Thermoelectric Performance of IV–VI Compounds with Octahedral‐Like Coordination: A Chemical‐Bonding Perspective

Abstract

Thermoelectric materials provide a challenge for materials design, since they require optimization of apparently conflicting properties. The resulting complexity has favored trial-and-error approaches over the development of simple and predictive design rules. In this work, the thermoelectric performance of IV-VI chalcogenides on the tie line between GeSe and GeTe is investigated. From a combination of optical reflectivity and electrical transport measurements, it is experimentally proved that the outstanding performance of IV-VI compounds with octahedral-like coordination is due to the anisotropy of the effective mass tensor of the relevant charge carriers. Such an anisotropy enables the simultaneous realization of high Seebeck coefficients, due to a large density-of-states effective mass, and high electrical conductivity, caused by a small conductivity effective mass. This behavior is associated to a unique bonding mechanism by means of a tight-binding model, which relates band structure and bond energies; tuning the latter enables tailoring of the effective mass tensor. The model thus provides atomistic design rules for thermoelectric chalcogenides.

Chalcogenide Thermoelectrics Empowered by an Unconventional Bonding Mechanism

“…A larger band anisotropy is beneficial for the thermoelectric performance, since it allows to maximize thermoelectric performance. A systematic study of the band anisotropy and power factor for (GeTe) 1− x –(GeSe) x solid solution may help answer this question . Figure a shows that with increasing the content of GeSe, the ternary alloy undergoes a phase transition from rhombohedral to hexagonal at x > 0.65, while the pure GeSe phase is orthorhombic.…”

“…Then, the trend of the effective mass tensor with respect to changes in bond parameters can be inferred from the corresponding changes in the bandwidths along the principal axes. This description has been successfully implemented by means of s–p 3 next‐nearest‐neighbor tight‐binding model . It has been shown that increasing s–p mixing between cation s and anion p states increases the bandwidth along Γ–L with respect to the one along L–W, hence reducing the ratio between longitudinal and transverse effective masses .…”

“…Then, materials whose trade‐off between stoichiometry and distortion effects leads to small effective s–p matrix element should have the most anisotropic carrier pockets. Those include materials with metavalent bonding, for example, GeTe, PbTe, Bi 2 Te 3 , and AgBiSe 2 , etc.…”

“…A systematic study of the band anisotropy and power factor for (GeTe) 1−x -(GeSe) x solid solution may help answer this question. [110] Figure 9a shows that with increasing the content of GeSe, the ternary alloy undergoes a phase transition from rhombohedral to hexagonal at x > 0.65, while the pure GeSe phase is orthorhombic. The phase transition from rhombohedral to hexagonal is accompanied also by a change of chemical bonding from metavalent to covalent as confirmed by previous studies.…”

“…This description has been successfully implemented by means of s-p 3 next-nearest-neighbor tight-binding model. [110] It has been shown that increasing s-p mixing between cation s and anion p states increases the bandwidth along Γ-L with respect to the one along L-W, hence reducing the ratio between longitudinal and transverse effective masses. [69] A similar effect is caused by larger ionicity, which leads to a shorter lattice constant and increased bond strength between adjacent p-orbitals in a σ-bonding configuration.…”