Trinuclear NiII-LnIII-NiII Complexes with Schiff Base Ligands: Synthesis, Structure, and Magnetic Properties

The reaction of the Schiff base ligand o-OH-C6H4-CH=N-C(CH2OH)3, H4L, with Ni(O2CMe)2·4H2O and lanthanide nitrate salts in a 4:2:1 ratio lead to the formation of the trinuclear complexes [Ni2Ln(H3L)4(O2CMe)2](NO3) (Ln = Sm (1), Eu (2), Gd (3), Tb (4)). The complex cations contain the strictly linear NiII-LnIII-NiII moiety. The central LnIII ion is bridged to each of the terminal NiII ions through two deprotonated phenolato groups from two different ligands. Each terminal NiII ion is bound to two ligands in distorted octahedral N2O4 environment. The central lanthanide ion is coordinated to four phenolato oxygen atoms from the four ligands, and four carboxylato oxygen atoms from two acetates which are bound in the bidentate chelate mode. The lattice structure of complex 4 consists of two interpenetrating, supramolecular diamond like lattices formed through hydrogen bonds among neighboring trinuclear clusters. The magnetic properties of 1–4 were studied. For 3 the best fit of the magnetic susceptibility and isothermal M(H) data gave JNiGd = +0.42 cm−1, D = +2.95 cm−1 with gNi = gGd = 1.98. The ferromagnetic nature of the intramolecular Ni···Gd interaction revealed ground state of total spin S = 11/2. The magnetocaloric effect (MCE) parameters for 3 show that the change of the magnetic entropy (−ΔSm) reaches a maximum of 14.2 J kg−1 K−1 at 2 K. A brief literature survey of complexes containing the NiII-LnIII-NiII moiety is discussed in terms of their structural properties.


Complexes
[{Ni(HL 12 ) 2 } 2 La(NO 3 [28] contain linear Ni-Ln-Ni moiety (Ni-Ln-Ni ≈ 169-180 • ) in which the central lanthanide ion is bridged to the two terminal Ni II ions through two phenolato oxygen atoms (Scheme 1b). The coordination environment around the Ni II ions consists of two imino nitrogen atoms and four oxygen atoms [25,26], or six oxygen atoms [28] and is distorted octahedral. In the case of [{NiL 14 (H 2 O)} 2 Ln(H 2 O)](trif) 3 the Ni II ion is five-coordinate, linked to the N 2 O 2 site of the ligand in the equatorial plane and to a water molecule in the apical position.
Complexes [Ni 2 (L28 ) 3  The magnetic susceptibility measurements of the above Ni II -Ln III -Ni II complexes reveal the presence of dominant antiferromagnetic interactions in the case of Ln III = Ce, Pr, Nd, and ferromagnetic interactions for Ln III = Gd, Tb, Dy, Ho, Er, and Yb. In the Ni 2 Gd complexes the presence of the isotropic Gd III ion facilitated the fitting of the magnetic data and the determination of the Ni···Gd exchange coupling constant which is approximately +0.5 cm −1 . Magnetization data vs. applied magnetic field at low temperatures revealed a ground state with total spin S = 11/2 in agreement with ferromagnetic interactions between two Ni II (S = 1) and one Gd III (S = 7/2) ions. In two cases, the magnetocaloric effect of the Ni 2 Gd complexes was determined by heat capacity and isothermal magnetization measurements yielding magnetic entropy change (−∆S m ) in the range~12-14 Jkg −1 K −1 [17,26]. In some cases, complexes with Ln III = Dy, Tb showed zero-field or field-induced slow relaxation of the magnetization.
With these considerations in mind we have explored a general reaction scheme involving

Synthesis and Spectroscopic Characterization
The reaction of two equivalents of Ni(O 2 CMe) 2 ·4H 2 O with one equivalent of Ln(NO 3 ) 3 ·6H 2 O (Sm III (1), Eu III (2), Gd III (3), Tb III (4)) and four equivalents of H 4 L in EtOH afforded trinuclear compounds of the general formula [Ni 2 Ln(H 3 L) 4 (O 2 CMe) 2 ](NO 3 ) according to Equation (1). Precipitation of the microcrystalline and/or single crystal X-ray diffraction quality crystals of 1-4 was achieved by layering of the reaction solution with mixture of Et 2 O/n-hexane. (1) The identity of 1-4 was confirmed by single crystal X-ray crystallography, infrared spectroscopy and microanalytical techniques. Initial attempts to prepare 1-4 from MeOH solutions gave microcrystalline products in very low yield (only few crystals). Subsequently, we changed the reaction solvent to EtOH and we managed to increase the yield of 1-4 and determine the crystal structure of 4·4EtOH·4H 2 O. The identity of 3 was confirmed by unit cell determination [46].
