Synthesis, structure, magnetic and catalytic behavior of a dinuclear copper(II) complex with triazendio ligands

Abstract

A dinuclear copper(II) complex [Cu₂L₄] has been prepared by the reaction of CuCl₂·2H₂O and 1-[(2-iodo)benzene]-3-[benzothiazole] triazene (HL). The complex has been characterized by X-ray crystallography and by physico-chemical and spectroscopic methods. In the solid state, there is a significant antiferromagnetic coupling between the copper(II) centers with a coupling constant (J) of − 558 cm⁻¹. In homogeneous solution, the complex shows electrocatalytic activities for hydrogen generation from both acetic acid and neutral buffer with a turnover frequency of 50 mol of H₂ per mole of catalyst per hour (mol H₂/mol catalyst/h) at an overpotential (OP) of 941.6 mV, and 502 mol H₂/mol catalyst/h at an OP of 836.7 mV.

Introduction

Much current research on transition metal complexes is focussed on their potential uses as catalysts, biological agents and inorganic materials [1,2,3]. By designing appropriate ligands, the electronic properties of the metal center can be tuned to achieve specific applications [4]. Our interests have focused on triazenide ligands, which serve as bridging ligands for the assembly of binuclear and multinuclear complexes with a range of functions, for example, magnetic coupling, catalytic and luminescent performance, etc. [5,6,7,8,9,10,11]. In particular, hydrogenase enzymes based on iron and nickel can efficiently catalyze both the generation and oxidation of hydrogen [12]. Structural investigations show that the metal ligands in these enzymes facilitate the cleavage/formation of the H–H bond and the transfer of protons to and from the distal metal center [13]. Considering that triazenido ligands show greater basicity on their [N···N···N]⁻ moieties relative to the neutral nitrogen, triazenido ligands should be more electron donating and hence amenable to binding to protons and hydrogen evolution. Hence, triazenido-metal complexes were selected as potential hydrogen evolution reaction (HER) catalysts [14, 15]. In this paper, we report the synthesis, structure, characterization and magnetic properties of a dinuclear copper(II) complex [Cu₂L₄], as well as its electrocatalytic properties for proton and water reduction.

Experimental

Physical measurements are given in the supplementary materials.

Synthesis of HL

A solution of 2-iodoaniline (Aladdin, Analytical, 2.19 g, 10 mmol) in water (5 mL) was mixed with 1 mol L⁻¹ HCl 30 mL (30 mmol) at 0 °C. An aqueous solution (15%) of sodium nitrite (1.04 g, 15 mmol) was added dropwise with stirring. Once the amine was dissolved, a 15% solution of 2-aminobenzothiazole (Aladdin, Analytical, 1.5 g, 10 mmol) in ethanol (10 mL) was added at 0 °C, and the mixture was stirred for 3 h at 25 °C. The mixture was then neutralized with 15% aqueous NaCH₃CO₂ to give a yellow precipitate. This was filtered off, and the solid was purified by crystallization at − 4 °C from 5:1 ethyl acetate/THF to obtain yellow crystals, which were collected and dried in vacuo, Yield 2.5 g (66%). Calcd for C₁₃H₉N₄SI: C, 41.06; H, 2.37; N, 14.74. Found: C, 41.57; H, 2.29; N, 14.56%. ¹H NMR (400 MHz, DMSO) δ 11.50 (s, 1H), 7.91 (d, J = 12.1 Hz, 2H), 7.54 (d, J = 7.4 Hz, 2H), 7.43 (t, J = 7.6 Hz, 2H), 7.01 (d, J = 40.6 Hz, 2H). ¹³C NMR (400 MHz, DMSO) δ 166.14, 152.69, 139.52, 129.50, 129.24, 126.32, 125.22, 123.46, 121.81, 120.63, 118.87, 117.45, 95.70. IR (KBr): ν 3700, 3058, 1731, 1578, 1497 cm⁻¹ (Fig. S1).

