Slow Magnetic Relaxation in a [Co₄O₄] Cubane Complex with Tridentate NNO-Schiff Base Ligands

Abstract

Two tetranuclear Co(II) complexes, [Co₄(pmab)₄Cl₄] (1) and [Co₄(pmab)₄(OBz)₂]Cl₂ (2) [Hpmab = 2-{(p-pyridinylmethylene)amino}benzenemethanol], have been synthesized and characterized through single-crystal X-ray diffraction, IR and UV-VIS spectroscopy, and magnetic measurements. Structural analysis revealed that both complexes possess a [Co₄O₄] cubane-like metal core connected by μ₃-alkoxo bridges. Magnetic measurements of Complex 1 indicate weak ferromagnetic interactions (J ~ +0.75 cm⁻¹) within the tetranuclear core, while Complex 2 exhibits antiferromagnetic behavior due to the presence of syn-syn bridging benzoate ligands. Alternating current (AC) magnetic measurements suggest that Complex 1 exhibits slow magnetic relaxation behavior.

1. Introduction

Since the discovery of the single-molecule magnet (SMM) properties of the Mn₁₂ cluster in 1993 [1], numerous magnetic molecules exhibiting slow magnetic relaxation have been synthesized and extensively studied over the past 30 years [2,3,4,5,6]. These SMMs, with their high relaxation barriers (U_eff) and blocking temperatures (T_B), hold promise for applications in quantum computing, spintronic devices, and high-density data storage [7,8]. This potential has driven the development of SMMs with enhanced performance [9,10,11].

Two key factors that determine the U_eff and T_B of SMMs are a large ground-state spin (S) and a large negative zero-field splitting parameter (D). While early SMM research focused on Mn(III) and Fe(III) [12,13], cobalt-based SMMs remained relatively unexplored due to their typically large positive D values. However, it has been suggested that in cubane structures, a negative D value can be achieved by designing the cubane motif to promote dominant ferromagnetic interactions within the tetranuclear metal core and optimize the orthogonality of the anisotropy axes of individual metal ions [14]. Based on this premise, the tetranuclear Co(II) complex [Co₄(hmp)₄(MeOH)₄Cl₄] (Hhmp = 2-hydroxymethylpyridine) was synthesized, exhibiting magnetic hysteresis due to the ground-state spin of S_T = 6 and negative magnetic anisotropy (D_mol < 0) [15]. Moreover, examples of Co(II) single-ion magnets (SIMs) exhibiting slow magnetic relaxation even when D_ion > 0 have been reported [16,17]. Since these discoveries, research on Co(II)-based SMMs and SIMs has progressed rapidly, and in particular, polynuclear clusters with the [Co₄O₄] cubane core as a structural motif have been reported [18,19,20,21]. In recent study, a [Co₇O₁₂] cluster [22] and a [Co₄O₄]₃ cluster [23], both featuring extended cubane structure, have garnered attention not only in the field of magnetochemistry but also as nanomaterials.

The cubane-type [M₄O₄] motif has been widely reported as a structural feature in the magnetochemistry of first-row transition metals, appearing in many complexes including [Mn₄O₄], [Ni₄O₄], and [Cu₄O₄] [24,25,26]. These complexes are often synthesized using chelating ligands with hydroxyl groups, which facilitate the formation of cubane-like structures due to their high bridging and chelating abilities. NO-type bidentate ligands, such as Hhmp used in [Co₄(hmp)₄(MeOH)₄Cl₄], are commonly employed due to their ease of synthesis and availability from commercial sources. However, these bidentate ligands cannot fully occupy the coordination sites of the cubane core, leading to the coordination of solvent molecules. These solvent molecules can dissociate from the metal ions under varying temperature or pressure conditions, potentially destabilizing the molecular structure. To address this issue and improve the stability of these complexes for practical applications as magnetic materials, it is essential to design complexes using tridentate or tetradentate chelating ligands rather than simple bidentate ligands [27,28]. These ligands can effectively occupy the coordination sites of the cubane core, preventing solvent coordination and enhancing the structural stability of the complexes.

