From data to catalysis: advances and prospects of machine learning-driven electrocatalytic CO2 reduction

Abstract

As global decarbonization imperatives intensify, the electrocatalytic CO₂ reduction reaction (CO₂RR) has emerged as a key area of research in sustainable energy research. Conventional experimental methods are constrained by factors such as prolonged catalyst screening cycles and unclear reaction mechanisms. Machine learning (ML), a data-driven modeling approach, has revolutionized catalyst development by significantly accelerating discovery and reducing associated costs. This review is structured as follows: It initially outlines the ML-assisted design workflow specifically for electrocatalytic CO₂RR. Subsequently, it reviews the current progress of ML applications in CO₂RR, with a particular emphasis on catalyst design and screening, optimization of reaction conditions, and mechanistic understanding. Finally, the article discusses the challenges and future perspectives of employing ML in this field, thereby aiming to provide useful insights for ongoing and future research efforts.