Synthesis of [Cu₂L₄]

To a solution of HL (0.37 g, 1 mmol) and triethylamine (0.10 g, 1 mmol) in dichloromethane/acetonitrile (20 mL, 1:1), CuCl₂·2H₂O (0.085 g, 0.5 mmol) was added and the mixture was stirred for 10 min. The solution was allowed to slowly evaporate, affording deep green crystals, which were collected and dried in vacuo, Yield 0.16 g (37%). Calcd for C₅₂H₃₂Cu₂I₄N₁₆S₄: C, 37.99; H, 1.97; N, 13.64. Found: C, 38.28; H, 1.94; N, 13.75%. ¹H NMR (400 MHz, DMSO) 7.93 (s, 2H), 7.55 (d, J = 8.0 Hz, 2H), 7.44 (t, J = 7.6 Hz, 2H), 7.02 (d, J = 32.6 Hz, 2H). IR (KBr): ν 3058, 1775, 1522, 1404 cm⁻¹ (Fig. S2).

X-ray crystallography

X-ray analysis of the complex was carried out with an Xcalibur Eos Gemini X-ray diffractometer using graphite monochromated Mo-Kα radiation (λ = 0.71073 Å) at room temperature. Empirical absorption corrections were applied using the SADABS program [16]. The structure was solved using direct methods. Non-hydrogen atoms were all refined anisotropically. Hydrogen atoms were included in geometrically idealized positions, and thermal parameters were fixed during structural refinement. All calculations were performed using the SHELXTL-2014 computer program [17]. Table S1 gives crystallographic data for the complex, and selected bond lengths are listed in Table 1.

Table 1 Selected bond distances (Å) for the complex

Results and discussion

The triazene HL was prepared by the reaction of 2-iodoaniline, sodium nitrite, and 2-aminobenzothiazole (Scheme 1 and Fig. S3). As shown in Fig. S4, the ¹H NMR spectrum showed a singlet at 11.5 ppm, assigned to the triazene group hydrogen atom, and a multiplet in the range of 8.0–6.9 ppm for the aromatic protons. In the presence of Et₃N, the reaction of HL with CuCl₂·2H₂O provided the copper complex, [Cu₂(L)₄].

Scheme 1

Synthesis of HL and [Cu₂(L)₄]. R = NNC₆H₄I, (i) NaNO₂, HCl; (ii) Et₃N, CuCl₂

As shown in Fig. 1, [Cu₂(L)₄] consists of two copper centers bridged by four anionic ligands (L⁻). The structure is in agreement with the ESI–MS analysis. As shown in Fig. S5, the complex gave a single ion at a mass-to-charge ratio (m/z) of 1644.7788, which can be attributed to [Cu₂(L)₃(L–H)]⁺. The two copper centers share the same coordination environment, in which each copper atom is coordinated by four nitrogen atoms from four anionic ligands (L⁻). In contrast to previously reported bridged complexes with triazenido ligands [18, 19], L⁻ serves as a monodentate ligand to each copper center. The bond distances between Cu and the benzothiazole N are in the range of 1.968(4)–2.050(4) Å, while the Cu–N(triazene) bond lengths are 1.964(4)–2.057(4) Å (Table 1), which are similar to those found in other copper(II) triazenide complexes 1.973–2.088 Å [20, 21]. The distance between the two copper centers (2.6925(8) Å) is longer than that found previously in the reported copper(II) complexes, Cu₂(dpt)₄ (dpt = 1,3-diphenyltriazenide) (2.441 Å) [21] and Cu₂(OAc)₄ (2.616 Å) [22].

