Thermoelectric Oxides

Most currently employed state-of-the-art thermoelectric (TE) materials are based on elements that are toxic, non-abundant and expensive. These include in particular materials containing tellurium (Te), lead (Pb) and to some extent bismuth (Bi), all of which are found in the best TE’s to date, Bi2Te3, PbTe and variations thereof. A lot of research efforts are therefore devoted to the identification of new thermoelectric candidate materials free of these elements in addition to the search for materials with larger figures of merit for operation at high temperatures in the range of 600-1000°C. These efforts have brought the sulfides, selenides and oxides back into the focus of research.

Oxide materials have long been discarded as potential thermoelectrics due to their prevalent ionic bonding character which implies low free charge carrier densities and consequently small values of the electrical conductivity. However, the oxide material class provides a large playground with a vast range of crystal structures and the option of engineering new structures with specifically tailored properties. In addition, their generally high melting points offer the potential for stable operation in air at high temperatures up to ~1000°C [1]. Several oxide families have been studied and identified as promising TE candidates to date, including the perovskite titanates and the sodium and calcium cobaltates. In these two families, different aspects have been identified as being relevant for their good thermoelectric performance. The perovskite titanates crystallizing in a simple cubic crystal structure are characterized by good electrical properties with high power factor upon elemental substitution but are generally limited by a high intrinsic thermal conductivity. The thermal transport and the role of lattice anharmonicity in these materials are still not fully understood. In the cobaltates, their layered crystal structure entails anisotropic thermal transport and low thermal conductivities beneficial for high figures of merit. Their anisotropic electronic properties are related to the interplay of spin and charge degrees of freedom of the Co ions within their specific environment and are of great current interest from both an experimental and theoretical point of view.

In our Working Group EM-AMCT we follow a model system approach, in which we select specific materials for the study of various fundamental physical principles and in particular aspects associated with a low lattice thermal conductivity. To this end, we focus on bulk polycrystalline and single crystalline samples. Most materials we are investigating are synthesized in our group by conventional solid state synthesis routes and by employing the optical floating zone technique for the growth of large high-quality single crystals. Single crystals are required for neutron scattering studies but are also required for transport property measurements, especially if anisotropic properties are to be investigated, or if one wishes to relate the macroscopic properties to the actual crystal structure on a microscopic level.

Our samples are routinely characterized using powder and single-crystal XRD available in HZB’s X-ray core lab as well as by using neutron diffraction available at HZB [E9, E2]. The more sophisticated measurements involving inelastic neutron or X-ray scattering techniques for the study of phonon dynamics and phonon lifetimes are then performed either at instruments at HZB [FLEXX] or at other facilities worldwide (FRMII, ORNL, ILL, ESRF; APS).

In addition, we can study a wide range of physical properties such as specific heat, magnetization, electrical conductivity, thermal conductivity, Seebeck effect, charge carrier concentration, by using equipment operated by our Working Group EM-AMCT, TE synthesis and TE characterization, or that available in other HZB labs.

One of our current research topics is related to the study of phonon lifetimes in the ABO3 perovskite titanate compound SrTiO3. This material has been identified as promising thermoelectric due to high power factors that can be achieved by partial substitution by La and Nb on the perovskite A and B sites, respectively. We are, however, interested in the low temperature properties, as SrTiO3 in its undoped form exhibits a structural phase transition at 105K accompanied by an incipient ferroelectric phase transition. The study of the related lattice dynamics in this material has been the topic of numerous theoretical studies, most of which predict anharmonic effects to have a large impact on the lattice dynamics and phonon mean free paths. However, to date, there have not been any experimental studies of the phonon lifetimes which provide a direct measure of the phonon scattering rates and phonon mode-dependent anharmonicity. Our Working Group is using the NRSE technique to extract temperature-dependent changes in the phonon linewidths.

StruKtur W-Bronze

Fig.1: Crystal structure of the archetype partially filled tetragonal tungsten bronze structure as discussed in the text.