The IR spectra of all complexes ( Figure S1) exhibit broad bands in the range 3437-3461 cm −1 attributed to the ν(OH) vibrations due to the presence of alkoxo groups of the ligand. The band at 1635 cm −1 in the free ligand is due to ν(C=N) vibration. The shifting of this band to lower frequency (~1600 cm −1 ) in the spectra of all complexes suggests coordination of the metal ions through the imino nitrogen. The ν(C-O) stretching frequency of the phenolic oxygen of the ligand is seen at 1395 cm −1 and shifts to lower frequency in the spectra of all complexes, in the range 1317-1320 cm −1 , indicating coordination to the metal ions [47]. The strong bands at~1555 and~1444 cm −1 are attributed to the ν as (CO 2 ) and ν s (CO 2 ) stretching vibrations of the bidentate chelate acetato ligands. The difference ∆ = ν as (CO 2 ) − ν s (CO 2 ) is 110-114 cm −1 and agrees with the low difference values found in rare earth acetates with chelating coordination mode [48][49][50]. The strong band at~1384 cm −1 in the spectra of all compounds is attributed to the presence of ν 3 (E ) [ν d (NO)] mode of the uncoordinated D 3h ionic nitrates [51].   Symmetry operation: ( ) 1.5-x, 0.5-y, z; (") x, 0.5-y, 1.5-z; ( ) 1.5-x, y, 1.5-z.
The lattice structure of 4 is built due to hydrogen bonding interactions ( Table 2). Each trinuclear cluster acts as a 4-connected node and is linked to four neighboring clusters through eight hydrogen bonds developed between the protonated pendant alkoxide groups of H 3 L − (Figure 2a) which expands to a 3D diamondlike network. Each trinuclear cluster is translated along a-axis resulting in a second 3D diamondlike network interacting weakly with the first one through Van der Waals forces, and thus the final architecture of the structure consists of two interpenetrating diamond lattices (Figure 2a,b). The topology of each independent lattice is described by the 6 6 Well point symbols or with Schläfli symbol (6,4) and the overall structure as 2-fold based on the Bratten-Robson classification scheme [53] or with topology dia belonging to Class Ia, with Translational Degree of interpenetration Z t = 2 and translation along [1,0,0] (10.0634 Å) according to Blatov et al. [54]. The characteristic interlocked adamantane units of two independent interpenetrating diamondlike lattices [53] observed in the structure of compound 4 are shown in Figure S2. Schiff bases have been proved to build very interesting supramolecular structures, with the involvement of solvent molecules or not [55]. Diamondlike interpenetrating lattices, and especially those created through supramolecular interactions, have attracted the interest of researchers as it is possible to control, both the density and the pore size of the materials [56]. The center/nodes in the diamondoid lattices are at 14.386 Å apart. In channels created along the a-axis of compound 4, the NO 3 − counteranions and the lattice solvents (water and ethanol molecules) are hosted (Figure 2b). The channels along the a-axis occupy the~38% of the total unit cell volume (with a volume of 1374.9 Å 3 out of a total of 3642.9 Å 3 ).

Magnetic Measurements
The χ M T product of 1 at room temperature is 2.27 cm 3 Kmol −1 , which is larger than the calculated value of 2.09 cm 3 Kmol −1 for two Ni II (S = 1, g = 2.0, χ M T = 2.0 cm 3 Kmol −1 ) and one Sm III non-interacting ions (χ M T = 0.09 cm 3 Kmol −1 ). The value of χ M T product decreases slightly upon lowering the temperature, reaches a value of 1.94 cm 3 Kmol −1 at 20 K and then drops to 1.30 cm 3 Kmol −1 at 2 K ( Figure 3). The susceptibility data were fitted considering (a) the intramolecular magnetic interaction between the two Ni II ions, J, and (b) the magnetic anisotropy of the Ni II ions, D, according to the spin Hamiltonian: The best fit gave J = −0.02 cm −1 , D = −7.67 cm −1 with g = 1.97 and TIP = 0.001 cm 3 mol −1 . Taking into account the negligible value of J, which is in agreement with the very large intramolecular Ni···Ni distance, the fit of the χ M T vs. T data was repeated considering only the magnetic anisotropy of the Ni II ions, and gave D = −8.07 cm −1 with g = 1.97 and TIP = 0.001 cm 3 mol −1 . These values for the zero field splitting parameter D are extremely large for this type of complexes, suggesting that antiferromagnetic intramolecular interactions are present and affect the decrease of χ M T product at low temperatures. The χ M T vs. T data could be nicely fitted considering only the magnetic interaction between the Ni II ions, leading to the value of J = −0.37 cm −1 with g = 1.97 and TIP = 0.001 cm 3 mol −1 (blue line in Figure 3).
