Abatement of Pollutants by Adsorption and Oxidative Catalytic Regeneration

4374 Ind. Eng. Chem. Res. 1997, 36, 4374-4380 Abatement of Pollutants by Adsorption and Oxidative Catalytic Regeneration Yurii I. Matatov-Meytal and Moshe Sheintuch* Department of Chemical Engineering, Technion-Israel Institute of Technology, Haifa, Israel 32000 The adsorption of phenol, and halogenated phenols on granular activated carbon (AC) modified with metal oxide catalyst, followed by catalytic oxidative regeneration, was studied as an efficient technology for the treatment of dilute wastewaters. The advantages of the combination of these technologies are as follows: (1) the process will be accelerated by the high concentrations of pollutants eluted from the adsorbent; (2) a large number of adsorption-regeneration cycles are expected without loss in capacity; and (3) the low-temperature regeneration will be conducted in situ, even in small units, thus improving the economy of the process. Oxidative catalytic regeneration of spent carbons, performed at 240-300 °C with air, completely restored the adsorption capacity of phenol on the ACs modified with catalyst, even after 10 cycles of regeneration. Under similar conditions, only partial recovery of the adsorption capacity was obtained for carbons loaded with p-chlorophenol and p-bromophenol. The adsorption capacity and the surface area of the AC (Filtrasorb-400) diminished somewhat with the impregnation of oxides (Fe2O3, CuO, and additivities of Cr2O3 or inert silica), but that did not affect the shape of the adsorption isotherms. Introduction The adsorption of organic pollutants by activated carbon (AC) is a well-established technology, yet its cost is still a prohibitive factor. For economic and environmental reasons, spent AC is not disposed of but undergoes several cycles of regeneration. Thermal regeneration of AC is the most common process, but it requires high temperatures (800-850 °C) and, consequently, is usually not conducted in situ, requiring shipment of the spent AC to special regeneration units and contributing significantly to its cost. Moreover, high-temperature regeneration is economically feasible only for large systems that use more than 500 000 lb of granular AC per year (Sonyheimer et al., 1988). We suggest here to combine an adsorption process of pollutants on well-established adsorbents like AC with periodic low-temperature catalytic regeneration of the adsorbent as an efficient technology for treatment of dilute wastewaters. The suggested advantages of the combination of these technologies are as follows: (1) the process will be accelerated by the high concentrations of pollutants eluted from the adsorbent; (2) a large number of adsorption-regeneration cycles are expected without loss in capacity; and (3) the low-temperature regeneration will be conducted in situ, even in small units, thus improving the economy of the process. The technological and scientific problems addressed here are as follows: (1) the strength, capacity, and reversibility of adsorption on AC impregnated with catalyst; (2) the choice of a catalyst that operates at relatively low temperatures but well below the ignition temperature of AC; (3) the engineering of contacting the pollutants with the AC and the catalyst; and, (4) the efficiency of regeneration in batch and fixed-bed studies. We review below the relevant information on AC structure, strength of adsorption, catalysts used in catalytic regeneration of AC, adsorbate-catalyst contacting modes, and alternative regeneration procedures. * To whom correspondence should be addressed. Telephone: 972 4 8292823. Fax: 972 4 8230476. E-mail: cermsll@ techunix.technion.ac.il. S0888-5885(96)00809-3 CCC: $14.00 Adsorption on and Thermal Regeneration of AC. The adsorption of phenols on ACs is normally characterized as physisorption at low temperatures and chemisorption at high temperatures. At relatively high adsorption temperatures, long adsorption contact times, and high oxygen content, phenols tend to irreversibly adsorb on the carbon surface (Grant and King, 1990; Vidic et al., 1993). Single-component adsorption equilibria of phenols on ACs are usually described by Langmuir (Harriott and Cheng, 1988; Koganovskii and Prodan, 1988; Nelson