Predicting the energetics and kinetics of Cr atoms in Fe-Ni-Cr alloys via physics-based machine learning

Abstract The energy and activation barrier distributions of Cr atoms in austenitic alloys are investigated over a multiplicity of modeling samples across a wide range of chemical (e.g. solid solutions vs. segregated states) and microstructural (e.g. bulk vs. grain boundaries) environments. Assisted with a physics-based machine learning algorithm, it is found that the thermodynamic and kinetic behaviors of Cr atoms can be reliably predicted according to the local electronegativity ( χ ) and free volume of local atomic packing (Vv). The corresponding predictive maps in the χ − V v parameter space are established, which are in line with existing experiments and validated by a parallel modeling with a different interatomic force field. The implications of the present study regarding its potential to guide the design of austenitic alloys with desired properties are also discussed.