The Thermal Hall Effect in Copper

Abstract

In comparison to the interest given to electrical effects, very little interest has been shown in the thermal magnetoresistance and Righi-Leduc (thermal Hall) effect in metals. There is no a priori reason to suppose that the Wiedemann-Franz law relating the electronic contributions to the thermal and electrical conductivity tensors should not hold in a magnetic field, and the lattice contribution to the thermal conductivity should be negligible at 2°K; in fact, Azbel et al.1 have stated explicitly that the Lorentz ratio for the transverse effects should be independent of Fermi surface and scattering mechanisms in the limit of high magnetic fields. There are a number of practical reasons, however, which weigh strongly in favor of the measurement of thermal effects. The first of these concerns noise levels; a typical sample carrying an electric current of a few amps will generate a transverse voltage of perhaps 10-7 V, which is not easy to measure accurately, especially when the magnetic field is not absolutely steady. A sample for thermal measurements can be considerably larger—about 5 mm square has been found satisfactory—and still generate temperature differences of 10-2 deg which can be measured reliably and accurately using carbon resistance thermometers. The effects of field fluctuations can be eliminated by using a thermal link between the sample and the thermometers, which can be placed outside the field.