and Yang, 1995) or Freundlich isotherms (Peel et al., 1981; Yong et al., 1985). This suggests reversible adsorption, but actual desorption experiments have been performed only in a limited number of studies. Desorption can be induced by displacement with a compound of a higher affinity to AC and/or by increasing the temperature. Suzuki et al. (1978) classified the adsorbed organics into three groups according to the characteristics of thermal regeneration of carbon: volatile compounds (type I), unstable compounds and refractory adsorbates (type II), and constituted aromatic compounds with side chains and adjacent OH groups (type III). Activated carbon with sorbed organics undergoes the following scenario with increasing temperature: drying and loss of highly volatile compounds occurs at temperatures below 200 °C, vaporization and decomposition of unstable compounds take place at 200 < T < 500 °C, and pyrolysis of nonvolatile adsorbates to form char occurs at 500 < T < 700 °C followed by oxidation of the residue at higher temperatures. Exposure to temperatures of 750-980 °C lead to oxidation of the residual material as well as that of the carbon itself. The latter step includes oxygen attack on the AC itself, which may alter the pore structures where small pores (<2 nm) are lost while larger pores are created. The degree of desorption and conversion to char depends on the nature of adsorbent and adsorbate and the rate of the process. Pyrolysis alone is not sufficient for repeated AC regeneration since the phenolic residue partly fills some of the pores (Wang and Smith, 1985; Harriott and Cheng, 1988). © 1997 American Chemical Society Ind. Eng. Chem. Res., Vol. 36, No. 10, 1997 4375 Other AC regeneration methods that have been explored, but have not been implemented commercially, include wet regeneration or oxidation under high oxygen pressures (10-100 atm) (Mishra et al., 1995), regeneration by extraction with supercritical carbon dioxide (Tan and Liou, 1989) or with organic solvents like ethanol (Tamon et al., 1990), and biological regeneration (Hatchinsonet and Robinson, 1990). Catalytic Regeneration. While adsorption and catalysis are well-established technologies, their combination for regeneration purposes has been employed in only a few cases. Three configurations of the AC/ catalyst contact were suggested: Impregnation of the adsorbent with a catalyst provides intimate contact in a single unit, reduces diffusion paths, and accelerates desorption rate, but it may diminish the adsorption capacity and may lead to metal elution. Organizing the AC and the catalyst in layers within the same bed is the next desirable mode of contact. The two may also be organized in separate units, a mode that requires regeneration by desorption. Crittenden and co-workers (1993) used the latter mode with photoreactive oxide catalysts (TiO2, SnO2, and ZrO2) impregnated on carbon and regenerated the carbon by irradiation. A technique which uses desorption of the toxic compounds from AC by hot water at temperatures up to 200 °C and elevated pressure has been suggested by Levec and Pintar (1995); the desorbed organics can be subsequently abated by catalytic liquid-phase oxidation either in a trickle bed or in a wet air oxidation reactor. Although AC is not usually employed as a support for oxidation catalysts, because of its potential pyrophority, the adsorbed organics may be catalytically oxidized directly on the carbon surface under conditions that do not favor oxidation of the carbon itself (Turk, 1955). Catalytic regeneration of AC saturated with hydrocarbon air-pollutants has been applied by Nwankwo and Turk (1975). The metal oxide catalysts (Cr2O3, CuO, Co3O4, V2O5, MoO3, WO3) were impregnated onto the carbon from their ammonium salt solution. High recovery of the adsorption capacities was obtained, and it varied with the nature of the adsorbate (with complete recovery for toluene) and catalyst. Carbons impregnated with oxides of manganese (Koganovskii and Kaninskaya, 1981), of iron (Koganovskii and Prodan, 1988), and a mixture of iron and copper (Prodan et al., 1988) were used for the adsorption of organics (dyes, surface active agents, phenol) from water solutions. The impregnation of the carbon surface with iron