[CrIII8NiII6]n+ Heterometallic Coordination Cubes

Three new heterometallic [CrIII8NiII6] coordination cubes of formulae [CrIII8NiII6L24(H2O)12](NO3)12 (1), [CrIII8NiII6L24(MeCN)7(H2O)5](ClO4)12 (2), and [CrIII8NiII6L24Cl12] (3) (where HL = 1-(4-pyridyl)butane-1,3-dione), were synthesised using the paramagnetic metalloligand [CrIIIL3] and the corresponding NiII salt. The magnetic skeleton of each capsule describes a face-centred cube in which the eight CrIII and six NiII ions occupy the eight vertices and six faces of the structure, respectively. Direct current magnetic susceptibility measurements on (1) reveal weak ferromagnetic interactions between the CrIII and NiII ions, with JCr-Ni = + 0.045 cm−1. EPR spectra are consistent with weak exchange, being dominated by the zero-field splitting of the CrIII ions. Excluding wheel-like structures, examples of large heterometallic clusters containing both CrIII and NiII ions are rather rare, and we demonstrate that the use of metalloligands with predictable bonding modes allows for a modular approach to building families of related polymetallic complexes. Compounds (1)–(3) join the previously published, structurally related family of [MIII8MII6] cubes, where MIII = Cr, Fe and MII = Cu, Co, Mn, Pd.

have previously reported a metalloligand approach that enabled us to sy nuclearity heterometallic coordination capsules of paramagnetic transition a modular and predictable fashion [26][27][28][29]. This metalloligand, based o [M III L3] moiety shown in Figure 1 (HL = 1-(4-pyridyl)butane-1,3-dion tris(acac) coordinated octahedral transition metal ion, in which the ligand ised with a p-pyridyl donor group. In the fac-isomer of this metalloligand donor groups are orientated in such a way that combination with a square ion leads to the entropically favoured self-assembly of a cubic structure [3 report the syntheses, structures, and magnetic properties of three novel tet [Cr III [26][27][28][29].

Synthesis
1-(4-pyridyl)butane-1,3-dione (HL) and the metalloligand [Cr III L3] wer previously published procedures [26,31,32]. All reactions were carried out conditions. Solvents and reagents were used as received from commercial s tion: perchlorate salts of metal complexes with organic ligands are potentia Synthesis of [Cr III 8Ni II 6L24(H2O)12](NO3)12 (1). To a solution of [Cr III L mmol) in 10 mL of dichloromethane, a solution of Ni(NO3)2•6H2O (30 mg, 0 added in 10 mL of methanol. The solution was stirred for 18 h before bein allowed to stand. Dark orange X-ray quality crystals were obtained from th diethyl ether into the mother liquor. Yield of (1) = 69%. Elemental analysis (found): C 46. 16 (2). To a solution o mg, 0.2 mmol) in 10 mL of dichloromethane, a solution of Ni(ClO4)2•6H2 mmol) was added in 10 mL of acetonitrile. The solution was stirred for 18 filtered and allowed to stand. Brown X-ray quality crystals were obtaine from the diffusion of pentane into the mother liquor. Yield of (2) = 81%. Elem (%) calculated (found): C 44. 34

