Electrocatalytic reduction of protons to dihydrogen by the cobalt tetraazamacrocyclic complex [Co(N₄H)Cl₂]⁺: mechanism and benchmarking of performances

Abstract

The cobalt tetraazamacrocyclic [Co(N₄H)Cl₂]⁺ complex is becoming a popular and versatile catalyst for the electrocatalytic evolution of hydrogen, because of its stability and superior activity in aqueous conditions. We present here a benchmarking of its performances based on the thorough analysis of cyclic voltammograms recorded under various catalytic regimes in non-aqueous conditions allowing control of the proton concentration. This allowed a detailed mechanism to be proposed with quantitative determination of the rate-constants for the various protonation steps, as well as identification of the amine function of the tetraazamacrocyclic ligand to act as a proton relay during H₂ evolution.

---

Molecular cobalt complexes are popular and versatile catalysts for the electrocatalytic evolution of hydrogen.^1–5 Recently, the cobalt complex [Co(N₄H)Cl₂]⁺ (Cat1, Fig. 1) based on the tetraazamacrocyclic 2,12-dimethyl-3,7,11,17-tetraazabicyclo[11.3.1]heptadeca-1(17),2,11,13,15-pentaene ligand,⁶ described by Karn and Busch in 1966, has received increased interest^7–9 namely because this catalyst proves active and stable for the evolution of H₂ from fully aqueous solutions.^7–17 A study carried out under homogeneous conditions using chemical reductants or photochemical activation confirmed the superior activity of Cat1 in fully aqueous media,¹⁸ and X-ray absorption spectroscopic monitoring of a homogeneous photocatalytic system for H₂ evolution based on Cat1 indicated an ECEC mechanism (E = monoelectronic electrochemical reduction, C = protonation step) starting from the bisaqua Co(II) complex.^13,14 However, very few metrics are currently available to benchmark the catalytic activity of this compound. Cat1 was included in a benchmarking study for H₂-evolving electrocatalysts carried out in aqueous electrolyte¹⁹ but under quite acidic conditions likely to induce reductive degradation of the ligand during the test and formation of metallic particles responsible for the observed HER activity.^20,21 To gain quantitative information on the H₂ evolution mechanism mediated by Cat1, we therefore revisited the non-aqueous conditions investigated by Lau and coworkers,²² where it is easier to control the concentration and chemical potential of protons.²³

Fig. 1 Structure of Cat1.

The cyclic voltammogram of Cat1 (perchlorate salt) in CH₃CN (with 0.1 M ⁿBu₄NBF₄, Fig. 2 and S1‡) displays two quasi-reversible systems at −0.47 V (ΔE_p = 92 mV) and −0.96 V (ΔE_p = 86 mV) vs. Fc⁺/Fc, corresponding to the Co^III/II and Co^II/I redox processes, respectively, based on previous literature. Of note, we will formally use the Co^I notation throughout this article, while the electronic state could also correspond to a reduced N₄H ligand (π-radical anion) antiferromagnetically coupled to a low-spin Co(II) ion.²⁴ A third ligand-centered process is observed at −1.89 V (ΔE_p = 69 mV) vs. Fc⁺/Fc.²² Upon addition of p-cyanoanilinium tetrafluoroborate (pK_a = 7.0 in CH₃CN)²⁵ acting as a proton source, a catalytic wave develops on the top of the Co^II/I wave as previously described.²² This electrocatalytic behavior corresponds to H₂ evolution with >90% faradaic yield²² and nicely mirrors the one observed for Cat1 in mildly acidic aqueous solution.^11,13,22 The addition of stronger acids such as HBF₄·Et₂O or CF₃SO₃H also triggers H₂ evolution catalysis but is detrimental to the stability of Cat1 at high concentration. Cat1 is unable to catalyze the reduction of acids with higher pK_a values, starting with p-toluenesulfonic acid (pK_a = 8.3 in CH₃CN).²⁶

