Methodology for Testing Thermoelectric Materials and Devices

A method was developed in the early 1960s by Harman and Honigt for testing the AC resistance and figure-of-merit, Z, of a thermoelectric material sample. This method provided a direct measurement of Z, but lacked precision and reproductibility, primarily due to the mechanical and electronic limitations of the instrumentation. A new method has been developed which is based on a similar concept but has some fundamental differences which give rise to improvements in accuracy and reproducibility. The fundamental similarity is that both techniques are designed to resolve the voltage components of a thermoelectric device. The fundamental difference is that the Harman1 method does this by measuring the resistive component, whereas the new method measures the Seebeck component. This new test method presented in this chapter is referred to as the "TRANSIENT" test method. The TRANSIENT method applies to completed, modular devices as well as individual thermoelectric material samples. It provides for the direct measurement of all the parameters needed to characterize the thermoelectric properties of materials or devices under test. The key to the TRANSIENT method is a computer-driven, high-speed, high-resolution, integrating voltage measurement system which is capable of accurately resolving the voltage components in an active thermoelectric device or sample. The subsequent computer analysis yields values for the Seebeck coefficient, a, electrical resistivity, p, thermal conductivity, X, and figure-of-merit, Z. These parameters are measured simultaneously on the same sample via an absolute method which requires no reference or standard material for comparison. Data are presented illustrating the application of this method over a wide temperature range with an extremely high level of reproducibility.