Exploring structures of Pd clusters on hydrogenated ceria surface using High-Dimensional neural network potential

HB Cai and YJ Zhang and QL Tang and ZK Han and Y Gao, APPLIED SURFACE SCIENCE, 692, 162780 (2025).

DOI: 10.1016/j.apsusc.2025.162780

Palladium supported by CeO2 is an attractive catalyst for hydrogenation reactions in which the hydrogenated oxide surface plays a crucial role. However, the effect of surface hydrogenation on the geometry and electronic structures of supported metal catalysts is still unclear due to the challenge of efficiently sampling their complex structures. Herein, we developed a neural network potential with the accuracy of density functional theory calculations to explore the configurations of Pd clusters on hydrogenated ceria. Using Pd6 clusters supported on hydrogenated CeO2(111) as an example, we demonstrated the most stable structure of the Pd cluster on the hydrogenated ceria surface was not necessarily the same as the most stable but even a significantly less stable structure on the pristine surface. Further analysis indicated that the surface hydrogen atoms on ceria could affect the charge transfer between the Pd cluster and CeO2, thus influencing the catalytic activity. The dynamic structural evolution of Pd clusters also varied with different surface H coverage and configurations. This work not only elucidated the unignored effect of surface H atoms on the supported metal catalysts, but also provided a benchmark for exploring the structures of supported metal catalysts under complex reaction conditions by neural network potentials.

Return to Publications page