oxide decreased the adsorption of phenol by 20-30% without affecting the adsorption of large molecules. These spent carbons were regenerated in the adsorption column, at 280-350 °C (Astachov et al., 1985; Prodan et al., 1988), and it was found that the adsorption capacity diminished by about 3% per cycle (Koganovskii and Kaninskaya, 1981). The catalyst of choice is one that yields high activity but at temperatures below the ignition point of carbon. The catalytic activity of the metal oxides during phenol liquid-phase oxidation was tested (Kochetkova et al., 1992) and showed the following typical order: CuO > CoO > Cr2O3 > NiO > MnO2 > Fe2O3 > YO2 > Cd2O3 > ZnO > TiO2 > Bi2O3. In gas-phase oxidation the following order of activity was determined within the temperature range 225-440 °C (Simonov, 1990); Pt > Cu > (Cu + Cr) > CuO > (CuO + Cr2O3) > V2O3 > (Cr2O3) > Co3O4 > Fe2O3 > MnO2 > ZnO. Experimental Section Materials. The granular AC employed in this study (Calgon Filtrasorb-400 supplied by Chemviron Carbon) is derived from bituminous coal and has been widely used in the chemical and food industries for adsorption, decolorization, and pollution abatement purposes. This AC had a specific surface area of 1200 m2/g and a broad pore-size distribution (<20 nm pore volume of 0.50 mL/g and <3 nm pore volume of 0.38 mL/g). Its properties were described elsewhere (Montgomery, 1985). The virgin AC was washed with distilled water to remove all fines and then dried at 105 °C for 4 h before being stored in sealed glass bottles for future use. The nitrate salts used for impregnation were analytical grade (Merck). Phenol (P) p-chlorophenol (CP), and p-bromophenol (BP) were purchased in their highest available purity (from Aldrich Chemical Co.) and were used without further purification to form solutions with distilled water. The solute concentrations were determined using UV spectroscopy (Unicam UV-2) at λmax ) 270 (P), 280 (CP), and 226 nm (BP). Oxygen (99.6%), nitrogen (99.99%), and air (zero grade) were used in the regeneration procedure with no further treatment. Impregnated Carbon Preparation. The metal oxide catalysts were prepared by wet impregnation (Satterfield, 1991). The carbon particles (2-5 g) were immersed into a calculated amount of a 10-15% aqueous solution of metal nitrates for 24 h, before filtering the solution and heating the carbon in a stream of air at 150 °C for 2 h. Following that, the carbon was treated with a calculated volume of a 20% ammonium carbonate solution to precipitate the hydroxides. The modified carbon was heated then in air at 160 °C for 2 h and was washed with water to eliminate unchanged residues of metal salts until the amount of metal ions in the wash waters did not exceed 1 mg/L. The AC samples were placed than in a Pyrex column and dried for 1 h at 180 °C, before heating it to 340 °C for 6 h in hot nitrogen to achieve calcination. Then the carbon particles were cooled, ground, and sieved to yield a narrow fraction of particle sizes (0.04-0.06 cm in diameter). Samples C and D underwent the following pretreatment before impregnation with oxide catalysts as described above: these samples were pretreated with a solution of Na2O‚nSiO2 for 18 h and then were filtered and heated in an air current at 160 °C for 2 h, after which they were contacted with a calculated volume of a 5% solution of HNO3 and then dried again in an air current at 180 °C for 2 h. The metal load applied to the AC was calculated from the difference in the metal salt solution concentrations before and after impregnation. These values were checked, in several cases, by grinding the carbon and dissolving it in concentrated hydrochloric acid. All metal analyses were duplicate, with a deviation of less than 5%. Carbon samples were characterized by measuring the specific surface areas due to N2 adsorption (Flowsorb II 2300, Micromeritics) and by conducting adsorption experiments, using a constant carbon mass and varying the solute concentration at 25 °C. The stopped-up Erlenmeyer flasks with carbon particles and phenolic solution were agitated in an automatic thermostatic shaker for at least 3 days. Spot checks of solute concentration were conducted to assure no detectable change after the 3-day equilibration period. A 