Regression Analyses of High Entropy Alloys

High entropy alloys are a novel class of materials with five or more elements alloyed together at relatively equiatomic ratios, unlike traditional alloys with one primary element and secondary additives. These alloys have shown potential for better mechanical and functional properties than traditional alloys. However, the new regions of the phase diagram they unlock are unfeasible to thoroughly map with current computational and experimental methods.

In this work, we collate published experimental measurements to form an exhaustive dataset of the lattice constant in FCC phase Cantor alloy across varying compositions. By utilizing data science and machine learning regression models, we successfully create a model to predict the lattice constant as a function of elemental composition. In doing so, we gain new insights into the mechanisms governing the formation of the uniform lattice in high entropy alloys. Such data science approaches enable greater utilization of experimental data, but are hindered by limited data availability.