Magnetocaloric Effect in 3D Gd(III)-Oxalate Coordination Framework

Abstract

Cryogenic magnetic refrigerants based on the magnetocaloric effect (MCE) hold significant potential as substitutes for the expensive and scarce He-3. Gd(III)-based complexes are considered excellent candidates for low-temperature magnetic refrigerants. We have synthesized a series of Ln(III)-based metal-organic framework (MOF) Ln-3D (Ln = Gd/Dy) by the slow release of oxalates in situ from organic ligands (disodium edetate dehydrate (EDTA-2Na) and thiodiglycolic acid). Structural analysis shows that the Ln-3D is a neutral 3D framework with one-dimensional channels connected by [Ln(H₂O)₃]³⁺ as nodes and C₂O₄²⁻ as linkers. Magnetic measurements show that Gd-3D exhibits very weak antiferromagnetic interactions with a maximum −ΔS_m value of 36.6 J kg⁻¹ K⁻¹ (−ΔS_v = 74.47 mJ cm⁻³ K⁻¹) at 2 K and 7 T. The −ΔS_m value is 28.4 J kg⁻¹ K⁻¹ at 2 K and 3 T, which is much larger than that of commercial Gd₃Ga₅O₁₂ (GGG), indicating its potential as a low-temperature magnetic refrigerant.

1. Introduction

In molecular magnets, molecule-based materials for magnetic refrigeration via the magnetocaloric effect (MCE) have garnered significant attention due to their benefits, including environmental sustainability and energy efficiency [1]. Magnetic entropy change (−ΔS_m) is a critical metric for assessing MCE performance. A large −ΔS_m value at low temperatures indicates superior performance of a cryomagnetic refrigerant [2]. High efficiency MCE requires magnetic molecules possessing negative magnetic anisotropy, large total spin ground states and low-lying excited spin states [3]. Among many metals, Gd³⁺ ion, characterized by a high spin state value (S = 7/2) and isotropic nature, makes Gd(III)-based complexes excellent candidates for magnetic refrigeration materials [4,5]. Additionally, weak magnetic interactions between Gd³⁺ ions are conductive to enhancing MCE. Consequently, the cryogenic magnetic refrigeration performance of many Gd(III)-based complexes, including 0D discrete clusters (−ΔS_m = 41.8 J kg⁻¹ K⁻¹ of Gd₂₇, −ΔS_m = 43.0 J kg⁻¹ K⁻¹ of Gd₃₂, −ΔS_m = 39.66 J kg⁻¹ K⁻¹ of Gd₃₆, −ΔS_m = 38.7 J kg⁻¹ K⁻¹ of Gd₃₇, −ΔS_m = 43.6 J kg⁻¹ K⁻¹ of Gd₄₈, −ΔS_m = 48.0 J kg⁻¹ K⁻¹ of Gd₆₀, −ΔS_m = 46.9 J kg⁻¹ K⁻¹ of Gd₁₀₄ and −ΔS_m = 38.0 J kg⁻¹ K⁻¹ of Gd₁₄₀) [6,7,8,9,10,11,12,13], 1D chains, 2D layers and 3D frameworks (such as −ΔS_m = 69.5 J kg⁻¹ K⁻¹ of Gd(CO₃)F, −ΔS_m = 76.2 J kg⁻¹ K⁻¹ of Gd(OH)F₂ etc.) [14,15,16,17,18,19,20,21,22] have been explored and reported one after another. These studies indicate that increasing magnetic density is one of the key strategies for enhancing −ΔS_m value.

