Evaluation of Figure of Merit of Thermoelectric Materials Using Machine Learning

Abstract

This work involves the use of a suitable machine learning algorithm to explore new thermoelectric (TE) materials by establishing a relationship between the figure of merit (ZT) of a compound and the different physical and chemical characteristics of thermoelectric compounds like Lead Telluride, Cobalt Antimonide and Germanium-based Clathrates etc. Thermoelectric materials have a wide range of research importance because of their capability of converting waste heat energy to electrical energy. The current scenario involves three approaches to find better thermoelectric materials: classical experimental approach, Density Functional Theory (DFT) and Machine Learning methods. Machine Learning can be used for material screening processes effectively in less computational time, with good efficiency and accurate data acquisition. This present study may be helpful for understanding the importance and use of machine learning in thermoelectric materials.