Slow Magnetic Relaxation in Cobalt(II) Complexes with One-Dimensional Hydrogen-Bonded Networks

Abstract

Two new cobalt(II) complexes with an unsymmetrical bidentate ligand, 2-(1,4,5,6-tetrahydropyrimidin-2-yl)-6-methoxyphenol (H₂mthp), were synthesized and crystallographically characterized. Tetra- and hexa-coordinate mononuclear complexes were selectively obtained by adjusting the stoichiometry of the base. The coordination geometry of hexa-coordinated complex was severely distorted from an ideal octahedron, due to the NO₅ coordination environment from the mixed coordination of one Hmthp⁻ and two H₂mthp ligands. Both complexes formed one-dimensional chain networks by hydrogen-bond and N-H···π interactions. Single-molecule magnet behavior was observed for the tetrahedral complex under zero magnetic field. The relatively short Co···Co distances induced non-zero intermolecular magnetic coupling, which split the ground ±M_s levels to suppress quantum-tunneling of magnetization. In the octahedral complex, by contrast, the distance was not short enough to induce the coupling. As a consequence, single-molecule magnetic behavior was observed for the octahedral complex only in the presence of an external static field.

1. Introduction

Single-molecule magnets (SMMs) [1] are of considerable current interest due to their potential applications in high-density storage, spintronics and quantum computing [2,3,4,5]. To exhibit slow relaxation of magnetization, a bistable spin ground state with a large negative axial zero-field splitting (ZFS) is important because the spin-reversal barrier U is proportionally dependent on the axial ZFS parameter D, where U = |D|S² for an integer spin system and |D|(S²−1/4) for a half-integer spin system). SMMs with a single paramagnetic center are called single-ion magnets (SIMs) [6]. Owing to the simplicity and facile design of the molecule, many lanthanide and first-row transition metal complexes have been studied to prepare SMMs with a strong magnetic anisotropy [7,8,9,10,11,12,13]. Cobalt(II) complexes are one of the most studied metal ions as SIMs because tetra- and hexa-coordinated cobalt(II) complexes often exhibit strong magnetic anisotropy [8,9,10,11,12,13,14,15,16,17]. In most of the examples of 3d metal-based SIMs, however, slow magnetic relaxation was not observed in the absence of an external magnetic field, due to fast relaxation via quantum-tunneling of magnetization (QTM). For a Kramers ion, QTM arises from the mixing of ±M_s level hyperfine or dipolar interactions [11]. Avoiding a close interaction between SMMs in the long distance is one of the possible solutions to avoid the QTM phenomena by dipolar interactions. On the other hand, we recently reported zero-field SIM behaviors in tetrahedral cobalt(II) complexes with one-dimensional hydrogen-bonded networks [12,13]. It is suggested that relatively short intermolecular distances (ca. 6 Å) and one-dimensional alignments of the complexes suppress the QTM by splitting the ground ±M_s levels by intermolecular magnetic coupling [18,19,20,21]. However, the correlation between intermolecular distances and the alignments of the SMMs within the chain structure remain unclear. In this paper, we synthesized an analogous tetrahedral cobalt(II) complex and a severely distorted octahedral cobalt(II) complex with one-dimensional networks (Scheme 1). The effects of intrachain distances and alignments of SIMs on QTM phenomena were investigated, and zero-field and field-induced SIM behaviors were observed in tetrahedral and octahedral complexes, respectively.

2. Results and Discussion

2.1. Preparation of the Ligand and Cobalt(II) Complexes

The ligand precursor 2-(1,4,5,6-tetrahydropyrimidin-2-yl)-6-methoxyphenol (H₂mthp) was prepared by a similar method to that reported for an analogous ligand in the literature [13]. The mononuclear bis-bidentate type complex [Co(Hmthp)₂]·C₂H₅OH was synthesized by a reaction of Co(BF₄)₂·6H₂O, H₂mthp and KO^tBu in 1:3:2 ratio in ethanol (Scheme 1). An excessive amount of the ligand precursor was required for the synthesis to prevent the formation of Co(OH)₂ as impurity. It should be noted that the use of other bases, such as triethylamine, was not successful, presumably because triethylamine was not strong enough to deprotonate H₂mthp. In the case of a smaller molar ratio of KO^tBu (1:3:1 ratio), a mononuclear tris-bidentate cobalt(II) complex [Co(Hmthp)(H₂mthp)₂]BF₄ (2BF₄) was obtained. Both cobalt(II) complexes tended to lose crystallinity in air by efflorescence, and solvent molecules of crystallization were substituted by water.

