Thermal Fatigue and Waste Heat Recovery via Thermoelectrics

Abstract

Using thermoelectrics to directly convert heat energy into useful electricity faces a number of challenges. In addition to the optimized thermal and electrical transport parameters that are needed to boost energy conversion efficiency, thermoelectric (TE) materials must possess sufficient mechanical integrity to survive hundreds or thousands of in-service heating and cooling cycles (thermal fatigue). The nature of TE materials themselves makes it problematic for them to survive under thermal fatigue conditions since TE materials typically (1) are brittle semiconductors or ceramics and (2) have low thermal conductivity which acts to enhance the dimensionless figure of merit ZT but also leads to high stress in response to thermal transients. In addition, many TE materials, including chalcogenide compounds and intermetallic compounds, have relatively high thermal expansion coefficients which also generates high stresses when TE materials are exposed to thermal gradients and transients. The problem of thermal fatigue has been discussed (but not studied directly) in terms of thermal shock parameters, where thermal shock refers to a single thermal cycle and thermal fatigue refers to many thermal cycles. The differences between thermal shock and thermal fatigue are profound in part since the physical mechanisms that lead to good thermal shock resistance are defeated by thermal fatigue. This paper presents a description of (i) the inadequacies of both the thermoelastic model and the energy balance model in addressing thermal fatigue problems as well as (ii) strategies that have been successful in reducing thermal fatigue damage in brittle materials.