Transition metal single-atom supported on PC3 monolayer for highly efficient hydrogen evolution reaction by combined density functional theory and machine learning study

TM@PC3 as single-atom catalyst with high HER activity by the joint study of DFT calculation and machine Learning. [Display omitted] •ΔG*H of V, Fe, Nb, Mo, and Pd@PC3 are lower than that of Pt (111) catalyst.•V, Fe, Nb, Mo, and Pd@PC3 follow Volmer-Tafel mechanism during HER.•Fe, Nb, and Mo@PC3 show lower activation barriers than of Pt metal.•The I, dTM−H and εd are more correlated with H atom adsorption. It is essential to develop non-precious metal-based alternatives used in hydrogen evolution reaction (HER) due to high cost and scarcity of Pt-based catalysts. Herein, through density functional theory (DFT) calculations, the HER activity over 26 single-atom anchored phosphorus carbide (PC3) monolayer (TM@PC3) has been systematically investigated. Results indicate that ΔG*H of V, Fe, Nb, Mo, and Pd@PC3 are lower than that of Pt (111) catalyst, with 0.03, −0.03, −0.07, −0.04, and − 0.02 eV, respectively. By imposing the criterion window (−0.2 ≤ ΔG*H ≤ 0.2 eV), the d band centre (εd) for catalysts with excellent HER ability is in the range of − 0.68–0.41 eV. Besides, the five promising HER catalysts follow Volmer-Tafel mechanism. Fe, Nb, and Mo@PC3 show activation barriers of 0.75, 0.74, and 0.55 eV, lower than that of Pt. Machine learning (ML) was employed to explore the intrinsic relationship between catalytic performance and feature parameters. We demonstrated that the first ionization energy, bond length of TM − H and d band center are more correlated with hydrogen adsorption behaviour. Our work not only predicts that Fe, Nb, and Mo@PC3 can be substitutes for Pt metal in HER, but also reveals that the intrinsic correlation between catalytic activity and feature parameters by combining DFT and ML investigations.