Acid-Induced Mechanism Change and Overpotential Decrease in Dioxygen Reduction Catalysis with a Dinuclear Copper Complex

Abstract

Catalytic four-electron reduction of O₂ by ferrocene (Fc) and 1,1′-dimethylferrocene (Me₂Fc) occurs efficiently with a dinuclear copper(II) complex [Cu^II₂(XYLO)(OH)]²⁺ (1), where XYLO is a m-xylene-linked bis[(2-(2-pyridyl)ethyl)amine dinucleating ligand with copper-bridging phenolate moiety], in the presence of perchloric acid (HClO₄) in acetone at 298 K. The hydroxide and phenoxo group in [Cu^II₂(XYLO)(OH)]²⁺ (1) undergo protonation with HClO₄ to produce [Cu^II₂(XYLOH)]⁴⁺ (2) where the two copper centers become independent and the reduction potential shifts from −0.68 V vs SCE in the absence of HClO₄ to 0.47 V_j this makes possible the use of relatively weak one-electron reductants such as Fc and Me₂Fc, significantly reducing the effective overpotential in the catalytic O₂-reduction reaction. The mechanism of the reaction has been clarified based on kinetic studies on the overall catalytic reaction as well as each step in the catalytic cycle and also by low-temperature detection of intermediates. The O₂-binding to the fully reduced complex [Cu^I₂(XYLOH)]²⁺ (3) results in the reversible formation of the hydroperoxo complex ([Cu^II₂(XYLO)(OOH)]²⁺) (4), followed by proton-coupled electron-transfer (PCET) reduction to complete the overall O₂-to-2H₂O catalytic conversion.

Introduction

The heme/copper (heme a3/Cu_B) heterodinuclear center in cytochrome c oxidases (CcO) catalyzes the four-electron and four-proton reduction of dioxygen (O₂) to water in the final stage of the respiratory chain (eq 1).^1-4

$${\mathrm{O}}_2+4e^{-}(\text{from}\mathrm{cyt}-c_{\text{reduced}})+8{\mathrm{H}}^{+}\left(\text{from outside membrane}\right)\to 2{\mathrm{H}}_2\mathrm{O}+4{\mathrm{H}}^{+}\left(\text{membrane translocated}\right)+4\mathrm{cyt}-c_{oxidized}$$ (1)

The catalytic four-electron reduction of O₂ to water has attracted much interest because of the important role in respiration⁵ and also potential application in fuel cell technology.⁶ The four-electron reduction of O₂ is catalyzed by platinum instilled in carbon at the cathode in fuel cells.^6-9 To achieve substantial activity, high loadings of this precious metal are required which have prompted research efforts to develop catalysts based on non-precious metals such as Co and Fe.^10-18 Cu catalysts have also merited considerable interest,^19-24 in relation with copper containing enzymes, the so called multi-copper oxidases (MCO’s), which efficiently effect the four-electron four-proton reduction to water as part of their function.^25-28

In contrast to such heterogeneous systems, investigations on the catalytic reduction of O₂ by metal complexes in homogeneous systems have provided deeper insight into the catalytic mechanism of the two-electron and four-electron reduction of O₂. Solution variable temperature kinetic studies and detection of reactive metal–O₂ intermediates reveal the controlling factors in the two- vs four-electron reduction of O₂ with metal complexes.^29-35 The key feature is the modification of supporting ligand environments which can allow for differing copper-dioxygen intermediates.

So far we found that two mononuclear copper complexes having tmpa (tmpa = tris(2-pyridylmethyl)amine) (A),³⁶ bzpy1 = N,N-bis[2-(2-pyridyl)ethyl]benzylamine (B)]³⁷ and one dinuclear copper complex with N3 ligand (N3 = −(CH₂)₃-linked bis[(2-(2-pyridyl)ethyl)amine])³⁷ (C) (Scheme 1) efficiently catalyze the four-electron reduction of dioxygen via the formation of [{(tmpa)Cu^II}₂(μ-1,2-O₂²⁻)]²⁺ (A1), [Cu^III₂(bzpy1)(μ-(O²⁻)₂)²⁺ (B1) and [Cu^II₂(N3)(μ-η²:η²-O₂²⁻)]²⁺ (C1) intermediates, respectively, which were prone to proton promoted reductive O–O cleavage to give water, in preference to simple protonation leading to H₂O₂^36,37. However, the dinuclear copper(II) complex [Cu^II₂(XYLO)-(OH)]²⁺ (1) (Scheme 1) {where XYLO is a m-xylene-linked bis[(2-(2-pyridyl)ethyl)amine dinucleating ligand with copper-bridging phenolate moiety] catalyzes the two-electron reduction of O₂ to hydrogen peroxide in the presence of trifluoroacetic acid (HOTF) via the hydroperoxo intermediate (D1).³⁸ In all cases, however, only a strong one-electron reductant such as decamethylferrocene (Fc*) could be used to reduce O₂ with the copper complexes. There has so far been no example for the catalytic reduction of O₂ by weaker one-electron reductants than Fc* with Cu complexes.

