Thermoelectric materials could play an important role in global sustainable energy. However, improving thermoelectric efficiency has proved difficult, largely due to the complex interdependence of electronic properties of solids. Early work by Ioffe has developed into the standard thermoelectric optimization paradigm of tuning the electronic carrier concentration in semiconductors. Although the localization theory of electrons by Anderson and Mott has developed in parallel, its potential for thermoelectrics optimization has not been explored. Here, we show that structural-disorder-induced electron localization also provides an effective optimization strategy for thermoelectric materials. By using a transport model that includes the relevant physics of localization, it is shown that the maximum thermoelectric figure of merit can be increased ∼20% by tuning both carrier concentration and disorder. The benefit of slight disorder is confirmed in two model Ge-Sb-Te material systems. Particularly for highly degenerate semiconductors, this bidimensional optimization strategy provides a new methodology to attain high thermoelectric performance.

热电材料可以在全球可持续能源中发挥重要作用。然而，事实证明，提高热电效率很困难，这主要是由于固体电子特性之间复杂的相互依赖关系。Ioffe 的早期工作已发展成为调节半导体中电子载流子浓度的标准热电优化范例。尽管安德森和莫特的电子定域理论已经并行发展，但尚未探索其在热电优化方面的潜力。在这里，我们展示了结构无序诱导的电子局域化也为热电材料提供了一种有效的优化策略。通过使用包含相关定位物理的传输模型，结果表明，通过调整载流子浓度和无序度可以将最大热电品质因数提高约 20%。在两种模型 Ge-Sb-Te 材料系统中证实了轻微无序的好处。特别是对于高度简并的半导体，这种二维优化策略提供了一种获得高热电性能的新方法。