Binuclear Nickel Complexes of a New Di(hydroxyphenyl)imidazolate

Abstract

Here, we report a new di(hydroxyphenyl)imidazolate ligand with an N₂O₂ donor set synthesized by a modified Debus–Radziszewski procedure. Its binuclear nickel(II) complexes feature a weak antiferromagnetic interaction with J₁₂ = −3.16 cm⁻¹ between the two nickel(II) ions identified by magnetometry measurements. As follows from cyclic voltammetry experiments and DFT calculations, they undergo ligand-centered oxidation via the formation of cation radicals with short lifetimes that can be potentially stabilized by bulkier t-butyl groups in the ortho-positions of the ligand. The reported ligand widens the range of the building blocks available to molecular magnetism community and thus provides new ways to the design of magnetic materials with switchable properties.

1. Introduction

Switchable molecular compounds that experience an intramolecular electron transfer if triggered by an external stimulus (temperature, pressure or light irradiation) are sought for their potential use in spintronics, sensorics and multifunctional devices [1,2]. This ability is common in mixed-valence systems [3,4] that contain at least two redox sites in different oxidation states and a bridging ligand to mediate metal–metal interactions. A redox active ligand, e.g., dioxolene [5,6], is often chosen as a second active site to promote metal–ligand electron-transfer. One-electron oxidation of some dioxolene complexes results in temperature- [7] or ligand geometry-dependent [8] products that feature valence tautomerism between a high-valent metal–dioxolene and a metal-semiquinone radical state [3,7,9]. While most mononuclear dioxolene complexes have only one characteristic oxidation state [9], salen-type complexes with N₂O₂ donor–ligands are prone to valence tautomerism [10,11]. Examples of di- or trinucleating phenolate-containing ligands are rare [12].

Metal complexes of stable phenoxyl radicals are targeted as catalysts that mimic galactose oxidase and glyoxal oxidase [13,14], a feature that boosted the interest in coordination compounds of phenoxyl radicals. Mixed-valence polymetallic complexes containing such radicals may be of use in information storage and integrated molecular-sized devices owing to their ability to reversibly change the electronic structure of a molecule [9].

Much of the recent efforts have been invested into understanding the influence of the bridging moiety on the rate of the electron transfer process [15]. Nitrogen-containing heterocyclic ligands are known to facilitate magnetic interactions between the metal ions through their π-system [15], as in ruthenium complexes of doubly chelating bridging ligands, 2,2 -bipyrimidine [16], 2,2-azobis(pyridine) [17], 1,2,5-oxadiazole [18] and imidazole [19].

Here, we synthesized a new di(hydroxyphenyl)imidazolate ligand with a N₂O₂ donor set, L, which contains an extended π-conjugated system and a bridging imidazole moiety (Scheme 1). A comprehensive study of its two binuclear nickel complexes [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl by dc-magnetometry, electrochemistry and DFT calculations showed them to feature a weak antiferromagnetic interaction between the two metal ions and to easily undergo two-step oxidation with an intermediate formation of a phenoxyl radical.

2. Results and Discussion

2.1. Synthesis

The target ligand L was obtained in a high yield by a four-step synthetic procedure. The first step is the acylation of 4-ethylphenol by oxalyl chloride in the presence of aluminum chloride (Scheme 2, see Figure S1 for NMR spectra). Attempts to create an imidazole moiety directly from the obtained diketone were unsuccessful due to a side condensation of phenol moiety and formaldehyde [20]. To avoid the latter from occurring, protecting methyl groups were introduced into the diketone by reacting it with dimethyl sulphate (Figure S2). A subsequent reaction with urotropine as a source of formaldehyde and ammonia acetate in a closed vessel at the boiling point of the solvent (acetic acid) produced methylated 4,5-di(2-hydroxyphenyl)imidazolate with the conversion of ca. 80% and no traces of side products (Figure S3). Note that a multicomponent Debus–Radziszewski reaction rarely proceeds with such a good conversion and a high yield [19,21]. When performing the reaction under reflux or in a closed vessel under the microwave radiation, the conversion becomes below 30%, and many by-products appear in the reaction mixture. A possible reason behind it is that the conditions of the conventional reflux method are too mild to give a high yield, and the microwave radiation causes a product to decompose.