Scheme 1.

We report herein that [Cu^II₂(XYLO)(OH)]²⁺ (1) can instead act as an efficient catalyst for the four-electron reduction of O₂ by weaker one-electron reductants than Fc* such as ferrocene (Fc) and 1,1′-dimethylferrocene (Me₂Fc) in the presence of HClO₄ in acetone. The reasons why the same catalyst can act in either the two-electron or four-electron reduction of O₂ by Fc* and Fc in the presence of CF₃COOH and HClO₄, respectively, are elucidated based on new kinetic studies on the overall catalytic reactions as well as each catalytic step and also by detection of copper-dioxygen derived intermediates which form during the catalytic cycle. The present study leads to successful achievement of two important features, which have never been previously observed:

• -We have been able to effect catalytic O₂ reduction using significantly less overpotential with a dinuclear copper complex, which is of course more desirable and more energy efficient.
• -We have been able to change the number of electrons in the catalytic reduction of O₂ from two electrons to four electrons by only increasing the acidity of the proton source employed.

The mechanistic insights obtained in this study should serve as useful and broadly applicable principles for future design of more efficient catalysts in fuel cells.

Results and Discussion

Catalytic Four-Electron Reduction of O₂ by Fc and Me₂Fc with 1 in the Presence of HClO₄

The addition of a catalytic amount of 1 to an air-saturated acetone solution of Me₂Fc and perchloric acid (HClO₄) results in the efficient reduction of O₂ by Me₂Fc to afford the corresponding dimethylferrocenium cation (Me₂Fc⁺). The spectral changes for the catalytic reduction of O₂ by Me₂Fc with 1 in the presence of HClO₄ in acetone at 298 K are shown in Figure S1 [see Supporting Information (SI)]. When more than four equivalents of Me₂Fc relative to O₂ (limiting [O₂]) were employed, four equivalents of Me₂Fc⁺ (λ_max = 650 nm, ε = 360 M⁻¹ cm⁻¹) were formed in the presence of excess HClO₄ (Figure 1). From iodometric titration experiments, it was confirmed that no H₂O₂ had formed after completion of the reaction (Figure S2 in SI). Thus, the four-electron reduction of O₂ by Me₂Fc occurs efficiently with a catalytic amount of 1 in the presence of HClO₄ (eq 2).

Figure 1.

UV-vis spectral changes observed in the four-electron reduction of O₂(1.0 mM) by Me₂Fc (6.0 mM) with HClO₄ (40 mm) catalyzed by 1 (0.20 mM) in acetone at 298 K. Inset shows the time profile of the absorbance at 650 nm due to Me₂Fc⁺.

$$4{\mathrm{Me}}_2\mathrm{Fc}+{\mathrm{O}}_2+4{\mathrm{H}}^{+}\to 4{\mathrm{Me}}_2{\mathrm{Fc}}^{+}+2{\mathrm{H}}_2\mathrm{O}$$ (2)

When Me₂Fc was replaced by the weaker reductant Fc, the four-electron reduction of O₂ by Fc also occurred efficiently with 1 (Figure S3 in SI). The rate of formation of Me₂Fc⁺ and Fc⁺ obeyed pseudo-first order kinetics under the conditions that [1] << [O₂] < [Me₂Fc] < [HClO₄] (Figure 1 inset). The time profiles of the absorbance at 650 nm due to Me₂Fc⁺ and at 620 nm due to Fc⁺ and the first-order plots by varying HClO₄, O₂ and catalyst are shown in Figures S4, S5, S6a and Figure 2a, respectively. The pseudo-first order rate constant (k_obs) increased linearly with increasing concentration of 1 (Figures 2b and S6b). It should be noted that no oxidation of Me₂Fc occurs by O₂ in the presence of HClO₄ without 1, under the present experimental conditions, even though ferrocene derivatives are known to be slowly oxidized by O₂ in the presence of strong acids.^39,40 It should also be noted that the use of a non-coordinating solvent (acetone) is essential for the catalytic reduction of O₂ by Me₂Fc with 1 in the presence of HClO₄, because a coordinating solvent such as acetonitrile prohibits such chemistry. The k_obs values were also proportional to concentrations of HClO₄ (Figures 2c and S6c) and O₂ (Figures 2d and S6d). Thus, the kinetic equation is given by eqs 3 and 4, where k_cat is the apparent fourth-order rate constant (k_cat) for the catalytic four-electron reduction of O₂ by Me₂Fc

