 123, 154
HVLP, see High volume, low pressure air spraying
Hybrid alkyd/acrylic latexes, 195 –196, 313, 530
Hybrid powder coatings, 550–551, 559
Hybrid urethane/acrylic PUDs, see Polyurethane
dispersions, hybrid acrylic/urethane
Hydrazine, 261, 263, 322
Hydrazine-ketone cross-linking, 263
Hydroabietyl alcohol, 557
Hydroxyacetophenone, 674
Hydroblasting, 126, 659, 667
Hydrocarbon resins, 668
Hydrochloric acid, reaction with epoxides, 289
Hydrodynamic chromatography, 184
Hydrodynamic volume, 56–59, 368
Hydrofluoric acid, 546
Hydrogen abstraction, 20 –21, 98, 104, 179, 182, 184,
290, 14– 3-5, 320, 336, 341, 372, 580
Hydrogenated bisphenol A, 225, 228
Hydrogenated BPA epoxy resins, 53, 276
Hydrogen bonding effects:
on abrasion, 231 –232
on adhesion, 128–129, 132, 630
on reaction rates of epoxy–amine coatings, 278,
284, 369
on reaction rates of isocyanates, see Isocyanates,
reactions with alcohols; solvent effects on
on solvent evaporation, 351–357
704
INDEX
Hydrogen bonding effects (Continued )
on solubility, 351–356
on viscosity, see Viscosity, effect of hydrogen
bonding
Hydrogen-bonding index, 354
Hydrogen peroxide, 546, 621
Hydrogen sulfide, 431, 609
Hydrolytic degradation, 97, 109–110
Hydrolytic stability:
of alkyds, see Alkyd resins, hydrolytic stability
of polycarbonate polyols, 258, 261, 347
of polyesters, see Esters, hydrolytic stability of
of silicone resins, 110, 331
of ureas, 109, 260
of urethanes, 109, 240 –241, 260
of vinyl acetate, 192
Hydroperoxides, 98–99, 103–104, 110, 115, 179,
298–300, 302, 313
Hydrophilically-modified polyisocyanates, 264–266,
529, 625, 628
Hydrophilically-modified alkali-swellable emulsions
(HASE), 63
Hydrophilically-modified ethoxylated aminoplast
(HEAT), 63
Hydrophilically-modified ethoxylated polyurethanes
(HEUR), 63, 646
2-Hydroxyalkylamides, 23, 165, 200, 204, 211,
342–343, 553, 559. See also
Bis-(N,N)-(di-2-hydroxyethyl)adipamide;
Tetra(2-hydroxyalkyl)bisamide
esterification, 342 –343
reaction with epoxies, 343
2-Hydroxybenzophenones, 101–104
4-Hydroxybutylvinyl ether, 554
Hydroxyethylcellulose (HEC), 63, 65, 182,
633, 640
2-Hydroxyethyloxazolidone, 260
Hydroxy free radicals, 98, 107, 179, 373–374
Hydroxy-functional poly(vinylidene fluoride), 147
2-Hydroxy-2-methyl-1-phenylpropan-1-one (HMMP),
578, 584, 586
2(2-Hydroxyphenyl)-2H-benztriazoles, 101
2(2-Hydroxyphenyl)-4,6-phenyl-1,3,5-triazine, 101
o-Hydroxyphenyl-s-triazine, 106
Hydroxypivalic acid, 552
5-Hydroxypoly[methyleneoxy]methyl-1-aza-3,7dioxabicyclo[3.3.0]octane, 645
Hydroxypropyl carbamate, 165, 602
2-Hydroxy-2-sulfinatoacetic acid disodium salt, 179
Hydroxy-terminated polyurethanes, 266 –267
Hyperdispersants, 440–441, 521
IBAC, Isobutyl acetate
IBMAA, see N-Isobutoxymethylacrylamide
ICI viscosity, 643
iep, see Isoelectric point
Illuminant, 383–384, 398, 400
Imidazole, 559
Imidazoline formation in polyamides, 282
Iminooxadiazenediones, see Isocyanates,
unsymmetrical trimers
Impact resistance, 79, 90, 93, 312, 552
Impedance, 153
Incineration, 375, 613
Indentation tests, 85, 87–88, 91–92
Inert pigments, 152, 430 –431, 437, 459, 623,
640, 645–647
Infrared backscattering, 458
Infrared heating, 112, 564, 575, 586
Infrared reflecting pigments, 422, 432, 616
Infrared (IR) spectroscopy
Fourier transform infrared (FTIR), see Fourier
transform infrared spectroscopy
photoacoustic FTIR, 115
Infrared thermography, 155
Inhibition of corrosion, 142, 149. See also
Passivation
Inhibitor, 32
Initiation, 18 –22, 160, 163, 178 –179, 183– 184
Initiators, 19, 21, 160, 164, 167, 178–179, 183–185,
576 –580, 585
Inks, see Printing inks
In-mold coating of plastics, 341, 628–629, 632
Instron tester, 87, 134
Intercoat adhesion, see Adhesion, intercoat
Interference pigments, 387, 394, 429
Interior flat wall paints, 191, 193, 378,
529, 643–649
Internal stresses, 80, 123–124, 138, 143,
154, 592
Intrinsic viscosity, 54–56
Iodine value, 298
Iodonium salts, 585
Ion-exchange latexes, 638
IPA, Isophthalic acid
IPDI, Isophorone diisocyanate
Iron, 137– 141, 537, 546
Iron blue, 426
Iron oxide, 445, 471
magnetic, 439
micaceous, 147, 514, 663
red, 425
transparent, 107, 425
yellow, 423, 425
transparent, 423, 447
Iron phosphide, 152
Isoascorbic acid, 179
Isobornyl methacrylate, 53, 164
N-(Isobutoxymethyl)methacrylamide
(IBMAA), 166, 229
Isobutyl isocyanate, 248
Isobutyl methyl urea, 248
Isobutyraldehyde, aldimine with bis(4-aminocyclohexyl)methane, 260
INDEX
Isocyanates. See also Urethane coatings
allophanate based polyisocyanates, 245, 252, 520
blocked, see Blocked isocyanates
catalysis, 237– 242, 246, 252, 255, 258–259
dimers (uretdiones), see Uretdiones
hydrophilically-modified, 264–266, 529, 625, 628
reaction with acetoacetylated resins, 344
reactions with alcohols, 232 –243, 245, 247–248
catalysis, 237 –242
kinetics, 234–241
solvent effects on, 238, 240, 248, 369
reaction with aldimines, 249, 260, 520, 624, 673
reaction with alkyds, 247, 315, 517, 672
reaction with amines, 233, 249, 260– 264, 661
reaction with aspartate esters, 249 –250, 674
reaction with carboxylic acids, 233, 261, 265
reaction with hindered amines, 249 –250, 661
reaction with ketimines, 248– 249, 260
reaction with nitrocellulose, 247, 622
reaction with phenols, 232
reaction with ureas, 233
reaction with urethanes, 233, 252
reaction with water, 233, 241, 244, 247, 249, 253,
258 –261, 264 –266
toxic hazards, see Toxic hazards, isocyanates
trimers, see Isocyanurates
unsymmetrical trimers, 245, 265 –266, 520,
529, 673
Isocyanate-functional acrylic resins, 165–166, 246
4-Isocyanatomethyl-1,8-diisocyanatononane, 266
Isoelectric point (iep), 446, 448
Isoeugenol, 485
Isoindoline yellow pigment, 425
Isophorone, 350, 360, 372, 379
Isophoronediamine, 260
Isophorone diisocyanate (IPDI) allophanate
derivatives, 245
Isophorone diisocyanate (IPDI) uretdione,
257– 258, 553
Isopropenyldimethylbenzyl isocyanate (TMI),
165– 166, 195, 244, 246, 254
acrylic copolymers, 165, 246
oxime blocked, see Blocked isocyanates, oxime
blocked
2-Isopropylthioxanthone, 579–580, 586
Jar mills, 453
Jetness, 427
2K coatings, see Epoxy–amine coatings; Urethane
coatings, 2K
2K urea coatings, 249–250, 661
Kaolin clay, 446, 471, 505, 645
Karl Fischer water analysis, 377
Ketimines:
use with acetoacetates, 345
use with polyacrylates, 342, 345
705
use with epoxies, 279 –280, 284
use with isocyanates, 248 –249, 260
KHN, see Knoop hardness numbers
Kinematic viscosity, 42, 47, 50– 51
Kinetics of cross-linking, 29–33, 251 –252, 581
N-(3-keto-2,2-dimethylbutyl)acrylamide, 196
Knoop hardness numbers (KHN), 91, 188
Koenig pendulum test, 92
Krebs units (KU), 51, 639
KTPP, see Potassium tripolyphosphate
KU, see Krebs units
Kubelka–Munk color equations, 391–392, 405
Kuhn–Mark–Houwink–Sakurada equation, 55
L a b color difference equations, 401
Lacquers:
acrylic, 12, 1 –8, 315, 604, 672
alkyd plasticizers for, 315, 622, 671
emulsions of, 532, 624
gloss control, 431, 623
nitrocellulose, 247, 315, 351, 358, 364,
621–626, 671
solvent blends for, 351, 358
Lake, 417
LAMMA, see Laser microprobe mass analysis
Lamp black, 427, 504
Lanthanum nitrate, 145
Lapping, 642, 649, 655
Laser light scattering, 458
Laser microprobe mass analysis (LAMMA), 501
Latent acids, 223
Latent solvents, 351
Latex, 34–38, 176–196
acetoacetate-functional, 195, 343
acrylic, see Acrylic latexes
alkyd/acrylic hybrid, 195 –196, 313, 530
amine-substituted, 666
carboxylic acid-functional, 332, 346
coagulation, 181, 422, 451, 642
coalescence, see Coalescence, of latexes
core-shell, 186, 189 –190, 192, 195–196,
346, 421, 654
cross-linking with polyaziridines, 195,
345–346
cross-linking with polycarbodiimides, 195, 346
film formation from, see Coalescence, of latex
particles
hydroxy-functional, 184, 186
inverse core-shell, 186, 192
lobed particles, 185, 641
molecular weight, 176, 183–185, 192–195
nano clay composite, 190–191, 648
particle size, see Particle size of latexes
phosphate-functional, 178
pigments, 420 –421, 430, 642, 647
polystyrene, 430, 647
seed latex, 178, 182, 184, 192
706
INDEX
Latex (Continued )
sequential polymerization of, see Emulsion
polymerization; sequential polymerization
stabilization of, 180 –182
styrene/acrylic, see Styrene/acrylic latexes
styrene/butadiene, 178, 647
thermosetting, 38, 193 –196, 332, 343, 346,
530– 532, 603, 624, 648, 655
trialkoxysilane-functional, 196, 332, 530
urethane/acrylic hybrids, 186, 262 –263, 322,
530– 531
vinyl acetate copolymer, see Vinyl acetate
copolymer latexes
vinylidene chloride copolymer, 178, 190, 546, 666
viscosity of, 60, 62–63, 176, 188– 189
wet adhesion, 178, 654, 666
Latex-based coatings, 529–532
auto:
OEM, 603
refinish, 671, 673
coil coatings, 614
container coatings, 610
CPVC (LCPVC) of, 464–466
dirt retention by, see Dirt retention
exterior house paint, see Exterior house paint
floating, 504 –505
glass transition temperature, see Glass transition
temperature, of latexes
gloss of, see Gloss, of latex paints
gloss enamels, 190, 447, 651 –655
interior flat wall paint, see Interior flat wall paint
leveling, 411, 496, 531, 637–638, 652–653
maintenance, 666–667
marine, 668
permeability, 108, 529, 637, 666
pigment dispersion for, 640
popping, 508
primers, 513
sagging, 63, 638, 642, 644, 653
saponification resistance, 647, 666
solvent evaporation from, 366–367, 654
spattering, 475, 647, 653
viscosity, see Viscosity of latex paints
water spotting, 181
Latex critical pigment volume concentration
(LCPVC), 464–466
LCPVC, see Latex critical pigment volume
concentration
LTDP, see Dilauryl thiodipropionate
Lead carbonate, 420
Lead chromate, 150, 423
Lead driers, 300–301, 505
Lead molybdate, 423
Lead octanoate, 241
Lead oxide, 423
Lead, red, 149, 431
Lead sulfate, 423
Lead toxicity, see Toxic hazards, lead
Lecithin, 330
Leveling, 23, 360, 393 474, 482, 486–487, 492–495,
529, 545, 555 –558, 570, 596, 598–599, 601, 614.
