Lal, M.A.; Singh, A.; Mzik, R.; Lanjan, A.; Srinivasan, S. AI-Based Nano-Scale Material Property Prediction for Li-Ion Batteries. Batteries2024, 10, 51.
Lal, M.A.; Singh, A.; Mzik, R.; Lanjan, A.; Srinivasan, S. AI-Based Nano-Scale Material Property Prediction for Li-Ion Batteries. Batteries 2024, 10, 51.
Lal, M.A.; Singh, A.; Mzik, R.; Lanjan, A.; Srinivasan, S. AI-Based Nano-Scale Material Property Prediction for Li-Ion Batteries. Batteries2024, 10, 51.
Lal, M.A.; Singh, A.; Mzik, R.; Lanjan, A.; Srinivasan, S. AI-Based Nano-Scale Material Property Prediction for Li-Ion Batteries. Batteries 2024, 10, 51.
Abstract
Molecular dynamics (MD) simulations present a sophisticated nano-scale computational approach that can play a critical role in material design for next generation batteries. One critical piece of information needed for MD simulations is the non-bonded potential parameters, which can be obtained through quantum mechanics (QM) calculations or experimental methods. However, experimental data is not available for exploratory and novel materials to derive the potential parameters. Also, the QM approach needs significant computational power and is too time consuming for large systems. On the other hand, current Artificial Intelligence techniques are faster, but they cannot be generalized across molecules and properties. As a result, there is a significant barrier to discovering and exploring new materials. Overcoming this barrier, in this work, we propose a Machine Learning (ML) based technique that can learn inter-atomic potential parameters for various particle-particle interactions employing QM calculations. This ML model can be used as an alternative for QM calculations for predicting non-bonded interactions in a computationally efficient manner. Using these parameters as input to MD simulations, we can predict a diverse range of properties, enabling researchers to design new and novel materials suitable for various applications in the absence of experimental data. We employ our ML-based technique to learn the Buckingham potential, a non-bonded interatomic potential. Subsequently, we utilize these predicted values to compute the densities of four distinct molecules, achieving an accuracy exceeding 93%. This serves as a strong demonstration of the efficacy of our proposed approach.
Chemistry and Materials Science, Materials Science and Technology
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
