Active learning-guided catalyst design for selective acetate production in CO electroreduction

Acetic acid is an essential chemical with industrial and consumer relevance, and global demand is expected to reach 24.5 million tonnes by 2025. Electrochemical CO reduction reaction (CORR) over Cu offers a sustainable route to acetate from waste carbon, but identifying catalysts with high CO-to-acetate selectivity remains challenging due to limited mechanistic understanding. Here, we establish an artificial intelligence (AI)-driven multi-scale simulation framework integrating grand-canonical density functional theory (GC-DFT), microkinetic modeling (MKM), and active learning to elucidate the CORR mechanism and guide catalyst discovery. The DFT-based MKM reveals that acetate forms via CO-CH coupling, with CH* binding energy identified as the key descriptor governing selectivity of acetate production from CORR. Active learning optimization predicts Cu/Pd (2:1) and Cu/Ag (3:1) as the most selective catalysts. Zero-gap electrolyzer experiments confirm their effective performance, achieving acetate Faradaic efficiencies of 50% and 47%, respectively, compared to 21% for pure Cu. This study demonstrates a data-driven catalyst design strategy for advancing selective electrocatalysis. CO electroreduction offers a sustainable route to acetate production but remains mechanistically unclear. Here, the authors develop an active learning multiscale framework revealing CH* binding strength as the key descriptor, predicting Cu/Pd and Cu/Ag as optimal catalysts.