Structural Diversity, XAS and Magnetism of Copper(II)-Nickel(II) Heterometallic Complexes Based on the [Ni(NCS)6]4− Unit

The new heterometallic compounds, [{Cu(pn)2}2Ni(NCS)6]n·2nH2O (1), [{CuII(trien)}2Ni(NCS)6CuI(NCS)]n (2) and [Cu(tren)(NCS)]4[Ni(NCS)6] (3) (pn = 1,2-diaminopropane, trien = triethylenetetramine and tren = tris(2-aminoethylo)amine), were obtained and characterized by X-ray analysis, IR spectra, XAS and magnetic measurements. Compounds 1, 2 and 3 show the structural diversity of 2D, 1D and 0D compounds, respectively. Depending on the polyamine used, different coordination polyhedron for Cu(II) was found, i.e., distorted octahedral (1), square pyramidal (2) and trigonal bipyramidal (3), whereas coordination polyhedron for nickel(II) was always octahedral. It provides an approach for tailoring magnetic properties by proper selection of auxiliary ligands determining the topology. In 1, thiocyanate ligands form bridges between the copper and nickel ions, creating 2D layers of sql topology with weak ferromagnetic interactions. Compound 2 is a mixed-valence copper coordination polymer and shows the rare ladder topology of 1D chains decorated with [CuII(tren)]2+ antennas as the side chains attached to nickel(II). The ladder rails are formed by alternately arranged Ni(II) and Cu(I) ions connected by N2 thiocyanate anions and rungs made by N3 thiocyanate. For the Cu(I) ions, the tetrahedral thiocyanate environment mixed N/S donor atoms was found, confirming significant coordination spheres rearrangement occurring at the copper precursor together with the reduction in some Cu(II) to Cu(I). Such topology enables significant simplification of the magnetic properties modeling by assuming magnetic coupling inside {NiIICuII2} trinuclear units separated by diamagnetic [Cu(NCS)(SCN)3]3− linkers. Compound 3 shows three discrete mononuclear units connected by N-H…N and N-H…S hydrogen bonds. Analysis of XAS proves that the average ligand character and the covalency of the unoccupied metal d-based orbitals for copper(II) and nickel(II) increase in the following order: 1 → 2 → 3. In 1 and 2, a weak ferromagnetic coupling between copper(II) and nickel(II) was found, but in 2, additional and stronger antiferromagnetic interaction between copper(II) ions prevailed. Compound 3, as an ionic pair, shows, as expected, a spin-only magnetic moment.


Materials and Methods
Materials: All reagents used in the synthesis were of analytical grade and used without further purification. All polyamine ligands were purchased from Sigma Aldrich (Darmstadt, Germany) and other reagents from Avantor Performance Materials Poland SA (formerly POCH, Gliwice, Poland) and Chempur (PiekaryŚląskie, Poland).

Synthesis of [{Cu(pn) 2 } 2 Ni(NCS) 6 ] n ·2nH 2 O (1)
An amount of 241 mg of Cu(NO 3 ) 2 ·3H 2 O (1 mmol) was dissolved in 10 mL of water. Then, 0.26 mL of pn (3 mmol) was slowly added with constant stirring. The violet solution was obtained. In other beakers, 295 mg of Ni(NO 3 ) 2 ·6H 2 O (1 mmol) and 324 mg of NaNCS (4 mmol) were dissolved in 2 and 1 mL of water, respectively. They were mixed and added to the copper-amine solution. The mixture was stirred at room temperature for 15 min. There was no precipitate, and the clear solution was left to evaporate. The violet crystals, suitable for X-ray analysis, were obtained after a few days. The analysis found the following: C, 25

Synthesis of [{Cu II (trien)} 2 Ni(NCS) 6 Cu I (NCS)] n (2)
A total of 341 mg of CuCl 2 ·2H 2 O (2 mmol) was dissolved in 10 mL of water. A solution of 0.30 mL trien (2 mmol) in 20 mL of water was prepared separately. Both solutions were mixed, and a violet solution was formed. An amount of 238 mg of NiCl 2 ·6H 2 O (1 mmol) in 10 mL of water H 2 O and 583 mg of KNCS (6 mmol) in 6 mL of water were mixed. Then, the nickel(II) solution was added to the copper(II) solution. A violet-navy solution was formed. The navy-blue crystals, suitable for X-ray analysis, were obtained after about a month. The analysis found the following: C, 24

Synthesis of [Cu(tren)(NCS)] 4 [Ni(NCS) 6 ] (3)
To 341 mg (2 mmol) of CuCl 2 ·2H 2 O in 10 mL of distilled water, 20 mL of tren (0.30 mL, 2 mmol) was added. The solution turned from light blue to blue. An amount of 238 mg (1 mmol) of NiCl 2 ·6H 2 O was dissolved in 10 mL of water, and 583 mg (6 mmol) of NaNCS in 6 mL of water was added. They were mixed and added to the copper-amine solution. The mixture was stirred at room temperature for 30 min, and then the solution was filtered. The green-blue crystals, suitable for X-ray analysis after one month, were obtained. The analysis found the following: C, 27

