High-Nuclearity Lanthanide-Containing Clusters as Potential

Feb 2, 2018 - Furthermore, we demonstrated using ample examples that the presence of small anions as templates is essential to the assembly of high-nu...
2 downloads 11 Views 5MB Size
Download PDF
Article Cite This: Acc. Chem. Res. 2018, 51, 517−525

pubs.acs.org/accounts

High-Nuclearity Lanthanide-Containing Clusters as Potential Molecular Magnetic Coolers Xiu-Ying Zheng,† Xiang-Jian Kong,*,† Zhiping Zheng,*,‡,§ La-Sheng Long,*,† and Lan-Sun Zheng† †

Collaborative Innovation Center of Chemistry for Energy Materials, State Key Laboratory of Physical Chemistry of Solid Surface and Department of Chemistry, College of Chemistry and Chemical Engineering, Xiamen University, Xiamen 361005, China ‡ Department of Chemistry and Biochemistry, University of Arizona, Tucson, Arizona 85721, United States § Department of Chemistry, Southern University of Science and Technology, Shenzhen 518055, China CONSPECTUS: High-nuclearity cluster-type metal complexes are a unique class of compounds, many of which have aesthetically pleasing molecular structures. Their interesting physical and chemical properties arise primarily from the electronic and/or magnetic interplay between the component metal ions. Among the extensive studies in the past two decades, those on lanthanide-containing clusters, lanthanide-exclusive or heterometallic with transition metal elements, are most notable. The research was driven by both the synthetic challenges for these generally elusive species and their intriguing magnetic properties, which are useful for the development of energy-efficient and environmentally friendly magnetic cooling technologies. Our efforts in this vein have been concentrated on developing rational synthetic methods for high-nuclearity lanthanide-containing clusters. By means of the now widely adopted approach of “ligand-controlled hydrolysis” of lanthanide ions, a great variety of cluster-type lanthanide hydroxide complexes had been prepared in the first half of this developing period (1999−2006). In this Account, our efforts since 2007 are summarized. These include (1) further development of synthetic strategies in order to expand the ligand scope and/or to increase the nuclearity (>25) of the cluster species and (2) magnetic studies pertinent to the pursuit of materials with a large magnetocaloric effect (MCE). Specifically, with the hope of expanding the family of ligands and producing clusters of previously unknown structures, we tested under hydrothermal or solvothermal conditions the use of readily available yet not commonly used ligands for controlling lanthanide hydrolysis; such ligands, carboxylates as mundane examples, tend to form insoluble complexes prior to any possible hydrolysis. We have also validated the use of preformed transition metal complexes as metalloligands for subsequent control of lanthanide hydrolysis toward heterometallic 3d−4f clusters. Furthermore, we demonstrated using ample examples that the presence of small anions as templates is essential to the assembly of high-nuclearity lanthanide-containing clusters and that maintaining a low concentration of the anion template(s) is a key to such success. It has been found that slow production/release of such anion templates by in situ ligand decomposition or absorption of atmospheric CO2 is effective in preventing precipitation of their lanthanide salts, allowing not only controllable lanthanide hydrolysis but also gradual and modular assembly of the giant cluster species. Magnetic studies targeting potential applications of such clusters as molecular magnetic coolers have also been conducted. The results are summarized in the second portion of this Account in an effort to establish a certain magneto−structure relationship. Of particular relevance is the possible correlation between MCE (evaluated using the isothermal magnetic entropy change, −ΔSM) and magnetic density, and the intracluster antiferromagnetic exchange coupling. We have also made some preliminary attempts at preparing processable and practically useful materials in the form of a monodisperse core−shell nanostructure. We succeeded in encapsulating a single nanosized heterometallic molecular cluster in a nanoshell of silica. It was found that such passivation not only helped stabilize the cluster but also reduced the magnetic interactions between individual clusters. These effects are reflected in the slightly enhanced value of −ΔSM for the core−shell composite over the parent unprotected cluster.

1. INTRODUCTION

magnetically achieving a cooling effect, commonly called

The cryogenic magnetocaloric effect (MCE) is a magnetothermodynamic phenomenon first discovered in 1881 when Warburg observed under adiabatic conditions a reversible temperature change in a piece of iron exposed to a changing magnetic field.1 Debye and Giauque suggested its first practical useadiabatic demagnetizationin reaching temperatures lower than that of liquid helium, which had been the lowest experimentally achievable temperature.2,3 This method of

magnetic refrigeration, has been heralded as a next-generation

© 2018 American Chemical Society

cooling technology that is more energy-efficient and environmentally friendly than the current vapor-cycle refrigeration technology.4 Received: November 16, 2017 Published: February 2, 2018 517

DOI: 10.1021/acs.accounts.7b00579 Acc. Chem. Res. 2018, 51, 517−525

Article

Accounts of Chemical Research

Figure 1. Anion-templated modular assembly of multicubane cluster cores. Ln (violet), O (red), Cl (bright green), I (rose red), C (light gray), hereinafter the same.

Giant MCEs were first observed in alloy-type solids and magnetic nanoparticles.4 However, the past decade has witnessed extensive research on molecular materials with giant MCEs due largely to the flexibility in structural control and ease of introducing various magnetic interactions within the molecular structure.5 Cluster-type complexes of Fe3+ and Mn2+ were first studied for their high ground-state spin values.6 However, strong magnetic couplings between such transition metal ions generally lead to small MCEs. Recent studies have concentrated on high-nuclearity Gd(III) clusters, as Gd(III) with its orbitally isotropic f7 electron configuration possesses the electronic and magnetic characteristics propitious for achieving giant MCEs.7 Research to date has revealed in these substance certain shared structural and functional features, including a large metal/ligand mass ratio to ensure a high magnetic density, a high-spin ground state, low-lying excited spin states, negligible magnetic anisotropy, and weak magnetic exchange coupling. 8 Toward the same goal, heterometallic transition metal−Gd(III) clusters have also been studied, in which the mitigating effect of Gd(III) ion on the typically strong magnetic coupling between transition metal ions has been demonstrated.9 How to incorporate a large number of Gd(III) ions into a single complex of high symmetry with a large metal/ligand ratio appears to be the key to the development of molecular magnetic coolers. The exploratory synthesis started about two decades ago when Zheng and co-workers obtained a series of lanthanide hydroxide clusters featuring a common cubane-like [Ln4(μ3‑OH)4]8+ unit (Figure 1) by using the now widely adopted approach of “ligand-controlled hydrolysis”.10−13 We found that the high-pH conditions and the use of multidentate ligands are essential for the assembly of the clusters and that small anions, presumably serving as templates, are present in almost all higher-nuclearity clusters that can be formally built up from smaller cluster motifs.14 Building on these initial successes,15 we and others have been pushing forward the synthetic frontiers as well as studying their magnetic properties with an eye on their potential applications in achieving cooling by magnetic means.16,17 The synthetic

