The Impact of Specifically Adsorbed Ions on the Copper-Catalyzed Electroreduction of CO₂

The free energy of surface bound species X^* (G_X^*) is calculated using Equation 1 and includes vibrational correction terms that correct the free energy from 0 K to 298 K. For the chemical potential of gas-phase species X⁰_(g), computed at standard conditions of 298 K and 1 atm pressure, all degrees of freedom are included in the enthalpy and entropy corrections as shown in Equation 2. For liquid water, the gas-phase free energy is determined at the room-temperature vapor pressure of water (p_H2O = 0.031 atm). The free energy of solid-phase species X_(s) (G_X(s)) is set equal to their DFT energy as shown in Equation 3:

G represents the free energy of constituent X that may be a surface bound (X^*), gas-phase (X_(g)) or solid (X_(s)) species. E^DFT and E^ZPVE refer to the 0 K DFT and zero-point vibrational energy with U, TS and PV as the entropy, internal energy and pressure-volume corrections, respectively. The free energy of aqueous-phase iodide, I⁻_(aq) (), and potassium ion, K⁺_(aq) (), are computed using the standard-state free energy of iodine gas, I_2(g) () (assuming ideal gas behavior), and solid potassium, K_(s) (), respectively, along with the experimental standard reduction potential (U⁰)^61,62 in V-SHE as shown in Equations 4–7. Equations 5 and 7 are valid when the free energy change of iodide and potassium dissolution is zero.

where e represents the absolute value of the elementary charge. Standard reduction potential values (at 25^○C) of U⁰ _I2 = 0.54 V-SHE and U⁰ _K = −2.93 V-SHE are used.

The specific adsorption of iodide and potassium ions onto the solvated copper surface is described in Equations 8 and 9 respectively:

where (H₂O)_n^* represents the copper surface with n H₂O molecules explicitly adsorbed, I(H₂O)_n^* and K(H₂O)_n^* represent surface-adsorbed iodide and potassium solvated by n H₂O molecules and I⁻_(aq) and K⁺_(aq) are the aqueous-phase ions. Owing to the difficulty in representing a dynamic ensemble of the H₂O molecules within the framework of DFT, we instead resort to the use of a static explicit structure of H₂O molecules at a local minimum. Final converged structures of solvated K^* and I^* on the copper surface are presented in the Supplementary Information. Detailed discussions on the adopted approach to model ion solvation was previously reported.^39,41 We neglect longer range solvent stabilization of adsorbed ions in this model, which would act only to further stabilize adsorbed species.

The thermodynamic free energy change (ΔG) of solvated ion adsorption as a function of the electrode potential on an NHE scale (U_NHE) is calculated using Equations 10 and 11:

where the free energy of aqueous-phase ions (G_(aq)) is calculated using Equations 5 and 7, μ represents the dipole moment in the surface normal direction, U_PZC is the potential of zero charge and d represents the Helmholtz layer thickness. The last term serves as a correction to the free energy change of ion adsorption and accounts for the interaction between the surface dipole moment (μ_X(H2O)n^* − μ_(H2O)n^*, due to charge retention of the adsorbed ion) and the electrode/electrolyte interfacial electric field (U_NHE – U_PZC)/d. This term effectively allows ion adsorption with partial electron transfer. The formalism for this correction has been previously reported,^39,41,63 and we approximate a U_PZC of 0 V-RHE and d of 3 Å as used in these previous studies. The equilibrium adsorption potential (U⁰_NHE) for iodide and potassium adsorption is calculated using Equations 10 and 11, when ΔG^ADS = 0, as shown in Equations 12 and 13:

The binding energies of all surface-bound CO₂ ER intermediates are calculated using Equation 14. Equation 15 is employed to estimate the effect of the co-adsorbed ion on the binding energy of the CO₂ intermediate.

where ^* denotes surface-adsorbed species, and is the gas-phase DFT energy of the CO₂ ER intermediates.

The reaction free energy change (ΔG) of CO₂ ER reactions as a function of electrode potential U can be calculated on a reversible hydrogen scale (RHE) using the computational hydrogen electrode (CHE) model.⁶³ The relative free energy of an intermediate CO_xH_y^* formed from CO_2(g) in the CO₂ ER pathway as shown in Equation 16 is calculated using Equation 17. To estimate the effect of a co-adsorbed ion on reaction free energies of CO₂ ER, Equation 17 is modified to account for the free energy of the adsorbed ion as shown in Equation 18.

