English

• Heterometallic 3d-4f cluster compounds have attracted widespread interest due to their interesting optical, electrical, magnetic, and catalytic properties resulting from the contribution of 3d and 4f electrons [1-9]. Considerable efforts in this field have resulted in the fabrication of a large number of 3d-4f clusters with fascinating structures, and unique physical and chemical properties based on various organic ligands. In these clusters, the organic ligands play an important protecting and bridging role in the separation and preparation of cluster structures [10-12]. 3d-4f Clusters protected by inorganic ligands have higher thermal stabilities and unique rigid structures compared to 3d-4f clusters protected by organic ligands [13, 14].

  Lacunary polyoxometalates (POMs) are a class of inorganic multidentate ligands with well-defined vacant sites and high negative charges. These are commonly used to prepare stable metal oxygen clusters or cluster-of-cluster aggregates due to their high nucleophilicity and high reactivity [15-19]. Several 3d metal clusters based on the trivacant POM {XW₉O₃₄} (heteroatom x = P, Si, Sb) have been reported [20-29]. However, the synthesis of POM-based 3d-4f clusters remains challenging due to the coordination competition between 3d metal ions and 4f ions in inorganic POM ligands. Moreover, the strong reaction between 4f ions and the oxygen-rich POM makes precipitation easy but crystallization difficult [13, 14, 30-34]. Therefore, reducing the reaction rate of lanthanide ions with POMs may be an effective strategy for the synthesis of POM-based 3d-4f clusters [35-37]. In this study, we have obtained three {FeW₉O₃₄}-based 3d-4f clusters, formulated—Na₈K₂[Fe₂Ln₂(H₂O)₄(B-α-FeW₉O₃₄)₂]·nH₂O (Ln₂Fe₄, where Ln = Dy and n = 17 for cluster 1; Ln = Ho and n = 15 for cluster 2; and Ln = Y and n = 14 for cluster 3). These clusters have classical sandwich-type structures, where the [(B-α-FeW₉O₃₄)¹¹⁻] unit is generated via the transformation of the [(B-α-SbW₉O₃₃)⁹⁻] precursor (Figs. 1a and b). Ln₂Fe₄ is the first 3d-4f cluster assembled from POMs containing d-metal heteroatoms. Interestingly, Dy₂Fe₄ exhibits single-molecule magnet (SMM) behavior with a high energy barrier.

  Figure 1

  

  Figure 1.  The structures of (a) [B-α-SbW₉O₃₃]⁹⁻ unit; (b) [B-α-FeW₉O₃₄]¹¹⁻ unit. (c, d) Ball-and-stick/polyhedral view of [(B-α-FeW₉O₃₄)₂Fe₂Dy₂(H₂O)₄]¹⁰⁻. (e) The metal core of 2Dy³⁺ and 4Fe³⁺ ions in cluster 1. (f-h) The coordination geometries of DyO₈/FeO₆/FeO₄. Color key: W/WO₆, blue; Fe/FeO₆/FeO_4, green; Dy/DyO₈, purple; Sb, black and O, pink.

