Pentacoordinate and Hexacoordinate Mn(III) Complexes of Tetradentate Schiff-Base Ligands Containing Tetracyanidoplatinate(II) Bridges and Revealing Uniaxial Magnetic Anisotropy

Crystal structures and magnetic properties of polymeric and trinuclear heterobimetallic MnIII···PtII···MnIII coordination compounds, prepared from the Ba[Pt(CN)4] and [Mn(L4A/B)(Cl)] (1a/b) precursor complexes, are reported. The polymeric complex [{Mn(L4A)}2{μ4-Pt(CN)4}]n (2a), where H2L4A = N,N’-ethylene-bis(salicylideneiminate), comprises the {Mn(L4A)} moieties covalently connected through the [Pt(CN)4]2− bridges, thus forming a square-grid polymeric structure with the hexacoordinate MnIII atoms. The trinuclear complex [{Mn(L4B)}2{μ-Pt(CN)4}] (2b), where H2L4B = N,N’-benzene-bis(4-aminodiethylene-salicylideneiminate), consists of two [{Mn(L4B)} moieties, involving pentacoordinate MnIII atoms, bridged through the tetracyanidoplatinate (II) bridges to which they are coordinated in a trans fashion. Both complexes possess uniaxial type of magnetic anisotropy, with D (the axial parameter of zero-field splitting) = −3.7(1) in 2a and −2.2(1) cm−1 in 2b. Furthermore, the parameters of magnetic anisotropy 2a and 2b were also thoroughly studied by theoretical complete active space self-consistent field (CASSCF) methods, which revealed that the former is much more sensitive to the ligand field strength of the axial ligands.


Introduction
Single-molecule magnets (SMMs) are compounds composed of the individual molecules capable of preserving their magnetic moment even after removing external magnetic field. This is possible due to an existence of energy barrier ∆ separating the states (formed by the crystal-field splitting of the ground spin state S, S > 1/2) with minimal and maximal M S under the condition of axial magnetic anisotropy. The potential applications of SMMs and their polymeric analogues, so called single-chain magnets (SCMs), include information processing, data storage, quantum computing, spintronics, or biomedical applications [1,2].
Bearing in mind the above mentioned conditions for SMM occurrence it is clear why so much of attention has been devoted to the research of the Mn III complexes involving tetradentate salen-type Schiff base ligands (e.g., L4A 2− = salen 2− = N,N'-ethylene-bis(salicylideneiminate) dianion; variously substituted salen ligands will be further abbreviated as H 2 L4). In compounds involving the [Mn(L4)] + moieties the central Mn III atom is in the hexacoordinate environment and, thus, it is the object of the Jahn-Teller effect (S = 2, 5 E crystal field ground term in weak ligand fields). This implies distortion of the coordination polyhedron (its prolongation or compression), which gives rise to significant magnetic anisotropy. These might be of axial (prolongation) or easy-plane (compression) character [3]. Thus, it is clear that compounds of this type are attractive building blocks coordination compounds exhibiting magnetic bistability [4,5]. A typical approach for synthesis of polymeric or polynuclear coordination compounds involving the [Mn(L4)] + moieties is associated with the reaction of the halide precursor complexes [Mn(L4)X], where X = the halido ligand, with various cyanidometallates. Such reactions lead to preparations of compounds related to Prussian blue [6][7][8][9][10][11]. Previously, we reported on crystal structures and magnetic properties of polymeric and polynuclear compounds involving the [Mn(L4)] + moieties which were bridged by [Fe(CN)5(NO)] 2− , [Pt(SCN)4] 2− and [Pt(SCN)6] 2 complex anions [12][13][14]. This work represents a continuation of our ongoing study of platinum cyanido/thiocyanido bridged heterometallic compounds and reports on two new Mn III (4-aminodiethylenesalicylideneiminate) (Scheme 1). The prepared complexes were characterized by elemental analysis, molar conductivity, infrared spectroscopy, single crystal X-ray analysis, and magnetic measurements. To the best of our knowledge this is the first report on the crystal structures and magnetic properties of Mn III ⋅⋅⋅Pt II ⋅⋅⋅Mn III compounds involving [Pt(CN)4] 2− as bridging units. Furthermore, the CASSCF/NEVPT2 calculations were performed on 2a and 2b and also on other model hexa/penta-coordinate compounds with the aim to more deeply understand variations in zero-field splitting (ZFS) parameters in Mn III Schiff-base complexes.