Fig. 1

ORTEP drawing of [Cu₂(L)₄] with thermal ellipsoids at the 50% probability level

Magnetic behavior of the copper complex

The magnetic properties of the complex were measured at a field of 2000 Oe over the temperature range from 2 to 300 K, with the results shown in Fig. 2. Copper(II), which has the d⁹ configuration, is expected to exhibit paramagnetic behavior. However, [Cu₂(L)₄] is diamagnetic at 300 K, because of a strong antiferromagnetic interaction between the copper atoms [23]. The magnetic analysis was carried out using the Bleaney–Bowers equation based on the Heisenberg model H = − 2JS₁S₂, where ρ is the fraction of monomeric impurity. The magnetism of the complex can be well reproduced by a modified Bleaney–Bowers Eq. (1) [24];

$$\chi_{M} T = \frac{{2N\beta^{2} g^{2} }}{k}\left[ {3 + \exp \left( { - \frac{J}{kT}} \right)} \right]^{ - 1} (1 - \rho ) + \frac{{N\beta^{2} g^{2} }}{2k}\rho + 2N\alpha T.$$

(1)

Fig. 2

Temperature dependence of χ_MT at 2000 Oe for [Cu₂L₄]. The solid line: the fitting data

Least-squares fitting of the experimental data led to J = − 558 cm⁻¹, Nα = 3.1 × 10⁻⁴ cm³ mol⁻¹ and ρ = 0.00065. Hence, there is a very strong antiferromagnetic exchange interaction between the copper atoms. This result is also in agreement with the NMR analysis. According to Fig. S6, the ¹H NMR spectrum of the copper complex is indicative of a single-ligand environment. ¹H resonances are found in the range of 8.0–6.9 ppm for the aromatic protons.

Electrocatalytic behavior of the complex

Since triazenido complexes can act as electrocatalysts for hydrogen generation [25,26,27], we have investigated the electrochemical behavior of [Cu₂(L)₄]. Cyclic voltammograms (CVs) of the complex (Fig. 3a), HL (Fig. S7) and CuCl₂ (Fig. S8) were measured in CH₃CN with 0.10 M [(n-Bu)₄N]ClO₄ as the supporting electrolyte. All values are reported versus Ag/AgNO₃ as a reference. As shown in Fig. S7, HL exhibited a quasireversible couple at − 1.51 V and a reduction wave at − 1.70 V. Compared to free HL, the copper complex displayed a quasireversible couple at − 1.49 V and a reduction wave at − 1.67 V (Fig. 3a). Hence, the potentials were shifted positively, and current strengths increased for the complex compared to the free ligand. According to Fig. S9, all the redox waves were characteristic of diffusion-controlled processes, as scan rate analyses of voltammograms exhibited linear dependences in plots of current versus ν^1/2.

Fig. 3

a Cyclic voltammogram (CV) of 0.56 mM [Cu₂L₄] in 0.10 M of [n-Bu₄N]ClO₄ CH₃CN solution at a glassy carbon electrode and a scan rate of 100 mV/s. b CVs of a 0.56 mM solution of [Cu₂L₄], with varying concentrations of acetic acid in CH₃CN. Conditions: 0.10 M [n-Bu₄N]ClO₄ as supporting electrolyte, scan rate: 100 mV/s, glassy carbon working electrode (1 mm diameter), Pt counter electrode, Ag/AgNO₃ reference electrode, Fc internal standard (*). (Color figure online)

Next, acetic acid was used as proton source to investigate the catalytic behavior of the copper complex for hydrogen evolution. According to the data plotted in Fig. 3b, with sequential increments of acetic acid concentration (from 0.00 to 4.64 mM), the potential strengths of the complex at − 1.69 and − 1.54 V were proportionately increased, which is attributed to the catalytic generation of H₂ from acetic acid [28]. These results suggest that the electrocatalytic generation of hydrogen by [Cu₂(L)₄] involves the reduction of Cu(II) to Cu(I) or of L⁻ to L, with associated protonations.

In order to investigate whether the triazene HL may play a role in the catalytic process, the electrochemical behavior of free HL was investigated under the same conditions. From Fig. S10, with sequential increments of acetic acid concentration from 0.00 to 4.64 mM, the peak currents at − 1.56 and − 1.71 V increased, indicating that HL can also catalyze proton reduction to H₂. When the concentration of acetic acid reached 3.99 mM, a new reduction wave appeared, suggesting that new components were formed.

