Measurement of the Thermoelectric Power of Germanium at Temperatures above 78°K

Abstract

A systematic investigation of the thermoelectric power of germanium samples has been made in the temperature range between 78\ifmmode^\circ\else\textdegree\fi{}K and about 925\ifmmode^\circ\else\textdegree\fi{}K. The general method of measurement is the determination, through thermocouple readings, of the ratio $\frac{\ensuremath{\Delta}E}{\ensuremath{\Delta}T}$, where $\ensuremath{\Delta}E$ is the thermal emf (relative to chromel) measured between two points in the sample having a temperature difference $\ensuremath{\Delta}T$. At first, difficulty was experienced in obtaining reliable and reproducible thermoelectric power data at the lower temperatures (78\ifmmode^\circ\else\textdegree\fi{}K to about 350\ifmmode^\circ\else\textdegree\fi{}K). A detailed discussion is given of the steps taken to eliminate this difficulty by reducing or removing the effects of thermal conduction by the thermocouple wires, convection currents around the the sample, poor contacts between thermocouple junctions and the sample, and inhomogeneities in the thermocouple wires. In the impurity range the sign of the thermoelectric power is the sign of the Hall coefficient, and the magnitudes of thermoelectric power, resistivity, and Hall coefficient increase with decreasing impurity content. With rising temperature the thermoelectric power increases in magnitude and passes through an extremum at a temperature somewhat lower than that corresponding to the end of the exhaustion range or flat portion of the Hall curve. The thermoelectric power of a $p$-type sample passes through zero at a temperature higher than that at which the Hall coefficient of the same sample becomes zero. The thermoelectric power curves of both $p$-type and $n$-type samples approach a common curve, having negative values in the intrinsic range.