Fig. 2 Top: cyclic voltammograms of Cat1 (0.5 mM) in CH₃CN (+0.1 M ⁿBu₄NBF₄) recorded at a glassy carbon electrode (1.6 mm diameter) in the absence (black trace) and in the presence of 5.0 (red trace), 10.0 (green trace), 12.5 (navy trace) and 15.0 mM (cyan trace) p-cyanoanilinium tetrafluoroborate; scan rate: 100 mV s⁻¹. A control voltammogram of 15 mM p-cyanoanilinium tetrafluoroborate without Cat1 (gray dashed trace) is shown for comparison; bottom: cyclic voltammograms of Cat1 (0.5 mM) in CH₃CN (+0.1 M ⁿBu₄NBF₄) recorded at a glassy carbon electrode in the absence (black) and in the presence (red) of 5 mM p-cyanoanilinium tetrafluoroborate_; scan rate: 0.1 (solid) and 10 V s⁻¹ (dashed). See Fig. S5‡ for control voltammograms without catalyst under similar conditions.

Noteworthily, the addition of acid also affects the Co^III/II system, which shifts to more positive potentials and partly loses reversibility. ¹H NMR experiments confirmed that the Co^III form of Cat1 is not protonated under these conditions (Fig. S2‡), neither does p-cyanoaniline coordinate to any of the Co^III, Co^II and Co^I states of Cat1 (Fig. S3‡) in the absence of acid. However, when 5 equivalents of p-cyanoanilinium tetrafluoroborate is added to an electrochemically-generated solution of the Co^II form of Cat1, the EPR signal is significantly changed, from a broad signal centered at g₁ = 2.240, g₂ = 2.130 and g₃ = 2.004 to a better resolved spectrum characteristic for a low spin d⁷ (S = ½) electronic configuration of Co (Fig. 3). It furthermore displays a 5-line superhyperfine structure with an intensity ratio of 1 : 2 : 3 : 2 : 1 and a coupling constant of 45 Hz in line with the coordination of two equivalent nitrogen-based ligands. Of note, a similar behavior was observed when the Co^II form of Cat1 was prepared by chemical reduction with cobaltocene instead of exhaustive bulk electrolysis (Fig. S4‡). No such change is observed when p-cyanoaniline is added (Fig. S4‡). When HBF₄·Et₂O is used as proton source instead of p-cyanoanilinium tetrafluoroborate, this superhyperfine structure is changed to a 3-line structure with a coupling constant of 40 Hz (Fig. 3), suggesting the coordination of a single nitrogen-based axial ligand, CH₃CN being the only plausible one under these conditions. In the former case, coordination of CH₃CN and p-cyanoanilinium (or p-cyanoaniline generated in situ upon protonation of the complex) can be considered without being possible to discriminate one from the other at the EPR level. Taken all together, these observations suggest (i) protonation of the ligand in Cat1 occurs upon reduction from the Co^III to the Co^II state in the presence of acid; (ii) at the same time, chloride axial ligands are displaced for nitrogen ligands; (iii) the nature and number of axial ligands depend on the nature of the acid employed. It should also be underlined that the binding of one vs. two ligands in the cobalt + II oxidation state likely depends on a subtle balance of their donating ability, as previously stated for Cat1 (ref. 11) and also found for cobaloximes and related cobalt diimine–dioxime complexes.^27–30 A possible structure of Co^IILH, in line with a previous proposition,¹³ is shown in Scheme 1.

Fig. 3 CW X-band EPR spectra (9.65 GHz) of the electrochemically generated Co^II form of Cat1 (0.5 mM) as prepared (top), with 5 eq. of the acids p-cyanoanilinium tetrafluoroborate (middle), HBF₄ (bottom) and their respective simulations in red. Experimental conditions: 30 K, 1 mW microwave power, 1600 G field sweep. Simulation parameters are reported in Table S1.‡

Scheme 1 Proposed ECEC mechanism for H₂ evolution mediated by Cat1. L and L′ indicate acetonitrile or p-cyanoaniline.

