Statistical modeling of high power microwave surface flashover delay times

Summary form only given. The development of high power microwave (HPM) systems and technologies has been inhibited by breakdown phenomena which limit their transmitting capabilities. For a system in which there is no obvious source of breakdown initiating electrons, HPM breakdown can be even less predictable. A system in which microwaves are generated in a vacuum environment for the purpose of radiating into atmosphere typically uses a dielectric window to separate the vacuum and atmospheric sides of the system. At sufficient field levels, surface flashover can occur across this dielectric window resulting in a severe drop in transmitted power. The time between the application of the HPM and the sharp drop in transmitted power is described as the delay time, which consists of a statistical waiting time for initiatory electrons combined with an electron amplification time, or formative time. The experimental setup for this project consists of a 4 MW HPM source operating a 2.85 GHz attached to a traveling wave structure and a dielectric window mounted on the output side of the system. Dielectric surface flashover has been observed in air and nitrogen with pressures ranging from 60 to 155 torr. To provide a constant source of seed electrons, a UV lamp is used to illuminate the window resulting in photo-emitted electrons appearing at the surface. Another way to provide seed electrons is the inclusion of metallic points on the window which provide a source of field emitted electrons, and also results in a field enhancement at the dielectric surface. In the absence of a constant source of seed electrons, it is expected that field detachment from ion clusters is the primary mechanism for providing the high field region with flashover initiating electrons. A statistical model has been developed for predicting surface flashover that takes into account relevant parameters such as field level, ionization rate, gas type, and pressure. This model has shown good agreement with e- perimental data in nitrogen with UV illumination providing a constant electron seed rate. Presented here is an adaptation of this statistical model to an environment consisting of field enhancing metallic points as well as a comparison of results for UV illumination and stochastic seeding through field detachment from ion clusters.