Revealing high-fidelity phase selection rules for high entropy alloys: A combined CALPHAD and machine learning study

We reveal high-fidelity new phase selection rules for high entropy alloys (HEAs) by combining CALPHAD calculations and the machine learning (ML) method. Employing Thermo-Calc and TCHEA3 database, we first generate more than 300,000 equilibrium phase data from 20 quinary families formed by the 8 elements of Al Co, Cr, Cu, Fe, Mn, Ni, and Ti, and choose initially 15 materials/physical descriptors. The eXtreme Gradient Boosting (XGBoost) method is then used to identify 5 most important descriptors that best delineate the single and mixed phases in the complex temperature-composition space of HEAs. The ML model trained by the 5 features is validated by 155 annealing experimental data points from 15 publications and then used to predict 213 new single-phase alloys with BCC and FCC structures of the alloy families of AlCrNiFeMn and AlCrCoNiFeTi. We also highlight the importance of equilibrium temperature and offer in-depth insights into the paradigm of composition-feature-phase of HEAs. On the basis of the 5 important features, we establish new phase selection rules for single FCC and BCC phases with a success rate above 90%, significantly outperforming all existing phase selection rules and providing a powerful tool for mapping single-phase in the complex temperature-composition space of HEAs. Graphical abstract [Display omitted] •300,000+ equilibrium data with 8 metal elements in the vast temperature-composition space were generated by CALPHAD calculations.•Machine learning model developed by generated data was validated by 155 experimental data and used to predict 213 new HEAs.•High fidelity phase selection rules were established based on large data and five important features identified by the ML model.•In-depth insights into the paradigm of composition-feature-phase of high entropy alloys were revealed.