[Tb 4 (OH) 4 ]-Cuboid Complex Dianion Stabilized with Six Carboxylate Bridges and Four Diketonate Caps

: A cubane-like complex dianion [Tb 4 (OH) 4 (tfa) 6 (hfac) 4 ] 2 − was synthesized, accompanied by two counter cations of [Ni(hfac)(2pyIN) 2 ] + , where Htfa, Hhfac, and 2pyIN stand for triﬂuoroacetic acid, 1,1,1,5,5,5-hexaﬂuoropentane-2,4-dione, and 4,4,5,5-tetramethyl-2-pyridylimidazolin-1-oxyl, respectively. In the complex anion, each Tb ion was capped with hfac, and each Tb ··· Tb diagonal was bridged with tfa, thus forming an approximate Td symmetry, though the whole molecular formula was crystallographically independent in an orthorhombic Pbca space group. The ionic charge was − 2 in total. The magnetic study revealed that the complex ions were magnetically isolated from each other. Practically no 4f–4f superexchange interaction was operative, while the 2p–3d ferromagnetic coupling seemed to be appreciable, as anticipated from the proposed magnetostructural relationship.


Preparation
The complex formation was conducted according to the method for the known lanthanide-nitroxide complexes [22,23]  but a highly symmetric architecture and dianionic valence is rare in the lanthanide cubane-like chemistry, to our knowledge.

Preparation
The complex formation was conducted according to the method for the known lanthanide-nitroxide complexes [22,23]

Structural Analysis
X-Ray diffraction data of a single crystal of 1 were collected on a Saturn 70 hybrid pixel array detector (Rigaku, Tokyo, Japan) with graphite monochromated Mo Kα radiation (λ = 0.71073 Å). The structure was solved by a heavy-atom method and expanded using Fourier techniques in the Olex2 program [26]. The parameters were refined on Shelxl [27]. The thermal displacement parameters of non-hydrogen atoms were refined anisotropically. Hydrogen atoms were located at calculated positions and treated as "riding". A disorder model was applied to four trifluoromethyl groups. Considerable residual electron densities were found in a void space, and solvent heptane molecules could not be placed, possibly owing to severe disorder. A "squeeze" treatment [28] was applied to 183 residual electrons in a void space of 1016 Å 3 per asymmetric cell, corresponding to three heptane molecules as a solvent mask. Selected data are as follows: C111H122F54N12Ni2O32Tb4, orthorhombic Pbca, a = 32.1908

Structural Analysis
X-Ray diffraction data of a single crystal of 1 were collected on a Saturn 70 hybrid pixel array detector (Rigaku, Tokyo, Japan) with graphite monochromated Mo Kα radiation (λ = 0.71073 Å). The structure was solved by a heavy-atom method and expanded using Fourier techniques in the Olex2 program [26]. The parameters were refined on Shelxl [27]. The thermal displacement parameters of non-hydrogen atoms were refined anisotropically. Hydrogen atoms were located at calculated positions and treated as "riding". A disorder model was applied to four trifluoromethyl groups. Considerable residual electron densities were found in a void space, and solvent heptane molecules could not be placed, possibly owing to severe disorder. A "squeeze" treatment [28] was applied to 183 residual electrons in a void space of 1016 Å 3 per asymmetric cell, corresponding to three heptane molecules as a solvent mask. Selected data are as follows: C 111 H 122 F 54 N 12 Ni 2 O 32 Tb 4 , orthorhombic Pbca, a = 32.1908(6), b = 29.0734(5), c = 32.2901(6) Å, V = 30220.1(9) Å 3 , Z = 8, d = 1.721 g cm −3 , µ(MoKα) = 2.228 mm −1 , R(F) (I > 2σ(I)) = 0.0526, wR(F 2 ) (all reflections) = 0.1185, goodnessof-fit parameter = 0.976, T = 96 K for 41279 reflections. For experimental details and geometrical parameter tables, see CCDC reference number 2153576.