efforts involve expanding the family of ligands to include transition metal complexes as ligands to support the assembly of the cluster complexes and exploring specific reaction conditions for the chosen ligand(s). Of particular note is the demonstration of the critical roles played by small anions in supporting the assembly of giant clusters. Using ample examples, we demonstrate the effective application of multiple anions, mixed or of the same kind, and slow release/production of anions in the templated assembly of giant lanthanidecontaining clusters. Although this Account focuses on the discussion of our own work, due credit is given to other researchers where appropriate.

2. STRATEGY FOR THE SYNTHESIS OF HIGH-NUCLEARITY LANTHANIDE-CONTAINING CLUSTERS The assembly of lanthanide-containing clusters is a complex process that is sensitively dependent on a number of factors. These include the nature of the hydrolysis-limiting ligands, the ligand/metal ratio, possible involvement of anion templates, pH, temperature, and duration.18 Despite the efforts made and progress achieved, the number of giant clusters (nuclearity >25) remains small.19 According to the structural characteristics of a representative cluster (Figure 2), we focused our synthetic strategy on the slow and templated buildup of the cluster core.

Figure 2. Common structural features of a lanthanide cluster complex with templating anion(s) and supporting ligands. 518

DOI: 10.1021/acs.accounts.7b00579 Acc. Chem. Res. 2018, 51, 517−525

Article

Accounts of Chemical Research

Figure 3. Ball-and-stick views of the clusters (a) Gd24Zn4, (b) Ln104, and (c) Ln27.

Figure 4. Ball-and-stick views of the clusters (a) Dy24Mn2, (b) Gd12Fe14, and (c) Ln54.

cohydrolysis of Zn(CH3COO)2 and Gd(ClO4)3.20 The cationic cluster core is formally constructed using four octahedral [Gd6(μ6‑O)(μ3‑OH)8(H2O)12]8+ units and one tetrahedral [Zn4(μ4‑O)]6+ unit, which are joined by 12 μ3‑OH groups and 12 CH3COO− ligands (Figure 3a). The success of using this small and simple carboxylate ligand culminated when we isolated the largest known lanthanide-exclusive clusters, [Ln104(ClO4)6(CH3COO)56(μ3‑OH)168(μ4‑O)30(H2O)112]·(ClO4)22 (2, Ln104, Ln = Gd, Nd), consisting of 24 square-pyramidal [Ln5(μ3‑OH)4(μ4‑O)]9+ units, by hydrolysis of Ln(ClO4)3 in the presence of acetate (Figure 3b).21 High-nuclearity clusters supported by other simple carboxylates include cagelike [(ClO 4)@Ln 27(μ3‑OH) 32 (CO3 ) 8(CH3CH2COO)20(H2O)40]·(ClO4)12 (3, Ln = Gd, Dy) (Figure 3c), 22 {[Gd 36 (NA) 36 (OH) 49 (O) 6 (NO 3 ) 6 (N 3 ) 3 (H2O)20]·Cl2}n (4, Gd36),23 [Gd38(μ-O)(μ8-ClO4)6(μ3‑OH)42(CAA)37(H2O)36(EtOH)6]·(ClO4)10·(OH)17 (5, Gd38),24 and heterometallic [Gd24Cu36(OH)72(NO3)6(O2CPh)60(MeOH)6(H2O)12]·(NO3)6 (6, Gd24Cu36),25 in which propionate, nicotinate (NA), chloroacetate (CAA), and benzoate serve as the supporting ligands, respectively. Polyalcohols are another class of O-donor ligands for the construction of high-nuclearity lanthanide clusters. However, their use in this capacity has largely been overlooked. Our attempts to utilize polyalcohols as supporting ligands for lanthanide hydrolysis under hydro/solvothermal conditions produced some positive results. With myo-inositol (cyclohexanehexol, H6L) and 1,2,3-cyclohexanetriol (H3L′) as ligands, the heterometallic cluster [Dy 24 Mn 2 (OH) 8 (CH3COO)12(H2L)6(H3L)6(H2O)51]·[Dy(H2O)9]·(ClO4)29 (7, Dy24Mn2) (Figure 4a)26 and the double-cage-like structure [(CO 3 ) 2 @Ln 37 (H 3 L′) 8 (CH 3 COO) 21 (CO 3 ) 12 (μ 3 ‑OH) 41 (μ2‑H2O)5(H2O)40]·(ClO4)21 (8, Ln37, Ln = Gd, Dy) were obtained, respectively.27 Recently, the largest 4f−Fe cluster, [Gd12Fe14(μ3‑OH)12(μ4‑OH)6(μ4‑O)12(TEOA)6(CH3COO)16(H2O)8]·(CH3COO)2 (9, Gd12Fe14), was reported with the use of acetate and tris(2-hydroxyethyl)amine (TEOA), a

Specifically, analogous to the growth of larger and higherquality single crystals by a slow crystallization process, larger clusters as opposed to their smaller analogues may be formed if the synthesis is carried out under a set of conditions leading to slow-paced assembly of the cluster. In such a design, the widespread hydro/solvothermal synthesis is the method of choice. This is of particular interest when simple carboxylates are used as the hydrolysis-limiting ligands. Under ambient pressure, lanthanide carboxylates typically precipitate before any hydrolysis may occur. Under hydro/solvothermal conditions, however, the otherwise insoluble lanthanide carboxylates are adequately soluble to allow hydrolysis. Therefore, those ligands, which generally are not applicable for lanthanide hydrolysis, may now be utilized, and the hydrolysis process is slow enough to ensure the gradual assembly and growth of giant clusters thanks to the still-limited solubility of the lanthanide carboxylates. On the other hand, we submit that multiple anions may support the formation of giant clusters, as each anion may template the assembly of a fragment of the overall structure as well as support the final structure by filling any voids within the cluster core. As the majority of lanthanide salts are insoluble in water, direct addition of potentially templating anions to the reaction mixture is likely to cause precipitation of the lanthanide ions. Our strategy to overcome this problem is to control the release of such anions via in situ ligand decomposition, as is frequently encountered in hydro/ solvothermal syntheses.18 2.1. Lanthanide Hydrolysis Supported by Carboxylates, Polyalcohols, and Their Derivatives