  DownLoad: Full-Size Img PowerPoint

  Clusters 1–3 were synthesized from Na₉[B-α-SbW₉O₃₃]·19.5H₂O, Ln₂O₃ (Ln = Dy/Ho/Y), FeCl₃·6H₂O and KH₂PO₄ under hydrothermal conditions (details in Supporting information). Single-crystal X-ray diffraction analysis showed that the isomorphic clusters 1–3 crystallized in the triclinic crystal system with a P-1 space groups (Table S1 in Supporting information). The valence bonding theory calculations of the three clusters (Tables S2-S4 in Supporting information) and the Mössbauer spectrum of cluster 3 indicate that the Fe ions are all high-spin Fe³⁺ (Fig. S3, Tables S5 and S6 in Supporting information) [38]. As shown in Figs. 1c and d, anion cluster 1 [(B-α-FeW₉O₃₄)₂Fe₂Dy₂(H₂O)₄]¹⁰⁻ exhibits a classic sandwich structural unit, which can be observed as a [Fe₂Dy₂(H₂O)₄]¹²⁺ unit sandwiched by two inorganic [(B-α-FeW₉O₃₄)¹¹⁻] ligands. The [(B-α-FeW₉O₃₄)¹¹⁻] unit is formed via the in situ transformation of the [(B-α-SbW₉O₃₃)⁹⁻] precursor. The Ln₂Fe₄ described in this study is the first 3d-4f cluster containing a trilacunary [(B-α-FeW₉O₃₄)¹¹⁻] unit [39-43]. The formation of the [(B-α-FeW₉O₃₄)¹¹⁻] species was investigated using single-crystal structure analysis, infrared spectroscopy, and energy dispersive spectroscopy (Figs. S1, S4-S6 in Supporting information). The metal core Fe₄Dy₂ in cluster 1 has exactly two tetrahedral structures with shared edges, with Fe⋯Fe distances of 3.1301(33)–3.3196(29) Å and Fe⋯Dy distances of 3.4987(22)–3.7597(26) Å (Fig. 1e, Fig. S18 and Table S12 in Supporting information). Continuous shape measurement (CShM) using Alvarez's SHAPE program (Table S8 in Supporting information) [44, 45] shows that Dy³⁺ in the [Fe₂Dy₂(H₂O)₄]¹²⁺ unit is 8-coordinated with triangular dodecahedral geometry (Fig. 1f), which is different from that reported for Fe₄Dy₂ [46-50]. The Dy-O length is between 2.243(12) Å and 2.514(15) Å (Table S7 in Supporting information). Six-coordinated Fe³⁺ has a classical octahedral geometry (Fig. 1g) with an Fe-O length of 1.991(10) Å–2.048(10) Å, whereas 4-coordinated Fe³⁺ in [(FeW₉O₃₄)¹¹⁻] has a special tetrahedral geometry (Fig. 1h) with an Fe-O length of 1.834(10) Å–1.856(10) Å (Table S2). Clusters 2 and 3 have the same coordination pattern and similar bond lengths and bond angles as cluster 1 (Tables S3, S4, S7 and S8 in Supporting information). Based on elemental analysis, thermogravimetric analysis (Fig. S2 in Supporting information), ion chromatography, and charge balance theory, there are 8 Na⁺ and 2 K⁺ counter cations. As shown in Figs. S7 and S8 (Supporting information), the [(B-α-FeW₉O₃₄)₂Fe₂Dy₂]¹⁰⁻ anions are connected through hydrated Na⁺ and K⁺ counter ions to form a two-dimensional framework. The minimum asymmetric units of clusters 1-3 are shown in Figs. S9-S11 (Supporting information).

  The use of insoluble lanthanide substances also plays a key role in the formation of clusters 1–3. Clusters 1–3 can be obtained using insoluble lanthanide salts such as Ln₂(CO₃)₃ and Ln₂(SO₄)₃ to replace Ln₂O₃ (Ln = Dy, Ho and Y). However, such clusters cannot be obtained using soluble LnCl₃ or Ln(NO₃)₃ salts. Ln₂O₃ can slowly release Ln³⁺ ions under mild acidic conditions (pH 5.5) and medium-high temperature hydrothermal conditions (140 ℃). The low equilibrium concentration of Ln³⁺ ions during the reaction can effectively slow the reaction rate between Ln³⁺ and POMs to prevent precipitation. Therefore, combining in situ POM transformation with the slow release of Ln³⁺ ions might be an effective synthetic strategy for the synthesis of POM-based 3d-4f clusters.