We have previously developed a simple, transferable method to approximate a potential-dependent activation barrier for a surface-mediated elementary electrochemical reaction .^44,48,64 DFT models are unable to accurately capture the chemical potential of an aqueous phase proton in bulk electrolyte, within limited unit cells and static H₂O structures. We assume that the reduction reaction in Equation 19 can be approximated as an inner shell reaction, with the transition state being local to the surface species (A^*) and that electron motion is rapid once the nuclei attain the transition state. Under these assumptions, we can use local models of adsorbed species and the ion being transferred such that the elementary electrochemical reaction (Equation 19) can be recast in the form of an analogous non-electrochemical reaction as shown in Equation 20. The main approximation of this method is that the transition state located for the analogous non-electrochemical reaction (Equation 20) is equivalent to the potential-dependent reaction (Equation 19) at a specific potential U⁰. The equilibrium potential for the addition of H into the DFT model cell (to form (A^*+H^*)), denoted by U⁰ in Equation 21, allows the determination of the electrode potential at which the proper reference reactant state, A^* + H⁺ + e⁻, and its intermediate state, (A + H)^* have the same chemical potential. This allows the energy of the transition state to be referenced back to the chemical potential of the ion in the bulk electrolyte.

The activation barrier corresponding to the transition state of the reaction in Equation 20 is labeled G⁰_act and is the barrier obtained from DFT with ZPVE corrections at U = U⁰. The barrier can be extrapolated to other potentials using the Butler-Volmer formalism as shown in Equation 22⁶⁴

where β denotes an effective reaction symmetry factor, similar to the Brønsted-Evans-Polanyi coefficient⁶⁵ and typically varies between 0.3 and 0.7.⁶⁶ In this study, we approximate β = 0.5 for all reactions. While future work will address improved estimation of the symmetry factor for various elementary steps, within the scope of this study, we assume that the relative activation barriers with and without ion co-adsorption will have minor variations in the values of β.

The equilibrium adsorption potentials, U⁰ (in V-NHE), for K⁺ and I⁻ adsorption in a low-coverage limit (θ = 1/9^th ML) on Cu (111), Cu (100) and Cu (211) surfaces were calculated using Equations 12 and 13 and are summarized in Table S1 of the Supplementary Information. A low-coverage limit is considered based on previous studies that concur that the favorability of adsorption decreases with increasing coverage, most likely due to the steric or Coulombic repulsion between adjacent ions at higher coverages.^41,67 The specific adsorption of K⁺ is favorable at any potential more negative of , while I⁻ adsorbs favorably at potentials more positive of . Iodide adsorption is more favorable on the Cu (100) (U⁰ = −0.89 V-NHE) and Cu (211) (U⁰ = −0.85 V-NHE) surfaces and least favorable on the Cu (111) surface (U⁰ = −0.59 V-NHE). The favorability of potassium adsorption on the copper surfaces follows the order: Cu (211) (U⁰ = −2.04 V-NHE) > Cu (111) (U⁰ = −2.50 V-NHE) > Cu (100) (U⁰ = −2.65 V-NHE).

The effect of surface solvation on the equilibrium potentials of K^* and I^* adsorption was investigated by explicitly solvating K^* and I^* with up to 4–6 water molecules. Surface solvation dependence on the equilibrium adsorption potential of ions was calculated by systematically varying the number of explicit water molecules and is reported in Figure S14 of the Supplementary Information. Based on the analyses, we observe that the equilibrium adsorption potential is converged with respect to the number of explicit water molecules included with 6 molecules, albeit with noise that remains significant on the 111 surface. The equilibrium adsorption potentials of the explicitly solvated ions (Table S1) and the final converged configurations are reported in the Supplementary Information. Solvation increases the favorability of specific adsorption of K⁺ by shifting to more positive potentials. The shift is greatest on Cu (100) (1.14 V) followed by Cu (211) (0.73 V) and Cu (111) (0.41 V). This significant increase in adsorption favorability when surface solvation is included matches approximately the solvation interactions during alkali cation specific adsorption onto Pt (111).³⁹ Relative to K⁺, the effect of solvation on I⁻ adsorption across the copper surfaces is minimal with a shift of 0.01–0.02 V-NHE in values on Cu (111) and Cu (211) surfaces and 0.16 V-NHE on Cu (100). The relative impact of solvation can be rationalized by examining the charge retained by the ion upon adsorption and the resulting surface normal dipole moment (Figures S4 and S5 of the Supplementary Information). The magnitude of the surface dipole moments are significantly larger for adsorbed K^* relative to I^*, indicative of the higher amount of charge retained by K^* than I^*. By explicitly solvating K^*, the surface dipole moment is lowered while K^* retains more charge due to the H₂O molecules that stabilize the adsorbate. Conversely, due to the small surface dipole moment and lower charge retention of I^*, the effect of solvation on the stability of I^* is minimal. This observation concurs with our previous investigations of the effect of solvation on the adsorption favorability of aqueous-phase halides on copper surfaces.⁴¹