In our study, we focused on synthesizing polynuclear complexes using tridentate ligands featuring an o-aminobenzyl alcohol structure [29,30]. In this research, we successfully synthesized two new tetranuclear Co(II) complexes with a [Co₄O₄] cubane core, [Co₄(pmab)₄Cl₄] (1) and [Co₄(pmab)₄(OBz)₂]Cl₂ (2), using the NNO-type tridentate ligand Hpmab [2-{(2-pyridinylmethylene)amino}benzenemethanol] (Figure 1). These complexes are unique in that they lack removable coordinating solvent molecules in their cubane cores, significantly reducing the risk of ligand dissociation and subsequent complex decomposition. This enhanced stability is crucial for the practical application of these complexes as magnetic materials. This paper reports the synthesis, molecular structures, and magnetic properties of these tetranuclear Co(II) complexes, as well as the slow magnetic relaxation phenomenon observed in Complex 1.

2. Materials and Methods

2.1. Preparation

All chemicals were used as received unless otherwise noted. Methanol was purified by distillation over magnesium turnings.

2.1.1. Hpmab

To a methanol solution (20 mL) of 2-aminobenzyl alcohol (2.432 g, 20 mmol), 2-pyridinecarboxaldehyde (2.142 g, 20 mmol) was added dropwise with stirring. The resulting mixture was refluxed for 4 h, then filtered and allowed to stand at room temperature. The resulting yellowish-orange solution was concentrated in vacuo to remove all solvent. The crude ligand, obtained as an oily substance, was used in the complexation reactions without further purification. IR (ATR) [cm⁻¹]: 3331 (w), 2842 (w), 1609 (m), 1590 (m), 1494 (s), 1470 (vs), 1424 (s), 1365 (m), 1303 (m), 1253 (m), 1048 (s), 748 (vs), 623 (m).

2.1.2. [Co₄(pmab)₄Cl₄] (1)

To a methanol solution (10 mL) containing cobalt(II) chloride hexahydrate (0.119 g, 0.5 mmol) and Hpmab (0.106 g, 0.5 mmol), a methanol solution (2 mL) of potassium t-butoxide (0.011 g, 0.1 mmol) was slowly added over 30 min. The solution turned dark orange, and after 20 min, orange microcrystals suitable for X-ray analysis were obtained. 1∙5H₂O; Yield: 0.078 g, 48%. C₅₂H₅₄Cl₄Co₄N₈O₉: C, 47.58; H, 4.15; N, 8.54%. Found: C, 47.51; H, 3.77; N, 8.50%. IR (ATR) [cm⁻¹]: 2873 (vw), 1597 (s), 1493 (m), 1442 (m), 1366 (m), 1241 (m), 1190 (m), 1156 (m), 1018 (vs), 776 (vs), 746 (s), 720 (m), 624 (s). UV-VIS (reflectance) [10³ cm⁻¹]: 8.7, 14.4, 14.9, 15.8.

2.1.3. [Co₄(pmab)₄(OBz)₂]Cl₂ (2)

A methanol solution (10 mL) containing cobalt(II) chloride hexahydrate (0.119 g, 0.5 mmol) and Hpmab (0.106 g, 0.5 mmol) was prepared. Sodium benzoate (0.072 g, 0.5 mmol) was then added to the solution, resulting in a dark orange color. The solution was slowly concentrated at room temperature over one day, yielding orange single crystals suitable for X-ray analysis. 2∙2CH₃OH; Yield: 0.018 g, 9.9%. C₆₆H₅₄Cl₂Co₄N₈O₈: C, 56.02; H, 4.29; N, 7.69%. Found: C, 55.90; H, 3.95; N, 7.43%. IR (ATR) [cm⁻¹]: 2856 (w), 1593 (s), 1551 (vs), 1492 (w), 1399 (vs), 1371 (s), 1304 (m), 1190 (m), 1017 (s), 773 (s), 723 (vs), 679 (m), 623 (m). UV-VIS (reflectance) [10³ cm⁻¹]: 8.9, 14.2, 15.0, 15.9.