Magnetic density can be improved by increasing the metal coordination number or reducing the content of non-metal components to raise the metal/ligand ratio. Low-molecular weight ligands with multiple negative charges and high coordination capacity are essential for satisfying the high coordination demands of Gd³⁺ ions and balancing the concentrated positive charges. Currently, Gd(III)-based clusters or Gd(III)-based frameworks constructed with formate, acetate, and carbonate ligands have shown excellent MCE [23,24,25], particularly the large-volume entropy (−ΔS_v) resulting from high metal/ligand ratios. Good candidates for cryogenic magnetic coolers are those complexes in which Gd³⁺ ions’ centers are sufficiently isolated from each other to maintain paramagnetic behavior even at temperatures as low as 2 K. Increasing the dimensionality of Gd-based complexes is an effective approach to enhance magnetic density while ensuring the isolation of Gd³⁺ ions from one another. Unlike these small carboxylate ligands, oxalate is a bridge-like carboxylate ligand featuring rigid coplanar and small steric hindrance. Oxalate offers various coordination modes and can display suitable configurations, forming extended structures by connecting with metal centers, and meeting the high coordination demands of Ln³⁺ ions and controllable structural dimensions of complexes. Investigations showed that the most reported oxalate-based Gd-containing complexes primarily exhibit 1D chain or 2D layer systems [26,27], while only a few cases about 3D metal-organic framework structures have been reported [28,29].

Here, a series of oxalate-based 3D frameworks Ln-3D with the formulas {[Ln(C₂O₄)_1.5(H₂O)₃]•xH₂O}_n and (Ln = Gd, x = 3, Gd-3D; Ln = Dy, x = 5, Dy-3D) were obtained using the solvent–thermal reaction of disodium edetate dehydrate (EDTA-2Na), thiodiglycolic acid and Ln(NO₃)₃. The C₂O₄²⁻ ions were derived from the in situ decomposition of EDTA-2Na and thiodiglycolic acid. Magnetic measurements show that Gd-3D exhibits a very weak paramagnetic interaction and Dy-3D displays weak antiferromagnetic interaction. Gd-3D shows a large −ΔS_m value of 36.65 J kg⁻¹ K⁻¹ (−ΔS_v = 74.47 mJ cm⁻³ K⁻¹) at 2 K and 7 T, contributing to the weak paramagnetic interaction and large metal/ligand ratio. Although the −ΔS_m value of 36.65 J kg⁻¹ K⁻¹ is slightly slower than that reported for previous Gd(III)-MOF, its −ΔS_m value of 28.37 J kg⁻¹ K⁻¹ at 2 K and 3 T is much larger than that of commercial Gd₃Ga₅O₁₂ (GGG), suggesting its potential as a low-temperature magnetic refrigeration material.

2. Materials and Methods

All materials and reagents used were obtained from commercial sources. Magnetic measurements were performed using a Quantum Design MPMSXL magnetometer (San Jose, CA, USA). Powder X-ray diffraction data were collected using a Rigaku Ultima IV powder X-ray diffractometer. The contents of C, H and N were determined using a Vario EL-3 elemental analyzer (EA, El Cajon, CA, USA). The element contents of Gd and Dy were determined using an inductively coupled plasma mass spectrometer (ICP-MS) based on the sample concentration of 20 μg mL⁻¹. Fourier transform infrared (FT-IR) spectra were performed on a VERTEX 80 FT−IR spectrophotometer (Boston, MA, USA) with KBr pellets. Thermogravimetric analysis (TGA) was carried out on a thermal analyzer (TGA-5500, TA Instruments, New Castle, DE, USA) under a N₂ flow rate of 100 L min⁻¹ with a heating rate of 3 °C min⁻¹.

Synthesis of {[Gd(C₂O₄)_1.5(H₂O)₃]•3H₂O}_n (Gd-3D). EDTA-2Na (0.3 mmol, 87.6 mg), thiodiglycolic acid (0.2 mmol, 30.0 mg), Na₂CO₃ (0.5 mmol, 52.9 mg), 2-methylimidazole (1.0 mmol, 82.1 mg) and Gd(NO₃)₃ (1.0 mL, 1.0 M) were dissolved in deionized water (7.0 mL). The resulting mixture was transferred into a polytetrafluoroethylene liner and then placed in a stainless-steel autoclave resistant to high temperature and high pressure for solvent–thermal reaction. The temperature program used is as follows: the temperature rises to 160 °C over 2 h and is maintained at 160 °C for 72 h; then it gradually cools down to room temperature over 12 h. The resulting reaction solution was filtered and evaporated at room temperature for about 1 week to yield rod-like colorless crystals. Anal. C₃H₁₆GdO₁₄: C, 9.06; H, 3.04; Gd, 7.91. Found (%): C, 8.90; H, 3.05; Gd, 8.16. The original data are shown in Figure S5. IR (KBr, cm⁻¹): 3356 (s), 1620 (s), 1418 (w), 1384 (m), 1319 (m), 1106 (m), 1053 (w), 1029 (w), 998 (w), 930 (w), 852 (w), 804 (m), 656 (w).