$$\mathrm{d}\left[{\mathrm{Me}}_2{\mathrm{Fc}}^{+}\right]∕\mathrm{d}t=k_{\mathrm{obs}}\left[{\mathrm{Me}}_2\mathrm{Fc}\right]$$ (3)

$$k_{\mathrm{obs}}=k_{\mathrm{cat}}\left[1\right]\left[{\mathrm{O}}_2\right]\left[{\mathrm{H}}^{+}\right]$$ (4)

when k_obs is given by eq 4. The k_cat values of Me₂Fc and Fc are listed in Table S1 in SI. The kinetic formulation in eqs 3 and 4 obtained in this study is quite unique because the rate is proportional to concentrations of not only 1 but also O₂, HClO₄ and Me₂Fc, in sharp contrast to the previously reported cases of Cu catalysis for O₂ reduction in which the rate was rather independent of O₂ or H⁺.^36-38 In such a case, the rate-determining step in the catalytic cycle should involve the reactions of 1 with Me₂Fc, O₂ and H⁺. In order to elucidate the catalytic mechanism that can explain such a unique kinetic formulation, we decided to examine each portion of the catalytic cycle, step by step.

Figure 2.

(a) Time profiles of the absorbance at 650 nm due to Me₂Fc⁺ in the four-electron reduction of O₂ catalyzed by 1 (0.080 mM (green), 0.10 mM (black), 0.12 mM (red), 0.20 mM (dark yellow) and 0.25 mM (blue)) with Me₂Fc (4.0 mM) in the presence of HClO₄ (40 mM) in an air-saturated ([O₂] = 2.2 mM) acetone solution at 298 K. (b) Plot of k_obs vs [1] for the four-electron reduction of O₂ catalyzed by 1 with Me₂Fc (4.0 mM) in the presence of HClO₄ (40 mM) in an air-saturated ([O₂] = 2.2 mM) acetone solution at 298 K. (c) Plot of k_obs vs [HClO₄] for the four-electron reduction of O₂ by Me₂Fc (4.0 mM) catalyzed by 1 (0.12 mM) in an acetone solution containing O₂ (2.2 mM) at 298 K. (d) Plot of k_obs vs [O₂] for the four-electron reduction of O₂ catalyzed by 1 (0.20 mM) with Me₂Fc (3.2 mM) in the presence of HClO₄ (40 mM) in an acetone solution at 298 K.

Protonation of 1 to Produce Two Independent Cu Centers

The catalytic reduction of O₂ by Me₂Fc and Fc with 1 was made possible only by the presence of HClO₄. The effect of protonation of 1 was first examined by the spectral titration of 1 with HClO₄ (Figure 3). The absorption band at 378 nm due to [Cu^II₂(XYLO)(OH)]²⁺ decreased with increasing concentration of HClO₄, and this was completely different from the spectral behavior of complex 1 with CF₃COOH (HOTF) where the absorption band was shifted to 420 nm and a clean isosbestic point was observed at 430 nm.³⁸ This spectral changes in the presence of HClO₄ indicate that not only the hydroxide group but also the phenoxo group of [Cu^II₂(XYLO)(OH)]²⁺ (1) is protonated with HClO₄, a much stronger acid than CF₃COOH, to produce [Cu^II₂(XYLOH)]⁴⁺ (2) (Scheme 2).

Figure 3.

(a) UV-visible spectral changes of [Cu^II₂(XYLO)(OH)](PF₆)₂ (1) (0.20 mM) upon addition of HClO₄ (0.0–6.0 mM) in acetone at 298 K. (b) Absorbance changes at 378 nm as a function of HClO₄ concentration.

Scheme 2.

The two-step protonation in Scheme 2 was confirmed by the EPR titration with HClO₄. The starting dinuclear copper(II) complex [Cu^II₂(XYLO)(OH)](PF₆)₂ (1) is EPR silent because of antiferromagnetic coupling of the two Cu(II) ions (Figure 4a). It should be noted that when the complex 1 was protonated with trifluoroacetic acid (HOTF), the protonated complex was EPR silent, indicating the two Cu(II) ions still maintain an electronic/magnetic interaction after the protonation of 1 with HOTF.³⁸ In the presence of one equiv of HClO₄ which can protonate the OH group to produce H₂O afforded the EPR silent species. In the presence of excess HClO₄, however, a typical axial Cu(II) EPR spectroscopic signal was observed, indicating that both the hydroxide and the phenoxo group were protonated to produce two independent Cu(II) sites (Figure 4b).

Figure 4.

