Oxides for thermoelectricity

In 1997, Ichiro Terasaki showed that high thermoelectric power could be achieved in a metal oxide, NaxCoO2 (x ~ 0.7). This oxide thus possesses a power factor S2/ρ close to that of Bi2Te3 [Phys. Rev. B56, R12685 (1997)], a material traditionally used for thermoelectric applications close to room temperature. This result has triggered a flurry of activity in the oxide field to understand the origin of these antagonistic properties, and to search for new thermoelectric oxides. NaxCoO2 has a lamellar structure, made up of sheets of CoO6 octahedra separated by planes partially filled with Na+, and various models have been proposed to analyze the thermoelectric properties, taking into account the material's characteristics: mixed valence of cobalt (Co3+ and Co4+), low spin states of cobalt, particular filling of t2g orbitals, disorder potential bound to Na+ and presence of strong electronic correlations. One of the particularities of oxides is the potential existence of these strong electronic correlations, likely to induce high values of Seebeck coefficients through renormalization effects and modification of the band structure. The aim of this seminar is to present the different families of oxides of interest for thermoelectricity, in particular NaxCoO2 for p-type oxides, and 'transparent conductor' oxides for n-type oxides, and to introduce the different mechanisms proposed to explain these properties. The specific features of oxides, compared with more 'conventional' thermoelectric materials, will be highlighted.