About this Abstract |
| Meeting |
2027 TMS Annual Meeting & Exhibition
|
| Symposium
|
Computational Thermodynamics and Kinetics
|
| Presentation Title |
Automating the First-Principles to CALPHAD Pipeline for Discovery of Multicomponent Alloys with Magnetic Components |
| Author(s) |
Prajna Jalagam, Zhigang Wu, Axel van de Walle, John Lawson |
| On-Site Speaker (Planned) |
Prajna Jalagam |
| Abstract Scope |
Successes such as NASA's GRX-810 — an additively manufactured NiCoCr-based superalloy discovered through CALPHAD — show how CALPHAD can compress alloy development timelines, yet extending this paradigm across multicomponent space is bottlenecked by database coverage. We present an end-to-end automated pipeline (orchestrating ATAT, spin-polarized VASP, and PyCALPHAD) that facilitates construction of thermodynamic databases from first-principles data alone. Upstream, the workflow autonomously prepares magnetic configurations where ground states are unknown or vary across composition, generating SQS with disordered-local-moment and ferromagnetic spins and automating convergence, relaxation, and phonon calculations. Downstream, it prunes erroneous fits, selects Redlich-Kister complexity while guarding against overfitting, and scores candidate databases against known phase equilibria, adaptively refining SQS coverage where scoring trends demand. Magnetic ordering enters via composition-dependent Curie/Néel temperatures and moments β(x) in the Inden-Hillert-Jarl formalism that are also computed autonomously. We apply this to Co-Cr-Ni (GRX-810's base ternary) and discuss extrapolation toward multicomponent alloy design. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Phase Transformations, Modeling and Simulation, Magnetic Materials |