Crystal Structures
The crystal structures of 2a and 2b were determined by single-crystal X-ray analysis ( Figure 1, Table 1). Both compounds comprise Pt II atoms coordinated by four carbon atoms from the cyanide groups, thus forming a {PtC4} chromophore (Figure 1). The Pt−C bond lengths are from the narrow range of 1.99-2.00 Å (Figure 1), and with the C−Pt−C angles close to the ideal straight angle as expected for the square planar chromophore geometry (179.6(5)° in 2a, and 180° in 2b). In both compounds, the Schiff base ligands (Scheme 1) coordinate Mn III atoms by two imino nitrogen (Nim) and two phenolate oxygen atoms with the metal−donor atom distances: Mn-Nim = 1.96-1.98 Å and Mn-O = 1.86-1.89 Å.

Crystal Structures
The crystal structures of 2a and 2b were determined by single-crystal X-ray analysis ( Figure 1, Table 1). Both compounds comprise Pt II atoms coordinated by four carbon atoms from the cyanide groups, thus forming a {PtC 4 } chromophore ( Figure 1). The Pt−C bond lengths are from the narrow range of 1.99-2.00 Å (Figure 1), and with the C−Pt−C angles close to the ideal straight angle as expected for the square planar chromophore geometry (179.6(5) • in 2a, and 180 • in 2b). In both compounds, the Schiff base ligands (Scheme 1) coordinate Mn III atoms by two imino nitrogen (N im ) and two phenolate oxygen atoms with the metal−donor atom distances: Mn-N im = 1.96-1.98 Å and Mn-O = 1.86-1.89 Å.    The reason for such difference is that the pentacoordinate Mn III center in 2b is not affected by Jahn-Teller effect. Furthermore, it must be stressed that the Mn III atom is involved into metal···π non-covalent interactions with carbon atoms from aromatic part of the L4B 2− ligands from the adjacent trinuclear complex molecules ( Figure 2). The Mn···C distances in such non-covalent contact are shorter (3.423(4) and 3.449(4) Å) than the sum of their van der Waals radii (3.7 Å) [20]. Other non-covalent interactions presented in 2b are of π-π stacking nature. They do extend the crystal structure of 2b to 2D supramolecular layers ( Figure 2). Two different stacking interactions can be distinguished: (a) intrachain, π 1 −π 1 , with the shortest C···C distance at 3.338(5) Å, and (b) interchain, π 2 −π 2 , with the shortest C···C distance of 3.403(5) Å ( Figure 2). The reason for such difference is that the pentacoordinate Mn III center in 2b is not affected by Jahn-Teller effect. Furthermore, it must be stressed that the Mn III atom is involved into metal···π non-covalent interactions with carbon atoms from aromatic part of the L4B 2− ligands from the adjacent trinuclear complex molecules ( Figure 2). The Mn···C distances in such non-covalent contact are shorter (3.423(4) and 3.449(4) Å) than the sum of their van der Waals radii (3.7 Å) [20]. Other non-covalent interactions presented in 2b are of π-π stacking nature. They do extend the crystal structure of 2b to 2D supramolecular layers ( Figure 2). Two different stacking interactions can be distinguished: (a) intrachain, π1−π1, with the shortest C···C distance at 3.338(5) Å, and (b) interchain, π2−π2, with the shortest C···C distance of 3.403(5) Å ( Figure 2).

Magnetic Properties
The magnetic data of complexes 2a and 2b are depicted in Figure 3 and conform with practically isolated paramagnetic Mn III cations. The effective magnetic moment is linearly decreasing from the room temperature value of 4.91 µB to the value of 4.78 µB at 30 K for 2a, and from 5.35 µB to the value of 4.89 µB at 30 K for 2b. The decrease of the effective magnetic moment in this temperature region is much more apparent for 2b and this can be ascribed to a presence of small amount of unknown magnetic impurity undetectable by other physical methods. Then, the effective magnetic moment drops to the value of 3.53 µB for 2a and to the value of 4.19 µB for 2b at the lowest available temperature (T = 1.9 K). This is mainly due to the zero-field splitting (ZFS) of Mn III atoms and partly also due to very weak intra/inter-molecular non-covalent interactions. Therefore, we postulated the following spin Hamiltonian: where the first term describes the magnetic anisotropy with single-ion axial ZFS parameter D, the second term is the Zeeman term, and the last term, represented with the zj variable, is the common molecular-field correction parameter, which is due to intermolecular interactions. The <Sa> is a thermal average of the molecular spin projection in the a direction of magnetic field defined as Ba = B·(sinθcosϕ, sinθsinϕ, cosθ) with the help of polar coordinates. Then, the molar magnetization in the a-direction of magnetic field can be numerically calculated as:

Magnetic Properties
The magnetic data of complexes 2a and 2b are depicted in Figure 3 and conform with practically isolated paramagnetic Mn III cations. The effective magnetic moment is linearly decreasing from the room temperature value of 4.91 µ B to the value of 4.78 µ B at 30 K for 2a, and from 5.35 µ B to the value of 4.89 µ B at 30 K for 2b. The decrease of the effective magnetic moment in this temperature region is much more apparent for 2b and this can be ascribed to a presence of small amount of unknown magnetic impurity undetectable by other physical methods. Then, the effective magnetic moment drops to the value of 3.53 µ B for 2a and to the value of 4.19 µ B for 2b at the lowest available temperature (T = 1.9 K). This is mainly due to the zero-field splitting (ZFS) of Mn III atoms and partly also due to very weak intra/inter-molecular non-covalent interactions. Therefore, we postulated the following spin Hamiltonian:Ĥ where the first term describes the magnetic anisotropy with single-ion axial ZFS parameter D, the second term is the Zeeman term, and the last term, represented with the zj variable, is the common molecular-field correction parameter, which is due to intermolecular interactions. The <S a > is a thermal average of the molecular spin projection in the a direction of magnetic field defined as B a = B·(sinθcosϕ, sinθsinϕ, cosθ) with the help of polar coordinates. Then, the molar magnetization in the a-direction of magnetic field can be numerically calculated as: where Z a is the matrix element of the Zeeman term for the a-direction of the magnetic field and C are the eigenvectors resulting from the diagonalization of the complete spin Hamiltonian matrix. The presence of zj means that iterative procedure must be used [21]. Then, the averaged molar magnetization of the powder sample was calculated as integral (orientational) average: In order to obtain reliable parameters, both temperature and field dependent magnetic data were fitted simultaneously, which resulted in these parameters: where the temperature-independent magnetism correction χ TIM was applied to describe the contribution of traces of magnetic impurities [22]. The sign of D-parameters suggests that in 2a and 2b there is axial magnetic anisotropy and that more negative D value found for 2a revealed larger anisotropy in the hexacoordinate Mn III compound as compared to that of pentacoordinate one. where Za is the matrix element of the Zeeman term for the a-direction of the magnetic field and C are the eigenvectors resulting from the diagonalization of the complete spin Hamiltonian matrix. The presence of zj means that iterative procedure must be used [21]. Then, the averaged molar magnetization of the powder sample was calculated as integral (orientational) average: In order to obtain reliable parameters, both temperature and field dependent magnetic data were fitted simultaneously, which resulted in these parameters: D = −3.7(1) cm −1 , g = 1.981(6), zj = −0.244(9) cm −1 , χTIM = 6(1)·10 −9 m 3 ·mol −1 for 2a, and D = −2.2(1) cm −1 , g = 1.975 (8), zj = −0.02(1) cm −1 , χTIM = 28(2)·10 −9 m 3 ·mol −1 for 2b, where the temperature-independent magnetism correction χTIM was applied to describe the contribution of traces of magnetic impurities [22]. The sign of D-parameters suggests that in 2a and 2b there is axial magnetic anisotropy and that more negative D value found for 2a revealed larger anisotropy in the hexacoordinate Mn III compound as compared to that of pentacoordinate one.

Theorethical Calculations
Furthermore, we supported our experimental magnetic study also by ab initio CASSCF/NEVPT2 calculations using ORCA computational package where also the relativistic effects were included as described in the Section 4.