  The variable-temperature magnetic susceptibilities (χ_mT) of clusters 1–3 were measured between 2 K and 300 K under a 1000 Oe applied magnetic field. The relevant magnetic data obtained from these measurements are listed in Table S9 (Supporting information). The experimental χ_mT values for clusters 1–3 at 300 K (33.05, 32.20 and 5.63 cm³ K/mol, respectively) are smaller than the expected values (45.83, 45.63 and 17.50 cm³ K/mol, respectively), which may be due to antiferromagnetic interactions in the clusters [51]. As shown in Fig. 2, with decreasing temperature, the χ_mT values of clusters 2 and 3 gradually decrease to 21.69 cm³ K/mol and 1.09 cm³ K/mol at 2 K, respectively, mainly due to the thermal depopulation of the excited states at the m_J sublevels of the Fe³⁺ and Ho³⁺ ions. The χ_mT values of cluster 1 gradually decreased to 25.22 cm³ K/mol at 9 K followed by a sudden rebound to 25.99 cm³ K/mol at 2 K. This indicates that the presence of weak dipole interactions and/or ferromagnetic interactions between Dy³⁺ and Fe³⁺ below 9 K, including the decrease above 9 K can be attributed to the thermal depopulation of the excited states at the m_J sublevels of the Fe³⁺ and Dy³⁺ ions [52]. Fitting the χ_m⁻¹-T data individually in the appropriate temperature interval for these three clusters according to the Curie-Weiss law gives C = 34.77 cm³/mol, θ = −17.57 K (1); C = 34.57 cm³/mol, θ = −16.87 K (2); and C = 34.45 cm³/mol, θ = −1554.21 K (3) (Figs. S12-S14 in Supporting information), which further indicates the presence of antiferromagnetic interactions. As shown in Figs. S15-S17 (Supporting information), the inter-cluster magnetic interactions are negligible due to the long distances. Cluster 3 has only Fe-Fe magnetic interactions, whereas cluster 1 (2) has Fe-Fe and Fe-Dy (Fe-Ho) magnetic interactions (Fig. S18 in Supporting information). Therefore, the difference in χ_mT values between isomorphic clusters 1, 2 and 3 can be attributed to the contribution of Fe-Dy (Fe-Ho) magnetic interactions (Fig. S19 in Supporting information). The field-dependent magnetization curves of 1–3 at 2 K are shown in Fig. S20 (Supporting information). The experimental values at 7 T are much smaller than their theoretical saturation values. Furthermore, the M vs. H/T curves at different temperatures are not superimposed on a single master curve (Fig. S20, Table S9), which suggests the existence of significant magnetic anisotropy and/or low-lying excited states [53].

  Figure 2

  

  Figure 2.  χ_mT vs. T in the range of 2–300 K under an applied field of 1000 Oe for clusters 1–3. Inset: Enlarged view of the upturn of χ_mT (cluster 1).

  DownLoad: Full-Size Img PowerPoint

  AC susceptibility measurements of clusters 1–3 were performed to explore their dynamic magnetic properties at indicated frequencies under zero DC field and a 3 Oe AC field at 2 K. As shown in Figs. S21 (Supporting information), the out-of-phase susceptibilities (χ") of clusters 2 and 3 exhibit temperature dependence at 100–1500 Hz, but no significant peaks are observed due to the quantum tunneling effect. The in-phase susceptibilities (χ') and the out-of-phase susceptibilities (χ") of cluster 1 exhibit obvious temperature dependence, and the out-of-phase susceptibilities (χ") have obvious peaks at 0.1–1500 Hz. The AC magnetic susceptibility of cluster 1 was further tested at the specified temperatures in the frequency range of 0.1–1000 Hz under zero DC field and a 3 Oe AC field. Figs. 3a and b show the frequency dependence plots of χ' and χ''. A series of single frequency-dependent peaks were observed between 2 K and 9 K, indicating the SMM behavior of cluster 1. The Cole-Cole plots were fitted using the generalized Debye model (Fig. 3c). The large α values under zero DC field indicates the presence of multiple relaxation processes (Table S10 in Supporting information) [54]. The lnτ vs. T⁻¹ plots are almost linear when T > 7 K, which can be fitted by considering only the Orbach relaxation process using Arrhenius' law (τ = τ₀exp(U_eff/k_BT)). The obtained parameters are U_eff/k_B = 43.68 K and τ₀=1.4 × 10⁻⁶ s, which is consistent with the reported ranges for other SMMs (10⁻⁶–10⁻¹¹ s). However, when T < 7 K, the lnτ vs. T⁻¹ plots deviate from linearity, indicating the presence of other magnetic relaxation processes. Combining the Obrach and quantum tunneling of magnetization (QTM) relaxation processes using the equation τ⁻¹ = τ₀⁻¹ exp(−U_eff/k_BT) + τ_QTM⁻¹ gives the best fit (R² = 0.99865), where τ_QTM is the rate of QTM. The fit gives τ₀ = 6.4 × 10⁻⁷ s, τ_QTM = 6.0 × 10⁻³ s, and U_eff/k_B= 50.24 K (Fig. 3d). To slow relaxation, the AC magnetic susceptibilities were measured under specific DC fields at 2 K. Under DC fields from 1000 Oe to 4000 Oe, the out-of-phase AC signal values increased significantly compared to those at 0 Oe (Fig. S22 in Supporting information), indicating that QTM is suppressed. Simultaneously, the frequency corresponding to the peak of the AC signal effectively decreases after applying the field and remains stable above 2500 Oe (Fig. S23 in Supporting information). The AC susceptibilities of cluster 1 were measured under an optimal 2500 Oe DC field. The temperature-dependent and frequency-dependent out-of-phase susceptibilities (χ") are significantly higher than that at zero field. The peak frequency is lower, and the peak temperature is slightly higher under the same frequency conditions. The Cole-Cole plot obtained by fitting the same generalized Debye model shows a semicircular shape (Fig. 4 and Fig. S24 in Supporting information). Similarly, the lnτ vs. T⁻¹ plot is almost linear in the high-temperature region; considering only the Orbach relaxation process fitted according to Arrhenius' law: U_eff/k_B = 64.55 K, and τ₀ = 1.7 × 10⁻⁷ s. In the low-temperature region, the lnτ vs. T⁻¹ plot deviates from linearity. As shown in Fig. 4d, combining the Obrach, direct and Raman relaxation processes using the equation τ⁻¹ = τ₀⁻¹exp(-U_eff/k_BT) + AT + CTⁿ, results in τ₀ = 9.3 × 10⁻⁸s, U_eff/k_B = 80.21 K, A = 1.97 K⁻¹ s⁻¹, C = 1.51 × 10⁻³ s⁻¹ K⁻ⁿ, n = 6.6 (R² = 0.99762). If the QTM relaxation process is also considered, the fitted τ_QTM = −8.05 × 10¹⁴ s, suggesting that QTM can be considered completely suppressed at 2500 Oe [33, 55].

