First-Principles\nCalculations and Machine Learning\nof Hydrogen Evolution Reaction Activity of Nonmetallic Doped β‑Mo<sub>2</sub>C Support Pt Single-Atom Catalysts

The most widely used catalyst for the hydrogen evolution\nreaction\n(HER) is Pt, but the high cost and low abundance of Pt need to be\nurgently addressed. Single-atom catalysts (SACs) have been an effective\nmeans of improving the utilization of Pt atoms. In this work, we used\na nonmetal (NM = B, N, O, F, Si, P, S, Cl, As, Se, Br, Te, and I)\ndoped β-Mo<sub>2</sub>C (100) C-termination surface as the support,\nwith Pt atoms dispersed on the support surface to construct Pt@NM-Mo<sub>2</sub>C. Using density functional theory (DFT) calculations, we\nselected catalysts with excellent HER activity. Among 117 candidate\ncatalysts, 49 catalysts exhibited ideal catalytic performance with\nGibbs free energy of hydrogen intermediate (H*) adsorption (Δ<i>G</i><sub>H*</sub>) values less than 0.2 eV. The Δ<i>G</i><sub>H*</sub> values of 16 catalysts were even lower than\nthat of Pt (Δ<i>G</i><sub>H*</sub> ≈ 0.09 eV),\nwith Pt<sub>I</sub>@N<sub>2/4‑a</sub>-Mo<sub>2</sub>C demonstrating\nthe best performance (Δ<i>G</i><sub>H*</sub> = −0.01\neV). Combined with electronic structure analysis, we could understand\nthe impact of charge transfer between Pt and the underlying NM atoms\non the strength of the Pt–H bond, thereby promoting HER activity.\nUsing machine learning (ML), we identified that the primary influencing\nfactors of the HER catalytic activity in the Pt@NM-Mo<sub>2</sub>C\nsystem were the Bader charge transfer of Pt (<i>N</i><sub>ePt</sub>), the d-band center of Pt (ε<sub>dPt</sub>), and\nthe atomic radius of NM (<i>R</i><sub>NM</sub>), with <i>N</i><sub>ePt</sub> having the greatest impact on the HER catalytic\nactivity.