Synthesis
1-(4-pyridyl)butane-1,3-dione (HL) and the metalloligand [Cr III L 3 ] were prepared by previously published procedures [26,31,32]. All reactions were carried out under aerobic conditions. Solvents and reagents were used as received from commercial suppliers.  (1). To a solution of [Cr III L 3 ] (54 mg, 0.1 mmol) in 10 mL of dichloromethane, a solution of Ni(NO 3 ) 2 ·6H 2 O (30 mg, 0.1 mmol) was added in 10 mL of methanol. The solution was stirred for 18 h before being filtered and allowed to stand. Dark orange X-ray quality crystals were obtained from the diffusion of diethyl ether into the mother liquor. Yield of (1) = 69%. Elemental analysis (%) calculated (found): C 46. 16 (2). To a solution of [Cr III L 3 ] (108 mg, 0.2 mmol) in 10 mL of dichloromethane, a solution of Ni(ClO 4 ) 2 ·6H 2 O (73 mg, 0.2 mmol) was added in 10 mL of acetonitrile. The solution was stirred for 18 h before being filtered and allowed to stand. Brown X-ray quality crystals were obtained after 5 days from the diffusion of pentane into the mother liquor. Yield of (2) = 81%. Elemental analysis (%) calculated (found): C 44. 34  . To a solution of [Cr III L 3 ] (108 mg, 0.2 mmol) in 10 mL of dichloromethane, a solution of NiCl 2 (20 mg, 0.15 mmol) was added in 10 mL of tetrahydrofuran. The solution was stirred for 18 h before being filtered and allowed to stand. Brown X-ray quality crystals were obtained after room temperature evaporation of the mother liquor for 5 days. Yield of (3) = 58%. Elemental analysis (%) calculated (found): C 51.01 (50.79) H 3.81 (3.71) N 6.61 (6.68).

Crystallographic Details
Single-crystal X-ray diffraction data were collected for (1)-(3) at T = 100 K on a Rigaku AFC12 goniometer equipped with an enhanced sensitivity (HG) Saturn 724+ detector mounted at the window of an FR-E+ Superbright MoKα rotating anode generator with HF Varimax optics (70 µm focus) [33]. The CrysalisPro software package was used for instrument control, unit cell determination, and data reduction [34]. Due to very weak scattering power, single-crystal X-ray diffraction data for (1) and (2) were collected at T = 30.15 K using a synchrotron source (λ = 0.6889 Å) on the I19 beam line at Diamond Light Source on an undulator insertion device with a combination of double crystal monochromator, vertical and horizontal focussing mirrors, and a series of beam slits. The same software as above was used for data refinement. Crystals of all samples were sensitive to solvent loss, which resulted in crystal delamination and poor-quality X-ray diffraction data. To slow down crystal degradation, crystals of (1)-(3) were "cold-mounted" on MiTeGen Micromounts TM at T = 203 K using Sigma-Aldrich Fomblin Y ® LVAC (3300 mol. wt.), with the X-Temp 2 crystal cooling system attached to the microscope [35]. This procedure protected crystal quality and permitted collection of usable X-ray data. Unit cell parameters in all cases were refined against all data. Crystal structures were solved using Intristic Phasing as implemented in SHELXT [36]. All non-hydrogen atoms were refined with anisotropic displacement parameters, and all hydrogen atoms were added at calculated positions and refined using a riding model with isotropic displacement parameters based on the equivalent isotropic displacement parameter (U eq ) of the parent atom. All three crystal structures contain large accessible voids and channels that are filled with diffuse electron density belonging to uncoordinated solvent, whose electron contribution was accounted for by the PLATON/SQUEEZE routine ( (1) and (2)) [37], or by the SMTBX solvent masking routine, as implemented in OLEX2 software (3). To maintain reasonable molecular geometry, DFIX restraints were used in all three complexes.
Crystal Data for [Cr III 8 Ni

Magnetic and Spectroscopic Measurements
Direct current (dc) susceptibility and magnetisation data were measured on powdered, polycrystalline samples of (1) using a Quantum Design SQUID MPMS-XL magnetometer, operating between 1.8 and 300 K for dc applied magnetic fields ranging from 0 to 5 T. X-band EPR spectra were collected on powdered microcrystalline samples of (1) using a Bruker EMX spectrometer at the EPSRC UK National EPR Facility at The University of Manchester.

