The transition to conduction in diffusion-controlled gas breakdown

Abstract

Measurements have been made of ultra-high-frequency (l80Mc/s) breakdown fields in hydrogen, nitrogen and neon, in a Rogowski-profiled spark gap, under conditions of pressure in which both the electron ambit and the mean-free-path were less than the gap width (0.494cm). The measurements were made with sustained fields and also with pulsed fields in overvolted conditions to obtain data on the variation of breakdown delays with overvoltage. The delays were measured by photographing an oscilloscope trace which displayed the gap voltage, and a range of 400 to 0.2 microseconds was observed with over voltages up to 100%. Breakdown was initiated during a pulse by irradiating the gas in the mid-gap region with a short (about 0.1 microsecond) burst of ultraviolet photons, A theory is proposed which determines the shape of the oscillograms, and hence gives breakdown delays. Good agreement was obtained between theoretical and experimental delays for hydrogen and nitrogen, but predicted delays were too short by factors up to ten for neon. A qualitative explanation is given for the anomalous results for neon. Maintaining voltages have been measured in hydrogen and nitrogen, as a function of the gas pressure, but were too low to measure in neon. In all three gases it is shown that the maintaining fields are consistent with losses by ambi-polar diffusion, although electrons which approach within one ambit of an electrode are driven into that electrode by the action of the field. Immediately after breakdown in overvolted conditions the voltage approached zero and then recovered to approach the maintaining voltage asymptotically. The recovery time- constants were of the order which might be expected if an initial excess electron population were being diminished by ambi-polar diffusion.