Predicting phase preferences of two-dimensional transition metal dichalcogenides using machine learning

Two-dimensional transition metal dichalcogenides (TMDs) can adopt one of\nseveral possible structures, with the most common being the trigonal prismatic\nand octahedral symmetry phases. Since the structure determines the electronic\nproperties, being able to predict phase-preferences of TMDs from just the\nknowledge of the constituent atoms is highly desired, but has remained a\nlong-standing problem. In this study, we applied high-throughput quantum\nmechanical computations with machine learning algorithms to solve this old\nproblem. Our analysis provides insights into determining physiochemical factors\nthat dictate the phase-preference of a TMD, identifying and going beyond the\nattributes considered by earlier researchers in predicting crystal structures.\nA knowledge of these underlying physiochemical factors not only helps us to\nrationalize, but also to accurately predict structural preferences. We show\nthat machine learning algorithms are powerful tools that can be used not only\nto find new materials with targeted properties, but also to find connections\nbetween elemental attributes and the target property/properties that were not\npreviously obvious.\n