Methods
Elemental analysis was performed on Vario Macro CHN Analyzer (Elementar Analysensysteme GmbH, Langenselbold, Germany). Infrared spectra were recorded on a Vertex 70v Spectrometer from Bruker Optik: GmbH (Ettlingen, Germany) in the range 4000-400 cm −1 . Magnetic measurements in the temperature range of 1.8-300 K were performed using SQUID MPMS-3 and MPMS-XL-5 magnetometers at the magnetic field of 0.1 and 0.5 T for 1 and 2, respectively. The data were corrected for the sample holder and the underlying diamagnetism using Pascal's constants [67] (−480 and −485 × 10 −6 cm 3 mol −1 , for compounds 1 and 2, respectively). The temperature-independent paramagnetism and all uncertainty in diamagnetic correction were treated as variable parameters within the fitting procedures (see below). The effective magnetic moment was calculated from the equation: µ eff = 2.828 (χ M corr · T) 1/2 . Magnetization versus magnetic field measurements were carried out at 1.8 K for 1 and 2 K for 2 in the magnetic field range 0-7 Tesla. Magnetic susceptibility for 3 was measured at room temperature by the Faraday method on a homemade balance at field strength up to 1.0 T. The magnetic field was calibrated with Hg[Co(NCS) 4 ] [68]. The X-ray absorption spectra were recorded at the National Synchrotron Radiation Centre SO-LARIS at the bending magnet PEEM/XAS beamline for N (350-450 eV) and O (500-550 eV) K-edge as well as Ni (850-900 eV) and Cu (910-1000 eV) L-edge. Samples of 1, 2 and 3 were finely ground and attached to double-sided adhesive conductive graphite tape. For all measurements, the step size of 0.2 eV was used for the edge regions and 0.4 eV for the remaining regions. The data sets were collected at room temperature under ultra-high vacuum (UHV) in the total electron yield mode (TEY). The data were processed using the PyMca 5.4.0 program package and were evaluated using the OriginPro 5.0 (Northampton, MA, USA) program as the data analysis and graphing software.

Single Crystal X-ray Diffraction Measurement
The diffraction experiments for the single crystal of 1, 2 and 3 were performed at room temperature using an Oxford Sapphire CCD diffractometer, MoKα radiation λ = 0.71073 Å. During the data processing, the numerical absorption correction was applied [69]. The structure was solved by the direct methods and refined with a full-matrix least-squares procedure on F 2 (SHELX-97 [70]). For all heavy atoms, a refinement procedure with anisotropic displacement parameters was applied. Positions of hydrogen atoms attached to carbon atoms were determined from geometrical conditions and refined with thermal displacement parameters fixed to a value of 20% or 50% higher than those of the corresponding carbon atoms. Hydrogen atoms from NH and NH 2 groups, as well as water molecules, were located in different electron density syntheses. In (1), in the refinement process, hydrogen atoms from the O7 water molecule were refined with constraints (DFIX and DANG) to ensure their reasonable geometry. In (2), ISOR restraints were applied to occupy the S7 Materials 2023, 16, 731 5 of 24 atom partially. All figures were prepared in DIAMOND [71] and ORTEP-3 [72]. The results of the data collection and refinement are summarized in Table 1. CCDC 2164900, 2164901 and 2164903 contain the supplementary crystallographic data for 1-3, respectively. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk.data_request/cif.

Infrared Spectra
The IR spectra are a good and convenient criterion to determine the coordination modes of thiocyanate ligands and to prove the presence of polyamine ligands and the lattice water [73][74][75]. In the infrared spectra of 1, 2 and 3, very intense bands at ca. 2100 cm −1 were detected, coming from CN stretching vibration of bridging thiocyanate. For 1 and 3, the band is very broad and does not show splitting, whereas for 2, two peaks at 2116 cm −1 and at 2071 cm −1 were distinguished, assignable to the NCS − bridge and to the terminal N-bonded thiocyanate, respectively. The NCS − bending vibration appears in all studied spectra as bands in the range 473-478 cm −1 . The bands occurring in the range 835-869 cm −1 are due to the vibration of the CS group, also coming from the thiocyanate ligand. Some of the expected low-intensity new CS bands can be masked by stronger bands from the rocking NH 2 vibrations. Several bands occurring in the range 3451-3129 and 2972-2872 cm −1 are characteristic of the polyamine group, NH and CH vibrations, respectively. The peak occurring in the IR spectrum of 1 at 3578 cm −1 is due to the stretching vibration of the OH group from water molecules.  [27,31,[76][77][78]. The axial bonds are distinctly longer than equatorial bonds and are equal to 2.9883(11) and 3.0224(10) Å for Cu-S5 and Cu-S6, respectively. Therefore Cu-N and Cu-S distances are characteristic of the geometry of the elongated octahedron [31,79]. Nickel(II) ion was found in a slightly distorted octahedral environment, with six Ni-N bonds being as follows: 2.060(3), 2.073(3) and 2.118(4) Å. These values correspond to those found in the literature [44,54]. The longest distance was found for the non-bridging thiocyanate ions indicating that a bridge formation resulted in a shortening of Ni-N bonds. The selected valence angle description is given in the Supplementary Materials.
The peak occurring in the IR spectrum of 1 at 3578 cm -1 is due to the stretching vibration of the OH group from water molecules.

