About this Abstract |
| Meeting |
2027 TMS Annual Meeting & Exhibition
|
| Symposium
|
2D Materials – Preparation, Properties, Modeling & Applications
|
| Presentation Title |
Atomistic Insights into the Graphite-to-Diamond Phase Transformation via Deep Neural Network Potential Molecular Dynamics |
| Author(s) |
Mehrab Lotfpour, Haoran Cui, John Crosby, Lei Cao, Yan Wang |
| On-Site Speaker (Planned) |
Mehrab Lotfpour |
| Abstract Scope |
We present a molecular dynamics (MD) investigation of the graphite-to-diamond phase transformation using a newly developed deep neural network (DNN) interatomic potential trained on ab initio molecular dynamics data. This high-fidelity potential captures the complex bonding environment of carbon under extreme conditions with quantum-level accuracy and MD-level efficiency. Large-scale simulations of graphite in AB and AA stacking configurations reveal that hydrostatic pressures exceeding 100 GPa produce polycrystalline cubic diamond with grain boundaries interspersed by hexagonal diamond regions, while non-hydrostatic compression — achieved by selectively reducing pressure along specific crystallographic directions — preferentially stabilizes pure hexagonal diamond, demonstrating a controllable, phase-selective transformation pathway. Mechanistically, cubic diamond forms via in-plane sliding of graphene layers, whereas hexagonal diamond arises from an out-of-plane buckling process, revealing distinct structural signatures for each polymorph. These findings demonstrate the power of machine-learned potentials for modeling high-pressure phase transitions and provide a predictive framework for stress-controlled diamond synthesis. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Phase Transformations, Computational Materials Science & Engineering, Modeling and Simulation |