Although the formation of the protonated Co^II derivative of Cat1 falls into the proton-coupled electron transfer classification, the cathodic peak potential does not follow the increase of 29 mV per decade of acid concentration expected for an irreversible EC process with a fully displaced protonation equilibrium (e.g. in the pure kinetic KP zone);³¹ rather the cathodic peak potential rapidly shifts to a new value upon addition of acid and then keeps this value unchanged upon further addition (Fig. 2). This behavior is characteristic of the extraordinary kinetic (KE) zone,³¹ with fast protonation of the Co^II species so that, even with few equivalents of acid added, the new wave is observed at a potential close to the apparent standard potential of the Co^III/Co^IILH couple, thus with a ∼250 mV shift compared to the original Co^III/Co^II couple. Recording the cyclic voltammograms at a significantly higher scan rate (10 V s⁻¹) did not allow approaching the pure kinetic KP zone although a slightly more progressive evolution of the cathodic peak potential was observed with an increase of ∼200 mV per decade of acid concentration (Table S2‡). Simulations using the DigiElch software allowed to reproduce the potential shift of such an EC process at both 0.1 and 10 mV s⁻¹ using a protonation equilibrium constant higher than 10⁴ and a bimolecular protonation rate of 10⁷ mol⁻¹ L s⁻¹, suggesting that the amine moiety can potentially act as a proton relay during catalysis.³²

Protonation of the Co^II state should also alter the standard potential of the Co^II/Co^I couple. At 100 mV s⁻¹, this redox process is hidden by the catalytic wave it triggers. However, measuring cyclic voltammograms at 10 V s⁻¹ enabled the catalysis to be outrun and a reversible wave to be recovered centred at −0.89 V vs. Fc⁺/Fc (Fig. 2 bottom) that likely corresponds to the actual Co^IILH/Co^ILH couple responsible for catalysis. Remarkably, raising acid concentration, we found that catalysis proceeds under pure kinetic conditions, where a catalytic current plateau, independent of the scan rate, is observed (Fig. 4) because substrate consumption is negligible (pure kinetic KS zone).^31,33 Under these conditions, the mid-wave potential of the catalytic process is found at −0.87 V vs. Fc⁺/Fc. Varying the catalyst concentration for a given acid concentration showed a linear dependence of the catalytic current with the catalyst concentration (Fig. S6‡). The catalytic plateau current also linearly varies with the acid concentration (Fig. S7‡).

Fig. 4 Cyclic voltammograms of Cat1 (1.6 mM) in CH₃CN (+0.1 M ⁿBu₄NBF₄) recorded at a glassy carbon electrode (1.6 mm diameter) in the presence of 60 mM p-cyanoanilinium tetrafluoroborate at various scan rates, scan rate ranging from 200 to 500 mV s⁻¹ and with ohmic drop compensation. A control voltammogram of 60 mM p-cyanoanilinium tetrafluoroborate without Cat1 at 100 mV s⁻¹ is shown for comparison.

Together these data are consistent with a mechanism for H₂ evolution catalyzed by Cat1 following an ECEC reaction scheme (Scheme 1), where E and C stand for electrochemical steps and chemical (i.e. protonation) steps and with the second reduction occurring at a potential more positive to that of the first one: reduction of the protonated Co^II (Co^IILH) state yields Co^ILH, which is further protonated to yield the Co^IIIHLH species. Further reduction of this derivative then produces the Co^II hydride Co^IIHLH species that is further protonated to evolve H₂ and regenerate the starting Co^IILH complex. Some of the distorded geometries displayed in Scheme 1 for transient intermediates involved in the catalytic cycle are approximate and we do realize that they do not correspond to ideal geometries for coordination complexes with corresponding electronic structures. Protonation of the amine residue of the macrocyclic ligand namely generates an ammonium moiety unable to coordinate the cobalt center, though with significant steric hindrance preventing for example the adoption of a perfect square planar (or octahedral) geometry favored by the d⁸ (or d⁶) configuration in the Co^ILH (or Co^IIIHLH) species, as demonstrated for the unprotonated Co^I (ref. 14) or Co^III-hydride species,³⁴ respectively. Of note this family of cobalt complexes can accommodate various coordination spheres as demonstrated by the heptacoordinated systems recently reported.^35,36

Eqn (1) gives the plateau current i_p for such a process, assuming that the two electrons required to complete catalytic turnover are transferred from the electrode to the catalyst.

where F is the Faraday constant, S is the geometric electrode surface area, C⁰_cat is the concentration of the catalyst and D_cat is the diffusion coefficient of the catalyst, determined to be 10⁻⁵ cm² s⁻¹ from the scan-rate dependence of the peak current of the first cathodic wave and the Randles–Sevcik equation (see ESI‡).

