WINDOW AND CAVITY BREAKDOWN CAUSED BY HIGH POWER MICROWAVES A. Neuber, J. Dickens, D. Hernmert, H. Krompholz, L.L. Hatfield, M. Kristiansen Departments of Electrical Engineering and Physics Texas Tech University, Lubbock, TX 79409-3102 Abstract Physical mechanisms leading to microwave breakdown on windows and in cavities are investigated for power levels on the order of 100 MW at 2.85 GHz. The test stand uses a 3 MW magnetron coupled to an S-band traveling wave resonator. Various configurations of dielectric windows are investigated. In a stan- dard pillbox geometry with a pressure of less than 10-8tom, surface discharges on an alumina window and multipactor-like discharges starting at the waveguide edges occur simultaneously. To clarify physical mechanisms, window breakdown with purely tangential electrical microwave fields is investigated for spe- cial geometries. Other configurations, such as air filled two window setups, relevant for vacuum-air inter- faces, can be investigated as well. Diagnostics include the measurement of incident.heflected power, meas- urement of local microwave fields, discharge luminosity, and x-ray emission. All quantities are recorded with 0.2 to 1 ns resolution. In addition, a framing camera with gating times of 5 ns is used. Based on the experimental results, methods to increase the power density which can be transmitted through windows, such as surface coatings and window profiles, will be investigated as well. Introduction Generation and transport of high power microwaves are usually limited by electric breakdown mechanisms, where the discharge plasma inhibits microwave propagation. Discharge thresholds are on the order of 100 kV/cm for multipactorl cavity breakdown, and lower for surface breakdown at interfaces. Up to now, mainly empirical methods have been used to increase breakdown thresholds. The aim of the pre- sent work is to clarify the physical mechanisms leading to breakdown for a variety of conditions. These mechanisms can be categorized into two parts: the generation of initial eleetrons, and amplification or avalanche mechanisms leading to a build-up of a high electron density plasma. Knowledge of these physical processes is a prerequisite to develop ways to increase breakdown thresholds. Satisfactory models have been developed to describe the multipactor process in cavities or waveguidesz , where electrons produce secondaries when hitting the metallic walls, and the lowest breakdown thresholds occur for resonance, i.e. if the electron transit time equals a multiple of half a wave period. The selection of materials andlor coatings is then based on field emission and secondary electron emission properties. For dielectric window break- down, however, no concise models have been developed so far. Elementary processes which are expected to domi- nate the breakdown process, i.e. secondary electron emission and associated charging of dielectrics, and electron in- duced outgassing3, have not been described quantitatively. Even for dc surface flashover, only qualitative models exist4. To explore the physical mechanisms leading to breakdown, we use high speed real-time diagnostics of eas- ily accessible phenomena. The combined results of these measurements yield a comprehensive picture of the events leading to breakdown, and they serve as starting points to develop methods to increase breakdown thresholds. Experimental Setup O-7e034214-3197/Sl 0,0001997 IEEE 135 To reach microwave power densities on the order of 107Wcm-2at a frequency of 2.85 GHz, a magnetron in combination with a traveling wave resonator5 is used (Fig. 1). The magnetron, a Varian type VMS 1143B, has a nominal output power of 3.2 MW and a pulse duration of 2 W. With a 14.6 dB directional coupler, a ring loss per pass of 0.22 dB, and a phase shifter matching the traveling and incoming phases, a stored power of 60 MW with a risetime of 0.9 ps and a pulse duration (FWHM) of 3 I.LS is obtained. The system is operated at a maximum pressure