Ligand-controlled lanthanide hydrolysis is the most successful route to hydroxide clusters.12 We are particularly interested in using acetate as a supporting ligand because it is one of the smallest organic ligands that may lead to clusters of high magnetic density. Supporting our strategy is the isolation of [Gd 24 Zn 4 (μ 6 ‑O) 4 (μ 4 ‑O)(OH) 44 (CH 3 COO) 12 (H 2 O) 48 ]· (ClO4)14 (1, Gd24Zn4), a 28-metal 3d−4f cluster obtained from 519

DOI: 10.1021/acs.accounts.7b00579 Acc. Chem. Res. 2018, 51, 517−525

Article

Accounts of Chemical Research

Figure 5. Ball-and-stick views of the clusters (a) La60Ni76 and (b) Ni64Ln96.

Figure 6. Crystal structures of (a) La20Ni30, (b) Pr20Ni21, and (c) Gd54Ni54.

polyalcohol ligand (Figure 4b).28 We also obtained nonanuclear clusters [Ln9(μ4‑O)(μ3‑OH)8(LH)4(OAc)4(H2O)12]·(ClO4)5 (10, Ln9, Ln = Gd, Dy; LH3 = Chromogen I) by heating a mixture of Ln(ClO4)3, NaOAc, NaOH, and the polyalcohol derivative N-acetyl-D-glucosamine (GlcNAc) (Figure 4c).29 The doubly deprotonated Chromogen Ia different polyalcohol ligand present in the final productwas apparently derived from GlcNAc upon dehydration. Recently, Zheng and co-workers obtained the nanoscopic cluster [Dy72(mda)24(mdaH)8(OH)120(O)8(NO3)16]·(NO3)8 (11, Dy72) with the use of N-methyldiethanolamine (mdaH2 ), a bisalcohol derivative of an amine, to control the hydrolysis of Dy(III).30

Gd52Ni56) was obtained.33 Though it is structurally very similar to Gd54Ni54, two Gd(III) and six CO32− ions in Gd54Ni54 are replaced by two Ni(II) ions and 10 OH− groups in the new cluster. With the same metalloligand, one giant 3d−4f cluster was obtained: [La60Ni76(IDA)68(μ3‑OH)158(NO3)4(H2O)44]· (NO3)34 (14, La60Ni76) (Figure 5a).34 The successful use of the metalloligand strategy has also been illustrated in the recent work by Zheng and co-workers, who obtained the largest 3d−4f clusters, [Ni64Ln96(μ3‑OH)156(IDA)66(DMPA)12(CH3COO)48(NO3)24(H2O)64]·Cl24 (15, Ln96Ni64, Ln = Gd, Dy, Y; DMPA = 2,2-dimethylol propionic acid) (Figure 5b), using a mixture of Ln(NO3)3, Ni(OAc)2, H2IDA, and DMPA.35 With 1,3-bis(2-pyridyl)-1,3-propanedione (pyacacH) as a ligand capable of forming a metalloligand, Tong and co-workers reported the wheel-shaped clusters [Co16Ln24(OH)50(pyacac)16(NO3)18(H2O)12][Ln(H2O)8]2(NO3)16(OH)10 (16, Ln = Dy, Gd).36

2.2. Assembly of Heterometallic 3d−4f Clusters Supported by Transition Metal Complexes as Metalloligands

Lanthanide and transition metal ions have distinctly different coordination behaviors. Using a preformed transition metal complex to control subsequent hydrolysis of a lanthanide ion provides a valid route to heterometallic 3d−4f cluster complexes.18 Key to the success of this approach is the use of chelating ligands with which a stable transition metal complex is first formed, followed by lanthanide hydrolysis controlled by the complex as a metalloligand. Previous work focused on the use of Schiff base ligands, but the use of the resulting metalloligands in lanthanide hydrolysis usually produced low-nuclearity complexes.31 Encouraged by the successful application of α-amino acids for the assembly of lanthanide hydroxide clusters, we turned to iminodiacetic acid (H2IDA) to test our hypothesis. A typical transition metal complex with IDA has the formula [M(IDA)]. With this metalloligand under hydrothermal conditions, we obtained the four-shell Russian-doll-like cluster [Gd54Ni54(IDA)48(OH)144(CO3)6(H2O)25]·(NO3)18 (12, Gd54Ni54).32 Interestingly, when the reaction was carried out under ambient pressure, [Gd52Ni56(IDA)48(OH)154(H2O)38]·(NO3)18 (13,

2.3. Anion Template Method

In the structures of the many lanthanide hydroxide clusters, the presence of one or more small anions is conspicuous. These anions are believed to template the organization of the recognizable smaller metal complex units or building blocks into the final giant clusters. Their functions include balancing the positive charge of the overall structure and providing geometrical support for the growing structure, among others.18 As anions vary in composition, charge, and/or geometry, the use of various anions is expected to lead to clusters of distinct structures and properties. The main challenge in using oxoanions as template is the possible precipitation of their lanthanide salts prior to hydrolysis. However, if the amount of oxoanions can be controlled at a low concentration, precipitation can be avoided, and slow growth or buildup of the cluster architecture may be anticipated. It has also been realized that the presence of 520

DOI: 10.1021/acs.accounts.7b00579 Acc. Chem. Res. 2018, 51, 517−525

Article

Accounts of Chemical Research

Figure 7. Ball-and-stick views of the metallic cores of (a) Gd36Ni12 and (b) Ln42M10.