Structure of [{Cu(pn)2}2Ni(NCS)6]n·2nH2O (1)
The reported complex crystallizes in a monoclinic C2/c space group with the nickel ions positioned at a twofold axis. In the asymmetric unit, there are copper and nickel ions, three linear thiocyanate ligands (N-C-S angles: 178.9(4)-179.6(4)°), two 1,2-diaminepropane molecules and one water molecule ( Figure 1). The selected bond lengths and angles are presented in Tables 2 and S1, Supplementary Materials. Carbon atoms in pn molecules reveal positional disorder with occupancy 0.55 (C12 and C13) and 0.45 (C14 and C15). Copper(II) ions adopt 4+2 coordination with four short Cu-N and two much longer Cu-S bonds. The Cu-N bonds are found in the range 2.003(3)-2.028(3) Å and are similar to the values found in other complexes [27,31,[76][77][78]. The axial bonds are distinctly longer than equatorial bonds and are equal to 2.9883(11) and 3.0224(10) Å for Cu-S5 and Cu-S6, respectively. Therefore Cu-N and Cu-S distances are characteristic of the geometry of the elongated octahedron [31,79]. Nickel(II) ion was found in a slightly distorted octahedral environment, with six Ni-N bonds being as follows: 2.060(3), 2.073(3) and 2.118(4) Å. These values correspond to those found in the literature [44,54]. The longest distance was found for the non-bridging thiocyanate ions indicating that a bridge formation resulted in a shortening of Ni-N bonds. The selected valence angle description is given in the Supplementary Materials.   i −x, y,−z + 1/2 ii −0.5 + x,−0.5+y, z.
In crystal packing, there are ab layers of sql topology [80] composed of covalently connected copper(II) and nickel(II) blocks, with four thiocyanate ions involved in bridge formation between paramagnetic centers (Figures 2 and S1, Supplementary Materials). In such a layer, every copper(II) ion is connected with two nickel(II) ions, and each nickel(II) cation serves as a node of the network interacting with four copper(II) cations. The ab layers form . . . ABAB . . . stack motifs with adjacent layers shifted by ca. [0, 0.45y, −0.5z], and hence, the nodes of A layer are not positioned above the cavity center of the B layer ( Figure S2, Supplementary Materials). It is noted that the layer topology is identical to that in the 2D compound of [{Cu(pn) 2 } 2 Mn(NCS) 6 ] n ·2nH 2 O [31].
formation between paramagnetic centers (Figures 2 and S1, Supplementary Materials). In such a layer, every copper(II) ion is connected with two nickel(II) ions, and each nickel(II) cation serves as a node of the network interacting with four copper(II) cations. The ab layers form …ABAB… stack motifs with adjacent layers shifted by ca. [0, 0.45y, −0.5z], and hence, the nodes of A layer are not positioned above the cavity center of the B layer ( Figure  S2, Supplementary Materials). It is noted that the layer topology is identical to that in the 2D compound of [{Cu(pn)2}2Mn(NCS)6]n·2nH2O [31]. The shortest heterometallic distances are 5.616 and 5.708 Å between Ni(II) and Cu(II) cations connected by four bridging thiocyanate ions located at equatorial positions, whereas the apical positions are occupied by the non-bridging thiocyanate forming the longest Ni-N bonds. These distances for Ni and Cu ions connected via thiocyanate bridge are shorter than in [CuLN4{Ni(NCS)4(H2O)2}]n (6.342 Å) [81] and [Cu(oxpn)Ni(μ-NCS)(H2O)(tmen)]2(X)2 (oxpn = N,N'-bis(3-aminopropyloxamide)) (6.313 and 6.335 Å for X = PF6 − and X = ClO4 − , respectively) [82]. In 1, the Cu-Cu distance between copper(II) ions from adjacent layers are similar to those inside the layer (e.g., are 7.284, 7.987 Å) and is 7.638 Å. The Ni-Ni distance between nickel(II) ions in one layer is 11.317 Å, while the distance between nickel(II) ions from the adjacent layers is significantly smaller (9.214 Å).
Most of the observed hydrogen bonds (N-H . . . N/S, O-H . . . N/S, N-H . . . O) were found inside the ab layer between N1, N2 and N12 NH 2 groups and N4 and S4 atoms of non-bridging thiocyanate anion as well as the O7 water molecule located in a cavity of ab layer (Table S2, Supplementary Materials). Moreover, these interactions may account for the slight elongation of Ni1-N4 according to Ni1-N5 and Ni1-N6 bonds. There were found only two interlayer hydrogen bonds: O7-H7A . . . S6[x, 1−y, −1/2+z] and N11-H11B . . . S5[−x, −y, 1−z]. In the former case, interactions are mediated by the O7 water molecule, whereas in the latter, direct contact between paramagnetic lattices exists. Thus, it can be assumed that the layers interact with each other affecting the observed antiferromagnetic coupling.
Two important points should be mentioned. First, analysis of coordination spheres and atomic charges indicated that Cu(II) occupies a pentacoordinate position, whereas Cu(I) is located in a tetrahedral environment. This indicates that a redox process occurs during the reaction and/or crystallization, associated with the oxidation of some thiocyanates and reduction in Cu(II). As a result, a stable [Cu I (NCS)(SCN) 3 ] 3− unit is formed, which is coordinated terminally (N2 and N3) to Ni(1) atom. Second, the ladder running along the b axis is formed due to three thiocyanates bridging Ni(II) and Cu(I) cations with two symmetrically related S2 thiocyanate anions involved in ladder rails and S3 forming rungs between Ni1 and Cu3 atoms ( Figure 4). Hence, [Cu II (trien)] 2+ moieties are located in pending side chains connected to Ni(II) via S4 thiocyanate with an S4 atom forming also a direct bridge between two Cu(II) units ( Figure 3). and atomic charges indicated that Cu(II) occupies a pentacoordinate position, whereas Cu(I) is located in a tetrahedral environment. This indicates that a redox process occurs during the reaction and/or crystallization, associated with the oxidation of some thiocyanates and reduction in Cu(II). As a result, a stable [Cu I (NCS)(SCN)3] 3− unit is formed, which is coordinated terminally (N2 and N3) to Ni(1) atom. Second, the ladder running along the b axis is formed due to three thiocyanates bridging Ni(II) and Cu(I) cations with two symmetrically related S2 thiocyanate anions involved in ladder rails and S3 forming rungs between Ni1 and Cu3 atoms ( Figure 4). Hence, [Cu II (trien)] 2+ moieties are located in pending side chains connected to Ni(II) via S4 thiocyanate with an S4 atom forming also a direct bridge between two Cu(II) units ( Figure 3). . We can presume that the {Ni II Cu II 2} trinuclear unit bridged by S4 thiocyanate anion should be sufficient. Three metals involved in interaction form T shape motif (with NCS(4) anion acting as a linker) with 90.08(2)° for C4-S4-Cu2 and 171.19 (5) In this structure, S1 and S5 thiocyanates are coordinated terminally to Ni(II), whereas S6 to Cu(I) and S2, S3 and S4 anions are involved in bridge formation. Thiocyanate coordination geometry might be described by Cu-S-C and Cu/Ni-N-C angles. Cu II -S-C angles range from 108.30 (9) to 109.57(16)°, whereas for T-shaped S4 linker this value is much smaller (90.08(2)°). Cu/Ni-N-C angles range from 154.1(2) to 178.7(4)° with larger values for the N-terminal ligand, while the smallest angle was observed for the N,S-bridging S2 thiocyanate involved in ladder propagation.
The shortest intermetallic distances are created between Cu1 and Cu3[−x, 1−y,−z] ions (5.653 Å), whereas for Cu-Ni, it is 6.913 Å. In the crystal network, the copper and nickel units are tightly packed (packing index: 67.1% [90]), forming columns running along the b axis, which are connected by three hydrogen bonds N17-H17A . . . S1   Table S6, Supplementary Materials). The differences in fingerprints for both Cu(II) units prove that they form slightly different interactions in the network ( Figure S3, Supplementary Materials). However, in both cases, the most numerous are H…H, S…H and N…H interaction (data in brackets) [95,96].

