Exploring the Stable Structures of Cerium Oxide Nanoclusters using High-Dimensional Neural Network Potential

Huabing Cai,Qinghua Ren,Yi Gao

DOI: https://doi.org/10.1039/d3na01119d

IF: 5.598

2024-04-05

Nanoscale Advances

Abstract:Ceria clusters have been extensively applied in the industry owing to the extraordinary properties in oxygen storage and redox catalytic activities. However, their atomically precise structures have not been studied, because of lacking a reliable method to sample their complex structures efficiently. Here, we combined the neural network algorithm with the density functional theory calculations to establish a high-dimensional potential to search for global minimums of cerium oxide clusters. Using Ce14O28 as well as its reduced state Ce14O27 and oxidized state Ce14O29 with the ultra-small dimensions of ~1.0 nm as examples, we found these three clusters adopt pyramid-like structures with the lowest energyies obtained by exploring 100,000 configurations in large feasible spaces. Further neural network potential enhanced molecular dynamics indicated these cluster structures are stable at high temperature. The electronic structure analysis suggested these clusters are highly active and good for oxygen storage. This work demonstrated that neural network potentials could be useful for exploring the stable structures of metal oxides nanoclusters in practical applications.

materials science, multidisciplinary,nanoscience & nanotechnology,chemistry