2-day safety factor was added to the equilibration period. The adsorbent particles were filtered (Whatman filter paper 4376 Ind. Eng. Chem. Res., Vol. 36, No. 10, 1997 Table 1. Initial Properties of Prepared Carbons content, wt % Figure 1. Experimental setup: 1, gas cylinder; 2, CO2 trap; 3, buffer vessel; 4, flowmeter; 5, gas heating furnace; 6, temperature controller; 7, adsorber/regenerator; 8, condensator, 9, silica gel; 10, CO2 analyzer; 11, tank with stock solution; 12, collected water. No. 1), and the carbon samples were dried in a stream of air at 25 °C to achieve constant weight. The quantity of adsorbate loading was calculated by mass balance, knowing the volume of solution, initial and final liquidphase solute concentrations, and mass of carbon. Apparatus and Procedure. A Pyrex column (0.6 cm in diameter) packed with 0.4-1.0 g of carbon particles was subject to periodic liquid-phase adsorption and oxidative gas-phase regeneration (Figure 1). In the adsorption step, a stock solution of 10 mg/L of phenol was fed to the column top at a flow rate 4 mL/min, and the outlet stream was sampled. The length-to-diameter ratio of the carbon bed (with bulk density of 0.425 g/cm3) was sufficient to minimize the end effects of the adsorption column (with empty bed contact time of 1 min). The operation times required to achieve about 99.9% of the phenol influent concentration were 6-7 days. The integral adsorption capacity until breakthrough (qb, at breakthrough of C ) 0.1 mg/L at t ) t*) and the total capacity (qt) were calculated by integrating the difference between the feed (C0) and exit (C) concentrations: qb ) 1 mc ∫0t*f(C0 - C) dt qt ) 1 mc ∫0∞f(C0 - C) dt (1) The regeneration step was carried out in the same column under atmospheric pressure. The oxidant gas (air or 5% oxygen in nitrogen) was fed through a preheater to the bottom of the column at a predetermined flow rate (0.5-1.0 L/min), and its temperature was then increased to the desired regeneration temperature (measured by thermocouples) for a period of about 0.5 h. The reactor and preheator were electrically heated. The carbon dioxide concentration in the gas effluent was determined by an infrared analyzer (Beckman 865) during the regeneration tests; the stream was cooled and dried over silica gel. In order to reduce the times involved in the column adsorption cycle, carbon samples were loaded also by batch adsorption similar to the procedure described above. The spent carbons were removed then from the flasks and were regenerated in the regeneration column. The overall degree of spent AC regeneration was estimated as the ratio of carbon adsorption capacity after each regeneration cycle (q) to its initial value (q0), both at a solute concentration (Ce) of 10 mg/L, while maintaining the same ratio of phenolic solution volume to carbon mass. The reproducibility of these measurements was found to be within 6%; the initial phenol loadings (q0, mg per g of dry carbon) were 97 (F), 87 (A), 85 (C), and 75 (B, D). carbon CuO Fe2O3 Cr2O3 A B C D 0.9 1.8 0.9 1.8 3.7 7.3 3.7 7.3 0.5 1.0 SiO2 So, m2/g 0.9 1.8 1000 920 1000 900 The regeneration efficiency of spent AC loaded with phenol was also estimated from the CO2 balance in the effluent stream and from the results of the column adsorption-regeneration experiments as the ratio of the adsorption bed utilization (qb) and capacity (qt) after regeneration to their values before regeneration. Results and Discussion Catalysts Screening. A preliminary screening of oxidation catalysts was conducted by determining the initial temperature at which detectable CO2 concentrations appear during regeneration of phenol-loaded carbon samples. AC samples impregnated with CuO, CoO, Cr2O3, Fe2O3, NiO, and their mixtures were examined. Fe2O3 and CuO were found to be the most active. A loading of at least 3 wt % (calculated as metal) of a single metal oxides on carbon Filtrasorb-400 was necessary to obtain a detectable CO2 concentration in the gas effluent. Mixtures of these metal oxides exhibit a greater activity than the single oxides. The most promising catalysts were found to be mixtures of Fe2O3 and CuO either with small additions of Cr2O3 (samples A and B) or with carbon pretreatment by inert silica, in order to