X-ray Absorption Spectroscopy
The XAS measurements showed the selected 3d metals (Ni, Cu) L-edge and the N and O K-edge structure for 1, 2 and 3 compounds (Figures 7,S4-S6, Supplementary Materials). The L-edge XAS spectra are supposed to allow for determining the differential orbital covalency and the electron structure of metal ion bonds in complexes, which allow for characterizing the metal-ligand interactions and may improve the interpretation of magnetic properties of newly synthesized materials. The normalized Cu L-edge X-ray absorption spectra of 1, 2 and 3 are presented in Figure 8. The Cu L-edge spectrum is related to the Cu 2p3/2→3d (L3) and 2p1/2→3d (L2) transitions and consists of two peaks split by ~20 eV with an intensity ratio above 2: 1 [97]. Three features of L-edge XAS spectra, namely total intensity, energy shift and spectral shape, represent the bonding properties. Within the series 1-3, the integrated intensity of the L3-edge transition peak decreases in the order 1  2  3 indicating that the average ligand character, and therefore the covalency of the unoccupied metal d-based orbitals, increases in the order 1  2  3 ( Table 5). The trend results from changes in the symmetry of copper(II) surrounded by nitrogen atoms of various amines and, in some cases, also by sulfur atoms of bridging thiocyanates. Metal Ledge energy pattern is dependent on three factors: the charge on the absorbing metal atom in the molecule (Zeff), the ligand field splitting of the d-orbitals, and any difference in the nature of the ligand valence orbitals. In the case of Cu(II), similarly to the high-spin Fe(III) [98], the contribution of the ligand field is weak and should be negligible; the energy shift for the 2p3/2→3d transitions decreases in the same order as the integrated intensity of L3edge: 1  2  3, and the maximum energy shift is only 0.4 eV in the L-edge spectra between complexes 1 and 3 (Table 5). This is in contrast to low-spin Fe(III) systems with t2g-eg energy splitting reaching even 1.4 eV [99].

