Metal-Free Homogeneous O₂ Reduction by an Iminium-Based Electrocatalyst

Abstract

The oxygen reduction reaction (ORR) is important for alternative energy and industrial oxidation processes. Herein, an iminium-based organoelectrocatalyst (im⁺) for the ORR with trifluoroacetic acid as a proton source in acetonitrile solution under both electrochemical and spectrochemical conditions using decamethylferrocene as a chemical reductant is reported. Under spectrochemical conditions, H₂O₂ is the primary reaction product, while under electrochemical conditions H₂O is produced. This difference in selectivity is attributed to the interception of the free superoxide intermediate under electrochemical conditions by the reduced catalyst, accessing an alternate inner-sphere pathway.

The increasing atmospheric carbon dioxide (CO₂) concentration has had detrimental impacts on our environment and creates a drastic need for alternative energy processes. The oxygen reduction reaction (ORR) is important in fuel cells and alternative energy devices such as zinc-air batteries, in addition to a green alternative for H₂O₂ production.^1,2 Researchers have focused on open-shell transition-metal complexes as catalysts for this reaction due to their often facile reactivity with the triplet ground state of dioxygen (O₂). There has also been some advancement in the development of carbon-based catalysts for the ORR.²⁻⁸ Comparatively, the use of homogeneous organic molecules for catalytic ORR has been less widely studied since reduced oxygen species (ROSs) formed as intermediates (e.g., superoxide O₂^•–) can degrade organic molecules.

In 2020, Karimi et al. reported the ORR activity of carbenium dications using decamethylferrocene (Cp*₂Fe) as a chemical reductant with methanesulfonic acid as a proton source (Figure 1).⁹ The reduced carbene radicals were found to rapidly react via an inner-sphere mechanism to form an intermediate peroxide, which is protonated to release H₂O₂. Recently, Tanjedrew et al.reported imidazole-benzimidazole electrocatalysts for ORR under aqueous conditions.¹⁰ They proposed that O₂ binds to the reduced catalyst to form a superoxide species, which is further reduced and protonated to produce H₂O₂. Homogeneous organic species have also been shown to be active for the outer-sphere reduction of O₂, with free O₂^•– as an intermediate. Electrocatalytic ORR by an outer-sphere mechanism was first reported in 1985 by Andrieux et al. using methylviologen to generate H₂O₂ in acidic dimethylsulfoxide (DMSO).¹¹ Outer-sphere electron transfer to generate O₂^•– was followed by protonation to HO₂^•, which was subsequently reduced by the regenerated methyl viologen monocation to HO₂^– and then protonated to form H₂O₂. Following this, a 1993 study by Audebert and Hapiot found that substituted 9-(4-X-phenyl)-N-methylacridinium salts in acidic DMSO also effectively reduce O₂ to H₂O₂ via an analogous mechanism.¹²

Figure 1.

Summary of previously reported organic-based catalysts for the ORR and the catalyst (im⁺) described here.

Previously, 3,4-dihydro-2,4,4-trimethyl-1-(trifluoromethyl)isoquinolinium tetrafluoroborate (im⁺, Figure 1) was reported to be a hydroxylation catalyst with H₂O₂ as an oxidant.^13,14 It was hypothesized that this iminium could mediate outer-sphere O₂ reduction similar to related organic cations and that its intrinsic stability to H₂O₂ would be beneficial to catalyst stability.¹⁵ Catalytic and mechanistic experiments reveal that the iminium salt is an efficient catalyst for ORR to H₂O and H₂O₂ in acetonitrile (MeCN) with trifluoroacetic acid (TFAH) as the proton source. Under spectrochemical conditions with Cp*₂Fe as a chemical reductant in solution, O₂ is quantitatively reduced via an outer-sphere mechanism to H₂O₂. Under electrochemical conditions, O₂^•– is intercepted by the reduced iminium in solution, accessing an inner-sphere mechanism to quantitatively produce H₂O. The difference in selectivity is proposed to be regulated by the simultaneous availability of O₂^•– and the reduced iminium in higher concentrations in the reaction-diffusion layer.

Im⁺ was analyzed by cyclic voltammetry (CV) with tetrabutylammonium hexafluorophosphate (TBAPF₆) as the supporting electrolyte in MeCN (Figure 2). Under Ar saturation conditions, there is a diffusion-limited irreversible reduction feature, which is attributed to the reduction of the iminium to a carbon-based radical species (im⁰, see below) at E_p = −0.82 V vs Fc⁺/Fc. Due to irreversibility at scan rates below 2 V/s, the E_1/2 was estimated to be −0.77 V vs Fc⁺/Fc by taking the first derivative of the current density (Figures S1 & S5).¹⁶ This irreversible feature is consistent with a radical–radical dimerization (RRD) mechanism occurring between two equivalents of im⁰, as evidenced by an evaluation of peak potential dependence on both scan rate and concentration (Figures S1 & S2).¹⁷ Upon saturation of the solution with O₂, there is an observed 150 mV positive potential shift of the irreversible reduction feature to E_p = −0.67 V vs Fc⁺/Fc, suggestive of a strong binding interaction. Interestingly, variable concentration and scan rate studies suggested a shift to a radical-substrate dimerization (RSD) mechanism, based on the observed dependence of E_p on scan rate and concentration under these conditions (Figures S3 & S4). This observation is proposed to correlate to a favorable interaction between im⁰ and O₂^•– at reducing potentials.¹⁷ Addition of TFAH (pK_a(MeCN) = 12.65)¹⁸ as a proton source under Ar saturation conditions resulted in an increase in current (Figures 2 & S5); the dependence of E_p on scan rate and concentration suggests an RSD mechanism, implying that a reaction with TFAH precedes the dimerization of the im⁰ radical (Figures S6 & S7). However, under O₂ saturation in the presence of TFAH, a large increase in current density is again observed 150 mV positive of the Faradaic response, consistent with catalytic O₂ reduction, implying that dimerization is not relevant under these conditions. A CV rinse test and control experiments demonstrated that the observed catalytic response of im⁺ is homogeneous (Figure S8).