Magnetic Measurements
Direct and alternating current (dc and ac, respectively) magnetic susceptibilities of 1 were measured on MPMS-XL7 and MPMS3 SQUID magnetometers (Quantum Design Inc., San Diego, CA, USA). For the dc measurements, the magnetic responses were corrected with diamagnetic blank data of the sample holder measured separately, and further diamagnetic contribution was estimated from Pascal's constants [29].

Results and Discussion
Compound 1 was prepared from the starting materials [Tb(hfac) 3 (H 2 O) 2 ], Ni(CF 3 CO 2 ) 2 , and 2pyIN in a one-pot manner. The product was stable under ambient conditions, but the crystal solvent molecules incorporated escaped easily from the crystal lattice. The composition was confirmed to be [Ni(hfac)(2pyIN) 2 ] 2 [Tb 4 (OH) 4 (tfa) 6 (hfac) 4 ] by means of spectroscopic, analytic, and X-ray diffraction analyses. The single-crystal X-ray diffraction study clarified that the space group was orthorhombic Pbca with a whole formula unit crystallographically independent (Figure 1a). Considerable residual electron densities were found in a void space in a lattice, being ascribable to heptane. The complex cations and anions were discrete and accordingly seem to be magnetically isolated from each other. On the other hand, the Ni 2+ ion and 2pyIN ligands must be magnetically correlated because of the direct metal-radical coordination bonds (see below).
As Figure 1b shows, the anionic portion has a tetranuclear Tb 3+ cage cluster in 1. Each Tb 3+ ion is coordinated with three OH − ions, and each OHion plays the role of a tripodal bridge, forming a cube-like structure. The intra-cuboid Tb-O bond lengths varied from 2.329 (3)  The peripheral ligands were characterized as follows: each Tb 3+ ion was capped with an hfac anion, and every Tb···Tb diagonal was bridged with a tfa anion. Several tfa-bridged lanthanide complexes are known in the literature [30]. Thus, the tetra(µ 3 -hydroxo)-and hexa(µ-η 2 (O,O')-tfa)-bridged structure in 1 can be regarded as an approximate Td symmetry. The Tb 3+ ions are eight-coordinate, and the SHAPE analysis [31] indicates that the TbO 8 coordination spheres are best described as a triangular dodecahedron (TDD-8). Hydrogen bonds were found between the anion and cation moieties; the O···O distances were 2.991(5) and 2.918(4) Å for O2-H2···O25 and O4-H4···O29, respectively. Such interactions brought about an L-shaped cation-dianion-cation arrangement, and the resultant bulkiness seems to afford a void space in a crystal lattice, where the crystal solvent molecules were incorporated.

Magnetic Measurements
Direct and alternating current (dc and ac, respectively) magnetic susceptibilities of 1 were measured on MPMS-XL7 and MPMS3 SQUID magnetometers (Quantum Design Inc., San Diego, CA, USA). For the dc measurements, the magnetic responses were corrected with diamagnetic blank data of the sample holder measured separately, and further diamagnetic contribution was estimated from Pascal's constants [29].