2.2. Crystal Structures of H₂mthp and Cobalt(II) Complexs

The ligand precursor H₂mthp was crystallized in chiral orthorhombic space group P2₁2₁2₁. In the crystal, the asymmetric unit consisted of two independent H₂mthp molecules and existed in zwitterionic form, with a deprotonated phenol group and a protonated imino group precedented in analogous compounds (Figure S1) [22]. The X-ray crystallographic analysis revealed that 1·C₂H₅OH and 2BF₄ were tetra-coordinated and hexa-coordinated cobalt(II) complexes, respectively (Figure 1 and Figure 2). In 1·C₂H₅OH, the crystallographic asymmetric unit consisted of 1 and an ethanol molecule of crystallization. The cobalt(II) ion was coordinated by phenolato-O and imino-N donors of Hmthp⁻ ligand in a bidentate fashion to afford a pseudo-tetrahedral coordination geometry (

Table 1. Selected bond parameters for 1·C₂H₅OH.

Atom-Atom	Length/Å	Atom-Atom-Atom	Angle/°
Co(1)-O(1)	1.913(2)	O(1)-Co(1)-N(2)	94.08(10)
Co(1)-O(3)	1.910(2)	O(3)-Co(1)-N(4)	93.27(10)
Co(1)-N(2)	1.982(3)	O(1)-Co(1)-O(3)	106.28(10)
Co(1)-N(4)	1.972(3)	O(1)-Co(1)-N(4)	126.84(10)
		O(3)-Co(1)-N(2)	125.07(11)
		N(2)-Co(1)-N(4)	114.22(11)

). The structural parameter τ₄ for 1 was calculated to be 0.77 (τ₄ = [360° − (α + β)]/141°, where α and β were the largest two angles in the coordination sphere), which was smaller than those of analogous complexes, indicating that the coordination geometry was more distorted from an ideal tetrahedron [23]. The mean plane angle between two bidentate ligands (O1-Co1-N2 plane to O3-Co1-N4 plane angle = 76.7°) deviated from 90°. Hydrogen-bonding interactions were observed between the N-H group of the ligand and phenolato-O atom of the neighboring molecule to construct one-dimensional hydrogen-bonded networks of 1. As the crystallographic inversion centers were located between two neighboring molecules in a chain, the magnetic anisotropy axes of the molecules should be in the same axis. Two intrachain Co···Co distances were not equivalent (5.979(4) and 6.106(4) Å). The interchain closest Co···Co distances were much longer (≥9 Å) than the intrachain ones, implying that the interchain magnetic interaction were negligible.

The 2BF₄ was crystallized in a triclinic system P$\overline{1}$ as 2BF₄·1.5C₂H₅OH (Figure 2). The hexa-coordinated geometry of the Co center was severely distorted by a NO₅ coordination mode. In 2⁺ cation, the Co center was coordinated by one Hmthp⁻ and two H₂mthp ligands. Hmthp⁻ acted as a bidentate ligand with phenolato-O atom and imino-N atoms as donor atoms, as was also observed in 1, whereas H₂mthp ligands coordinated via phenolato-O and methoxy-O atoms in zwitterionic form, as observed in the crystal of H₂mthp. The protonation state of the ligands was consistent with the stoichiometry of the reaction condition (H₂mthp:KO^tBu = 3:1). The coordination bonds with methoxy-O atoms were significantly longer than those with the other donor atoms (by 0.3 Å), implying the weak coordination of the methoxy-O atoms. The bond parameters of Co ion were within the range of a typical divalent high-spin cobalt(II) complex (

Table 2. Selected bond parameters for 2BF₄·1.5C₂H₅OH.