suppress the carbon ignition (samples C and D). The metal oxide contents of the selected carbon samples used in this study are presented in Table 1. In a related article, we report the temperatures at which phenol oxidation commences on the AC/catalyst samples (typically at 210-230 °C) and the temperatures at which carbon oxidation commences (320-360 °C). These were determined by thermogravimetric and differential thermal analysis (DTA/TGA) (Matatov-Meytal et al., 1997). In the presence of active catalysts, both temperatures (phenol oxidation and AC gasification) decline but the interval where catalytic regeneration is possible becomes too narrow. In this study, all regeneration runs were conducted at temperatures of 240290 °C. The transient CO2 generation curve, during phenol oxidation in the regeneration of 1 g of spent carbon sample (A-D) in an air stream of 500 mL/min, is presented in Figure 2. The curves exhibit three temporal zones with a slow and accelerating rate in the first zone, a constant rate in the second, and a final section with declining CO2 concentrations. It is evident from Figure 2a that the initial production rate of CO2 is almost independent of the metal oxide used, while the constant rate of the second zone varies with the type of catalyst and its loading. Table 2 compares the rates and the phenol conversions, calculated from the CO2 balance for carbons A-D; conversions were in the range 87-95%. With increasing regeneration temperature, the initial rates and the maximal rates are bigger, and consequently, the regeneration process is faster (Figure 2b). The conversions were calculated to be 83% at 240 °C, 88% at 250 °C, 92% at 270 °C and 96% at 290 °C. Based on the results shown in Figure 2 and Table 2, two impregnated carbons (A and C) were mainly used in the remaining studies and were considered adequate in an effort to optimize the catalytic regeneration Ind. Eng. Chem. Res., Vol. 36, No. 10, 1997 4377 Table 3. Langmuir Characteristics of the Carbons Used P Figure 2. Temporal CO2 concentration profiles in the outlet stream during regeneration of different carbon samples (a, top) and for carbon A at different temperatures (b, bottom). Initial phenol loadings (q0, mg per g of dry carbon) were 97 (F), 87 (A), 85 (C), and 75 (B, D). Table 2. Regeneration Rates and Conversions carbon q0, mmol/g of C max CO2 concn, vol % amt of CO2 eluted, mmol/g of C phenol conversn, % A B C D 0.925 0.797 0.921 0.797 0.25 0.32 0.23 0.27 4.90 4.56 4.80 4.55 88.5 95.3 86.7 95.1 process. In regeneration runs without CO2 analysis, spot checks of the recovered phenol adsorption capacity of a given mass carbon (1 g) showed that a 3-h regeneration period was sufficient to restore the adsorption capacity, and further exposure did not improve the recovered capacity. Adsorption Isotherms. The single-solute adsorption isotherms were determined for the three phenols (Figure 3a), using the virgin and the four selected impregnated carbon samples, A-D (Figure 3b). All the materials showed strong adsorption even at low solute concentrations. The isotherms correlate well with the Langmuir adsorption isotherm bCe qe ) Q0 1 + bCe (2) and the estimated lines are presented in Figure 3, while the estimated constants are tabulated in Table 3. CP BP sample Q0, mg/g b, L/mg Q0, mg/g b, L/mg Q0, mg/g b, L/mg F A B C D 169 149 130 138 128 0.62 0.62 0.55 0.65 0.48 281 270 249 266 252 0.58 0.60 0.56 0.60 0.52 416 398 366 388 372 0.66 0.67 0.58 0.66 0.57 The incorporation of about 5 and 10 wt % metal oxides onto the carbon led to losses of about 12-18% and 2325% in the adsorption capacity (Table 3) and in the specific surface area (Table 1), as compared with the virgin carbon (denoted as F). It is pertinent to note that Filtrasorb-400 has a well-developed structure in the pore size range of less than 2 nm; these pores constitutes 90-95% of the total surface area, with larger pores providing access to the interior of the particles. The high adsorption capacity for small phenolic molecules, with a typical molecular radius of approximately 0.350.5 nm (as calculated from the molar volume), has been related to these pores (Montgomery, 1985). The catalytic effect of the metal oxide depended on its phase constitution. Studies by Astachov