Discussion
In order to improve our understanding of magnetic anisotropy in these complexes, we

Theorethical Calculations
Furthermore, we supported our experimental magnetic study also by ab initio CASSCF/NEVPT2 calculations using ORCA computational package where also the relativistic effects were included as described in the Section 4.2.3. The resulting values of ZFS parameters were as follows: D = −3.57 cm −1 and E/D = 0.030 for 2a, and D = −2.75 cm −1 and E/D = 0.021 for 2b, while the components of g-tensors for both compounds were in narrow range from 1.977 to 1.994. Thus, the calculations supported our conclusions following from the experimental results showing that there is lower magnetic anisotropy in the pentacoordinate complex 2b than in the hexacoordinate one (2a).

Discussion
In order to improve our understanding of magnetic anisotropy in these complexes, we performed magneto-structural correlations for pentacoordinate [Mn(L4A)(MeCN)] + (MnL 5 ) and hexacoordinate [Mn(L4A)(MeCN) 2 ] + (MnL 6 ) model compounds, where we studied the impact of axial ligand(s) field strength on ZFS parameters by varying Mn−N MeCN distance(s) from 1.9 to 2.8 Å (MeCN = acetonitrile). The geometries of the complexes were optimized with the PBE functional also incorporating the COSMO model (COSMO = COnductor-like Screening MOdel). The only geometrical constrain applied was Mn-N MeCN distance and in case of hexacoordinate MnL6 model compound, the Mn-N MeCN distances were both equal. Subsequently, the single-point energy calculations were done for each optimized geometry using B3LYP/ZORA/def2-TZVP(-f) followed by extracting the information about the splitting of d-orbitals, which is visualized in Figure 4 (top). This figure resembles the well-known crystal-field theory schemes outlined also in Figure 4 (bottom). In the case of hexacoordinate MnL 6 model compound with d(Mn-N MeCN ) = 1.9 Å, the d-orbitals are evidently split according to scheme outlined for a compressed square-bipyramidal coordination geometry, whereas for increasing Mn-N MeCN distance, the pattern for elongated square-bipyramidal arrangement is observed. Upon further increase in the Mn-N MeCN distance, the d-orbitals are split similarly to square-planar geometry. An analogous situation can be found for the pentacoordinate MnL 5 model compound, where an increase in the Mn-N MeCN distance follows d-orbitals splitting from square-pyramidal to square-planar geometry. Of course, DFT-calculated (DFT = Density Functional Theory) energies of d-orbitals are not exactly following simplified schemes derived from the crystal-field theory due to non-equivalent ligand field strengths of Schiff base nitrogen and oxygen donor atoms. Then, ZFS parameters for each geometry were calculated by the same procedure with the CASSCF/NEVPT2 method. The results are depicted in Figure 5. constrain applied was Mn-NMeCN distance and in case of hexacoordinate MnL6 model compound, the Mn-NMeCN distances were both equal. Subsequently, the single-point energy calculations were done for each optimized geometry using B3LYP/ZORA/def2-TZVP(-f) followed by extracting the information about the splitting of d-orbitals, which is visualized in Figure 4 (top). This figure resembles the well-known crystal-field theory schemes outlined also in Figure 4 (bottom). In the case of hexacoordinate MnL6 model compound with d(Mn-NMeCN) = 1.9 Å, the d-orbitals are evidently split according to scheme outlined for a compressed square-bipyramidal coordination geometry, whereas for increasing Mn-NMeCN distance, the pattern for elongated square-bipyramidal arrangement is observed. Upon further increase in the Mn-NMeCN distance, the d-orbitals are split similarly to square-planar geometry. An analogous situation can be found for the pentacoordinate MnL5 model compound, where an increase in the Mn-NMeCN distance follows d-orbitals splitting from square -pyramidal to square-planar geometry. Of course, DFT-calculated (DFT = Density Functional Theory) energies of d-orbitals are not exactly following simplified schemes derived from the crystal-field theory due to non-equivalent ligand field strengths of Schiff base nitrogen and oxygen donor atoms. Then, ZFS parameters for each geometry were calculated by the same procedure with the CASSCF/NEVPT2 method. The results are depicted in Figure 5.   