First-Principles Insights into Proton-Coupled Electron Transfer versus Hydrogen Evolution Reaction Selectivity from a Base-Appended Cobaltocene Mediator

Abstract

Performing selective proton-coupled electron transfer (PCET) to substrates such as N₂, CO₂, and unsaturated organic molecules under electrochemical conditions requires the suppression of the competing hydrogen evolution reaction (HER). To address this challenge, our laboratory previously demonstrated a PCET mediator strategy using a dimethylaniline-appended cobaltocene complex, [(CpCoCp^NMe2)­H]⁺, which performs selective reductive chemistry while suppressing the HER. However, the origin of the suppressed, yet still observable, HER has not been thoroughly established. In this work, we perform density functional theory (DFT) calculations to elucidate the HER mechanism involving this redox mediator and to provide atomistic insights into the bifurcation between the PCET and HER pathways. We find that protonation of the aniline moiety to form [CpCoCp^NMe2H]⁺ is more favorable, both kinetically and thermodynamically, than formation of the ring-protonated species [(CpCo­(Cp-H)^NMe2)]⁺. Furthermore, PCET to acetophenone is energetically more favorable via [CpCoCp^NMe2H]⁺ than via [(CpCo­(Cp-H)^NMe2)]^+1/0. In contrast, the most favorable HER pathway involves the ring-protonated Co­(I) species. These results offer mechanistic insights into HER versus PCET bifurcation and establish guiding principles for designing PCET mediators for selective electroreductive transformations.

Introduction

Electrochemical reductions of small molecules, such as CO₂, N₂, and organic molecules containing unsaturated π-bonds, can benefit from selective proton-coupled electron transfer (PCET) steps. − Thermodynamic and kinetic challenges remain toward more efficient reductive chemistry under electrochemical conditions. On a thermodynamic basis, bond dissociation free energy (BDFE) matching is needed between a net H–atom source and a highly reactive X–H intermediate (e.g., M–N₂ H, ·C­(R)­(R’)–OH, where X is CR₂, NR, or O) to ensure facile PCET steps. Kinetically, these transformations invariably compete with the hydrogen evolution reaction (HER) mediated by the electrode and/or a cocatalyst. − Catalytic strategies that suppress HER toward selective H–atom delivery to substrates are needed.

A PCET shuttle that colocalizes protons and electronspresenting a reactive H–atom equivalent for substrate deliveryoffers an attractive strategy for selective reductions. ,,− In this context, a dimethylaniline-appended cobaltocene complex serves as an electrocatalytic PCET (ePCET) mediator and has been shown to reduce unsaturated CC/O bonds and function in tandem with transition metal cocatalysts for N₂-to-NH₃ conversion and the semireduction of alkynes (Figure ). ,− We have previously shown that the reduced, protonated form of this mediator, ([CpCo^IICp^NMe2H]⁺), serves as the PCET donor. However, the parasitic HER pathway is an important issue to address, especially to guide future mediator designs, motivating us to undertake theoretical and experimental studies to differentiate it from the productive PCET pathway.

1.

Schematic outline of the anilinium-appended cobaltocene PCET mediator in its reduced, N-protonated form, [CpCoCp^NMe2H]⁺, and two other isomers explored theoretically herein, pertaining to the origin of PCET-to-substrate versus HER selectivity.

With respect to HER pathway(s), one can envision a range of plausible scenarios to consider that vary in kinetic order with respect to the mediator and the acid, and depend on the specific protonation site(s). Isomers of the reduced, protonated mediatorincluding N-protonated, Cp-ring-protonated (at various positions), and cobalt-hydride (Co–H) formsmust be energetically evaluated to assess their roles in PCET versus HER reactivity. HER pathways have been previously suggested for a range of metallocenes, involving both ring-protonated and metal hydride isomers. ,− In addition, the relative concentrations of mediator and acid during electrocatalysis will influence the preferred pathway. First-principles theoretical studies offer a valuable tool for assessing the feasibility and competitiveness of various HER pathways.

