Machine learning-assisted design of transition metal-doped 2D WSn₂N₄ electrocatalysts for enhanced hydrogen evolution reaction

Abstract

Hydrogen energy is characterized as environmentally friendly and resourceful. The hydrogen evolution reaction (HER) is a crucial process for hydrogen production and is essential for achieving a transition to sustainable and clean energy. In this work, we conducted a systematic investigation into the hydrogen catalytic activity of two-dimensional WSn 2 N 4 materials. The transition metal atoms (Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn) were substituted at the N and Sn sites in WSn 2 N 4 , respectively. The catalyst's performance was evaluated to verify its catalytic activity for the HER. The results indicate that the Gibbs free energy changes (ΔG H∗ ) of Ti@W–WSn 2 N 4 , V@W–WSn 2 N 4 , Fe@W–WSn 2 N 4 , Co@W–WSn 2 N 4 , and Ni@W–WSn 2 N 4 are close to zero. Among the structures examined, Ti@W–WSn 2 N 4 exhibited the lowest Gibbs free energy change (ΔG H∗ = −0.01 eV), indicating a high degree of catalytic activity. Through machine learning analysis, key features affecting catalytic activity could be directly identified, and a framework for rapid screening was established. This study lays a solid foundation for the design and development of potential HER catalysts in the future.