A new series of Schiff base Ni(II) 4 cubanes: Evaluation of magnetic coupling via carboxylate bridges

The Schiff base ligand H 2 L has been synthesized by condensation of 2-aminophenol with 3-methyl-1-phenyl-4-formylpyrazol-5-one and reaction between H 2 L and Ni(OOCR 3 ) 2 , R = H in 1 , –CH 3 in 2 , Cl in 3 , yielded three new tetranuclear Ni(II) complexes. The complexes have been characterized by elemental analysis, IR-and ES-MS spectroscopy. Their structures as well as structure of ligand were determined by single-crystal X-ray diffraction. Compounds 1 – 3 possesses tetranuclear cubane-like structures containing [Ni 4 L 4 (R 3 COO) 2 ] 2 (cid:1) complex anions, which are charge balanced by two triethyl ammonium cations. Furthermore, the crystal structure of the Ni(II) cubane compound containing trichloroacetate bridging ligands, is reported for the ﬁrst time. Variable temperature magnetic susceptibility measurements revealed interplay between ferromagnetic and antiferromagnetic exchange in the tetranuclear cubane-like compounds 1 – 3 , in which ferromagnetic interactions were enhanced by introducing carboxylate bridging groups. DFT calculations supported the analysis of magnetic data. (cid:1) 2021 The Authors. Published by Elsevier Ltd. ThisisanopenaccessarticleundertheCCBYlicense(http


Introduction
Tetranuclear cubane-like M 4 O 4 complexes including high-spin cobalt(II) and nickel(II) metal centers have been a very important class of molecules due to their potential application as a new type of magnetic materials [1,2]. The magnetic behavior of Co(II) and Ni (II) complexes with a cubane-like inner core, typically with Schiff base ligands, is very intriguing and depends on several structural factors. In general, the Ni 4 O 4 cubane may exhibit either ferromagnetic or antiferromagnetic interactions among the nickel ions, depending upon the combination of the cubane topology and Ni-O-Ni angles. The structure of the cubane cores can be modified by introducing alterations in the coordinating ligands and the change in crystallization conditions such as solvent, temperature, guest molecules and pH [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20]. This is apparently helpful for the synthesis of new molecular magnets allowing preparation of the Co(II) or Ni(II) cubanes with the dominant ferromagnetic interactions among the metal centers and non-negligible magnetic anisotropy of the ground state. Such compounds can exhibit slow relaxation of magnetization of molecular origin characteristic for Single-Molecule Magnets (SMMs) [3][4][5][6].
A great number of the cubane structures based on Schiff bases are formed on the basis of derivatives of salicylic aldehyde [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20]. Such complexes are widely investigated due to the simplicity of their synthesis. Recently, we have started to develop new metal complexes based on polydentate Schiff-base ligands on 4-acylpyrazol-5-ones basis [21], which could be considered as a structural analog of salicylic aldehyde. In the present study, we have synthesized new Schiff-base ligand H 2 L by the condensation of oaminophenol with 3-methyl-1-phenyl-4-formylpyrazol-5-one (Scheme 1). We have also prepared three new cubane complexes by a reaction between H 2 L and Ni(OOCR) 2 ÁxH 2 O (R = CH 3 , (CH 3 ) 3 C and CCl 3 ). The alteration of the carboxylate ligands was done in an attempt to modify the cubane structures and thus also the Ni-O-Ni exchange pathways. Furthermore, it must be stressed that trichloroacetate is a rather uncommon ligand in the nickel coordination compounds and its bridging function has been structurally confirmed only in a couple of compounds yet [22]. None of them possessed the cubane like structure and thus, here we present the first Ni cubane compound bridged by trichloroacetate ligand.
Moreover, we present a detailed magnetic study explaining the observed unexpected magnetic behavior of complexes in correlation to their structures.

Materials and methods
All chemicals and solvents used for the synthesis were of analytical grade. 2-Aminophenol and Ni(CH 3 COO) 2 Á4H 2 O were commercially available and used as received without further purification. The starting nickel pivalate and trichloroacetate were synthesized according to known procedures [23]. The synthesis of new compounds was carried out with the use of trimethylacetic ant trichloracetic acid (Acros Organics). 3-methyl-1-phenyl-4formylpyrazol-5-one was synthesized as described in the literature [24,25]. The purity and composition of the prepared complexes were confirmed by means of elemental analysis, electrospray-ionization mass-spectrometry (ESI-MS), and FT-IR spectroscopy, and single-crystal X-ray structure analysis.
Elemental analyses of C, H, and N were performed with a Perkin-Elmer 240 C analyzer. 1 H NMR spectra were recorded on a Bruker VXR-400 spectrometer at 400 MHz from solutions in DMSO-d 6 . IR spectra were measured with a Spectrum Two PerkinElmer Inc spectrometer in the range 4000-400 cm À1 . Electrospray mass spectra of complexes were measured with the Finnagan TSQ 700 mass spectrometer in positive ion mode. Samples were prepared at a concentration of~2 mg/ml MeOH. The mass spectra were acquired over the m/z range of 50-2000; several scans were averaged to provide the final spectrum.
Magnetic measurements were carried out on a SQUID MPMS 5S Quantum Design magnetometer in the 2-300 K temperature range. The diamagnetic contributions of the samples were estimated from Pascal's constants.