Atom-Atom	Length/Å	Atom-Atom-Atom	Angle/°
Co(1)-O(1)	2.002(3)	O(1)-Co(1)-N(2)	91.44(14)
Co(1)-O(3)	1.982(3)	O(3)-Co(1)-O(4)	72.58(11)
Co(1)-O(4)	2.354(3)	O(5)-Co(1)-O(6)	71.15(12)
Co(1)-O(5)	1.998(3)	O(1)-Co(1)-O(3)	90.51(11)
Co(1)-O(6)	2.387(3)	O(1)-Co(1)-O(5)	108.44(12)
Co(1)—N(2)	2.041(4)	O(1)-Co(1)-O(6)	91.16(11)
		N(2)-Co(1)-O(3)	111.63(15)
		N(2)-Co(1)-O(4)	92.64(14)
		N(2)-Co(1)-O(5)	93.62(14)
		O(3)-Co(1)-O(6)	83.54(12)
		O(4)-Co(1)-O(5)	87.83(12)
		O(1)-Co(1)-O(4)	162.93(11)
		N(2)-Co(1)-O(6)	164.58(13)
		O(3)-Co(1)-O(5)	148.22(14)
		O(4)-Co(1)-O(6)	89.29(12)

). The considerable distortion of the coordination geometry can be explained by the severe steric requirement of the six-membered chelate mode of the ligand in an octahedral geometry. The mean plane angles between the phenyl and tetrahydropyrimidinyl 6-membered rings of a Hmthp⁻ ligand was 38.6(3)°, indicating that the ligand underwent a strong steric hindrance on O-N chelate mode in an octahedral coordination geometry [24]. It was noted that such distortion of the ligand was not observed in 1 because a tetrahedral coordination geometry accepted much larger bite angles than an octahedral one (109.5° for tetrahedral and 90° for octahedral geometry) and, the mean plane angles observed in 1 (15.4(3)° and 5.5(2)°) were comparable to those of H₂mthp. As a result, the bite angle of the six-membered chelate mode (phenolato- and imino-N atoms) in 2⁺ was restricted to 91.44(14)°, although the angle of the five-membered chelate rings (phenolato-O and methoxy) accepted much smaller bite angles (72.58(11)° or 71.15(12)°). Although one-dimensional chain networks were formed by hydrogen-bond and N-H···π interactions, the shortest Co···Co distance was not the intrachain one but the interchain one (8.239(1) Å). This long intermolecular Co···Co distance suggested that the intermolecular magnetic interaction was negligible.

2.3. Magnetic Properties

2.3.1. Static Magnetic Properties

The temperature dependence of magnetic susceptibility was measured for 1·C₂H₅OH and 2BF₄·1.5C₂H₅OH to reveal their magnetic properties (Figure 3). The χ_MT product at 300 K for 1·C₂H₅OH (ca. 2.5 cm³ mol⁻¹ K), which was higher than the spin-only value 1.876 for the S = 3/2 system, was a typical value for tetrahedral cobalt(II) complexes. Upon cooling, the χ_MT products were almost constant, down to 100 K, followed by a steep drop of value because of ZFS. To determine the g-factors and the axial ZFS parameter (D), the temperature dependence of the χ_MT data and the field dependence of the magnetization were simultaneously fitted to the following spin Hamiltonian:

$$\widehat{H}=g\beta H\widehat{S}+D\left[{\widehat{\mathrm{S}}}_{\mathrm{z}}^2-\frac{S\left(S+1\right)}{3}\right]+zJ{S}_z\widehat{S}$$

(1)

where β and zJ are Bohr magneton and the intermolecular interaction, respectively (

Table 3. Experimental g-factors (g_x, g_y, g_z), axial ZFS parameters (D), intermolecular magnetic interactions (zJ), and temperature-independent paramagnetism (TIP).

	1·C2H5OH	2BF4·1.5C2H5OH	[Co(Hthp)2] i	[Co(Hmimn)2]·CH3OH ii
gx, gy	2.16	2.35	2.18	2.18
gz	2.56	2.42	2.48	2.55
D/cm−1	–49	–31	–30	–30
zJ/cm−1	–0.56	0	–1.5	–0.25
TIP/cm3 mol−1	0.0003	0.0005	0.0003	0.0003

i Hthp⁻ = 2-(1,4,5,6-tetrahydropyrimidin-2-yl)phenolate [13]. ii Hmimn⁻ = 2-(2-imidazolinyl)-6-methoxyphenolate.