Figure 1 shows the equilibrium adsorption potentials, U⁰ (in V-NHE), for explicitly solvated K^* and I^* in a low-coverage limit (θ = 1/9^th ML) on Cu (111), Cu (100) and Cu (211) surfaces. This Pourbaix diagram thus indicates potential-pH regions at which specific adsorption of potassium and iodide on the copper surfaces is thermodynamically favorable. The solid black line in Figure 1 shows the thermodynamic equilibrium potential (U⁰ = 0.17 V-RHE)⁴⁶ for CO₂ reduction to CH₄ (CO₂ + 8 H⁺ + 8 e⁻ → CH₄ + 2 H₂O). Previous experiments, however, have typically reported significant operating overvoltages of more than 0.5 V on copper electrodes to produce appreciable product yields.^68,69 Under CO₂ ER conditions, we expect a low coverage of iodide to adsorb on all three surface facets, with I^* stability extending to larger CO₂ reduction overpotentials at low pH and on Cu (211) and Cu (100) facets. A low coverage of K^* can adsorb favorably at higher pH values on all three facets, though a significant CO₂ ER overpotential is required for K^* adsorption on the Cu (111) facet.

Zoom In Zoom Out Reset image size

Though the potential ranges of isolated K^* and I^* formation do not overlap, the expected attraction between them could create a potential window in which the co-adsorption of K^* + I^* is favorable. We, therefore, also compute the equilibrium potential range that favors their co-adsorption. Figure 1 also shows the potential window within which the co-adsorption of explicitly solvated K^* and I^* is favorable. On Cu (111), the value shifts from −2.09 to −1.24 V-NHE due to I^* co-adsorption, while the value shifts from −0.58 to −1.45 V-NHE when K^* is co-adsorbed, giving rise to a potential window within which K^* and I^* can co-adsorb on the Cu (111) surface. Under high pH CO₂ ER operating conditions, co-adsorption of K^* and I^* is only favorable on the Cu (111) surface when explicit solvation effects are considered, and only under a narrow potential range (−1.45 < < −1.24 V-NHE). Similar narrow potential ranges for K^* + I^* co-adsorption are observed for Cu (211) and Cu (100), as shown in Figure 1. Since the co-adsorbed ions are stabilized due to their attractive interaction, as evidenced by a higher charge retention (see Figure S5), it is plausible that higher coverages of co-adsorbed K^* + I^* could be favorable and expand the co-adsorption window.

To calculate the effect of co-adsorbed ions on the binding strength of the reaction intermediates, the ions and the reaction intermediates were adsorbed at their preferred site in a 3 × 3 unit cell (θ = 1/9^TH ML/adsorbate). The ions were placed in the preferred site nearest to the reaction intermediate to quantify their interaction. Figures 2–3 illustrate a top view of the configurations used in modeling CO^* reduction to CHO^* and COH^* (with explicit solvation) with and without ion co-adsorption on Cu (111), the results of which are elaborated in greater detail in the subsequent sections. The final converged configurations of the reaction intermediates with co-adsorbed ions on Cu (100) and Cu (211) surfaces are reported in the Supplementary Information. Figure 4 shows the variation in the binding energy of CO₂ ER intermediates on three copper facets with ion co-adsorption. The binding energies of the CO₂ ER intermediates are calculated using the gas-phase molecular energies as reference. More negative binding energy values correspond to stronger binding of the intermediates with the copper surface. The binding strength of the CO₂ ER intermediates on all three copper surfaces varies in the order: COH^* > CHO^* > CO^*. The O-terminated species, including CO^* and CHO^*, bind slightly stronger to the stepped Cu (211) facet than the close-packed Cu (100) and Cu (111) facets (Figures 4a–4b). This trend is reversed in the case of COH^* (Figure 4c) which binds weakest to Cu (211) and strongest to the Cu (100) surface.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