To produce the ligand L, the protective methyl groups were removed by reacting methylated 4,5-di(2-hydroxyphenyl)imidazolate with boron tribromide. The formation of the target product was confirmed by NMR spectroscopy; its ¹H NMR spectrum feature a singlet at 7.68 ppm, which is characteristic of the CH proton of the imidazole moiety (Figure S4), and no signals of the methoxy groups, which confirms the deprotection.

Binuclear Ni²⁺ complexes [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl (Scheme 1) were synthesized by the template reaction of the obtained ligand L and complexes bipy₂NiCl₂ and phen₂NiCl₂, respectively, in the presence of DBU as a noncoordinating base in methanol. Attempts to use trimethylamine for this purpose failed due to the formation of trimethylamine metal complexes [22]. The successful formation of the target Ni²⁺ complexes was confirmed by elemental analysis, NMR spectroscopy (although a full assignment of signals in the ¹H NMR spectrum was complicated by the paramagnetic nature of the complexes; Figures S5 and S6) and X-ray diffraction. [(bipy)₄Ni₂(L)]Cl is soluble in methanol and DMF; [(phen)₄Ni₂(L)]Cl is slightly soluble only in CH₃OH.

The choice of other metal salts as a source of metal ion M (M = Fe³⁺, Cu²⁺ or Mn²⁺) produced, in the same reaction conditions, the unwanted homoleptic complex [(bipy)₃M]ⁿ⁺ as the main product, which apparently resulted from the intramolecular disproportionation reaction occurring in the binuclear complexes. Using a Co²⁺ salt produced inseparable mixtures of dia- and paramagnetic complexes of an unknown composition.

2.2. X-ray Crystallography

The formation of the target complexes [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl was confirmed by X-ray diffraction analysis (Figure 1) of their solvates [(bipy)₄Ni₂(L)]Cl•2CH₃OH•0.75Et₂O and [(phen)₄Ni₂(L)]Cl•3CH₃OH obtained by recrystallization from methanol and diethyl ether (1/1). They crystallize in the triclinic space group P-1, thereby adopting a rac form (no meso form was isolated upon fractional crystallization [19]).

In the two complexes, the imidazole moiety and both the hydroxyl groups of the ligand L are deprotonated; the corresponding C-O bond lengths (C-O 1.330(6)–1.340(6) Å) fall between the values typical for the double and single C–O bond (1.24 and 1.38–1.42 Å) [23]. Each of the two Ni²⁺ ions coordinates four nitrogen atoms from the two bipyridine or phenanthroline ligands (Ni–N 2.061(5)–2.146(6) Å) and two heteroatoms of the ligand L, the nitrogen atom of its imidazole moiety and the oxygen atom of the C-O group (Ni–N 2.033(5)–2.050(4) Å, Ni–O 2.014(4)–2.050(4) Å), to produce a nearly perfect octahedral environment, as gauged by shape measures [24] (

Table 1. Main geometrical parameters * of [(bipy)₄Ni₂(L)]Cl•CH₃OH•0.75Et₂O and [(phen)₄Ni₂(L)]Cl •3CH₃OH.

Parameters	[(bipy)4Ni2(L)]Cl•2CH3OH•0.75Et2O		[(phen)4Ni2(L)]Cl•3CH3OH
	Ni(1)	Ni(2)	Ni(1)	Ni(2)
Ni–NPy, Å	2.074(5)–2.132(2)	2.061(5)–2.095(5)	-	-
Ni–NPhen, Å	-	-	2.076(5)–2.146(6)	2.080(5)–2.135(6)
Ni–NL, Å	2.040(4)	2.050(4)	2.036(6)	2.033(5)
Ni–OL, Å	2.014(4)	2.050(4)	2.028(4)	2.036(5)
S(TP-6)	12.997	15.215	13.850	12.366
S(OC-6)	0.958	0.733	0.658	0.885

* The N_Py (N_Phen), N_L and O_L are the nitrogen atoms of bipyridine (phenanthroline) ligands and the nitrogen and oxygen atoms of the ligand L, respectively. S(TP-6) and S(OC-6) are trigonal-prismatic and octahedral shape measures, respectively.

). These measures quantify how far the shape of the coordination polyhedron is from an ideal shape, such as an ideal octahedron (OC-6). For [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl, the octahedral shape measure S(OC-6) estimated from the X-ray diffraction data for these complexes [24] is as low as 0.733 and 0.658, respectively. For comparison, the shape measure of the NiN₄O₂ coordination polyhedron respective to another ideal shape with six vertices, the trigonal prism (TR-6), adopts much higher values of 12.366–15.215.