). The transversal ZFS parameter (E) was not considered as it was difficult to determine from the magnetic data. Intermolecular magnetic interactions were considered for 1·C₂H₅OH to analyze the effect of intermolecular coupling on the magnetic relaxation dynamics. Although Hmthp⁻ was a derivative of the Hthp⁻ ligand, with the same coordination mode, the obtained g-factors for 1·C₂H₅OH were rather closer to those of [Co(Hmimn)₂]·CH₃OH than [Co(Hthp)₂] [12,13]. This suggested that the electronic structure of the phenolate donor had stronger influence on the g-factors. The large negative D value, owing to the second order spin-orbit coupling, was suggestive of SIM behavior with a spin-reversal barrier of ca. 100 cm⁻¹ (2|D|).

The χ_MT product for 2BF₄·1.5C₂H₅OH (ca. 2.8 cm³ mol⁻¹ K) was also higher than the spin-only value because of unquenched orbital contribution on the magnetic moment. The simultaneous fitting of the χ_MT vs. T and M vs. H plots was performed. Intermolecular interaction was not considered, because the intermolecular Co···Co distance was significantly long. The obtained zero-field splitting parameter D was negative and relatively small among octahedral Co^II SIMs, presumably due to severely distorted octahedral geometry [9,16,17].

2.3.2. Dynamic magnetic properties

Alternating current (ac) susceptibility measurements for 1·C₂H₅OH exhibited significant frequency dependence on the out-of-phase signal (χ_M”) in the absence of an external magnetic field at 1.9 K (Figure 4). This was indicative of slow magnetic relaxation in the absence of an external field. It should be noted that zero-field SIM behavior in a 3d metal complex is still very rare owing to fast relaxation via QTM. Upon heating, however, the χ_M” vs. frequency plot showed no change except the weakening of the χ_M” signal. This temperature independence is a characteristic feature of QTM. By applying an external field, on the other hand, another relaxation process appeared in the lower frequency region in the χ_M” vs. frequency plot (Figure S3 and Figure 4b). The temperature dependence was measured in the presence of 3.5 kOe, where the QTM relaxation pathway was completely suppressed and relatively small contribution of the direct relaxation process was expected. To elucidate the relaxation dynamics, the relaxation time τ was extracted by the generalized Debye-fit model:

$${\mathsf{\chi }}_{ac}\left(\mathsf{\omega }\right)={\mathsf{\chi }}_S+\frac{{\mathsf{\chi }}_T-{\mathsf{\chi }}_S}{1+{\left(i\mathsf{\omega }\tau \right)}^{1-\mathsf{\alpha }}}$$

(2)

The relaxation dynamics could not be fitted with a single relaxation process, as it showed different trends below and above ca. 4 K (Figure 5). The dynamics was nicely fitted with a combination of two processes, direct and Raman relaxation processes:

$${\tau }^{-1}=A{H}^{4}T+C{T}^n$$

(3)

where A, C, and n are coefficients, H is the magnetic field, T is the temperature (A = 0.667 K⁻¹ kOe⁻⁴ s⁻¹, C = 1.23×10⁻¹ s⁻¹ K^−5.2, n = 5.2). In the low temperature region (<4 K), the direct process, which involves relaxation from −M_s to +M_s with emission of a single lattice phonon, was the predominant relaxation process. This predominance could be elucidated by the presence of a relatively strong external field (3.5 kOe) as the τ was inversely proportional to the powers of an external field. Above 4 K, the relaxation dynamics were taken over by the Raman process, in which relaxation between ±M_s states occurred via virtual state. The n value should be ≤9 for a Kramers ion when both an acoustic and an optical phonon are considered, and the obtained n value for 1·C₂H₅OH was consistent with the expected value [25,26].

