Molecular Dynamics Study of OH-Induced Disintegration of Cu/ZnO Catalysts Based on Machine Learning Potentials

Supported metal catalysts have attracted significant attention in the field of heterogeneous catalysis due to their excellent catalytic performance and wide-ranging industrial applications. Among them, Cu/ZnO catalysts, as a typical supported metal system, exhibit outstanding activity in key industrial reactions such as the water-gas shift reaction (WGSR) and methanol synthesis. However, their stability under H2O-rich conditions remains a critical issue that affects both the catalyst's long-term durability and catalytic efficiency. In this study, we investigate the thermodynamic stability and dynamic evolution of Cu/ZnO catalysts in the presence of H2O using first principles calculation combined with molecular dynamics (MD) simulations based on machine learning potentials. Our results reveal that H2O molecules can interact with isolated Cu atoms to form Cu–OH complexes, significantly lowering the formation energy of Cu single atoms on the ZnO(101 ̅0) surface. Further MD simulations under given atmospheric and simulation conditions were performed. Small Cu clusters (Cu8 and Cu18) are more prone to forming Cu(OH)2 complex, whereas no Cu–OH complexes are observed in larger clusters (Cu55). This work elucidates the destabilizing role of water molecules on Cu/ZnO catalysts and deepens our understanding of their structural evolution under H2O-rich environments. The findings provide valuable theoretical insights and data support for optimizing the stability of Cu/ZnO catalysts under reaction conditions.