Synthesis of {[Dy(C₂O₄)_1.5(H₂O)₃]•5H₂O}_n (Dy-3D). The synthesis steps are the same as for Gd-3D with Gd(NO₃)₃ replaced by Dy(NO₃)₃. Anal. C₃H₁₆DyO₁₄: C, 8.21; H, 3.67; Dy, 7.41. Found (%): C, 8.78; H, 2.90; Dy, 7.70. The original data are shown in Figure S5. IR (KBr, cm⁻¹): 3356 (s), 1620 (s), 1418 (w), 1384 (m), 1319 (m), 1106 (m), 1053 (w), 1029 (w), 998 (w), 930 (w), 852 (w), 804 (m), 656 (w).

The single crystal data of compounds Gd-3D and Dy-3D were collected at 120 K using a STOE STADIVARI detector equipped with Cu K_α radiation (λ = 1.54184 Å). Absorption corrections were applied via the multi-scan program STOE LANA (2.4 8.1). The structures were solved by direct methods with hydrogen atoms of the organic ligands generated geometrically (C-H, 0.96 Å). Non-hydrogen atoms were refined anisotropically by least-squares fitting on F² using the OLEX2 software suite [30,31,32]. Detailed crystallographic data and selected bond lengths and angles are provided in Tables S1–S3. The CCDC numbers of 2401699–2401700 correspond to Gd-3D and Dy-3D, respectively, and can be accessed free of charge via http://www.ccdc.cam.ac.uk/data_request/cif (accessed on 28 November 2024).

3. Results

3.1. Synthetic Strategy

The anion-templated method is an effective approach for synthesizing lanthanide clusters. In the process of cluster formation, anions not only act as templates to induce the formation of cluster framework, but also act as counterions to balance the excessive positive charges caused by the aggregation of Ln³⁺ ions, thereby further stabilizing the lanthanide clusters. Particularly, the slow-release anion-templated method, by controlling the rate of anion template formation, facilitates the formation of high-nuclearity metal clusters or high-dimensional cluster-based frameworks. Ligand in situ decomposition is a common method of slow-release anion templating. In this study, although the raw materials EDTA-2Na and thiodiglycolic acid do not directly participate in coordination, the C₂O₄²⁻ anions slowly generated through their decomposition under high temperature and pressure act as the sole ligand binding with Ln³⁺ anions. When only H₂C₂O₄ or Na₂C₂O₄ is used in the reaction with Ln³⁺ ions without the presence of EDTA-2Na and thiodiglycolic acid, the target product cannot be obtained. However, when EDTA-2Na, thiodiglycolic acid and Na₂C₂O₄ are present simultaneously, the yield of the target product is significantly improved. Therefore, EDTA-2Na and thiodiglycolic acid also function as buffers during the reaction process.

3.2. Crystal Structures

The compounds Gd-3D and Dy-3D are isostructural, both crystallizing in a trigonal crystal system and belonging to the R-3 space group. Here, Gd-3D is a representative example and its structure will be described in detail. Single-crystal structural analysis shows that Gd-3D is a neutral 3D metal framework {[Gd(C₂O₄)_1.5(H₂O)₃]•3H₂O}_n. As shown in Figure 1a, the asymmetric unit [Gd(C₂O₄)_1.5(H₂O)₃] consists of one Gd³⁺ ion, 1.5 C₂O₄²⁻ anions and three coordinating water molecules. Single-crystal analysis, combined with EA and TGA data (Figure S1), reveals that each asymmetric unit contains three free water molecules. In Gd-3D, C₂O₄²⁻ acts both as counter-anions and as linkers, connecting Gd³⁺ ions in the mode of μ₄: η¹: η¹: η¹: η¹ (Figure 1b). Figure 1c shows that Gd³⁺ ion has a nine-coordination with tricapped trigonal prism geometry, surrounded by three O atoms from three coordinating water molecules and six O atoms from three C₂O₄²⁻ anions. The bond length of Gd-O ranges from 2.377(11) to 2.536(11) Å, consistent with the previously reported values for oxalate-based lanthanide complexes [26,27]. The compound forms a 3D framework with a 1D channel, constructed from [Gd(H₂O)₃]³⁺ as nodes and C₂O₄²⁻ as linkers (Figure 1d). The 3D neutral metal framework not only enhances the dimensionality of the compounds but also eliminates free counter-ions, supporting an improved metal/ ligand ratio.