As observed in previously reported three analogous compounds [12,13], 1·C₂H₅OH exhibited one-dimensional hydrogen-bonded networks as well as zero-field slow magnetic relaxation. We reported that the relatively short intermolecular Co···Co distance induced a static exchange bias field to split the ground ±M_s levels. As this splitting of the ground ±M_s levels effectively suppressed the QTM phenomena, formation of hydrogen-bonded networks was a feasible approach for zero-field SIMs. In the case of one-dimensional chain, the efficiency of suppression of QTM was dependent on the molecular orientation and symmetry of the alignment of the chain. As the Co centers in the chain structures were antiferromagnetically coupled, alternating spin orientation should be the ground state. In this state, QTM would be suppressed, because the two neighboring molecules in a chain induced a dipolar field to split the ground ±M_s levels. When the dipolar field of the two neighboring molecules were stochastically oriented to opposite direction, on the other hand, the dipolar field would be canceled and QTM could not be suppressed, as observed in [Co(Hthp)₂] (Himn⁻ = 2-(1,4,5,6-tetrahydropyrimidin-2-yl)phenolate). This was not the case when the two intrachain Co···Co distances were significantly different because the dipolar field from the two neighboring molecules would not be canceled, as observed in [Co(Himn)₂] (Himn⁻ = 2-(2-midazlinyl)phenolate). In 1·C₂H₅OH, two intrachain distances were slightly different (ca. 0.1 Å) but the difference was only half the value of [Co(Himn)₂]. Consequently, the dipolar field was almost canceled, and the QTM was only partially suppressed in the absence of an external field. These results indicated that the difference in the intrachain Co···Co distances was an important factor to suppress the QTM as well as the intermolecular magnetic coupling.

Ac susceptibility measurement was also performed on complex 2BF₄·1.5C₂H₅OH. Unlike 1·C₂H₅OH, this complex did not exhibit slow magnetic relaxation in the absence of an external magnetic field, presumably because the Co···Co distance in 2BF₄·1.5C₂H₅OH was not short enough to cause intermolecular magnetic coupling to suppress the QTM (Figure S6). In the presence of a magnetic field, on the other hand, QTM was well suppressed, meaning that 2⁺ cation was a field-induced SIM (Figure 6). To investigate the relaxation dynamics, temperature dependence of the ac susceptibility was measured under a static field of 1.0 kOe. The τ vs. T plot suggested that there were two relaxation processes present and they switched at ca. 4 K. Below 4 K, the double-logarithmic plot showed a large degree of linearity suggesting a power law (τ ∝ T⁻ⁿ). The relaxation dynamics was well-fitted by combining the Raman and Orbach relaxation processes:

$${\tau }^{-1}=C{T}^n+{\tau }_0^{-1}\exp \left(\frac{{\Delta }_{\mathrm{Orbach}}}{k_{B}T}\right)$$

(4)

where τ₀ is the pre-exponential factor, ∆_Orbach is the relaxation barrier, and k_B is the Boltzmann constant (C = 45.5 s⁻¹ K^−2.6, n = 2.6, τ_o = 2.45×10⁻⁹ s, ∆_Orbach = 36.7 cm⁻¹). The n value was within the range of expected value for the Raman process. It was noted that the combination of the phonon-bottlenecked direct process (low temperature region) and the Raman process (high temperature region) did not give a reasonable fit. The relaxation barrier for the Orbach process ∆_Orbach was slightly smaller than the expected value from the ZFS parameter 2|D|. Thus, the relaxation dynamics of high-temperature region of 2BF₄·1.5C₂H₅OH were dominated by the Orbach process.

4. Materials and Methods

4.1. General Consideration

All the chemicals were used as received without further purification, except for Co(BF₄)₂·6H₂O, which was recrystallized from a water/ethanol mixed solvent before use. Elemental analyses (C, H, and N) were performed at the Research Institute for Instrumental Analysis, Kanazawa University. ¹H NMR measurements were carried out at 22 °C on a JEOL 400SS spectrometer. Chemical shifts were referenced to the solvent residual peak [27]. Infrared spectra were measured on a JASCO FT/IR-4200 spectrometer.