Results and Discussion
Compound 1 was prepared from the starting materials [Tb(hfac)3(H2O)2], Ni(CF3CO2)2, and 2pyIN in a one-pot manner. The product was stable under ambient conditions, but the crystal solvent molecules incorporated escaped easily from the crystal lattice. The composition was confirmed to be [Ni(hfac)(2pyIN)2]2[Tb4(OH)4(tfa)6(hfac)4] by means of spectroscopic, analytic, and X-ray diffraction analyses. The single-crystal X-ray diffraction study clarified that the space group was orthorhombic Pbca with a whole formula unit crystallographically independent (Figure 1a). Considerable residual electron densities were found in a void space in a lattice, being ascribable to heptane. The complex cations and anions were discrete and accordingly seem to be magnetically isolated from each other. On the other hand, the Ni 2+ ion and 2pyIN ligands must be magnetically correlated because of the direct metal-radical coordination bonds (see below). As Figure 1b shows, the anionic portion has a tetranuclear Tb 3+ cage cluster in 1. Each Tb 3+ ion is coordinated with three OH − ions, and each OHion plays the role of a tripodal bridge, forming a cube-like structure. The intra-cuboid Tb-O bond lengths varied from 2.329(3) to 2.429(3) Å, the Tb-O-Tb angles from 105.06(11) to 108.67(11)°, and the O-Tb-O angles from 69.04(10) to 71.47(10)°. Thus, the cubane-like cluster was considerably distorted with the wider Tb-O-Tb angles and narrower O-Tb-O angles, compared to 90°. The peripheral ligands were characterized as follows: each Tb 3+ ion was capped with an hfac anion, and every Tb···Tb diagonal was bridged with a tfa anion. Several tfa-bridged lanthanide complexes are known in the literature [30]. Thus, the tetra(µ3-hydroxo)and hexa(µ-η 2 (O,O')-tfa)-bridged structure in 1 can be regarded as an approximate Td symmetry. The Tb 3+ ions are eight-coordinate, and the SHAPE analysis [31] indicates that the TbO8 coordination spheres are best described as a triangular dodecahedron (TDD-8). Hydrogen bonds were found between the anion and cation moieties; the O···O distances were 2.991(5) and 2.918(4) Å for O2-H2···O25 and O4-H4···O29, respectively. Such interactions brought about an L-shaped cation-dianion-cation arrangement, and the resultant bulkiness seems to afford a void space in a crystal lattice, where the crystal solvent molecules were incorporated.
The [Tb4(OH)4] 8+ core in 1 was similar to that of the known lanthanide cuboids [32][33][34][35]. A possible formation mechanism has already been discussed elsewhere [35]. In comparison with the known compounds, it should be noted that the present complex had peripheral ligands located in a highly symmetrical manner. Furthermore, a dianionic valence of the cuboidal complex ion is rather rare.
As for the cation moiety of 1 (Figure 1c,d), we can find two 2pyIN chelate rings around the Ni 2+ (3d 8 ) ion. The C 2py -C 2im -N-Ni torsion angle, φ, is proposed to be a convenient metric to assess magnetostructural relationship. The torsion angles were 1.1(6), 1.7(6), −2.5(6), and 0.7(5) • around the C40-C33-N1-Ni1, C52-C45-N4-Ni1, C69-C62-N7-Ni2, and C74-C81-N12-Ni2 bonds, respectively. These small torsions guarantee the absence of d z 2/d x 2 -y 2-π* orbital overlap. A closely related 2p/3d = 1/1 complex [Ni(hfac) 2 (2pyIN)] was reported to display orthogonal d-π* orbital arrangement and ferromagnetic coupling [36]. According to the magnetostructural relationship, the critical |φ C |, where the exchange coupling alters from ferro-to antiferromagnetic, was reported to be 12.8(8) • for the Cu 2+ and Ni 2+ compounds [7] and 21(1) or 26(3) • for the Ni 2+ compounds [37]. We can safely predict that the [Ni(hfac)(2pyIN) 2 ] + portions in 1 should behave as a ground quintet species. Figure 2a displays the dc magnetic susceptibility result on 1. A relatively temperatureinsensitive character appeared, indicating that each complex ion was magnetically isolated from each other. This finding also implies that the Tb-O-Tb superexchange coupling was negligible. At 300 K, the χ m T value was 50.5 cm 3 K mol −1 , in good agreement with the theoretical value 50.7 cm 3 K mol −1 expected from the sum of four Tb 3+ ions (11.81 cm 3 K mol −1 each with the angular moment J = L + S = 6 and the Landé factor g J = 3/2), two Ni 2+ ions (1.00 cm 3 K mol −1 each), and four radicals (0.375 cm 3 K mol −1 each). On cooling, the χ m T value gradually decreased and reached 45.6 cm 3 K mol −1 at 1.8 K, and this profile can partly be explained in terms of depopulation effect regarding the J z sublevels in the Tb 3+ ion. Such a lanthanide single-ion property sometimes disturbs exchange coupling analysis. In the present case, the 2p-3d interaction in the cation portions could not be evaluated. cm K mol each with the angular moment J = L + S = 6 and the Landé factor gJ = 3/2), two Ni 2+ ions (1.00 cm 3 K mol −1 each), and four radicals (0.375 cm 3 K mol −1 each). On cooling, the χmT value gradually decreased and reached 45.6 cm 3 K mol −1 at 1.8 K, and this profile can partly be explained in terms of depopulation effect regarding the J z sublevels in the Tb 3+ ion. Such a lanthanide single-ion property sometimes disturbs exchange coupling analysis. In the present case, the 2p-3d interaction in the cation portions could not be evaluated. The M-H curve (Figure 2a, inset) displays a positive slope even at 7 T, and the magnetization (34.4 NAµB at 7 T and 1.8 K when the sample was unfixed) was lower than the theoretical maximum value 44 NAµB, mainly because of the strong magnetic anisotropy of the Tb 3+ moment. Figure 2b shows the in-phase (χ') and out-of-phase parts (χ") of the molar ac magnetic susceptibilities for 1, measured in no dc bias field. In further experiments with applied fields of 1000 and 2000 Oe, no appreciable χ" was recorded either. There have been several reports on Dy-based SMMs including Dy4(OH)4 cuboids [33]. The Kramers theorem suggests that Dy 3+ ion has a double-well potential surface suitable for SMMs but Tb 3+ ion does not [38,39]. After combining the results of the Tb4-and Dy4-based cuboids, the single-ion magnetic property of Tb 3+ and Dy 3+ seems to regulate the whole molecular magnetism. This conclusion is compatible with the negligible 4f-4f coupling in the cuboid.