See also Latex-based coatings, leveling
Lifting, 320, 512, 672
Light, 383–385
absorption, 387–388, 391, 418–419
interaction with observers, 382, 394–395
reflection, 385 –387, 391 –394, 398, 407–408, 411,
413–414
refraction, 385 –387
scattering, see Scattering of light
standard sources, 383 –384, 398, 400
transmission, 386, 402
wavelength sensitivity of eye, 383, 385
Light fastness, 422, 424 –425
Lignin, 626
Limiting viscosity number, 54
Linear power feed emulsion polymerization, 189
Linoleic acid, 296–297, 301, 309, 312, 316
Linolenic acid, 296– 297, 309. 312
Linseed oil, 292, 297 –300, 302 –304, 309, 319, 321,
444 –445, 463, 516, 621, 651
epoxidized, see Epoxidized linseed oil
fatty acids, 303, 322–323, 663
maleated, 304, 536
Liquid crystals, 78
Litharge, 316
Lithium hydroxide, 316
Lithium ricinoleate, 316
Lithium silicate, 152, 333, 664
Lithopone, 420
Living radical polymerization, 20, 22
Loose end, 16, 76, 163
Loss modulus, 73 –74, 76, 79, 92
Loss tangent, 73–74. 79 –80, 88
Low temperature loss peaks, 79
Low volume, low pressure air spraying
(LVLP), 478, 521
Luan, 626
Luminance, 399
Luster, 408
LVLP, see Low volume, low pressure air spraying
(LVLP)
MA, Methyl acrylate
MAA, Methacrylic acid
Magnesium, 137, 142, 668
Magnesium silicate, 430, 446
MAK, Methyl n-amyl ketone
Maintenance paints, 321, 378, 658–667
MALDI-ToF-MS, see Matrix-assisted laser desorption
ionization time-of-flight mass spectrometry
Maleated alkyds, 314
Maleated drying oils, 303–304
Maleated epoxy esters, 323– 324, 529, 536
INDEX
Maleated rosin, 622
Maleated uralkyds, 322, 663
Maleic acid, 335
Maleic anhydride, 166, 303 –304, 314, 322–323,
337, 340–341, 447, 630, 663
Maleimides, 580
Mandrel bend test, 92
Manganese driers, 300–301
Mannich bases, 282–283
Marangoni effect, 492–493, 500, 504. See also
Surface tension differential flow
Marine coatings, 283, 292, 667– 670
Mar resistance, 81, 85 –86, 165– 166, 432, 601–603
effect of nano-pigments, 81, 432, 469, 602 –603,
626
Mass spectroscopy, 12, 299
Matrix-assisted laser desorption ionization time-offlight mass spectrometry (MALDI-ToF-MS), 204
Matte, 407
MDI, Bis(isocyanatophenyl)methane
Mechanical effects on adhesion, 122, 127, 132
Mechanical properties of films, 68–94, 152, 224–225,
319, 466–469
measurement of, 86–89. See also Dynamic
mechanical analysis (DMA)
Media mills, 453 –455
MEK, Methyl ethyl ketone
MEK double rub test, 94, 234, 253
MEKO, Methyl ethyl ketone oxime
MEKO (cobalt ion complex), 301
Melamine, 214, 262. See also Melamine–
formaldehyde resins
Melamine–formaldehyde resins (MF resins), 214–217
acrylated, 582
acrylamide derivatives, 582, 651
catalysis, 214– 215, 217– 215, 221– 224, 516, 604
class I, 27, 78, 164–165, 173, 218– 223, 225–227,
290, 517, 527, 536
class II, 173, 218–222, 226–227, 516 –517
co-condensation, 222, 224 –225, 517
cross-linking of acetoacetylated resins, 344
cross-linking of acrylics, 76, 83, 85 –86, 161,
164 –165, 173, 220, 226 –227, 517, 599–602,
605, 610
cross-linking of alkyds, 220, 306, 315, 516, 605,
613 –615
cross-linking of carbamate-functional resins, 84,
110, 164, 220, 227, 518, 602
cross-linking of epoxy esters, 323
cross-linking of epoxy resins, 288, 290, 613
cross-linking of fluorinated resins, 338
cross-linking of hydroxy-functional urethanes, 227,
266 –267, 518, 601
cross-linking of latexes, 194
cross-linking of polyesters, 76, 83, 220–221, 516,
521, 610, 614
cross-linking of silicone-modified resins, 331, 614
707
cross-linking of hybrid acrylic/urethane
dispersions, 262, 600
environmental etch resistance, 83–85, 109–110,
165, 227, 332, 517, 601
functionality, 80
hydrolysis, 109– 110, 221, 223, 229, 601
in dual cure UV coatings, 589
in liquid crystal coatings, 80
mar resistance, see Mar resistance
mechanism of cross-linking, 224 –225
reaction with carboxylic acid groups, 214, 227
reaction with malonic ester blocked isocyanates,
227, 257
reaction with phenols, 214, 220 –221
reaction with urethanes, see Urethanes, reaction
with MF resins
self-condensation reactions, 79, 222, 224, 226,
519–520
stabilization, 8–15, 226
steric hindrance effects, 79, 221, 227
synthesis, 214–218
toluene sulfonamide-modified, 553
transetherification with solvents, 225
with water-reducible acrylics, 173, 527, 600,
610, 627
Mer, 7
Mercaptan chain-transfer agents, 21, 164, 194– 195
Mercaptan cross-linking of epoxy resins, 288–289
2-Mercaptobenzimidazole, 151
2-Mercaptoethanol, 164
Mercury vapor UV lamps, 575
Mesogens, 78
Metals. See also Aluminum; Steel
adhesion to, see Adhesion to metals
cleaning, 125 –126, 144, 611, 659, 667, 671
surface treatment, see Conversion coatings
Metal decorating, see Container coatings
Metallic color coatings, 159, 162, 340, 393– 394, 406,
425–426, 428–429, 570, 598– 600, 603– 604,
672–673
Metal marking, 86, 614
Metamerism, 397, 402 –403, 405 –406
Methacrylamide, 229
Methacrylamidoethylethyleneurea, 179, 654, 666
Methacrylated alkyds, 672
Methacrylated oligomers, 342
N-(Methacryloxyethyl)ethyleneurea, 654
3-Methacryloxypropyltriisobutoxysilane, 332
3-Methacryloxypropyltrimethoxysilane, 332
Methoxypolypropylene glycol monomethacrylate, 446
1-Methoxy-2-propyl acetate, 350, 356–357
1-Methoxy-2-propyl carbamate, 329
Methyl acetoacetate, 343
Methyl n-amyl ketone (MAK), chain transfer to TSA
resins, 160
Methyl a-bromopropionate, 23
Methyl carbamate, 205
708
INDEX
N-Methyl-N,N-diethanolamine, 580
Methyl p-(dimethylamino)benzoate, 579
2,20 -Methylenebis(4-methyl-6-t-butylphenol), 103
Methylene chloride, 660, 676
Methyl ethyl ketone peroxide, 340–341
Methyl ethyl ketone ketimines, 279
Methyl ethyl ketone oxime (MEKO) blocked
isocyanates, see Blocked isocyanates, oxime
blocked
MEKO antiskinning, 301, 651
Methyl 2-hydroxypropyl ether, 446
2-Methylimidazole, 280, 550, 553
Methyl isobutyl ketone (MIBK) ketimines, 260
Methyl linoleate, 300, 304
Methyl linolenate, 300
N-Methylmaleimide, 580
N-Methylmorpholine, 528
Methyl oleate, 300
Methylphenyldimethoxysilane, 329
3-Methylphthalic anhydride, 287
N-Methylpyrrolidone, (NMP) 262–263
Methyl silicone resins, 328 –329, 643
Methyltoluene sulfonimide, 553
Methyltrichlorosilane, 326
Methyltrimethoxysilane, 127, 152
MF, see Melamine-formaldehyde resins
MFFT, see Minimum film formation temperature
MGEBPA, see Monoglycidyl ether of BPA
MIBK, Methyl isobutyl ketone
Mica, 147, 430, 514, 663, 666
Micaceous iron oxide, 147, 514, 663
Micelles, 180, 183
Michael reactions:
with acetoacetylated resins, 344
with acrylate double bonds, 345– 346
Microgels in alkyds, 319
Microgels for sag control, 61, 498, 599
Mildew, 111, 641
Mill bases, 435 –436, 440, 442–445, 451–456
Miller–Macosko equations, 77
Minimum film formation temperature (MFFT),
35–38, 187–190, 262, 530, 649, 655
Minimum incremental ozone reactivity value (MIR),
374, 378
MIR, see Minimum incremental ozone reactivity value
Mist coat, 664
Misting, 65
MMA, Methyl methacrylate
Mobility effect on film formation, 33 –34, 247, 283,
660
Modulus:
elastic modulus, see Modulus; Storage
loss, 73–74, 76, 79, 92
shear, 74, 76
storage, 73–74, 78, 80– 81, 86, 468
tensile, 74, 76, 87, 89, 241
Moisture-curable epoxy–amine coatings, 279
Moisture-curable polyurethane coatings, 250,
259 –261, 506 –507, 624, 628, 632,
660 –662, 664
Moisture-curable trialkoxysilyl coatings, 165, 260,
332, 602, 673
Moisture vapor permeability, see Water, permeability
of films
Molar absorptivity, 387, 576
Mold releases, 131, 499, 630
Molecular sieves, 253
Molecular weight (MW), 7, 9 –13, 16, 19–22
acrylic lacquer polymers, 12
determination, 12, 54
effect on compatibility, 520
effect on intrinsic viscosity, 55
effect on relative viscosity, 56, 58
effect on solubility, 352, 355, 368
effect on solution viscosity, 54–58, 62. 65
effect on stability of pigment dispersions,
439–441, 447
effect on Tg, 15, 267
effect on viscosity of resin solutions, 58–59
effect on viscosity of water-reducible
coatings, 168
in emulsion polymerization, see Latex, molecular
weight
epoxy resins, 272– 275, 277, 279, 290
factors controlling in free radical polymerization,
19–21
number average, 184 –185, 193
of polyesters, 24– 27, 200–203, 206–207, 210
in solution chain-growth polymerization, 24– 27
weight average, 9 –12
Molecular weight distribution, 9–12, 15, 19–21,
23, 183, 192, 200, 206 –207, 311, 312,
446, 519
Molybdate orange, 423, 425
Molybdate passivating pigments, 150, 666
Monoazo yellow pigments, 424