X-band EPR spectra of [Cu^II₂(XYLO)(OH)](PF₆)₂ (1) (1.0 mM) in the (a) absence and (b) presence of HClO₄(5.0 mM) recorded in acetone at 5 K. The experimental parameters: microwave frequency = 9.646 GHz, microwave power = 1.0 mW and modulation frequency = 100 kHz.

Once the phenoxo group is protonated, the two Cu sites become independent and more electron deficient because of the lack of the coordination of the anionic phenoxo donor ligand. This was confirmed by the cyclic voltammetry (CV) and difference pulse voltammetry (DPV) measurements on 1 in the absence and presence of HClO₄ in acetone as shown in Figure 5. An irreversible cathodic peak current was observed at −1.08 V vs SCE at a sweep rate of 0.10 V s⁻¹, while the DPV exhibits the cathodic peak at −0.68 V vs SCE. The cathodic peak is much more negative as compared to the one-electron oxidation potential of Me₂Fc (E_ox = 0.26 V vs SCE) and Fc (E_ox = 0.37 V vs SCE).^31,41 This is the reason why no electron transfer from Me₂Fc or Fc to 1 occurs in the absence of acid. In the presence of HClO₄, however, the one-electron reduction potential determined by DPV is shifted to the positive direction (E_red = 0.47 V vs SCE) (Figure 5b) which is now more positive than the oxidation potential of Fc (E_ox = 0.37 V vs SCE) and Me₂Fc (E_ox = 0.26 V vs SCE). Thus, electron transfer from Fc and Me₂Fc to 1 becomes energetically feasible in the presence of HClO₄. This was confirmed by examination of electron transfer from Fc and Me₂Fc to 1 in the presence of excess HClO₄ in acetone (vide infra).

Figure 5.

Cyclic voltammograms (CV, solid line) and differential pulse voltammograms (DPV, dotted line) of 1 (2.0 mM) in the (a) absence and (b) presence of HClO₄ (50 mM) in deaerated acetone at 298 K. TBAPF₆ (0.20 M) was used as an electrolyte.

Electron Transfer from Ferrocene Derivatives to 1 in the Presence of HClO₄

No electron transfer from Me₂Fc and Fc to [Cu^II₂(XYLO)(OH)]²⁺(1) occurs in the absence of HClO₄ in acetone at 298 K, whereas the electron transfer occurs to completion to produce the corresponding ferrocenium ions in the presence of HClO₄ (Scheme 3). The rates of electron transfer from Me₂Fc and Fc to 1 were determined in presence of HClO₄ at 298 K under an argon atmosphere. The rate of formation of Me₂Fc⁺ and Fc⁺ obeyed pseudo-first kinetics under the conditions that [HClO₄] >> [Me₂Fc] and [Fc] > [1] as shown in Figure S8. The amount of Me₂Fc⁺ produced by electron transfer from Me₂Fc to 1 in presence of HClO₄ is twice that observed for complex 1 (0.10 mM). This result indicates that the two Cu sites of protonated 1 (Scheme 2) act independently without any interaction between them. The observed pseudo-first-order rate constant (k_obs) increased linearly with increasing concentration of Me₂Fc and Fc (Figure 6). The second-order rate constants (k_et) of electron transfer from Me₂Fc and Fc to 1 at 298 K were determined to be 2.7 × 10² M⁻¹ s⁻¹ and 1.8 × 10² M⁻¹ s⁻¹ from the slopes of linear plots of k_obs vs [Me₂Fc] and [Fc], respectively. The k_et values are listed in Table S1 together with the k_cat values.

Scheme 3.

Figure 6.

(a) Plot of k_obs vs [Me₂Fc] in the electron transfer from Me₂Fc to [Cu^II₂(XYLO)(OH)](PF₆)₂ (1) (0.10 mM) in presence of HClO₄(40 mM) in acetone at 298 K. (b) Plot of k_obs vs [Fc] in the electron transfer from Fc to 1 (0.10 mM) in presence of HClO₄(40 mM) in acetone at 298 K.

The k_et values (2.7 × 10² M⁻¹ s⁻¹ and 1.8 × 10² M⁻¹ s⁻¹) of electron transfer from Me₂Fc and Fc to 1 in the presence of 40 mM HClO₄ are significantly larger than the corresponding k_obs/[1] (see eqs 3 and 4) values (42 M⁻¹ s⁻¹ in Figure 2b and 10 M⁻¹ s⁻¹ in Figure S6b) in the presence of 40 mM HClO₄, respectively. Thus, the electron-transfer step is not the rate-determining step in the catalytic cycle. The electron transfer from Me₂Fc and Fc to 1 in the presence of excess HClO₄ without O₂ results in formation of the protonated dinuclear Cu(I) complex, [Cu^I₂(XYLOH)]²⁺, which can reduce O₂. Next we examined the reaction of [Cu^I₂(XYLOH)]²⁺ with O₂ at low temperatures to detect any copper-dioxygen intermediate.