Motivated by these considerations, we have performed density functional theory (DFT) calculations to disentangle the PCET versus HER pathway(s) using [(CpCoCp^NMe2)­H]⁺ (Figure ). This theoretical study complements a recent experimental study of a related set of cobaltocene-derived PCET mediators from our laboratory. DFT calculations show that the N-protonated anilinium isomer [CpCoCp^NMe2H]⁺ is the most stable monoprotonated form of the mediator; this isomer is selective for PCET-to-substrate reactivity and is found to be unreactive toward HER at room temperature (Figure ). Our calculations also point to a preferred HER pathway that is first-order in acid and first-order in the cobaltocene mediator, involving a Cp-ring-protonated intermediate. HER versus PCET pathways involve orthogonal species, suggesting that these mechanistically distinct pathways can be exploited in future designs toward more selective PCET mediators.

Results

Relative Stabilities of the Protonated Isomers

First, the monoprotonated [(CpCoCp^NMe2)­H]⁺ isomers investigated are shown in Figure (top). The nomenclature adopted throughout for these species is as follows: (a) −N-protonated species, [CpCoCp^NMe2H]⁺ abbreviated as [1 _N–H ⁺]; (b) −Cp-ring protonation ([CpCo­(Cp-H)^NMe2]⁺), with five distinct isomers denoted as [2 _Cp‑H ⁺] for ortho conformer, [2 _Cp‑H(meta) ⁺], [2 _{Cp‑H(distal)} ⁺], and the two ipso conformers [2 ^endoCp-H­(ipso)⁺] and [2 ^exoCp-H­(ipso)⁺]; (c) −Co protonated (Co–H) isomer [Cp­(Co–H)­Cp^NMe2]⁺, [3 _Co–H ⁺]. [1 _N–H ⁺] is predicted to be the most energetically stable isomer. This accords with selective protonation at the −N position reported in the original experimental study, and more recent EPR confirmation of the assignment. , The Co–H form is predicted to be the least stable, in accord with our previous experimental and theoretical study of the protonation of Cp₂*Co, where a ring-protonated isomer (exo or endo) was found to be far more stable than a Co–H isomer. The first protonation of [1 _N] at the N and Cp sites is thermodynamically downhill, with ΔG values of −7.4 and −6.5 kcal/mol, respectively (Figure ). In contrast, the second protonation of the corresponding monoprotonated species, [1 _N–H ⁺] and [2 _Cp‑H ⁺], is thermodynamically uphill by 9.2 and 8.3 kcal/mol, respectively (Figure S1); therefore, the diprotonated Co­(II) and Co­(I) species were excluded from further reactivity analysis. Additional details are provided in the discussion of Figure S1.

2.

Speciation of the anilinium-appended cobaltocene mediator in the presence of tosylic acid, and corresponding PCET reaction coordinates with acetophenone as substrate. (top) Various [(CpCoCp^NMe2)­H]⁺ isomers are considered and their relative stabilities in terms of electronic energy and Gibbs free energy. Shorthand nomenclature for these isomers is also noted. (bottom) Acid association and protonation energetics, and subsequent acetophenone association and PCET reaction coordinate, focusing on the two most stable isomers: [1 _N–H ⁺] and [2 _Cp‑H]. Calculations were performed at the CPCM­(acetonitrile)/TPSS/def2TZVP level of theory with D3BJ dispersion correction. Values are given in kcal/mol. ** denotes constrained optimization since a stable Co–H species could not be obtained at the same level of theory. Color code for transition state atoms: C (black), H (white), O (red), N (light blue), S (yellow), and Co (deep blue). Blue dotted ovals highlight regions of bond-making/breaking in the transition state. The net charge and spin state of each transition state are shown in braces, and the key distances are reported in Å;. After protonation, the acid is shown only for chemical balance and is not included explicitly in the calculated structures unless chemically involved; the energy of the free acid is included to complete the thermodynamic cycle. See the SI for enlarged depictions of the chemical structures.

Protonation Barriers to [1_N–H ⁺] and [2_Cp‑H ⁺]

We studied the protonation barriers leading to the N–protonated isomer [1 _N–H ⁺] and the most stable ring-protonated isomer [2 _Cp‑H ⁺]. Tosylic acid (TsOH) was used as the proton source in calculating the protonation energetics (Figure , bottom). We find that formation of [1 _N–H ⁺] has a lower barrier (4.2 kcal/mol) than formation of [2 _Cp‑H ⁺] (9.0 kcal/mol). We note that the barrier for the formation of [1 _N–H ⁺] represents the precomplexation barrier only (i.e., barrierless proton transfer following acid association).

