AI-based Nano-scale Material Property Prediction For Li-Ion Batteries

Introducing a novel Machine Learning (ML) approach to address the challenge of deriving non-bonded potential parameters for Molecular Dynamics (MD) simulations. Traditional methods like Quantum Mechanics (QM) and experimental data are limited for new materials. Our ML model efficiently learns inter-atomic potential parameters using QM calculations, serving as a faster alternative. We focus on predicting the Buckingham potential, a non-bonded interatomic potential, and demonstrate its effectiveness by achieving over 93% accuracy in predicting the densities of four distinct molecules. This approach opens doors for designing materials without extensive computational or experimental barriers.