Reversible Binding of O₂ to [Cu^I₂(XYLOH)]²⁺

When O₂ was bubbled into an acetone solution of separately synthesized dicopper(I) complex, [Cu^I₂(XYLOH)]²⁺ at 193 K, the absorption band at 395 nm due to the hydroperoxo complex, [Cu^II₂(XYLO)(OOH)]²⁺,⁴⁷ appeared immediately as shown in Figure 7 and Scheme 4. The yield of hydroperoxo complex, [Cu^II₂(XYLO)(OOH)]²⁺ (4) was determined to be 100 % at 193 K. As the temperature increased, the absorption band at 395 nm due to [Cu^II₂(XYLO)(OOH)]²⁺ decreased. This process was reversible at low temperatures up to 223 K (Figure 8a).⁴² The temperature dependence of K_eq was examined (Figure S9) and the van’t Hoff plot (Figure 8b) afforded ΔH = −31 kJ mol⁻¹ and ΔS = −86 J K⁻¹ mol⁻¹. The equilibrium constant at 298 K was estimated to be 11 M⁻¹ from the extrapolation of the van’t Hoff plot. The equilibrium lies to the reactant side at 298 K when only a small portion of [Cu^I₂(XYLOH)]²⁺ is converted to [Cu^II₂(XYLO)(OOH)]²⁺ (~10%).

Figure 7.

Formation of the hydroperoxo complex, [Cu^II₂(XYLO)(OOH)]²⁺ (λ_max= 395 nm) in the reaction of [Cu^I₂(XYL-OH)]²⁺(0.11 mM) with O₂ in acetone at 193 K.

Scheme 4.

Figure 8.

(a) UV-visible spectra indicating the reversible nature of dioxygen binding to [Cu₂^I(XYLOH)]²⁺(3). Bubbling O₂ into an acetone solution of 3 produces [Cu₂^II(XYLO)(OOH)]²⁺ (4) at 193 K (black, solid line) (inset zoom view). Increasing the temperature up to 223 K produces dark yellow solid spectrum. After cooling to 193 K again gives black dotted spectrum. (b) van’t Hoff plot to determine the activation parameters, enthalpy and entropy, in the dioxygen binding to [Cu₂^I(XYLOH)]²⁺ in acetone.

This intermediate was further reduced by decamethylferrocene (Fc*) in the presence of HClO₄ at 193 K (Scheme 5, Figure 9). Fc* was used because the reactions with Me₂Fc and Fc were too slow to be followed at 193 K.

Scheme 5.

Figure 9.

Formation of Fc*⁺ by addition of Fc* (0.25 mM) and HClO₄ (40 mM) to the hydroperoxo complex (0.060 mM) (generated by O₂ bubbling to the solution of [Cu₂^I(XYLOH)]²⁺ (0.060 mM)) at 193 K.

An alternate reaction pathway that may contribute in a small way to the overall chemistry and catalytic cycle comes about if the hydroperoxo complex [Cu^II₂(XYLO)(OOH)]²⁺ (4) is protonated in the presence of HClO₄ to yield hydrogen peroxide and protonated dicopper(II) complex, [Cu^II₂(XYLOH)]⁴⁺ (2) (Scheme 6a). This was separately demonstrated using excess HClO₄ (10 equiv) (Figure S10). However, we find that if hydrogen peroxide forms in this manner, it would be readily reduced to H₂O by the dicopper(I) complex [Cu^I₂(XYLOH)]²⁺ (3) (Scheme 6b and Figure 10; see also Figure S11 in SI). Although H₂O₂ and Cu(II)-H₂O₂ adducts have been previously reported to react directly with acetone,^43,44 clean and fast conversion of 3 with H₂O₂ to 1 in Figure 10 suggests the reaction of H₂O₂ and the Cu complex with acetone may be negligible under the present reaction conditions. The observed first-order rate constant (k_obs) increased linearly with increasing concentration of H₂O₂ (Figure 10b). The second-order rate constant (k₂) was determined to be 2.5 × 10³ M⁻¹ s⁻¹ at 298 K, which is much larger than the k_obs/[1] values (vide supra). Thus, under the catalytic conditions, the H₂O₂ produced, if any, is rapidly reduced by [Cu₂^I(XYLOH)]²⁺, contributing to the overall four-electron four-proton reduction of O₂.

Scheme 6.

Figure 10.

