| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Bridging Scale Gaps in Multiscale Materials Modeling in the Age of Artificial Intelligence
|
| Presentation Title |
An Ultra-Fast Machine-Learning Potentials to Investigate the Phonon-Dislocation Interaction of Lead Selenide |
| Author(s) |
Md Rakib Hossain, Jason B Gibson, Ajinkya C Hire, Youping Chen, Richard Hennig, Phillpot Simon R |
| On-Site Speaker (Planned) |
Md Rakib Hossain |
| Abstract Scope |
Dislocations, common crystallographic defects, influence phonon dynamics in crystals. To investigate phonon-dislocation interaction in lead selenide (PbSe), we employ a multiscale strategy integrating quantum and atomistic simulations. This study focuses on developing the Ultra-Fast Force Fields (UF3) for PbSe where effective 2- and 3-body terms are fitted using B-spline basis sets with regularized linear regression. For structural diversity, training data is generated through high-pressure relaxations, ab initio MD calculations at various temperatures, and defect system relaxations. Further, we sample the relaxation trajectories of structures produced from a genetic algorithm structure search to provide structurally and compositionally diverse data. Also, proper weighting of our training data ensures efficient and accurate potentials, avoiding overfitting. Our findings suggest that, for lattice parameter, system energy, defect formation energies, and elastic constants calculations, the UF3 force fields predicts reasonable results for PbSe crystal compared to DFT with speeds that are thousands of times faster. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Computational Materials Science & Engineering, Machine Learning, Modeling and Simulation |