As an example, the CO32− ions in [Gd24(DMF)36(μ4‑CO3)18(μ3‑H2O)2] (19, Gd24) were believed to be from the decomposition of DMF.40 The cationic cluster core Gd24 can be formally assembled by using eight triangles and 18 quadrangles with each of the quadrilateral faces being templated by a CO32− ion. Other lanthanide-exclusive hydroxide clusters formed with the assistance of low-concentration anion templates produced under hydro/solvothermal conditions are described in our recent review.14 2.3.2. Effect of Multiple Anion Templates, Including Mixed Templating Anions. Crystallographic studies have established the location of individual templating anions within a number of giant lanthanide-containing clusters. Analysis of the local coordination environments suggested the specific function(s) and the necessity of multiple anions for the assembly and stabilization of the cluster. As an example, structural analysis of Dy24Mn2 mentioned above revealed that three templating ClO4− anions are immobilized within the cluster cage through hydrogen bonding.26 Obviously, multiple anion templates may involve different types of anions. Mixed anions not only can act as complementary templates to induce the assembly of the cluster skeletons but also can enrich the structures of the clusters eventually realized. Mixed-anion templates of Cl− and NO3− are found in [Gd 36 Ni 12 (CH 3 COO) 18 (μ 3 ‑OH) 84 (μ 4 ‑O) 6 (H2O)54(NO3)Cl2]·(NO3)6·Cl9 (20, Gd36Ni12),41 a heterometallic cluster whose cationic core features a sandwich of two different kinds of cluster wheels (Figure 7a). The wheel-like structure is composed of six vertex-sharing square-pyramidal [Ln5(μ3‑OH)4(μ4‑O)]9+ units templated by one NO3− anion, while each of the outer “bread” layers is composed of six vertexsharing cubane-like [Gd3Ni(μ3‑OH)4]7+ units templated by two Cl− anions. The same mixed set of anion templates were also found in the similar tubular clusters [Gd48(μ4‑O)6(μ3‑OH)84(CAA) 36 (NO 3 ) 6 (H 2 O) 24 (EtOH) 12 (NO 3 )(Cl) 2 ]·Cl 3 (21, Gd 48 ) 24 and {[Cl 2 &(NO 3 )]@[Er 48 (NA) 44 (OH) 90 (N 3 )(H2O)24]}n (22, Er48).42 Interestingly, when the combination of ClO4− and CO32− was used, 52-metal clusters [(ClO4)3@ Ln42M10(μ3‑OH)68(CO3)12(CH3COO)30(H2O)70]·(ClO4)21 (23, Ln42M10, Ln = Gd, Dy; M = Co, Ni) were obtained.43 The bowl-like cationic core can be viewed as being constructed from [Ln8(μ3‑OH)9]15+, [Ln6M2(μ3‑OH)12]10+, and [M4(μ3‑OH)(CO3)3]+ units joined by nine CO32− ions (Figure 7b). In addition, three ClO4− anions are located in the three cavities of the bowl-like core. The cooperation of ClO4− and CO32− as mixed-anion templates also led to the cagelike cluster Gd27.29 Its [Gd27(μ3‑OH)32(CO3)8]32+ core can be built formally by

multiple templating anions is beneficial to the assembly of higher-nuclearity clusters, as each anion can locally template the formation of a piece of the cluster; multiple units of such smaller pieces can then be joined to generate a much larger clusters. Frequently, large cluster frameworks are created with significant internal voids, and the templating anions are found in these voids to stabilize the final product. The following examples from our research clearly demonstrate these individual or synergetic effects. 2.3.1. Slow Release of Anion Templates. In situ ligand decomposition under hydro/solvothermal conditions is one effective approach for achieving slow production of anions. Possibly because of the low concentration of the anion template, the system may self-adjust to achieve a balance between the metal ions, the templating anions, and the supporting ligands, leading to the assembly of higher-nuclearity clusters in a way similar to the growth of large single crystals via slow crystallization. As an example, the solvothermal cohydrolysis of Ln(III) and Ni(II) ions in the presence of H2IDA afforded different kinds of heterometallic clusters depending on the nature of the lanthanide ions used. The first type, formulated as [La 20 Ni 30 (IDA) 30 (CO 3 ) 6 (NO 3 ) 6 (OH) 30 (H2O)12]·(CO3)6 (17, La20Ni30),37 possesses a fascinating double-shell Keplerate structure featuring an icosidodecahedron encapsulating a dodecahedron (Figure 6a). As templates, six NO3− and six CO32− anions are located on the 12 pentagonal faces of the inner dodecahedron; the CO32− anions are believed to be from in situ decomposition of the IDA ligand. Interestingly, when the smaller Pr(III) ion was used, the sandwichlike cluster [Pr20 Ni21(IDA) 21(OH) 24(C 2 H2 O3 ) 6(C2O4)3(NO3)9(H2O)12]·(NO3)9 (18, Pr20Ni21) was obtained.38 Two new types of anions, C2O42− and glycolate (C2H2O3−), both believed to be derived from the original H2IDA under hydrothermal conditions, were found in a double-shell structure that is different from yet topologically related to the one described above (Figure 6b). When the reaction pH was increased, only six CO32− ions were observed in the giant cluster Gd54Ni54.32 The metal framework features a nesting Russian-doll-like four-shell structure, described as Gd2Ni6@Gd20@Gd32@Ni48 (Figure 6c). Six bridging CO32− ions contribute a total of 12 O atoms to the coordination spheres of the edge Gd(III) ions. This observation again indicates that the outcome of lanthanide hydrolysis is sensitively dependent on many factors, including a small change in pH of the reaction solution. Interestingly, the templating CO32− ions can also be generated from in situ decomposition of DMF or absorption of atmospheric CO2.39,40 521

DOI: 10.1021/acs.accounts.7b00579 Acc. Chem. Res. 2018, 51, 517−525

Article

Accounts of Chemical Research

43.6 J kg−1 K−1, respectively.22,27,29 The larger and heavier organic ligands and the incorporation of large anions both contribute to the reduced magnetic density and smaller MCEs observed for three clusters. The −ΔSM value is linearly related to the magnetic spin density. Data points above the line are due to weaker magnetic interactions, whereas those below the line are the result of stronger magnetic interactions. These observations suggest that magnetic exchange coupling between metal ions within a cluster complex also plays an important role in affecting the MCE, with weak magnetic interactions being much more desirable. For example, the Gd24 cluster displayed a maximum −ΔSM of 46.1 J kg−1 K−1, which is close to the theoretical value of 52.1 J kg−1 K−1.40 Such a large entropy change is probably due to weak antiferromagnetic interactions (θ = −0.16 K). Although it possesses a high spin density, the Gd36 cluster displayed a smaller −ΔSM value of 39.6 J kg−1 K−1 than the Gd24 cluster (−ΔSM = 46.1 J kg−1 K−1),23 which is consistent with the stronger antiferromagnetic interactions in Gd36 (θ = −2.41 K) than in Gd24.23,40 A similar situation was identified for Gd48, with −ΔSM = 41.8 J kg−1 K−1 and θ = −3.57.24

using multiple [Gd4(μ3‑OH)4]8+, [Gd4(μ3‑OH)7]5+, and [Gd7(μ3‑OH)5(CO3)4]8+ units; one ClO4− template is found in the center of the cage.