et al. (1985) showed that the Mossbauer spectrum of iron oxide prepared by impregnation of AC from nitrate salt solution and calcined with flowing helium at 200-600 °C coincide with those of highly dispersed R-Fe2O3, with crystallite dimensions smaller than 6 nm. The joint impregnation of iron(III) and copper(II) oxides has practically no effect on the dimensions of the catalyst crystallites. With crystallite dimensions of 3-6 nm, the catalyst can be housed only in mesopores and macropores. The crystallites, when located in narrow mesopores, are likely to block a part of the microporic structure of AC. This accounts for the decline in the adsorption capacity. The presence of metal oxides in the AC pores did not affect the shapes of the equilibrium adsorption isotherms, as evident by the equal slope of the three AC/catalysts. The isotherms for the virgin and two modified carbons (A, C) are plotted in Figure 3d, after rescaling the amount of adsorbed phenol (qe) with respect to the saturation adsorption capacity of that sample (Q0), showing that the three lines overlap. This observation suggests that the metal oxide crystallites are not impregnated in the micropores but only in the mesopores of the carbon. Regeneration Efficiency. The low-temperature oxidative regeneration of the AC/catalyst saturated with phenol can restore almost all of its original adsorption capacity and its surface area (Figure 4) even after 10 cycles of regeneration (carbon A). Other carbons, with approximately the same catalyst content, showed similar degrees of the recovered adsorption capacity and recovered surface areas even after five cycles. Regeneration of virgin carbon F, performed for comparison under the same conditions, restored less than 40% of the adsorption capacity per cycle, and the capacity declined from one cycle to another. Regeneration runs of spent carbons A and C were carried out using gas with different oxygen contents. No significant differences were found between the regeneration of carbons in air and in a stream of 5% oxygen in nitrogen (Figure 4b). The latter composition may be more desirable to avoid carbon ignition. A slight increase in the surface area (Figure 4b) and adsorption capacity is evident in the regenerated AC/catalyst 4378 Ind. Eng. Chem. Res., Vol. 36, No. 10, 1997 Figure 3. Batch adsorption isotherms at 25 °C: (a, top left) the three single solutes on the base carbon (F); (b, top right) phenol adsorption on the impregnated carbons A-D; (c, middle left) linear representation of data according to eq 2; (d, middle right) data rescaled with respect to Q0 of the respective carbon for fresh samples; and (e, bottom center) data rescaled with respect to Q0 for regenerated carbons after the third cycle. Regeneration carried out at 250 °C with air. sample. This is probably due to the partial oxidation of the carbon itself and due to unblocking of the carbon pores by metal oxide loss during the regenerationadsorption runs. The regeneration process did not affect the nature of adsorption: the shapes of the phenol adsorption isotherms on regenerated carbons A and C, when plotted as qe/Q0 vs Ce, did not differ after three regeneration cycles from that of the carbon before regeneration (Figure 3e). While the results for regeneration of AC preadsorbed with phenol are excellent, poor regeneration was achieved with halogenated phenols (Figure 5). The mean loss of adsorption capacity for p-chlorophenol and for p-bromophenol amounted to almost 50% per regeneration cycle. This is probably due to the formation of halides (Cl-, Br-) on the carbon, which can be poisonous to the metal oxide catalyst (Simonov, 1990), or due of the buildup of HCl or HBr on the carbon surface, which diminishes the phenolics loading. The adsorption breakthrough curves of phenol from several fresh and regenerated beds of modified AC are presented in Figure 6 and Table 4. We analyze the trends evident from these curves. The loss of phenol adsorption capacity is evident from the early breakthrough in a bed of AC/catalyst as compared with fresh Filtrasorb-400. Further loss is evident in beds that went through several regenerative cycles when compared with a freshly adsorbed bed of AC/catalyst. Breakthrough from the regenerated carbon bed of