In another project, we are focusing on ferroelectric and relaxor ferroelectric oxides with complex crystal structures based on the tetragonal tungsten bronze type structure. This structure is characterized by a 3D network of distorted transition metal-oxide octahedra running in chains along one axis of the structure (also the axis for spontaneous ferroelectric polarization). Triangular, square and pentagonal channels are left between the octahedra and can be partially occupied by alkaline earth metal ions (see Fig. 1).

This structure, due to its complexity and inherent degree of chemical disorder exhibits an intrinsically low lattice thermal conductivity. Moreover, it was recently shown that the electrical properties of these materials can be tuned by the changing the oxygen content, which leads to a significant increase in the power factor [2]. We are interested in the interplay between the crystal structure, the lattice dynamics and the thermal conductivity upon changing the oxygen stoichiometry in these materials, and in particular in the subtle effects on the ferroelectric/structural phase transitions. For this project, we have grown single crystals using the optical floating zone technique (see Fig. 2) and have characterized and aligned them using the new FALCON neutron Laue diffractometer (cf. Fig. 3). In single crystal X-ray and neutron diffraction experiments, we observe an evolution of structural superlattice peaks and the appearance of additional diffuse peaks at incommensurate positions in these materials upon varying the oxygen stoichiometry (see Fig. 4).

A relatively new research topic is based on the recent (re-)discovery of a class of minerals called tetrahedrites. These materials can attain ZT values of ~1 to ~1.7, comparable to the current state-of-the-art TEs, and their crystal structures involve the non-toxic, abundant and cheap elements Cu, Sb, S and Se. They exhibit an extremely low lattice thermal conductivity close to the amorphous limit which is attributed to an intrinsic bonding asymmetry related to the presence of lone-pair electrons. This bonding asymmetry is considered to be responsible for largely anharmonic phonon scattering [3]. These effects and the underlying microscopic mechanisms are, however, not well understood and currently under debate. One of our current projects is devoted to oxide materials with crystal structures that contain lone-pair electrons and to the study of the lone-pair electrons’ influence on the phonon dynamics.

TE-Oxid Float-Zone

Fig. 2: Optical floating zone growth of a semi-transparent calcium-barium-niobate (CBN) single crystal under oxygen atmosphere. Upon reduction under Argon, the transparent crystal turns black and the electrical and thermal properties change.

TE-Oxid Laue-Diagramm

Fig. 3a: Neutron Laue diffraction on CBN acquired on E11 (FALCON), illustrating the high crystal quality of our sample (raw data).


Fig. 3b: Orientation of the CBN crystal.

TE oxides diffr. CBN (4a)

Fig. 4a: Single crystal X-ray diffraction scan in CBN at room temperature along the [H,H,0] direction in reciprocal space around (2.5,2.5,0.5). Sharp resolution-limited peaks at (2,2,0.5), (2.5,2.5,0.5) and (3,3,0.5) are contrasted to broad diffuse peaks at incommensurate positions (2.24,2.24,0.5) and (2.76,2.76,0.5). This diffuse scattering is associated with the distortion and short-range ordering of oxygen octahedra in the structure.

CBN Einkristall Neutron

Fig. 4b: Single crystal neutron diffraction in the (H,H,L) scattering plane of CBN from E2 (HZB). Diffuse streaks of scattering are visible along [H,H,0], especially at L=1.5 r.l.u. The visible rings are due to scattering from the aluminum sample holder.


[1] I. Terasaki, Research Update: Oxide thermoelectrics: Beyond the conventional design rules, APL Mater. 4 (2016) 104501.

[2] S. Lee, R.H.T. Wilke, S. Trolier-McKinstry, S. Zhang, C.A. Randall, SrxBa1−xNb2O6-δ Ferroelectric-thermoelectrics: Crystal anisotropy, conduction mechanism, and power factor, Appl. Phys. Lett. 96 (2010) 031910.

[3] W. Lei, Y. Wang, D.T. Morelli, X. Lu, From Bonding Asymmetry to Anharmonic Rattling in Cu12Sb4S13 Tetrahedrites: When Lone-Pair Electrons Are Not So Lonely, Adv. Funct. Mater. 25 (2015) 3648-3657.