molecular dynamics, kinetic Monte Carlo, and machine learning-accelerated simulations. In this Review, we provide a concise overview of how these approaches have advanced the mechanistic understanding of electrocatalyst reconstruction across multiple time and length scales. We highlight how different computational strategies offer complementary insights into thermodynamically accessible states, atomic-scale restructuring pathways, and kinetically relevant structural evolution under reaction conditions. We emphasize the recent progress in machine learning potentials and multiscale modeling frameworks, which are extending simulations towards increasingly realistic chemical complexity. We also discuss the key challenges that remain, especially the need for open-system models that explicitly account for interfacial exchange, dissolution-redeposition processes, and the coupling between the local chemical environment and structural dynamics. This Review aims to provide a practical guide to the current computational toolbox and to outline directions for the next generation of predictive modeling for electrocatalyst reconstruction.
Liu et al. (Thu,) studied this question.