About this Abstract |
| Meeting |
2027 TMS Annual Meeting & Exhibition
|
| Symposium
|
Chemistry and Physics of Interfaces
|
| Presentation Title |
Simulating Epitaxial Growth and Dislocation Formation in GaN/AlN Heterostructures with Machine-Learned Interatomic Potentials |
| Author(s) |
Nicholas Taormina, Emir Bilgili, Simon Phillpot, Youping Chen |
| On-Site Speaker (Planned) |
Nicholas Taormina |
| Abstract Scope |
In this talk, we present large-scale molecular dynamics simulations investigating the heteroepitaxial growth of GaN on the wurtzite phase of AlN. These simulations are powered by a highly efficient and accurate machine-learned interatomic potential we developed specifically for Ga-Al-N semiconductors. The MLIP outperforms existing empirical potentials in accuracy and operates significantly faster than other MLIP, successfully reproducing mechanical, thermal, and surface properties in close agreement with density functional theory calculations. Leveraging this potential, we perform atomistic simulations of the epitaxial growth process. We successfully reproduce the experimentally observed wurtzite structure within the GaN overlayer. Furthermore, we capture the dynamic formation and evolution of dislocations as the epilayer grows. The resulting dislocation networks formed during the growth process are consistent with experimental observations. This work offers mechanistic insights into the formation, movement, and annihilation of dislocations during the manufacturing of wide bandgap semiconductors. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Modeling and Simulation, Thin Films and Interfaces |