In the crystal of the solvated complex [(bipy)₄Ni₂(L)]Cl, both the C-O bonds of the ligand L form hydrogen bonds with the lattice methanol (O…O 2.621(6) and 2.569(5) Å, OHO 162.5(4) and 171.4(3)°). In the case of [(phen)₄Ni₂(L)]Cl, only one of them is involved in such a bond (O…O 2.695(7) Å, OHO 162.7(3)°); the two other solvent molecules of methanol are hydrogen-bonded to the first methanol molecule (O…O 2.721(7) Å, OHO 161.3(4)°) and to the chloride anion (O…Cl 3.065(7) Å, OHCl 175.5(3)°). In [(bipy)₄Ni₂(L)]Cl, the chloride anion forms only weak C-H…Cl contacts.

2.3. Magnetic Susceptibility Measurements

To probe if the extended π-conjugated system of the ligand L mediates magnetic interactions between the two Ni²⁺ ions in the obtained complexes, magnetic properties of a finely ground powder of desolvated [(bipy)₄Ni₂(L)]Cl were accessed in the solid state by the dc-magnetometry (Figure 2, left); a solid sample of [(phen)₄Ni₂(L)]Cl suffered amorphization upon grinding to avoid magnetic torquing under the applied magnetic field [25] and therefore was not probed by this method. For [(bipy)₄Ni₂(L)]Cl, the behavior of the magnetic susceptibility in the temperature range of 2–300 K is consistent with a weak antiferromagnetic interaction between the two Ni²⁺ ions. At the highest accessible temperature of 300 K, the χ_mT product adopts the value of 2.21 cm³K/mol, which is close to the spin-only value (2.0 cm³K/mol) for two independent Ni²⁺ ions with no significant zero-field splitting or exchange effects. Upon cooling to 100 K, the magnetic susceptibility gradually decreases to a nearly constant value of 2.2 cm³K/mol. The drop below 100 K is a sign of antiferromagnetic exchange interactions between the two Ni²⁺ ions (Figure 2, left). The best fit of the resulting temperature-dependence of the χ_mT product with the isotropic g_iso value of the g-tensor [26] to avoid overparameterization was obtained with g_iso of 2.09 and the exchange constant J₁₂ of −3.16 cm⁻¹; a weak antiferromagnetic interaction with J₁₂ of −2.15 to −1.43 cm⁻¹ also follows from DFT calculations (see below).

The magnetic susceptibility for both the complexes, [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl, was also measured in their methanol solutions (Figure 2, right) by the Evans technique [27] of the NMR spectroscopy often used for this purpose [28]. For [(bipy)₄Ni₂(L)]Cl, the χ_mT product is constant at ca. 2.5 cm³K/mol over the entire temperature range of 200–330 K accessible in methanol. The observed value is higher than the spin-only value for two independent Ni²⁺ ions; however, with a claimed accuracy of the Evans technique of ~11% [29], it is consistent with dc-magnetometry data.

2.4. Redox Properties

To check for the possible formation of mixed-valence species, redox properties of the ligand L and its complexes [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl were probed by cyclic voltammetry (Figure 3). In the corresponding measurements, two cycles were used to highlight the stability of possible intermediates. For L, an irreversible one-electron oxidation wave with the highest oxidation potential at Ep = 0.378 V vs. Fc/Fc⁺ indicates a low stability of the phenoxyl radical produced upon the ligand oxidation [30]. The cyclic voltammograms of the Ni²⁺ complexes feature two reversible oxidation waves at E¹_1/2 = −0.194 V and E²_1/2 = 0.120 V for [(bipy)₄Ni₂(L)]Cl, hinting on the formation of two phenoxyl radicals on both sides of a binuclear complex (Scheme 3), as gauged by higher potentials of the redox couple Ni²⁺/Ni³⁺ [31]. An increase in the degree of reversibility of the oxidation process as compared to the parent ligand L is a sign of the stabilization of the phenoxyl radicals and of a substantial contribution of the orbitals of the metal ions to the overall stability of the complex, which increases the lifetime of the radical [30]. It is, however, a common wisdom that bulky groups (especially those located in ortho- and para-positions of the phenol moiety [12,30]) stabilize the radicals, as is, probably, the case of the complex [(bipy)₄Ni₂(L)]Cl with bulky para-ethyl groups in the ligand L and the bipy ligands.

