Statistics of high power microwave induced window flashover

Summary form only given as follows. For flashover in air, nitrogen, and SF₆, we have previously shown that the delay time between microwave pulse application and breakdown increases with pressure in the regime where the elastic electron collision frequency is larger than the microwave frequency (roughly >10 torr for 2.85 GHz microwave frequency), which also coincides with conditions found at the right hand side of the Paschen curve. The specific window flashover geometry was carefully chosen to avoid local field enhancement. That is, no metallic parts are exposed to high fields. Hence, only the window surface itself, the gas, and interaction processes between surface and volume contribute to flashover. A Monte Carlo based electron motion code developed for the flashover conditions predicts formative flashover delay times reasonably well in the pressure regime between 100 to 600 torr (10,000 Pa to 80,000 Pa). However, the statistical delay time, that is the time interval required for the initiatory electron(s) to appear, is unaccounted for in the code. Further computational efforts investigating seed electron production via collisional detachment from, for instance, negative ions in the gas have shown that while effective at unipolar fields, collisional detachment is unlikely to contribute to the production of seed electrons at higher microwave frequencies above several GHz. Experiments show that illuminating the surface with light/photons (180 nm < lambda < 350 nm) reduces the observed statistical delay considerably indicating the importance of seed electron production from the surface. This paper will discuss the key processes of high power microwave surface flashover and present experimental flashover data along with continued investigation into the statistics of possible seed electron sources, including trace contaminates present in the gas or on the dielectric surface.