From plateau currents measured over a range of concentrations for both catalyst and acid (Fig. S6 and S7‡), we found a value of 5.3 ± 0.1 × 10³ mol⁻¹ L s⁻¹ for the second order rate constant k₂. The rate constant for the first protonation step k₁ of ECEC processes is accessible from the Foot-of-the-Wave Analysis (FOWA).^37,38 This analysis requires the knowledge of the apparent standard potential of the redox couple involved, i.e.Co^IILH/Co^ILH, which could be determined to −0.89 V vs. Fc⁺/Fc using high scan rates (Fig. 2 bottom). FOWA was performed at 3 different scan rates (100, 400 and 1000 mV s⁻¹) for two different acid concentrations (5 and 25 mM) and gave a value of 2.5 ± 0.4 × 10⁴ mol⁻¹ L s⁻¹ for the second order rate constant k₁ (Fig. S8‡). The higher value of k₁ compared to k₂ is in line with the midwave of the catalytic process being shifted positively compared to the standard potential of the Co^IILH/Co^ILH couple.³⁸ Analysis of this shift using eqn (2) for the data shown in Fig. S6 and S7‡ also leads to an average value of 2.5 × 10⁴ mol⁻¹ L s⁻¹ for the second order rate constant k₁, although with a much larger error margin.

The k₁ value can finally be confirmed from the analysis of cyclic voltammograms recorded at low acid concentration where the system belongs to the “total catalysis” regime (Fig. 5).^31,33 In this regime, catalysis is so fast that all the acid present in the diffusion layer is consumed during the sweep of the catalytic wave. As a consequence, the unprotonated Co(II) form of Cat1 is regenerated after catalysis and its reduction is observed at −0.96 V vs. Fc⁺/Fc.

Fig. 5 Cyclic voltammograms of Cat1 (2 mM) in CH₃CN (+0.1 M ⁿBu₄NBF₄) recorded at a glassy carbon electrode (1.6 mm diameter) in the presence of various concentrations of p-cyanoanilinium tetrafluoroborate at 100 mV s⁻¹: 0 (black), 4 (red), 8 (navy), 12 (magenta), 16 (green), 20 (cyan) mM.

Based on simulations, Dempsey and coworkers could propose eqn (3) to describe the variation of the catalytic peak potential E_p.³⁹

As the intensity of the catalytic wave in the “total catalysis” regime is controlled by the diffusion of the acid, the D_cat/D_AH ratio between the diffusion coefficient of the catalyst and p-cyanoanilinium cation can be obtained from the ratio between the catalytic peak current and the current of the monoelectronic Co^II/Co^I wave measured in the absence of acid, according to eqn (4).

Applying eqn (3) to the data shown in Fig. 5 yields a k₁ value of 2.05 ± 0.30 × 10⁴ mol⁻¹ L s⁻¹ for the second order rate constant k₁. This value is in good agreement with the one obtained by FOWA although neither k₁ nor k₂ comply with the condition of being greater than 10⁷ mol⁻¹ L s⁻¹ required for eqn (3) to be valid.³⁹