X-ray Absorption Spectroscopy
The XAS measurements showed the selected 3d metals (Ni, Cu) L-edge and the N and O K-edge structure for 1, 2 and 3 compounds (Figure 7, Figures S4-S6, Supplementary Materials). The L-edge XAS spectra are supposed to allow for determining the differential orbital covalency and the electron structure of metal ion bonds in complexes, which allow for characterizing the metal-ligand interactions and may improve the interpretation of magnetic properties of newly synthesized materials. The normalized Cu L-edge X-ray absorption spectra of 1, 2 and 3 are presented in Figure 8. The Cu L-edge spectrum is related to the Cu 2p 3/2 →3d (L 3 ) and 2p 1/2 →3d (L 2 ) transitions and consists of two peaks split by~20 eV with an intensity ratio above 2: 1 [97]. Three features of L-edge XAS spectra, namely total intensity, energy shift and spectral shape, represent the bonding properties. Within the series 1-3, the integrated intensity of the L 3 -edge transition peak decreases in the order 1 → 2 → 3 indicating that the average ligand character, and therefore the covalency of the unoccupied metal d-based orbitals, increases in the order 1 → 2 → 3 ( Table 5). The trend results from changes in the symmetry of copper(II) surrounded by nitrogen atoms of various amines and, in some cases, also by sulfur atoms of bridging thiocyanates. Metal L-edge energy pattern is dependent on three factors: the charge on the absorbing metal atom in the molecule (Z eff ), the ligand field splitting of the d-orbitals, and any difference in the nature of the ligand valence orbitals. In the case of Cu(II), similarly to the high-spin Fe(III) [98], the contribution of the ligand field is weak and should be negligible; the energy shift for the 2p 3/2 →3d transitions decreases in the same order as the integrated intensity of L 3 -edge: 1 → 2 → 3, and the maximum energy shift is only 0.4 eV in the L-edge spectra between complexes 1 and 3 (Table 5). This is in contrast to low-spin Fe(III) systems with t 2g -e g energy splitting reaching even 1.4 eV [99].   The normalized Ni L-edge X-ray absorption spectra of 1, 2 and 3 are presented in Figure 9. The spectra are split into an L3-edge (2p3/2) around 853 eV and an L2-edge (2p1/2) around 870 eV due to the 2p spin-orbit coupling [100]. The structure of L3-edge and L2edge multiplets is characteristic of high-spin pseudo-Oh Ni(II) complexes. In contrast to Cu(II), the L3-edge and L2-edge peaks are split into two features due to distortion of Oh geometry of the [Ni(NCS)6] 4− unit, which is observed mainly in the case of complexes 1 and 2 with thiocyanates as bridging ligands. The structure of 1 was formed as a result of bridging four thiocyanate ions in equatorial positions to copper centers. The other two ions in the trans position in [Ni(NCS)6] 4− remain non-bridging ligands. Complex 2 is the first example of bridging the thiocyanate ion as μ1,3,3-NCS that is not found in the {NiCu2} unit so far. In these complexes, the half-field eg orbitals are split, and two peaks in the main part of the L-edge are observed with the energy of 1.0 eV and 1.2 eV for L3-edge and L2-edge features, respectively. The integrated intensity of the L3-edge transition multiplet decreases (Table 6) in the order: 1  2  3, indicating that the average ligand character, and therefore the covalency of the unoccupied metal d-based orbitals, increases in the same order. Nickel L-edge energy did not change and was constant for all the complexes, confirming that nickel retained its configuration and coordination with the six thiocyanates, as shown in the crystallographic section.  The normalized Ni L-edge X-ray absorption spectra of 1, 2 and 3 are presented in Figure 9. The spectra are split into an L 3 -edge (2p 3/2 ) around 853 eV and an L 2 -edge (2p 1/2 ) around 870 eV due to the 2p spin-orbit coupling [100]. The structure of L 3 -edge and L 2 -edge multiplets is characteristic of high-spin pseudo-O h Ni(II) complexes. In contrast to Cu(II), the L 3 -edge and L 2 -edge peaks are split into two features due to distortion of O h geometry of the [Ni(NCS) 6 ] 4− unit, which is observed mainly in the case of complexes 1 and 2 with thiocyanates as bridging ligands. The structure of 1 was formed as a result of bridging four thiocyanate ions in equatorial positions to copper centers. The other two ions in the trans position in [Ni(NCS) 6 ] 4− remain non-bridging ligands. Complex 2 is the first example of bridging the thiocyanate ion as µ 1,3,3 -NCS that is not found in the {NiCu 2 } unit so far. In these complexes, the half-field e g orbitals are split, and two peaks in the main part of the L-edge are observed with the energy of 1.0 eV and 1.2 eV for L 3 -edge and L 2 -edge features, respectively. The integrated intensity of the L 3 -edge transition multiplet decreases (Table 6) in the order: 1 → 2 → 3, indicating that the average ligand character, and therefore the covalency of the unoccupied metal d-based orbitals, increases in the same order. Nickel L-edge energy did not change and was constant for all the complexes, confirming that nickel retained its configuration and coordination with the six thiocyanates, as shown in the crystallographic section.

Magnetism
The plot of χT for 1 is given in Figure 10. In the whole measured range, it obeys the Curie-Weiss law with C = 2.104 cm 3 mol −1 K and θ = 0.17 K. The value of χT at 298 K is equal to 2.25 cm 3 ·mol −1 ·K (4.24 B.M.), which is slightly larger than 1.75 cm 3 ·mol −1 ·K expected for two uncoupled Cu(II) ions (0.75; S = 1 / 2 ; g = 2.0) and one Ni(II) ion (1; S = 1; g = 2). The value of χT product decreases slightly with temperature, and at 50 K, reaches a minimum of 2.14 cm 3 ·mol −1 ·K and then increases with the lowering temperature to the value 2.19 cm 3 ·mol −1 ·K at 3.2 K. This is followed by a rapid decline of the χT product to 2.155 cm 3 ·mol −1 ·K at 1.8 K. The increase in the value χT in the low-temperature range may be due to ferromagnetic interactions between metal ions through SCN − ligands. This is supported by the positive Weiss constant. The small decrease up to 50 K can be caused by measurement inaccuracies due to small spins or temperature-independent paramagnetism, especially for Ni(II) ions [101,102], while the rapid lowering of the χT product at 1.8 K can be caused by antiferromagnetic interactions through hydrogen bonds and zero-field splitting for Ni(II).