When I^* is co-adsorbed on the copper surface, the CO₂ ER intermediates bind weaker, on average by 0.14 eV. For all CO₂ ER intermediates, the weakening effect due to I^* co-adsorption is most pronounced on Cu (211) (ΔBE ∼0.26 eV), then Cu (100) (ΔBE ∼0.11 eV) and least on Cu (111) (ΔBE ∼0.08 eV). Conversely, the effect of co-adsorbed K^* is opposite but similar in magnitude to co-adsorbed I^*. K^* co-adsorption strengthens the binding of the CO₂ intermediates on the copper surfaces, on average by 0.16 eV. The effect of co-adsorbed K^* is more apparent on the binding strength of O-terminated species (CO^* and CHO^*) than COH^*, indicated by the smaller surface dipole moments (Figure S10), with the effect most pronounced on the Cu (111) surface (ΔBE ∼0.20 eV) and the Cu (100) (ΔBE ∼0.16 eV), least on Cu (211) (ΔBE ∼0.01 eV) surfaces. Thus, while co-adsorbed I^* weakens the binding of all CO₂ ER intermediates, co-adsorbed K^* generally promotes stronger binding of the intermediates, more so the O^*-terminated species (CO^* and CHO^*).

Under operating CO₂ ER conditions, solvation coupled with co-adsorbed ions likely play a crucial role in affecting the binding strength and stability of the CO₂ ER intermediates. Explicitly solvated K^* (shown in Figure 4) serves to strengthen the binding of CO₂ ER intermediates by 0.29 eV on average relative to co-adsorbed K^* alone. Since solvation effects on I^* adsorption is minimal, as reported in the previous section, we only explore the impact of explicitly solvated K^* on the binding of the CO₂ ER intermediates. Solvation alone serves to stabilize the CO₂ ER intermediates on the copper surfaces, with a significant effect on COH^* (0.58 eV) relative to CHO^* (0.35 eV) and CO^* (0.15 eV) on the copper surfaces. The preferential stability of COH^* with inclusion of explicit solvation can be attributed to a strong H-bond network. When the effects of K^* co-adsorption are taken into account, however, the stability of COH^* due to explicit solvation is reduced from 0.58 eV to 0.34 eV. Since co-adsorbed K^* retains most of its charge, we expect strong cation-water interactions to disrupt the H-bond network and lower the binding stability of COH^* on the copper surfaces.

Figure 5 shows the reaction free energy diagram for CO^* reduction on Cu (111), Cu (100) and Cu (211) at U = −0.5 V-RHE with and without ion co-adsorption. Before considering the effects of co-adsorbed ions, we summarize the reaction free energies of CO^* formation and reduction without co-adsorbed ions. The relative reaction energy of CO₂ gas reduction to CO^* formation is downhill in energy on all three copper facets, suggesting that the CO^* formation is thermodynamically viable at U = −0.5 V-RHE. The reaction energy of CO^* reduction to CHO^* on copper is more favorable by 0.34 eV on average than that of CO^*→COH^*, with reduction being more favorable on Cu (211) > Cu (100) > Cu (111). Previous DFT studies on the reaction mechanism of CO₂ ER on Cu (111) report that solvation plays a crucial role in lowering the activation barriers of OH bond formation and preferentially reducing CO^*→COH^*.^44,48 Differences in the reaction energetics and selectivity of CO^* reduction due to solvation and explicitly solvated ions are addressed in the latter part of the section.

Zoom In Zoom Out Reset image size

The reaction energies of the CO₂ ER intermediates are significantly affected by the presence of co-adsorbed K^* and I^*, as seen in Figure 5. K^* promotes CO^* formation, whereas co-adsorbed I^* weakens the relative energy of CO^* formation from CO_{2 (gas)} with the effect most pronounced on Cu (211) (0.34 eV) > Cu (100) (0.14 eV) ∼ Cu (111) (0.067 eV). In our recent study, we confirmed that a tendency to form a high coverage of CO^* on transition metals obstructed access of H^* to the metal surface sites and prevented subsequent C-H bond formation.⁷⁰ The presence of co-adsorbed I^* could lead to more CO gas phase products by reducing the CO^* binding strength. This is evident in Figure 5c, where the reaction energy of CO^* formation on Cu (211) at U = −0.5 V-RHE differs by only 0.1 eV from the standard gas-phase CO formation energy.