X-ray crystallography
Diffraction data were collected using the standard rotational method on a Bruker SMART APEX II automated diffractometer equipped with a CCD detector and a monochromatic radiation source (H 2 L, MoKa radiation, k = 0.71073 Å) or a Bruker D8 Quest diffractometer equipped with a Photon 100 CMOS detector using the Mo-Ka radiation (compounds 1-2) or a Super-Nova diffractometer equipped with a HyPix-3000 detector (Cu-Ka radiation k = 1.54184 Å) (compound 3). Data collection, data reduction, and cell parameters refinements were performed using the Bruker Apex III software package [26]. The molecular structures were solved by direct methods SHELXS-2014 and all non-hydrogen atoms were refined anisotropically on F 2 using full-matrix leastsquares procedure SHELXL-2014 [27]. All hydrogen atoms were found in differential Fourier maps and their parameters were refined using a riding model with U iso (H) = 1.2(CH) or 1.5(CH 3 ) U eq . The crystallographic parameters and X-ray diffraction experimental parameters are given in Table S1. Non-routine aspects of refinement: The crystals of 2 and 3 suffer from solvent loss and therefore, despite numerous attempts, we were not able to get diffraction data of sufficient quality to reasonably model the electron density corresponding to triethylammonium cations and cocrystallized solvent molecules. However, the positions of the Et 3 -NH + cations in 2 and 3 were localized and the figures of the (approximate) second coordination spheres in these compounds are shown in supplementary. Prior the final refinement, a SQUEEZE procedure incorporated to PLATON was used to subtract the electronic density corresponding to cations and solvent molecules [28].

Results and discussion
New ligand H 2 L was synthesized by the condensation of 2aminophenol with the corresponding pyrazolone in refluxing anhydrous methanol and in the presence of a catalytic amount of acetic acid. An important aspect of this type of ligand is related to its tautomerism, both in the solid state and solution [29]. These molecules may exist in four tautomeric forms differing by the positions of the hydrogen atom, which could be located at the oxygen (imine-ol (form -A)), nitrogen (imine-one (form -B)), and carbon (imine-one (form -C)) atoms of the pyrazolone ring or at the azomethine N atom (amine-one) (form -D) -Scheme 2.
IR, 1 H NMR, and XRD data reveal that ketoamine tautomeric Dform predominates for the azomethines of H 2 L both in solutions and in the solid state. Indeed. 1 H NMR spectra of H 2 L (Fig. S1) demonstrate double signals of CH groups of the aminomethylene fragment @CHANHA at 8.68 ppm a doublet of the NH-groups at 11.57 ppm. Moreover, IR spectra of H 2 L show intense bands at 1668 cm À1 attributed to m(C pyr @O) vibrations (Fig. S3a). The crystal structure of H 2 L was determined by single-crystal X-ray diffraction. H 2 L1 crystallized in the orthorhombic space group Pca2 1 . The structure of H 2 L molecule is shown in Fig. 1. The crystallographic data are listed in Tables S1. The X-ray data on these compounds confirm that the present ligand exists in the amine-form.

