Modulation of Hydrogen Evolution Reaction Performance of MXenes by Doped Transition Metals: Comprehensive Exploration of High-Throughput Computing and Machine Learning

Due to the unique properties of MXenes, the doping of transition metals can modulate their catalytic properties and make them potential materials for hydrogen evolution reaction (HER). Nevertheless, the extensive combinatorial space poses a challenge for rapid screening of catalysts. To address this issue, we conducted high-throughput calculations on a series of transition metal atom-doped Ti₃CNO₂ and Zr₂HfCNO₂. Furthermore, the local structure and the corresponding electronic structure changes are analyzed, focusing on their influence on the HER properties. Furthermore, site identification features were introduced to train a multisite prediction model with a final model accuracy of R² = 0.97 and predicted the trend of hydrogen adsorption Gibbs free energy (ΔG_H*) across a range of MXenes structures, which were doped with TM atoms. The results show that Nb, Sc, Rh, W, Ti, and V doping resulted in |ΔG_H*| < 0.2 eV for more than 38 M′₂M″CNO₂, respectively, and they are effective dopant atoms for enhancing the catalytic ability of M′₂M″CNO₂. This study not only demonstrates the potential of doped TM atoms in enhancing the performance of MXenes HER but also highlights the importance of multisite prediction models in the rapid identification and development of efficient HER catalysts.