(a) UV-vis spectral changes observed in the reaction of H₂O₂ (0.033 mM) with [Cu₂^I(OH)]²⁺ (0.025 mM) in acetone at 298 K. Inset shows the time profile monitored at 378 nm due to the formation of [Cu₂^II(XYLO)(OH)]²⁺. (b) Plot of k_obs vs [H₂O₂] in the reaction of H₂O₂ with [Cu₂^I(OH)]²⁺ (0.025 mM) in acetone at 298 K.

The overall catalytic cycle is summarized in Scheme 7. The protonation of [Cu^II₂(XYLO)(OH)]²⁺ (1) results in formation of [Cu^II₂(XYLOH)]⁴⁺ (2), which can be reduced by two equiv of Fc and Me₂Fc to produce fully reduced dicopper(I) complex [Cu^I₂(XYLOH)]²⁺(3). The O₂-binding to (3) to produce hydroperoxo complex [Cu^II₂(XYLO)(OOH)]²⁺ (4) is an equilibrium process. At 298 K, only a small portion of [Cu^I₂(XYLOH)]²⁺ is converted to [Cu^II₂(XYLO)(OOH)]²⁺, the concentration of which is proportional to the O₂ concentration. [Cu^II₂(XYLO)-(OOH)]²⁺ which is formed undergoes further reduction by PCET from Fc and Me₂Fc, leading to the four-electron reduction of O₂. In this final two-electron peroxide reduction, the first PCET reduction may be the rate-determining step followed by the much faster second PCET reduction to produce H₂O. In such a case, the overall catalytic rate is proportional to concentrations of 1, O₂, HClO₄ and electron donors (Fc and Me₂Fc) as observed in Figures 2 and S4-S6. The PCET reduction of [Cu^II₂(XYLO)(OOH)]²⁺ may compete with the protonation of [Cu^II₂(XYLO)(OOH)]²⁺ (4) to generate H₂O₂ (Scheme 6); however, H₂O₂ thus produced is rapidly reduced to H₂O by [Cu^I₂(XYLOH)]²⁺ (vide supra).

Scheme 7.

According to Scheme 7, the equilibrium between [Cu^I₂(XYLOH)]²⁺ (3) with O₂ and [Cu^II₂(XYLO)(OOH)]²⁺ (4) at 298 K lies to the side of [Cu^I₂(XYLOH)]²⁺ (vide infra), the [Cu₂^II(XYLOH)]⁴⁺ complex is being converted to [Cu^I₂(XYLOH)]²⁺ during the catalytic reaction but [Cu₂^II(XYLOH)]⁴⁺ (2) may be regenerated after the completion of the catalytic reaction when all ferrocenes were consumed. This was confirmed as the change in the EPR spectra as shown in Figure 11, where the EPR signal due to [Cu₂^II(XYLOH)]⁴⁺ (2) observed before the reaction disappeared during the catalytic reaction but reappeared after completion of the reaction without exhibiting any decomposition.

Figure 11.

EPR spectra of (a) [Cu^II₂(XYLO)(OH)](PF₆)₂ (1) (0.040 mM) with HClO₄(40 mM) and (b, c) the reaction solution of 1 (0.040 mM) with Me₂Fc (10 mM) in the presence of HClO₄(40 mM) in O₂-saturated acetone at 298 K [(b) during the reaction and (c) after completion of the reaction]. Spectra were recorded at 20 K. The experimental parameters: microwave frequency = 9.654 GHz, microwave power = 1.0 mW and modulation frequency = 100 kHz.

The kinetic results (as described by eqs 3 and 4) and the absence of the EPR signal due to 2 during the catalytic reaction indicate that the reaction of 3 with O₂, Me₂Fc and H⁺ involves all of these species in the rate-determining step. As described above, we have also separately shown that 4 is formed by the reaction of 2 and O₂ at 193 K (Figure 7).⁴⁷ At 298 K, the equilibrium for the formation of 4 lies to 2 (vide supra), when the concentration of 4 may be proportional to [O₂]. Then, the PCET reduction of 4 by Me₂Fc may compete with the formation of H₂O₂ by the protonation of 4. At 193 K, the protonation of 4 to produce H₂O₂ may be the major pathway as indicated by the results in Figure 9. At 298 K, however, the PCET reduction of 4 by Me₂Fc may be the major pathway when the rate of formation of Me₂Fc⁺ is derived as given by eq 5, where k_PCET is the rate constant of PCET reduction of 4

$$\mathrm{d}\left[{\mathrm{Me}}_2{\mathrm{Fc}}^{+}\right]∕\mathrm{d}t=k_{\mathrm{PCET}}K\left[{\mathrm{Me}}_2\mathrm{Fc}\right]\left[1\right]\left[{\mathrm{O}}_2\right]\left[{\mathrm{H}}^{+}\right]$$ (5)