Monobutyltin dioxide, 26, 319
Monocoats, 598– 599, 603
Monoglycerides, 315– 316, 321
Monoglyceride process, 316
Monoglycidyl ether of BPA (MGEBPA), 272
Monomer-starved copolymerization, see
Copolymerization, monomer-starved
Monomethylether of diethylene glycol, 195
Monomethylether of polyethylene glycol, 195, 264
Monomethylether of polyethylene glycol
methacrylate, 446
Mooney equation, 60, 177, 441–442
Morpholine, 314, 652
Morphology, 13 –17, 183 –184
of water-reducible resin coatings, 167–171
MRI, see Nuclear magnetic imaging
Munsell color system, 397–398, 400
MW, see Molecular weight
INDEX
Nano(aluminum oxide), 628
Nano(aluminum silicate), 433, 469
Nano(ceramic spherical) pigment, 603
Nano-clay pigment, 190 –191
Nano(zinc zirconium silicate), 106
Nano-pigments, 80, 106, 190–191, 432–433, 458,
468– 469, 602– 603, 626, 628, 648
effect on abrasion resistance, 432, 469, 628
effect on mechanical properties, 80, 190,
469, 603
effect on mar resistance, see Mar resistance, effect
of nano-pigments
UV absorbers, 106
Nanomechanical analyzer, 85
Nano scratch test, 85
1,5-Naphthalene diisocyanate, blocked with
imidazole, 280
1-Naphthalene sulfonyl chloride, 446
Naphthol red pigments, 425 –426
NEM, see N-ethylmorpholine; Nitroxide mediated
polymerization
Neodymium driers, 313
Nepheline, 640
Newman’s rule, 206
Newtonian flow, 42–43, 47, 50–51, 56, 79, 81, 443,
450, 457, 483, 486
NGR, see Nongrain raising stains
Nickel azo yellow pigment, 425
Nitric acid, 215, 333, 339
Nitroalkanes, see Nitroparaffins
Nitrocellulose, 338–339, 439, 619. See also Lacquers,
nitrocellulose
alkyd plasticizers for, 315, 623, 671
cross-linking with isocyanates, 247, 622
Nitroethane, 532
Nitrogen oxides (NOx), 373– 374
Nitroisophthalic acid, zinc salts, 151
Nitroparaffins, 350, 370
amine salts with epoxies, 285, 532
conductivity, 350, 370, 480
2-Nitropropane, 350, 379
Nitroxide compounds, 22
Nitroxide mediated polymerization (NEM), 22
Nitroxyl derivatives of 2,2,6,6-tetramethylpiperidines,
105, 113
NMP, N-Methylpyrrolidone
NMR, see Nuclear magnetic resonance
Nonafluorohexylmethylsiloxane side chains, 670
Nonconjugated drying oils, 297–304
Non-drying oils, 297
Non-Newtonian flow, see Shear thinning; Thixotropy
Non-oxidizing alkyds, see Alkyds, non-oxidizing
p-Nonylphenol, 292
Nonylphenylethoxylates, 35, 179–180, 182, 185
sulfates, ammonium salts, 186
Normal force flow, 64
Novolac epoxy resins, 276, 284 –285, 660, 675
709
Novolac phenolic resins, 276, 292–293, 660
NPG, Neopentyl glycol
Nuclear magnetic resonance (NMR) or
(MRI), 77, 299
Number average molecular weight, 164, 184,185
NVP, N-Vinylpyrrolidone
NVV, Nonvolatile volume
NVW, Nonvolatile weight
Nylon, 555, 630
OA, see Oil absorption
2,7-Octadienyl maleate-fumarate, 313
2-(2,7-Octadienyloxy)succinate, 313
n-Octane, 366
Octyl acrylate polymers, 495, 501, 508, 557
n-Octyl alcohol, 125
Odor, 193, 349–350, 529, 644, 650
OEM coatings, see Product coatings
Oils, see Drying oils; Semi-drying oils
Oil absorption (OA), 444–445
calculation of CPVC from, 463–464
determination of, 444–445
Oil-free polyesters, 199
Oil length, 306 –309, 312, 316, 321
Oleic acid, 296 –297
Oligomers, 7–9, 17, 43, 53, 56, 58, 342, 368, 376
Onium salts, 585, 589
Open cup flammability testing, 369
Opposed jet mills, 562
Optical fibers, 590–591, 634
Orange peel, 494 –495, 502, 557, 570, 601, 601,
605, 624
Orchard equation, 492–494, 557
Organosols, 336, 609, 614
Organotin biocides, 669
Organotin catalysts, 26, 200, 203, 207, 237–242, 246,
252, 255, 258–259, 675. See also Dibutyltin
dilaurate
Orientation of aluminum flakes in metallics, see Flop
of metallic colors
Ortho-phosphoric acid, 144
Orthosilicates, 322–333
Oscillating plate rheometer, 64, 225
Osmotic pressure, 143, 146, 149, 154–155, 514
Osmotic repulsion, 181, 437
Outdoor durability, see Exterior durability
Out-of-dust time, 672
Oven sagging, 498, 522, 544
Overbaking, 316
Overspray, 475, 479 –480, 482, 484, 567, 531
Overspray disposal, 482, 484
Oxanilides, 105
Oxazolidines in urethane coatings, 260
Oxidizing alkyds, see Alkyds, oxidizing
Oxime blocked isocyanates, see Blocked isocyanates,
oxime blocked
Oxiranes, 271
710
INDEX
Oxygen:
in corrosion, 138–143, 146–148
inhibition, 32, 341, 506, 580– 581, 583, 589
permeability, 144, 146– 149, 513, 659–661,
663, 666
quenching of photoinitiators, 580
Ozone as air pollutant, 372–375, 377–378
Ozone formation in the atmosphere, 373
PA, Phthalic anhydride
PAA, Polyacrylic acid
Package stability, 29–33. See also Pot life
of blocked isocyanate coatings, 256
of epoxy coatings, 279– 280, 284– 287
of MF resin coatings, 226, 369
by nonkinetic means, 32– 33, 226, 257
of phenolic resins, 261, 290
of powder coatings, 23, 165, 558
of self cross-linking resins, 287
thermosetting latexes, 195–196, 530 –532
of UV cure coatings, 578– 579
Packing factor in dispersed phase systems, 36, 60, 169,
176, 185, 462, 641
Pad application, 474
Padding stains, 621
Paddle viscometers, 51 –52
PAG, see Photoacid generators
Paints, 2, 473 –474, 636 –655
alkyd, see Alkyd coatings
ceiling, 420, 471, 646
exterior house, see Exterior house paints
gloss, 531, 649 –655
interior flat wall, see Interior flat wall paint
latex, see Latex-based coatings
Palmitic acid, 296 –297
Panel finishing, 528, 626–628
Particle board, 590, 623
Particle size distribution effect on CPVC, 462, 466
Particle size distribution effect on packing factor, 36,
61, 185, 462
Particle size distribution of powder coatings, 557, 562,
564, 567
Particle size of latexes, 36, 183–185, 465, 636, 646,
648, 652 –654
Passivating pigments, see Corrosion inhibiting
pigments
Passivation, see Corrosion, passivation
PCBTF, see p-Chlorobenzotrifluoride
PD, see Polydispersity
PDI, see Polydispersity index
PE, Pentaerythritol
Pearlescent pigments, 406, 429
Pebble mills, 451–452
Pelargonic acid, 315, 551
Pencil hardness tests, 92, 94
Pendulum hardness tests, 92
Pen knife adhesion testing, 133
Pentaerythritol triacrylate, 582
propyleneimine adduct, 195
Pentaerythritol tetrakis-3-mercaptopropionate, 673
Pentaerythritol tris-(3-aziridinopropionate), 195, 346,
628
Pentaerythritol trivinyl ether, 584
Pentamethyldipropylenetriamine (PMPTA), 238
N-2,2,6,6-Pentamethylhydroxypiperidinyl diester
of 4-methoxyphenylmethylenemalonic acid, 105
2,4-Pentanedione, 241 –242, 258
Peracetic acid, 277
Peracids in photodegradation, 98, 104, 110
Perchloric acid, 288
Perfluoroalkoxyalkane, 569
Perfluoroalkyl acrylates, 337, 583, 626
Performic acid, 110
Periodate titration of hydroperoxides, 115
Permanent red pigment, 425–426
Permeability, 144, 147–149, 529, 632–633
effect of pigmentation, 147 –148, 428, 430, 432,
446, 470, 513
Persoz pendulum test, 92
PET, see Polyethylene terephthalate
pH:
effect on corrosion, 139–140, 146, 150
effect of pigment dispersion stability, 446–447
in electrodeposition, 536, 536, 538, 544
of latex paints, 188 –189, 192
of water-reducible resin coatings, 167, 169, 526
Phase shift (angle), 72 –73
1,10-Phenanthroline, 301
Phenol-blocked isocyanates, 254–255
Phenolic antioxidants, 104, 106, 292, 485
Phenolic resins, 9, 146, 290–293, 512, 532, 550, 609,
611, 667
etherified, 293
novolac resins, 276, 292–293, 660
alcohol-soluble, nonheat-reactive, 292
oil-soluble, nonheat-reactive, 292
resole resins, 291–292, 609
alcohol-soluble, heat reactive, 291–292
oil-soluble, heat reactive, 292
rosin-modified, 292– 293
Phenoxy resins, 272, 439
Phenyl acid phosphate, 239
m-Phenylenediamine, 281
Phenyl isocyanate, 235, 240
Phenylmethyldichlorosilane, 326
p-Phenylphenol, 292–294
Phenyl silicone resins, 328–320
Phenyltrichlorosilane, 326
Phenyltrimethoxysilane, 326
Phosphate conversion coatings, see Conversion
coatings
Phosphate-functional acrylic resins, 130, 146, 165
Phosphate group effect on adhesion, 130, 144, 146,
165, 178, 290, 512, 659
INDEX
Phosphophyllite, 126, 540
Phosphoric acid, 150, 286, 290, 293, 609
Photoacid generator (PAG), 590
Photoacoustic FTIR, 115
Photochemically reactive solvents, 349, 373
Photochemical ozone creation potential
(POCP), 372, 374
Photoinitiated oxidative degradation, 97– 100
Photoxidation rate (PR), 114– 115
Photoinitiators, 576– 580, 585
bimolecular, 579–580
cationic, 585
unimolecular, 578–579
Photosensitizers, 579–580
Photostabilization, 100–108
Phthalate esters, hydrolytic stability, 116, 311
Phthalocyanine blue, 419, 426–427, 440, 505
Phthalocyanine green, 427
Physical ageing, 81–82, 87