Conclusions
A dianionic cuboidal complex ion [Tb4(OH)4(tfa)6(hfac)4] 2− was synthesized, accompanied by two counter cations of [Ni(hfac)(2pyIN)2] + in a crystal lattice. In the complex anion, each Tb ion is capped with hfac and bridged with tfa in a highly symmetrical The M-H curve (Figure 2a, inset) displays a positive slope even at 7 T, and the magnetization (34.4 N A µ B at 7 T and 1.8 K when the sample was unfixed) was lower than the theoretical maximum value 44 N A µ B , mainly because of the strong magnetic anisotropy of the Tb 3+ moment. Figure 2b shows the in-phase (χ') and out-of-phase parts (χ") of the molar ac magnetic susceptibilities for 1, measured in no dc bias field. In further experiments with applied fields of 1000 and 2000 Oe, no appreciable χ" was recorded either. There have been several reports on Dy-based SMMs including Dy 4 (OH) 4 cuboids [33]. The Kramers theorem suggests that Dy 3+ ion has a double-well potential surface suitable for SMMs but Tb 3+ ion does not [38,39]. After combining the results of the Tb 4 -and Dy 4 -based cuboids, the single-ion magnetic property of Tb 3+ and Dy 3+ seems to regulate the whole molecular magnetism. This conclusion is compatible with the negligible 4f-4f coupling in the cuboid.

Conclusions
A dianionic cuboidal complex ion [Tb 4 (OH) 4 (tfa) 6 (hfac) 4 ] 2− was synthesized, accompanied by two counter cations of [Ni(hfac)(2pyIN) 2 ] + in a crystal lattice. In the complex anion, each Tb ion is capped with hfac and bridged with tfa in a highly symmetrical manner. The magnetic study revealed that the complex ions were magnetically isolated and that the Tb-O-Tb superexchange coupling was negligible, although the 2p-3d-4f heterospin material was realized Building upon the present architectural study, the use of a wide diversity of carboxylates and diketonates will be of interest to supramolecular chemistry involving a lanthanide cuboid building block.

Conflicts of Interest:
The authors declare no conflict of interest.