Devlopment of a Machine-Learned Density Functional Tight Binding for TiH₂ Bulk and Surface Chemistry

Knowledge of chemical defect energies and kinetics is essential for assessing potential hydrogen storage materials like TiH₂, where hydrogen point defects need to be assessed accurately and efficiently. The Density Functional Tight Binding (DFTB) method is a highly efficient semi-empirical quantum approach that can accurately probe these properties, but can be challenging to parameterize for each system of interest due to the different bonding types that can occur. Here, we have created a machine learning-based approach for determining the DFTB models which is both rapidly optimized, systematically improvable, and highly transferable. Our method leverages the Chebyshev Interaction Model for Efficient Simulation (ChIMES), a reactive many-body molecular dynamics force field where interactions are represented by linear combinations of Chebyshev polynomials. In this work, we discuss our ChIMES/DFTB models for TiH₂, and show its accuracy for both bulk and surface properties. Our approach is easy to implement and can yield accurate DFTB models for a number of challenging materials and conditions where chemical events can be difficult to model with standard quantum approaches alone.