Interestingly, these data allow ruling out homolytic H₂ evolution mediated by Cat1. In such a mechanism, the first protonation step generating the Co^III–H derivative with the k₁ rate constant still exists but it is followed by reductive elimination of H₂ from two Co^III–H species with a k_d second-order rate constant. The linear dependency of the plateau current with the catalyst concentration (Fig. S6‡) rules out homolytic H₂ evolution from this Co^III–H derivative in the non-steady state where the rate-determining step is the reductive elimination step.^38,40 The identification of two distinct rate constants with k₁ being the largest one also allows to rule out steady state homolytic H₂ evolution with the rate-determining step being the formation of the Co^III–H derivative, therefore implying that k₁ < k_d. Finally, while the mechanism proposed in Scheme 1 does not formally exclude that formation of the Co^IIIHLH species may proceed through intramolecular protonation of the Co^I center of Co^ILH followed by reprotonation of the ligand, we believe that such a possibility is unlikely as Et₃NH⁺, an acid with a pK_a similar to that of the protonated N₄H₂⁺ ligand in Co^ILH, is unable to protonate Co^I complexes with similar value such as [Co^I(dmgBF₂)₂(CH₃CN)] (dmgH₂ = dimethylglyoxime) in CH₃CN.⁴¹

The mechanism proposed here for H₂ evolution differs from the one proposed by Llobet and Gimbert-Suriñach for aqueous conditions.^13,14 First we clearly evidenced that the Co^II form of Cat1 is protonated under the conditions investigated here. Ligand protonation also occurs in aqueous electrolyte, as shown by a 59 mV pH unit ⁻¹ shift of the redox process reported by Peters and coworkers.¹⁹ Second, we propose that H₂ is formed from a Co^II–H and not a Co^III–H intermediate. We recognize that the electrochemical responses of ECEC and ECCE sequences, both starting from a Co^II derivative and implying Co^II–H and Co^III–H active species, respectively, are similar. However DFT calculations clearly demonstrated that the standard potential of the Co^III–H/Co^II–H is more positive than that of the Co^II/Co^I couple,¹³ a feature also observed for cobaloximes and cobalt diimine–dioxime complexes for which the ECEC mechanism is now accepted. Importantly, the mechanism shown in Scheme 1 involves the Co^IIHLH species that was proposed by Llobet and Gimbert-Suriñach under photocatalytic conditions,¹³ therefore unifying the mechanistic understanding of this catalyst. DFT calculations previously indicated a near-thermoneutral intramolecular H₂ evolution step from this protonated hydride intermediate.¹³ Our analysis shows that this step is also the rate-determining one, which explains why catalysis is so fast. Still, the observation that the rate constant of this step is first order in acid concentration suggests that intramolecular H₂ formation is coupled with protonation, either in a concerted manner or through kinetic coupling with the fast reprotonation of the amine group of the ligand.

The maximal turnover frequency (TOF_max) of Cat1 therefore approximates k₂ × [acid] and a TOF_max value of 5.3 × 10³ s⁻¹ can be extrapolated for 1 M p-cyanoanilinium tetrafluoroborate concentration. Based on this value and an apparent equilibrium potential of the H⁺/H₂ couple of −0.47 V vs. Fc⁺/Fc at 1 M p-cyanoanilinium tetrafluoroborate concentration and taking homoconjugation into account,²⁶ we could derive the red trace in the catalytic Tafel plot^42,43 shown in Fig. 6. Interestingly, Cat1 displays significant catalytic activity (log(TOF/s⁻¹) > 1) at low overpotential values, a property shared by very few other molecular complexes including cobaloximes,^42,44 DuBois' nickel bisdiphosphine catalysts^42,45 and bis(thiosemicarbazone) cobalt⁴⁶ and nickel⁴⁷ complexes. Its overpotential requirement, estimated to be ∼400 mV at the catalytic half-wave potential (and corresponding to the inflexion point in the red catalytic Tafel plot on Fig. 6) is ∼100 mV higher than that of cobaloxime, cobalt diimine–dioxime⁴⁸ or DuBois' complexes. Importantly, the high TOF_max value also places Cat1 in an intermediate position between cobaloximes and DuBois' complexes, the two champion H₂-evolving molecular catalyst series identified so far in non-aqueous solvents.⁴² It is noteworthy that all three catalysts possess proton relays in their second-coordination sphere and the protonation of these relays is coupled with the metal-centered reduction step that sets the potential of the catalytic wave. The same conclusion can be reached with thiosemicarbazone nickel and cobalt complexes, with ligand-centered reduction occurring in these cases at quite positive potentials.^46,47