It is evident that there is a negligible effect of the axial ligand strength on magnetic anisotropy in the case of the pentacoordinate MnL5 complex, the D varied only between −2.9 and −2.5 cm −1 . This is in the stark contrast with the hexacoordinate MnL6 complex, where shortening of the Mn−N distances (axial elongation → nearly ideal octahedron), in other words, increasing the ligand field strength of the axial ligands, led to immense increase of the negative value of the D-parameter (D = −2.6 → −6.5 cm −1 ) with the maximal absolute value at the chromophore geometry close to the ideal octahedral arrangement ( Figure 5). Further increase in the axial ligand field strength (i.e., the axial compression) would lead to crossover to compressed square-bipyramid and consequently to a positive value of the D-parameter, which is evidenced here by increase of E/D ratio approaching the value of 1/3 ( Figure 5). Such a relationship between the sign of D and the axial elongation/compression was already proposed by Maurice et al. for a simple [Mn(NCH)6] 3+ model complex [23]. Furthermore, we plotted the ligand field terms, quintets and triplets, arising from CASSCF/NEVPT2 calculations for both MnL5 and MnL6 model compounds ( Figure 6). It is evident that the increase in the axial ligand field strength induced by decreasing of Mn-NMeCN distances led to significant lowering of the excited quintets and especially triplet states for MnL6, whereas energies of excited triplets and quintets states are not varied to such extent for MnL5. This different behaviour had dominant impact on the value of the D-parameter, because the contribution of quintet states to D is more or less the same for all Mn-NMeCN distances and moreover for both hexa-and pentacoordinate model complexes ( Figure 6). Furthermore, the comparison of D-values for MnL6 and MnL5 model compounds at d(Mn-NMeCN) = 2.2-2.3 Å (Figure 5), that means for axial Mn-N distances found in 2a and 2b, clearly showed that more negative D is expected for hexacoordinate compound as also confirmed from the analysis of the experimental magnetic data of the reported compounds. It is evident that there is a negligible effect of the axial ligand strength on magnetic anisotropy in the case of the pentacoordinate MnL 5 complex, the D varied only between −2.9 and −2.5 cm −1 . This is in the stark contrast with the hexacoordinate MnL 6 complex, where shortening of the Mn−N distances (axial elongation → nearly ideal octahedron), in other words, increasing the ligand field strength of the axial ligands, led to immense increase of the negative value of the D-parameter (D = −2.6 → −6.5 cm −1 ) with the maximal absolute value at the chromophore geometry close to the ideal octahedral arrangement ( Figure 5). Further increase in the axial ligand field strength (i.e., the axial compression) would lead to crossover to compressed square-bipyramid and consequently to a positive value of the D-parameter, which is evidenced here by increase of E/D ratio approaching the value of 1/3 ( Figure 5). Such a relationship between the sign of D and the axial elongation/compression was already proposed by Maurice et al. for a simple [Mn(NCH) 6 ] 3+ model complex [23]. Furthermore, we plotted the ligand field terms, quintets and triplets, arising from CASSCF/NEVPT2 calculations for both MnL 5 and MnL 6 model compounds ( Figure 6). It is evident that the increase in the axial ligand field strength induced by decreasing of Mn-N MeCN distances led to significant lowering of the excited quintets and especially triplet states for MnL 6 , whereas energies of excited triplets and quintets states are not varied to such extent for MnL 5 . This different behaviour had dominant impact on the value of the D-parameter, because the contribution of quintet states to D is more or less the same for all Mn-N MeCN distances and moreover for both hexa-and pentacoordinate model complexes ( Figure 6).

