Synthesis, crystal structures, HF-EPR, and magnetic properties of six-coordinate transition metal (Co, Ni, and Cu) compounds with a 4-amino-1,2,4-triazole Schiff-base ligand

We have synthesized a series of transition metal compounds [M(L)2(H2O)2] (M = Co (1), Ni (2), and Cu (3)) by using the 4-amino-1,2,4-triazole Schiff-base ligand via the hydrothermal methods. They are all mononuclear compounds with the octahedral geometry. Direct-current magnetic and HF-EPR measurements were combined to reveal the negative D values (–28.78 cm−1, –10.79 cm−1) of complexes 1 and 2, showing the easy-axis magnetic anisotropies of compounds 1 and 2. Applying a dc field of 800 Oe at 2.0 K, the slow magnetic relaxation effects were observed in compound 1, which is a remarkable feature of single-ion magnets.


Introduction
Single-ion magnets (SIMs), as a kind of single molecule magnets (SMMs), have been intensely studied in the past few decades due to their broad application prospects in the high-density magnetic information storage, quantum computation, and molecular spintronics. [1][2][3] They also provide an ideal model for understanding the quantum phenomena in the mesoscopic world. As is known, the large ground-state spin S and zero-eld splitting (ZFS) parameter D decide the relaxation energy barriers (U eff ), which are treated as the reason for the slow magnetic relaxation of SMMs. 4 We usually describe the energy barrier of transition metals as U eff ¼ |D|S 2 or U eff ¼ (S 2 À 1/4)|D|. 5 The SIMs of transition metal ions and single paramagnetic lanthanide and actinide ions have been widely investigated, such as Mn(III,IV), 6,7 Fe(II), 8,9 Fe(III), 10 Co(I,II), 11 Ni(II), 14,15 Dy(III) 16,17 and Tb(III). 18,19 Noteworthily, the Co(II) ion is an admirable candidate for constructing SIMs due to its large ground-state spin (S) and large magnetic anisotropy (D) on the basis of experimental and theoretical calculations. Moreover, the magnetic anisotropy of SIMs can be easily adjusted by the interactions between the ligand eld splitting and the spin-orbit coupling, and the ZFS parameters are determined by the coordination structural pattern and the distortion degree of its congurations. So far, a mass of Co(II)-based SIMs with variable coordination numbers from 2 to 8 and different coordination geometries, such as the distorted trigonal-planar, square-pyramidal, octahedral polyhedron and the like, have been investigated. 11 It is noted that most of the Co(II)-based SIMs with distorted octahedron geometries show the positive zero-eld splitting parameter 11b,11c,11g,11i (D > 0). However, six-coordinate Co(II)-SIMs with negative D values are rarely reported. 11k,11f, 12,13,20 The typical example is a chiral star-shaped Co II Co 3 III compound (HNEt 3 ) + (Co II Co 3 III L 6 ) À [H 2 L ¼ R-4-bromo-2-((2-hydroxy-1phenylethylimino)methyl)phenol] reported by Gao et al. 11a This compound has a distorted trigonal prismatic conguration around Co II ion, showing the slow magnetic relaxation behaviours with a large negative D value. Novikov et al. have also reported a negative D in a six-coordinate mononuclear Co(II)-SIM with trigonal prism geometry. 11k Recently, Zhang and coworkers observed that a series of hexa-coordinate Co(II) complexes with trigonal antiprismatic geometries exhibit the slow relaxation behaviours with the negative D values. 20 As the sign of the magnetic anisotropy parameter D is associated with the coordinated geometries of the hexa-coordinate Co(II)-based SIMs, it was proposed by Gomez-Coca et al. 11f that the six-coordinate high-spin systems (S ¼ 3/2) with the twisted octahedral geometry could lead to positive ZFS parameters (D > 0), while trigonal prism and antiprismatic geometries are in accordance with negative ones (D < 0). In this context, we report a six-coordinate Co(II) compound with the 4-amino-1,2,4-triazole Schiff-base ligand having a slightly distorted octahedral geometry (1). The compound shows a negative D value and the slow magnetic relaxations. As the Schiff base ligands can easily form stable compounds with most of the transition metallic ions in diverse valence states and different manners of coordination, 21 we also chose the 4-amino-1,2,4-triazole Schiff-base ligand to synthesize a series of transition metal compounds with nickel(II) (2) and copper(II) (3) ions.
In our study, all of the three compounds with the octahedral geometry structure. 1 and 2 possess the negative magnetic anisotropy, and 1 shows the eld-induced slow magnetic relaxation. Magnetization and HF-EPR measurements were adopted to characterize the magnetic properties of the three compounds.

