Discovered more than 200 years ago in 1821, thermoelectricity is nowadays of global interest as it enables direct interconversion of thermal and electrical energy via the Seebeck/Peltier effect. In their seminal work, Mahan and Sofo mathematically derived the conditions for ’the best thermoelectric’—a delta-distribution-shaped electronic transport function, where charge carriers contribute to transport only in an infinitely narrow energy interval. So far, however, only approximations to this concept were expected to exist in nature. Here, we propose the Anderson transition in a narrow impurity band as a physical realisation of this seemingly unrealisable scenario. An innovative approach of continuous disorder tuning allows us to drive the Anderson transition within a single sample: variable amounts of antisite defects are introduced in a controlled fashion by thermal quenching from high temperatures. Consequently, we obtain a significant enhancement and dramatic change of the thermoelectric properties from p-type to n-type in stoichiometric Fe₂VAl, which we assign to a narrow region of delocalised electrons in the energy spectrum near the Fermi energy. Based on our electronic transport and magnetisation experiments, supported by Monte-Carlo and density functional theory calculations, we present a novel strategy to enhance the performance of thermoelectric materials.

200 多年前的 1821 年发现的热电现在引起了全球的关注，因为它可以通过塞贝克/珀尔帖效应实现热能和电能的直接相互转换。在他们的开创性工作中，Mahan 和 Sofo 在数学上推导出了“最佳热电”的条件——一种 delta 分布形状的电子传输函数，其中电荷载流子仅在无限窄的能量区间内对传输做出贡献。然而，到目前为止，预计自然界中只存在这个概念的近似值。在这里，我们提出了窄杂质带中的安德森跃迁作为这种看似无法实现的场景的物理实现。连续无序调整的创新方法使我们能够在单个样本中驱动安德森转变：通过从高温进行的热淬火以受控方式引入了可变数量的反位缺陷。因此，我们从化学计量 Fe _{2 VAL 中的}p型到n型，我们将其分配给靠近费米能量的能谱中的离域电子的狭窄区域。基于我们的电子传输和磁化实验，在蒙特卡罗和密度泛函理论计算的支持下，我们提出了一种提高热电材料性能的新策略。