| About this Abstract |
| Meeting |
MS&T24: Materials Science & Technology
|
| Symposium
|
Machine Learning and Simulations
|
| Presentation Title |
Estimation of Thermal Hysteresis in Zirconia Using Machine Learning Molecular Dynamics and Transition State Modelling |
| Author(s) |
Owen Thomas Rettenmaier, Srikanth Patala, Christopher Schuh |
| On-Site Speaker (Planned) |
Owen Thomas Rettenmaier |
| Abstract Scope |
Zirconia-based alloys can undergo reversible martensitic transformations and, when alloyed, demonstrate thermally activated shape memory behavior. The thermal hysteresis in the phase transformation has proven to be key when designing these materials for enhanced performance. In this work, we use atomistic simulations to develop a metric that is shown to estimate the hysteresis of the phase transformation. We use uncertainty-driven dynamics and iterative learning to train a machine learning inter-atomic potential with excellent stability and low force errors. This potential is employed in medium-scale nudged elastic band calculations to model nucleation and growth. The calculated energy barriers are then compared to the results of free energy calculations to compute an effective hysteresis, closely matching experimental trends. The methodology developed here has the potential to accelerate search within the composition space and design shape memory ceramics with enhanced reversibility. |