While co-adsorbed I^* weakens the stability of both CHO^* and COH^* on the copper surface, co-adsorbed K^* increases the stability of CHO^* on the copper surfaces relative to COH^*. On Cu (211), the decreased stability of CO^* due to co-adsorbed I^* reduces the reaction free energy of CO^*→CHO^* reduction from 0.21 eV to 0.06 eV and from 0.77 eV to 0.68 eV for the CO^*→COH^* reduction step. On Cu (100), the relative reaction free energy of CO^*→COH^* remains largely unaltered with K^* co-adsorption, but the reaction free energy of CO^*→CHO^* is reduced by 0.20 eV, making the pathway exergonic at U = −0.5 V-RHE. Table I summarizes the differences in the relative free energy of COH^* and CHO^* with and without ion co-adsorption. Positive values reflect G_COH^* > G_CHO^*, with higher stability towards CHO^* formation. Co-adsorbed I^* weakens the stability of both CHO^* and COH^* slightly, while co-adsorbed K^* increases the relative stability of CHO^* over COH^* significantly on all copper surfaces. The preferred stability of CHO^* over COH^* with K^* adsorption can be attributed to the higher sensitivity of CHO^* BE to co-adsorbed K^* than that of COH^*.

Table I also compares the relative stability of CHO^*/COH^* when solvation effects are considered. Solvation preferentially stabilizes COH^* with a strong H-bond network, with the effect most pronounced on Cu (100) > Cu (111) ≫ Cu (211). The use of a single explicit H₂O molecule in our DFT models to calculate the relative stability of CHO^*/COH^* is an approximation to the fully solvated electrode/electrolyte interface, with the stability of the adsorbed species being subject to the structure of explicit solvation used. In our previous study, including explicit surface solvation using a single water molecule was found to shift the relative trends in CHO^* and COH^* binding on Pt, Ni, Co, Cu and Au surfaces by 0.41 eV and 0.60 on average respectively.⁴⁷ Adding additional water molecules up to a water bilayer has been previously shown to further stabilize the adsorbates by 0.1-0.2 eV,⁴⁸ though with additional computational demands. The use of a single explicit water molecule serves to capture the major effect of surface solvation on the relative stability of CHO^*/COH^*. When explicit solvation and K^* co-adsorption are considered, the stability of CHO^* over COH^* is accentuated on the copper surfaces, with the effect most pronounced on Cu (211). Thus, co-adsorbed K^* with the inclusion of solvation significantly promotes a CHO^* path and generally stronger binding of all three reaction intermediates to the copper surfaces. In contrast, I^* does little to affect the selectivity of CHO^*/COH^* formation, but will promote CO^* desorption on the copper surfaces. The differences in binding strength and relative stability of CHO^* and COH^* species with ion co-adsorption play a crucial role in dictating the CO^* reduction selectivity towards CHO^*/COH^* on the copper surfaces. While these conclusions are arrived at from a thermodynamic standpoint, elementary kinetics and competitive surface coverage effects will impact the formation of CO^* and selectivity of CO^* reduction to CHO^* or COH^*. In the next section, we address the impact of co-adsorbed ions on the potential-dependent activation barriers of the CO^* reduction to CHO^* and COH^*.

Figure 6 shows the reaction energy diagram and the activation barriers for CO^* reduction to CHO^* (Figure 6a) and COH^* (Figure 6b) on Cu (111) at U = −0.5 V-RHE with and without ion co-adsorption. The relative energies of CO^* and CHO^* formation shown in Figure 6a are the same as those in Figure 5a, with the corresponding activation barriers now included. Figure 6b shows the relative energies of CO^* and COH^* formation modeled with using one explicit water molecule. The corresponding initial, transition and final state configurations of both elementary reduction steps are illustrated in Figure 2 and Figure 3. The calculated activation energies (G⁰_act), initial state formation potentials (U⁰) and activation energies at U = −0.5 V-RHE computed using Equation 22 are reported in Table II. Imaginary frequencies along the reactive modes are also given in Table II.