Pickup on brushes, 474
Picture framing, 502
Pigments, 5, 417 –433
acetylene black, 428, 623, 630
aluminum, see Aluminum pigments
black, see Carbon black
bleeding, 424– 425, 448
blue, 426–427
bronze, 429
cholesteric, 406, 429
chromate, see Chromate pigments
corrosion inhibiting, see Corrosion inhibiting
pigments
dispersion, see Dispersion of pigments
effect on catalysis of MF resins, 224, 419, 471
effect on exterior durability, see Exterior durability,
effect of pigments
effect on gloss, see Gloss, effect of pigmentation
effect on mechanical properties, 80–81,
466 –468
effect on permeability, 147–148, 513, 660
functional, 431 –432
green, 426–427
inert, see Inert pigments
interference, 394, 406, 429
IR reflecting, 422, 432, 616
Latex, see Latex, pigments
metallic, 428. See also Aluminum pigments
nano-pigments, see Nano-pigments
orange, 423, 425
passivating, see Corrosion inhibiting pigments
pearlescent, 406, 428
in primers, see Primers, pigmentation
in radiation cure coatings, 419, 586–588
red, 425–426, 513
surface treatments, 107, 115, 224, 419, 412, 428,
436, 439, 441, 445–446
UV absorption, 106–107, 418, 586 –587
711
white, 390 –391, 418 –421. See also Titanium
dioxide
yellow, 423–425
Pigment volume concentration (PVC), 147, 150–152,
461–471, 542, 559, 588, 639 –640, 642– 643,
646–647. See also Critical pigment volume
concentration
effect on light scattering, 390, 646
effect on mechanical properties, 81
Pin disk mill, 562
Pinholes, 506, 558, 570,665
Pin mills, 454–455
Pinolenic acid, 296–297
Pipe coatings, 550, 566
Piperidine nitroxyethers, 447
pKa, 174, 223
Plasma discharge, 128, 631
Plasma treatment:
of aluminum, 145
of steel, 128
Plastic flow, 8
Plastics:
abrasion resistant coatings for, 590, 632–633
adhesion to, 131– 132, 337, 580, 630–633
coating of, 507, 555, 589, 590– 592,
628–633
crazing, 132, 632
in-mold coating, 341, 628–629, 632
post-mold coating, 629 –633
Plastisols, 258, 336, 614
acrylic based, 168
Plywood, 620
PMA, Poly(methyl acrylate)
PMMA, Poly(methyl methacrylate)
PMTA, see Pentamethyldipropylenetriamine
PnBA, see Propylene glycol monobutyl ether
POCP, see Photochemical ozone creation potential
Poisson’s ratio, 82
Polarization, 137 –138, 141, 154
Polyacrylic acid (PAA), 652
Polyamides:
amine-terminated, 146, 281–282, 284
imidazoline formation, 282
Polyamide-modified alkyds, 10, 315
Polyaniline, 141
Polyaziridine cross-linkers, 195, 345–346, 625, 628
Poly[2,5-bis(N-methyl-N-propylamino)phenylenevinylene], 141
Polycaprolactone-polyethyleneimine block
copolymers, 441
Polycaprolactone polyols, see Caprolactone polyols
Polycarbodiimide cross-linkers, 195, 346–347, 553
water emulsifiable, 195, 346
Polycarbonates, 590
brittle–ductile transition temperature, 76
Polycarbonate polyols, 248, 259, 261, 347
Polycarbonate polyol dimethacrylate, 554
712
INDEX
Polycarbonate-modified water-reducible acrylic
resin, 170
Poly(dimethylsiloxanes), 326–327, 590, 643
fluids, 17, 326, 495, 501, 504, 508, 520, 624
hydroxy-functional, 284, 329
modified, 86
polyether block copolymers, 327
polyether-terminated, 327, 502
vinyl ether-termiated, 586
Polydispersity (PD), 10–11, 23, 163–164, 312,
518, 553
Polydispersity index (PDI), 10
Polyester resins, 4, 24 –27, 199– 211, 517
acrylated, 554, 584
acrylic resin blends, 79, 520
allyl-functional, 552
carbamate-functional, 209
carboxylic acid-functional, 204 –205,
551– 553, 559
functionality, 25, 199 –202, 206, 208, 211
high solids, 205– 208, 520–521
hydrolytic stability, 109, 200, 202–205,
209– 210, 517
hydroxy-functional, 24–27, 200–208, 220–221,
517, 551–553
with isocyanates, 239, 241, 247 –248, 259 –264,
266
isocyanate-terminated, 259
kinetics of formation, 25– 27
in powder coatings, 211, 551–553, 559
silicone-modified, 99, 110, 330, 518, 554
transesterification by alcohol solvents, 209–210,
369
unsaturated, 583. See also Styrene-unsaturated
polyester resins
water-reducible, 209–210, 350, 369, 526
Polyethers, 15, 242, 247, 259, 264
isocyanate-terminated, 259
Polyethylene, 15, 633. See also Polyolefins
Polyethylene glycols, 53, 585
Polyethylene glycol diglycidyl ether, 277
Polyethyleneimine, 144, 345
Polyethylene/poly(vinyl acetate) wax, 672
Poly(ethylene terephthalate) (PET), recycled, 211,
312, 555
Polyhydroxystearic acid, 441
Poly(4-hydroxystyrene), 670
Polymeric dispersants, 440, 446
Polymerizable surfactants, 181 –182, 648
Polymerization, 17–27
cationic, see Cationic polymerization
chain-growth, 17– 23, 159, 177, 299
controlled free radical (CFRP), 22–23, 164–165,
446– 447, 602
emulsion, see Emulsion polymerization
free radical initiated, see Free radical
polymerization
photoinitiated, 123, 574–593
reversible addition-fragmentation-chain transfer
(RAFT), 22
stable free radical polymerization (SFRP), 22
step-growth (condensation), 2, 10, 199, 225, 307
suspension, 184, 336
Poly(octyl acrylate), 495, 501, 508, 557
Polyolefins, 555. See also Polypropylene-EPDM
rubber blend (TPO)
adhesion to, 131–132, 336–337
Poly(phenylene oxide), 631
Polypropylene, 430, 661. See also Polyolefins
Polypropylene-EPDM rubber blend (TPO), 134, 337,
513, 630–631
Polypropylene glycol, 258, 261
Propyleneimine, 345
Polypropylene powder pigments, 430
Polysiloxanes, see Poly(dimethylsiloxanes); Silicones
Polysiloxane–polyether block copolymers, 259, 327
Polystyrene latex pigments, 430
Polytetrafluoroethylene (PTFE), 337
Polytetramethylene glycol, 261, 632
Polythiophene, 141
Polyurea 2K coatings, 249–250, 661
Polyurethanes, 231, 266–267. See also Urethane
coatings
hydroxy-functional, 266–267
Polyurethane dispersions (PUD), 37–38, 261–263,
346, 530, 584, 628, 633
autoxidizable, 262, 322, 628, 663
hybrid acrylate/urethane, 186, 262–263,
322, 530– 531, 584, 628, 631, 649
Poly(vinyl acetate), 4, 356
latexes, 182, 185–186, 191 –193, 465
Poly(vinyl alcohol) (PVA), 182, 336, 621
Poly(vinyl butyral), 201
Poly(vinyl chloride) (PVC), 8, 108–109, 147. See also
Vinyl chloride copolymers
brittle –ductile transition temperature, 76
stabilizers, 108–109, 336
Poly(vinylidene chloride) latex, 146, 178,
190, 546
Poly(vinylidene fluoride) (PVDF), 167, 337,
555, 569
Popping, 349, 360, 366, 483, 506–508, 527–528,
552 –553, 600, 627, 632
Poppy seed oil, 470 –471
Porosity of films, 513, 640, 647
Porosity of wood, 642
Post-mold coating of plastics, 629 –633
Potassium fluoride, 144
Potassium hexafluorosilicate, 144
Potassium hexafluorozirconate, 144
Potassium hydroxide, 144
Potassium manganate, 148
Potassium silicate, 152, 333, 664
Potassium tripolyphosphate (KTPP), 448, 640
INDEX
Pot life, 32. See also Package stability
of epoxy–amine coatings, see Epoxy–amine
coatings, pot life
of 2K urethane, see Urethane coatings, 2K, pot life
of thermosetting latexes, 194–195, 344, 346, 530
of trialkoxysilyl-functional latexes, 332, 530
Powder coatings, 548–571, 629
acrylic, 23, 166, 553 –554
application, 554, 564 –569, 604, 615
clear coats, 553, 604
coalescence, 550, 555– 557
color matching, 563–564, 570
cross-linking with blocked isocyanates, 250, 255,
257, 552–553, 557
epoxy, 274, 276, 280, 287, 550, 597
formulation, 555–559
hybrid, 550–551, 559
leveling, 555–558, 570
low gloss, 559 –560, 570
manufacture, 560 –564
polyester, 211, 551 –553
primers, 550, 553, 549, 597
quality control, 563–564
silicone, 554
size distribution, 557, 562 –564, 566
thermoplastic, 555, 562, 558–559
UV cure, 554 –555, 570
waterborne, 571
PR, see Photoxidation rate
Precision coating, 486, 627
Precision paint collider, 90
Premixing, 559, 451, 455–456, 550
Prepolymers, 242–243, 246–247, 250, 260 –262
Prepolymer mixing process, 261 –262
Preprimed hardboard, 627
Primers, 323–324, 512–515
adhesion to, 127, 133, 470, 513, 545, 597,
631, 660
binders for, 512 –513, 667, 671, 675
color, 513
conductive, 540, 623, 630
corrosion protecting, see Epoxy–amine coatings,
corrosion protection
effect of PVC, 133, 470, 513, 542, 545,
597, 660
electrodeposition, 250, 254, 259, 285, 324, 507,
538 –545, 596 –597
epoxy, 286, 289, 292
high solids, 514 –515
latex, see Latex-based coatings, primer
with passivating pigments, 149– 151, 514,
665 –667, 675
pigmentation, 470, 513– 514, 542
powder, 549–551, 553, 597
red lead in oil, 149
stain blocking, 638
wash, see Wash primers
713
waterborne, 150 –152, 286, 529, 604, 671
UV cure, 671
zinc-rich, see Zinc-rich primers
Primer-surfacers, 378, 545, 597–597, 671
Printing inks, 293, 579, 586, 590–591, 623