Experimental section
General All the reagents were commercially purchased without further purication. The IR spectra were recorded on a Nicolet 5DX spectrometer with the wavenumber in the range of 400-4000 cm À1 (KBr pellets). Powder X-ray diffraction (PXRD) patterns were measured on a X'Pert PRO automated diffractometer at the room temperature (Cu Ka, l ¼ 1.5406 A). Thermogravimetric analyses (TGA) were performed in a ow of nitrogen at a heating rate of 10 C min À1 using a NETZSCH TG 209 F3. Magnetic properties of polycrystalline samples were measured in the temperature range of 2-300 K and the eld up to 7 T using a Quantum Design VSM SQUID magnetometer. High frequency/eld electron paramagnetic resonance (HF-EPR) were measured on a locally developed instruments with pulsed magnetic elds. 22

Synthesis
The ligand was synthesized by methods previously reported in the literature 23,24 (Scheme S1 †). were synthesized on the basis of the reported procedure. 24 PXRD spectra (see the details in ESI, Fig. S1 †) were used to verify the phase purity of compounds 1-3. As is seen, the experimental patterns agree well with the calculated patterns. TGA (thermogravimetric analysis) results (Fig. S2 †) show that there are no guest molecules in compound 1-3. In the IR spectra, the broad peak at 3519 cm À1 in free ligand was due to characteristic vibrations of O-H, while it was not present in compound 1-3, suggesting the loss of H atom on the HL. The absorptions (a strong band at 1461 cm À1 in 1, 1464 cm À1 in 2, and 1463 cm À1 in 3) conrm the intense vibrations of C]N, which appeared at 1480 cm À1 in HL. The red shi of absorption peak was attributed to the coordination of the N atoms of the azomethine groups with the central metal ions.

X-ray structural determination
The diffraction data were collected on a Bruker APEX-II CCD diffractometer with graphite-monochromatized Mo Ka radiation (l ¼ 0.071073 nm). The structures of the complexes were solved by direct methods using SHELXS-97 and rened by full-matrix least-squares on F 2 using SHELXS-97. All nonhydrogen atoms were rened anisotropically. All hydrogen atoms were positioned geometrically and rened as riding. Experimental details of crystal data, data collection parameters and renement statistics for compounds 1-3 are summarized in Table 1, while the selected bond lengths and angles are presented in Table S1. †

Structural description
Single-crystal X-ray diffraction was performed to determine the structures of compound 1-3. It revealed that 1-3 were isomorphous with similar structural parameters. Among them, the structures of 1 and 3 have already been reported. 23 Here, we would briey highlight 1 to discuss the following magnetic structure correlations. Their crucial crystallographic data are shown in Tables 1 and S1. † In Fig. 1, 1 is a mononuclear compound crystallized in monoclinic (P2 1 /n) space group. Regarding the crystallographic structure of 1, the central Co II ion demonstrates a six-coordinate geometry and lies in the central location of the reverse symmetry (symmetry code A: Àx + 2, Ày + 1, Àz + 1) with two ligand anions and two coordinated water molecules. Two nitrogen atoms (N1, N1A) and two oxygen atoms (O1, O1A) provided by two ligands occupy the equatorial planes, while the axial position is occupied by two oxygen atoms (O3, O3A) from the coordinated water molecules, forming a slightly twisted octahedral geometry.