Acetophenone Reduction via PCET

Next, we explored the reaction coordinate for PCET to acetophenone from [1 _N–H ⁺] and [2 _Cp‑H ⁺] (Figure , bottom), and established the potential energy surface for the PCET-mediated reductive transformation to generate α–hydroxy radical PhMe­(OH)­C^•. Precomplexation of acetophenone to either [1 _N–H ⁺] or [2 _Cp‑H ⁺] is thermodynamically uphill (7.0 kcal/mol versus 7.1 kcal/mol, respectively). The kinetic barrier for transferring the net H atom from [1 _N–H ⁺] to PhC­(O)­CH₃ is 10.4 kcal/mol, much lower than the 19.6 kcal/mol barrier for transfer from [2 _Cp‑H ⁺]. Overall, the 0.9 kcal/mol lower free energy of [1 _N–H ⁺] relative to the closest ring-protonated isomer, [2 _Cp‑H ⁺], corresponds to an estimated Boltzmann population of ∼82% at 298 K, while the lower net H-transfer barrier from [1 _N–H ⁺] further supports PCET to acetophenone as the favored pathway.

Hydrogen Evolution Reaction Pathways

We next discuss plausible hydrogen evolution reaction (HER) pathways for this mediator system in the presence of TsOH, using Schemes and and Figures and to guide the discussion. We focus on bimolecular pathways because intramolecular HER would require two H equivalents to accumulate on the same mediator; however, the second protonation is thermodynamically uphill, making the corresponding diprotonated species thermodynamically inaccessible. Two types of bimolecular pathways are explored, including a pathway whereby both H atom equivalents derive from two corresponding equivalents of some [(CpCoCp^NMe2)­H]^+1/0 species reacting in a fashion that is bimolecular in cobalt, and a distinct pathway whereby one hydride (H^–) equivalent derives from a [(CpCoCp^NMe2)­H]^+1/0 species and the H⁺ equivalent comes from bimolecular reaction with TsOH.

1. Considered HER Pathway from [1 _N–H ⁺]­ .

^a The energetics of the considered pathways, HER pathway I–IV, are shown in Figure . [*A minimum corresponding to [3 _Co–H] could not be obtained; however, IM3 was found to be a minimum, suggesting that precomplexation stabilizes the [3 _Co–H] species.].

2. Considered HER Pathway from the Ring-Protonated [2 _Cp‑H]^+/0 .

^a The energetics of the considered pathways, HER pathway V–IX, are shown in Figure .

3.

Hydrogen evolution reaction pathway from [1 _N–H ⁺] as given in Scheme . The calculations were performed at the level of CPCM­(acetonitrile)/TPSS/def2TZVP with D3BJ dispersion correction. Values are given in kcal/mol. The color code for the transition states: C (black), H (white), O (red), N (light blue), S (yellow), and Co (deep blue). Blue dotted ovals are used to highlight bond-making/breaking regions at the transition states and the key distances are reported in Å. See the SI for enlarged depictions of the chemical structures.

4.

Hydrogen evolution reaction pathway from the ring-protonated Co­(II) and Co­(I) species [2 _Cp‑H]^+/0 as given in Scheme . The calculations were performed at the level of CPCM­(acetonitrile)/TPSS/def2TZVP with D3BJ dispersion correction. Values are given in kcal/mol, and distances are given in Å. The color code for the transition states: C (black), H (white), O (red), N (light blue), S (yellow), and Co (deep blue). Blue dotted ovals are used to highlight bond-making/breaking regions at the transition states. See the SI for enlarged depictions of the chemical structures.

HER via Bimolecular Coupling of [1_N–H ⁺] (HER Pathway I in Scheme )

HER pathway I (Scheme ) explores bimolecular hydrogen evolution from the most stable reduced Co­(II) protonated isomer [1 _N–H ⁺]. We find that the triplet state for two such species (IM1) is more stable than the corresponding singlet by 9.3 kcal/mol. An activation barrier to forming H₂ from the triplet state of [1 _N–H ⁺] is 32.8 kcal/mol (Figure A). Notably, spin crossover must occur to achieve the stable singlet products (2 [1 _N ⁺] + H₂). The activation barrier for the alternative singlet reaction coordinate (not shown) was 33.5 kcal/mol. Both of these barriers are sufficiently high as to be kinetically inconsequential and can hence be dismissed. See the Supporting Information for further details (Figure S2).