Based on the above-mentioned results and previous studies [29, 30], we can propose a possible catalytic H₂ evolution mechanism for the present complex. As shown in Scheme 2, a putative [Cu ^I₂ (L)₄]²⁻ species is formed by two one-electron reductions of [Cu ^II₂ (L)₄]. Then, the introduction of two protons gives a three-coordinate dinuclear copper complex, [H–L–Cu^I–L]₂. Further addition of two protons provides a copper(III)-hydride species, {[H–L–Cu^III(H)–L]₂}²⁺. Finally, two one-electron reductions of the {[H–L–Cu^III(H)–L]₂}²⁺ species affords H₂, with regeneration of the starting complex.

Scheme 2

Possible catalytic mechanism for proton reduction by the complex. R = NNC₆H₄I

A series of bulk electrolysis experiments was carried out to further characterize the electrocatalytic activity of the complex, with the results listed in Fig. S11. For example, during 2 min of electrolysis at − 1.45 V, a 0.10 M [n-Bu₄N]ClO₄ solution of the complex exhibited a charge of 55 mC (Fig. S11a), accompanied by the generation of a gas which was confirmed as H₂ by gas chromatography. According to Fig. S12, ~ 0.035 mL of H₂ was produced over an electrolysis period of 1 h. Under the same conditions, this system only gave 10 mC of charge (Fig. S11b), indicating that this complex does serve as a catalyst for hydrogen production from acetic acid. The catalytic activity was estimated using Eqs. (2) and (3) [10] (Eq. S1), giving the results listed in Fig. S13. For instance, at an OP of 941.6 mV, this electrocatalytic system could produce 50 mol of hydrogen per mole of catalyst per hour (mol H₂/mol catalyst/h).

$${\text{TOF}} = {\Delta\text{C/}}(F \ast n_{1} \ast n_{2} \ast t)$$

(2)

$$\begin{aligned} {\text{Overpotential}} &= {\text{Applied}}\;{\text{potential}} - E^{ \odot } \,_{\text{HA}} \\ &= {\text{Applied}}\;{\text{potential}} - (E^{ \odot } \,_{{{\text{H}}^{ + } }} - (2.303\;RT /F)pK_{\text{aHA}} ) \\ \end{aligned}$$

(3)

where ΔC is the charge from the catalyst solution during controlled-potential electrolysis (CPE), minus the charge from an equivalent solution without catalyst during CPE; F is Faraday’s constant, n₁ is the number of moles of electrons required to generate one mole of H₂, n₂ is the number of moles of catalyst in solution, and t is the duration of electrolysis in seconds. E^{⊙ +}_H is the standard potential for the solvated proton/dihydrogen couple, and K_aHA is the dissociation constant of acetic acid.

To explore the catalytic function of the complex in aqueous media, its electrochemical behavior was investigated in aqueous buffer (CH₃CN: H₂O = 2:5 as solvent). As shown in Fig. S14, in the absence of the complex, the catalytic current was not apparent until a potential of − 1.84 V versus Ag/AgCl was attained. Upon addition of the complex, the onset of catalytic current was observed at about − 1.40 V, showing that the complex can efficiently reduce the potential. Moreover, the current strength increased significantly with buffer of decreasing pH values from 7.0 to 4.6 μM (Fig. S15). According to Fig. S15, the onset of this catalytic current was clearly influenced by the aqueous solution pH, such that the applied potential declined with increasing pH, which is indicative of a catalytic process [15].

Bulk electrolyses catalyzed by [Cu₂(L)₄] and HL in 0.25 M aqueous buffer (pH 7.0) were then used to study the catalytic activities. From Fig. 4a, when the applied potential was − 1.45 V, the maximum charge was only 22 mC during 2 min of electrolysis in the absence of the complex. However, addition of the complex afforded 678 mC during 2 min of electrolysis (Fig. 4b), accompanied by a large amount of gas bubbles, which were confirmed as H₂ by gas chromatography. According to Fig. S16a, ~ 4.26 mL of H₂ was afforded over an electrolysis period of 1 h with a Faradaic efficiency of 98.6% for H₂ (Fig. S16b). Turnover frequencies (TOFs) and overpotentials (OPs) were calculated from Eqs. (2) and (4) (Eq. S2) [15], giving the results listed in Fig. S17. For example, this system could afford 502 mol H₂/mol catalyst/h at an OP of 837.6 mV, where

$${\text{OP}} = {\text{Applied}}\;{\text{potential}} - E({\text{pH}}) = {\text{Applied}}\;{\text{potential}} - ( - \,0.059\;{\text{pH}}).$$

