基于全第一性原理计算预测几类潜在热电材料的热电优值

Parameter-free first-principles calculations have been playing increasingly important role in exploring new materials with desired properties. However, fully first-principles calculations of thermoelectric properties are still lacking due to the limitation of electrical transport calculation. The calculation of electron-phonon coupling, which dominates electron scattering, is extremely challenging. Therefore, conventionally the carrier lifetime, one critical quantity affecting transport properties, is simply approximated with parameterized constant time. This type of calculation is not fully first-principles, and does not allow for practical guidance to experiments. In this proposal, we aim to realize fully first-principles calculations of thermoelectric figure of merit ZT with high accuracy by developing methods for calculating electron-phonon coupling, impurity scattering, electronic thermal conductivity and Seebeck coefficient and combining with first-principles calculation of lattice thermal conductivity. We further apply this fully-principles method to several classes of materials with ultralow thermal conductivity including FeSb3, half-Heusler and full-Heusler, in the hope of discovering some new thermoelectric materials. Through calculations we also aim to understand complex transport behaviors, reveal mechanism leading to high ZT and provide theoretical guidance to experiments searching for high ZT.

不需要任何可调参数的第一性原理计算在新材料探索方面起到了越来越重要的作用，但是由于电输运计算所限，当前还缺乏对热电输运性质的真正意义上的全第一性原理计算。由于对主导电子散射的电声耦合的计算具有相当的挑战性，通常对决定输运性质的核心物理量载流子寿命直接做参数化常数时间近似，这种半第一性原理计算缺乏对实验的指导意义。本项目将发展高效电声耦合输运计算方法，并考虑杂质散射，计算电导率、电子热导率和Seebeck系数，同时结合晶格热导率的第一性原理计算，实现高精度的热电优值ZT的全第一性原理计算。进而研究FeSb3、half-Heusler 和full-Heusler几类具有极低热导率的潜在热电材料的ZT，期望从这些材料中发现一些新型热电材料。通过计算理解材料的复杂输运性质，探索热电材料ZT优化的关键机理，为高性能热电材料的实验研究提供有效的理论支持。

不需要任何可调参数的第一性原理计算在新材料探索方面起到了越来越重要的作用，但是由于决定电输运性质的电子声子耦合计算具有相当的挑战性，此前还缺乏对热电输运性质的真正意义上的全第一性原理计算。本项目发展了高效电声耦合输运计算方法，同时结合晶格热导率的第一性原理计算，实现了高精度的热电优值ZT的全第一性原理计算。进而预测了几种ZT大于1.5的新型热电材料，包括p型Mg3Sb2, 二维碲烯和full Heusler 材料Ba2AuX (X=As, Sb, and Bi)。通过计算理解了包括反常Lorenz数在内的一些材料复杂输运性质。在本项目资助下共发表通讯作者论文14篇。