HER via Reaction of [1_N–H ⁺] with TsOH (HER Pathway II in Scheme )

H₂ evolution from direct reaction between the N-protonated isomer [1 _N–H ⁺] and TsOH has a significant barrier of 32.5 kcal/mol (HER pathway II in Scheme ; also see Figure ), rendering it kinetically inconsequential at room temperature. This finding helps explain why [1 _N–H ⁺] is selective for acetophenone reduction via PCET.

HER via [1_N–H ⁺] and [3_Co–H] (HER Pathway III in Scheme )

Bimolecular hydrogen evolution was examined between [1 _N–H ⁺] and the neutral cobalt hydride [3 _Co–H], while the cationic cobalt hydride [3 _Co–H ⁺] was excluded as it is thermodynamically inaccessible (Figure , top). The H₂ evolution pathway involving [3 _Co–H] is highly unlikely as the precomplexation is 29.7 kcal/mol uphill (IM3 in Figure ). The precomplex structure resembles the TS but lies at a higher energy, suggesting alternative intermediates. Moreover, a comparison of the energies of IM3 and IM4 shows that IM4 is 23.4 kcal/mol more stable than IM3. Therefore, any formation of IM3 would convert almost instantaneously to IM4, resulting in a negligible population of IM3 under the reaction conditions. Given the inaccessible activation barrier, this pathway was not pursued.

HER via [1_N–H ⁺] and [2_Cp‑H] (HER Pathway IV in Scheme )

Bimolecular H₂ formation was investigated between [1 _N–H ⁺] and [2 _Cp‑H]. Experimentally, the electrode potential required to generate [1 _N–H ⁺] from [1 _N–H ²⁺] is ∼−1.2 V versus Fc^+/0. The calculated reduction potential of [2 _Cp‑H]^+/0 is −0.6 V versus Fc^+/0, making [2 _Cp‑H] the dominant species under the applied electrode potential. For reference, a reduction potential of −0.4 V versus Fc^+/0 was recently reported for ring-protonated cobaltocene. A kinetic barrier of 21.6 kcal/mol makes the HER pathway IV plausible (Figure ).

HER via [2_Cp‑H]^+/0 (HER Pathway V, VII, IX in Scheme and Figure )

Bimolecular HER pathways from [2 _Cp‑H]^+/0 are discussed next (Scheme and Figure ), wherein three pathways distinguished by the charge states of the participating ring-protonated species were studied. Pathways involving species of the same charge have higher barriers, with H₂ evolution via the bimolecular coupling of two [2 _Cp‑H ⁺] (HER Pathway V) and two [2 _Cp‑H] (HER pathway VII) having barriers of 30.0 and 40.8 kcal/mol, respectively. In contrast, the H₂ evolution barrier from [2 _Cp‑H ⁺] and [2 _Cp‑H] is 14.1 kcal/mol (HER Pathway IX). This trend is consistent with hydricity values. [2 _Cp‑H] has a hydricity of 22.7 kcal/mol, compared to 48.7 kcal/mol for [2 _Cp‑H ⁺]. Therefore, in the H₂ evolution step, [2 _Cp‑H] can act as an effective H^– donor, which then combines with H⁺ from [2 _Cp‑H ⁺] to produce H₂. H₂ evolution is kinetically favored via Co­(II)/Co­(I) coupling. Thermodynamically, however, the coexistence of Co­(II) and Co­(I) is unlikely at the same applied electrode potential. Since the calculated reduction potential of [2 _Cp‑H]^+/0 couple is ∼−0.7 V versus Fc^+/0, conditions of −1.2 V versus Fc^+/0 (where [1 _N–H ⁺] is present) would render [2 _Cp‑H] dominant, creating a thermodynamic bottleneck of HER Pathway IX.

HER from TsOH and [2_Cp‑H]^+/0 (HER Pathway VI and VIII in Scheme and Figure )

H₂ evolution from TsOH and [2 _Cp‑H]^+/0 was examined. The activation barrier of H₂ formation from [2 _Cp‑H ⁺] and TsOH is 31.8 kcal/mol (HER pathway VI), whereas that from [2 _Cp‑H] and TsOH is 16.5 kcal/mol (HER pathway VIII). The lower barrier for the HER pathway VIII can be attributed to the higher hydricity of 2 _Cp‑H.