Fig. 6 Catalytic Tafel plots. Comparison of performances for H₂ evolution catalyzed by Cat1 in CH₃CN in the presence of 1 M p-cyanoanilinium (red line), with other catalysts reported in the literature: black: Fe^IITPP, DMF, Et₃NH⁺;^42,49 blue: Co^II(dmgH)₂py, DMF, Et₃NH⁺;⁴² green: , X = CH₂P(O)(OEt)₂, MeCN, DMFH⁺;^42,45 orange: 4-{bis[4-(p-methoxyphenyl)thiosemicarbazone]}-2,3-butane cobalt, DMF, Et₃NH⁺;⁴⁶ purple: [Co^II(bapbpy)Cl], DMF, Et₃NH⁺.²³

Conclusions

While Cat1 is becoming more and more popular as a molecular H₂-evolving catalyst for the design of aqueous systems, it is increasingly important to advance the understanding of its H₂ evolution catalysis mechanism¹³ and performance by providing insight into the catalytic steps involved. This is especially the case when considering structural modification⁵⁰ or molecular engineering⁵¹ in order to either enhance catalytic activity or stability¹⁶ or immobilise a catalytic centre for integration in photoelectrodes^7–9 or devices. In the societal context, both of these objectives are ultimately necessary to achieve industrial relevancy and technological maturity of hydrogen-producing electrolysers based on molecular catalysts made from earth-abundant elements.^2,4

In this study, new EPR evidence for the structure of the Co^II state of the catalyst and its dependence on the presence and nature of the acid has been presented, as well as NMR and cyclic voltammetry data indicating that fast protonation of the ligand occurs at the Co^II stage. Carrying on from previous work in the literature,^13,14,22 these results reconcile mechanisms at play under electrochemical and photochemical conditions. Furthermore, a proper utilisation of the most advanced analytical methods available for molecular catalysis of electrochemical reactions has allowed definitively ruling out a homolytic H₂ evolution mechanism and substantiated the proposed heterolytic ECEC mechanism, for which the rate constants for the two successive protonation steps could be determined. The rate-determining step has been confirmed as the intramolecular H₂ evolution step, surmised to be coupled to the second protonation, thereby regenerating a protonated ligand and thus acting as a proton relay in catalysis,³² which, interestingly, is at variance with the behaviour of the dioxime bridge in cobalt diimine–dioxime complexes.⁴⁸ Catalytic Tafel plots could be derived to enable the benchmarking of the H₂ evolution performance of Cat1 alongside other highly efficient catalysts and confirm its place on the podium.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work received funding from the French National Research Agency (Labex ARCANE, CBH-EUR-GS, ANR-17-EURE-0003), the European Research Council and European Commission's Seventh Framework Program (FP7/2007–2013) under grant agreement no. 306398 (project PhotocatH₂ode), the European Union's Horizon 2020 Research and Innovation program under grant agreement no. 765376 (eSCALED Marie Curie ITN project). C-B. L. acknowledges support from the Chinese Scientific Council for a post-doctoral scholarship.

