Designing and screening single‐atom alloy catalysts for CO2 reduction to CH3OH via DFT and machine learning

Abstract Carbon dioxide (CO 2 ) utilization technology is of great significance for achieving carbon neutrality, in which the catalytic materials play crucial roles, and among them, single‐atom alloys (SAAs) are of particular interests. In this study, density functional theory (DFT) calculations and machine learning are employed to assess the effectiveness of Cu‐, Ag‐, and Ni‐host SAAs as catalysts for electrochemical CO 2 reduction to CH 3 OH. The Gibbs free energies of 477 elementary reactions across 35 SAAs involved in CO 2 reduction are calculated, and by utilizing this dataset, a trained gradient boosting regression model is established with an excellent accuracy. Subsequently, the properties of 46 unknown SAAs are predicted, including their pathways, products, potential‐determining steps (PDS), and corresponding Gibbs free energies of the PDS ( G PDS ). Three promising candidates, ZnCu, AuAg and MoNi, stand out due to their lowest G PDS among Cu‐, Ag‐ and Ni‐ hosted SAAs, respectively.