Abstract
The paper summarizes the present situation concerning the measurement and interpretation of the thermopower in ionic crystals. We begin by recalling the physical origin of the thermopower and by showing how it can be phenomenologically expressed in terms of the heats of transport of the lattice defects. The heat of transport gives a measure of the thermodynamic force exerted on a defect by a temperature gradient; according to the Onsager reciprocal relations they also measure the heat flux which accompanies the diffusion of defects in the absence of a temperature gradient. We then review the experimental method of measuring thermopowers, noting that the measured quantity is a sum of two parts; the so-called homogeneous and heterogeneous contributions; it is the former which contains information about the heats of transport. Next, we turn to the question of atomistic interpretation. We show that simple physical ideas can give a rough estimate of the magnitude of the heats of transport, but that they are incapable of giving an adequate explanation of the experimental results. This leads to a brief description of recent theoretical progress, which indicates that it is fruitful to relate the heat of transport to the atomistic heat flux associated with the elementary jump process of a defect. We show that the calculation of this heat flux by computer simulation allows one, in principle, to obtain theoretical results for the heats of transport. We conclude by mentioning some preliminary steps which have been taken towards implementing such a claculation for vacancies and interstitials in fluorite crystals.