A Robust Machine Learned Interatomic Potential for Nb: Collision Cascade Simulations with accurate Defect Configurations

Abstract: Niobium (Nb) and its alloys are extensively used in various technological applications owing to their favorable mechanical, thermal and irradiation properties. Accurately modeling Nb under irradiation is essential for predicting microstructural changes, defect evolution, and overall material performance. Traditional interatomic potentials for Nb fail to predict the correct self-interstitial atom (SIA) configuration, a critical factor in radiation damage simulations. We develop a machine learning interatomic potential (MLIP) using the Spectral Neighbor Analysis Potential (SNAP) framework, trained on ab-initio Density Functional Theory (DFT) calculations, which accurately captures the relative stability of different SIA dumbbell configurations. The resulting potential reproduces DFT-level accuracy while maintaining computational efficiency for large-scale Molecular Dynamics (MD) simulations. Through a series of validation tests involving elastic, thermal, and defect properties -- including collision cascade simulations -- we show that our SNAP potential resolves persistent limitations in existing Embedded Atom Method (EAM) and Finnis--Sinclair (FS) potentials and is effective for MD simulations of collision cascades. Notably, it accurately captures the ground-state SIA configuration of Nb in the primary damage of a collision cascade, offering a robust tool for predictive irradiation studies.