Magnetism
The plot of χT for 1 is given in Figure 10. In the whole measured range, it obeys the Curie-Weiss law with C = 2.104 cm 3 mol -1 K and θ = 0.17 K. The value of χT at 298 K is equal to 2.25 cm 3 ·mol −1 ·K (4.24 B.M.), which is slightly larger than 1.75 cm 3 ·mol −1 ·K expected for two uncoupled Cu(II) ions (0.75; S = 1 /2; g = 2.0) and one Ni(II) ion (1; S = 1; g = 2). The value of χT product decreases slightly with temperature, and at 50 K, reaches a minimum of 2.14 cm 3 ·mol −1 ·K and then increases with the lowering temperature to the value 2.19 cm 3 ·mol −1 ·K at 3.2 K. This is followed by a rapid decline of the χT product to 2.155 cm 3 ·mol −1 ·K at 1.8 K. The increase in the value χT in the low-temperature range may be due to ferromagnetic interactions between metal ions through SCNligands. This is supported by the positive Weiss constant. The small decrease up to 50 K can be caused by measurement inaccuracies due to small spins or temperature-independent paramagnetism, especially for Ni(II) ions [101,102], while the rapid lowering of the χT product at 1.8 K can be caused by antiferromagnetic interactions through hydrogen bonds and zerofield splitting for Ni(II).   Due to the 2D structure of 1, it was possible to perform only approximate calculations of the magnetic parameters. Both the Cu-S bond lengths and the Cu-S-C angles of the two thiocyanate bridges, which are of key importance for the magnetic interactions, do not differ significantly (see above). Therefore, the {NiCu2} trimetallic unit was taken as a representative model for the analysis of magnetic properties. The Hamiltonian of the type H = −2JCuNi(SCu1 SNi + SNi SCu2) was applied. The JCuNi is the Heisenberg exchange coupling constant between adjacent Cu(II)-Ni(II) ions. The magnetic data were fitted using the PHI program [103], taking into account also other parameters such as TIP, nickel(II) zero-field splitting and the molecular field. The best fit was obtained for the following parameters: gav = 2.201, JCuNi = 0.059 cm −1 , zJ' = 0.019 cm −1 , TIP = 461 × 10 −6 cm 3 mol −1 and DNi = 2.271 cm −1 . We made various modifications to the model (Table S7, Supplementary Materials), which do not affect the sign of the JCuNi exchange parameter, which is always positive with values ranging from 0.06 to 0.48 cm −1 . Doubling of the considered interacting unit to the six-metal {Ni2Cu4} and introduction of an additional interaction between the trimetallic subunits leads to similar results. All the modeling tests carried out indicate that the interactions are indeed very weak but predominantly ferromagnetic. This can be explained as follows: The unpaired electrons occupy the eg orbital of nickel(II) ion and dx2-y2 of copper(II) ion. Both dz2 and dx2-y2 Ni(II) magnetic orbitals overlap with the σ orbitals of the bridging thiocyanate ion. We also suggest that the spin density of dx 2 -y 2 of Cu delocalizes on π(SCN -) systems, which leads to the orthogonal arrangement of both SOMO systems in line with effective weak ferromagnetic interactions derived from the fits of magnetic data ( Figure 12  Due to the 2D structure of 1, it was possible to perform only approximate calculations of the magnetic parameters. Both the Cu-S bond lengths and the Cu-S-C angles of the two thiocyanate bridges, which are of key importance for the magnetic interactions, do not differ significantly (see above). Therefore, the {NiCu 2 } trimetallic unit was taken as a representative model for the analysis of magnetic properties. The Hamiltonian of the type H = −2J CuNi (S Cu1 S Ni + S Ni S Cu2 ) was applied. The J CuNi is the Heisenberg exchange coupling constant between adjacent Cu(II)-Ni(II) ions. The magnetic data were fitted using the PHI program [103], taking into account also other parameters such as TIP, nickel(II) zero-field splitting and the molecular field. The best fit was obtained for the following parameters: g av = 2.201, J CuNi = 0.059 cm −1 , zJ' = 0.019 cm −1 , TIP = 461 × 10 −6 cm 3 mol −1 and D Ni = 2.271 cm −1 . We made various modifications to the model (Table S7, Supplementary Materials), which do not affect the sign of the J CuNi exchange parameter, which is always positive with values ranging from 0.06 to 0.48 cm −1 . Doubling of the considered interacting unit to the six-metal {Ni 2 Cu 4 } and introduction of an additional interaction between the trimetallic subunits leads to similar results. All the modeling tests carried out indicate that the interactions are indeed very weak but predominantly ferromagnetic. This can be explained as follows: The unpaired electrons occupy the e g orbital of nickel(II) ion and d x2-y2 of copper(II) ion. Both d z2 and d x2-y2 Ni(II) magnetic orbitals overlap with the σ orbitals of the bridging thiocyanate ion. We also suggest that the spin density of d x  The plot of χT for 2 is given in Figure 13. The χT(T) curve in the whole measured range obeys the Curie-Weiss law in the whole measured range with C = 2.049 cm 3 mol -1 K and θ = −1.93 K. The value of χT at 295 K is equal to 2.07 cm 3 ·mol −1 ·K (4.06 B.M.), which is slightly bigger than the value (1.75 cm 3 ·mol −1 ·K) expected for two uncoupled Cu(II) ions (0.75; S = 1 /2; g = 2.0) and one Ni(II) ion (1; S = 1; g = 2). Cu(I) ion is diamagnetic and gives no contribution to the magnetism except for the negligible contribution to the diamagnetic correction. The χT product decreases slowly with temperature to 10 K and then more rapidly reaches 1.69 cm 3 ·mol −1 ·K at 1.8 K. The small decrease up to 10 K can be caused by measurement inaccuracies due to small spins or temperature-independent paramagnetism, especially for Ni(II) ions [101,102], while the rapid lowering of the χT product below 10 K can be caused by antiferromagnetic interactions through thiocyanate bridges and zero-field splitting for Ni(II). This is supported by the negative Weiss constant. The orthogonal orientation of the resultant natural magnetic orbital systems is in line with the weak ferromagnetic Ni(II)-Cu(II) interactions found from the fits of magnetic data.
