Machine learning screening of high-performance single-atom electrocatalysts for two-electron oxygen reduction reaction

Xuqian Zhang; Jiming Liu; Rui Li; Xuan Jian; Xiaoming Gao; Zhongli Lu; Xiuping Yue

doi:10.1016/j.jcis.2023.05.011

Machine learning screening of high-performance single-atom electrocatalysts for two-electron oxygen reduction reaction

J Colloid Interface Sci. 2023 Sep:645:956-963. doi: 10.1016/j.jcis.2023.05.011. Epub 2023 May 9.

Authors

Xuqian Zhang¹, Jiming Liu², Rui Li¹, Xuan Jian³, Xiaoming Gao⁴, Zhongli Lu¹, Xiuping Yue¹

Affiliations

¹ College of Environmental Science and Engineering, Taiyuan University of Technology, Taiyuan 030024, Shanxi Province, People's Republic of China.
² College of Environmental Science and Engineering, Taiyuan University of Technology, Taiyuan 030024, Shanxi Province, People's Republic of China. Electronic address: liujiming@tyut.edu.cn.
³ College of Environmental Science and Engineering, Taiyuan University of Technology, Taiyuan 030024, Shanxi Province, People's Republic of China; College of Chemistry and Chemical Engineering, Shaanxi Key Laboratory of Chemical Reaction Engineering, Yan'an University, Yan'an 716000, Shaanxi Province, People's Republic of China. Electronic address: jianxuan@yau.edu.cn.
⁴ College of Chemistry and Chemical Engineering, Shaanxi Key Laboratory of Chemical Reaction Engineering, Yan'an University, Yan'an 716000, Shaanxi Province, People's Republic of China.

PMID: 37182327
DOI: 10.1016/j.jcis.2023.05.011

Abstract

Electrocatalysis has emerged as one of the most promising alternatives to conventional anthraquinone for preparing hydrogen peroxide (H₂O₂) with high energy consumption and pollution because of its simplicity, convenience, and environmental friendliness. However, the oxygen reduction reaction (ORR) generating H₂O₂viathe2e^- path is acompetitive path for 4e^-ORR to generate H₂O. Therefore, it is crucial to identify an electrocatalyst with high selectivity and activity of 2e^-ORR. Here, we established five machine learning (ML) models based on the adsorption free energy of O* (△G (O*)) of 149 single-atom catalysts (SACs) collected and the limiting potential (U_L) of 31 SACs calculated using density functional theory (DFT) from the literature. We then obtained descriptors that could accurately describe SACs. Furthermore, 690 unknown SACs' 2e^-ORR catalytic performance was well predicted. Four 2e^-ORR materials with high selectivity and activity were screened: Zn@Pc-N₃C₁, Au@Pd-N₄, Au@Pd-N₁C₃, and Au@Py-N₃C₁. We verified the U_L of these SACs through DFT calculation, which was higher than the standard value, proving the ML model's validity. The ML-based method to predict the material properties with highly selective and active electrocatalysts provides an efficient, rapid, and low-cost method for discovering and designing more valuable SACs catalysts.

Keywords: 2e(-)ORR; H(2)O(2); High activity; High selectivity; Random forest; Single-atom catalysts.