First Principles Investigation of Thermoelectric Materials

Abstract

Thermoelectric (TE) materials have attracted particular attention in the last decade because they act as a ‘green’ way of converting waste heat energy to electrical energy through the Seebeck effect. The efficiency of a thermoelectric device is measured by the dimensionless figure of merit (ZT), which is directly proportional to the square of Seebeck coefficient, electrical conductivity, and inversely proportional to thermal conductivity. Depending on the operational temperatures, materials like bismuth telluride, lead halides, MgAgSb, skutterudites, and copper and tin chalcogenides are promising candidates for use in thermoelectric devices. To the best of our knowledge, the highest reported value of ZT is about 2.6 at 573 K, which is observed for cadmium-doped AgSbTe2. Hence, there are still efforts to improve/design novel materials with high ZT. Improving ZT implies that one needs to increase the Seebeck coefficient and electrical conductivity and reduce thermal conductivity. However, the mechanisms that improve one of them deteriorate the other, thereby, making it challenging to design novel materials with improved ZT. Over the last several years, many strategies like nano-structuring, band structure engineering, heterostructure formation, dimensionality reduction, etc. have been developed to improve ZT.

In addition to experimental techniques, computational materials design is also an important tool for the discovery of novel thermoelectric materials. In this thesis, using computational tools like density functional theory, semiclassical Boltzmann theory, and many-body electron-phonon coupling, we have studied two aspects of computational research in thermoelectric materials.

In the first part of the thesis we have used computational tools to study the effect of structural modifications on transport properties of bulk and layered materials. Amongst bulk materials, we have studied how doping copper chalcogenides (BaCu2Se2) with suitable dopants introduce resonant states (band engineering) to improve the power factor. Between layered materials, we have shown how the formation of a heterostructure of BiI3 and ZrS2 improves the transport properties compared to the individual monolayers.

In the second part of this thesis, we have critically examined the validation of some approximations that are typically made in the calculations. For predicting transport properties of novel materials, it is important to compute carrier relaxation time. Usually, this is computed using deformation potential theory where only coupling of electrons/holes and acoustic phonons are considered. In our work, we have shown that for ionic solids the coupling of charge carriers with optical phonons is significantly strong. Neglecting them, while computing relaxation times not only results in quantitative errors (by orders of magnitude) but also gives qualitative incorrect trends.