4.2. Preparations

2-(1,4,5,6-tetrahydropyrimidin-2-yl)-6-methoxyphenol (H₂mthp). A mixture of methyl 3-methoxysalicylate (9.11g, 50 mmol) and 1,3-diaminopropane (12 mL) was refluxed overnight. The unreacted 1,3-diaminopropane was evaporated off under ambient pressure, followed by the addition of 5 mL ethanol. After cooling, the yellowish residue was collected by filtration and washed with ethanol. Yield: 7.02 g, 68%. ¹H NMR (399 MHz, Methanol-d₄) δ 7.02 (d, J = 8.8 Hz, 1H, aryl-H), 6.83 (d, J = 7.7 Hz, 1H, aryl-H), 6.45 – 6.38 (m, 1H, aryl-H), 3.78 (s, 3H, -CH₃), 3.59 – 3.46 (m, 4H, -CH₂-), 2.02 (quin, J = 5.8 Hz, 2H, -CH₂-).

[Co(Hmthp)₂]·C₂H₅OH (1·C₂H₅OH). An ethanol solution (10 mL) of Co(BF₄)₂·6H₂O (34.1 mg, 0.10 mmol) was slowly added to a mixture of ethanol (10 mL), H₂mthp (62.0 mg, 0.30 mmol) and KO^tBu (23.0 mg, 0.20 mmol) in a Schlenk flask under Ar atmosphere. The reaction mixture was allowed to stand at room temperature for a few days, and red crystals were obtained. Yield: 11.3 mg, 22%. The crystallinity of this compound was easily lost in air, due to its efflorescent nature, and solvent molecules of crystallization were substituted by water. Anal Calcd for [Co(Hmthp)₂]·0.7H₂O = C₂₂H₂₇CoN₄O_4.7: C, 54.82; H, 5.73; N, 11.62%. Found: C,54.64; H, 5.58; 11.62%.

[Co(Hmthp)(H₂mthp)₂]BF₄·1.5C₂H₅OH (2BF₄·1.5C₂H₅OH). An ethanol solution (5 mL) of Co(BF₄)₂·6H₂O (35.0 mg, 0.10 mmol) was slowly added to a mixture of ethanol (5 mL), H₂mthp (62.2 mg, 0.30 mmol) and KO^tBu (10.4 mg, 0.09 mmol) in a Schlenk flask under Ar atmosphere. The reaction mixture was allowed to stand at room temperature for a few weeks, and purple crystals were obtained. Yield: 46.9 mg, 61%. The crystallinity of this compound was easily lost in air, due to its efflorescent nature, and solvent molecules of crystallization were substituted by water. Anal Calcd for [Co(Hmthp)(H₂mthp)₂]BF₄·2H₂O = C₃₃H₄₅BCoF₄N₆O₈: C, 49.58; H, 5.67; N, 10.51%. Found: C, 49.59; H, 5.76; 10.08%.

4.3. Crystallography

Crystallographic data are summarized in

Table 4. Crystallographic data and refinement parameters of 1, 2BF₄·1.5C₂H₅OH and H₂mthp.

Complex	1·C2H5OH	2BF4·1.5C2H5OH	H2mthp
Empirical formula	C24H32CoN4O5	C36H50BCoF4N6O7.5	C11H14N2O2
Formula weight	515.46	832.56	206.24
Crystal system	Monoclinic	Triclinic	Orthorhombic
Crystal dimensions/mm	0.08 × 0.05 × 0.02	0.15 × 0.09 × 0.07	0.16 × 0.11 × 0.07
Space group	P21/c	P$\overline{1}$	P212121
a/Å	10.3216(6)	12.1036(6)	7.5402(2)
b/Å	19.6289(11)	13.6859(6)	12.0806(3)
c/Å	11.8343(6)	13.9026(8)	22.6040(8)
α/°		114.527(5)
β/°	96.137(5)	104.484(4)
γ/°		92.660(4)
V/Å3	2383.9(2)	1998.83(19)	2059.00(10)
Z	4	2	8
T/K	100(2)	100(2)	100(2)
ρcalcd/g·cm−3	1.436	1.383	1.331
µ/mm−1	0.763	0.503	0.093
F(000)	1084	872	880
2θmax/◦	55	55	55
No. of reflections measured	20466	26576	12996
No. of independent reflections	5452 (Rint = 0.0866)	9145 (Rint = 0.0482)	4588 (Rint = 0.0339)
Data/restraints/parameters	5452/0/319	9145/83/552	4588/4/289
R1 [I > 2.00 σ(I)] i	0.0621	0.0852	0.0376
wR2 (all reflections) ii	0.1230	0.2784	0.0850
Goodness of fit indicator	1.043	1.040	1.030
Highest peak, deepest hole/e Å−3	0.437, −0.575	1.432, −0.799	0.181, −0.214
CCDC deposition number	2216987	2216988	2216989

