Thermoelectric clathrates for waste heat recovery

Intermetallic clathrates are fascinating solids that offer multiple opportunities for both fundamental science and applications. Their crystallographic structure - invented by nature and first found for gas hydrates in arctic ice - is stabilized by two different types of chemical bonding: An essentially covalently bonded cavity-forming framework (host) encapsulates atoms via much weaker, predominantly ionic bonding (guests). This leads to a partial decoupling of some of the degrees of freedom of the resulting solid. For instance, distinctly different branches in the phonon spectrum (Debye- and Einstein-like modes) may be associated with the host and guest, respectively. Such decoupling, and a controlled interaction via materials design, is of great interest for thermoelectric applications, where a combination of materials properties is required that can hardly be met by simple solids. In this project affordable and highly tunable clathrates with potential for large scale use in waste heat recovery are at focus. With various new approaches we aim at reaching a level of understanding of the composition - structure - property relationship that will enable us to tailor these materials to distinctly higher thermoelectric figures-of-merit than achieved before. We expect this project to move clathrates a big step forward on the value chain, and to pave the way for their application in thermoelectric generators.

Type-I clathrates are a fascinating class of materials that offer multiple opportunities for both fundamental and application-oriented research. In particular, the extremely low thermal conductivities but comparably high electrical conductivities have made clathrates candidate materials for applications in thermoelectric waste-heat recovery. Germanium-based clathrates show the highest thermoelectric figures of merit. However, their price limits the applicability to niche markets. The present project focused on Silicon-based clathrates as much cheaper variants, with potential for mass-market applications.Two design strategies were explored to improve their thermoelectric performance: phonon engineering (the suppression of the thermal conductivity by scattering from mass disorder, vacancies, grain boundaries, dislocations, electrons, and phonons) and electron engineering (the enhancement of the thermopower by resonant states and the Kondo effect). The goal in all our investigations was to unravel the dominant microscopic mechanisms. For this purpose, the materials studied were carefully selected well-controlled sample series, frequently grown as single crystals. In addition to a standard thermoelectric characterization, many other advanced measurements were used whenever needed (magnetotransport, thermodynamic measurements, optical conductivity, NMR, neutron or X-ray diffraction and inelastic scattering). To improve the experimental accuracy of thermal conductivity measurements at high temperatures and to be able to measure also thin films, a high-end 3-omega technique was developed.Key results are that structural defects and electrons are efficient phonon scatterers at low temperatures, but that phonon Umklapp scattering becomes increasingly important at higher temperatures. In single crystalline or large-grain clathrates, a universal scaling of the thermal conductivity with only two parameters characteristic for the phonon spectrum could be demonstrated. Further lowering of the thermal conductivity can be achieved by nanostructuring or the introduction of dense dislocation arrays. Both ball milling and melt spinning were shown to produce nanogranular clathrates under appropriate conditions. Process control agents and composite formation can protect the nanostructure from coarsening during compaction. A drastic enhancement of the thermoelectric figure of merit was achieved by a partial substitution of guest atoms with Cerium, which was successful for the first time worldwide by a newly-developed synthesis technique.In summary, we have developed new design strategies that have the potential to make even Silicon-based clathrates useful thermoelectrics.