Electronic Structure Insights into the HER Activation Barriers

Electronic structure analysis of the HER barrier trends was carried out using a simplified model based on Löwdin charge (q) calculations (Figure A,B). − For each pathway, relative changes in q of the participating H atoms in H₂ formation between the initial state (IS) and the TS were computed, yielding the descriptor ΔΔq:

$$\mathrm{\Delta }\mathrm{\Delta }q=\mathrm{\Delta }{q}_{{\mathrm{H}}_1}+\mathrm{\Delta }{q}_{{\mathrm{H}}_2}=[q_{{\mathrm{H}}_1}(\mathrm{IS})-q_{{\mathrm{H}}_1}(\mathrm{TS})]+[q_{{\mathrm{H}}_2}(\mathrm{IS})-q_{{\mathrm{H}}_2}(\mathrm{TS})]$$

5.

Charge descriptor and HER barriers: (A) Löwdin charges of H (q _H) in different species, H is shown in red color. (B) Change in charge between the initial and transition state (ΔΔq) versus the activation barrier of different HER pathways.

Chemically, ΔΔq quantifies the extent of charge reorganization required to reach the TS, with larger values indicating more proton-like H atoms are participating in the H₂ formation. In the transition states analyzed, H charges range from 0.06 to 0.16, with an average of 0.11 and a median of 0.12 (Table S1).

Plotting ΔΔq against the activation barriers (Figure B) reveals several trends. Among HER pathways involving [2 _Cp‑H]^+/0 (i.e., HER pathway V, VII, and IX), lower ΔΔq correlates with lower activation barriers. Relative Cp­(C–H) distances in the TS also contribute, suggesting covalent effects beyond this electrostatic model. For instance, HER pathway VII exhibits longer Cp­(C–H) distances in the TS than HER pathway IX (1.70 Å vs 1.44 Å), consistent with its higher activation barrier (40.8 kcal/mol vs 14.1 kcal/mol).

For pathways involving TsOH, more proton-like counter-species (Figure A) correspond to higher activation barriers. HER pathway VIII, involving [2 _Cp‑H] (lowest q in Figure A), has the lowest barrier among the pathways involving TsOH (Figure B). A similar trend holds for HER pathway I versus HER pathway IV, where the participation of more proton-like [1 _N–H ⁺] leads to a higher barrier.

Overall, the ΔΔq descriptor indicates that HER pathways involving two H atoms with “proton-like” character have higher activation barriers; greater electrostatic repulsions must be overcome to drive association and H₂ evolution. Thus, a mediator state with the most proton-like character, i.e., [1 _N–H ⁺], is expected to exhibit the lowest HER activitya prediction consistent with experimental observations. In other words, ΔΔq descriptor offers a quantitative rationale for the expected observed trend that protonation occurs more readily for the more basic species.

Discussions

Summary of HER Pathways and Discussions on HER versus PCET Bifurcation

Our calculations point to two HER pathways as viable at room temperature: (a) HER pathway IX, which is second-order in Co and proceeds via Co­(II)/Co­(I) coupling, and (b) HER pathway VIII, which is first-order in both Co and TsOH (Figure ). As discussed earlier, HER pathway IX faces a thermodynamic bottleneck since [2 _Cp‑H ⁺] cannot attain an appreciable population under the applied electrode potentials in experiments. On the other hand, HER pathway VIII via [2 _Cp‑H] can gain favorability, as [2 _Cp‑H] is the dominant ring-protonated species. The presence of excess TsOH in the experiments further promotes the feasibility of HER pathway VIII. Hence, we conjecture that the most favored HER pathway is first-order in Co and first-order in TsOH. In accord with this argument, our recent experimental kinetic study of a modified cobaltocene mediator identified its HER pathway to be first-order in both mediator and acid, with PCET to substrate being zero-order in acid.

6.

Summary of HER versus PCET bifurcation via [(CpCoCp)^NMe2H]^+1/0. The most favorable HER pathway is shown in red arrows and the most favorable PCET pathway is shown in green arrows. Values in blue are given in kcal/mol (activation barriers) and the value in black (reduction potential versus Fc^0/+1) is given in V. TsOH is the source of H⁺ in all cases.