  Figure 3

  

  Figure 3.  (a) In-phase susceptibility (χ') and (b) out-of-phase susceptibility (χ'') vs. v under zero Oe DC field for 1. (c) Cole-Cole plots for AC susceptibilities under zero DC field. The colored solid lines are the best fit to the generalized Debye model. (d) Arrhenius plots of lnτ vs. T^–1 for 1 under zero DC field.

  DownLoad: Full-Size Img PowerPoint

  Figure 4

  

  Figure 4.  (a) In-phase (χ') and (b) out-of-phase susceptibility (χ'') vs. v at indicated temperatures under a 2500 Oe DC field for cluster 1. (c) Cole-Cole plots for AC susceptibilities under a 2500 Oe DC field. The colored solid lines are the best fit to the generalized Debye model. (d) Arrhenius plots of lnτ vs. T^–1 for cluster 1.

  DownLoad: Full-Size Img PowerPoint

  POMs are redox reversible and maintain high structural stability while rapidly gaining and losing multiple electrons [56]. The cyclic voltammograms and controlled-potential coulometry experiments showed that the isomorphic clusters 1-3 have three redox processes, peak shapes, and potentials similar to those of the Fe₆W₁₈ cluster, which indicates the gradual reduction of the four Fe^Ⅲ centers. Based on the peak reduction currents, the three redox processes are correspond to the two-electron process, the one-electron process and the one-electron process, respectively (Figs. S25 and S26 in Supporting information). Clusters 1-3 are structurally similar to the reported cluster Fe₆W₁₈, which have only two Fe ions replaced by lanthanide ions at the periphery of the sandwich part [43]. Therefore, the Fe heteroatoms in clusters 1–3 (Ln₂Fe₄) can participate in the electrochemical redox process, whereas the Fe heteroatoms in the Fe₆W₁₈ cluster are electrochemically inert. Notably, among the POMs containing d-metal heteroatoms, only a few heteroatoms exhibit electrochemical activity [57]. The stability of cluster 1-3 in solution was confirmed by ESI-MS (Figs. S27-S29 in Supporting information).

  In summary, three {FeW₉O₃₄}-based 3d-4f clusters were obtained via a strategy that employs the slow release of lanthanide ions. The sandwich-like Ln₂Fe₄ clusters are the first 3d-4f clusters assembled from POMs containing d-metal heteroatoms. The in situ formation of [B-α-FeW₉O₃₄]¹¹⁻ and the slow release of Ln³⁺ play important roles in the formation of Ln₂Fe₄. Interestingly, the Dy₂Fe₄ cluster exhibits single molecule magnet properties with an 80 K energy barrier under an optimal DC field. The assembly of lacunary polyoxometalate-based 3d-4f SMMs is currently in progress.