Print resistance, 622, 625
Product coatings, 2– 3, 595–616, 619–634
Prohesion testing, 145, 154
Propagation, 18, 98, 104, 114, 183
1,3-Propanediamine, 34
1-Propenyl ethers in UV curing, 583
Propoxylated ethylenediamine, 624
1-(n-Propoxy)-2-propanol, 166, 350, 528
Propylene glycol monobutyl ether (PnBA), 37, 654
Propylene glycol monoester of sunflower fatty
acids, 648
Propylene oxide, 161, 340
Protective colloid, 182
PTFE, see Polytetrafluoroethylene
pTSA, see p-Toluenesulfonic acid
PUD, see Polyurethane dispersions
Pulverization, 562 –564
Purity (color), 399
PVA, see Poly(vinyl alcohol)
PVAc, see Poly(vinyl acetate)
PVC, see Pigment volume concentration;
Poly(vinyl chloride)
PVDF, see Poly(vinylidene fluoride)
2-Pyrrolidone, 255
Pyruvic acid ester oximes, 255
Quality control tests, 89 –96, 393, 414, 563–564
Quenching, 81, 100 –102, 576, 580
Quenching volume, 101
Quickee mill, 453
Quinacridone pigments, 426, 448
Quinone methide, 292
Quinuclidine, 237
QUV weather testing, 81, 114, 116–117
Radiation curing, 574–592. See also Electron beam
curing; UV curing
RAFT, see Reversible addition-fragmentation-chain
transfer polymerization
Random copolymers, 8, 163, 169
Raoult’s law, 361
Reaction rates, 29–33
Reactive diluents, 161, 207, 249, 276–277, 289–290,
312–313, 520, 581, 651
Reactive organic compounds (ROC), 372, 374, 378
Reactive organic gases (ROG), 372
Reactive silanes, 128, 130–131, 260–261, 264,
331–332, 634
Recoat adhesion, 133
Red iron oxide, see Iron oxide, red
Red lead, 149, 431
Red pigments, 425–426
714
INDEX
Redox polymerization, 179
Redwood, 638
Refinish automobile paints, 378, 670–674
UV cure primer, 671
Refractive index, 189, 206, 386, 389, 410, 418,
420, 471, 645
effect of difference on light scattering, see
Scattering of light
Relative evaporation rates, see Evaporation of
solvents, relative
Relative humidity (RH), 241, 249, 260, 362, 366,
373, 498, 525, 529, 662, 666, 673
critical (CRH), 373, 498
Relative viscosity, 54–58, 368
Relaxation, 72, 82, 84, 86
Reliability theory, 70, 112, 659, 681
R enamel, 609
Repair coatings, 597, 604–605
Resole phenolics, 286, 288, 291–292
Retained solvent, 28
Retraction, see Crawling
Reversible addition-fragmentation-chain transfer
polymerization (RAFT) 22
Reverse impact resistance, 93
Reverse roll coating, 485 –487, 611, 614, 629
RH, see Relative humidity
Rheology, 41–66, 68–94. See also Viscosity;
Mechanical properties
Rheometers, 48–49, 64
Ricinoleic acid, 296–297, 303, 315
Rippling of direct current, 541
ROC, see Reactive organic compounds
ROG, see Reactive organic gases
Roll coating, 246, 485–487, 528
direct, 65, 485–486, 611
hand, 474– 475, 508, 647, 653
reverse, 485– 487, 611, 614
Ropaque latex pigment (Rhoplexw multilobeTM latex),
185, 642, 647
Rosin, 292– 293, 296–297, 301, 309, 622, 669, 671
rosin-modified phenolic resins, 292–293
Rota-rub test, 85
Rotor/stator mixers, 451
Rubber particles, effect on abrasion resistance,
84–85
Rubbing, 622 –623
Rule 66, 373
Rutile titanium dioxide, see Titanium dioxide, rutile
S. Styrene
SAA, see Styrene/allyl alcohol copolymers
Safflower oil, 309, 312, 321, 644
Sagging, 45, 52, 62–63, 360, 365–366, 431, 478,
481–483, 496–498, 521, 589, 599, 638, 642,
644, 653
hot, 498
oven, 498, 522, 544
Salt spray (fog) test, 154–155
Sandblasting, 125 –126, 659, 676
Sanding sealers, 621, 624
Sand mills, 453
Saponification resistance, 127, 148–149, 152, 323,
512, 647
Satin paint, 643
Scanlan equation, 77
Scanning electron microscope (SEM), 114
Scanning probe microscopy, 85
Scattering coefficient, 390, 416
Scattering of light, 189, 388–393, 409, 420, 423, 425,
430, 436, 458, 471
effect of refractive index difference, 389–390, 410,
418, 646
Scattering of UV radiation, 586– 587
Scratch resistance, 321
Scrub resistance, 193, 645– 646, 648, 654
testing, 85
Scumming of epoxy–amine coatings, 282
Sebacic acid, 205
Seed latex, 178, 182, 184, 192
Self-polishing marine coatings, 669–670
SEM, see Scanning electron microscope
Semi-drying oils, 297–298, 303, 307 –308, 319
Semigloss paint, 645, 649
Sequence length in free radical polymerization, 163
Settling of pigments, 437, 447, 457
SFRP, see Stable free radical polymerization
Shading stains, 621
Shear flow, 41–63
Shear modulus, 74, 76
Shear rate, 41– 46, 48–50, 61–62, 436,
439, 443
in brush application, 474, 641 –642
in spray application, 476, 483
Shear strain, 74
Shear stress, 42–44, 48, 60–61, 74, 181, 436,
449 –451
Shear thickening, 43, 49
Shear thinning, 43– 44, 46, 49, 62–63, 167, 169,
180 –182, 185, 188, 439, 450, 483, 531,
641 –642, 653
Sheen, 407–408, 413
Sheet molding compound (SMC), 502
Shot mills, 453
Shrinkage during film formation, 80, 123, 588, 590,
621, 633
effect on adhesion 80, 123, 584, 592
Side stripers, 607
Siding, 68–69, 74, 80, 627
Silanes, reactive, see Reactive silanes
Silanols, 130, 326, 330– 332, 631
Silicon dioxide pigments, 430, 432, 508, 550
effect on viscosity, 45, 61, 432, 522
effect on abrasion resistance, 84
gloss effects, 410, 431, 588, 623
INDEX
Silicones, 99, 110, 326–331
abrasion resistant coatings on plastics, 84
amino-functional, 329
in antifouling compounds, 669–670
carbamate-functional, 329
epoxy-functional, 329, 586, 671
elastomers, 327, 669
fluids (oils), see Poly(dimethylsiloxane), fluids
hydroxy-functional, 152, 329
intermediates, 330
powder coatings, 554
resins, 327 –329
hydrolysis, 110, 330– 331
exterior durability, 99, 110, 328, 330
silicone-modified acrylic resins, 86, 110, 330–331,
518, 614
silicone-modified alkyd resins, 329 –330,
528, 667
silicone-modified polyester resins, 110, 330,
518, 614
waterborne, 329
Silicone-modified resins, 99, 110, 329–331, 518,
528, 554, 628
Silicotungstates, 144
Siloxanes:
hydroxy-functional, 284
modified epoxy resin, 284
polyether-terminated, 327
Silsesquioxanes, 327
Simulated tests, 89–90
Size exclusion chromatography (SEC), 12
Skid proof paints, 668
Skinning, 301, 651
Small’s molar attraction constants, 352 –353
SMC, see Sheet molding compound
Sodium benzoate, 665
Sodium bicarbonate, 192
Sodium carbonate, 23
Sodium dichromate, 150
Sodium dodecylbenzenesulfonate, 185
Sodium erythroborate, 179
Sodium formaldehyde sulfoxylate, 179
Sodium hydrosulfite, 144
Sodium hypochlorite, 631
Sodium lauryl sulfate, 180
Sodium methoxide, 256
Sodium nitrite, 665
Sodium montmorillonite clay, 191
Sodium polyacrylate, 447
Sodium silicate, 152, 333, 664
Sodium sulfite, 179
Sodium thiosulfate, 179
Sodium vinyl sulfonate, 192, 646
Soft drink cans, see Beverage cans
Softening point, 87, 272
Sol–gel coatings, 333
Solubility, 351– 357
715
Solubility parameters, 351–356
limitations of, 355
three-dimensional, 354 –356
Solutions:
glass transition temperatures, 17
viscosity of, 52 –59, 65, 367–369
Solvents, 349– 379
air pollution, see Volatile organic compounds
aminolysis of esters, 284, 350
coalescing, see Coalescing solvents
conductivity, 350, 370, 480
density, 352, 356, 367–368, 370
effect on adhesion to plastics, 132, 134–135, 632
effect on intrinsic viscosity, 55
effect on pigment dispersion, 438– 440,
443–444, 448
effect on viscosity, 54– 59, 367–369
evaporation, see Evaporation of solvents
effect on surface tension, 491, 495, 500–502
evaporation from films, 28, 361, 363–367, 376,
644, 648, 650, 654
evaporation from high solids coatings, 365–366,
497
evaporation from waterborne coatings, 366–367,
507–508, 525, 642
flammability, 369–370, 485, 524, 529, 644, 668
glass transition temperature, 17
hydrogen-bond acceptor, see Hydrogen-bond
acceptor solvents
hydrogen-bond acceptor-donor, see Hydrogen-bond
acceptor-donor solvents
mixture, computer calculation, 356
odor, 349–350, 644
retention, 17, 365
silicone fluids, 327, 377, 669
solubility theories, 356
surface tension, 370–371, 491, 521
toxic hazards, 371 –372
transesterification by, 209–210, 314
urethane grade, 369
viscosity, 56, 59, 367– 368
VOC exempt, 284, 313, 321, 327, 350, 377, 669
Solventborne coatings, 511–522
Solvent recovery, 375
Solvent resistance, 94, 195–196
Solvent shock, 440
Sorbitol, 277
Soybean oil, 263, 297–298, 303–304, 307–309,
312, 319, 516, 672
blown, 621
epoxidized, see Epoxidized soybean oil
maleated, 304
fatty acids, 307, 319, 516
epoxidized, 557
Spacers, 640, 646
Spar varnish, 292, 667
Spattering, 65, 475, 647, 653
716
INDEX
Special purpose coatings, 2–3, 658 –676
Spectrophotometer, 398
Specular gloss, 407–408, 411
Specular gloss meters, 412 –414