3. MAGNETOCALORIC EFFECT OF HIGH-NUCLEARITY LANTHANIDE-CONTAINING CLUSTERS Key to the potential magnetic cooling application of a substance is its large MCE, which can be evaluated by the isothermal magnetic entropy change (−ΔSM).6 For a magnetic coolant, the molar entropy associated with the magnetic degrees of freedom can be calculated using the equation SM = nR ln(2S + 1).8 Recent studies have shown that molecular magnets with large MCEs usually have high-spin ground states, negligible magnetic anisotropy, and low-lying excited spin states. For these considerations, Gd(III) ion with its f7 electron configuration is a perfect choice for making molecular clusters with large MCEs. Despite the extensive efforts made, the design and synthesis such molecules remain a great challenge. Our own efforts in this vein have concentrated on improving the magnetic spin density. By increasing the metal nuclearity and using small ligands, we obtained a series of high-nuclearity Gd(III)-exclusive clusters having large MCEs. In addition, the effect of magnetic coupling on the MCE was studied. Introducing Gd(III) to mitigate any intracluster magnetic coupling or increasing the separation between cluster units by encapsulating one single molecular cluster in a nanoshell of silica led to enhanced MCEs. In general, high magnetic spin density and weak magnetic coupling are key factors for achieving large MCEs in molecular magnetic coolers. The studies of the MCEs of Gd(III)-exclusive and heterometallic 3d−Gd(III) clusters are summarized below.

3.2. High-Nuclearity Heterometallic 3d−Gd(III) Clusters for Cryogenic Magnetic Cooling

The generally weak magnetic coupling between Gd(III) and 3d transition metal ions, combined with the ability of Gd(III) to mitigate the otherwise relatively strong 3d−3d magnetic exchange, makes 3d−Gd(III) cluster complexes a valid class of synthetic targets for magnetic cooling applications. Analogous to the situation for Gd(III)-exclusive clusters, a high spin density of a 3d−Gd(III) complex generally correlates to a large −ΔSM value (Figure 9). Accordingly, Gd24Co16 and

3.1. High-Nuclearity Gd(III)-Exclusive Clusters for Cryogenic Magnetic Cooling

High magnetic spin density and negligible anisotropy are the primary requirements to obtain high MCEs. Generally speaking, the higher the spin density, the larger is the −ΔSM value (Figure 8). Therefore, the most straightforward strategy is

Figure 9. Plots of gravimetric −ΔSM values vs unpaired electron density for reported high-nuclearity 3d−4f clusters at ΔH = 70 kOe.

Gd24Cu36 clusters supported by large and heavy organic ligands showed small −ΔSM values of 26.036 and 21.0 J kg−1 K−1,25 respectively. In comparison, Gd24Zn4 has a maximum −ΔSM value of 31.4 J kg−1 K−1 based on its high magnetic density that is directly associated with the use of the smaller and lighter acetate ligand.20 With an increased number of Gd(III) ions, the acetate-supported cluster Gd36Ni12 exhibited a reasonably large −ΔSM of 36.3 J kg−1 K−1.41 Changing the anion templates afforded two more acetate-supported clusters with high spin density, Gd42Ni10 and Gd42Co10, having −ΔSM = 38.2 and 41.3 J kg−1 K−1, respectively.42 The slightly larger value of −ΔSM for Gd42Co10 may be understood in terms of the higher groundstate spin of Co(II) ion. Increasing the percentage of Gd(III) in a heterometallic cluster complex may help enhance its MCE.

Figure 8. Plots of gravimetric −ΔSM values vs unpaired electron density for reported high-nuclearity Gd(III) clusters at ΔH = 70 kOe.

to select small and light ligands to construct high-nuclearity Gd(III)-based clusters. For example, acetate-supported Gd104 has the highest spin density (calculated by unpaired electrons/ mol g−1) and displays a −ΔSM value of 46.9 J kg−1 K−1, which is the largest among any known lanthanide-exclusive cluster compounds.21 In comparison, the propionate-supported Gd27, 1,2,3-cyclohexanetriol-based Gd37, and chloroacetate-protected Gd48 clusters display smaller −ΔSM values of 41.8, 38.7, and 522

DOI: 10.1021/acs.accounts.7b00579 Acc. Chem. Res. 2018, 51, 517−525

Article

Accounts of Chemical Research

Figure 10. (a) TEM image and schematic diagram of the Gd52Ni56@SiO2 nanoparticles. (b, c) Plots of −ΔSM vs temperature for (b) Gd52Ni56 and (c) Gd52Ni56@SiO2.

tunable platform for investigating magnetic exchange coupling within a particular cluster unit.28

Some results suggest the validity of such a submission. The largest 3d−4f metal cluster, Gd96Ni64, produced a −ΔSM value of 42.8 J kg−1 K−1, which is the largest among all reported 3d− Gd clusters.35 Analogous to Gd(III)-exclusive clusters, magnetic exchange coupling in 3d−Gd(III) clusters also significantly affects the MCE. As shown in Figure 9, although Gd36Ni12 exhibited a reasonably large magnetic entropy change of 36.3 J kg−1 K−1,41 this value is only about half of the theoretical value (64.8 J kg−1 K−1) as a result of both the antiferromagnetic interactions within the cluster and the Ni(II) crystal-field effect. Compared with Gd36Ni12 (θ = −3.95 K), Gd42Ni10 has a larger −ΔSM value of 38.2 J kg−1 K−1 because of its weaker antiferromagnetic interactions (θ = −3.79 K) despite its lower magnetic density.43 These results are consistent with the general understanding that weak magnetic interactions are beneficial for improving the MCEs of 3d−4f clusters. Lastly, for practical applications, magnetic interactions between individual clusters within the bulk material are important. To evaluate such interactions, we prepared by the microemulsion method a core−shell Gd52Ni56@SiO2 hybrid; monodisperse core−shell nanoparticles were obtained, each consisting of one core of cluster encapsulated by one nanoshell of silica (Figure 10).33 Its maximum −ΔSM of 44.6 J kg−1 K−1 reflects a 10% enhancement over the value for the corresponding parent cluster (40.5 J kg−1 K−1). Theoretical studies suggested that shelling the molecular cluster should help reduce magnetic interactions between the cluster units. Interestingly, experimental and theoretical studies on the Gd12Fe14 and Gd12Fe14@SiO2 systems revealed that the shielding effect of the SiO2 not only effectively decreases the intermolecular magnetic interactions but also significantly increases the zero-field splitting effect of the outer-layer Fe(III) ions. Significantly, forming the core−shell nanostructure not only helps stabilize the cluster and improve the processability of the materials for practical applications but also provides a