Figure 2.

CVs of im⁺ under Ar and O₂ saturation conditions with and without acid. Conditions: 1.3 mM im⁺, 0.1 M TBAPF₆/MeCN; glassy carbon working electrode, glassy carbon counter electrode, Ag/AgCl pseudoreference electrode; 100 mV/s scan rate.

Independently varying the concentrations of im⁺, O₂ and TFAH revealed a first-order concentration dependence on im⁺, O₂, and TFAH in the current response (eq 1, Figures S9–S11). As the concentration of im⁺ decreases, there is an observed negative potential shift (Figure S9), which conforms to the behavior expected for an outer-sphere electron transfer followed by rapid catalytic reaction steps.¹¹ As the concentration of O₂ decreases, a decrease in the catalytic wave at E_p = −0.67 V vs Fc⁺/Fc is observed, as well as the recovery of the im⁺/im⁰ redox feature at E_p = −0.82 V vs Fc⁺/Fc. The reappearance of the intrinsic redox feature suggests the system is operating under total catalysis conditions, where O₂ is rapidly consumed within the reaction-diffusion layer and excess im⁺ is available for reduction at more negative potentials. The recovery of the Faradaic redox response has also been observed for ORR mediated by methyl viologen and phenylacridinium salts, which are proposed to have an outer-sphere mechanism.^11,12

1

Rotating ring-disk electrode (RRDE) methods with a glassy carbon disk and roughened gold ring¹⁹ were used to determine the selectivity of ORR by im⁺ under electrochemical conditions. Under air saturation, this system was found to be 92.6 ± 1.3% selective for H₂O (see Supporting Information, Figures S14–S16). Control CV studies with added urea·H₂O₂ showed a slight decrease in current density (Figures S12 & S13) under Ar saturation conditions without a shift to positive potentials, which suggests a relatively slower reaction between im⁰ and H₂O₂. Consistent with this, under O₂ saturation the catalytic current density is recovered, as is the shift to more positive potentials, confirming that im⁰ preferably reacts with O₂ over H₂O₂.

Catalytic ORR activity of im⁺ was then studied by stopped-flow UV–vis methods using Cp*₂Fe as a chemical reductant and TFAH as the proton source. The growth of the spectral handle of [Cp*₂Fe]⁺ at 780 nm was monitored to extract the kinetic parameters for the reaction (Figure 3). The rate law under these conditions (eq 2) was determined by independently varying the concentration of im⁺, TFAH, O₂, and Cp*₂Fe (Figures S18–S21). These studies revealed that ORR shows a first-order concentration dependence on [im⁺] and is independent of TFAH, O₂, and Cp*₂Fe concentration, indicative of saturation kinetics at low catalyst concentration (4 μM). A Ti(O)SO₄ colorimetric assay was used to determine the selectivity of ORR by im⁺ under spectrochemical conditions, finding that this system is 102 ± 8.4% selective for H₂O₂ (n_cat = 2; Figure S22), in contrast to the electrochemical studies. Additional control testing showed no degradation of H₂O₂ by disproportionation or catalytic H₂O₂ reduction with Cp*₂Fe as the reductant (Figures S23 & S24). Therefore, based on the apparent rate law, the slope of variable [im⁺] studies (Figure S18) could be used to estimate an apparent TOF of 6.66 × 10³ s^–1.

2

Figure 3.

Formation of [Cp*₂Fe]⁺ at 780 nm from the ORR catalyzed by im⁺ (red trace) and control (black trace). Concentrations: im⁺ = 4 μM, TFAH = 25 mM, O₂ = 4.05 mM, Cp*₂Fe = 1 mM; control: no im⁺.

To evaluate the thermodynamics of the reaction, computational studies of likely intermediates during O₂ reduction were undertaken (see Supporting Information). Evaluation of spin density showed that the radical character of im⁰ is localized on the imine C atom with some electron density shared by the N atom (Figure S25). Positioning the neutral radical and O₂ within a sufficient radius for a covalent interaction did not result in bond formation; however, electron transfer between the two occurred resulting in the formation of superoxide O₂^•– and im⁺, although free energy of the reaction was endergonic overall (+8.8 kcal/mol). In the presence of TFAH (and considering the exergonicity of homoconjugation between TFAH and trifluoroacetate), the formation of protonated superoxide HO₂^• is favorable by −33.2 kcal/mol. Subsequent disproportionation of two equiv of HO₂^• to form O₂ and H₂O₂ is comparably favorable at −33.1 kcal/mol. The reduction potential of HO₂^• is estimated to be 0.58 V negative of the im^+/0 couple, excluding outer-sphere reduction.