ⁱ R₁ = Σ||Fo| − |Fc||/Σ|Fo|, ⁱⁱ wR₂ = [Σ(w(Fo² − Fc²)²)/Σw(Fo²)²]^1/2.

. Single-crystal X-ray diffraction data were obtained with a Rigaku XtaLAB AFC11 diffractometer with graphite-monochromated Mo Kα radiation (λ = 0.71073 Å). A single crystal was mounted with a glass capillary and flash-cooled with a cold N₂ gas stream. Data were processed using the CrysAlisPro software packages. The structures were solved by intrinsic phasing methods using the SHELXT [28] software packages, and refined on F₂ (with all independent reflections) using the SHELXL [29] software packages. The non-hydrogen atoms were refined anisotropically. In 2BF₄·1.5C₂H₅OH, a BF₄⁻ anion and ethanol molecules of crystallization were disordered at two possible positions. The rigid-bond restraint (RIGU) and SIMU command were employed for these disordered atoms except for one of two parts of BF₄⁻ anion (B1, F1, F2, F3, and F4). All B-F bonds were restrained to the same distances by the SADI command. The C-O bonds for ethanol molecules and B-F bonds for the minor part of BF₄⁻ (B2, F5, F6, F7, and F8) were constrained by the DFIX command. The DANG command was employed to restrain C-C-O angles of the ethanol molecules.

4.4. Magnetic Measurements

Magnetic susceptibility measurements were performed with a MPMS-7 or MPMS-XL7 SQUID magnetometer. Susceptibility data were recorded in the temperature range from 1.9 to 300 K with a static field of 5.0 kOe. The polycrystalline samples were grounded into fine powders by an agate mortar in dried condition. The samples were loaded into a gelatin capsule and covered in liquid paraffin to prevent field-induced orientation of crystals. All data were corrected for diamagnetism of the sample by means of Pascal’s constants [30]. Temperature dependence of the magnetic susceptibility and field dependence of the magnetization data were fitted using the MagSakiTetra W0913 program [31]. The dynamic susceptibility was measured with alternating-current (ac) fields of 3 Oe magnitude and a constant direct current (dc) field of 0–4.0 kOe in the frequency range from 1 to 1500 Hz. The relaxation time τ was extracted from fitting to the generalized Debye model. The fit was performed using the CC-Fit program [32].

5. Conclusions

In this study, tetra-coordinated and a hexa-coordinated Co^II complexes, containing the same ligand, were prepared and characterized by X-ray crystallography and magnetometry. In the crystal, the tetracoordinated complex 1·C₂H₅OH possessed pseudo-tetrahedral coordination geometry and formed one-dimensional chain networks by intermolecular hydrogen-bonding interactions. Two intrachain Co···Co distances were slightly different (ca. 0.1 Å) and were short enough to induce a weak intermolecular magnetic coupling. Zero-field SIM behavior was observed for 1·C₂H₅OH because QTM was partially suppressed by the non-zero intermolecular magnetic coupling. The partial suppression of QTM indicated that the difference in intrachain Co···Co distances is important to suppress QTM completely. On the other hand, the hexacoordinated complex 2BF₄·1.5C₂H₅OH exhibited a severely distorted octahedral coordination geometry with one-dimensional networks by N-H···π and hydrogen-bonding interactions. The intermolecular Co···Co distances were above the range of intermolecular magnetic coupling, and SIM behavior was observed only in the presence of an external field. Thus, these results indicated that not only non-zero magnetic coupling, but also the difference of intrachain distances are important to achieve zero-field SIM.