High-Throughput Exploration\nof Ti–V–Nb–Mo\nCarbide MXenes Using Neural Network Potentials and Their Evaluation\nas Catalysts for Hydrogen Evolution Reaction

Realization of a sustainable hydrogen economy in the\nfuture requires\nthe development of efficient and cost-effective catalysts for its\nproduction at scale. MXenes (M<sub>n+1</sub>X<sub>n</sub>) are a class\nof 2D materials with ‘n’ layers of carbon or nitrogen\n(X) interleaved by ‘n+1’ layers of transition metal\n(M) and have emerged as promising materials for various applications\nincluding catalysts for hydrogen evolution reaction (HER). Their properties\nare intimately related to both their composition and their atomic\nstructure. Recently, high entropy MXenes were synthesized, opening\na vast compositional space of potentially stable and functionally\nsuperior materials. Detailed atomistic modeling enables us to systematically\nexplore this extensive design space, which is otherwise infeasible\nin experiments. We have developed a Neural Network Potential (NNP)\nto model (Ti<sub><i>x</i></sub>V<sub><i>y</i></sub>Nb<sub><i>z</i></sub>Mo<sub>p</sub>)<sub>n+1</sub>C<sub>n</sub> MXenes (x+y+z+p = 1; n = 1,2,3) by training against Density\nFunctional Theory (DFT) data in an active learning fashion. We then\nused the developed NNP to perform hybrid Monte Carlo-Molecular Dynamics\n(MC-MD) simulations to identify thermodynamically stable compositions\nand investigate the relative arrangement of transition metal atoms\nwithin and across layers. Thermodynamic stability increased with Mo\ncontent and its presence on the surface layer. We further investigated\nthe catalytic performance of stable MXenes for the HER and observed\nthat the center of the oxygen p-band (ε<sub>p</sub>) correlated\nwell with the energy of adsorption of a hydrogen atom ΔG(*H).\nSubsurface metal atoms significantly influenced the ΔG(*H) values\nat the surface via both ligand and strain effects. Our work expands\nthe space of potentially stable MXene compositions, providing targets\nfor synthesis and their evaluation in various applications.