Thermoelectric cage compounds: a combined experimental and ab-initio effort

Research on the materials family of filled skutterudites turned out to be a fruitful playground regarding both a rich variety of exciting ground state properties as well as a promising potential for thermoelectric applications. Skutterudites crystallise in a cubic structure where e.g., rare earths, actinoids or alkaline earth atoms are filled as a guest atoms into the cage-like voids of the basic structure. Many of the already observed properties derive from the unique crystal structure of skutterudites, which property also allows tailoring physical properties by a specific choice of the crystallographic cages and their filler atoms. Recently, the thermoelectric capability of filled skutterudites has experienced a significant progress. Work carried out by part of the proposers has hitherto achieved remarkably figures of merit ZT of up to 1.7 for severe plastically deformed skutterudites. In general, ZT can be enhanced by increasing the Seebeck coefficient S and by an appropriate reduction of the thermal conductivity and the electrical resistivity. Since the electronic contribution to the thermal conductivity is inverse proportional to the electrical resistivity, the remaining parameter - apart from S - to increase ZT is a reduction of the lattice thermal conductivity. Thus, the knowledge of the lattice thermal conductivity is of fundamental importance for appropriately reducing this parameter. Therefore, the primary goal of the present research project is to gain a profound understanding of the relevant mechanisms contributing to the heat transport via lattice vibrations. This will be achieved by a combined effort of experimental studies and theoretical model calculations based on first-principles density functional theory (DFT) calculations in the framework of the Boltzmann transport equations and the Green-Kubo formalism. The experimental and theoretical studies focus on skutterudites where the cage forming structure is derived entirely by germanium as a technologically stable element. Members of this novel class of materials are essentially composed as MPt4Ge12 with M = Sr, Ba, La, Ce, Pr and Nd. Following predictions of DFT calculations, it is then experimentally possible by Ge/Sb substitutions to drive these compounds towards a metal-to-insulator transition, thus providing favourable conditions to achieve ZT values of relevant order. The main challenges of the project proposed are: i) tuning of the charge carrier density by chemical substitution and/or by varying the filler elements to drive the system from a superconductor towards a semiconducting thermoelectric material; a reduction of the charge carrier density is expected to concomitantly cause modifications of the thermal conductivity, too; ii) reducing the phonon contribution ph to the thermal conductivity by tuning the vibrational properties via the filler atoms;. iii) tuning of the charge carrier density by chemical substitution [see (i)] will simultaneously tune the electrical conductivity and the Seebeck coefficient; iv) searching for thermodynamically stable phases as a function of composition with promising thermoelectric properties under the guidance of DFT results.

The aim of the present research project was an experimental and theoretical study focusing on so-called thermoelectric materials. The latter have the ability to convert heat directly into electricity (and vice versa). This characteristic, if successful, provides a possibility to recover otherwise lost waste heat by converting it into valuable electricity. In many respects, it might help saving energy and reducing the impact on energy related processes to our environment. The work carried out in course of action had two integral parts: i) experimental studies on different families of promising thermoelectric materials and ii) first principal calculations of the thermal conductivity as one of those physical quantities, governing the performance of a distinct thermoelectric material. The group of materials selected were filled skutterudites and (half)Heusler materials. Certain members of these families exhibit very large values of the so-called figure of merit, ZT, which is a dimensionless numerical value, allowing a comparison and ranking of such materials. Note that for technical applications this quantity should reach, at least, values of the order of one. A breakthrough in the material class of half-Heusler systems was the removal of elemental hafnium - as a very expensive metal - by a combination of much cheaper elements without losing the superior thermoelectric performance. With respect to the family of filled skutterudites, we have shown that through a modification of the electronic structure by appropriately substituting some of the elements in the starting material a crossover of physical properties can be reached, from a superconducting state (without substitution) towards an insulating state with high thermopower values, which indicates that the system loses its metallic character and approaches the metal-to-insulator transition, where good thermoelectric materials are found in general. It was, however, not possible to cross this line as it is proposed from so-called DFT calculations, although sophisticated sample preparation methods were applied. The theoretical part of the project consisted in the development and application of parameter-free methods for the calculation of the lattice thermal conductivity. Starting point was Boltzmann's transport theory, in which the so-called relaxation time is the most difficult part to be handled. Based on volume dependent density functional calculations for volume dependent lattice properties, a variety of materials was investigated. Doing so, a number of technical and physical problems had to be solved for the development of an applicable program package. First results reveal good agreement with experimental measurements both on bench-mark systems and actual thermoelectric materials, which confirms the predictive power of the present approach.