  Declaration of competing interest

  The authors declare that there is no interest for this manuscript.

  Acknowledgments

  This work was supported by the National Natural Science Foundation of China (Nos. 21871224, 92161104, 92161203 and 21721001) and Innovation Laboratory for Sciences and Technologies of Energy Materials of Fujian Province (IKKEM No. RD2021040301).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2022.02.056.

  
  
• 1. [1]

     L.Z. Cai, Q.S. Chen, C.J. Zhang, et al., J. Am. Chem. Soc. 137 (2015) 10882–10885. doi: 10.1021/jacs.5b05320

  2. [2]

     R. Chen, C.L. Chen, M.H. Du, et al., Chem. Commun. 57 (2021) 3611–3614. doi: 10.1039/d0cc08132a

  3. [3]

     R. Chen, Z.H. Yan, X.J. Kong, et al., Angew. Chem. Int. Ed. 57 (2018) 16796–16800. doi: 10.1002/anie.201811211

  4. [4]

     W.P. Chen, G.J. Zhou, Z.L. Gou, et al., Chin. Chem. Lett. 32 (2021) 838–841. doi: 10.1016/j.cclet.2020.05.018

  5. [5]

     M.H. Du, S.H. Xu, G.J. Li, et al., Angew. Chem. Int. Ed. 61 (2022) e202116296.

  6. [6]

     N.F. Li, Q.F. Lin, Y.M. Han, et al., Chin. Chem. Lett. 32 (2021) 3803–3806. doi: 10.1016/j.cclet.2021.04.042

  7. [7]

     H.J. Lun, M.H. Du, D.H. Wang, et al., Inorg. Chem. 59 (2020) 7900–7904. doi: 10.1021/acs.inorgchem.0c00613

  8. [8]

     X. Wang, M.H. Du, H. Xu, et al., Inorg. Chem. 60 (2021) 5925–5930. doi: 10.1021/acs.inorgchem.1c00333

  9. [9]

     H.L. Zhang, Y.Q. Zhai, L. Qin, et al., Matter 2 (2020) 1481–1493. doi: 10.1016/j.matt.2020.02.021

  10. [10]

      R. Chen, Z.F. Hong, Y.R. Zhao, et al., Inorg. Chem. 58 (2019) 15008–15012. doi: 10.1021/acs.inorgchem.9b02112

  11. [11]

      H. Han, Y.S. Ding, X. Zhu, et al., Inorg. Chem. Front. 7 (2020) 4070–4076. doi: 10.1039/d0qi00792g

  12. [12]

      M.G.F. Vaz, M. Andruh, et al., Coord. Chem. Rev. 427 (2021) 213611–213634. doi: 10.1016/j.ccr.2020.213611

  13. [13]

      V. Das, R. Kaushik, F. Hussain, Coord. Chem. Rev. 143 (2020) 213271–213291.

  14. [14]

      J.W. Zhao, Y.Z. Li, L.J. Chen, et al., Chem. Commun. 52 (2016) 4418–4445. doi: 10.1039/C5CC10447E

  15. [15]

      J. Cai, X.Y. Zheng, J. Xie, et al., Inorg. Chem. 56 (2017) 8439–8445. doi: 10.1021/acs.inorgchem.7b01104

  16. [16]

      C. Dey, J. Cluster Sci. 33 (2022) 1839–1856. doi: 10.1007/s10876-021-02110-8

  17. [17]

      C. Li, N. Mizuno, K. Yamaguchi, et al., J. Am. Chem. Soc. 141 (2019) 7687–7692. doi: 10.1021/jacs.9b02541

  18. [18]

      J.X. Liu, X.B. Zhang, Y.L. Li, et al., Coord. Chem. Rev. 414 (2020) 213260–213276. doi: 10.1016/j.ccr.2020.213260

  19. [19]

      H.Y. Wang, X.Y. Zheng, L.S. Long, et al., Inorg. Chem. 60 (2021) 6790–6795. doi: 10.1021/acs.inorgchem.1c00622

  20. [20]

      B.S. Bassil, M. Ibrahim, R. Al-Oweini, et al., Angew. Chem. Int. Ed. 50 (2011) 5961–5964. doi: 10.1002/anie.201007617