Magnetic measurements
The direct-current magnetic susceptibilities for 1-3 were measured over the temperature range of 2-300 K ( Fig. 2 and 3). For compound 1, at 300 K, the resulting c M T value is 2.733 cm 3 K mol À1 , larger than the expected value (1.875 cm 3 K mol À1 , g ¼ 2.00) for one single high spin Co(II) ion (S ¼ 3/2), which can be due to the strong orbital contribution. 9b With the reduction of temperature, in the temperature range of 60-300 K, the c M T value decreases slightly and then downs sharply to reach the minimum value of 1.60 cm 3 K mol À1 at 2 K, indicating the existence of largely unquenched orbital angular momentum. 11e,11h,27,28 The correlative eld-dependent magnetization measurements were carried out within 2.5-10 K at the eld range of 0-7 T (inset in Fig. 2). When the eld up to 7 T, the maximum magnetization value is 2.1 Nm B at 2 K, which is smaller compared with the theoretical saturation value (3 Nm B for S ¼ 3/2, g ¼ 2). The unsaturated M vs. H plot at high eld supports the existence of the magnetic anisotropy in 1.
As for 2, the c M T product is measured to be 1.194 cm 3 K mol À1 at 300 K, which is larger than the theoretical spin-only c M T value (1.00 cm 3 K mol À1 , g ¼ 2.00) and is expected for six-coordinate Ni(II) ions with the momentous spin-orbit contribution. Upon cooling, the c M T value decreases slowly before 30 K, and then sharply drops down to 0.38 cm 3 K mol À1 at 2 K, which might be due to ZFS of the Ni(II) ion and Zeeman effect. The eld-dependence magnetization of 2 was performed with the magnetic eld of 0-7 T and at temperatures of 2-5 K respectively (inset in Fig. 3). As increasing the magnetic eld, the magnetization gradually increases and then reaches 1.55 Nm B at 7 T and 2 K, which thus is smaller than the theoretical saturation value of 2.0 Nm B (S ¼ 1, g ¼ 2.0), indicating the fundamental magnetic anisotropy of 2.
As for 3, c M T value is about 0.43 cm 3 K mol À1 at 300 K, lower than the theoretical value (0.45 cm 3 K mol À1 ) for one single magnetic Cu(II) ion (g ¼ 2. 2). 29 The c M T vs. T plot is nearly at   This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 12833-12840 | 12835 between 20 and 300 K before the sharp decrease to 0.37 cm 3 K mol À1 at 2 K.
To gain further insight into magnetic properties, c M T vs. T and M vs. H curves of compound 1-3 were tted by PHI program 30 using the following spin Hamiltonian: where m B , D, E, S, and B represent the Bohr magneton, axial and rhombic ZFS parameters, the spin operator, the magnetic eld vector, respectively. The best-tting parameters are presented in Table 2.

High-eld EPR studies
It is common known that the sign of magnetic anisotropy plays a signicant role in the magnetic behaviours of high-spin Co(II) and Ni(II) compounds. The six-coordinate Co(II) and Ni(II) compounds with the octahedral geometric structure have been investigated to show ne both positive and negative easy-axis magnetic anisotropies. 11b, [31][32][33][34][35][36] The high eld/frequency electron paramagnetic resonance (HF-EPR) measurements were carried out to conrm the sign of D because the sign of D obtained from the magnetic data is not always reliable. The polycrystalline powders of 1-3 were measured in the frequency range of 60-500 GHz (Fig. 4, 5, and S5 †). As the amplitude of D for 1 ($28.78 cm À1 ) is much larger than the available microwave quantum energy (500 GHz $ 16.7 cm À1 ), no transition between Kramers doublets M S ¼ AE1/2 and M S ¼ AE3/2 can be seen, 11j,37,38 so the HF-EPR spectra of 1 were simulated base on the amplitude of D and the g values obtained from SQUID measurements and adjusting the ZFS parameter E, yielding |E| ¼ 4.78 cm À1 (Fig. 4b). As shown in Fig. 4a, two simulations were done with different signs of D, showing that the negative D value are well in accord to the experimental data. But beyond that, there are only two peaks in the spectra for 1, which is typical for high-spin 3/2 Co(II) systems with large negative D values due to the limit of magnetic eld. 39 As for 2, the well-resolved powder-pattern spectrum at 480 GHz of a triplet state (S ¼ 1) were received as shown in Fig. 5a (black trace). Simulations with positive (blue trace) and negative (red trace) D values evidence that the D value is negative, indicating the easy-axis magnetic anisotropy of 2. The correlative parameters are given as follows: D ¼ À10.79(5) cm À1 , E ¼ 3.08(5) cm À1 , g x ¼ g y ¼ 2.15(2), and g z ¼ 2.05 (2). These parameters are close to the obtained values from magnetic studies   Plots of HF-EPR resonances vs. frequency for 1 at 2 K. Squares are the experimental data, and blue, green, and red lines are the calculated lines using the spin-Hamiltonian parameters for the magnetic fields parallel to the molecular axes X, Y, and Z, respectively. The vertical dashed line is the frequency (170 GHz) at which the spectrum was taken and simulated.
above (Table 1) and the values of hexa-coordinate Ni(II) compounds in literature. [40][41][42][43][44][45] The entire resonance eld vs. frequency plot of 2 is demonstrated in Fig. 5b. Resonances drawn in this way forms characteristic branches, which are labelled according to Wasserman's terminology. 46 The calculation lines through these points are based on the combination of automatic nonlinear least square tting which uses articial judgment to eliminate the physical unreasonable results. The HF-EPR spectrum of 3 was recorded at 154 GHz and 4.2 K (Fig. S5 †). Three main peaks were observed which correspond to the anisotropic g values of the Cu(II) ion with S ¼ 1/2. The simulation to the spectrum results in the g values as g x ¼ 2.03(2), g y ¼ 2.07(2), and g z ¼ 2.27(2) with g iso ¼ 2.13, which agrees well with the magnetic studies (Table 1).

