Prediction of CO<sub>2</sub> Reduction Reaction Intermediates and Products on Transition Metal-Doped γ-GeSe Monolayers: A Combined DFT and Machine Learning Approach
Xuxin Kang, Wenjing Zhou, Ziyuan Li, Zhaoqin Chu, Hanqing Yin, Shan Gao, Aijun Du, Xiangmei Duan
Abstract
Accurate prediction of free energy changes (Δ G ) for the vast network of reaction intermediates in the electrocatalytic CO 2 reduction reaction (CO 2 RR) is essential for evaluating catalytic performance. We combined density functional theory (DFT) and machine learning (ML) to screen 25 single-atom catalysts (SACs) on defective γ-GeSe monolayers for CO 2 reduction to methanol, methane, and formic acid. Among nine ML models evaluated with 14 features, the XGBoost performed best (R 2 = 0.92 and MAE = 0.24 eV), identifying Ni, Ru, and Rh@GeSe as prospective catalysts. Feature importance analysis highlighted CO 2 activation with ∠O–C–O and IP C–O1 as the key attributes. The trained ML model’s Δ G predictions closely match DFT-calculated values for the reported Ti@N 4 –C, Fe@g-C 6 N 6, and Co@g-C 6 N 6 . Incorporating non-DFT-based features enabled rapid predictions while retaining model performance. This work identifies effective SACs for CO 2 RR and offers insights for efficient catalyst design.