The χ M T product of 2 at 300 K is 3.17 cm 3 Kmol −1 , significantly higher than expected for two non-interacting Ni II (S = 1, g = 2.0) ions and one Eu III ion (S = 0). This behavior can be explained if we assume that the first excited states for the Eu III ion are energetically close enough to the ground state so that can be thermally populated at 300 K. As the temperature decreases, the χ M T product of 2, decreases to~2 cm 3 Kmol −1 at 10 K (Figure 3). This is expected due to the progressive and final depopulation of the magnetic excited states of the Eu III ions. Below 10 K, the χ M T product of 2 decreases to 1.41 cm 3 Kmol −1 at 2 K, which is close to the value of the χ M T product of 1 at 2 K, suggesting that both compounds exhibit similar ground states and that the Eu III low-lying excited states in 2 are completely depopulated at 2 K.
Field dependent magnetization measurements were performed up to 5 T for 1 and 2 at 2 K; these are shown as insets in Figure 3. In both cases, the magnetization increases gradually in low magnetic fields and reaches 2.80 Nµ B for 1 and 2.98 Nµ B for 2 at 5 T, without reaching saturation.
The χ M T product of 3 at 300 K is 9.63 cm 3 Kmol −1 which is in very good agreement with the theoretic value (9.88 cm 3 Kmol −1 ) expected for two non-interacting Ni II (S = 1, g = 2.0) and one Gd III (S = 7/2, J = 7/2, g = 2.0) ions. Between 300 and 60 K, the χ M T product remains practically constant, and then increases slightly up to~30 K; below that temperature the χ M T product increases rapidly and reaches the value of 15.92 cm 3 Kmol −1 at 2 K (Figure 4). The overall behavior is consistent with the presence of dominant ferromagnetic Ni···Gd interactions within the trinuclear complex. Magnetization measurements at 2 K show a rapid increase upon increasing of the magnetic field reaching a value of 10.64 Nµ B at 5 T (Figure 4, inset) very close to the theoretical value for a S = 11/2 spin ground state corresponding to ferromagnetic coupling between two Ni II and one Gd III ions. Complex 3 is isomorphous to 4 as determined by unit cell measurements. For the latter, the molecular symmetry implies strictly linear metal arrangement and two equal Ni···Ln distances, which would require only one exchange parameter J NiGd . The magnetic susceptibility and isothermal M(H) data were fitted by using the spin Hamiltonian: where S Ni = 1 and S Gd = 7/2 and D the magnetic anisotropy of the Ni II ions. The best fit gave J NiGd = +0.42 cm −1 , D = +2.95 cm −1 with g Ni = g Gd = 1.98 (solid lines in Figure 4). These values agree with those found in other linear trinuclear Ni 2 Gd complexes [13,20]. The magnetocaloric effect of 3 was determined by isothermal magnetization measurements in the temperature range 2-12.5 K under applied magnetic fields up to 5 T ( Figure 5). The magnetic entropy change can be obtained by using Maxwell's relation where B is the applied magnetic induction in Tesla, B i and B f are the initial and final applied magnetic induction. A simple numerical approach can be used to obtain the experimental value of the entropy from the M vs. H curves using Equation (5) where M i and M i+1 are the magnetization values measured in a field H at temperatures T i and T i+1 , respectively. The change of the magnetic entropy was calculated |S(0,5T)| = 14.2 Jkg −1 K −1 at T = 2 K in agreement with the values found in the literature [17]. Moreover, the estimated value of the |S(0,5T)| is lower than the theoretically expected for three independent ions (two Ni II S = 1 and one Gd III S = 7/2), ∆S = 2Rln(3) + Rln(8) = 4.276R Jmol −1 K −1 ≡ 35 Jmol −1 K −1 . With a molecular weight of 1.35 kgmol −1 for 3, an entropy change of 26.33 Jkg −1 K −1 is expected. This difference can be attributed to the presence of non-zero exchange interactions. Nevertheless, our results are comparable with the values found in the literature in a similar Ni 2 Gd cluster [17].
The χ M T product of 4 at 300 K is 13.88 cm 3 Kmol −1 , which is close to the theoretical value of 13.81cm 3 Kmol −1 , for two Ni II (S = 1, g = 2.0) and one Tb III (S = 3, J = 6, g = 3/2) non-interacting ions (Figure 4). The χ M T product of 4 remains practically constant in the temperature range 300-80 K and then decreases slightly to the value of 13.54 cm 3 Kmol −1 at 25 K. This behavior indicates dominant intramolecular antiferromagnetic interactions between the metal ions and/or thermal depopulation of the Tb III excited states. Between 25 and 8 K, the χ M T product increases slightly to reach the value of 13.88 cm 3 Kmol −1 and then drops to 11.79 cm 3 Kmol −1 at 2 K. This behavior could be attributed to ferromagnetic interactions between the metal ions that could result in a high spin ground state.
