Song-De
Han
,
Xiao-Hong
Miao
,
Sui-Jun
Liu
and
Xian-He
Bu
*
Department of Chemistry, TKL of Metal- and Molecule-Based Material Chemistry and Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), Nankai University, Tianjin 300071, P.R. China. E-mail: buxh@nankai.edu.cn; Fax: +86-22-23502458
First published on 24th June 2014
Two (3,12)-connected complexes with distorted cubic [Ln4O4] (Ln = Gd (1), Dy (2)) building units were synthesized. The magnetic studies reveal that 1 features a large magnetocaloric effect with −ΔSmaxm = 51.29 J kg−1 K−1, and 2 displays slow magnetic relaxation behavior.
To design molecular magnetic coolers with remarkable MCE, GdIII with seven unpaired 4f electrons is a promising candidate as the intrinsic nature of GdIII fulfils the requirements for improving the MCE well, such as a large spin ground state S, negligible magnetic anisotropy (Dion = 0) and low-lying excited spin states.4 Furthermore, due to the efficient shielding of the 4f orbitals of the GdIII ion, the magnetic interactions of GdIII–GdIII/3d–GdIII are usually anticipated to be weak, which is also helpful to promote the MCE. Hence, the assembly of GdIII/3d–GdIII with light or polydentate ligands is a feasible strategy to achieve the aforementioned target, which has been well corroborated in recent publications.5,6 To our knowledge, most of the reported Gd-based complexes with significant MCE are discrete [Gdn]4a,5 and [3d–Gd]6 (3d = Cr, Mn, Fe, Co, Ni, Cu) clusters and a few cases are low dimensional Gd-based coordination polymers.4a,7 Research into the MCE in 3D coordination polymers is still rare.8 The potential advantages of the 3D framework are not only that the adjacent metal ions/metal clusters share the bridging ligands, which can enhance the magnetic density and result in a large MCE, but also that they have relatively higher thermal and/or solvent stabilities than discrete molecular clusters, which provides a solid foundation for future applications.3e,f,8b,c However, to date, most of the reported Gd-based complexes with significant MCE from 0D discrete clusters to 3D frameworks are constructed via organic ligands,5,6 and cases fabricated by inorganic ligands are very rare.8c,i In fact, some inorganic ligands like sulfate, phosphate, carbonate are also polydentate. For example, the tetradentate sulfate with rich coordination modes preferably bonds to lanthanide ions, has been widely utilized to synthesize metal–organic frameworks (MOFs).9 Moreover, weak exchange interactions and relatively high density may be expected in Gd-containing complexes bridged by sulfate.9b,10 Both of these characteristics are helpful in enhancing the performance in MCE, which was well demonstrated by the early studied molecular magnetic cooler [Gd2(SO4)3·8H2O].11
On the other hand, Dy clusters are being intensively explored in single-molecule magnets (SMMs) due to the large single-ion magnetic anisotropy of DyIII. Hitherto, various [Dyn] clusters with slow magnetic relaxation behavior have been reported, including Dy2,12a Dy3,12b Dy4,12c Dy5,12d Dy6,12e Dy7,12f Dy8,12g Dy9,12h Dy10,12i Dy11,12j Dy12,12k Dy14,12l Dy24.5b However, observations of such behavior in 2D or 3D complexes utilizing Dy clusters as nodes remain scarce.8c,13 Therefore, it is interesting and worthwhile to investigate the magnetic properties of high dimensional complexes based on Dy clusters as nodes.
Herein, we report the synthesis, structures, and magnetic properties of two (3,12)-connected sulfate-based lanthanide complexes [Ln4(SO4)4(μ3-OH)4(H2O)n] (Ln = Gd, n = 4 for 1, and Ln = Dy, n = 3 for 2). Magnetic measurements demonstrate that complex 1 displays a magnetic entropy change −ΔSmaxm = 51.29 J kg−1 K−1, being the third largest value to date, and simultaneously presents high solvent stabilities. Complex 2 exhibits slow magnetic relaxation behavior at low temperature, which is rare in 2D or 3D complexes based on Dy clusters as nodes. Colorless crystals of 1 and 2 were obtained by the solvothermal reaction of Hbms·HCl (Hbms = (1H-benzimidazol-2-yl)methanethiol), and Ln(NO3)3·6H2O in a CH3OH–H2O mixed solvent at 140 °C. It is notable that the sulfate is generated in situ due to the oxidation of thiol under acid conditions in the presence of NO3−. The in situ transformation of thiol to sulfate has been documented in a recent publication.14 The SO42− in 1 and 2 can also be unequivocally demonstrated by the IR characteristic peaks at about 1100 cm−1 (Fig. S12†).
Two similar complexes (Ln = Y, Er) with differences only in the terminal ligands on the GdIII ions for 1 have been previously reported.15 Therefore, the structure of 1 is discussed briefly here. Complex 1 crystallizes in the orthorhombic space group P212121 and contains hydroxyl bridged tetranuclear units [Gd4(μ3-OH)4(H2O)4]8+ ([Gd4], for short), which are further connected via SO42− ligands to generate a 3D framework. Each GdIII connects three neighboring GdIII ions with four μ3-OH− bridging ligands, forming a distorted cubic [Gd4] unit, in which four GdIII ions are located at the corners of the tetrahedron. Each SO42− bridges three [Gd4] units, and each [Gd4] unit links twelve SO42− groups (Fig. 1a and 1b). Thus the SO42− and [Gd4] cluster can be considered as 3-connected and 12-connected nodes, respectively, giving rise to a unique (3,12)-connected topological network with the point (Schläfli) symbol of (420·628·818)(43)4 calculated by TOPOS (Fig. 1c).16 To our knowledge, several (3,12)-connected frameworks have been reported, while only four examples are based on multinuclear lanthanide clusters.8c,14,17
Both 1 and 2 are stable in air, confirmed by their simulated and experimental PXRD (powder X-ray diffraction) patterns (Fig. S1†). Importantly, the simulated PXRD pattern of 1 from single crystal data reaches good agreement with the corresponding experimental ones after 1 is soaked in common solvents, such as H2O, CH3OH, CH3CH2OH, CH3CN, CH2Cl2, acetone, DMF, tetrahydrofuran, and cyclohexane solution for 48 hours (Fig. S1†), which notably shows that 1 possesses excellent and extensive solvent stability.