The same stabilization occurs for [(phen)₄Ni₂(L)]Cl, as two reversible oxidation waves are observed at even lower values of E¹_1/2 = −0.270 V and E²_1/2 = 0.116 V. Unfortunately, all our attempts to chemically oxidize the complexes [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl with Ag(BF₄), Br₂ or Fc[PF₆] resulted in mixtures of paramagnetic products, the major of them being the initial complex (bipy)₂NiCl₂ as identified by NMR spectroscopy. This failure can be explained by the formation of a ligand radical with a short lifetime due to the absence of t-butyl groups in its ortho-positions, the common stabilizing substituents that protect phenoxy radicals from bis(μ-phenolate) dimerization [32].

2.5. DFT Calculations

To confirm the ligand-centered nature of the oxidation of the complexes [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl that proceeds via the formation of unstable cation radicals, DFT calculations with the popular hybrid functional B3LYP [33] were performed for the complex cation [(bipy)₄Ni₂(L)]⁺ and the product of its oxidation, [(bipy)₄Ni₂(L)]²⁺. The chosen functional is the best for describing one-electron-oxidized nickel(II) complexes with the phenol-containing ligands, such as a salen-type complex featuring di-t-butyl groups in its ligand [34].

The frontier computed MOs of the cation [(bipy)₄Ni₂(L)]⁺ are of a π-character; α-HOMO and β-HOMO (Figure 4) are localized on the imidazole and ethylphenol moieties and the oxygen atoms of the phenolate moieties of the ligand L with only a small contribution from the d-orbitals of the Ni²⁺ ions, thereby supporting the above conclusion on the ligand-centered nature of the oxidation of the complexes [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl. In contrast, α-LUMO and β-LUMO are localized on the bipyridine ligands (Figure S8). The spin density is distributed over two metal Ni²⁺ ions, residing on the nitrogen and oxygen atoms of the ligands and both the metal ions (Figure 4). The exchange constant J₁₂ calculated with the broken symmetry approach, which allows extracting a phenomenological constant in Heisenberg–Dirac–van Vleck Hamiltonian [35,36,37], is in the range of −2.15 to −1.43 cm⁻¹.

The calculated energy level diagrams for [(bipy)₄Ni₂(L)]⁺ and [(bipy)₄Ni₂(L)]²⁺ (Figure 5) show that β-MOs in the nonoxidized species lie lower than α-MOs. The same pattern is observed for the oxidized species. In this case, however, the energy difference between α- and β-MOs is larger, e.g., |ΔE| for α-273 and β-268 is 0.0096 a.u. The MOs of [(bipy)₄Ni₂(L)]²⁺ are lower in energy than those in [(bipy)₄Ni₂(L)]⁺, with |ΔE^HOMO| equal to 0.032 a.u.

For the oxidized species [(bipy)₄Ni₂(L)]²⁺, the frontier α-orbitals are very similar to those in the complex cation [(bipy)₄Ni₂(L)]⁺ (Figure 4, Figure 6, Figures S8 and S9); however, the spin density distribution is quite different (Figure 5). In [(bipy)₄Ni₂(L)]²⁺, it has an out-of-plane π-character and is located on the carbon atoms of the phenyl rings, the oxygen atoms of the phenolate moieties of the ligand L and the nitrogen atoms of the bipy ligands (Figure 6). Some of the spin density is also located on the two nickel ions. The same trend is observed for [(phen)₄Ni₂(L)]²⁺ (see Figures S10–S12). These results confirms that upon one-electron oxidation, the electron leaves the ligand site rather than a Ni²⁺ ion, thereby explaining a short lifetime of the oxidized species and the difficulties that we experienced while attempting to chemically oxidize the complexes [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl.

3. Materials and Methods

Synthesis. All synthetic manipulations were carried out on air unless stated otherwise. Solvents were purchased from commercial sources and purified by distilling from conventional drying agents under an argon atmosphere prior to use. 2,2′-Bipyridine and 1,10-phenanthroline were obtained from Aldrich, and (bipy)₂NiCl₂ and (phen)₂NiCl₂ were synthesized by the reported procedure [38].