2.2. Measurements

Elemental analyses for C, H, and N were obtained at the Elemental Analysis Service Center, Kyushu University. Infrared spectra were recorded on a Bruker VERTEX70-S FT-IR Spectrometer (Bruker Corp., Billerica, MA, USA) using the attenuated total reflection (ATR) method. UV-VIS reflection spectra were obtained using a PERKIN ELMER Lambda900Z UV/VIS/NIR Spectrometer (PerkinElmer Inc., Waltham, MA, USA) and an Ocean Optics USB2000+ Fiber Optic Spectrometer (Ocean Optics Inc., Dunedin, FL, USA). Magnetic susceptibilities were measured using a Quantum Design MPMS-XL5R SQUID (Superconducting Quantum Interference Device) susceptometer (Quantum Design Inc., San Diego, CA, USA) under an applied magnetic field of 0.1 T in the temperature range 2–300 K. The susceptibilities were corrected for the diamagnetism of the constituent atoms using Pascal’s constant [31].

2.3. Single Crystal X-Ray Diffraction

Diffraction data were collected on a Rigaku Vari-Max Saturn CCD 724 diffractometer (Rigaku Corp, Tokyo, Japan) with graphite-monochromated Mo Ka radiation (λ = 0.71069 Å) at the Analytical Research Center for Experimental Sciences, Saga University. Data collection was performed using CrystalClear SM 2.0 r16 [32]. The crystal was kept at 113 K during data collection, and multi-scan correction for absorption was applied. Data processing was performed using CrysAlisPro 42.49 [33] for 1 and CrystalClear for 2. The crystal data and experimental parameters are summarized in Table S1. Structures were solved by direct methods (ShelXT-2016/6) and expanded using Fourier techniques [34]. Non-hydrogen atoms were refined anisotropically, while hydrogen atoms were placed geometrically in calculated positions and refined with a riding model. For Complex 2, BASF/TWIN refinement was applied. Solvent masks were calculated for disordered solvent molecules [35]. For Complex 1, 360 electrons were found in a volume of 1360 ų in one void per unit cell, and 854 electrons were found in a volume of 3271 Å in one void per unit cell for 2. These are consistent with the presence of 8[H₂O] and 12[CH₃OH] per molecular formula for 1 and 2, respectively. The final cycle of full-matrix least-squares refinement on F² using ShelXL-2016/6 [36] was based on observed reflections and variable parameters and converged with unweighted and weighted agreement factors of R and R_w. Olex2-1.5 [37] was used as an interface to the ShelX program package. The molecular structures were drawn using Mercury-2022.1.0 [38].

3. Results and Discussion

3.1. Synthetic Outcomes and Characterization

The tridentate ligand Hpmab was isolated as a crude product, and its formation was confirmed by FT-IR spectroscopy (Figure S1). The disappearance of the C=O stretching peak at 1710 cm⁻¹, characteristic of 2-pyridinecarboxaldehyde, and the appearance of the C=N stretching vibration at 1609 cm⁻¹, indicated the formation of a Schiff base. Additionally, the C–O stretching vibration at 1048 cm⁻¹, attributed to the benzyl alcohol moiety, was observed. In the FT-IR spectrum of 1 (Figure S2), the strong and characteristic bands corresponding to the azomethine ν(C=N) and alcoholic ν(C–O) stretching vibrations of the ligand were shifted to lower frequencies, appearing at 1597 and 1018 cm⁻¹, respectively [22]. This significant shift in the ν(C–O) band suggests the formation of a μ-alkoxo bridged structure [27]. For Complex 2, similar IR bands were observed at 1593 and 1017 cm⁻¹ (Figure S3). Additionally, the asymmetric and symmetric carboxylate stretching bands, ν_as(COO) and ν_s(COO), were observed at 1551 and 1399 cm⁻¹, respectively. The separation value (Δ) between the ν_as(COO) and ν_s(COO) frequencies was 152 cm⁻¹, indicating that the benzoate ion coordinates to the Co(II) ions in a syn-syn bridging mode [39]. Elemental analysis results for 1 and 2 showed good agreements between the found and calculated values for the composition of [Co₄Cl₄(pmab)₄(H₂O)₅] and [Co₄Cl₂(OBz)₂(pmab)₄(CH₃OH)₂], respectively.