The phase purity of complexes 1 and 2 was confirmed by PXRD patterns (Fig. S1 and S2†). Variable-temperature magnetic susceptibility measurements were investigated on polycrystalline samples of complexes 1 and 2 with an applied dc field of 1000 Oe (Fig. S3†). The observed χMT products of 31.61 and 55.73 cm3 mol−1 K at 300 K for 1 and 2 are in good agreement with the theoretical values for the unit of four non-interacting GdIII (31.50 cm3 K mol−1, 8S7/2, g = 2) in 1 and four isolated DyIII (56.67 cm3 K mol−1, 6H15/2, g = 4/3) in 2. Upon cooling, the χMT value of 1 stays essentially constant until approximately 25 K, while the χMT value of 2 gradually decreases from 300 K to 50 K, followed by an obvious decrease to the minimum value of 19.74 cm3 K mol−1 for 1 and 38.88 cm3 K mol−1 for 2 at 2 K. The decline of the curve for 1 indicates that antiferromagnetic coupling exists between adjacent GdIII ions, which is further corroborated by the negative Weiss constant θ = −1.57 K from Curie–Weiss fitting (Fig. S4†).
Magnetization measurements were investigated in the range 0–7 T at 2–10 K for 1 and 0–7 T at 2 K for 2 (Fig. S5 and S6†). The plots of M versus H display a steady increase with the increasing field. For 1, M reaches a value of 28.30Nβ at 2 K and 7 T, which agrees with the theoretical value of 28Nβ for four GdIII (g = 2, S = 7/2). For 2, the M value is 25.20Nβ at 2 K and 7 T, far from the theoretical saturated value of 40Nβ for four DyIII (g = 4/3, J = 15/2). The magnetic unsaturation even at 7 T may be attributed to the magnetic anisotropy and/or low-lying excited states of DyIII, which is supported by the non-superposition of magnetization curves at different temperatures (Fig. S7†). The beautiful superposition of the field-cooled (FC) curve and the zero-field-cooled (ZFC) curve preclude the existence of long-range magnetic ordering above 2 K in 1 (Fig. S8†).
The weak magnetic interactions and large metal/ligand mass ratio make complex 1 a promising candidate for low-temperature magnetic cooling, as magnetic entropy change ΔSm, a key parameter in evaluating the MCE, can be derived by applying the Maxwell equation ΔSm(T)ΔH = ∫[∂M(T,H)/∂T]HdH to the experimentally obtained magnetization data.5–8 The entropy changes at various magnetic fields and temperatures are summarized in Fig. 2, with an impressive −ΔSmaxm of 51.29 J kg−1 K−1 for T = 2 K and ΔH = 7 T. The value of −ΔSmaxm is smaller than the value of 59.96 J kg−1 K−1 for four uncoupled GdIII (judged by 4Rln(2S + 1), where R is the gas constant and S is the spin state). The gap between the experimental data and the theoretical value mainly originates from the intracluster antiferromagnetic interactions in 1. Among reported molecule-based magnetic cryogens, magnetic entropy changes −ΔSmaxm above 40.0 J kg−1 K−1 are limited, as listed in Table S1.† It is notable that only three cases with magnetic entropy change −ΔSmaxm above 50.0 J kg−1 K−1 have been reported. It is still very competitive when considered from the volumetric aspect with −ΔSmaxm = 198.85 mJ cm−3 K−1, being among the highest throughout the GdIII-containing complexes (Table S1†). Indeed, evaluating molecule-based magnetocaloric materials from a volumetric aspect is more meaningful for practical application.18 The large MCE per unit mass and/or per unit volume of 1 is primarily ascribed to the high magnetic density and the small linker between the clusters. Furthermore, the −ΔSm per unit mass reaches a value of 42.63 J kg−1 K−1 for T = 2 K and ΔH = 40 kG, which is larger than for most reported GdIII-based complexes.4–7,8a–g
To further explore the magnetic dynamics of 2, the frequency and temperature dependencies of the alternating current (AC) susceptibilities were collected under a zero direct current (DC) field and a 3 Oe AC magnetic field (Fig. 3 and Fig. S9†). The out-of-phase constituent of the AC susceptibilities exhibits frequency dependent signals below 15 K, suggesting the slow magnetic relaxation behaviour of 2, which might be an indication of SMM behaviour. However, similar to some reported Dy-based complexes,19 the peaks of the out-of-phase signals were not observed in the technically available temperature range due to fast quantum tunneling. To reduce the quantum tunneling effect, a proper DC field needs to be exerted. The optimal field was found from field-dependent AC susceptibility measurements (0–5 kOe). As shown in Fig. S10,† the peak of χ′′M appeared at ca. 800 Oe. Thus, an 800 Oe DC field was exerted. However, the peak of χ′′M still could not be observed because the strong quantum tunneling effect was not effectively suppressed (Fig. S11†).
Fig. 3 Frequency dependence of the out-of-phase (χ′′) ac susceptibility components for 2 at zero dc field. |
Footnote |
† Electronic supplementary information (ESI) available. CCDC 981569 and 981570. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c4qi00073k |
This journal is © the Partner Organisations 2014 |