Specular reflectance, 385–386
Spontaneous combustion, 644
Spray booths, 358, 361, 371, 375, 482–483, 559
Spraying, 475–483
air, 358, 476–478, 482 –483
air-assisted airless, 476, 478, 483
airless, 358, 476, 478 –479, 482 –483, 528
atomization during, 476–483
dual feed, 246– 247, 250, 266, 476, 478 –479, 674
effect on popping, 483
effect on sagging, 478, 481–483, 497, 521
electrostatic, 350, 476, 479–481, 521
bell, 476, 479 –480, 47, 496
disk, 479–480
effect on leveling, 495
Faraday cage effect, 481, 566 –558, 570
on plastics, 632, 630– 631
of powder, 564–567
triboelectric, 567, 570
of waterborne coatings, 480 –481
extensional viscosity effects, 65– 66, 478
flame, 568–569
formulation considerations, 482–483
of high solids coatings, 481, 483, 498, 521 –522
high volume, low pressure air (HVLP), 476–478,
483, 521, 674
hot, 481, 521, 623
low volume, low pressure air (LVLP), 478
solvent evaporation during, 358, 361–362, 365,
478, 482–483, 497
with supercritical fluids, 374, 481 –483, 521
transfer efficiency, see Transfer efficiency
Spreading, 124 –125, 131
Stabilization:
of aqueous pigment dispersions, 446–449
of latexes, 180 –181, 641
of organic media pigment dispersions, 436 –442
by surfactants, see Surfactants, for stabilization of
pigment dispersions
Stable free radical polymerization (SFRP), 22
Stains, 417, 425, 621
Stain blocking, 641
Stainless steel, 127–128, 141, 145, 147, 525
Stain resistance, 470, 590, 646
Stannous chloride, 242
Stannous methane sulfonate, 559
Stannous octanoate, 241
Statistics, 252, 683–684
STATRAM, 134
Stearic acid, 296–297, 428
Steel, 127–128
black plate, 608
cold-rolled, 127, 138, 155
corrosion protection, see Corrosion
galvanized, see Galvanized steel
hot-rolled, 138
stainless, see Stainless steel
surface treatment, see Conversion coatings;
Plasma coatings
tin-coated, 142, 607–608
zinc-coated, see Galvanized steel
Steel ball mills, 451 –452
Steel Structures Painting Council, 659
Step-growth polymerization, see Polymerization,
step-growth
Steric repulsion, 180–181, 437. See also Entropic
repulsion
Stern layer, 180
Storage modulus, see Modulus, storage
Storage stability, see Package stability
Stormer viscometer, 51–52, 643, 653
Strain, 70 –75, 80, 82–84, 86 –87
Stress, 70–78, 80, 82–88, 91, 138
internal, see Internal stress
Stress-strain tests, 70–72, 74–75, 84
Strontium chromate, 150, 431, 666, 675
Strontium soap stabilizers, 109, 336
Styrenated alkyds, 320, 512
Styrene/acrylic latexes, 111, 151, 185, 187, 329,
529, 638, 643, 646, 647, 666
Styrene/acrylic acid copolymers, 652
Styrene/allyl alcohol copolymers (SAA), 58–59, 310
Styrene/butadiene latexes, 178, 529, 647
Styrene-maleic anhydride copolymer, 559
Styrene-unsaturated polyester resins, 340–342, 581,
623, 626, 629, 633
waterborne, 341–342
Subtractive color mixing, 401–402
Succinic acid, 205
Sulfonic acid catalysis, 217, 221 –224, 288–289, 293
blocked, 223, 288, 506
Sulfonium salts, 585, 589
Sun-10 exposure, 113
Sunflower oil, 296–297, 309, 313, 645, 648
Sunlight, 97–99, 112 –113, 116 –117, 383
Supercritical fluid spraying, 374, 481–482, 521
Superdispersants, 440
Surface driers, 300
Surface free energy, 124, 131, 436, 448
Surface resistivity, 632
Surface tension, 327
dynamic, 13, 491, 502, 633
effect on abrasion, 84, 86
effect on adhesion, 122, 124 –125, 131, 630
effect on friction, 84, 602, 610
effect on foaming, 182
effects on latex coalescence, 36
effect on leveling, 492–506, 557, 653
effect on pigment dispersion, 436, 448
effect on powder coalescence, 556 –557
INDEX
effect on spray atomization, 475, 477
effect of temperature, 370, 491, 502, 558
effect on wetting, 122, 124–125, 131, 436, 448,
499, 558
factors controlling, 491
of high solids coatings, 500 –501
of solvents, 370 –371, 491, 521
of waterborne coatings, 525
Surface tension differential driven flow, 370– 371,
491– 492, 500– 501, 516, 557, 650, 653
effect on cratering, see Cratering
effect on crawling, see Crawling
effects in curtain coating, 488, 502
effect on edge coverage, 545, 556
effect on floating, 504 –505
effect on leveling, 493 –495, 556, 650, 653
Surface tension gradient driven flow, see Surface
tension differential driven flow
Surface treatment of plastics, 131–132, 337, 631–632
Surface treatment of pigments, see Pigments, surface
treatment
Surfactants:
anionic, 9–4-6, 179 –181, 533, 645, 652
for cleaning metal, 125
fluorocarbon, 84, 494, 499, 610
in latex preparation, 179–182
nonionic, 35, 179–181, 532, 609, 640, 645, 652
polymerizable, 181 –182, 532, 648
for stabilization of pigment dispersions, 440, 442,
445 –448, 640
Suspension polymerization, 184, 336
Sward rocker, 92
Swelling of cross-linked polymers, 77, 94, 675
calculation of cross-link density, 77
Synergism, 104 –105, 114
Taber abraser, 83, 93
Taffy process for epoxy resins, 273–274
Talc, 430, 505
Tall oil fatty acids (TOFA), 296–297, 303, 309, 316,
318, 323, 516, 624, 672
Tan delta, 73–74, 79–80, 88, 467
T-bend test, 93
TDI, Toluene diisocyanate
TEA, Triethylamine
Telegraphing, 499, 502, 544
Temperature:
brittle –ductile transition, 76, 86, 187
critical, 481
dependence of surface tension, 370, 491, 502, 558
dependence of viscosity, 53 –54, 481, 556–557
effect on corrosion, 140, 146, 148, 153–154
effect on mechanical properties, 71–81, 83–84,
86 –88, 90–93
glass transition, see Glass transition temperature
Tensile-at-break, 71–72, 74, 84
effect of pigmentation, 467
717
Tensile modulus, see Modulus, tensile
Tensile strength, see Tensile-at-break
Tensile tests, 71, 87
TEOS, see Tetraethyl ortho silicate
Termination, 18 –20, 25, 109 –110, 192, 581
by combination, 18, 20, 183
by disproportionation, 18, 20, 183
Tests on coatings, 89 –94
empirical tests, 90–94
field exposure tests, 89
laboratory simulation tests, 89–90
TETA, see Triethylenetetramine
Tetrabutylammonium bromide, 551
Tetrabutylphosphonium bromide, 559
Tetraethoxysilane, 145
Tetraethyl orthosilicate (TEOS), 152, 333, 673
Tetra(2-hydroxyethyl)bisamide, 23, 165. See also
Bis-(N,N)-di-(2-hydroxyethyl)adipamide;
Hydroxyalkylamides
Tetraisobutyl titanate, 330
Tetraisopropyltitanate, 211, 316, 330
Tetramethoxymethylglycoluril (TMMGU), 211, 229,
553, 559, 605
Tetramethylammonium octonoate, 238
2,4,7,9-Tetramethyl-5-decyne-4,7-diol alkoxylates,
508, 601
Tetramethylguanidine (TMG), 344
Tetramethylolglycoluril, 228– 229
2,2,6,6-Tetramethylpiperidinol, 104, 553. See also
Hindered amine light stabilizers
Tetramethyltetravinyl-cyclotetrasiloxane, 333
Tetramethyl-m-xylidene diisocyanate (TMXDI), 244,
246, 552
prepolymer with trimethylolpropane, 246, 552
Texanol, see 2,2,4-Trimethylpentane-1,3-diol
isobutyric ester
Tg, see Glass transition temperature
TGA, see Thermogravimetric analysis
TGIC, see Triglycidylisocyanurate
THEIC, see Tris(2-hydroxyethyl)isocyanurate
Thermal coefficient of reactivity, see Activation
energy
Thermal expansion coefficient, 14, 53, 468
Thermal mechanical analyzer (TMA), 87, 123
Thermal stability, 100, 328, 422–423, 614
Thermogravimetric analysis (TGA), 225,
362, 558
Thermoplastic acrylic polymers (TPA), 65, 159–160,
517, 529, 672
Thermoplastic polymers:
brittle–ductile transition temperature, 76,
86, 187
film formation with, 28
glass transition temperatures, 15 –17
Thermosetting acrylic polymers (TSA), see Acrylic
resins, thermosetting
Thermosetting latexes, see Latex, thermosetting
718
INDEX
Thermosetting polymers:
brittle–ductile transition temperature, 76
film formation with, 28 –34
glass transition temperatures, 16–17, 77 –79
Theta conditions, 55
Thickeners, 62–63, 640. See also Associative
thickeners
Thioglycolic acid, 603
Thioindigo maroon, 107
Thiol-ene UV coatings, 583– 584
Thiosulfate, 179
Thioxanthones, 579, 587
Thixotropy, 44– 45, 60–61, 182, 315, 432, 437, 474,
483, 486, 497–498, 522, 599, 640 643
Three roll mills, 455 –456
Through driers, 300
Throw power, 541 –542, 544
Tie coats, 132, 336
Time of flight secondary ion mass spectroscopy
(ToF-SIMS), 501
Time-temperature superimposing, 72, 84
Tin, 137, 142, 607
Tin biocides, 669
Tin catalysts:
for esterification, 26, 200, 203, 207, 312, 319
for isocyanate reactions, 237 –242, 246, 252, 255,
258– 259, 539, 673–675
inhibition by carboxylic acids, 239, 248, 259
Tin octanoate, 238, 243
Tin plate, 142, 608
Tint bases, 644
Tinting colors, 449, 637, 644
Tinting strength, 426, 456–457, 459, 463, 644
Titanate orthoesters, 200, 203, 441. See also Tetraisobutyl titanate; Tetraisopropyl titanate