  21. [21]

      B. Godin, Y.G. Chen, J. Vaissermann, et al., Angew. Chem. Int. Ed. 44 (2005) 3072–3075. doi: 10.1002/anie.200463033

  22. [22]

      X.B. Han, Y.G. Li, Z.M. Zhang, et al., J. Am. Chem. Soc. 137 (2015) 5486–5493. doi: 10.1021/jacs.5b01329

  23. [23]

      X.B. Han, Z.M. Zhang, T. Zhang, et al., J. Am. Chem. Soc. 136 (2014) 5359–5366. doi: 10.1021/ja412886e

  24. [24]

      L. Huang, S.S. Wang, J.W. Zhao, et al., J. Am. Chem. Soc. 136 (2014) 7637–7642. doi: 10.1021/ja413134w

  25. [25]

      M. Ibrahim, Y. Lan, B.S. Bassil, et al., Angew. Chem. Int. Ed. 50 (2011) 4708–4711. doi: 10.1002/anie.201100280

  26. [26]

      Z.J. Liu, X.L. Wang, C. Qin, et al., Coord. Chem. Rev. 313 (2016) 94–110. doi: 10.1016/j.ccr.2015.12.006

  27. [27]

      S.S. Mal, U. Kortz, Angew. Chem. Int. Ed. 44 (2005) 3777–3780. doi: 10.1002/anie.200500682

  28. [28]

      L. Qiao, M. Song, A. Geng, et al., Chin. Chem. Lett. 30 (2019) 1273–1276. doi: 10.1016/j.cclet.2019.01.024

  29. [29]

      Y. Sakai, K. Yoza, C.N. Kato, et al., Chem. Eur. J. 9 (2003) 4077–4083. doi: 10.1002/chem.200305182

  30. [30]

      X. Feng, W. Zhou, Y. Li, et al., Inorg. Chem. 51 (2012) 2722–2724. doi: 10.1021/ic202418y

  31. [31]

      M. Ibrahim, V. Mereacre, N. Leblanc, et al., Angew. Chem. Int. Ed. 54 (2015) 15574–15578. doi: 10.1002/anie.201504663

  32. [32]

      Y.D. Lin, R. Ge, C.B. Tian, et al., Chem. Commun. 57 (2021) 8624–8627. doi: 10.1039/d1cc03262c

  33. [33]

      E. Tanuhadi, E. Al-Sayed, G. Novitchi, et al., Inorg. Chem. 59 (2020) 8461–8467. doi: 10.1021/acs.inorgchem.0c00890

  34. [34]

      Y. Chen, Z.W. Guo, X.X. Li, et al., CCS Chem. 3 (2021) 1232–1241.

  35. [35]

      S.R. Li, H.Y. Wang, H.F. Su, et al., Small. Methods 5 (2020) 2000777.

  36. [36]

      H.H. Wu, S. Yao, Z.M. Zhang, et al., Dalton. Trans. 42 (2013) 342–346. doi: 10.1039/C2DT32318D

  37. [37]

      L.L. Li, G.J. Cao, J.W. Zhao, et al., Inorg. Chem. 55 (2016) 5671–5683. doi: 10.1021/acs.inorgchem.6b00754

  38. [38]

      G. Abbas, Y. Lan, V. Mereacre, et al., Inorg. Chem. 48 (2009) 9345–9355. doi: 10.1021/ic901248r

  39. [39]

      J. Bai, F. Li, Z. Sun, et al., Inorg. Chim. Acta 458 (2017) 1–7. doi: 10.1016/j.ica.2016.12.010

  40. [40]

      D. Barats, G. Leitus, R. Popovitz-Biro, et al., Angew. Chem. Int. Ed. 47 (2008) 9908–9912. doi: 10.1002/anie.200803966

  41. [41]

      M.P.E.M. Limanski, E. Droste, K. Burgemeister, et al., J. Cluster Sci. 13 (2002) 369–379. doi: 10.1023/A:1020551016077

  42. [42]

      P.T. M, J.P. Wang, Y. Shen, et al., Cryst. Growth. Des. 7 (2007) 603–605. doi: 10.1021/cg060751y

  43. [43]

      J.D. Compain, P. Mialane, A. Dolbecq, et al., Angew. Chem. Int. Ed. 48 (2009) 3007–3081.

