The thermionic constants of metals and semi-conductors I. Graphite

Abstract

The thermionic constants of graphite are determined by finding by Knudsen's effusion method the saturation vapour pressure of the electron gas in a graphite chamber, at different known temperatures, and applying the well-known Clapeyron-Clausius thermodynamic relation. The work function $\phi $ of graphite determined in this manner is 4$\cdot $62 $\pm $ 0$\cdot $02 eV, and the effusion constant A, analogous to the emission constant in Richardson's equation, is 60 $\pm $ 2 A cm$^{-2}$ deg.$^{-2}$. The introduction of metals like platinum or tungsten into the graphite chamber does not affect the rate of effusion detectably, i.e. does not affect either $\phi $ or A determined therefrom, and these constants are also quite insensitive to any contamination of the graphite surface by normal adsorption, even in conditions in which the emission of electrons from the surface is greatly suppressed. These observations emphasize the advantages of this method of determining the thermionic constants of graphite over the usual methods. They further suggest that (1) the contact potential between graphite and the metal introduced into the chamber almost exactly compensates for the difference between their work functions; (2) the effusion coefficients A of these metals are practically the same as for graphite, namely, 60 A cm$^{-2}$ deg.$^{-2}$; (3) surface contamination of the walls of the chamber behaves in the same way as the metals introduced in it; (4) any observed differences in the emission coefficients of these metals should be attributed to differences in the reflexion coefficients of their surfaces for electrons; and (5) contamination of the surface affects its reflexion coefficient very markedly. Measurements are also reported incidentally for the spectral emissivity of graphite for electromagnetic radiations in the neighbourhood of 6550 angstrom at different temperatures.