npj Computational Materials

High-throughput deformation potential and electrical transport calculations

Yeqing Jin, Xiangdong Wang, Mingjia Yao, Di Qiu, David J. Singh, Jinyang Xi, Jiong Yang, & Lili Xi

npj Computational Materials 9: 190 (2023)

doi.org/10.1038/s41524-023-01153-x

Published online: 14 October 2023
Abstract| Full Text | PDF OPEN

Abstract: The deformation potential plays an important role in electrical transport properties, but in the context of high-throughput searches requires a consistent and readily computable reference level. Here, we design a high-throughput method for calculating the deformation potential for semiconductors in the MatHub-3d database. The deformation potential is obtained from the volume-dependent band edge (valence band maximum and conduction band minimum) variations with respect to the reference level. In order to adapt the calculation of reference level to a high-throughput manner, the average value of the first valence electron energy band associated with each compound is used as the reference level. Deformation potentials for 11,993 materials are calculated using the above-mentioned method. By considering the calculated deformation potentials, electronic structures, and bulk moduli, 9957 compounds with converged electrical transport properties are obtained. 332 promising p-type thermoelectric materials, and 321 n-type entries are then proposed. Among them, 156 compounds are screened to having both potentially good n- and p-type thermoelectric transport properties. The band structures and chemical bonding information for several typical compounds are further presented to reveal favorable band and bonding features for thermoelectrics.

摘要: 形变势在电输运特性中起着重要作用，但在高通量计算下，参考能级的选择至关重要。在这里，我们设计了一种高通量方法，用于计算MatHub-3d数据库中半导体的形变势。形变势是从相对于参考能级的体积相关的带边位置（价带顶和导带底）变化中获得的。为了使参考能级的计算适应高通量计算，采用与每种化合物相关的第一条价电子能带的平均值作为参考能级。用上述方法计算了11993种材料的形变势。通过考虑计算的形变势、电子结构和体积模量，获得了9957种化合物的电输运性质。进一步筛选出332种有前景的p型热电材料和321种n型材料，其中，156种化合物被筛选出具有潜在良好的n型和p型热电输运性质。对于几种典型结构的能带结构和化学键性质的分析进一步揭示了对热电有利能带特征和化学键特点。

Editorial Summary

High throughput calculation of deformation potential

The uncertainty of the relaxation time of the electrical transport property of thermoelectric materials makes its prediction results difficult to believe, and the deformation potential method based on the electron-phonon coupling approximation is considered an efficient and reliable calculation method. This study proposes a high-throughput deformation potential calculation method based on the average value of the first valence electron energy band. This method can efficiently and quickly calculate the deformation potential of materials at the valence band maximum (VMB) and conduction band minimum (CBM), and based on this, electrical transport properties are calculated, and new high-performance thermoelectric materials are predicted by efficiently and accurately high-throughput way. This work was completed by Professor Yang Jiong and Associate Researcher Xi Lili from the Materials Genome Institute of Shanghai University. In the high-throughput processing of deformation potential, the band alignment method selected the average value of the first valence electron energy band as the reference level, avoiding the uncertainty of different element core energy levels in complex compounds, and obtained over 10000 bar deformation potential data and electrical transport properties. 332 potential n-type thermoelectric materials and 321 p-type thermoelectric materials were identified, resulting in 156 new high-performance thermoelectric materials with excellent n-type and p-type performance. The high-throughput deformation potential processing method proposed in this work has obtained a large amount of deformation potential data and electrical transport performance parameters, which can systematically analyze the distribution pattern and influencing factors of deformation potential. It also obtains the relationship between deformation potential and electrical transport performance, which is of great significance for accelerating the discovery of high-performance materials and providing a large amount of data support for further machine learning research.

形变势的高通量计算

热电材料电输运性能弛豫时间的不确定性，导致其预测结果可靠性较差，即使采用基于电声耦合近似的形变势方法，即所谓高效可靠的计算方法，也难逃可靠性不足的魔咒。上海大学材料基因组工程研究院的杨炯教授与席丽丽副研究员等，研究提出基于第一条能带平均值为基准的高通量形变势方法，可克服上述缺点，高效快速地计算出材料的价带顶和导带底的形变势，并以此为基础进行电输运性能计算研究，从而高效准确地预测新型高性能热电材料。他们建立的形变势高通量处理方法中，能带对齐方式选用第一条能带平均值为基准，避免了复杂化合物中不同元素的芯能级为基准的不确定性，获得了1万多条准确的形变势数据及电输运性能数据。再通过高通量预测出332种潜在的n型热电材料和321种p型热电材料，最终获得156种潜在的n型、p型性能均优异的新型高性能热电材料。作者提出的高通量形变势处理方法，所获得的大量形变势数据及电输运性能参数，可进一步用于系统地分析形变势的分布规律及影响因素，并获得了形变势与电输运性能的关系，对加速高性能材料的发现具有重要的意义，同时也为进一步的机器学习研究提供了大量的数据基础。