High-throughput computational evaluation of lattice thermal conductivity using an optimized Slack model

Abstract

High-throughput computational screening of materials with targeted thermal conductivity (κ) plays an important role in promoting the advancement of material design and enormous applications. The Slack model has been widely applied for the fast evaluation of κ with minimal time and resources, showing the potential capability of high-throughput screening of κ. However, after examining the Slack model on a large set of 353 materials, a huge discrepancy is found between the predicted κ and the correspondingly measured κ in experiments for some materials in addition to the generally overestimated κ by the Slack model. Thus, it is necessary to optimize the Slack model for efficiently and accurately evaluating κ. In this study, based on the high-throughput comparison of the κ predicted by the Slack model using elastic properties and those measured in experiments, an optimized Slack model is proposed. As a result, the κ predicted by the optimized Slack model agrees reasonably with the κ measured in experiments, which is much better than the previous prediction. The optimized Slack model proposed in this study can be used for further high-throughput computational evaluation of κ, which would be helpful for finding materials of ultrahigh or ultralow κ with broad applications.