Dynamic (ac) magnetic properties of 1
In order to investigate the dynamic magnetic behaviours at low temperatures, ac susceptibilities were measured at 2-6 K for 1.
When no dc eld was applied, no out-of-phase ac susceptibility (c 00 ) signal could be observed at 2 K, which can be attributed to the existence of the magnetized quantum tunnelling (QTM). With the external magnetic eld increases, the c 00 signal could be seen, suggesting that the QTM effect could be suppressed by the applied dc eld. At 2.0 K and 1.0-996.5 Hz, the variable dc elds were performed to measure the out-of-phase ac susceptibilities for 1 to nd an optimum applied magnetic eld which can suppress the QTM phenomenon (Fig. S6 †). Consequently, an optimum eld of 800 Oe (at which the c 00 signals in a lower frequency range and with enough amplitude) was applied to detect the frequency and temperature-dependent ac susceptibilities within 2-6 K ( Fig. 6 and S7 †), and the eld-induced slow magnetic relaxation of 1 was observed.
From the ac data collected above, the Cole-Cole diagrams were charted as a form of c M 00 vs. c M 0 at 2-6 K and 800 Oe dc eld (Fig. 7). Fitting was performed by using the generalized Debye model as following (eqn (2)): where c T and c S represent the isothermal and adiabatic susceptibilities, respectively, u represents the angular frequency, s represents the relaxation time, and a indicates the deviation from a pure  Debye model. 47,48 The correlative tting parameters are listed in Table S3. † The parameter a is from 0.025 to 0.22, which indicates that the distribution of relaxation time of compound 1 is small.
Fitting the values of relaxation time within 2.0-10.0 K in the light of the Arrhenius law s À1 ¼ s 0 À1 exp(ÀU eff /kT) results in U eff ¼ 43.70 K and s 0 ¼ 1.12 Â 10 À7 s (Fig. S8 †). Comparing with the energy gap assessed from the value of |2D| (57.56 cm À1 ), the obtained U eff is rather small, and the evident curvature in the Arrhenius plot of 1 at low temperature manifests that there could exist multiple relaxation pathways, including non-negligible Raman or Raman-like, direct, and Orbach mechanisms. It is worth noting that the contributions of a Raman process cannot be neglected and a direct one-phonon will make some contributions at lower temperatures in terms of most of the Co(II)-based SIMs reported. 11g,11j,49 Therefore, a model including two possible mechanisms were applied to analyse this relaxation behaviour by eqn (3): 29,49a where the two terms in eqn (3) relate to the contributions of Raman or Raman-like and Direct mechanisms. The best t of experimental data is shown in Fig. S9 † with the parameters n ¼ 7.97, C ¼ 0.0037 K À7.97 S À1 , and A ¼ 99.12 K À1 S À1 . Compared with the anticipated value 9 of the Raman process of Kramers ions, the n obtained value of 7.97 is smaller, stating possibly a photoacoustic Raman process refer to virtual states, in which both of the optical and acoustic phonons are all considered. 49, 50 We can conclude that the slow relaxation behaviour is the result of synergistic effect of Raman-like and Direct processes.
No ac signal was observed on 2 and 3 under the dc elds of 0-0.1 T were applied (Fig. S10 †), indicating that 2 and 3 don't possess the properties of SIM or spin-phonon trapping properties. [51][52][53] Conclusions To sum up, we have successfully synthesized three 3d transition metal compounds based on 4-amino-1,2,4-triazole Schiff-base ligand via hydrothermal method and their structures were further characterized. All of 1, 2, and 3 are mononuclear complexes with the octahedral geometry. Magnetic measurements and HF-EPR studies manifest the easy-axis anisotropy with negative D parameters in 1 and 2, which might be due to the elongated octahedron of metal geometry. Ac magnetic susceptibility measurements conrm that 1 exhibit the typical eld-induced slow magnetic relaxation behaviours, while no slow magnetic relaxations were observed in 2 and 3. We have added one more example to the six-coordinate cobalt(II) compounds showing the eld-induced SIM behaviours with the negative magnetic anisotropy, which was rarely reported. Its dynamic magnetic behaviours could be explained via a Ramanlike process in high-temperature zone while the relaxation occurs via a direct process within the lower temperature range.

Conflicts of interest
There are no conicts to declare.