Understanding the hydrogen evolution reaction activity of doped single-atom catalysts on two-dimensional GaPS4 by DFT and machine learning

As a zero-carbon fuel, hydrogen can be produced via electrochemical water splitting using clean electric energy by the hydrogen evolution reaction (HER) process. The ultimate goal of HER catalyst is to replace the expensive Pt metal benchmark with a cheap one with equivalent activities. In this work, we investigated the possibility of HER process on single-atom catalysts (SACs) doped on two-dimensional (2D) GaPS4 materials, which have a large intrinsic band gap that can be regulated by doping and tensile strain. Based on the machine learning regression analysis, we can expand the prediction of HER performance to more catalysts without expensive DFT calculation. The electron affinity and first ionization energy are the two most important descriptors related to the HER behavior. Furthermore, constrain molecular dynamics with solvation models and constant potentials were applied to understand the dynamics barrier of HER process of Pt SAC on GaPS4 materials. These findings not only provide important insights into the catalytic properties of single-atom catalysts on GaPS4 2D materials, but also provides theoretical guidance paradigm for exploration of new catalysts.