1,2-bis(5-ethyl-2-hydroxyphenyl)ethane-1,2-dione. 4-Ethylphenol (25.65 g, 0.21 mol) was dissolved in 50 mL of methylene chloride, and the solution was added dropwise to a suspension of aluminum chloride (56 g, 0,42 mol) in 300 mL of dichloromethane. The resulting mixture was stirred for 30 min. At some point, a solution of oxalyl chloride (8.6 mL, 0.1 mol) in 50 mL of dichloromethane was added dropwise to produce a dark-colored solution that was then stirred at r.t. for 10 h. A 12 M HCl solution was carefully added while cooling with an ice bath to quench the reaction mixture and thereby avoid its overheating. The resulting suspension was filtered, and the filtrate was washed with dichloromethane. The mother liquor that contained both the organic and aqueous phases was extracted with dichloromethane, dried over magnesium sulfate (MgSO₄), filtered and evaporated. The dry residue was recrystallized from ethanol at −10 °C. Yield: 59% (36.96 g). ¹H NMR (CDCl₃, 400 MHz): δ (ppm) = 11.21 (s, 2H, OH), 7.45 (dd, 2H, ²J_HH = 8.7 Hz, Ph), 7.24 (d, 2H, ²J_HH = 2.2 Hz, Ph), 7.04 (d, 2H, ²J_HH = 8.6 Hz, Ph,), 2.54 (q, 2H, ²J_HH = 7.6 Hz, 2CH₂), 1.15 (t, 6H, ²J_HH = 7.6 Hz, 2CH₃,). ¹³C NMR (CDCl₃, 101 MHz): δ (ppm) = 196.49, 161.84, 138.54, 135.70, 130.69, 118.69, 27.74, 15.58.

1,2-bis(5-ethyl-2-methoxyphenyl)ethane-1,2-dione. In a 250 mL round-bottom flask, a solution of 1,2-bis(5-ethyl-2-hydroxyphenyl)ethane-1,2-dione (5.97 g, 0.02 mol) in THF and aqueous lithium hydroxide LiOH∙H₂O (1.84 g, 0.04 mol) was stirred at r.t. for one hour. Dimethyl sulfate (3.13 mL, 0.04 mol) was added dropwise, and the reaction mixture was kept under stirring for 24 h. The solvent was distilled off, and the solid residue was dissolved in 2 M solution of sodium hydroxide, extracted with dichloromethane, dried over magnesium sulfate and filtered off. The solvent was evaporated, and the residue was dried under high vacuum. Yield: 82% (5.35 g). ¹H NMR (CDCl₃, 400 MHz): δ (ppm) = 7.96 (d, 2H, ²J_HH = 2.4 Hz, Ph), 7.45 (dd, 2H, ²J_HH = 8.5 Hz, Ph), 6.94 (d, 2H, ²J_HH = 8.5 Hz, Ph), 3.63 (s, 6H, 2CH₃), 2.72 (c, 4H, ²J_HH = 7.6 Hz, 2CH₂), 1.31 (t, 6H, ²J_HH = 7.6 Hz, 2CH₃). Calcd for C₂₀H₂₂O₄ %: C 73.60%, H 6.79%. Found, %: C 73.58%, H 6.82%.

4,5-bis(5-ethyl-2-methoxyphenyl)-1H-imidazole. In a 25 mL thick-walled Schlenk flask with a Teflon screw stopper and a magnetic stirrer, 1,2-bis(5-ethyl-2-methoxyphenyl)ethane-1,2-dione (1.68 g, 0.005 mol) was added to 20 mL of acetic acid together with urotropine (1.59 g, 0.011 mol) and ammonium acetate (7.95 g, 0.1 mol). The mixture was stirred for 8 h at 120 °C and then poured into a 30% aqueous ammonia solution, extracted with ethyl acetate and dried over magnesium sulfate; the solvent was distilled off. The product was purified by flash chromatography; the starting compound was first eluted with ethyl acetate, and the final product with ethyl acetate-methanol (1:5). Yield: 72% (1.21 g). ¹H NMR (CDCl₃, 400 MHz): δ (ppm) = 7.68 (s, 1H, CH=N), 7.17 (d, 2H, ²J_HH = 2.2 Hz, Ph), 7.04 (d.d, 2H, ²J_HH = 8.4 Hz, Ph), 6.82 (d, 2H, ²J_HH = 8.3 Hz, Ph), 3.67 (s, 6H, 2CH₃), 2.46 (q, 4H, ²J_HH = 7.6 Hz, 2CH₂), 1.07 (t, 6H, ²J_HH = 7.6 Hz, 2CH₃). ¹³C NMR (CDCl₃, 151 MHz): δ (ppm) = 154.05, 136.32, 133.68, 129.67, 127.58, 110.92, 55.51, 27.87, 15.71. Calcd for C₂₁H₂₄N₂O₂ %: C 74.97%, H 7.19%, N 8.33%. Found, %: C 74.91%, H 7.22%, N 8.27%.

