High-throughput prediction of the carrier relaxation time via data-driven descriptor

Zizhen Zhou,Huijun Liu,Guohua Cao,Jianghui Liu

doi:10.1038/s41524-020-00417-0

Abstract

It has been demonstrated that many promising thermoelectric materials, such as tetradymite compounds are also three-dimensional topological insulators. In both cases, a fundamental question is the evaluation of carrier relaxation time, which is usually a rough task due to the complicated scattering mechanisms. Previous works using the simple deformation potential theory or considering complete electron-phonon coupling are, however, restricted to small systems. By adopting a data-driven method named SISSO (Sure Independence Screening and Sparsifying Operator) with the training data obtained via deformation potential theory, we propose an efficient and physically interpretable descriptor to evaluate the relaxation time, using tetradymites as prototypical examples. Without any input from first-principles calculations, the descriptor contains only several elemental properties of the constituent atoms, and could be utilized to quickly and reliably predict the carrier relaxation time of a substantial number of tetradymites with arbitrary stoichiometry.

Highlights

Substantial advances have been made in screening or discovering functional materials via high-throughput computational method, which involves physics, statistics, computer science, and artificial intelligence[1]
By employing the data-driven SISSO approach, we report two simple descriptors for predicting the carrier relaxation time of normal insulators (NIs) and topological insulators (TIs) in the tetradymite family, where the training sets are obtained from accurate first-principles calculations
Unlike those from the deformation potential (DP) theory or electron-phonon coupling (EPC) calculations which is forbidden for systems with large unit cell, our descriptors enable an effective prediction for the relaxation time of the tetradymite compounds with either integer or fractional stoichiometry at negligible computational cost

Summary

Introduction

Substantial advances have been made in screening or discovering functional materials via high-throughput computational method, which involves physics, statistics, computer science, and artificial intelligence[1]. There have been a few works focusing on the high-throughput prediction of compounds with low thermal conductivity[10,11,12], which is essential for high TE conversion efficiency Another critical transport coefficient closely related to the TE performance is the so-called power factor[7], where a troublesome issue is suitable treatment of the relaxation time owing to the complex scattering mechanisms involved. The energy dependent k-resolved relaxation time can be obtained by a complete electron-phonon coupling (EPC) calculation[15,16] based on the Wannier interpolation techniques[17] Both approaches involve quite complex first-principles calculations and huge computational efforts are needed. A high-throughput investigation combined with the compressed sensing approach[18,19] is employed to quickly predict the carrier relaxation time of the tetradymite compounds

Methods

Results

Conclusion