4. SUMMARY AND OUTLOOK Molecule-based cryogenic magnetic cooling materials have received much interest in the past two decades, with highnuclearity lanthanide clusters dominating related research. Exploratory synthesis produced numerous cluster species with appealing molecular structures and promising magnetic properties. In this Account, we have highlighted our own efforts in the past decade by focusing on the synthetic strategies targeting giant clusters with high magnetic density and large MCEs. Toward this goal, hydro/solvothermal synthesis has been applied in the case of common but less studied ligands such as carboxylates and polyalcohols. In the course of such studies, we have realized the critical importance of small anions as templates to support the cluster assembly. To avoid the precipitation of lanthanide salts with such anions, we developed the strategies of slow release/production of such anions. Studies aiming at establishing a magneto−structure relationship have also been presented, from which the importance of high magnetic spin density and paramagnetic or weak magnetic coupling for achieving large MCEs in molecular magnetic coolers has been clearly established. Although great progress has been made in the synthesis and magnetic studies of cluster species for possible applications in achieving magnetic cooling, one profound lesson from the substantial body of work is that the chemistry of lanthanide hydroxide clusters is sensitive to a number of chemical forces. Consequently, one should be cautious about any prognostication in this field. From the molecular structure viewpoint, we believe that certain common structural features of the lanthanide clusters have been recognized. Specifically, the seemingly modular structure suggests the real possibility of creating even larger “clusters of clusters”. It is entirely reasonable to envision the creation of structurally more complex clusters using various combinations of smaller building 523

DOI: 10.1021/acs.accounts.7b00579 Acc. Chem. Res. 2018, 51, 517−525

Article

Accounts of Chemical Research units. As an example, a supercluster cage with the same architecture as the seminal Buckminsterfullerene (C60) could be built from wheel-like cluster units that are composed of fiveand six-vertex-sharing [Ln4(μ3‑OH)4]8+ cubanes. It is indeed an immediate synthetic target of our ongoing efforts. For the ultimate applications of lanthanide clusters as molecular magnetic coolers, we have investigated the factors believed to be behind the much-sought large MCEs. In addition to increasing the number of isotropic Gd(III) ions within the cluster core, incorporating other high ground-state spin metal ions in the form of heterometallic 3d−4f clusters has also been pursued. Other strategies toward the making of a molecular cooler substance include the use of small organic ligands to help increase the magnetic density. Last but not least, with all of the molecular design strategies attempted, we have made our initial efforts in minimizing the intercluster magnetic coupling by separating individual magnetic clusters via encapsulation using a nanoshell of silica. With these combined efforts, we hope to see in the near future the real applications of highnuclearity lanthanide clusters for energy-efficient and environmentally benign magnetic cooling technology.

■

His current research interest is focused on studies of clusters and related materials.

■

ACKNOWLEDGMENTS This work was supported by the 973 Project (Grant 2014CB845601) from the Ministry of Science and Technology of China, the National Natural Science Foundation of China (Grants 21422106, 21673184, 21371144, 21431005, and 21390391), the Fok Ying Tong Education Foundation (151013), the Recruitment Program for Leading Talent Team of Anhui Province, and the U.S. National Science Foundation (Grant CHE-1152609).

■

REFERENCES

(1) Warburg, E. Magnetische Untersuchungen. Ann. Phys. 1881, 249, 141−164. (2) Debye, P. Einige Bemerkungen zur Magnetisierung bei tiefer Temperatur. Ann. Phys. 1926, 386, 1154−1160. (3) Giauque, W. F. A thermodynamic treatment of certain magnetic effects. A proposed method of producing temperatures considerably below 1° absolute. J. Am. Chem. Soc. 1927, 49, 1864−1870. (4) Tegus, O.; Bruck, E.; Buschow, K. H. J.; de Boer, F. R. Transition-metal-based magnetic refrigerants for room-temperature applications. Nature 2002, 415, 150−152. (5) Evangelisti, M.; Luis, F.; de Jongh, L. J.; Affronte, M. Magnetothermal properties of molecule-based materials. J. Mater. Chem. 2006, 16, 2534−2549. (6) Torres, F.; Hernández, J. M.; Bohigas, X.; Tejada, J. Giant and time-dependent magnetocaloric effect in high-spin molecular magnets. Appl. Phys. Lett. 2000, 77, 3248−3250. (7) Sharples, J. W.; Zheng, Y. Z.; Tuna, F.; McInnes, E. J.; Collison, D. Lanthanide discs chill well and relax slowly. Chem. Commun. 2011, 47, 7650−7652. (8) Sessoli, R. Chilling with Magnetic Molecules. Angew. Chem., Int. Ed. 2012, 51, 43−45. (9) Zheng, Y.-Z.; Evangelisti, M.; Winpenny, R. E. P. Large Magnetocaloric Effect in a Wells−Dawson Type {Ni6Gd6P6} Cage. Angew. Chem., Int. Ed. 2011, 50, 3692−3695. (10) Wang, R.; Zheng, Z.; Jin, T.; Staples, R. J. Coordination Chemistry of Lanthanides at “High” pH: Synthesis and Structure of the Pentadecanuclear Complex of Europium(III) with Tyrosine. Angew. Chem., Int. Ed. 1999, 38, 1813−1815. (11) Zheng, Z. Ligand-controlled self-assembly of polynuclear lanthanide−oxo/hydroxo complexes: from synthetic serendipity to rational supramolecular design. Chem. Commun. 2001, 2521−2529. (12) Wang, R.; Selby, H. D.; Liu, H.; Carducci, M. D.; Jin, T.; Zheng, Z.; Anthis, J. W.; Staples, R. J. Halide-templated assembly of polynuclear lanthanide-hydroxo complexes. Inorg. Chem. 2002, 41, 278−286. (13) Kong, X.-J.; Wu, Y.; Long, L.-S.; Zheng, L.-S.; Zheng, Z. A chiral 60-metal sodalite cage featuring 24 vertex-sharing [Er4(μ3-OH)4] cubanes. J. Am. Chem. Soc. 2009, 131, 6918−6919. (14) Zheng, X.-Y.; Xie, Y.; Kong, X.-J.; Long, L.-S.; Zheng, L.-S. Recent Advances in the Assembly of High-Nuclearity Lanthanide Clusters. Coord. Chem. Rev. 2017, DOI: 10.1016/j.ccr.2017.10.023. (15) Kong, X.-J.; Long, L.-S.; Zheng, Z.; Huang, R.-B.; Zheng, L.-S. Keeping the ball rolling: fullerene-like molecular clusters. Acc. Chem. Res. 2010, 43, 201−209. (16) Zheng, Y.-Z.; Zheng, Z.; Chen, X.-M. A symbol approach for classification of molecule-based magnetic materials exemplified by coordination polymers of metal carboxylates. Coord. Chem. Rev. 2014, 258−259, 1−15. (17) Liu, J.-L.; Chen, Y.-C.; Guo, F.-S.; Tong, M.-L. Recent advances in the design of magnetic molecules for use as cryogenic magnetic coolants. Coord. Chem. Rev. 2014, 281, 26−49. (18) Zheng, X.-Y.; Kong, X.-J.; Long, L.-S. Synthesis and Structures of Lanthanide−Transition Metal Clusters. In Recent Development in

