Circumventing the scaling relationship on bimetallic monolayer electrocatalysts for selective CO2 reduction

Electrochemical conversion of CO2 into value-added chemicals continues to draw interest in renewable energy applications. Although many metal catalysts are active in the CO2 reduction reaction (CO2RR), their reactivity and selectivity are nonetheless hindered by the competing hydrogen evolution reaction (HER). The competition of the HER and CO2RR stems from the energy scaling relationship between their reaction intermediates. Herein, we predict that bimetallic monolayer electrocatalysts (BMEs) – a monolayer of transition metals on top of extended metal substrates – could produce dual-functional active sites that circumvent the scaling relationship between the adsorption energies of HER and CO2RR intermediates. The antibonding interaction between the adsorbed H and the metal substrate is revealed to be responsible for circumventing the scaling relationship. Based on extensive density functional theory (DFT) calculations, we identify 11 BMEs which are highly active and selective toward the formation of formic acid with a much suppressed HER. The H–substrate antibonding interaction also leads to superior CO2RR performance on monolayer-coated penta-twinned nanowires.

G = E B + ZPE + ∫C P dT -TS (1) All the free energy corrections are calculated based on the molecular vibration analysis and assuming that the changes in the vibrations of the surface caused by the intermediate are minimal.
We applied approximate solvation corrections to the reaction intermediates proposed by Peterson et al. 2  According to the computational hydrogen electrode (CHE) model, the limiting potential (U L ) for the reaction step, for example, CO 2 → *HCOO, is defined as the change of the free energies between *HCOO and CO 2 , in addition to the chemical potential of a proton-electron pair μ(H + + e -), calculated as half of the free energy of gas-phase H 2 at zero applied voltage: The free energies of non-adsorbed species such as CO 2 and HCOOH are taken from previous work. 2 Note that to correct the inconsistency between the theoretical and experimental gas-phase reaction enthalpies, +0.45 eV is added to the energy of CO 2 and HCOOH, as also proposed in previous work. 2 The following reaction pathways for CO 2 hydrogenation have been considered in this work: As discussed in the main paper, the reduction of CO 2 into *COOH is disfavored on the proposed BMEs and CO 2 is only expected to be reduced into *HCOO. Thus, the reactions after *COOH S3 are excluded. In addition, *CO cannot be reduced from *HCOO because the reduction requires breaking of a C-O bond and a C-H bond in *HCOO, which is energetically unfavorable. Thus, the reduction of *HCOO to *CO and other intermediates after *CO is also excluded.
Furthermore, because the desorption of *HCOOH is exothermic on the proposed BMEs, *HCOOH cannot be reduced further and is the final product. Hence, no products other than HCOOH can be produced on the proposed BMEs. Based on the two-step reaction (CO 2 → *HCOO → HCOOH), the more negative U L among the two steps is defined as the reaction overpotential U OP in this work.

b. Activation barrier calculations
The activation barrier for the hydrogenation reaction is calculated based on the model proposed by Nie et al. [3][4] In the model, the activation barrier for an elementary electrochemical reaction (A* + H + + e -→ AH*) is derived from analogous surface hydrogenation reaction (A* + H* → AH*). The activation barrier as a function of the electrode potential U is calculated as: where E act 0 is the reaction barrier calculated from DFT plus the ZPE correction. U 0 is set so that the chemical potential of the adsorbed H* is equal to that of a proton-electron pair. β' is an effective symmetry factor calculated by: where μ TS -μ reactant represents the variation in the surface dipole moments between the reactant and the transition state.

c. Water-Assisted reaction model
In the activation barrier calculations, we considered the presence of one water molecule below the H-down ice-like water bilayer to assist the hydrogenation reactions. More specifically, water can assist the reaction in two manners: (1) the surface proton is transferred directly to the adsorbate, assisted by the hydrogen bond between the water molecule and the adsorbate; (2) the surface proton is transferred to the water molecule, which concurrently shuttles another proton to the adsorbate, analogous to the Grotthuss mechanism.
We note that the above water-assisted activation barrier calculation model was able to reproduce experimentally identified CO 2 RR species on Cu, and also predicted a correct methanol product for CH 2 O reduction on Cu. [3][4] Besides, the model was also used to examine C 2 product selectivity on Cu(100) for CO 2 RR, 5 to predict Cu monolayer catalysts and bimetallic alloys for S4 CO 2 RR, [6][7] to study binary metal catalyst for electrochemical nitrogen reduction reaction, 8 to design Au 22 (L 8 ) 6 nano-clusters for oxygen reduction reaction, 9 and to predict bimetallenes for selective CO 2 RR to HCOOH. 10  Figure S1 Radial distribution function and crystal structure of the selected BMEs at 300 K in the presence of HCOO* on the surface after a 4 ps MD simulation. The vertical lines indicate the equilibrium nearest-neighbor (NN) distance at 0 K. The proposed BMEs are found to be stable because of the negligible changes in the NN distance between 0 K and 300 K. S10 Figure S2 The adsorption structure and differential charge density isosurfaces for *H on