2,2′-(1H-imidazole-4,5-diyl)bis(4-ethylphenol) (L). In a Schlenk flask (50 mL) under argon, 4,5-bis(5-ethyl-2-methoxyphenyl)-1H-imidazole (1.01 g, 0.003 mol) was dissolved in 15 mL of dry dichloromethane. The solution was cooled to −10 °C, and a solution of boron tribromide (1.13 mL, 0.012 mol) in 15 mL of dry dichloromethane was added dropwise. The resulting mixture was warmed to r.t. and stirred for 24 h at this temperature. Then, ethanol (2.7 mL) was carefully added to the mixture at −10 °C that was stirred for another 2 h. The precipitate was filtered off, washed with dichloromethane, added to a solution of Na₂CO₃, and then with distilled water. The product was dried under high vacuum. Yield: 59% (0.55 g). ¹H NMR (DMSO-d6, 400 MHz): δ (ppm) = 10.01 (s, 2H, OH, NH), 9.20 (s, 1H, CH=N), 7.15–7.02 (m, 2H, Ph), 6.93 (dd, 4H, ²J_HH = 8.3 Hz, Ph), 2.37 (q, 4H, ²J_HH = 7.6 Hz, CH₂), 0.97 (t, 6H, ²J_HH = 7.5 Hz, CH₃). ¹³C NMR (DMSO-d6, 101 MHz): δ (ppm) = 153.71, 134.49, 130.69, 129.92, 126.39, 116.26, 27.43, 16.03. Calcd for C₁₉H₂₀N₂O₂ %: C 74.00%, H 6.54%, N 9.08%. Found, %: C 73.96%, H 6.57%, N 9.05%.

[(bipy)₄Ni₂(L)]Cl. The ligand L (100 mg, 0.32 mmol) was dissolved with DBU (0.145 mL, 0.97 mmol) in 15 mL of methanol and stirred for 10 min. (bipy)₂NiCl₂ (322 mg, 0.68 mmol) in 10 mL of methanol was quickly added to this solution. The resulting mixture was stirred for 3 h at r.t., evaporated under reduced pressure and recrystallized from a mixture of methanol/diethyl ether (1/1). Yield: 121 mg (35%). ¹H NMR (CD₃OD; 300 MHz): δ (ppm) = 134.5–120.3 (br. m), 63.47–59.44 (br. m), 56.76 (br. s), 46.94 (br. s), 43.80–41.33 (br. m), 40.05 (br. s), 39.46 (br. s), 38.03 (br. s), 14.75 (br. s), 13.66–13.46 (br. m),12.46 (br. s), 11.83–10.99 (br. m), 9.63 (br. s), −2.4 (br. s), −5.2 (br. s). UV-vis in CH₃OH: λ(max) = 340 nm. Calcd for (C₅₉H₅₁ClN₁₀Ni₂O₂•2CH₃OH•0.75C₄H₁₀O): C 63.92%; H 5.41%; N 11.65%. Found: C 63.49%; H 5.03%; N 11.32%.

[(phen)₄Ni₂(L)]Cl. The ligand L (197 mg, 0.64 mmol) was dissolved with DBU (0.29 mL, 1.91 mmol) in 20 mL of methanol and stirred for 10 min. (phen)₂NiCl₂ (626 mg, 1.28 mmol) in 20 mL of methanol was quickly added to this solution. The resulting mixture was stirred for 3 h at r.t., and then the product was filtered off and recrystallized from methanol at –10 °C. Yield: 421 mg (55%). ¹H NMR (CD₃OD; 300 MHz): δ (ppm) = 133.5–122.3 (br. m), 45.82–44.04 (br. m), 42.20 (br. s), 41.22 (br. s), 23.84 (br. s), 23.06–22.26 (br. m), 21, 49 (br. s), 16.25–15.28 (br. m), 15.45 (br. s), 15.10 (br. s), 14.60 (br. s), 11.77 (br. s), 11.28 (br. s), 10.46–9.93 (br. m), −3.31 (br. m), −5.73 (br. m). UV-vis in CH₃OH: λ(max) = 366 nm. Calcd for (C₆₇H₅₁ClN₁₀Ni₂O₂•3CH₃OH): C 65.93%; H 4.82%; N 10.98%. Found: C 65.64%; H 4.62%; N 10.65%.