3.2. Structural Studies

The molecular structure of 1 is illustrated in Figure 2a, and selected bond distances and angles are summarized in

Table 1. Selected bond distances and angles of 1.

Bond	Distance/Å	Angle	Angle/°
Co1–Cl1	2.4763(12)	Cl1–Co1–O1′	170.11(7)
Co1–O1	2.028(3)	O1–Co1–N2	162.00(12)
Co1–O1′	2.194(3)	O1′–Co1–N1	168.09(12)
Co1–O1′’	2.109(2)	O1–Co1–N1	87.13(11)
Co1–N1	2.169(3)	N1–Co1–N2	77.81(13)
Co1–N2	2.103(3)	O1–Co1–O1″	81.19(10)
Co1···Co1′	3.1398(11)	O1″–Co1–N2	114.09(12)
Co1···Co1″	3.2371(7)	Co1–O1–Co1′	96.03(10)
		Co1–O1–Co1‴	102.96(11)
		Co1′–O1–Co1‴	97.56(10)

Symmetry code: (′) 1 − x, 3/2 − y, +z; (″) 5/4 − y, 1/4 + x, 5/4 − z; (‴) −1/4 + y, 5/4 − x, 5/4 − z.

. Complex 1 crystallizes in the tetragonal space group I4₁/a, where the asymmetric unit contains one [Co(pmab)Cl] unit. The tridentate ligand coordinated to Co(II) ion in a meridional mode. The metal core of this tetranuclear complex consists of four Co(II) ions and four μ₃-bridging oxygen atoms from the deprotonated (pmab)⁻ ligands, forming a [Co₄O₄] cubane structure. Each Co(II) ion adopts an octahedral geometry, with its six coordination sites occupied by three μ₃-oxygen atoms, two nitrogen atoms, and one chloride ion. The Co–O and Co–N bond lengths range from 2.028(3) to 2.194(3) Å and 2.103(3) to 2.169(3) Å, respectively. In contrast, the Co1–Cl1 bond length of 2.4762(12) Å. Notable differences exist in the lengths of the three principal axes in the octahedral geometry around Co1. Specifically, the distances between coordinated atoms along these axes are 4.083(4) Å for O1–Co1–N2, 4.255(4) Å for O1′–Co1–N1 [symmetry code: (’) 1 − x, 3/2 − y, +z], and 4.653(3) Å for O1″–Co1–Cl1 [symmetry code: (″) 5/4 − y, 1/4 + x, 5/4 − z], suggesting a pronounced rhombic distortion in the octahedral environment of Co1. SHAPE analysis values for Co1, with OC-6 and TPR-6 indices of 2.817 and 9.101, respectively, further confirm this significant distortion [40].

