Novel concept for thermal conductivity in thermoelectrica

Thermoelectric materials are able to directly convert thermal energy into electrical energy and reversibly electrical energy into thermal energy. Thermoelectric devices are ideally suited to applications were simplicity and reliability, as well as the absence of moving parts and silent operation outweigh their relatively high cost and low efficiency: spot cooling of electronics, air conditioning, water cooler, cooling of superconducting electronics and infrared detectors or home refrigerators. Moreover, applications are essential, wherever electric energy can be obtained from waste heat gradients. The potential of a material for thermoelectric application is determined by the dimensionless figure of merit, given by the square of the Seebeck coefficient divided by the product of electrical resistivity and thermal conductivity. As the thermal conductivity enters the figure of merit in the denominator, the primary goal of the present proposal is a thorough study of the temperature dependent thermal conductivity of solids, as a key parameter controlling the thermoelectric performance. To increase the efficiency, the absolute values of the thermal conductivity have to be diminished in order to prevent a significant portion of the heat flowing down the temperature gradient. This can be achieved by boosting the number and the intensity of scattering processes. Beside mechanisms which in the past have been employed independently, such as i) atom disorder via simple substitution and ii) phonon scattering on lattice sites by inserting atoms in large cages of the structure (rattling modes); we propose a novel concept for efficient reduction of thermal conductivity simultaneously combining the above mentioned mechanisms with scattering of heat carrying phonons on valence fluctuating electrons of atoms with valence instabilities (breathing modes). In this context, the proposed research will focus on families of compounds characterised by strong electron correlations providing extraordinary high density of states at the Fermi level (large thermopower). Promising systems selected under these constraints are i) intermetallics with perovskite type AuCu3, ii) with BaAl4 type and iii) with filled skutterudite type (LaFe4P12). Although some of the materials intended to be studied in detail are already described in literature, the respective thermoelectric properties and thus the thermoelelectric performance is still unexplored in most cases.

Thermoelectric materials are able to directly convert thermal energy into electricity (Seebeck effect, 1821; thermoelectric generator) and reversibly, electrical energy into thermal energy (Peltier effect, 1835; thermoelectric heat pump or Peltier cooling). The potential of these materials covers a large spectrum of applications varying from space engineering to civil markets. Major advantages over conservative competitive systems are in particular: high reliability, silent motionless operation, saving waste energies and being environment-friendly. The basic idea of the research project P16370 was the search for novel mechanisms, able to efficiently reduce the lattice thermal conductivity. Keeping other parameters unchanged, the thermoelectric performance could be enhanced. The most important concept was a combined interaction of so-called "rattling modes" in the context of scattering of the heat carrying phonons and strong valence fluctuations present in compounds having pronounced electron correlation. Main emphasise was laid on materials based on CePd3 and hypothetical CePt 3 , where substitution effects caused a dramatic change of physical properties, attending with substantial changes of transport coefficients and thus of the thermoelectric performance. Moreover, Mischmetall (mostly Ce, minor contributions of La, Pr, Nd, and Sm) was introduced into the cages of skutterudites as a novel possibility of electropositive filling elements. The most important findings made in course of the project P16370 are: (a) Discovery of heavy fermion superconductivity in CePt 3 Si in absence of inversion symmetry. The latter feature is responsible for significant constraints of the superconducting order parameter and should not allow spin-triplet pairing to occur. In fact, the large value of the upper critical field found for CePt 3 Si, requires involving spin triplet pairing in the superconducting condensate. As a result, many theoretical scenarios were developed and a very promising one takes into account a distinct mixture of spin-singlet and spin-triplet components forming the order parameter driven by a strong spin-orbit coupling of the electrons. This extraordinary feature associated with CePt 3 Si causes the appearance of rather strange temperature dependences of various physical quantities such as the behaviour of the London penetration depth or the 1/T 1 relaxation time in NMR measurements, which, in part, have never seen before in any other superconductors. (b) The thermoelectric behaviour of systems based on the intermetallic compound CePd3 has been tuned by substitution on the Pd site and by doping. Strong electron correlations are a source of substantial reductions of the lattice thermal conductivity, proven in this family of compounds as a main goal of this research project. The series of skutterudites Pr(Fe,Ni)4 Sb12 as well as Pr(Fe,Co)4 Sb12 were studied in detail. Promising thermoelectric features were observed because of two reasons; a) the exchange of Fe/Ni causes a decrease of the number of holes and as a consequence, the systems shifts towards a semiconducting state. The reduction of the carrier number is expected to significantly enhance the Seebeck coefficient. In fact, a compound Pr y Fe2.5 Ni1.5 Sb12 exhibits a Seebeck coefficient of almost 250 V/K, one of the highest values ever found in skutterudite systems. Adding cheap Mischmetall instead of expensive Pr does not significantly alter this behaviour and brings technical applications with respect to conversion of waste heat into electricity within reach.