About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Algorithms Development in Materials Science and Engineering
|
| Presentation Title |
SAGA: A Simulated Annealing / Genetic Algorithm Approach to Atom Probe Tomography Rectification |
| Author(s) |
Dylan Miley, Manisha Tripathy, Jeremy Mason, Gregory Thompson |
| On-Site Speaker (Planned) |
Dylan Miley |
| Abstract Scope |
Atom probe tomography (APT) provides atomic-level structural and chemical information throughout a material volume. However, the positional uncertainty and limited detection efficiency pose major challenges to the study of chemical short range order using APT. We propose a method to rectify APT data using open-source atomistic molecular-dynamics software. The standard interatomic potential energy is amended to include a harmonic potential that tethers APT atoms to their reconstructed positions. Filler atoms are placed randomly to bring the sample to the expected density and atomic proportions. A simulated annealing schedule creates crystal nuclei within the sample and a subsequent genetic algorithm promotes growth of nuclei at lower temperatures and disfavors unfavorable nuclei at higher temperatures. This results in a plausible, low-energy atomic configuration that is consistent with the reconstructed atomic positions and positional uncertainty, but notably does not require prior assumptions about the symmetry, number, or orientation of crystals in the sample. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Computational Materials Science & Engineering, Characterization, Modeling and Simulation |