The plot of χT for 2 is given in Figure 13. The χT(T) curve in the whole measured range obeys the Curie-Weiss law in the whole measured range with C = 2.049 cm 3 mol −1 K and θ = −1.93 K. The value of χT at 295 K is equal to 2.07 cm 3 ·mol −1 ·K (4.06 B.M.), which is slightly bigger than the value (1.75 cm 3 ·mol −1 ·K) expected for two uncoupled Cu(II) ions (0.75; S = 1 / 2 ; g = 2.0) and one Ni(II) ion (1; S = 1; g = 2). Cu(I) ion is diamagnetic and gives no contribution to the magnetism except for the negligible contribution to the diamagnetic correction. The χT product decreases slowly with temperature to 10 K and then more rapidly reaches 1.69 cm 3 ·mol −1 ·K at 1.8 K. The small decrease up to 10 K can be caused by measurement inaccuracies due to small spins or temperature-independent paramagnetism, especially for Ni(II) ions [101,102], while the rapid lowering of the χT product below 10 K can be caused by antiferromagnetic interactions through thiocyanate bridges and zero-field splitting for Ni(II). This is supported by the negative Weiss constant. The JCuNi and JCuCu are the Heisenberg exchange coupling constants between adjacent Cu(II)-Ni(II) and Cu(II)-Cu(II) ions, respectively. Magnetic data were fitted using the PHI program [103], taking into account also other parameters such as TIP, ZFS for Ni(II) ion and molecular field correction. The best fit (red lines in Figures 13 and 14) was obtained for the following parameters: gav = 2.085, JCuNi = 0.057 cm −1 , JCuCu = -0.718 cm −1 , zJ' = 0.009 cm −1 , TIP = 567 × 10 −6 cm 3 mol −1 , DNi = 0.735 cm −1 .  Due to structural reasons, the analysis of magnetic properties was based on the trimetallic {NiCu II 2 } unit, ignoring the bridges through [Cu(SCN) 3 (NCS)] 3− diamagnetic unit. The Hamiltonian of the type H = −2J CuNi (S Cu1 S Ni + S Ni S Cu2 ) -2J CuCu (S Cu1 S Cu2 ) was applied. The J CuNi and J CuCu are the Heisenberg exchange coupling constants between adjacent Cu(II)-Ni(II) and Cu(II)-Cu(II) ions, respectively. Magnetic data were fitted using the PHI program [103], taking into account also other parameters such as TIP, ZFS for Ni(II) ion and molecular field correction. The best fit (red lines in Figures 13 and 14) was obtained for the following parameters: g av = 2.085, J CuNi = 0.057 cm −1 , J CuCu = -0.718 cm −1 , zJ' = 0.009 cm −1 , TIP = 567 × 10 −6 cm 3 mol −1 , D Ni = 0.735 cm −1 . The JCuNi and JCuCu are the Heisenberg exchange coupling constants between adjacent Cu(II)-Ni(II) and Cu(II)-Cu(II) ions, respectively. Magnetic data were fitted using the PHI program [103], taking into account also other parameters such as TIP, ZFS for Ni(II) ion and molecular field correction. The best fit (red lines in Figures 13 and 14) was obtained for the following parameters: gav = 2.085, JCuNi = 0.057 cm −1 , JCuCu = -0.718 cm −1 , zJ' = 0.009 cm −1 , TIP = 567 × 10 −6 cm 3 mol −1 , DNi = 0.735 cm −1 .  The magnetization curve at 2 K ( Figure 14) shows no hysteresis loop. Saturation magnetization equals 3.96 B.M. at 7 T corresponds to the ground state of metal ions with the total spin equal 2 for the {NiCu 2 } unit.
The unpaired electrons are located on the e g orbitals of nickel(II) ions and d x copper(II) ions are linked by µ 1,3,3 -NCS anions in the manner that enables contacts between two d z 2 -y 2 orbitals of both Cu complexes through the part of π(SCN − ) system located on S atom (Figure 15). The magnetic interaction Cu-Ni paths are similar to 1, but another antiferromagnetic Cu-Cu path appears with much shorter Cu-S distances (ca. 2.36 Å) compared to 1, which thus dominates the scheme of magnetic exchange interactions. The unpaired electrons are located on the eg orbitals of nickel(II) ions and dx 2 -y 2 orbital of copper(II) ions. However, in this case, octahedral nickel(II) and both square pyramidal copper(II) ions are linked by µ1,3,3-NCS anions in the manner that enables contacts between two dz 2 -y 2 orbitals of both Cu complexes through the part of π(SCN -) system located on S atom ( Figure 15). The magnetic interaction Cu-Ni paths are similar to 1, but another antiferromagnetic Cu-Cu path appears with much shorter Cu-S distances (ca. 2.36 Å) compared to 1, which thus dominates the scheme of magnetic exchange interactions. Figure 15. Schematic representation of spin density delocalization from the dx 2 −y 2 orbital of Ni(II) onto the σ(SCN − ) system (a) and from the dx 2 −y 2 orbital of Cu(II) onto the π(SCN − ) system (b) in 2. The magnetic interaction Cu-Ni paths in 2 are similar to that of 1 depicted in Figure 12. However, due to significant overlap of natural magnetic orbitals of Cu(II) systems of end-on thiocyanatobridged Cu(II) moieties (dCu-S ca. 2.36 Å)(b), the Cu-Cu antiferromagnetic interaction prevailed in 2, as was shown by the fit of magnetic data.
The room temperature (293 K) effective magnetic moment of 3 is 4.91 B.M., which is slightly higher than the expected spin-only value for uncoupled four Cu(II) and one Ni(II) ions, i.e., 4.47 B.M. (4 × 0.5; S = 2; g = 2.0 and S = 1; g = 2). By taking into account that for both copper and nickel, the g factors usually are well over 2, we have an agreement for both g = 2.19.  The magnetic interaction Cu-Ni paths in 2 are similar to that of 1 depicted in Figure 12. However, due to significant overlap of natural magnetic orbitals of Cu(II) systems of end-on thiocyanato-bridged Cu(II) moieties (d Cu-S ca. 2.36 Å) (b), the Cu-Cu antiferromagnetic interaction prevailed in 2, as was shown by the fit of magnetic data.

