Temperature Dependence of the Work Function of Tungsten from Measurement of Contact Potentials by the Kelvin Method

The Kelvin method for the determination of contact potentials is adapted for measurements between filaments in vacuum. Measurements between a heated and a cool tungsten filament indicate an increase in the work function of tungsten, with temperature, of 6.3×${10}^{-5\mathrm{}}$ volt per degree centigrade. This effect is distinguished from changes in work function arising from variations, with temperature, in contaminating layers on the tungsten, which tend to obscure the true temperature coefficient even under the best vacuum conditions. The observed true temperature coefficient appears to be comparatively independent of the existing degree of surface contamination. This suggests that the temperature effect must reside primarily in the thermodynamic potential of the electrons inside the metal and not in the potential barrier at the surface. The coefficient obtained resolves the discrepancy between the experimental value of $A$ in the Richardson equation and the theoretical factor of 120 without the introduction of a reflection coefficient.