References

1. V. Artero, M. Chavarot-Kerlidou and M. Fontecave, Angew. Chem., Int. Ed., 2011, 50, 7238–7266.
2. J. R. McKone, S. C. Marinescu, B. S. Brunschwig, J. R. Winkler and H. B. Gray, Chem. Sci., 2014, 5, 865–878.
3. N. Queyriaux, R. T. Jane, J. Massin, V. Artero and M. Chavarot-Kerlidou, Coord. Chem. Rev., 2015, 304–305, 3–19.
4. L. Tong, L. Duan, A. Zhou and R. P. Thummel, Coord. Chem. Rev., 2020, 402, 213079.
5. V. S. Thoi, Y. J. Sun, J. R. Long and C. J. Chang, Chem. Soc. Rev., 2013, 42, 2388–2400.
6. J. L. Karn and D. H. Busch, Nature, 1966, 211, 160–162.
7. S. Bold, J. Massin, E. Giannoudis, M. Koepf, V. Artero, B. Dietzek and M. Chavarot-Kerlidou, ACS Catal., 2021, 11, 3662–3678.
8. C. Nie, W. Ni, L. Gong, J. Jiang, J. Wang and M. Wang, J. Mater. Chem. A, 2019, 7, 27432–27440.
9. C. Nie, C. Liu, L. Gong and M. Wang, J. Mater. Chem. A, 2021, 9, 234–238.
10. R. Gueret, C. E. Castillo, M. Rebarz, F. Thomas, A. A. Hargrove, J. Pecaut, M. Sliwa, J. Fortage and M. N. Collomb, J. Photochem. Photobiol., B, 2015, 152, 82–94.
11. S. Varma, C. E. Castillo, T. Stoll, J. Fortage, A. G. Blackman, F. Molton, A. Deronzier and M. N. Collomb, Phys. Chem. Chem. Phys., 2013, 15, 17544–17552.
12. C. Gimbert-Surinach, J. Albero, T. Stoll, J. Fortage, M. N. Collomb, A. Deronzier, E. Palomares and A. Llobet, J. Am. Chem. Soc., 2014, 136, 7655–7661.
13. S. Grau, M. Schilling, D. Moonshiram, J. Benet-Buchholz, S. Luber, A. Llobet and C. Gimbert-Suriñach, ChemSusChem, 2020, 13, 2745–2752.
14. D. Moonshiram, C. Gimbert-Suriñach, A. Guda, A. Picon, C. S. Lehmann, X. Zhang, G. Doumy, A. M. March, J. Benet-Buchholz, A. Soldatov, A. Llobet and S. H. Southworth, J. Am. Chem. Soc., 2016, 138, 10586–10596.
15. C. H. Lee, D. Villagran, T. R. Cook, J. C. Peters and D. G. Nocera, ChemSusChem, 2013, 6, 1541–1544.
16. M. Sandroni, R. Gueret, K. D. Wegner, P. Reiss, J. Fortage, D. Aldakov and M. N. Collomb, Energy Environ. Sci., 2018, 11, 1752–1761.
17. R. Gueret, L. Poulard, M. Oshinowo, J. Chauvin, M. Dahmane, G. Dupeyre, P. P. Laine, J. Fortage and M. N. Collomb, ACS Catal., 2018, 8, 3792–3802.
18. S. Roy, M. Bacchi, G. Berggren and V. Artero, ChemSusChem, 2015, 8, 3632–3638.
19. C. C. L. McCrory, C. Uyeda and J. C. Peters, J. Am. Chem. Soc., 2012, 134, 3164–3170.
20. N. Kaeffer, A. Morozan, J. Fize, E. Martinez, L. Guetaz and V. Artero, ACS Catal., 2016, 6, 3727–3737.
21. V. Artero and M. Fontecave, Chem. Soc. Rev., 2013, 42, 2338–2356.
22. C. F. Leung, Y. Z. Chen, H. Q. Yu, S. M. Yiu, C. C. Ko and T. C. Lau, Int. J. Hydrogen Energy, 2011, 36, 11640–11645.
23. N. Queyriaux, D. Y. Sun, J. Fize, J. Pecaut, M. J. Field, M. Chavarot-Kerlidou and V. Artero, J. Am. Chem. Soc., 2020, 142, 274–282.
24. C. Römelt, T. Weyhermüller and K. Wieghardt, Coord. Chem. Rev., 2019, 380, 287–317.
25. A. M. Appel, S. J. Lee, J. A. Franz, D. L. DuBois, M. R. DuBois and B. Twamley, Organometallics, 2009, 28, 749–754.