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (X.-J. Kong). *E-mail: [email protected] (Z. Zheng). *E-mail: [email protected] (L.-S. Long). ORCID

Xiang-Jian Kong: 0000-0003-0676-6923 La-Sheng Long: 0000-0002-0398-4709 Notes

The authors declare no competing financial interest. Biographies Xiu-Ying Zheng received her B.S. in Chemistry from Anqing Normal University. She is pursuing her graduate studies on the exploratory synthesis and magnetic studies of heterometallic lanthanide−transition metal clusters under the direction of Profs. Xiang-Jian Kong and LaSheng Long. Prof. Xiang-Jian Kong received his B.S. in Chemistry from Liaocheng University and his Ph.D. from Xiamen University in 2009 with Professor La-Sheng Long. He is now a professor at Xiamen University. His research is focused on lanthanide-containing clusters. Prof. Zhiping Zheng received his B.S. and M.S. in Chemistry from Peking University in China and his Ph.D. from UCLA in 1995 with Professor M. Frederick Hawthorne. After conducting postdoctoral research with Professor Richard H. Holm at Harvard University, he joined the faculty of the University of Arizona in 1997. His current research is focused on the synthetic and materials chemistry of cluster compounds of both lanthanide and transition metal elements. Prof. La-Sheng Long received his B.S. and M.S. in Chemistry from Anhui Normal University in 1986 and Lanzhou University in 1989, respectively, and his Ph.D. from Zhongshan University in 1999. He is now a professor at Xiamen University. His current research is focused on the synthetic and materials chemistry of cluster compounds containing lanthanide and/or transition metal elements. Prof. Lan-Sun Zheng received his B.S. in chemistry from Xiamen University in China and his Ph.D. from Rice University in 1986 with Professor Richard E. Smalley. He is a professor at Xiamen University. 524