and K is the equilibrium constant of formation of 4 with O₂ from 3. The derived kinetic equation (eq 5) agrees with the experimental observations in eqs 3 and 4. If the protonation of 4 to produce H₂O₂ was the major pathway at 298 K, the rate would not be dependent of Me₂Fc, because electron transfer from Me₂Fc to 2 was shown to be too fast to be involved in the rate-determining step. Because the rate is proportional to [Me₂Fc], the rate-determining step must be the PCET reduction of 4 by Me₂Fc. It should be noted that the protonation of 1 is completed in the presence of HClO₄ (> 10 mM) as shown in Figure 2, when the linear dependence of the rate on concentration of HClO₄ (> 10 mM) in Figure 2c results from the rate-determining PCET reduction of 4.

Conclusion

A dinuclear copper(II) complex ([Cu^II₂(XYLO)(OH)]²⁺) acts as an efficient catalyst for the four-electron reduction of O₂ by Me₂Fc and Fc with HClO₄ in acetone as shown in Scheme 7. The hydroxide group as well as the phenoxo group of [Cu^II₂(XYLO)(OH)]²⁺ (1) were protonated with HClO₄ to produce [Cu^II₂(XYLOH)]⁴⁺ (2) which can be reduced by Me₂Fc and Fc to produce [Cu^I₂(XYLOH)]²⁺ (3). The dinuclear Cu(I) complex [Cu^I₂(XYLOH)]²⁺ (3) reacts with O₂ to produce the hydroperoxo complex ([Cu^II₂(XYLO)(OOH)]⁺ (4)), and this is followed by PCET reduction, leading to the catalytic four-electron reduction of O₂ by Fc and Me₂Fc.

It is instructive to compare and contrast the chemistry described here with that previously reported,³⁸ both with exactly the same catalyst, [Cu^II₂(XYLO)(OH)]²⁺ (1) (Schemes 1 & 8) but having very differing behaviors. As indicated in the summary in Scheme 8, 1 is quite difficult to reduce, but in the presence of HClO₄, the bridging hydroxide ligand is displaced (as H₂O) and now, reduction to a dicopper(I) (or a mixed-valent form, [Cu^IICu^I(XYLO)]²⁺)³⁸ is possible; it is this/these forms which are required for O₂-binding and initial reduction to the peroxide level. With HOTF, however, the phenoxo O-atom still bridges the Cu(II) ions leaving the redox potential negative enough to require stronger reductants such as Me₈Fc or Fc*. A key coordination chemistry aspect is that HClO₄ as proton source is strong enough to break the phenoxide bridge between copper ions, allowing facile reduction of the Cu(II) ions with Me₂Fc or even Fc itself;the complex produced, [Cu^II₂(XYLOH)]⁴⁺ (2) now has Cu(II) ions possessing only N₃ bis[(2-(2-pyridyl)ethyl)amine chelation (Scheme 8). Thus, the change to perchloric acid facilitates a drop in effective overpotential of ~ 0.30 V, or more (Scheme 8).

Scheme 8.

Perchloric acid effects another dramatic change; the reaction mechanism switches from the catalytic two-electron two-proton reduction of O₂ to H₂O₂ with HOTF, to the catalytic four-electron four-proton reduction of O₂ to water with HClO₄. Firstly, for the HOTF case, dicopper(II) reduction is rate limiting, but with HClO₄, PCET reduction/protonation of [Cu^II₂(XYLO)-(OOH)]²⁺ (4) is the rate-determining step. Secondly, note that in both systems, the hydroperoxo complex 4 is the key oxygen-intermediate which is formed. HOTF readily protonates off the bound –OOH ligand giving H₂O₂, but it is not strong enough to allow PCET hydroperoxide reduction to water. Perchloric acid does facilitate the latter hydroperoxide reductive cleavage to water, accounting for the differing stoichiometries of catalytic O₂-reduction chemistry.

Although the mechanism of the PCET reduction of 4 by Me₂Fc has yet to be clarified, the chemistry described here provides the first example of four-electron reduction of O₂ by weaker one-electron reductants than Fc* such as Fc and Me₂Fc in the presence of HClO₄ in acetone by a copper complex acting as a catalyst. The present study opens a new approach to achieve less overpotential which is more desirable and more energy efficient for the catalytic four-electron reduction of O₂ using copper complexes.

