Elucidating fundamental governing principles of CO2 reduction on single-atom catalysts through interpretable machine learning

Abstract

Understanding the effect of local coordination and electronic structure on the catalytic activity and selectivity is crucial to design CO2 reduction reaction (CO2RR) catalysts. Combined hierarchical high-throughput density functional theory (HT-DFT) screening and interpretable machine learning (ML), we aim to accelerate and rationalize the discovery of efficient SACs for CO2RR. Using a BC2 monolayer as a prototypical substrate, we constructed a library of transition-metal (TM) SACs featuring diverse vacancy types and N-coordination. A four-stage HT-DFT workflow, namely assessing stability, CO2 adsorption, key elementary steps, and hydrogen evolution reaction (HER) selectivity, was used, and 16 promising candidates with low limiting potentials (0.35 V) were identified. By integrating ML regression and feature-importance analysis, five fundamental physicochemical features that control activity, including TM d-electron count (Nd), ionization energy (IE1), electronegativity (χTM) and local electronegativity metrics (χn, Wv) were extracted. Furthermore, six analytical descriptors (φ1-φ6) were derived to quantify atomic features with catalytic energetics, specially, the descriptor framework is transferability to different graphene-like supports. Mechanistic analysis revealed that increasing Nd enhances CO-related adsorption while weakens O-TM interactions, and higher IE1 suppresses HER by favoring CO2 activation. This study establishes a physically interpretable ML-DFT strategy that unifies predictive efficiency with mechanistic insight, providing a blueprint for the rational design of atomically dispersed electrocatalysts for CO2 conversion.