3.3. Magnetic Properties

Powder samples of Gd-3D and Dy-3D were used to measure their magnetic properties. The sample purity was confirmed by powder X-ray diffraction (PXRD). As shown in Figure S2, the experimental PXRD patterns closely match the simulated ones derived from single-crystal data. Magnetic susceptibility measurements of Gd-3D and Dy-3D were conducted in the temperature range of 2–300 K under a direct current field of 1000 Oe (Figure 2a). At 300 K, the obtained experimental χ_MT values of 7.58 cm³ K mol⁻¹ for Gd-3D and 14.10 cm³ K mol⁻¹ for Dy-3D closely align with the calculated values of 7.87 cm³ K mol⁻¹ for uncoupled Gd³⁺ (g = 2, J = 7/2) and 14.16 cm³ K mol⁻¹ for uncoupled Dy³⁺ (g = 4/3, J = 15/2), respectively. For Gd-3D, the χ_MT value remained almost constant as the temperature decreased to 10 K, and then gradually decreased to 6.9 cm³ K mol⁻¹ at 2 K, indicating an extremely weak magnetic interaction. For Dy-3D, the χ_MT value was almost unchanged from 300 K to 100 K and then slowly decreased to a minimum value of 10.07 cm³ K mol⁻¹ as the temperature continued to decrease to 2 K, suggesting a weak antiferromagnetic interaction. The χ_M⁻¹ vs. T curves of Gd-3D and Dy-3D in the range of 2–300 K were fitted based on the Curie–Weiss law (Figure 2b), with Curie constants of 7.68 cm³ mol⁻¹ for Gd-3D and 14.27 cm³ mol⁻¹ for Dy-3D and Weiss constants of −0.09 for Gd-3D and −2.84 for Dy-3D. The near-zero Weiss constant of −0.09 confirms the extremely weak antiferromagnetic interaction in Gd-3D, which is conductive to the improvement of MCE. For Dy-3D, the negative Weiss constant of −2.84 verifies its antiferromagnetic interaction and zero-field splitting effect.

The field-dependent magnetizations of Gd-3D and Dy-3D were measured in 0–7 T at different temperatures. The magnetization value of Dy-3D was 5.75 Nμ_B at 2 K and 7 T, which is significantly lower than the theoretical value of 10.00 Nμ_B. Additionally, the magnetization curves of Dy-3D at 2 K, 5 K and 7 K do not overlap into a single curve, further confirming the strong magnetic anisotropy and zero-splitting effect in Dy-3D (Figure 3). However, the results of alternating current magnetic susceptibilities show that Dy-3D shows no frequency-dependent behavior (Figure S3). In contrast, the obtained magnetization value of Gd-3D stabilized at approximately 6.76 Nμ_B at 2 K and 7 T, closely matching the theoretical saturation value of 7.00 Nμ_B, indicating the isotropic nature and extremely weak antiferromagnetic interaction of Gd-3D (Figure 4).