Since the minimal electrocatalytic current observed with H₂O₂ in comparison to that with O₂ and im⁺ was not observed to reduce H₂O₂ under spectrochemical conditions (Figure S24), subsequent calculations focused on alternative pathways to produce water. The reaction between 1 equiv of im⁰ and O₂^•– to produce a monoanionic C-bound end-on peroxide species is exergonic by −21.7 kcal/mol. Protonation by TFAH to generate a neutral hydroperoxide is further downhill by −42.0 kcal/mol, considering homoconjugation. The alternative generation of this intermediate by the reaction of im⁰ and HO₂^• is favorable by −30.6 kcal/mol.

Protonation of the hydroperoxide to generate an oxaziridinium with water coproduct is downhill by an additional −18.2 kcal/mol. The reduction potential of the oxaziridinium is calculated to be approximately 0.21 V more negative than im^+/0; however, subsequent protonation to produce a cationic C–OH is favorable by −49.0 kcal/mol. Given the rate of catalysis observed electrochemically, it is probable that these steps occur as a single-proton-coupled electron transfer step, which would be favored overall by −44.2 kcal/mol (+1.92 V vs Fc⁺/Fc). Reduction of this cationic species is expected to be facile, with a calculated reduction potential of +1.34 V versus Fc^+/0 (Figure S26). The protonation of the resultant neutral C–OH group to generate water and reform im⁺ is then favorable, with an estimated free energy change of −10.4 kcal/mol.

Based on electrochemical, spectrochemical, and computation analyses, separate cycles for the reduction of dioxygen by im⁺ under electrochemical and spectrochemical conditions can be proposed (Scheme 1). Starting at i, an electron transfer to form carbon-centered radical species, ii, which reacts with O₂ to reform i and an equivalent of O₂^•–. The product O₂^•– is then protonated by TFAH to form two equivalents of HO₂^•, which favorably disproportionate to one equiv each of O₂ and H₂O₂.²⁰ Under spectroscopic conditions, the catalytic cycle closes here, as supported by the observed quantitative selectivity for H₂O₂ (Figures S22 & S27). Under electrochemical conditions, control experiments show that reactivity with H₂O₂ is slow relative to that with O₂, suggesting that these conditions have a divergent mechanistic pathway. Instead, it is proposed that under electrochemical conditions, the neutral radical ii is available in sufficient concentrations in the reaction-diffusion layer to bind available O₂^•–, which is supported by the RSD pathway observed in electrochemical studies (Figures S3 & S4). Based on these data and the empirically determined rate law, it is likely that species iv represents the resting state of the catalytic cycle, with the protonation reaction to generate v representing the rate-determining step (Figure S28).

Scheme 1. Proposed Catalytic Cycle for ORR by im⁺.

The change in accessible pathways can then be ascribed to the concentration differences under each reaction condition. The reaction-diffusion layer during electrochemical experiments is likely to contain both an excess of im⁰ because of rapid consumption of O₂ and suitable concentrations of superoxide from rapid outer-sphere reduction. Comparatively, under spectrochemical conditions, the relative concentrations of im⁰ and O₂^•– are significantly more dilute, allowing thermodynamically viable disproportionation pathways to generate H₂O₂. Outer-sphere reduction of HO₂^• by im⁰ is excluded based on the difference in calculated reduction potentials.

Here, catalytic ORR conditions for H₂O or H₂O₂ production with an iminium-based catalyst have been reported. The difference in accessible reaction pathways under electrochemical and spectrochemical conditions, where the primary product shifts from H₂O to H₂O₂, respectively, is the result of the relative available concentrations of the key im⁰ and O₂^•– intermediates available under the respective reaction conditions. Since catalysis is initiated by an outer-sphere electron transfer, the O₂/O₂^•– reduction potential of −1.29 vs Fc⁺/Fc in MeCN defines the overall ORR reaction mediated by im⁺.²¹ However, the favorable pre-equilibrium reaction between im⁰ and O₂^•– causes a positive potential shift from this redox couple, indicating that further optimization of the operating potential could be possible.²² The work described here reports a novel mechanism whereby the electrocatalytic reduction of O₂ occurs via both an inner-sphere and outer-sphere mechanism, resulting in product selectivity being controlled by the nature of electron delivery. Given that there are few known organoelectrocatalysts for the ORR,²³ mechanistic understanding will enable the development of additional examples as well as inform the development of new classes of doped carbons as heterogeneous catalysts.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.3c14549.

• Description of experimental details and methods, electrochemical data, and stopped-flow data (PDF)

• Computational Coordinates (XYZ)

Author Contributions

All authors have given approval to the final version of the manuscript.

The authors declare no competing financial interest.