  44. [44]

      M. Llunell, D. Casanova, J. Girera, P. Alemany, S. Alvarez, SHAPE, Version 2.0, Shape Software, Barcelona, 2010.

  45. [45]

      H. Zabrodsky, S. Peleg, D. Avnir, J. Am. Chem. Soc. 115 (1993) 8278–8289. doi: 10.1021/ja00071a042

  46. [46]

      M.N. Akhtar, V. Mereacre, G. Novitchi, et al., Chem. Eur. J. 15 (2009) 7278–7282. doi: 10.1002/chem.200900758

  47. [47]

      A. Baniodeh, C.E. Anson, A.K. Powell, Chem. Sci. 4 (2013) 4354–4361. doi: 10.1039/c3sc52041b

  48. [48]

      L.J. Chen, F. Zhang, X. Ma, et al., Dalton 44 (2015) 12598–12612. doi: 10.1039/C5DT00696A

  49. [49]

      S.F.M. Schmidt, C. Koo, V. Mereacre, J. Park, D.W. Heermann, et al., Inorg. Chem. 56 (2017) 4796–4806. doi: 10.1021/acs.inorgchem.6b02682

  50. [50]

      J.J. Jiang, Y.H. Chen, L.L. Liu, et al., Inorg. Chem. 58 (2019) 15853–15863. doi: 10.1021/acs.inorgchem.9b02273

  51. [51]

      X.Y. Zheng, H. Zhang, Z. Wang, et al., Angew. Chem. Int. Ed. 56 (2017) 11475–11479. doi: 10.1002/anie.201705697

  52. [52]

      G. Peng, Y.Y. Zhang, B. Li, et al., Dalton. Trans. 47 (2018) 17349–17356. doi: 10.1039/c8dt03613f

  53. [53]

      Y. Huo, Y.C. Chen, S.G. Wu, et al., Inorg. Chem. 58 (2019) 1301–1308. doi: 10.1021/acs.inorgchem.8b02788

  54. [54]

      K.C. Mondal, A. Sundt, Y.H. Lan, et al., Angew. Chem. Int. Ed. 51 (2012) 7550–7554. doi: 10.1002/anie.201201478

  55. [55]

      F. Shao, B. Cahier, Y.T. Wang, et al., Chem. Asian. J. 15 (2020) 391–397. doi: 10.1002/asia.201901511

  56. [56]

      J.J. Chen, M. Symes, L. Cronin, Nat. Chem. 10 (2018) 1042–1047. doi: 10.1038/s41557-018-0109-5

  57. [57]

      Y.Y. Liu, S.F. Lu, H.N. Wang, Adv. Energy Mater. 7 (2017) 1601224. doi: 10.1002/aenm.201601224

• 

  Figure 1  The structures of (a) [B-α-SbW₉O₃₃]⁹⁻ unit; (b) [B-α-FeW₉O₃₄]¹¹⁻ unit. (c, d) Ball-and-stick/polyhedral view of [(B-α-FeW₉O₃₄)₂Fe₂Dy₂(H₂O)₄]¹⁰⁻. (e) The metal core of 2Dy³⁺ and 4Fe³⁺ ions in cluster 1. (f-h) The coordination geometries of DyO₈/FeO₆/FeO₄. Color key: W/WO₆, blue; Fe/FeO₆/FeO_4, green; Dy/DyO₈, purple; Sb, black and O, pink.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  χ_mT vs. T in the range of 2–300 K under an applied field of 1000 Oe for clusters 1–3. Inset: Enlarged view of the upturn of χ_mT (cluster 1).

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) In-phase susceptibility (χ') and (b) out-of-phase susceptibility (χ'') vs. v under zero Oe DC field for 1. (c) Cole-Cole plots for AC susceptibilities under zero DC field. The colored solid lines are the best fit to the generalized Debye model. (d) Arrhenius plots of lnτ vs. T^–1 for 1 under zero DC field.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) In-phase (χ') and (b) out-of-phase susceptibility (χ'') vs. v at indicated temperatures under a 2500 Oe DC field for cluster 1. (c) Cole-Cole plots for AC susceptibilities under a 2500 Oe DC field. The colored solid lines are the best fit to the generalized Debye model. (d) Arrhenius plots of lnτ vs. T^–1 for cluster 1.

  下载: 全尺寸图片 幻灯片
•