The weak magnetic interactions and high metal-to-ligand ratio in Gd-3D indicate its potential as an excellent candidate for magnetic refrigeration applications. The −ΔS_m value is a key parameter in assessing MCE, and it can be estimated from the temperature-dependent magnetizations at 2–7 T using the Maxwell equation ∆S_m(T)∆H = ∫[∂M(T, H)/∂T]_HdH [33]. As shown in Figure 4b, the maximum −ΔS_m value of Gd-3D is 36.67 J kg⁻¹ K⁻¹ at 2 K and 7 T, and the corresponding volume entropy is 74.47 mJ cm⁻³ K⁻¹. Compared with previously reported oxalate-based magnetic coolers (Table 1), the −ΔS_m value of Gd-3D is slightly lower than that of the 2D [Gd₂(C₂O₄)₃(H₂O)₆•0.6H₂O] (−ΔS_m = 46.6 J kg⁻¹ K⁻¹) [26], [Gd(C₂O₄)(H₂O)₃Cl] (−ΔS_m = 48.0 J kg⁻¹ K⁻¹) [27], and 3D [Gd(C₂O₄)_0.5(CO₃)H₂O] (−ΔS_m = 50.7 J kg⁻¹ K⁻¹) [29]. Gd-3D exhibits weak antiferromagnetic interactions, while in 3D [Gd(C₂O₄)_0.5(CO₃)H₂O] [29], the mixed bridging of CO₃^2- and C₂O₄^2- between Gd³⁺ ions leads to weak ferromagnetic interactions. For 2D [Gd₂(C₂O₄)₃(H₂O)₆•0.6H₂O] [26], the planar structure formed by C₂O₄²⁻ bridging Gd³⁺ ions ensures good isolation between magnetic centers within the lattice, resulting in weak paramagnetic behavior. These features are beneficial for enhancing the MCE. The −ΔS_m value of Gd-3D is comparable to that of 3D (H₆edte)_0.5[Gd(C₂O₄)₂(H₂O)] [28]. By comparison, both Gd-3D and 3D (H₆edte)_0.5[Gd(C₂O₄)₂(H₂O)] contain C₂O₄²⁻-bridged Gd³⁺ ions, and 3D (H₆edte)_0.5[Gd(C₂O₄)₂(H₂O)] has a Weiss constant of −0.09 K [28], almost identical to that of Gd-3D, showing weak antiferromagnetic interactions in both. Additionally, the crystal density is related to the metal content. The volume entropy (−ΔS_V) is determined by the −ΔS_m and crystal density. As shown in Table 1, the −ΔS_V value of Gd-3D at 2 K, 7 T is 74.47 mJ cm⁻³ K⁻¹, which is lower than that of the reported Gd-based framework materials [25,26,28,29]. This reduction may result from the presence of three free guest water molecules in Gd-3D, leading to decreased crystal density and magnetic density.

Relative cooling power (RCP) is another critical parameter used for evaluating magnetocaloric performance, which is calculated as the product of the maximum −ΔS_m value and the corresponding half-peak width [34]. RCP at low temperatures and low fields is particularly significant for practical applications. At 2 K, the −ΔS_m value of Gd-3D rapidly increases to 20.37 J kg⁻¹ K⁻¹in the field range from 0.5 to 2.0 T with a corresponding RCP of 68.45 J kg⁻¹, comparable to that of the commercial Gd₃Ga₅O₁₂ (GGG) at 1.2 K and 2 T [1]. The −ΔS_m value of Gd-3D rises to 28.37 J kg⁻¹ K⁻¹at 2 K and 3 T, exceeding that of GGG in the same field [35,36]. The RCP of Gd-3D at 2 K and 3 T is 112 J kg⁻¹, which is slightly higher than that of other Gd-based complexes but significantly lower than that of GGG. This discrepancy may be attributed to the higher non-metallic content in Gd-3D compared to GGG. Nonetheless, the RCP of 198 J kg⁻¹for Gd-3D at 2 K and 7 T indicates that it remains a potential cryogenic refrigerant.

Table 1. Comparison of parameters of selected Gd-based frameworks and other refrigerant materials in the cryogenic zone.

