Data-driven discovery of single-atom catalysts for CO2 reduction considering the pH-dependency at the reversible hydrogen electrode scale

The electrochemical carbon dioxide reduction reaction (CO2RR) represents a promising approach to mitigating climate change and addressing energy challenges by converting CO2 into value-added chemicals. Among various CO2RR products, CO is attractive due to its economic viability and industrial relevance. By integrating large-scale data mining (with 939 experimental performance data), we reveal that the catalytic performance of d-block transition metal-based single-atom catalysts (SACs) for CO2RR is influenced not only by the coordination environment but also significantly by pH. However, the unified model that could accurately depict the pH-dependent CO2RR to CO activity of d-block SACs is urgently needed. Herein, we conducted pH-dependent microkinetic modeling based upon density functional theory calculations and pH-electric field coupled microkinetic modeling to analyze CO2RR performance of 101 SACs. Our data-driven screening identifies 12 high-performance SACs with promising CO selectivity across different pH conditions, primarily based on Fe, Cu, and Ni centers. We establish a scaling relation between key intermediates (*COOH and *CO) and analyze their adsorption behaviors under varying electrochemical conditions. Furthermore, our pH-dependent microkinetic modeling reveals the critical role of electric field effects in determining catalytic performance, aligning well with experimental turnover frequency values. Most importantly, our theoretical model accurately captures the pH-dependent performance of CO2RR-to-CO on d-block SACs, which is experimentally validated and serves as a general theoretical framework for the rational design of high-performance CO2RR catalysts. Based on this model, we identify a series of promising M-N-C catalysts, providing a universal design principle for optimizing CO2-to-CO conversion.