Speaker
Description
Two-dimensional (2D) transition-metal dichalcogenides (TMDCs) exhibit a broad range of attractive electronic, optical, catalytic, and tribological properties. Incorporating multiple transition-metal species into the metal sublattice offers further opportunities for tuning these properties through the high-entropy alloy (HEA) concept. However, the vast configurational space of multicomponent 2D TMDCs makes their systematic investigation using density functional theory (DFT) alone computationally demanding.
In this work, we develop a machine-learning interatomic potential (MLIP) for the pentanary 2D high-entropy TMDC using the Gaussian Approximation Potential (GAP) framework. The performance of the MLIP is assessed by comparing its predicted energies and forces with DFT results and by evaluating its ability to reproduce DFT-optimized atomic structures.
The developed MLIP provides a computationally efficient framework for exploring the large configurational space of 2D TMDC high-entropy alloys beyond the practical limits of direct DFT calculations. The work establishes a foundation for large-scale atomistic simulations and further investigation of the structural and thermodynamic behavior of these complex 2D materials.