AC (with no catalyst) occurs very early. While these trends are expected from the data on batch adsorption experiments (Figure 4), the changes in the breakthrough curves seem to be more pronounced. The data listed in Ind. Eng. Chem. Res., Vol. 36, No. 10, 1997 4379 Table 4. Adsorption Bed Utilization (qb) and Capacity (qt) Calculated from Phenol Breakthrough Curves carbon F for N qb, mg/g qb/qbatch qt, mg/g qt/qbatch carbon A for N carbon A for N carbon A for N carbon A for N 0 3 0 7 10 0 3 0 3 0 3 85 0.88 88 0.9 8 0.7 9 0.8 62 0.7 73 0.84 58 0.66 62 0.7 48 0.54 50 0.54 56 0.75 58 0.77 49 0.7 50 0.7 62 0.73 67 0.78 54 0.73 55 0.74 55 0.73 56 0.75 45 0.66 48 0.7 Figure 5. Degree of recovered capacity (q/q0) vs regeneration cycle number. Regeneration of spent AC/catalysts saturated with halogenated phenols was applied at 270 °C with air; initial loadings (q0, mg per g of dry carbon) were CP-136 (A), CP-132 (C), BP-194 (A), and BP-192 (C). Figure 4. Degree of recovered capacity (q/q0) and of recovered surface area (S/S0) vs regeneration cycle number. Regeneration was applied at 250 °C with 5% oxygen (open symbols) or with air (full symbols); initial phenol loadings (q0, mg per g of dry carbon) were 97 (F), 87 (A), 85 (C), and 75 (B, D). Figure 6. Phenol breakthrough curves from fresh (F) and virgin and regenerated AC/catalyst beds. (Regeneration was applied at 250 °C with air; C0 and C denote feed and effluent concentrations.) Table 5. Metal Oxides Content on Impregnated Carbons after N Adsorption-Regenerationa Cycles metal oxide content, mg/g of C Table 4 show that the breakthrough capacity (qb) is 7080% of that in a batch adsorption unit, while the total capacity (qt) is somewhat higher. The decline in capacity with the number of regeneration cycles is similar to that observed in a batch unit. Note, however, that the breakthrough curves from the regenerated bed are steeper than those of the fresh bed, indicating improved diffusity in the regenerated bed. During successive adsorption regeneration cycles, the impregnated metal oxides content declined by a few percent per cycle due to metal losses (Table 5). The corresponding concentrations of these metal ions can be estimated from the metal loss and the filtrate specific volume per cycle (Vs). Nevertheless, in almost all cycles, the concentration of metal ions passing into the filtrate did not exceed the values of the maximal permissible concentration level. sample N A A A A A C C C C initial 1 3 7 10 initial 1 3 5 Vs, L/g 2.4 2.4 2.2 1.9 2.4 2.2 1.9 Fe2O3 CuO Cr2O3 mcat 37.0 35.5 34.5 33.7 33.5 37.0 35.6 35.1 34.8 9.0 8.8 8.7 8.5 8.4 9.0 8.9 8.8 8.8 5.0 4.9 4.8 4.7 4.6 51.0 49.2 48.0 46.9 46.5 46.0 44.5 43.9 43.6 a At 250 °C with 5% oxygen for sample A and 270 °C with air for sample C. Summary The results presented here demonstrate that destruction of phenols in dilute solutions can be achieved in two steps of adsorption on AC impregnated with oxide catalyst, followed by gas-phase catalytic oxidation at low 4380 Ind. Eng. Chem. Res., Vol. 36, No. 10, 1997 temperatures. The adsorption capacity and the surface area of the granular AC Filtrasorb-400 diminished somewhat (by 12-25%) with the impregnation of oxides (5-10 wt % Fe2O3, CuO with additivities of Cr2O3, or inert SiO2), but that did not affect the shape of the adsorption isotherms. Oxidative catalytic regeneration performed at 240-300 °C completely restored the adsorption capacity of phenol on the carbons modified with catalyst, even after 10 cycles of regeneration. Regeneration temperatures are significantly below the AC ignition point. Under similar conditions, only partial recovery of the adsorption capacity was obtained for spent carbons loaded with p-chlorophenol and pbromophenol, but future work will show that this can be overcome by a similar technology. Acknowledgment This research was supported by The Water Reseach Institute, The Center for Absorption in Science, (Ministry of Immigrant Absorption), and the Ministry of the Environment, State of Israel. Nomenclature b ) Langmuir coefficient Ce ) equilibrium concentration f ) flow rate mcat ) content of metal oxides on carbon, wt % N ) regeneration cycle number qe ) amount of solute adsorbed at equilibrium Ce Q0 ) Langmuir adsorption capacity