High power density metal junction thermoelectric (TE) power generators

High efficiency direct energy conversion is still a daunting challenge. The efficiency of state-of-the-art, semiconductor based thermoelectrics (TE) is in the range of only about 7~10% efficiency.  The poor performance of conventional TE devices is due to the fundamentally limited potential well that is intrinsically determined by the population of doping materials. Developmental efforts in TE materials have so far only focused on lowering thermal conductivity in order to increase the figure of merit. Implicitly, this practice also lowers the overall energy flow through the TE system, resulting in low power density.  A new approach based on metallic junctions will be implemented in this project to enhance thermoelectric performance beyond the state-of-the-art limit of the conventional concept and approach. 

The performance of conventional TE devices is determined by the figure of merit (ZT): ZT = S²σT/k (S: Seebeck coefficient, σ: electrical conductivity, k: thermal conductivity, and T: temperature). It must be noted that this figure of merit relates to the fundamental performance of the thermoelectric domain, and not necessarily to actual power delivered by a thermoelectric generator. The challenge in developing conventional thermoelectric materials with enhanced performance is being able to engineer the thermal and electrical parameters separately. Research to date has focused on morphological design and nanometer sized structures to lower thermal conductivity (k) in order to improve ZT. The reduction of the lattice contribution to the thermal conductivity directly improves ZT, and can be achieved through nanostructural design. However, lowering the thermal conductivity by introducing nanostructured dislocations and boundary conditions for phonon scattering, which disrupts the thermal transport, has not improved the figure of merit as anticipated. More importantly, these physical dislocations have not enhanced the power delivery of the thermoelectric devices.

This is largely due to a fact that lowering the thermal conductivity also reduces heat flow into a TE domain. Since the initial loading of thermal energy into a domain is substantially reduced by lowered thermal conductivity, there is a reduction in thermal energy in the domain to be converted into power. This is why the conventional TE principle has NOT worked favorably, despite the expenditure of significant resources and enormous effort. The charge carrier density in semiconductor-based TE materials is limited to the electron densities of 1 C/cm³.

The aim of this project is to develop a new TE concept based on metal junctions, which will substantially increase the TE efficiency and power density resulting in substantially increased power delivery by TE generators. Metals have intrinsically high electron densities >10³-10⁴ C/cm³, which is several orders of magnitude higher than the best semiconductors. Metal junctions are also easier to make than semiconductor junctions, which need an additional doping process. These factors combine to enable the fabrication of high efficiency, high power density energy conversion systems at reasonable cost. The actual power and power density delivered by the thermoelectric generator, rather than ZT, is going to be the target metric in this study. The power density predicted for metal junctions TE will be greater than thermoelectrics with semiconductor junctions for the reasons given above.