Zoom In Zoom Out Reset image size

A direct surface hydrogenation mechanism is employed in calculating the barrier for CO^*→CHO^* (C-H bond formation in Figure 2). The mechanism transfers H^* to surface-adsorbed CO^* species. Previous DFT electrokinetic studies revealed that all C-H bond formation reactions proceeded over lower barriers through surface hydrogenation paths, rather than proton shuttling through water.^44,48 This conclusion is intuitive given the expected role of the metal in C-H dissociation steps and the principle of microscopic reversibility.

A solvation assisted proton shuttling mechanism is used in calculating the barrier for CO^*→COH^* (O-H bond formation in Figure 3) based on previous reports wherein the barriers associated with O-H bond formation and C-OH bond dissociation were significantly lowered via the H-shuttling H^* transfer mechanism.^44,48 Modeling the solvation assisted proton shuttling mechanism using a single water molecule in our work is an approximation. Previous reports indicate that the use of one explicit water molecule can lower the activation barrier of CO₂→COOH^* on Cu (111) by 0.8 eV and inclusion of a water bilayer can further stabilize the barrier by 0.1–0.2 eV.⁴⁸ Although the water bilayer generally serves to improve the accuracy of the solvation model and can influence the energetics of CO₂ ER, the use of a single water molecule in our model is expected to capture the relative variations in the activation barriers of CO^* reduction with and without ion co-adsorption.

Neglecting ion adsorption effects, the presented reaction energetics in Figure 6 and Table II suggest that the CO^* reduction to COH^* on Cu (111) at U = −0.5 V-RHE has a lower barrier of 0.33 eV compared to a barrier of 0.73 eV for CO^*→CHO^* reduction. The relative reaction energy for CO^*→COH^* is also lower by 0.48 eV relative to CO^*→CHO^*, indicating that solvation plays a crucial role in lowering the activation barriers of O-H bond formation and the preferential selectivity of CO^* reduction to COH^*.

Co-adsorbed I^* slightly weakens the stability of the CO₂ intermediates and the activation barriers of CO^* reduction to CHO^* and COH^* on Cu (111) in the presence of co-adsorbed I^* are mainly unaltered (∼0.05 eV) relative to without ion co-adsorption. Since co-adsorbed I^* loses most of its charge, we expect it to have a low impact on the electron transfer at the transition state for both mechanisms. Co-adsorbed K^* preferentially increases the stability of CO^* (0.23 eV) and CHO^* (0.36 eV) on Cu (111) as shown in Figure 6a, however, decreases the stability of solvated COH^* by 0.33 eV (Figure 6b) by disrupting the H-bond network between COH^* and the water molecules. The activation barrier for CO^*→CHO^* reduction is lowered by 0.18 eV (at U = −0.5 V-RHE) in the presence of co-adsorbed K^*. The barrier for CO^*→COH^* reduction, via the water-assisted proton shuttling mechanism, increases by 0.70 eV with co-adsorbed K^*, inverting the selectivity of CO^* reduction to CHO^* when the effects of co-adsorbed K^* are taken into account. The preferred selectivity towards CHO^* formation with co-adsorbed K^* can be rationalized in terms of the higher charge retained by K^* that increases the surface dipole moment, thereby facilitating the electron transfer process from the surface to the transition state in a surface hydrogenation mechanism. For the water-assisted proton shuttling mechanism, the extreme 0.70 eV barrier increase occurs due to the difficulty in shuttling a positive charge through a water molecule interacting with a partially positively charged K^* species. The high charge on K^* (q_K = +0.82 e) and the shuttling H^* (q_H = +1.00 e) in the transition state destabilizes the state and increase the corresponding activation barrier, rendering the CO^*→COH^* path kinetically unfavorable relative to CO^*→CHO^*. This specific 0.70 eV impact of K^* is reliant on the relative positions of K^*, CO^*, and H₂O, however, the qualitative effect of K^* retaining positive charge, slowing H⁺ transfer to CO^* adsorbed in its vicinity, is expected to be robust. Although the solvation model, reaction mechanism and the approximations in estimating potential-dependent activation barriers can give rise to quantitative variations in the activation barriers, the qualitative impact of co-adsorbed K^* on the kinetics and selectivity of CO^* reduction are emphasized here.