Titanium complex with 2,4-pentanedione
(Ti(AcAc)4), 258
Titanium dioxide, 418– 420, 439, 447, 459, 465,
470–471, 504–505, 558
anatase, 107, 115, 389, 418 –419, 586
chalking, 100, 107, 419
effect on catalysis of MF resins, 224, 419, 471
rutile, 107, 115, 389–390, 418–419, 470–471,
586– 587, 640, 642, 645–647, 653
scattering coefficient-particle size, 389–390, 471
slurry, 419, 447, 640
surface treated, 107, 115, 224, 284, 419, 439, 465,
640, 645
tinting of, 418– 419
transparent, 419
UV absorption, 107, 418– 419, 586
in UV cure coatings, 419, 586 –587
Titanium 2-ethylhexanoate, 252
Titanium yellow pigment, 423
TMA, see Thermal mechanical analyzer; Trimellitic
anhydride
TMB, see 2,2,4-Trimethyl-1,3-pentanediol isobutyrate
TMG, see Tetramethylguanidine
TMHDI, see 2,2,5-Trimethylhexane diisocyanate
TMI, see Isopropenyldimethylbenzyl isocyanate
TMMM, see Trimethoxymethylmelamine
TMMG, see Tetramethoxymethylglycoluril
TMMLRP, see Transition metal mediated living
radical polymerizations
TMP, Trimethylolpropane
TMPD, see 2,2,4-Trimethyl-1,3-pentanediol
TMXDI, see Tetramethyl-m-xylidene diisocyanate
TOFA, see Tall oil fatty acids
ToF-SIMS, see Time of flight secondary ion mass
spectroscopy
Toluene sulfonamide-modified melamineformaldehyde resins, 553
p-Toluenesulfonylisocyanate, 241
Toluidine red, 425–426
Toluylpropionic acid, morpholine salt, 151
Top driers, 300
Torsional pendulum, 88–89
Torture track testing, 89
TOTO, see Bis-trioctyltin oxide
Toxic hazards. See also Hazardous air pollutants
(HAP)
amines, 281
aziridines, 345, 625
benzene, 350
bisphenol A, 286, 610
in can coatings, 286, 606, 610
chromates, 127, 144, 150, 423 –424, 611, 675
coal tar, 668
cobalt, 301
epoxy–amine coatings, 281, 660
ethoxyethyl esters, 356
formaldehyde, 220
hydroxyethyl acrylate, 583
iodonium salts, 585
isocyanates, 232, 242–244, 256, 259, 672
lead, 186, 301, 316, 420, 423, 431, 607
maleimides, 580
mercury, 87
methylene chloride, 676
methyl ethyl ketone oxime, 255, 539
nickel compounds, 144
nonylphenylethoxylates, 180
organotin compounds, 669
oximes, 552
ozone, 575
perfluoroalkyl acids and alcohols, 338
pigments, 422
silicon dioxide, 126
solvents, 371–372
triacrylate esters, 583
triglycidylisocyanurate, 276
UV radiation, 575
vinyl chloride copolymers, 809
N-vinylpyrrolidone, 583
INDEX
TPA, see Terephthalic acid; Acrylic resins,
thermoplastic
TPO, thermoplastic polyolefin, see PolypropyleneEPDM rubber blend
Track marks, 475, 486
Trade sales paints, 636
Traffic striping paints, 89, 194–195, 336, 666– 667
Transesterification, 202, 207, 200 –210, 224, 296, 312,
314– 316, 321– 322, 343, 347, 551
Transetherification, 214, 216–217, 220, 223 –224
Transfer efficiency, 374 –375, 475 –478, 480, 674
Transition metal mediated living radical
polymerizations (TMMLRP), 22
Transmission electron microscopy, 185
Transmission of light, 386, 402
Triacetoxysilane, 144
Trialkoxysilyl-functional resins, 165, 196, 332, 530,
590, 631, 673
Trialkoxysilylalkyl methacrylate, 332
2,4,6,-Triamino-1,3,5-triazine, see Melamine
Triarylsulfonium salts, 585, 589
Triarylsulfonium salt of hexafluoroantimonic acid,
585
1,2,4-Triazole blocked isocyanates, see Blocked
isocyanates, 1,2,4-triazole blocked
Triboelectric spraying, 567, 570
Tribology, 83
Tributyltin acrylate polymers, 669
Triethylenetetramine (TETA), 281, 441
Triflic acid (trifluoromethylsulfonic acid), 289, 521
Trifluoroethyl methacrylate resin, 670
Trifluoropropylnethylsiloxane side chains, 670
Triglycerides, 295–296, 303
Triglycidylisocyanurate (TGIC), 276, 551 –553,
558– 559, 606
Triisobutoxysilylpropyl methacrylate, 196
Triisocyanic acid salt of dicyclohexylamine, 151
Triisopentylphosphine, 243
Trilinolein, 298
Trilinolenein, 298
Trimellitic anhydride (TMA), 24–25, 208–210, 288,
314, 550
Trimethoxymethylmelamine (TMMM), 214, 220
Trimethoxymethylsilane, 144
3-Trimethoxysilylpropyl methacrylate, 628, 673
2,4,6-Trimethylbenzoyldiphenylphosphineoxide, 579,
591
3,3,5-Trimethylcyclohexyl methacrylate, 53, 164
Trimethylene carbonate, 248
2,2,5-Trimethylhexane diisocyanate (TMHDI), 244
Trimethylolmelamine, 214
Trimethylolpropane diallyl ether, 342
Trimethylolpropane monoallyl ether, 329
Trimethylolpropane triacetoacetate, 671
Trimethylolpropane triacrylate, 250, 344, 582
Trimethylolpropane tri-(3-aziridinopropionate), 346
Trimethylolpropane triglycidyl ether, 276, 660
719
Trimethylolpropane tri-(3-meraptopropionate), 583
Trimethylolpropane tri-[3-(2-methylaziridino)
propionate], 346
2,2,4-Trimethyl-1,3-pentanediol (TMPD), 204, 210
isobutyrate (Texanol, TMB), 37, 642, 654
Trimethylsilane, 128, 145
Trimethylsilyl methacrylate, 670
Triolein, 298
Tripentaerythritol, 303, 310, 312
Triphenylphosphine, 287
Triphenylphosphite, 103, 329
Tripropylamine, 171 –172, 526
2,4,6-[Tris(dimethylamino)]phenol, 279
Tris(2-hydroxyethyl)isocyanurate (THEIC), 260
Tristimulus values, 398, 401, 404
True solvents, 351
TSA, see Acrylic resins, thermosetting
Tukon tester, 91–92
Tungsten light, 384
Turbulent flow, 63–64
Two package (2K) coatings, see Epoxy-amine
coatings; Latexes, thermosetting; Urethane
coatings
Two roll mills, 456
UCPVC, see Ultimate critical pigment volume
concentration
UF, see Urea-formaldehyde resins
Ultimate critical pigment volume concentration
(UCPVC), 452
Ultrafiltration, 542–543
Ultra high pressure hydroblasting, 126, 659, 667
Ultrasound pigment dispersion, 456
Undecanoic acid, 446
UNIFAC, 361, 370
Universal colorants, 442, 644
Unsaturated polyesters, see Styrene-unsaturated
polyester resins
Uralkyds, 320–322, 628, 662, 667
maleated, 322, 663
Ureas, 109, 213, 228, 233, 246–250, 252, 255–256,
258, 260, 265–266
hydrolytic stability, 109, 200
reaction with acrylic resins, 165
reactions with isocyanates, 233, 248–249
Urea-formaldehyde resins (UF), 194, 228, 252, 256,
260, 288, 624–625
Ureidotriethoxysilane, 145
Uretdiones, 236, 245, 257–258, 263, 265, 529, 553
Urethanes, 231 –233. See also Urethane coatings
acrylated, 582
hydrolytic stability, 109, 241, 260
reactions with isocyanates, 233
reaction with MF resins, 214, 227, 266, 518, 600
Urethane coatings, 4
abrasion resistance, see Abrasion resistance,
polyurethanes
720
INDEX
Urethane coatings (Continued )
acrylate-terminated, 582
aqueous dispersions (PUDs), see Polyurethane
dispersions
acrylated, see Polyurethane dispersions, hybrid
acrylate/urethane
autoxidizable PUDs, see Polyurethane dispersions,
autoxidizable
blends with acrylic latexes, 38, 261–265, 663
blocked isocyanate coatings, see Blocked
isocyanates
exterior durability, 100, 232, 242, 245 –247, 250,
257, 259–261, 263–264
high solids, 245, 248, 249– 250, 520, 675
2K solventborne, 246–249, 597, 601, 604–605,
624, 631, 633, 660–661, 673 –675
pot life; 239 –240, 246 –250, 260, 263, 344, 674
2K urea coatings, 249 –250, 661
2K waterborne, 174, 241, 245, 264–266, 529,
625, 662
pot life, 265, 529
indexing, 247, 265
mar resistances, see Mar resistance
moisture-curable, see Moisture-curable
polyurethane coatings
thermal stability, 232, 241
uralkyds, see Uralkyds
Urethane diols, 266–267, 520
Urethane grade solvents, 369
Urethane oils, see Uralkyds
UV absorbers, 82, 100 –103, 105– 107, 109, 115, 410,
425, 427, 539, 592, 622, 631
nano(zinc zirconium silicate), 106
UV cure coatings, 574–592
acrylic, 581– 583
adhesion to metals, 123, 583–584, 590, 592
bimolecular initiation, 579– 580
cationic, 506, 554, 585–586, 589–590
clear coats for automobiles, 589–592, 603
coil coating, 590, 30
dual cure systems, 589, 603, 633
effect of pigmentation, 419, 471, 586–588, 623
free radical initiated, 508, 578–584
furniture, 621, 623, 625– 626
gloss of, 582, 584, 588–589, 592, 626
oxygen inhibition, see Oxygen, inhibition
particle board fillers, 623
primer for refinish, 671
powder coatings, 554–555
rate of polymerization, 577
thiol-ene, 583–584
unimolecular initiation, 578– 579
vehicles for, 581 –586
waterborne, 195, 263, 584, 592, 603, 626, 628
for wood, 626
UV degradation, see Exterior durability
UV sources, 575 –576, 580, 583, 587
VAc, see Vinyl acetate
VaE, see Vinyl acetate/ethylene latexes
Value (color dimension), 397
Vanadium tetrafluorborate, 144
Van Krevelen equation, 16
Vapor degreasing, 125
Vapor pressure, 358, 361
Vapor pressure depression, 497
Vapor pressure osmometry, 12
Varnish, 292, 302–303, 312–313, 609, 667
conversion, 624
finishing, 606, 610
Vat yellow pigments, 425