26. V. Fourmond, P. A. Jacques, M. Fontecave and V. Artero, Inorg. Chem., 2010, 49, 10338–10347.
27. J. Niklas, K. L. Mardis, R. R. Rakhimov, K. L. Mulfort, D. M. Tiede and O. G. Poluektov, J. Phys. Chem. B, 2012, 116, 2943–2957.
28. T. Arcos, B. de Castro, M. J. Ferreira, M. Rangel and J. B. Raynor, J. Chem. Soc., Dalton Trans., 1994, 369–377.
29. A. Bakac and J. H. Espenson, J. Am. Chem. Soc., 1984, 106, 5197–5202.
30. A. Bhattacharjee, E. S. Andreiadis, M. Chavarot-Kerlidou, M. Fontecave, M. J. Field and V. Artero, Chem.–Eur. J., 2013, 19, 15166–15174.
31. J. M. Savéant, Elements of Molecular and Biomolecular Electrochemistry, Wiley, 2006.
32. J. M. Saveant, Angew. Chem., Int. Ed., 2019, 58, 2125–2128.
33. J. M. Saveant and K. B. Su, J. Electroanal. Chem., 1984, 171, 341–349.
34. A. F. M. M. Rahman, W. G. Jackson, A. C. Willis and A. D. Rae, Chem. Commun., 2003, 2748–2749.
35. J.-W. Wang, K. Yamauchi, H.-H. Huang, J.-K. Sun, Z.-M. Luo, D.-C. Zhong, T.-B. Lu and K. Sakai, Angew. Chem., Int. Ed., 2019, 58, 10923–10927.
36. R. Gueret, C. E. Castillo, M. Rebarz, F. Thomas, M. Sliwa, J. Chauvin, B. Dautreppe, J. Pécaut, J. Fortage and M.-N. Collomb, Inorg. Chem., 2019, 58, 9043–9056.
37. C. Costentin, S. Drouet, M. Robert and J.-M. Savéant, J. Am. Chem. Soc., 2012, 134, 11235–11242.
38. C. Costentin and J.-M. Savéant, ChemElectroChem, 2014, 1, 1226–1236.
39. E. S. Rountree, D. J. Martin, B. D. McCarthy and J. L. Dempsey, ACS Catal., 2016, 6, 3326–3335.
40. C. Costentin, H. Dridi and J.-M. Savéant, J. Am. Chem. Soc., 2014, 136, 13727–13734.
41. C. Baffert, V. Artero and M. Fontecave, Inorg. Chem., 2007, 46, 1817–1824.
42. V. Artero and J.-M. Saveant, Energy Environ. Sci., 2014, 7, 3808–3814.
43. C. Costentin, S. Drouet, M. Robert and J. M. Saveant, Science, 2012, 338, 90–94.
44. M. Razavet, V. Artero and M. Fontecave, Inorg. Chem., 2005, 44, 4786–4795.
45. U. J. Kilgore, J. A. S. Roberts, D. H. Pool, A. M. Appel, M. P. Stewart, M. R. DuBois, W. G. Dougherty, W. S. Kassel, R. M. Bullock and D. L. DuBois, J. Am. Chem. Soc., 2011, 133, 5861–5872.
46. T. Straistari, R. Hardre, J. Fize, S. Shova, M. Giorgi, M. Reglier, V. Artero and M. Orio, Chem.–Eur. J., 2018, 24, 8779–8786.
47. T. Straistari, J. Fize, S. Shova, M. Réglier, V. Artero and M. Orio, ChemCatChem, 2016, 2262–2268.
48. D. Sun, A. K. Harshan, J. Pecaut, S. Hammes-Schiffer, C. Costentin and V. Artero, Chemelectrochem, 2021, 8, 2671–2679.
49. I. Bhugun, D. Lexa and J. M. Saveant, J. Am. Chem. Soc., 1996, 118, 3982–3983.
50. W. X. Nie, D. E. Tarnopol and C. C. L. McCrory, J. Am. Chem. Soc., 2021, 143, 3764–3778.
51. N. Coutard, N. Kaeffer and V. Artero, Chem. Commun., 2016, 52, 13728–13748.

---

Footnotes

† In memory of Dr Jean-Michel Savéant.

‡ Electronic supplementary information (ESI) available. See DOI: 10.1039/d1se01267c

§ Equal contribution.

---