Classical methods when applied to photoelectric and thermionic emission from composite surfaces often lead to inconsistent results. A study of the thermionic emission, with a new thyratron circuit for heating tungsten and thoriated tungsten filaments, reveals that many points can be explained if we take into consideration the complex form of the potential barrier produced by an atom layer of electropositive metal when deposited on an "electronegative" base. From an experimental knowledge of the transmission coefficient of the barrier as a function of electron energy and the use of the Wentzel approximation for the transmission coefficient $D\left(v\right)=\exp -\left(\frac{4\pi }{h}\right)\int {x_1}^{x_2}{\mathrm{}\left(2mH\right)\mathrm{}}^{\frac{1}{2}}\mathrm{dx}$), it is possible to calculate the shape of the barrier which turns out to be practically parabolic and about 4.3×${10}^{-8\mathrm{}}$ cm in width one electron-volt below the top of the hill. Since the width and perhaps the height of the barrier seem to depend upon the temperature, it becomes clear that the Richardson-Dushman equation $I=A{T}^2\exp (-\frac{b_0}{T})$ is applicable only as an empirical representation of the data obtained with composite surfaces as emitters. This surface model explains very naturally the fact that the electrons emitted from a composite have an apparently Maxwellian distribution of velocities corresponding to a temperature fifty percent higher than that of the filament as reported by Koller and Rothe. It also gives a qualitative explanation of the observed dependence of the photoelectric long wave-length limit on the applied potential for thin films of sodium on nickel using very low fields.

• Received 6 August 1932

DOI:https://doi.org/10.1103/PhysRev.41.793