q, q0 ) amount of solute loading for regenerated and virgin carbon, respectively qb ) breakthrough capacity, bed adsorption capacity up to exit concentration that is 1% of the feed concentration qt ) total bed adsorption capacity S, S0 ) specific surface area of regenerated and virgin carbon samples, respectively t* ) breakthrough time Vs ) filtrate specific volume (L/g of carbon) Literature Cited Astachov, M. V.; Koganovskii, A. M.; Kaninskaya, R. L.; Chashnik, A. A. Constitution of the Iron Oxide Impregnated on Active Carbon. Khim. Tekhnol. Vody. 1985, 7, 45. Crittenden, J. C.; Notthakun, S.; Hand, D. W.; Perram, D. L. Regeneration of Adsorbents Using Advanced Oxidation. U.S. Patent 5,182,030, 1993. Grant, T. M.; King, C. J. Mechanism of Irreversible Adsorption of Phenolic Compounds by Activated Carbons. Ind. Eng. Chem. Res. 1990, 29, 264. Harriott, P.; Cheng, A. T.-Y. Kinetics of Spent Activated Carbon Regeneration. AIChE J. 1988, 34, 1656. Hatchinsonet, D. H.; Robinson, C. W. A. Microbial Regeneration Process for Granular Activated Carbon- II. Regeneration Studies. Wat. Res. 1990, 24, 1217. Koganovskii, A. M.; Kaninskaya, R. L. Low-Temperature Regeneration of Activated Carbon Modified with Manganese Oxides. Khim. Tekhnol. Vody. 1981, 3, 437. Koganovskii, A. M.; Prodan, L. N. Influence of iron oxide deposited on activated carbon to phenol adsorption. Khim. Tekhnol. Vody. 1988, 10, 229. Kochetkova, R. P.; Babikov, A. F.; Shiplevskaya, L. I.; Shiverskaya, I. P.; Eppel, S. A.; Smidt, F. K. Liquid-Phase Oxidation of Phenol. Khim. Tekhnol. Topl. Masel. 1992, 4, 31. Levec, J.; Pintar, A. Catalytic Oxidation of Aqueous Solutions of Organics. An Effective Method for Removal of Toxic Pollutants from Wastewater. Catal. Today 1995, 24, 51. Matatov-Meytal, Y. I.; Sheintuch, M.; Shter, G. E.; Grader, G. S. Optimal Temperatures for Catalytic Regeneration of Activated Carbon. Carbon 1997, in press. Mishra, V. S.; Mahajani, V. V.; Joshi, J. B. Wet Air Oxidation. Ind. Eng. Chem. Res. 1995, 34, 2. Montgomery, J. M. Water Treatment: Principles and Design; Consulting Engineers, Inc., Wiley-Interscience: New York, 1985. Nelson, P. O.; Yang, M. Equilibrium Adsorption of Chlorophenols on Granular Activated Carbon. Water Environ. Res. 1995, 67, 892. Nwankwo, J.; Turk, A. Catalytic Oxidation of Adsorbed Hydrocarbons. Carbon 1975, 13, 397. Peel, R. G.; Benedek, A.; Crowe, C. M. A. Branched Pore Kinetic Model for Activated Carbon Adsorption. AIChE J. 1981, 27, 26. Prodan, L. N.; Koganovskii, A. M.; Kofanov, V. I. Catalytic Regeneration of Activated Carbon Saturated with Surface Active Agents, Dyes and Phenol. Khim. Tekhnol. Vody. 1988, 10, 278. Satterfield, C. N. Heterogeneous catalysis in industrial practice; McGraw-Hill: New York, 1991. Simonov, A. D. Catalytic Purification of Wastewaters. J. Wsesous. Khim. Obshestva D. I. Mendeleewa. 1990, 1, 102. Sonyheimer, J.; Crittenden, J. C.; Summers, R. S. Activated Carbon for Water treatment, 2nd ed.; DVGW-Forschungsstelle: Karslruhe, Germany, 1988. Suzuki, M.; Misic, D. M.; Koyama, O.; Kawazoe, K. Study of Thermal Regeneration of Spent Activated Carbons: Thermogravimetric Measurement of Various Single Component Organic Loaded on Activated Carbons. Chem. Eng. Sci. 1978, 88, 271. Tamon, H.; Saito, T.; Kishimura, M.; Okazaki, M.; Toei, R. Solvent Regeneration of Spent Activated Carbon in Wastewater Treatment. J. Chem. Eng. Jpn. 1990, 23, 426. Tan, C. S.; Liou, D. S. Regeneration of Activated Carbon Loaded with Toluene by Supercritical Carbon Dioxide. Sep. Sci. Technol. 1989, 24, 111. Turk, A. Catalytic Reactivation of Activated Carbon in Air Purification Systems. Ind. Eng Chem. 1955, 47, 966. Wang, J.-H.; Smith, J. M. Thermal Regeneration in PhenolCarbon System. AIChE J. 1985, 31, 496. Vidic, R. D.; Suidan, M. T.; Brenner, R. C. Oxidative Coupling of Phenols on Activated Carbon: Impact on Adsorption Equilibrium. Environ. Sci. Technol. 1993, 27, 2079. Yonge, D. R.; Kelnath, T. M.; Poznanska, K.; Jang, Z. P. Single Solute Irreversible Adsorption on Granular Activated Carbon. Environ. Sci. Technol. 1985, 19, 690. Received for review December 18, 1996 Revised manuscript received April 28, 1997 Accepted June 30, 1997X IE960809D X Abstract published in Advance ACS Abstracts, September 1, 1997.