The overall structure can fundamentally be regarded as a dimer-of-dimers configuration, composed of bis-μ₂-alkoxo-bridged dicobalt units. The structure of the dicobalt unit is depicted in Figure 2b. Generally, in dinuclear structures with planar tridentate ligands, these ligands typically lie within the dinuclear plane. However, in Complex 1, the chloride ion is positioned within the dinuclear plane, causing the (pmab)⁻ ligand to adopt a meridional coordination mode via the axial direction relative to the dinuclear plane. The coordination of the chloride ion within the dinuclear plane also results in variations in the Co–O distances within the Co₂O₂ unit. The Co–O distance from Co1 to the chelated oxygen O1 is 2.028(3) Å, while the distance to the bridging oxygen O1′ trans to the chloride ion is Co1–O1′ = 2.194(3) Å, approximately 0.17 Å longer. Consequently, the Co₂O₂ square in the dinuclear unit of 1 takes the shape of a parallelogram. Additionally, the Co1–O1″ distance, which corresponds to the stacking distance between the dinuclear units, is 2.109(2) Å, differing from any of the Co–O distances within the dinuclear plane. The intramolecular metal-to-metal distances are Co1···Co1′ = 3.1398(11) Å and Co1···Co1″ = 3.2371(7) Å. The bridging angles around the O1 atom are Co1–O1–Co1′ = 96.03(10)°, Co1–O1–Co1‴ = 102.96(11)° [symmetry code: (‴) −1/4 + y, 5/4 − x, 5/4 − z], and Co1′–O1–Co1‴ = 97.56(10)°. Since all these angles are all greater than 90°, the [Co₄O₄] cubane core is highly distorted, forming a triakis tetrahedron with four Co vertices.

Each of the four (pmab)⁻ ligands in 1 is engages in π-π stacking with four different neighboring molecules. The two aromatic rings within the tridentate ligand—the phenyl ring of the benzyl alcohol moiety and the pyridine ring—stack with the pyridine and phenyl rings of neighboring molecules, respectively, at a distance of 3.804(3) Å (Figure 3a). As a result, the cubane complexes function as building bricks, forming an ordered 3-D structure, as shown in Figure 3b and Figure S4.

Figure 4a shows the structure of the complex cation of 2, and selected bond distances and angles are summarized in

Table 2. Selected bond distances and angles of 2.

Bond	Distance/Å	Angle	Angle/°
Co1–O1	2.138(5)	O1–Co1–N2	165.8(2)
Co1–O11	2.127(5)	O12–Co1–O2	170.5(2)
Co1–O12	2.091(5)	O11–Co1–N1	175.7(2)
Co1–O2	2.077(5)	O1–Co1–N1	88.4(2)
Co1–N1	2.144(7)	N1–Co1–N2	78.6(3)
Co1–N2	2.184(7)	O1–Co1–O12	87.5(2)
		O12–Co1–N2	105.6(2)
Co1···Co11	2.978(2)	Co1–O1–Co11	88.57(18)
Co1···Co12	3.2561(17)	Co1–O1–Co13	100.7(2)
		Co11–O1–Co13	101.1(2)

Symmetry code: (¹) 1 − x, 2 − y, +z; (²) 3/2 − y, 1/2 + x, 1/2 − z; (³) −1/2 + y, 3/2 − x, 1/2 − z.

. Complex 2 is an ionic compound containing two chloride anions and crystallizes in the tetragonal chiral space group I$\overline{4}2$d with Z = 4. A Flack parameter of 0.48 suggests racemic twinning. The asymmetric unit of 2 contains half of the [Co₂(pmab)₂(OBz)]⁺Cl⁻ binuclear unit. Similar to 1, Complex 2 features a [Co₄O₄] cubane core and a dimer-of-dimers structure. However, a key difference is the presence of syn-syn bridging benzoate ions. The dicobalt structure is illustrated in Figure 4b. Unlike Complex 1, the ligands in 2 are coplanar with the Co₂O₂ dinuclear plane, and the benzoate anion binds to the Co atoms axially. The bond distances between the metal ions and coordinating atoms within the dinuclear plane range from 2.127(5) to 2.184(7) Å, while the axial bond distances are slightly shorter, ranging from 2.077(5) to 2.091(5) Å, indicating compressed axial distortion. The SHAPE calculation result (OC-6) of 1.696, smaller than that of 1, suggests a less distorted geometry overall.