Structural Description
The heterometallic cubes [ with the corresponding Ni II salt in CH 2 Cl 2 /MeOH, CH 2 Cl 2 /MeCN, and CH 2 Cl 2 /THF, respectively. All three structures ( Figure 2) reveal a similar [Cr III 8 Ni II 6 ] cube-like metallic skeleton, with the eight Cr III ions located at the corners and the six Ni II ions located on the faces, approximately 1.4-2.3 Å above the Cr· · · Cr· · · Cr· · · Cr plane. The internal cavity volume of the cube is approximately 1400 Å 3 .
Molecules 2020, 25 (1) and (2), the axial positions are occupied by twelve water and twelve acetonitrile/water molecules (Ni II -O ≈2.1 Å; Ni II -N ≈2.1 Å), respectively. The cubes are therefore cationic (12+). The charge balancing nitrate or perchlorate anions for (1) and (2) respectively, are located both within the central cavity of the cube and in the void spaces between cubes. In contrast to (1) and (2) There are several close intermolecular contacts ( Figure 3) between the cages in the extended structures of (1)-(3). In (1), the closest inter-cluster contact is between the aromatic protons of the pyridyl group and the O-atom (2.3 Å) of a neighbouring Lligand. In (2) and (3)   There are several close intermolecular contacts (Figure 3) between the cages in the extended structures of (1)-(3). In (1), the closest inter-cluster contact is between the aromatic protons of the pyridyl group and the O-atom (2.3 Å ) of a neighbouring Lligand. In (2) and (3) down the a-, a-, and c-axis, respectively. Colour code as in Figure 1.

Magnetic Properties
As complexes (1)-(3) are structurally analogous, and for the sake of brevity, we discuss only the behaviour of a representative example, complex (1). The dc molar magnetic    down the a-, a-, and c-axis, respectively. Colour code as in Figure 1.

Magnetic Properties
As complexes (1)-(3) are structurally analogous, and for the sake of brevity, we discuss only the behaviour of a representative example, complex (1). The dc molar magnetic There are several close intermolecular contacts (Figure 3) between the cages in the extended structures of (1)-(3). In (1), the closest inter-cluster contact is between the aromatic protons of the pyridyl group and the O-atom (2.3 Å) of a neighbouring L − ligand. In (2) and (3), the closest contact is between the protons of the metalloligand methyl group, and the O-atom of a neighbouring L − ligand (2.3 Å) and the protons of a neighbouring methyl group (2.3 Å), respectively. Several other close inter-cluster contacts between neighbouring cubes exist, for example: Ar-H···O ≈2.5 Å and C-H···O ≈2.7 Å for (1), H 2 C-H···O ≈2.5 Å and H 2 O···H-CH 2 ≈2.7 Å for (2), and Ar-H···Cl ≈2.7 Å and C-H···Cl ≈2.8 Å for (3).  There are several close intermolecular contacts (Figure 3) between the cages in the extended structures of (1)-(3). In (1), the closest inter-cluster contact is between the aromatic protons of the pyridyl group and the O-atom (2.3 Å) of a neighbouring Lligand. In (2) and (3)