Materials
All of the starting chemicals were of analytical reagent grade and were used as received. All of the chemicals were purchased from commercial sources (Sigma Aldrich, St. Louis, MO, USA).

Materials
All of the starting chemicals were of analytical reagent grade and were used as received. All of the chemicals were purchased from commercial sources (Sigma Aldrich, St. Louis, MO, USA).

Preparation of Ba[Pt(CN) 4 ]
A solution of PtCl 2 (1 mmol, in 10 mL of water) was added to a cold, saturated solution of KCN (4 mmol, in 2 mL of water  0.2 mmol). The reaction mixture was stirred for 60 min at room temperature, and then was kept undisturbed in the dark. After twelve days, single-crystals suitable for X-ray analysis were formed. The resulting crystals were filtered off from the mother liquor, washed with water, diethyl ether, dried in a drying kiln (at 50 • C), and stored in a desiccator.

General Methods
Elemental analysis (CHN) was performed on a FLASH 2000 CHN Analyser (ThermoFisher Scientific, Waltham, MA, USA). Infrared spectra of the complexes were recorded on a NEXUS 670 FT-IR spectrometer (ThermoNicolet, Waltham, MA, USA) using the ATR technique on a diamond plate in the range 600-4000 cm −1 . The reported FT-IR intensities were defined as w = weak, m = medium, s = strong, and vs = very strong. The magnetic data were measured on powdered samples pressed into pellets using a MPMS XL-7 Quantum Design SQUID magnetometer (Quantum Design Inc., San Diego, CA, USA). The experimental data were corrected for the diamagnetism of the constituent atoms by using Pascal's constants.

Single Crystal X-ray Analysis Details
Single crystal X-ray diffraction data of 2a and 2b were collected on an Oxford diffractometer Xcalibur2 (Oxford Diffraction Ltd., Oxford, UK) with the Sapphire CCD detector and fine-focused sealed tube (Mo Kα radiation, λ = 0.71073 Å) source and equipped with an Oxford Cryosystem nitrogen gas flow apparatus. All structures were solved by direct methods using SHELXS-2014 [25] incorporated into the WinGX program package [26]. For each structure, its space group was checked by the ADSYMM procedure with PLATON software [27,28]. All structures were refined using full-matrix least-square procedures on F 2 with SHELXL-2014 [25] with anisotropic displacement parameters for all non-hydrogen atoms. The hydrogen atoms were placed into the calculated positions and they were included into the riding model approximation, with U iso = 1.2 or 1.5 U eq . All of the crystal structures were visualized using Mercury software [16].

Theoretical Methods
Ab initio CASSCF/NEVPT2 calculations [29,30] using ORCA 3.0 [31] with active space defined by four electrons in five d-atomic orbitals, CAS(4,5) and taking into account five quintets and 45 triplets. The relativistic effects were also included in the calculations with zero order regular approximation (ZORA) [32,33] together with the scalar relativistic contracted version of def2-TZVP(-f) basis functions [34]. The calculations utilized the RI approximation with the decontracted auxiliary def2-TZV/C Coulomb fitting basis sets and the chain-of-spheres (RIJCOSX) approximation to exact exchange [35]. Increased integration grids (Grid5 in ORCA convention) and tight SCF convergence criteria were used. The ZFS parameters, based on dominant spin-orbit coupling contributions from excited states, were calculated through quasi-degenerate perturbation theory (QDPT) [36], in which approximations to the Breit-Pauli form of the spin-orbit coupling operator (Spin-Orbit Mean-Field (SOMF) approximation) [37] and the effective Hamiltonian theory [38] were utilized.

Conclusions
In conclusion, this work presents synthesis and characterization of two new Mn III ···Pt II ···Mn III complexes with the general formula [{Mn(L4A)} 2 {µ 4 -Pt(CN) 4 }] n (2a) and [{Mn(L4B)} 2 {µ-Pt(CN) 4 }] (2b). Both compounds were investigated by single crystal X-ray diffraction and their magnetic properties were studied by temperature and field dependent measurements of magnetic moment. These studies revealed that compound 2a has a two-dimensional polymeric structure with the hexacoordinate Mn III atoms, while 2b possesses a trimeric structure with both Mn III atoms being pentacoordinate. The analysis of magnetic data showed that the magnetic anisotropy is uniaxial in both compounds, with D = −3.7(1) for 2a and −2.2(1) cm −1 for 2b. These findings were also supported by CASSCF/NEVPT2 calculations (D = −3.57 cm −1 and E/D = 0.030 for 2a, and D = −2.75 cm −1 and E/D = 0.021 for 2b). Further theoretical modelling of other pentacoordinate [Mn(L4A)(MeCN)] + and hexacoordinate [Mn(L4A)(MeCN) 2 ] + model compounds elucidated why the axial ZFS parameter D is larger in the case of the latter. The next important consequence is that magnetic anisotropy cannot be practically tuned in pentacoordinate square-pyramidal complexes by modifying ligands in axial position. On the contrary, hexacoordinate square-bipyramidal complexes are sensitive to the axial ligand field, thus prone to tuning of magnetic anisotropy, and for that reason they indicate themselves as promising for the preparation of the single-molecule magnets (SMMs).