The surface ionisation of potassium by tungsten

Abstract

1. It is well known that atoms of low ionisation potential may be ionised by contact with a hot metal surface. If, for instance, a positively charged hot tungsten filament be surrounded by the vapour of potassium, rubidium or cæsium the filament will lose positive charge at a rate governed by the vapour density of the alkali metal and very nearly independent of the temperature of the filament if this temperature be not too low. From this independence the conclusion is drawn that all (or very nearly all) atoms which strike the filament evaporate in the ionic state; in other words, the efficiency of the surface as an ionising agent is nearly perfect. The objection might perhaps be raised that this conclusion is not strictly justified by the experimental evidence, which merely shows the efficiency to be constant . It would, of course, be remarkable if the constant were other than unity, but further evidence is clearly desirable and has been given by Langmuir and Kingdon in their discussion and extension of Saha’s theory of the equilibrium of ions, electrons and atoms. Their calculations account for the low efficiency and its large temperature variation in cases where the “electron work-function” of the surface is less than the ionisation potential of the atom, and predict perfect efficiency when the work-function greatly exceeds the ionisation potential. A direct verification that the efficiency of surface ionisation is, in a favourable case, indeed unity, is provided by the observation made by one of us that no net current flows to a sufficiently hot nickel surface bombarded by slow Cs + ions. For fast ions, however, the efficiency was found to decrease. It is the purpose of this paper to show how surface ionisation may provide the source of a beam of positive ions which has great intensity, steadiness and length of life, and to describe some experiments in which we have used this source for the more detailed investigation of the drop in ionisation efficiency which occurs (as already noted in the case of Cs + ) when fast K + ions strike a hot target. Moon’s previous observation dealt only with the equilibrium state of the re-emission from the target; in the present work we have studied the building up and breaking down of this equilibrium when the incident beam is switched on or off. It should be stated at once that we have found the process to be complicated and have attempted no more than a qualitative study of its more important features. Tungsten was chosen as the target material on account of the ease with which it may be cleaned by flashing to a high temperature, while K ÷ ions were used in preference to Cs + because the temperatures at which the phenomena under investigation occur are higher and more readily measured by optical pyrometry.