(4)

Fig. 4

a Charge buildup versus time from electrolysis of a 0.25 M buffer (pH 7.0) under − 1.45 V versus Ag/AgCl. b Charge buildup versus time from electrolysis of a 4.18 μM [Cu₂L₄] in 0.25 M buffer solution (pH 7.0). All data have been deducted blank. (Color figure online)

In a similar procedure to that used for the complex, bulk electrolyses from HL was conducted in 0.25 M neutral buffer. According to Fig. S18, during 2 min of electrolysis, the addition of HL gave 178 mC of charge. The TOF was calculated as 146 mol of hydrogen per mole of catalyst per hour at an OP of 837.6 mV (Eq. S3, Fig. S19). Clearly, the complex exhibits much higher efficiency for hydrogen production than the free ligand HL.

In terms of TOF values, the results plotted in Fig. S17 show that the catalytic activity of the present complex (502 mol H₂/mol catalyst/h at an OP of 869 mV) is higher than those of some previously reported molecular catalysts based on copper complexes. These include a copper complex of a tetradentate aminophenol ligand with a TOF of 300 mol H₂/mol catalyst/h at an OP of 869 mV [31], a copper complex of a dicyano acetic acid methyl ester ligand giving 285 mol H₂/mol catalyst/h at an OP of 787.6 mV [32], a copper(II) complex of 1,3-bis[(4-chloro)benzene]triazene ligand with a TOF of 82 mol H₂/mol catalyst/h at an OP of 837.6 mV [25], and a copper(I) complex of 1-[(2-methoxy) benzene]-3-[2-(chloro)benzene] triazene (272 mol H₂/mol catalyst/h at an OP of 839 mV) [26]. However, the TOF of the present complex is lower than that of [Cu ^I₄ (L′)₄] (HL′ = [1,3-bis(2-methoxy)benzene]triazene) with a TOF of 1034.31 mol H₂/mol catalyst/h at an OP of 837.6 mV [10]. These findings show that the donor type and electronic properties of the ligands play a vital role in determining the structure and catalytic activity of the corresponding copper complexes.

Controlled-potential electrolysis (CPE) experiments on buffer solutions of HL, CuCl₂, a mixture of HL and CuCl₂ or, [Cu₂(L)₄] were each measured under the same conditions. According to Fig. 5, a neutral buffer solution (pH 7.0) only afforded 32 mC of charge during 2 min of electrolysis under − 1.45 V. The use of solely HL or CuCl₂ as a catalyst afforded hydrogen with 157 and 187 mC of charge, respectively. However, when [Cu₂(L)₄] was employed as catalyst, the system provided H₂ with 525 mC of charge. Thus, prior formation of the copper complex is essential for catalytic activity in this system.

Fig. 5

Charge buildups versus time from [Cu₂L₄] and the related components: A neutral buffer (pH 7.0) (black), 4.18 μM HL (red), 4.18 μM CuCl₂·2H₂O (blue) and 4.18 μM [Cu₂L₄] (green) under − 1.45 V versus Ag/AgCl in a 0.25 M buffer (pH 7.0). (Color figure online)

Conclusion

The dinuclear triazenido-copper complex [Cu₂L₄] was formed by the reaction of CuCl₂ with 1-[(2-iodo)benzene]-3-[benzothiazole] triazene. The solid copper complex shows a magnetic exchange interaction between the copper atoms mediated by triazenido bridges. In homogeneous solution, the complex can act as an electrocatalytic system for hydrogen production both from acetic acid and aqueous buffer. The present system is broadly comparable to other examples available in the literature.