DOI: 10.1021/acs.accounts.7b00579 Acc. Chem. Res. 2018, 51, 517−525

Article

Accounts of Chemical Research Clusters of Rare Earths and Actinides: Chemistry and Materials; Zheng, Z., Ed.; Springer: Berlin, 2017; pp 51−96. (19) Papatriantafyllopoulou, C.; Moushi, E. E.; Christou, G.; Tasiopoulos, A. J. Filling the gap between the quantum and classical worlds of nanoscale magnetism: giant molecular aggregates based on paramagnetic 3d metal ions. Chem. Soc. Rev. 2016, 45, 1597−1628. (20) Zheng, X. Y.; Wang, S. Q.; Tang, W.; Zhuang, G. L.; Kong, X. J.; Ren, Y. P.; Long, L. S.; Zheng, L. S. Two nanosized 3d−4f clusters featuring four Ln6 octahedra encapsulating a Zn4 tetrahedron. Chem. Commun. 2015, 51, 10687−10690. (21) Peng, J. B.; Kong, X. J.; Zhang, Q. C.; Orendac, M.; Prokleska, J.; Ren, Y. P.; Long, L. S.; Zheng, Z.; Zheng, L. S. Beauty, Symmetry, and Magnetocaloric Effect-Four-Shell Keplerates with 104 Lanthanide Atoms. J. Am. Chem. Soc. 2014, 136, 17938−17941. (22) Zheng, X.-Y.; Peng, J.-B.; Kong, X.-J.; Long, L.-S.; Zheng, L.-S. Mixed-anion templated cage-like lanthanide clusters: Gd27 and Dy27. Inorg. Chem. Front. 2016, 3, 320−325. (23) Wu, M.; Jiang, F.; Kong, X.; Yuan, D.; Long, L.; Al-Thabaiti, S. A.; Hong, M. Two polymeric 36-metal pure lanthanide nanosize clusters. Chem. Sci. 2013, 4, 3104−3109. (24) Guo, F. S.; Chen, Y. C.; Mao, L. L.; Lin, W. Q.; Leng, J. D.; Tarasenko, R.; Orendac, M.; Prokleska, J.; Sechovsky, V.; Tong, M. L. Anion-Templated Assembly and Magnetocaloric Properties of a Nanoscale {Gd38} Cage versus a {Gd48} Barrel. Chem. - Eur. J. 2013, 19, 14876−14885. (25) Leng, J. D.; Liu, J. L.; Tong, M. L. Unique nanoscale {CuII36LnIII24} (Ln = Dy and Gd) metallo-rings. Chem. Commun. 2012, 48, 5286−5288. (26) Liu, D.-P.; Peng, J.-B.; Lin, X.-P.; Huang, Q.; Kong, X.-J.; Long, L.-S.; Huang, R.-B.; Zheng, L.-S. Myo-inositol supported heterometallic Dy24M2 (M = Ni, Mn) cages. CrystEngComm 2014, 16, 5527− 5530. (27) Zhou, Y.; Zheng, X. Y.; Cai, J.; Hong, Z. F.; Yan, Z. H.; Kong, X. J.; Ren, Y. P.; Long, L. S.; Zheng, L. S. Three Giant Lanthanide Clusters Ln37 (Ln = Gd, Tb, and Eu) Featuring a Double-Cage Structure. Inorg. Chem. 2017, 56, 2037−2041. (28) Zheng, X.-Y.; Zhang, H.; Wang, Z. X.; Liu, P. X.; Du, M. H.; Han, Y. Z.; Wei, R. J.; Ouyang, Z. W.; Kong, X.-J.; Zhuang, G.-L.; Long, L.-S.; Zheng, L.-S. New Insight into Magnetic Interaction in Monodisperse Gd12Fe14 Metal Cluster. Angew. Chem., Int. Ed. 2017, 56, 11475−11479. (29) Zheng, X. Y.; Peng, J. B.; Livera, M. M.; Luo, Y.; Wang, Y. Y.; Kong, X. J.; Long, L. S.; Zheng, Z.; Zheng, L. S. Selective Formation of Chromogen I from N-Acetyl-D-glucosamine upon Lanthanide Coordination. Inorg. Chem. 2017, 56, 110−113. (30) Qin, L.; Yu, Y.-Z.; Liao, P.-Q.; Xue, W.; Zheng, Z.; Chen, X.-M.; Zheng, Y.-Z. A “Molecular Water Pipe”: A Giant Tubular Cluster {Dy72} Exhibits Fast Proton Transport and Slow Magnetic Relaxation. Adv. Mater. 2016, 28, 10772−10779. (31) Liu, K.; Shi, W.; Cheng, P. Toward heterometallic singlemolecule magnets: Synthetic strategy, structures and properties of 3d− 4f discrete complexes. Coord. Chem. Rev. 2015, 289−290, 74−122. (32) Kong, X. J.; Ren, Y. P.; Chen, W. X.; Long, L. S.; Zheng, Z.; Huang, R. B.; Zheng, L. S. A four-shell, nesting doll-like 3d−4f cluster containing 108 metal ions. Angew. Chem., Int. Ed. 2008, 47, 2398−401. (33) Liu, D. P.; Lin, X. P.; Zhang, H.; Zheng, X. Y.; Zhuang, G. L.; Kong, X. J.; Long, L. S.; Zheng, L. S. Magnetic Properties of a SingleMolecule Lanthanide-Transition-Metal Compound Containing 52 Gadolinium and 56 Nickel Atoms. Angew. Chem., Int. Ed. 2016, 55, 4532−4536. (34) Kong, X. J.; Long, L. S.; Huang, R. B.; Zheng, L. S.; Harris, T. D.; Zheng, Z. A four-shell, 136-metal 3d−4f heterometallic cluster approximating a rectangular parallelepiped. Chem. Commun. 2009, 4354−4356. (35) Chen, W. P.; Liao, P. Q.; Yu, Y.; Zheng, Z.; Chen, X. M.; Zheng, Y. Z. A Mixed-Ligand Approach for a Gigantic and Hollow Heterometallic Cage {Ni64RE96} for Gas Separation and Magnetic Cooling Applications. Angew. Chem., Int. Ed. 2016, 55, 9375−9379.

(36) Zhang, Z. M.; Pan, L. Y.; Lin, W. Q.; Leng, J. D.; Guo, F. S.; Chen, Y. C.; Liu, J. L.; Tong, M. L. Wheel-shaped nanoscale 3d−4f {CoII16LnIII24} clusters (Ln = Dy and Gd). Chem. Commun. 2013, 49, 8081−8083. (37) Kong, X.-J.; Ren, Y.-P.; Long, L.-S.; Zheng, Z.; Nichol, G.; Huang, R.-B.; Zheng, L.-S. Dual Shell-like Magnetic Clusters Containing NiII and LnIII (Ln = La, Pr, and Nd). Inorg. Chem. 2008, 47, 2728−2739. (38) Kong, X.-J.; Ren, Y.-P.; Long, L.-S.; Zheng, Z.; Huang, R.-B.; Zheng, L.-S. A Keplerate Magnetic Cluster Featuring an Icosidodecahedron of Ni(II) Ions Encapsulating a Dodecahedron of La(III) Ions. J. Am. Chem. Soc. 2007, 129, 7016−7017. (39) Hooper, T. N.; Inglis, R.; Palacios, M. A.; Nichol, G. S.; Pitak, M. B.; Coles, S. J.; Lorusso, G.; Evangelisti, M.; Brechin, E. K. CO2 as a reaction ingredient for the construction of metal cages: a carbonatepanelled [Gd6Cu3] tridiminished icosahedron. Chem. Commun. 2014, 50, 3498−3500. (40) Chang, L. X.; Xiong, G.; Wang, L.; Cheng, P.; Zhao, B. A 24-Gd nanocapsule with a large magnetocaloric effect. Chem. Commun. 2013, 49, 1055−1057. (41) Peng, J. B.; Zhang, Q. C.; Kong, X. J.; Ren, Y. P.; Long, L. S.; Huang, R. B.; Zheng, L. S.; Zheng, Z. A 48-metal cluster exhibiting a large magnetocaloric effect. Angew. Chem., Int. Ed. 2011, 50, 10649− 10652. (42) Wu, M.; Jiang, F.; Yuan, D.; Pang, J.; Qian, J.; Al-Thabaiti, S. A.; Hong, M. Polymeric double-anion templated Er48 nanotubes. Chem. Commun. 2014, 50, 1113−1115. (43) Peng, J. B.; Zhang, Q. C.; Kong, X. J.; Zheng, Y. Z.; Ren, Y. P.; Long, L. S.; Huang, R. B.; Zheng, L. S.; Zheng, Z. High-nuclearity 3d− 4f clusters as enhanced magnetic coolers and molecular magnets. J. Am. Chem. Soc. 2012, 134, 3314−3317.

525

DOI: 10.1021/acs.accounts.7b00579 Acc. Chem. Res. 2018, 51, 517−525