HER versus PCET Bifurcation and Future Directions

We find that PCET proceeds via [1 _N–H ⁺] (Figure ), whereas HER occurs through the ring-protonated species [2 _Cp‑H]. The combination of facile protonation kinetics and the highest thermodynamic stability makes [1 _N–H ⁺] the dominant species under experimental conditions, explaining why [(CpCo^IICp^NMe2)­H]⁺¹ performs selective PCET while suppressing HER in prior studies. ,−

Thus, HER versus PCET selectivity is governed by the speciation between N- and C-protonated mediator forms, which differ in their thermodynamic accessibility and kinetic reactivity toward H₂ evolution and substrate PCET. This conclusion comports well with the results and implications of a related experimental study by our laboratory. As a design principle for the next generation of redox mediators, minimizing –C protonationand thereby reducing the population of H^– donorsis prudent. The bifurcation ratio between −N and −C-protonated species can be tuned by varying external acid strength. Additionally, modification of the mediator scaffold, such as introducing steric effects to slow protonation at the −C center, can be explored. While modifying the mediator scaffold is challenging due to multiple competing factors, including synthetic accessibility, isolation of the desired mediator structure, and the possibility of altering competing HER pathways, tuning the external acid provides a comparatively simpler and more experimentally accessible strategy for controlling the PCET/HER bifurcation.

Conclusion

DFT calculations elucidate the origin of the high PCET selectivity of the dimethylaniline-appended cobaltocene redox mediator and explain the suppression of the HER. The results show that speciation governs HER versus PCET, with the PCET-active species being thermodynamically more stable. Furthermore, our results indicate that HER mitigation can be achieved by slowing ring protonation, thereby suppressing the propensity of the hydride pathway. This can be attained with weaker acids and covalent modification of the cobaltocene scaffold. These mechanistic insights can hence guide the design of next-generation PCET mediators for reductive transformations.

Methods

DFT calculations were performed with the ORCA quantum chemistry program package. Geometries of reactants, intermediates, and transition states (TSs) were optimized by using the TPSS density functional with the def2TZVP basis set. , CPCM solvation model was used to incorporate the solvent effects, and acetonitrile was selected as the continuum dielectric. The TS structures are identified by only one imaginary frequency. Grimme’s D3 dispersion correction was implemented in the calculations and the Gibbs free energies of the intermediates and TSs were obtained by adding the thermal and entropic corrections using Harmonic approximations. Overall, the Gibbs free energies were obtained at the CPCM­(acetonitrile)/TPSS/def2TZVP level of theory with D3 dispersion correction at 298.15 K and 1 atm pressure. To assess the sensitivity of the computed energetics to the chosen level of theory, we performed additional benchmark calculations on representative intermediates and transition states using TPSSh and B3LYP, along with several basis sets, including 6-311G*, def2-SVP, def2TZVP, and def2-TZVPP. These benchmarks showed that the main relative energetic trends and the qualitative PCET versus HER bifurcation are preserved across the tested functional/basis-set combinations. The benchmark results are provided in Tables S2 and S4. Note that for the PCET pathway, a net H atom transfer is modeled in silico. Similarly, when we describe H₂ evolution as involving a net H^– and/or H⁺, the calculations do not treat these fragments explicitly. Instead, the overall charge is assigned to the full system without enforcing charge localization. For example, in HER pathway VIII, we refer to [2 _Cp‑H] as delivering a net H^– and TsOH as providing a net H⁺; while this is the chemical picture, in the calculations H₂ evolution is modeled simply as the combination of two H atoms. Although disentangling the exact proton- and electron-transfer coordinates involved in such processes is mechanistically important, such an analysis is beyond the scope of the present study. Here, our focus is to provide a qualitative picture of the PCET versus HER bifurcation. Accordingly, because the present adiabatic DFT treatment does not explicitly resolve coupled proton–electron coordinates or possible nonadiabatic effects, we do not assign the PCET transition states as synchronous or asynchronous based on transition-state charge populations alone.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.inorgchem.6c03389.

• XYZ coordinates. (ZIP)

• Calculated structures, enlarged graphics. (PDF)

The authors declare no competing financial interest.