Thermoelectric figure-of-merit from Peltier cooling

We use Peltier cooling to experimentally determine the thermoelectric figure-of-merit of a material, where the maximum temperature depression, delta T-max, is simply related to Z of a thermocouple. However, for thermocouples comprising good thermoelectric materials with ZT >= 1, delta T-max can be large (>= 100 K), making it difficult to assign the measurement to a particular temperature. This is less of a problem if the thermocouple consists of a semiconductor and a metal since ZT and delta T-max are then much smaller. We developed a procedure for measuring the dimensionless figure-of-merit of a semiconductor using semiconductor-metal thermocouples in which the form factor (i.e., the ratio of areas to lengths of the two branches) is optimized. We also show valuable aspects in the delta T-max measurement when the material dimensions are varied. Our experiments on couples consisting of a bismuth telluride alloy and constantan, and the corresponding theoretical analysis, show that the radiation losses are significant at ordinary temperatures. A good estimate of thermal conductivity and zT of the sample can be obtained when radiation losses are minimized.

Automated summary

We experimentally determine the thermoelectric figure-of-merit of a material using Peltier cooling, where the maximum temperature depression, delta T-max, is related to Z of a thermocouple. However, for thermocouples made of good thermoelectric materials, delta T-max can be large, making it difficult to assign the measurement to a specific temperature. In this study, we propose a procedure to measure the dimensionless figure-of-merit of a semiconductor using semiconductor-metal thermocouples, optimizing the form factor. We also investigate the impact of varying material dimensions on the delta T-max measurement. Our experiments show the significant radiation losses at ordinary temperatures and suggest minimizing radiation losses for obtaining accurate thermal conductivity and zT estimation.