The intramolecular metal-to-metal distances in 2 are Co1···Co1¹ = 2.978(2) Å and Co1···Co1² = 3.2561(17)Å [symmetry code: (¹) 1 − x, 2 − y, +z; (²) 3/2 − y, 1/2 + x, 1/2 − z]. The bridging angles around the O1 atom are Co1–O1–Co1¹ = 88.57(18)°, Co1–O1–Co1³ = 100.7(2)°, and Co1¹–O1–Co1³ = 101.1(2)° [symmetry code: (³) −1/2 + y, 3/2 − x, 1/2 − z]. Compared to 1, these values suggest that each face of the cubane core in 2 is closer to a square, and that the distortion into a triakis tetrahedron is smaller. Additionally, intra- and intermolecular π–π stacking interactions occur between phenyl rings, with a centroid-to-centroid distance of 3.6284(8) and 3.6729(8) Å (Figure 5a). In the crystal packing, the complex cations are arranged in stacks along the c-axis, forming an ordered 3D structure (Figure S5). This arrangement creates channels along the c-axis, within which the counter anions (Cl⁻) are positioned at intervals of 11.424(6) Å (Figure 5b).

No clear signals were observed in the powder X-ray diffraction (PXRD) measurements of bulk samples for 1 and 2, which prevented comparisons with the patterns simulated from single-crystal X-ray diffraction (SCXRD) data. The solvent mask method indicated that both complexes contain crystal solvents within the voids of their 3D structures formed by π–π stacking. However, the absence of PXRD signals is likely due to crystal efflorescence caused by the evaporation of these solvent molecules.

3.3. Magnetic Properties

Magnetic susceptibility measurements were performed using a SQUID (superconducting quantum interference device) magnetometer. The temperature dependence of the χ_MT values for 1 and 2 is presented in Figure 6. At 300 K, the χ_MT values for 1 and 2 are 10.73 and 10.56 cm³ mol⁻¹ K, respectively, which are significantly higher than the spin-only value of 7.5 cm³ mol⁻¹ K expected for four uncoupled high-spin Co(II) ions (S = 3/2). This increase is attributed to the contribution from orbital angular momentum, resulting from the distorted octahedral geometry of the Co(II) ions. For Complex 1, the χ_MT values remain constant down to 70 K, after which they decrease steadily to 7.92 cm³ mol⁻¹ K at 2 K. This gradual decrease is caused by spin–orbit coupling or zero-field splitting effects. In contrast, the χ_MT values of 2 show a steady decline as the temperature decreases, reaching 0.63 cm³ mol⁻¹ K at 2 K. This significant decrease suggests that antiferromagnetic interactions dominate in 2.

The interpretation of magnetic exchange interactions in Co(II) complexes is known to be difficult to estimate accurately due to the effects of spin–orbit coupling and zero-field splitting. We initially attempted to fit the data using a simple S = 3/2 tetranuclear cubane model, including the zero-field splitting parameter D [15]. However, the calculations failed to converge, yielding nonsensical results, as shown in Figure S6. Consequently, we applied the molecular field approximation, treating the cubane structure as a dimer-of-dimers. In this approach, J represents the interaction within the dinuclear unit, while zJ′ represents the interaction between the dinuclear units. The spin Hamiltonian for the dinuclear unit is expressed in Equation (1), and the corrected susceptibility (χ_corr) obtained through the molecular field approximation is given by Equation (2) [41]. The analysis was performed using the PHI 3.1.6 program [42].

H = −JS₁S₂ + Δ(L_z² − 2/3) − (3/2)κλLS + β[−(3/2)κL + g_eS]H

(1)

$${\chi }_{\mathrm{corr}}={\displaystyle \frac{\chi }{1-\left(\frac{z{J}^{\prime }}{N{\mu }^2}\right)\chi }}$$

(2)

The obtained magnetic parameters for 1 and 2 are summarized in

Table 3. Magnetic parameters for 1 and 2.

