(526d) Electrocatalyst Design Via Reaction Pathway Analysis: Integrating Density Functional Theory and Machine Learning

Electrochemical reactions offer sustainable pathways for numerous chemical transformations, notably through the electrocatalytic conversion of atmospheric molecules such as CO₂ into valuable chemicals. However, catalyst design for CO₂ reduction reaction (CO₂RR) remains challenging due to the complexity introduced by diverse reaction pathways, product selectivity, and intricate dependencies on operational conditions like applied potentials and solvation environments. Our work specifically targets the electrocatalytic conversion of CO₂ to formic acid, employing a robust integration of quantum mechanical modeling and machine learning-driven inverse catalyst design by studying reaction pathways. Initially, we utilized density functional theory (DFT) combined with string methods to elucidate detailed reaction pathways and energy barriers for N-heterocyclic-based catalysts. Quantum mechanical insights enabled us to identify structure-property relationships critical for optimizing catalyst performance, particularly selectivity and activation energy and study transition state structures. To overcome traditional trial-and-error methods, we deployed a genetic algorithm (GA) for inverse catalyst design. This evolutionary optimization algorithm systematically explores the catalyst design space, leveraging quantum-derived descriptors to iteratively generate and evaluate thousands of new catalyst structures and reaction pathways. Promising candidates identified via GA are subjected to comprehensive reaction pathway analysis and integrated back into an active learning framework, progressively refining the catalyst selection. Through iterative optimization cycles, our approach has successfully developed catalysts significantly reducing activation energies for CO₂ conversion to formic acid—achieving a barrier as low as 10 kcal/mol. This combined QM and ML-driven strategy demonstrates a powerful methodology to streamline electrocatalyst design and reaction pathway analysis, paving the way toward more sustainable and efficient electrochemical processes.