Magnetic Properties
As complexes (1)-(3) are structurally analogous, and for the sake of brevity, we discuss only the behaviour of a representative example, complex (1). The dc molar magnetic susceptibility, χ M , of a polycrystalline sample of (1) was measured in an applied magnetic field, B, of 0.1 T, over the 2-300 K temperature, T, range. The experimental results are shown in Figure 4 in the form of the χ M T product versus temperature, where χ M = M/B, and M is the magnetization of the sample. Due to the loss of lattice solvent during the evacuation of the sample chamber of the SQUID magnetometer, leading to an uncertainty in the molar mass of the measured sample, the T = 300 K χ M T product of (1) was scaled to 21.00 cm 3 mol −1 K, the expected value from the sum of Curie constants for a [Cr III 8 Ni II 6 ] unit, with g Cr = g Ni = 2.0, where g Cr and g Ni are the g-factors of Cr III and Ni II , respectively. Note that this rescaled value has a maximum deviation of 15% from the unscaled data.
lecules 2020, 25, x; doi: FOR PEER REVIEW As complexes (1)-(3) are structurally analogous, and for the sake cuss only the behaviour of a representative example, complex (1). The susceptibility, χM, of a polycrystalline sample of (1) was measured in a field, B, of 0.1 T, over the 2-300 K temperature, T, range. The exper shown in Figure 4 in the form of the χMT product versus temperatur and M is the magnetization of the sample. Due to the loss of lattice evacuation of the sample chamber of the SQUID magnetometer, leadin in the molar mass of the measured sample, the T = 300 K χMT product 21.00 cm 3 mol -1 K, the expected value from the sum of Curie consta unit, with gCr = gNi = 2.0, where gCr and gNi are the g-factors of Cr III an Note that this rescaled value has a maximum deviation of 15% from t Upon cooling, the value of χMT remains essentially constant to ap K, where it begins to increase, reaching a maximum of 21.8 cm 3 mol -1 this temperature, χMT falls rapidly to a minimum value of 18.5 cm 3 m The behaviour is suggestive of weak ferromagnetic exchange betwe ions, with the decrease in χMT below 6 K attributed to intermolecula exchange interactions, and/or zero-field splitting (zfs) effects primarily Ni II ions. Quantitative analysis of the susceptibility data via standard tion techniques is non-trivial due to the large nuclearity of the cluster enormous dimensions of the spin-Hamiltonian matrices. Even the tota trices used in approaches based on Irreducible Tensor Operator alge mensions than what is realistic for exact numerical matrix diagonaliza reported the use of computational techniques, known in theoretical statistical spectroscopy [38], to analyse the structurally similar [M III 8M M II = Co, Cu, Ni; n = 0-12) cubes [26][27][28]. We now extend this methodo exchange interactions present in (1). Due to the fact that the influence ions will mainly affect the measured properties at low temperatures Upon cooling, the value of χ M T remains essentially constant to approximately T = 75 K, where it begins to increase, reaching a maximum of 21.8 cm 3 mol −1 K at T = 6 K. Below this temperature, χ M T falls rapidly to a minimum value of 18.5 cm 3 mol −1 K at T = 2.0 K. The behaviour is suggestive of weak ferromagnetic exchange between the Cr III and Ni II ions, with the decrease in χ M T below 6 K attributed to intermolecular antiferromagnetic exchange interactions, and/or zero-field splitting (zfs) effects primarily associated with the Ni II ions. Quantitative analysis of the susceptibility data via standard matrix diagonalization techniques is non-trivial due to the large nuclearity of the cluster and the associated enormous dimensions of the spin-Hamiltonian matrices. Even the total spin (S) block matrices used in approaches based on Irreducible Tensor Operator algebra are of larger dimensions than what is realistic for exact numerical matrix diagonalization. Previously, we reported the use of computational techniques, known in theoretical nuclear physics as statistical spectroscopy [38], to analyse the structurally similar [M III 8 M II 6 ] n+ (M III = Cr, Fe; M II = Co, Cu, Ni; n = 0-12) cubes [26][27][28]. We now extend this methodology to quantify the exchange interactions present in (1). Due to the fact that the influence of the zfs of the Ni II ions will mainly affect the measured properties at low temperatures, the use of the isotropic spin-Hamiltonian (1) is sufficient to model the exchange interactions between Cr III and Ni II ions in the T = 300-6 K region: with i running over all constitutive metal centres, g is the isotropic g-factor,Ŝ is a spinoperator, J Cr-M is the isotropic exchange parameter between Cr III and M II centres, and µ B is the Bohr magneton. We assume common g-factors for both Cr III and Ni II (g Cr = g Ni = 2.0) since the 300 K χ M T product of (1) was scaled to the sum of its Curie constants, as explained above. We neglect any J Cr-Cr and J Ni-Ni terms as these centres are not connected as first neighbours. Using Hamiltonian (1), J Cr-Ni was determined to be +0.045 cm −1 . Variable-temperature and variable-field (VTVB) magnetization studies of (1) collected in the T = 2-10 K and B = 0.5-5 T temperature and field ranges ( Figure 5) are consistent with this picture. M reaches a value of 32.8 µ B at B = 5 T and T = 2 K, approaching the saturation value of 36 µ B , consistent with relatively small exchange-induced splittings that lead to the m S = 18 projection of the S = 18 total spin state, being the ground state at the highest measured magnetic field. The weak ferromagnetic exchange between the d 3 Cr III ions and the d 8 Ni II ions is as one would expect, mediated via the 1-(4-pyridyl)butane-1,3-dione ligand [26][27][28].
lecules 2020, 25, x; doi: FOR PEER REVIEW first neighbours. Using Hamiltonian (1), JCr-Ni was determined to be +0 temperature and variable-field (VTVB) magnetization studies of (1) co 10 K and B = 0.5-5 T temperature and field ranges ( Figure 5) are cons ture. M reaches a value of 32.8 μB at B = 5 T and T = 2 K, approaching t of 36 μB, consistent with relatively small exchange-induced splittings 18 projection of the S = 18 total spin state, being the ground state at th magnetic field. The weak ferromagnetic exchange between the d 3 Cr III ions is as one would expect, mediated via the 1-(4-pyridyl)butane-1, 28].