Experimental Section

Materials

Grade quality solvents and chemicals were obtained commercially and used without further purification unless otherwise noted. Decamethylferrocene (Fc*) (97%), 1,1′-dimethylferrocene (Me₂Fc), ferrocene (Fc), hydrogen peroxide (30%) and HClO₄ (70%) were purchased from Aldrich Co., U.S., and NaI (99.5%) was from Junsei Chemical Co., Japan. Acetone was purchased from JT Baker, U.S., and used either without further purification for non-air-sensitive experiment or dried and distilled under argon and then deoxygenated by bubbling with argon for 30–45min and kept over activated molecular sieve (4 Å) for air-sensitive experiments.⁴⁵ Preparation and handling of air-sensitive compounds were performed under Ar atmosphere (< 1 ppm O₂ and < 1 ppm H₂O) in a glove box (Korea Kiyon Co., Ltd.). The copper complexes, [Cu^I₂(XYLH)(CH₃CN)₂](PF₆)₂ (XYLH = m-xylene-linked bis[(2-(2-pyridyl)ethyl)amine]),⁴⁶ which is a precursor complex, [Cu^II₂(XYLO)(OH)](PF₆)₂ (1) and [Cu^I₂(XYLOH)](PF₆)₂, were prepared according to the literature procedures.⁴⁷

Instrumentation

UV–vis spectra were recorded on a Hewlett-Packard 8453 diode array spectrophotometer equipped with a UNISOKU Scientific Instruments Cryostat USP-203A for low-temperature experiments or an UNISOKU RSP-601 stopped-flow spectrometer equipped with a MOS-type highly sensitive photodiode array. Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) measurements were performed on an ALS 630B electrochemical analyzer, and voltammograms were measured in deaerated acetone containing 0.20 M TBAPF₆ as a supporting electrolyte at room temperature. A conventional three-electrode cell was used with a gold working electrode (surface area of 0.3 mm²), and a platinum wire was the counter electrode. The gold working electrode was routinely polished with BAS polishing alumina suspension and rinsed with acetone before use. The potentials were measured with respect to the Ag/AgNO₃ (10 mM) reference electrode and were converted to values vs SCE by adding 0.29 V.⁴⁸ All electrochemical measurements were carried out under an atmospheric pressure of nitrogen. X-band EPR spectra were recorded at 5 K or 20 K using an X-band Bruker EMX-plus spectrometer equipped with a dual mode cavity (ER 4116DM). Low temperature was achieved and controlled with an Oxford Instruments ESR900 liquid He quartz cryostat with an Oxford InstrumentsITC503 temperature and gas flow controller. The experimental parameters for EPR spectra were as follows: microwave frequency = 9.646 GHz at 5K and 9.654 GHz at 20 K, microwave power = 1.0 mW, modulation amplitude = 10 G, gain = 5 × 10³, modulation frequency = 100 kHz, time constant = 81.92 ms, and conversion time = 81.00 ms.

Kinetic Measurements

The UV-vis spectral changes were recorded on a Hewlett-Packard 8453 diode array spectrophotometer equipped with Unisoku thermostatted cell holder for low temperature experiments. Rate constants in the oxidation reaction of ferrocene derivatives by O₂ in the presence of catalytic amount of 1 and excess amount of perchloric acid (HClO₄) in acetone at 298 K were determined by monitoring the appearance of the absorption band due to the corresponding ferrocenium ions (Fc⁺: λ_max = 620 nm, ε_max = 430 M⁻¹ cm⁻¹. Me₂Fc⁺ : λ_max = 650 nm, ε_max = 360 M⁻¹ cm⁻¹. Fc*⁺: λ_max = 780 nm, ε_max = 520 M⁻¹ cm⁻¹). The limiting concentration of O₂ in an acetone solution was prepared by a mixed gas flow of O₂ and Ar. The mixed gas controlled by using a gas mixer (SMTEK, Korea), which can mix O₂ and Ar gases at a certain pressure and flow rate for the rate measurements with different concentrations of O₂ in Figure 2c. The rate constants as determined from the first-order plots are provided in the SI (Table S1).

Spectroscopic Measurements

The amount of H₂O₂ was determined by titration with iodide ion.⁴⁹ The diluted acetone solution of the reduced product of O₂ was treated with an excess of NaI. The amount of I₃⁻ formed was then quantified using its visible spectrum (λ_max = 361 nm, ε = 2.5 × 10⁴ M⁻¹ cm⁻¹).

Low-Temperature Experiments Concerning the Generation of [Cu^II₂(XYLO)(OOH)]²⁺ (4)

Under an argon atmosphere within a glove box, [Cu^I₂(XYLOH)](PF₆)₂ (0.11 mM) was dissolved in 3.0 mL of O₂-free acetone. The cuvette was fully sealed with a septum and quickly removed from the glove box and cooled to −80 °C in the UV-vis spectrophotometer equipped with a thermostatted cell holder. O₂ was gently bubbled through the reaction solution, the formation of the hydroperoxo species was followed by the change in the absorbance at 395 nm.