| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Chemistry and Physics of Interfaces
|
| Presentation Title |
Machine Learning Guided Prediction of Solute Segregation at Metal/Oxide
Interfaces |
| Author(s) |
Yizhou Lu, Blas Uberuaga, Samrat Choudhury |
| On-Site Speaker (Planned) |
Yizhou Lu |
| Abstract Scope |
Investigation of semi-coherent metal/oxide interfaces with misfit dislocations using density
functional theory (DFT) is computationally intensive to the point of being prohibitive, as it
involves several hundreds to thousands of atoms. In this study, we examined the solute
segregation behavior at the Fe/Y2O3 interface—a model interface for cladding applications in
nuclear fission reactors—using a combination of DFT calculations and machine learning (ML)
approaches. ML models were trained on DFT-calculated segregation energies to identify the key
chemical and geometric features that govern solute segregation behavior at coherent and semi-
coherent metal/oxide interfaces. Furthermore, it was found that ML models when trained with
DFT calculated segregation energy of elements at a coherent interface, comprising of about a
hundred-atom supercell, can predict the segregation energy of elements at a semi-coherent
Fe/Y2O3 interface, with DFT level accuracy at a fraction of the computational cost needed for
similar DFT calculations. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Thin Films and Interfaces, Computational Materials Science & Engineering, Machine Learning |