EPR Spectroscopy
X-band EPR spectra of a powdered sample of (1) at 5 and 10 K feature at ca. 2 kG (Figure 6). This is similar to spectra from the isolate and related [Cr III 8M II 6] and [Cr III 2M II 3] species [26,29], and arises from th with a near-axial zero-field splitting of |DCr| ca. 0.5-0.6 cm −1 . This is o a weak exchange interaction |JCr-Ni| with respect to |DCr|, and hence magnetic data. There are no clear features arising from the Ni II (S = 1) that |DNi| must be much larger than the microwave energy. We also related [Fe III 8M II 6] cube, which only showed EPR features due to Fe III [28 with |DNi| values of 5-10 cm −1 determined from magnetization

EPR Spectroscopy
X-band EPR spectra of a powdered sample of (1) at 5 and 10 K are dominated by a feature at ca. 2 kG (Figure 6). This is similar to spectra from the isolated [ [26,29], and arises from the Cr III (S = 3/2) ions with a near-axial zero-field splitting of |D Cr | ca. 0.5-0.6 cm −1 . This is only consistent with a weak exchange interaction |J Cr-Ni | with respect to |D Cr |, and hence consistent with the magnetic data. There are no clear features arising from the Ni II (S = 1) ions, which implies that |D Ni | must be much larger than the microwave energy. We also observed this for a related [Fe III 8 M II 6 ] cube, which only showed EPR features due to Fe III [28]. This is consistent with |D Ni | values of 5-10 cm −1 determined from magnetization studies of isolated [Ni II (pyridine) 4 X 2 ] complexes [39], and with high-field EPR studies of Ni II complexes with mixed N,O-donor sets [40]. magnetic data. There are no clear features arising from the Ni II (S = 1) ion that |DNi| must be much larger than the microwave energy. We also ob related [Fe III 8M II 6] cube, which only showed EPR features due to Fe III [28]. with |DNi| values of 5-10 cm −1 determined from magnetization stu [Ni II (pyridine)4X2] complexes [39], and with high-field EPR studies of Ni I mixed N,O-donor sets [40].

Conclusions
We have shown that the modular self-assembly of [M III L 3 ] metalloligands with simple M II salts can be exploited to construct large heterometallic coordination compounds of Cr III and Ni II . Compounds (1)-(3) join a growing family of [M III 8 M II 6 ] cubes, where M III = Cr and Fe and M II = Cu, Co, Mn, Pd, and Ni. The ability to build families of isostructural complexes containing different combinations of paramagnetic (and diamagnetic) metal centres aids the qualitative and quantitative understanding of magnetic properties and the underlying structural parameters that govern behaviour. Examples of large, heterometallic cages in which the 3d metal ions can be exchanged with other 3d metal ions are extremely rare.
Magnetic susceptibility and magnetization data show the presence of weak ferromagnetic exchange between the Cr III and Ni II ions, with J Cr-Ni = +0.045 cm −1 . EPR spectroscopy is consistent with the exchange interactions being much weaker than the zero-field splittings of both the Cr III and Ni II ions. Funding: This research was funded by the EPSRC (UK), grant numbers EP/N01331X/1 and EP/P025986/1, and by the VILLUM FONDEN (Denmark), grant 13376. We also thank the EP-SRC for funding the UK National EPR Facility.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding authors.