Analytical and Spectroscopic Measurements. ¹H and ¹³C NMR spectra were recorded at 298 K with Bruker Avance 300 and 600 FT-spectrometers (300.15 and 600.22 MHz ¹H frequency, respectively). UV-vis spectra were recorded in methanol at room temperature with a Shimadzu UV-2600i double-beam spectrophotometer in the wavelength range from 240 to 1400 nm. The amount of carbon, nitrogen and hydrogen contained was estimated on a Carlo Erba analyzer, model 1106. Cyclic voltammograms (CV) were recorded at r.t. in the nitrogen atmosphere on an Autolab PGSTAT128N potentiostat-galvanostat in a standard three-electrode cell using a 3 mm thick glassy carbon working electrode, a platinum wire counter electrode and an Ag/AgCl reference electrode in a saturated KCl solution. The required conductivity of the solution was achieved by adding a TBAPF₆ to the test solution to the concentration of 0.1 M. All CV curves were measured at a potential sweep rate of 100 mV/s.

X-ray crystallography. X-ray diffraction data for the single crystals of [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl, which were grown by slow vapor diffusion of diethyl ether into a methanol solution, were collected with Bruker APEXII DUO and Quest D8 CMOS diffractometers, using graphite monochromated Mo-Kα radiation (λ = 0.71073 Å). Using Olex2 [39], the structures were solved with the ShelXT [40] structure solution program using Intrinsic Phasing and refined with the XL [41] refinement package using least-squares minimization against F² in anisotropic approximation for nonhydrogen atoms. Hydrogen atoms of methanol molecules in [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl were located via difference Fourier synthesis, while the positions of other hydrogen atoms were calculated, and they all were refined in isotropic approximation within the riding model. Crystal data and structure refinement parameters are given in

Table 2. Crystal data and structure refinement parameters for [(bipy)₄Ni₂(L)]Cl•2CH₃OH•0.75Et₂O and [(phen)₄Ni₂(L)]Cl•3CH₃OH.

	[(bipy)4Ni2(L)]Cl•2CH3OH•0.75Et2O	[(phen)4Ni2(L)]Cl•3CH3OH
Empirical formula	C64H64.5ClN10Ni2O4.75	C70H61ClN10Ni2O5
Formula weight	1202.62	1275.15
T, K	120	100
Crystal system	triclinic	triclinic
Space group	P-1	P-1
Z	2	2
a, Å	13.555(2)	12.2327(3)
b, Å	15.389(3)	13.0491(3)
c, Å	17.007(3)	20.7742(5)
α, °	77.732(3)	72.9350(10)
β, °	70.231(3)	77.2100(10)
γ, °	81.830(3)	89.9820(10)
V, Å3	3253.0(9)	3084.00(13)
Dcalc, (g cm−1)	1.228	1.373
μ, cm−1	6.73	7.15
F(000)	1259	1328
2θmax, °	54	50
Reflections measured	45,304	41,247
Independent reflections	14,150	10,859
Observed reflections [I > 2σ(I)]	6921	5967
Parameters	756	811
R1	0.0794	0.0811
wR2	0.2188	0.2702
GOF	0.955	1.100
Δρmax/Δρmin (e Å−3)	0.955/−0.694	2.109/−0.415

. CCDC 2,207,058 and 2,207,059 contain the supplementary crystallographic data for [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl, respectively.

Magnetochemistry. Magnetic susceptibility of [(bipy)₄Ni₂(L)]Cl was measured in the temperature range 2–300 K with a Quantum Design PPMS-9 device under the dc-magnetic field of 5 kOe. A finely ground and dried microcrystalline powder (23 mg) was immobilized in mineral oil inside a polyethylene capsule. Sweep rate was 74 s/K below 100 K and 54.6 s/K above this temperature. The data were corrected for the sample holder, the mineral oil and the diamagnetic contribution, the latter by using the Pascal constants. Smoothing was performed by the Savitzky–Golay method with five points of window and polynomial order of two. To avoid overparameterization, the observed variable-temperature dc-magnetic susceptibility was fitted in PHI software [19] using the spin Hamiltonian:

$$\begin{array}{l}\widehat{H}={\mu }_{\beta }{g}_{iso}(B_x{\widehat{S}}_{1x}+B_y{\widehat{S}}_{1y}+B_z{\widehat{S}}_{1z})\\ +{\mu }_{\beta }{g}_{iso}\left(B_x{\widehat{S}}_{2x}+B_y{\widehat{S}}_{2y}+B_z{\widehat{S}}_{2z}\right)+J_{12}{\widehat{S}}_1{\widehat{S}}_2\end{array}$$

(1)

where S = 1 is the spin of the Ni²⁺ ion, g_iso is the isotropic value of the g-tensor, μ_B is the Bohr magneton and J₁₂ is the exchange constant.