Conclusions
The room temperature (293 K) effective magnetic moment of 3 is 4.91 B.M., which is slightly higher than the expected spin-only value for uncoupled four Cu(II) and one Ni(II) ions, i.e., 4.47 B.M. (4 × 0.5; S = 2; g = 2.0 and S = 1; g = 2). By taking into account that for both copper and nickel, the g factors usually are well over 2, we have an agreement for both g = 2.19.

Conclusions
We prepared [{Cu(pn) 2 } 2 Ni(NCS) 6 ] n ·2nH 2 O (1), [{Cu II (trien)} 2 Ni(NCS) 6 Cu I (NCS)] n (2) and [Cu(tren)(NCS)] 4 [Ni(NCS) 6 ] (3) showing different coordination topologies-2D layer and 1D ladder together with a complex salt composed of discrete mononuclear species, respectively. In the structure of 1, there are layers with paramagnetic centers connected by four thiocyanate anions. Hydrogen bond inspections revealed that the O7 water molecule plays a crucial role in crystal network formation, which is a pivot point for many interlayer hydrogen bonds. Compound 1 is a coordination polymer presenting a relatively scarce example of ferromagnetic coupling between nickel(II) and copper(II) ions. Compound 2 shows a rare ladder topology with rails and rungs formed by thiocyanate anions bridging arranged alternately by Cu(I) and Ni(II) ions, whereas [Cu II (trien)] 2+ moieties are antennas decorating the ladder and acting as side chains connected to Ni(II) cations. Hence, in the description of the magnetic interaction, we took into account {NiCu II 2 } trinuclear units featuring µ 1,3,3 -NCS bridging mode that enables dominating antiferromagnetic interactions between the Cu moieties and separated by Cu(I) units. The unique ladders in this structure form a tightly packed system without solvent molecules and are connected by hydrogen bonds. The highest Cu:Ni ratio was reached in 3, showing discrete mononuclear complexes forming columns connected by hydrogen bonds. In all three complexes, the topology was tailored due to the proper selection of the amine, and it affected the magnetic couplings. In (1), the ferromagnetic properties resulted from the spin density of Cu d x 2 -y 2 delocalized onto π(SCN − ) systems and hence, from the orthogonal arrangement of both SOMO systems in line with effective weak ferromagnetic interactions derived from the fits of magnetic data. In (2), the overall weak antiferromagnetic effect comes from two concurrent Cu-Ni (ferromagnetic) and Cu-Cu (antiferromagnetic) pathways, with the latter predominating. Compound (3), as a compound without a thiocyanato bridge, shows a spin-only magnetic moment. The differences in the topology also resulted in changes in Cu(II) coordination sphere geometry-6 coordinated 4+2 in 1, square pyramid in 2 and trigonal bipyramid in 3. Analysis of XAS proves that the average ligand character and the covalency of the unoccupied metal d-based orbitals for copper(II) and nickel(II) increase in the following order: 1 → 2 → 3.