Coordination compounds on the H 2 L basis
The synthesis of 1-3 was carried out by reacting H 2 L with the nickel carboxylates in methanol in presence of triethylamine as a base at ambient temperature. Detailed synthetic procedures are given in the experimental section. Complexes 1-3 were characterized by standard analytical/spectroscopic techniques and their solid-state structures were established by single-crystal X-ray diffraction analysis.
According to the element analysis data, these complexes possess Ni:L = 1:1 M ratio. ESI -mass spectroscopic data clearly suggests this ratio. Electrospray mass spectrum revealed a sequence of peaks corresponding to {NiL + H + }, and {Ni 2 L 2 + H + } fragments with the correct isotopic spacings consistent with the charge (Fig. S2). ESI-data indicate that the tetranuclear cluster is fragile in solution and broke up under the conditions of the MS experiment The infrared spectra for all complexes have been analyzed and compared with the free ligand spectrum (Fig. S3a-d). No stretching vibrations of the @CRANHA fragments were found in IR-spectra of 1-3, while the strong and sharp bands in the region of 1595-1599 cm -1 that can be assigned to the azomethine C@N stretching vibrations of the coordinated ligand were observed. A new intense absorption band, which is absent in the spectra of free ligand, appeared in the spectra of 1-3 with a maximum at 1352-1354 cm À1 . This could be assigned to C pyr -O stretching vibrations confirming thus the formation of the imine-ole form of ligand upon coordination. Furthermore, two intense bands of the characteristic of the carboxylate-anion stretching are found in the regions 1621-1625 cm À1 and 1519-1524 cm À1 .
The crystal structures of 1-3 have been determined by X-ray single-crystal diffraction. Complexes 1-3 crystallizes in the monoclinic space groups P2 1 /c (2, 3) and P2 1 /n (1). The crystallographic data are listed in Tables S1. The asymmetric unit of 1-3 contains anionic tetranuclear complex molecules and in the case of 1, the positions of the triethylammonium cations, co-crystallized water and acetic acid molecules were determined. However, in 2 and 3, the electron density outside of the complex anions was affected by partial solvent loss and thus it was not possible to be modeled reasonably and thus, it was removed using SQUEZZE procedure [28].
Molecular structures of complex anions are shown in Fig. 1. Complex anions exhibit distorted cubane topology of their coordination polyhedral in which the Ni(II) and the l 3 -O oxygen atoms occupy alternate vertices of the cube (Fig. 1). The Ni 4 O 4 cubane cores are encapsulated by an organic shell of four Schiff base ligands (in double deprotonated form, L 2À ) and two coordinated carboxylate-anions. Each Ni(II) atom is six-coordinated by five oxygen atoms and one imino nitrogen atom. Four out of five oxygen atoms originate from the Schiff base ligands and one is from the bridging carboxylate ligand. The H 2 L ligand coordinates nickel atoms in a tridentate manner (the ONO atom donor set) by one imine nitrogen atom, oxygen atom from pyrazole ring and phenolate oxygen atom, which acts as l 3 -bridging atom. The metal-ligand bond lengths are very similar for both types of bonds, whereas the Ni-N bonds are a bit shorter than Ni-O bonds (Fig. 1). Remarkably, the longest bond lengths exceeding 2.10 Å are observed for some of the Ni-l 3 -O bonds.
Considering only the Ni 4 O 4 cubane cores and the nature of the bridging ligands for the different faces of the cube, the D 2d symmetry could be assumed. The cubane cores are elongated and the NiÁÁÁNi distances between the Ni atoms bridged by the carboxylate ligands (forming opposite faces of cubane) are shorter (e.g. 2.95-2.98 Å for 1) than in four other faces (d(NiÁÁÁNi) = 3.20-3.27 Å for 1).
In crystal structure of 1, the Et 3 NH + cations are not linked to the complex anions by hydrogen bonding involving amine groups but remarkably, by weaker CAHÁÁÁO hydrogen bonding (d(CÁÁÁO) = 3.37-3.98 Å). The amine groups form rather strong NAHÁÁÁO hydrogen bonds with co-crystallized water molecules (d(NÁÁÁO) = 2.687(4) Å). These form OAHÁÁÁO hydrogen bonds between the coordinated acetate ligands, other water molecules and non-coordinated mole- cules of acetic acid. Some of them provide non-covalent bridging between pyrazolone rings (OAHÁÁÁN) and coordinated acetato ligands (non-covalent interactions are summarized in Supplementary Table S3).
In the case of crystal structures of 2 and 3, the detailed discussion of the non-covalent interactions is prevented by a heavy disorder of the Et 3 NH + cations and co-crystallized solvent molecules caused by the partial solvent loss. The positions of the cations and some of the solvent molecules can be located from difference maps and it can be concluded that in both compounds the present non-covalent interactions are similar to those observed in 1 with Et 3 NH + cations forming NAHÁÁÁO hydrogen bonds with co-crystallized water molecules ( Fig. S5a and S5b). The noticeable difference is the absence of the co-crystallized acid molecule when compared to the crystal structure of 1.