Evans method. The temperature dependence of the molar magnetic susceptibility (χ_M) of the complexes [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl in methanol-d4 was evaluated in the Evans experiment. The appropriate ¹H NMR spectra were recorded in the temperature a range of 200–330 K using an NMR tube with a coaxial inset. The inner (control) tube contained ~1% Me₄Si in methanol-d4, and the outer tube, a solution of the complex (~1–5 mg/cm³) in the same solvent with the same concentration of Me₄Si. The value of χ_M was calculated from the difference between the chemical shifts of Me₄Si in pure dichloromethane-d2 and in a solution of the complex (Δδ, Hz) using the following equation:

$${\chi }_M=\frac{\Delta \delta M}{{\upsilon }_0{S}_{f}c}-{\chi }_M^{dia}$$

(2)

M is the molar mass of the Ni²⁺ complex, g/mol; ν₀ is the spectrometer frequency, Hz; S_f is the coefficient of the magnet shape (4π/3); c is the concentration of the complex, g/cm³; ${\chi }_M^{\mathit{dia}}$ is the molar diamagnetic contribution to the paramagnetic susceptibility calculated using Pascal’s constants. The concentration (c) was recalculated for each temperature according to the changes in the solvent density (ρ): c_T = m_cρ/m_sol, where mc is the weight of the complex, and msol is the weight of the solution.

DFT calculations. All calculations were performed with the ORCA 5.0.3 program package [42]. Geometry of the complex cations [(bipy)₄Ni₂(L)]⁺ and [(phen)₄Ni₂(L)]⁺ and the products of their oxidation, [(bipy)₄Ni₂(L)]²⁺ and [(phen)₄Ni₂(L)]²⁺, was optimized using the B3LYP functional [33] with RIJCOSX approximation for Coulomb and HF integrals and def2-tzvp basis set. Solvation effects were accounted with CPCM continuum solvation model with dichloromethane as a solvent. The exchange constant for [(bipy)₄Ni₂(L)]⁺ was estimated by the broken symmetry approach as implemented in the ORCA 5.0.3 program package [43].

4. Conclusions

A modified Debus–Radziszewski procedure allowed us to produce a new ligand L with a N₂O₂ donor set that contains two phenolic and bridging imidazole moieties to provide an extended π-conjugated ligand framework. In the two binuclear complexes, [(bipy)₄Ni₂(L)]Cl and [(phen)₄Ni₂(L)]Cl, it mediates a weak antiferromagnetic interactions between the two Ni²⁺ ions with J = −2.15 cm⁻¹, as identified by variable-temperature dc-magnetometry. Attempts to introduce other metal ions, such as iron(III), copper(II), manganese(II) or cobalt(II), failed due to the disproportionation reaction of the hypothetical heteroleptic complexes.

One-electron electrochemical oxidation of both the ligand and the binuclear complexes produces unstable cation radicals with the lifetimes that can be potentially improved by the use of bulkier t-butyl groups in the ortho-positions of the ligand, often used to stabilize phenoxy radicals and prevent them from bis(μ-phenolate) dimerization [32]. As follows from quantum chemical calculations, the redox process is indeed centered on the ligand, with the spin density in the resulting cation-radical [(bipy)₄Ni₂(L)]²⁺ mostly localized on the phenolic and imidazole moieties of the ligand.

To produce mixed-valence complexes from the synthesized di(hydroxyphenyl)imidazolate L or similar ligands, other metal ions may be introduced, such as platinum or palladium, given the limited scope of substrates that we discovered. As an alternative pathway toward such complexes that may find use in molecular-sized devices [9], the bipyridine and phenanthroline co-ligands may be substituted for less sterically hindered ligands, such as ethylenediamine, to obtain binuclear metal complexes from a di(hydroxyphenyl)imidazolate ligand with t-butyl groups; otherwise, a steric bulk of these co-ligands does not allow such a complex to form.