Fusing machine learning strategy with density functional theory to hasten the discovery of MXenes for hydrogen generation

Abstract

The complexity of the topological and combinatorial configuration space of MXenes can give rise to gigantic design challenges that cannot be addressed through traditional experimental or routine theoretical approaches. To this end, we establish a robust and more broadly applicable multistep workflow from the toolbox of supervised machine learning (ML) algorithms for predicting the hydrogen evolution reaction (HER) activity over 4,500 MM$^{\prime}$XT$_2$-type MXenes, where 25\% of the material space (1125 systems) is randomly selected to evaluate the HER performance using density functional theory (DFT) calculations. As the most desirable ML model, the random forest regression method with recursive feature elimination and hyperparameter optimization accurately and rapidly predicts the Gibbs free energy of hydrogen adsorption ($Δ$G$_{H}$) with a low predictive mean absolute error of 0.374 eV. Based on these observations, the H-atom adsorbed directly on top of the outermost metal atomic layer of the MM$^{\prime}$XT$_2$-type MXenes (site-2) with Nb, V, Mo, Cr and Ti metals composed of carbon based O-functionalization are discovered to be highly stable and active catalysts, surpassing that of commercially available platinum based counterparts. Overall, the physically meaningful predictions and insights of the developed ML/DFT-based multistep workflow will open new avenues for accelerated screening, rational design and discovery of potential HER catalysts.