Magnetic properties
The thermal variation of the molar magnetization of 1-3 was measured over the range of 2-300 K and additionally, the magnetic  Table S2. field dependent isothermal magnetizations were acquired up to 5T. X-ray diffraction (XRD) measurements on the powder of 1-3 confirmed the retention of purity and crystallinity of related samples for magnetic measurement (Fig. S4 a-c). The tetranuclear compounds 1-3 have the room temperature value of the effective magnetic moment in the 6-6.5 l B range, which corresponds to four magnetic centers with S = 1 having the g-factor slightly above free-electron value, which is typical for hexacoordinate Ni(II) ions [30]. At temperatures below 50 K, the increase of l eff is observed followed by its decrease on cooling to 2 K. Such behavior is most probably caused by the interplay of ferromagnetic and antiferromagnetic exchange interactions among Ni(II) ions influenced also by the single-ion zero-field splitting (ZFS). This complex situation is also reflected in the isothermal magnetization data, which does not follow Brillouin function and does not saturate at the maximal magnetic field.
Thus, the spin Hamiltonian comprising the isotropic exchange, ZFS and Zeeman term was postulated as followŝ where a-direction of the magnetic field is defined as B a = B(sin(h)cos (u), sin(h)sin(u), cos(h)) [31]. Such general spin Hamiltonian possesses too many free parameters and therefore the theoretical calculations at DFT level were performed to estimate the isotropic exchange parameters J kl (vide infra). Furthermore, we assumed all ZFS parameters equal for the given tetranuclear compound and also the g-tensors were treated as isotropic. The DFT calculations showed that Ni(II) ions bridged also by the carboxylic groups pos- Fig. 2. Magnetic data for compounds 1-3. Temperature dependence of the effective magnetic moment and the isothermal magnetizations measured at T = 2, 4, and 6 K. The empty circles represent the experimental data points and the lines represent the best fits calculated by using Eq. (1) with parameters listed in Table 1. Whereas other interactions mediated solely by phenolato groups are either weakly antiferromagnetic or weakly ferromagnetic (J ph ) and as expected, there is linear correlation between J ph and average <(Ni-O-Ni) angle. Therefore, these J ph (J 13, J 14 , J 23 , J 24 ) were calculated during fitting procedure as J kl = a J + b J Â<(Ni-O-Ni) kl , which significantly reduced the number of varied parameters. We employed similar approach previously for analogous systems [3]. Next, both temperature and field-dependent data were fitted simultaneously for both possible signs of the D-parameters, and the best-fits results are summarized in Table 1 [32]. The better agreement with the experimental data was acquired for a positive value of D-parameters, and these fits are shown in Fig. 2. It is evident that the modification of bridging carboxylate acid led to a variation of both the isotropic exchange parameters and ZFS parameters - Table 1. Next, the summary of similar Ni 4 O 4 cubanes having octahedrally coordinated Ni atoms and two carboxylato-bridging ligands similarly to compounds 1-3 is presented in Table 2. First, it is obvious that various carboxylato ligands were already utilized in this part of the coordination chemistry, but this is first time when trichloroacetato ligand was used. Secondly, both antiferromagnetic and ferromagnetic exchange is reported for these carboxylatobridging ligands ranging from À10 to +17 cm À1 and no clear correlation with selected structural parameters, like interatomic distance between two Ni atoms or average Ni-O-Ni angle, was found. Moreover, it must be pointed out that in all previously published results in Table 2, the authors used temperature dependent magnetic data (cT vs T) for analyzing magnetic properties. Furthermore, only in case of three compounds, magnetic data were analyzed together with ZFS terms. The fitted D-values were 3.1 cm À1 or À2.2 cm À1 for IBOGAI, 7.8 cm À1 for XIWFUG and 8.5 cm À1 for XIWGAN. However, even large values of |D| up to 10-15 cm À1 can be expected for octahedral Ni(II) ions having the heterogenous coordination sphere and the large deviations from the ideal octahedron geometry [33,34]. To conclude, the nature of carboxylato bridging ligand has deep impact on the magnetic properties, albeit there is no obvious magneto-structural correlation.

DFT calculations
The complexity of the isotropic exchange interactions in 1-3 was assessed also by DFT calculations employing ORCA 4.2 ab initio computation package [44]. The respective molecular structures were extracted from the experimental X-ray data followed by the optimization of the hydrogen atom's positions with PBE functional [45]. Next, B3LYP functional [46] was utilized for the calculations of high-spin (HS) states and several broken-symmetry spin states (BS) in order to calculate individual J-parameters using Ruiz's approach [47]. In all calculations, def2-TZVP basis set was used for all atoms except for carbon and hydrogen atoms, for which def2-SVP basis set was used [48]. Moreover, RIJCOSX approximation [49] was used to speed up the calculations together with def2/J auxiliary basis set [50]. The calculations for 1-3 were based on the following spin Hamiltonian and the values of the calculated J-parameters are listed in Table 3. An example of BS calculation for 3 is visualized in Fig. 3 [51]. Evidently, a relatively strong ferromagnetic exchange is triggered by the carboxylate-bridging ligands OOCR 3 and its strength is only slightly modified by the alternation of carboxylic groups. Comparison with isotropic exchange parameters (Table 1) determined from the experimental data shows that DFT calculations overestimates this type of the exchange. It is well-known that the isotropic exchange in Ni 4 cubane is dependent on Ni-O-Ni angle, [3][4][5]52] and also in this series of compounds 1-3, the magneto-structural linear correlation can be established - Fig. 4.

Conclusions
In summary, we have reported on the synthesis of three new cubane complexes 1-3 with were prepared by reactions between nickel carboxylates and Schiff base ligand H 2 L. Determination of the crystal structures confirmed cubane-like structure with the

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.