VC, see Vinyl chloride
VDC, see Vinylidene chloride
VDF, see Vinylidene fluoride
Vehicles, 5
Veneers, 620, 623, 626
Vinyl acetate copolymer latexes, 182, 185– 186,
191 –193, 332, 642–643, 645–646
Vinyl acetate/ethylene latexes (VaE), 193, 645
Vinyl acetate/ethylene/vinyl neodecanoate latexes,
193
Vinyl chloride copolymers, 4, 8, 335, 555
dispersion grade, 336, 614
solution grade, 28, 335, 609
thermal and UV degradation, 100, 108–109,
335–336
Vinyl ether copolymers, 338, 447
Vinyl ethers, photopolymerization, 583– 584, 586
Vinyl 2-ethylhexanoate, 193
Vinylidene chloride (VDC) latexes, 178, 190, 546, 666
Vinylidene fluoride (VDF) polymers, 147, 337
Vinyl pivalate, 193
N-Vinylpyrrolidone (NVP), 583
Vinyltoluene, 304, 320
Vinyltriacetoxysilane, 145, 332
Vinyltriethoxylsilane, 196
Vinyltrimethoxysilane, 166
Vinyl versatatew (vinyl neo-decanoate), 192–193,
642, 646
Viscoelasticity, 45, 70–745
Viscometers:
Brookfield, 49–50
bubble, 50, 318
capillary, 47– 48
cone and plate, 48, 52, 318
efflux cups, 50–51, 483
falling ball, 339
ICI, 643
paddle, 51–52
rolling ball, 556
rotating disk, 49–50
Stormer, 51 –52, 643, 653
Viscosity, 41–66
of acrylic resins, 160, 164 –165, 167
of alkyd resins, 308 –309, 311 –312, 317–319
INDEX
in brush application, 474
of concentrated solutions, 55– 59
concentration dependence, 55, 58–59, 367–369
determination, 46–52, 316
of dilute solutions, 54 –55
of dispersed phase coatings, 59–63, 164–165, 167,
177, 210, 442, 465
effect on adhesion, 122, 145, 512
effect on film thickness, 474, 484, 642, 652–653
effect of flocculation, see Flocculation, effect on
viscosity
effect of free volume, 28, 52, 56–57
effect of gel particles, 61, 498, 599
effect of hydrogen bonding, 57– 59, 208, 312, 388,
650
effect on leveling, 474, 482, 486 –487, 545, 555,
588, 642, 653
effect of molecular weight, 54– 56, 58, 62, 65
effect of molecular weight distribution, 160–162,
164 –165, 312, 319, 553
effect of pigments, 43, 45, 59–60, 62, 428,
431 –432, 514 –515, 588
effect of resin-solvent interaction, 55–59, 368
effect on sagging, 45, 52, 62–63, 431, 437, 474,
478, 481–483, 486–487, 496–498, 642, 644
effect on settling, 63, 431, 437, 447, 457, 642
epoxy resins, 53–54, 273
extensional, see Flow, extensional
of external phases, 62, 122, 145, 512, 642,
647, 664
glass transition temperature dependence, 53, 58
intrinsic, 54–56
kinematic, 42, 47, 50–51
of latexes, 60, 176, 180–182, 185, 188– 189,193
of latex paints, 45, 52, 63, 65, 640–643, 647,
652 –653
of MF resins, 220
Mooney equation, 60, 62, 441 –442
Newtonian, see Newtonian flow
Non-Newtonian, see Shear thickening; Shear
thinning; Thixotropy
pH dependence in latexes, 188
of pigment dispersions, 62, 441 –442, 450, 452, 455
of pigmented high solids coatings, 441–442
of powder coatings, 556–557
relative, 54–58, 368
of resin solutions, 52 –59, 367– 368
shear rate dependence, see Shear thinning;
Thixotropy
shear thickening, 43, 49
shear thinning, see Shear thinning
of solvents, 56, 59, 367– 368
in spray application, 478, 480–483
temperature dependence, 56, 59, 481, 556 –557
of water-reducible resin coatings, 164–165,
525 –526
weight intrinsic, 56
721
Vitrification, 33
VOC, see Volatile organic compounds
Volatile organic compounds (VOC), 159, 165–167,
170, 174, 190, 205– 206, 209, 245, 248– 250, 308,
312–313, 322, 544, 546, 518 –521, 624– 625,
645, 648 –649, 654 –655
determination, 376–377
regulations, 377 –379, 313, 321, 624, 643, 649, 670
VOC exempt solvents, see Solvents, VOC exempt
Wash coats, 621
Wash primers, 150, 378
Water:
analysis for, 377
evaporation rate, 362
evaporation from waterborne coatings, 358,
366–367, 406, 498, 507–508, 527, 531, 600
heat of vaporization, 362– 363, 508
permeability of films, 143– 144, 112, 153,
190–191, 193, 321, 514, 529, 637,
640, 643, 661
pigment dispersions in, 445– 449, 640, 652
as a plasticizer, 182, 187, 190, 232, 261
reaction with isocyanates, see Isocyanates, reaction
with water
solubility parameter, 352
Waterborne coatings, 524– 533
application by electrostatic spray, 480–481
popping, 507–508
powder coatings, 571
sagging, 366, 498
Waterborne epoxy coatings, see Epoxy-amine
coatings, waterborne
Waterborne urethane coatings, see Polyurethane
dispersions; Urethane coatings, 2K waterborne
Waterborne UV cure coatings, see UV cure coatings,
waterborne
Water-reducible acrylic coatings, see Acrylic resins,
water-reducible
Water-reducible alkyd resins, 313– 314, 526, 528, 651
Water-reducible epoxy esters, 323, 526
Water-reducible polyester resins, see Polyester resins,
water-reducible
Water-thinnable coatings, 210, 521
Wave guides, 590– 591
Wax, 84, 341, 559, 580, 600, 672
Weatherability, see Exterior durability
Weight average molecular weight, 10, 20–21, 176,
184–185
Weight intrinsic viscosity, 56
Wet adhesion, see Adhesion, wet; Corrosion, effect of
wet adhesion
Wet hiding, 646, 652
Wet lapping, 642, 655
Wetting, 435–436. See also Surface tension, effect on
wetting
White lead, 420, 422
722
INDEX
White pigments, 418– 421. See also Titanium dioxide
Williams-Landel-Ferry equation (WLF), 27, 53–54
Wire coatings, 250, 252
WLF equation, see Williams-Landel-Ferry equation
Wood coatings, 101, 106, 111, 186, 190, 263 –264,
321, 471, 555, 584, 588, 619 –628, 637 –638
Wood furniture, see Furniture coatings
Work-to-break, 71, 84
Wrap around, 479
Wrinkle finishes, 505 –506, 559
Wrinkling, 173, 411, 505– 506, 559
Xanthones, 587
XDI, see 1,3-Xylylene diisocyanate
XLD, see Cross-link density
XPS, see X-ray photoelectron spectroscopy
X-ray photoelectron spectroscopy (XPS), 5 –18, 6–9
1,3-Xylylene diisocyanate (XDI), 244, 246
Yellowing, 300, 309, 312, 644
Yellow pigments, 423–425
Yield point (stress), 8, 70–71, 74 –75
Yield strength, 71, 76
Yield stress, 86
Yield value, 43
Zeolite, 14
Zinc, see Galvanized steel; Zinc-rich primer
Zinc acetate, 200
Zinc aluminum phosphate, 150
Zinc-calcium molybdate, 150, 666
Zinc carbonate, 127, 142, 146, 152, 512, 668
Zinc chromate, 150, 422, 431. See also Zinc
tetroxychromate; Zinc yellow
Zinc-coated steel, see Galvanized steel
Zinc complex with 2,4-pentanedione, Zn(AcAc)2, 238
Zinc ferrite, 151
Zinc hydroxide, 127, 142, 146, 152, 512, 665
Zinc iron phosphate, 150
Zinc/manganese phosphate, 150
Zinc molybdate, 150
Zinc molybdenum phosphate, 150
Zinc nitrophthalate, 151
Zinc octanoate, 328
Zinc oxide, 127, 146, 152, 420, 431, 512, 641
Zinc phosphate, 126–127, 150–151, 431, 540, 666
Zinc-rich primers, 151 –153, 333, 431, 470,
663 –665, 667
top coats for, 152, 663–665
waterborne, 152, 333, 664
Zinc salt of benzothiazoylthiosuccinic acid, 151
Zinc salt of hydrogenated rosin, 669
Zinc stearate, 522, 621, 630
Zinc sulfide, 420, 431, 609
Zinc tetroxychromate, 150
Zinc yellow, 150, 665, 675
Zinc zirconium silicate, nano-sized UV
absorber, 106
Zirconium complex with 2,4-pentanedione,
Zr(AcAc)4, 241, 265, 673
Zirconium driers, 300–301, 320, 505
Zn-Mn-Ni phosphate conversion coatings, 127, 540
Related papers
عادات النعي والأحزان والاحداد على الميت في اقليم كوردستان-العراق دراسة فقهية
Journal of Tikrit University for Humanities
LOS NUEVOS ANTISEMITAS Y LA CUESTA DE LA GUERRA
Diario de santiago, 2024
Today's Problems in The Minds of The Great Economists
Today's Problems in The MInds of The Great Economists, 2021
Савети српским средњовековним владарима за успешну владавину [The Advice to Serbian Medieval Rulers for Successful Rule]
Средњи век у српској науци, историји, књижевности и уметности XIII, 2023
Repressive Entsublimierung im Wandel: Über die kontrollierte Freigabe des Selbstgenusses im Postfordismus
Leviathan - Berliner Zeitschrift für Sozialwissenschaft, 2024
Románia mezőgazdaságának vizsgálata, különös tekintettel az EU-csatlakozás utáni időszakra
Acta Carolus Robertus, 2021
Optimum Fracture Treatment Design Minimizes the Impact of Non-Darcy Flow Effects
Proceedings of SPE Annual Technical Conference and Exhibition, 2004
A note on width-parameterized SAT: An exact machine-model characterization
Information Processing Letters, 2009
Türkçede Cümle Ögelerinin Sınıflandırılması Üzerine Bir Ölçüt Önerisi
KORKUT ATA TÜRKİYAT ARAŞTIRMALARI DERGİSİ, 2024
The Sensory Experience of Toilet Training and Its Implications for Autism Intervention
Recent Advances in Autism Spectrum Disorders - Volume II, 2013
Efficient estimation of free energy differences from Monte Carlo data
Journal of Computational Physics, 1976