Complex	J/cm−1	J′/cm−1	g	λ/cm−1	κ	Δ/cm−1
1	+0.75	−0.015	2.22	−180	0.955	−44.3
2	−6.08	−0.629	2.29	−140	1.09	−6.13

, where λ represents the spin–orbit coupling parameter, κ is the orbital reduction factor, and Δ is the axial distortion parameter corresponding to the zero-field splitting parameter. In Complex 1, the magnetic interaction within the [Co₄O₄] core is weak, but the presence of an exceedingly small ferromagnetic interaction is suggested, similar to what has been observed in related complexes [22,28,43,44]. In contrast, Complex 2 is dominated by antiferromagnetic interactions. Typically, simple [Co₄O₄] clusters, such as in 1, exhibit ferromagnetic interactions due to the orthogonality of the magnetic orbitals, which promotes parallel spin alignment. However, when additional bridging ligands like carboxylate ions are present, as in 2, antiferromagnetic interactions become predominant due to superexchange pathways facilitated by these ligands [45,46]. Therefore, the differing χ_MT behavior of 2 compared to 1 is attributed to antiferromagnetic interactions caused by the presence of syn-syn bridging benzoate ligands.

The relatively large negative Δ value of 1 is supported by magnetization measurements conducted in the temperature range of 2 to 10 K. The magnetization curve at 2 K does not saturate, even up to 5 T, and the observed value of 6.7 Nμ_B is smaller than the expected value 12 Nμ_B for an S = 6 ground state (Figure 7a). This indicates the presence of strong magnetic anisotropy in the Co(II) ions. The anisotropy is further confirmed by the M vs. H/T plot shown in Figure 7b, where the isofield lines do not superimpose, indicating significant magnetic anisotropy in the ground state for 1. In contrast, the magnetization curve of 2 at 2 K reaches only 1.7 Nμ_B, without saturating even at 5 T (Figure S7a). The curve displays a subtle S-shape, indicative of antiferromagnetic interactions. Due to the small magnetization values, measurements were conducted only at 2 K and 4 K. However, the M vs. H/T plots do not overlap on a single curve, suggesting the presence of weak antiferromagnetic interactions or magnetic anisotropy (Figure S7b).

Frequency-dependent alternating current (AC) magnetization measurements were performed at frequencies of 1, 10, 100, 500, and 1000 Hz (Figure 8a,b). In the absence of a static DC field, only subtle out-of-phase (χ″_M) signals were detected (Figure S8). However, when an external magnetic field of 2000 Oe was applied, the χ″_M component exhibited a significantly larger response and a more pronounced frequency dependence below 5 K, indicating field-induced slow magnetic relaxation. The χ″_M responses are similar to those for reported complexes as weak SMMs [22,27,47].

Since no maximum value was observed in the χ″_M graph, a rough estimate of the energy barrier (E_a) and pre-exponential factor (τ₀) was attempted using the Debye model via Equation (3) [48,49,50].

$$\ln {\displaystyle \frac{\chi ″}{{\chi }^{\prime }}}=\ln \left(\omega {\tau }_0\right)+{\displaystyle \frac{E_a}{k_{B}T}}$$

(3)

Figure 9 shows plots of ln(χ″/χ′) versus 1/T for each frequency, based on Equation (3). The resulting lines are not parallel, likely due to correlation errors caused by the ridiculously small χ″ values or the possibility of multi-relaxation processes. Consequently, it was not possible to estimate the relaxation parameters using this method.

4. Conclusions

New tetranuclear Co(II) complexes 1 and 2, featuring a [Co₄O₄] cubane core without coordinating solvent molecules, were synthesized via the reaction of CoCl₂∙6H₂O with the NNO Schiff base ligand Hpmab. The crystal structures of these complexes were elucidated through single-crystal X-ray analysis, revealing that both complexes form tetranuclear cluster based on a dimer-of-dimers structure. The magnetic properties of 1 and 2 were investigated by DC magnetometry, which showed that ferromagnetic interactions dominate in 1, while antiferromagnetic interactions dominate in 2. This difference can be attributed to the presence or syn-syn bridging benzoates in 2, absent in 1. Additionally, 1 exhibited frequency-dependent out-of-phase signals below 5 K in AC studies, indicating relatively slow magnetic relaxation in magnetization.