The aim of this work was the design and test of an apparatus for the measurement of the thermoelectric figure of merit ZT of solids. The thermal and the electric conductivity as well as the determination of the Seebeck coefficient of cylindrical bulk-samples can be determined using a newly designed and built set-up. Measurements of the different quantities are performed simultaneously to prevent deviations due to sample degradation during several temperature cycles. The focus is on the measurement of the thermal conductivity. Two different steady-state methods and one transient measurement method are used to determine the thermal conductivity. The first steady-state method is a comparative approach, where the heat flux through the sample is measured using sensors built from a known material. The second steady-state method is an absolute method, where the heat flux is determined from the electric power of the heater generating the heat flux. Since those measurement approaches show long measurement durations a transient method was developed and implemented. Here, the temperatures measured inside the set-up are fitted to a numerical model and then the thermal conductivity and the thermal diffusivity of the sample extracted from the fit. Since a numerical model was implemented the heater power can be an arbitrary function of time. The main benefit of the transient approach is the much higher measurement speed, since the waiting periods where the steady-state is approached are omitted. By increasing the base temperature of the set-up continuously a sample can be characterized quickly over a wide temperature range. The transient measurement method has been investigated further using Monte-Carlo simulations. With those simulations conclusions about the influence of certain measurement parameters like the heater waveform or measurement durations on the informative value of the measurement itself can be drawn. Furthermore, the effect of e.g. the sample geometry or heat radiation has been analyzed. First experimental results of the set-up confirm that the transient mode is working properly. Test samples have been investigated with the different modes of the set-up as well as with other set-ups and the results are found to agree within the error limits of the set-up. Also, the measurements of the Seebeck coefficient are working properly. However, the determination of the electric conductivity of the sample has to be revised, since the contact resistance to the electrical leads is too large, especially in combination with well conducting samples. Further sources of measurement error inside the set-up have been investigated and quantified. Different improvements of the set-up have been suggested for the future.
Verknüpfung zu Publikationen oder weiteren Datensätzen
