Machine Learning Guided CALPHAD Design of Ru-Stabilized BCC B2 Refractory Alloys
Abstract
Refractory alloys with a ductile body-centered-cubic (BCC) matrix strengthened by ordered B2 precipitates offer a high-temperature analogue to the gamma/gamma-prime architecture of Ni-based superalloys. Ruthenium is particularly attractive as a B2 stabilizer because RuHf, RuZr, and RuTi can retain ordered phases well above 1300 C. In this work, equilibrium CALPHAD calculations were coupled with random-forest-guided active learning to explore a ten-element Nb-based, Ru-bearing composition space containing Nb, Ta, Mo, V, Ru, Ti, Zr, Hf, Al, and Y at 1 at.% resolution. Across 500 CALPHAD-evaluated alloys, the calculations reproduced the principal trends reported for the Ru-B2 design space. RuHf and RuZr remained stable to the solidus, RuTi commonly exhibited a solutionizing window, and Al-containing alloys preferentially formed competing sigma and A15 phases. The upper bound of the BCC+B2 field increased from a median of approximately 1570 C at 5 at.% Ru to approximately 1980 C near 9-10 at.% Ru. Among the group-IV additions, Hf, Zr, and Ti produced progressively lower two-phase stability. Re-screening using physically motivated criteria identified 100 alloys satisfying requirements for high-temperature BCC+B2 stability, absence of liquid, phase purity, and appropriate secondary-phase fraction, including 19 Ru-lean compositions and two independently reported HfRu-B2 alloys. The results establish practical compositional design rules for Ru-stabilized dual-phase refractory alloys and identify phase-specific BCC/B2 lattice misfit as a key target for future design cycles.
Related
- Machine-Learning-Compatible CALPHAD-type Optimization from Phase Equilibria by Auto-differentiation
- Benchmarking Universal Machine-Learning Interatomic Potentials for Temperature-Dependent Elasticity of Binary and High-Entropy Refractory Carbides
- Physics-guided machine learning for predicting complete tool influence functions in flexible ball-end polishing
- Machine-learning-guided exploration of domain walls in the hybrid improper ferroelectric Ca3Ti2O7
- Knowledge-Guided Construction of Machine Learning Potentials for Reactive Behavior in Energetic Materials
BibTeX
@misc{peterson2026machine,
title = {{Machine Learning Guided CALPHAD Design of Ru-Stabilized BCC B2 Refractory Alloys}},
author = {Nathan Peterson and Avik Mahata and Nick Beaver and Mohsen Kivy},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.14925},
eprint = {2609.14925},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.14925},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Machine Learning Guided CALPHAD Design of Ru-Stabilized BCC B2 Refractory Alloys
AU - Nathan Peterson
AU - Avik Mahata
AU - Nick Beaver
AU - Mohsen Kivy
PY - 2026
DA - 2026/09/14
PB - arXiv
DO - 10.48550/arxiv.2609.14925
UR - https://arxiv.org/abs/2609.14925
AB - Refractory alloys with a ductile body-centered-cubic (BCC) matrix strengthened by ordered B2 precipitates offer a high-temperature analogue to the gamma/gamma-prime architecture of Ni-based superalloys. Ruthenium is particularly attractive as a B2 stabilizer because RuHf, RuZr, and RuTi can retain ordered phases well above 1300 C. In this work, equilibrium CALPHAD calculations were coupled with random-forest-guided active learning to explore a ten-element Nb-based, Ru-bearing composition space containing Nb, Ta, Mo, V, Ru, Ti, Zr, Hf, Al, and Y at 1 at.% resolution. Across 500 CALPHAD-evaluated alloys, the calculations reproduced the principal trends reported for the Ru-B2 design space. RuHf and RuZr remained stable to the solidus, RuTi commonly exhibited a solutionizing window, and Al-containing alloys preferentially formed competing sigma and A15 phases. The upper bound of the BCC+B2 field increased from a median of approximately 1570 C at 5 at.% Ru to approximately 1980 C near 9-10 at.% Ru. Among the group-IV additions, Hf, Zr, and Ti produced progressively lower two-phase stability. Re-screening using physically motivated criteria identified 100 alloys satisfying requirements for high-temperature BCC+B2 stability, absence of liquid, phase purity, and appropriate secondary-phase fraction, including 19 Ru-lean compositions and two independently reported HfRu-B2 alloys. The results establish practical compositional design rules for Ru-stabilized dual-phase refractory alloys and identify phase-specific BCC/B2 lattice misfit as a key target for future design cycles.
ER -