The field dependence of the magnetization for 4 is shown in Figure 4 inset. The magnetization of 4 at 2 K reaches 7.92 Nµ B at 5 T. The curve does not level out as magnetization does not saturate suggesting that ground spin state is not fully populated because other excited states remain populated to some extent.

General and Spectroscopic Measurements
All manipulations were performed under aerobic conditions using materials as received (Aldrich Co).
All chemicals and solvents were of reagent grade. The ligand OH-C 6 H 4 -CH=NC(CH 2 OH) 3 , H 4 L was synthesized as described previously [57]. Elemental analysis for carbon, hydrogen, and nitrogen was performed on a Perkin Elmer 2400/II automatic analyzer. Infrared spectra were recorded as KBr pellets in the range 4000-400 cm −1 on a Bruker Equinox 55/S FT-IR spectrophotometer. Variable-temperature magnetic susceptibility measurements were carried out on polycrystalline samples of 1-4 by using a SQUID magnetometer (Quantum Design MPMS 5.5). Diamagnetic corrections were estimated from Pascal s constants. The program PHI was used for simulations of the magnetic susceptibility data of 3 [58].

Conclusions
We have demonstrated the preparation of the trinuclear complexes of general formula [Ni 2 Ln(H 3 L) 4 (O 2 CMe) 2 ](NO 3 ), Ln = Sm (1), Eu (2), Gd (3), Tb (4), which contain the monoanion of the tetradentate Schiff base ligand H 4 L: o-OH-C 6 H 4 -CH=N-C(CH 2 OH) 3 . The crystal structure of 4 revealed the presence of complex cations and nitrate anions. The complex cations contain two terminal Ni II ions and one central Tb III ion in linear arrangement. Both Ni II ions present distorted octahedral geometry and are bound to two H 3 L − ligands through their phenolato oxygen, the imino nitrogen atoms and one of the protonated alkoxo groups. The Tb III ion is coordinated to four phenolato oxygen atoms from the four ligands, and four carboxylato oxygen atoms from two acetates which are bound in the bidentate chelate mode. The use of Continuous Shape Measures approach (CShM) revealed that the coordination polyhedron around the terbium ion in 4 is square antiprism, SAPR-8. The pendant alkoxide arms of the ligands participate in an extensive network of hydrogen bonds. The final architecture of the lattice structure consists of two interpenetrating 3D diamond lattices which host the NO 3 − counteranions and the lattice solvents. The identity of 3 was confirmed by unit cell determination and it was found isomorphous to complex 4. The chemical identity and similarity of 1-4 were confirmed by infrared spectroscopy. The magnetic study of these complexes demonstrated the nature of the magnetic exchange between the metal ions in 1-4. The magnetic susceptibility measurements revealed weak antiferromagnetic coupling between the Ni II ions in 1. The contribution of the magnetic anisotropy of the Ni II ions (S = 1) during the fit of the susceptibility data of 1 led to unrealistic high values for the parameter D and was not taken into account. The significantly high value of the χ M T product of 2 at 300 K is consistent with the presence of first excited states which are sufficiently low in energy and are thermally populated at r.t. The experimental χ M T vs. T curve for 4 indicates the presence of intramolecular antiferromagnetic interactions between the metal ions; the decrease of the χ M T products at low temperatures are consistent with the thermal depopulation of the Tb III excited states. The magnetic study of 3 revealed dominant ferromagnetic interactions between the Ni II and Gd III ions with J NiGd = +0.42 cm −1 , D = +2.95 cm −1 (g Ni = g Gd = 1.98), resulting in S = 11/2 spin ground state. The change of the magnetic entropy of 3 was calculated |∆S(0,5T)| = 14.2 Jkg −1 K −1 at T = 2 K in agreement with the values found in the literature. Further work is in progress in our lab with other lanthanides in order to systematically delve into the magnetic properties of these complexes.
Supplementary Materials: The following are available online. Figure S1: The FT-IR spectra of complexes 1-4. Figure S2: The characteristic interlocked adamantane units of two independent interpenetrating diamondlike lattices observed in the structure of compound 4.
Author Contributions: A.N.G. performed the synthesis of 1-4; M.P. and Y.S. performed and interpreted the magnetic properties of 1-4; V.P. and C.P.R. performed the crystallographic analysis of 1-4. C.P.R. was responsible for coordinating the work and measurements and for writing-review-editing of the manuscript. All authors contributed to the writing-review and editing of the manuscript. All authors have read and agreed to the published version of this manuscript.
Funding: This research received no external funding.