Compound	H (T)	T (K)	−ΔSm (J kg−1 K−1)	−ΔSv (mJ cm−3 K−1)	RCP (J kg−1)	Ref.
Gd-3D	1	2	8.32	16.91	23.71	this work
Gd-3D	2	2	20.22	41.09	68.45	this work
Gd-3D	3	2	28.37	57.65	112.01	this work
Gd-3D	4	2	31.96	64.94	146.25	this work
Gd-3D	5	2	33.44	67.95	171.17	this work
Gd-3D	6	2	35.27	71.67	188.99	this work
Gd-3D	7	2	36.67	74.47	198.25	this work
3D (NH4)[Gd(C2O4)(SO4)(H2O)]	7	2	42.4	137	―	[29]
3D [Gd(C2O4)0.5(CO3)(H2O)]	7	2	50.7	165	―	[29]
3D (H6edte)0.5[Gd(C2O4)2(H2O)]	9	2	36.8	84.4	―	[28]
2D [Gd2(C2O4)3(H2O)6•0.6H2O]	7	2	46.6	139.9	―	[26]
2D [Gd(C2O4)(H2O)3Cl]	7	2.2	48.0	144	―	[25]
Gd3Ga5O12 (GGG)	3	1.2	24	173	―	[35,36]
Gd(HCOO)3	2	1.1	43.6	168.5	135.47	[23]
[{Gd(OAc)3(H2O)2}2]•4H2O	2	0.9	32.6	66.5	104.41	[37]
[Gd(HCOO)(OAc)2(H2O)2]	2	0.9	37.0	88.9	118.6	[24]
Gd2(fum)3(H2O)4•3H2O	2	1.0	18.0	45.3	43.2	[16]
Gd3Ga5O12 (GGG)	2	1.2	20.4	145	67.58	[1]

H₄edte = N,N,N’,N’’-tetrakis(2-hydroxyethyl)ethlenediamine; fum = fumarate.

Table 1. Comparison of parameters of selected Gd-based frameworks and other refrigerant materials in the cryogenic zone.

Compound	H (T)	T (K)	−ΔSm (J kg−1 K−1)	−ΔSv (mJ cm−3 K−1)	RCP (J kg−1)	Ref.
Gd-3D	1	2	8.32	16.91	23.71	this work
Gd-3D	2	2	20.22	41.09	68.45	this work
Gd-3D	3	2	28.37	57.65	112.01	this work
Gd-3D	4	2	31.96	64.94	146.25	this work
Gd-3D	5	2	33.44	67.95	171.17	this work
Gd-3D	6	2	35.27	71.67	188.99	this work
Gd-3D	7	2	36.67	74.47	198.25	this work
3D (NH4)[Gd(C2O4)(SO4)(H2O)]	7	2	42.4	137	―	[29]
3D [Gd(C2O4)0.5(CO3)(H2O)]	7	2	50.7	165	―	[29]
3D (H6edte)0.5[Gd(C2O4)2(H2O)]	9	2	36.8	84.4	―	[28]
2D [Gd2(C2O4)3(H2O)6•0.6H2O]	7	2	46.6	139.9	―	[26]
2D [Gd(C2O4)(H2O)3Cl]	7	2.2	48.0	144	―	[25]
Gd3Ga5O12 (GGG)	3	1.2	24	173	―	[35,36]
Gd(HCOO)3	2	1.1	43.6	168.5	135.47	[23]
[{Gd(OAc)3(H2O)2}2]•4H2O	2	0.9	32.6	66.5	104.41	[37]
[Gd(HCOO)(OAc)2(H2O)2]	2	0.9	37.0	88.9	118.6	[24]
Gd2(fum)3(H2O)4•3H2O	2	1.0	18.0	45.3	43.2	[16]
Gd3Ga5O12 (GGG)	2	1.2	20.4	145	67.58	[1]

H₄edte = N,N,N’,N’’-tetrakis(2-hydroxyethyl)ethlenediamine; fum = fumarate.

4. Conclusions

In this study, we have investigated the influence of magnetic interactions and magnetic density on the MCE of Gd-3D. The results of magnetic susceptibilities indicate that Gd-3D exhibits very weak antiferromagnetic interactions, which is beneficial for enhancing MCE. The −ΔS_m value of Gd-3D is 36.67 J kg⁻¹ K⁻¹, which is slightly lower than that of the reported oxalate-based Gd(III) complexes, attributed to the presence of additional free water molecules reducing its magnetic density. However, finding low-temperature magnetic refrigerants with an excellent MCE at low fields is even more critical. Compound Gd-3D displays a −ΔS_m value of 28.37 J kg⁻¹ K⁻¹ at 2 K and 3 T, surpassing that of commercial GGG, indicating that it still meets requirements as a potential low-temperature magnetic refrigerant.