John T. Krile

Monte Carlo simulation of high power microwave window breakdown at atmospheric conditions

A Monte Carlo-type electron motion simulation program was developed to calculate the increasing electron density for pulsed high power microwave window flashover in air and nitrogen at atmospheric pressures, i.e., >90torr. Through comparison of experimental and simulated results several processes such as flashover delay time’s strong dependence on pressure and the lack of significant surface charge buildup have been confirmed. The quantitative agreement of the code results with the experiment is a clear step towards predicting high power microwave flashover under a wide range of atmospheric conditions as well as for different gases and more complex window geometries.

Interface Breakdown During High-Power Microwave Transmission

The major limiting factor in the transmission of narrowband high-power microwaves (HPM) has been the interface between vacuum-vacuum or even more severely between vacuum-air if HPM are to be radiated into the atmosphere. Extensive studies have identified the physical mechanisms associated with vacuum/dielectric flashover, as opposed to the mechanisms associated with dielectric/air flashover, which are not as well known. Due to the high electron collision frequencies (in the terahertz range) with the background gas molecules, established mitigation methods and concepts of vacuum/dielectric flashover will have to be re-evaluated. The primarily limiting factors of HPM transmission through a dielectric/air interface are presented based on recent experiments at 2.85 GHz. The physics of the involved mechanisms and their practical ramifications are discussed. The potential of surface roughness/geometry for flashover mitigation is addressed as well

Dielectric surface flashover at atmospheric conditions under high-power microwave excitation

Due to recent advances in the peak output power densities and pulse widths of high-power microwave (HPM) devices, the ability to radiate this power into the atmosphere is limited by surface plasma formation at the vacuum-air interface. Very little is known about this window flashover under HPM excitation at “air” side pressures from atmospheric down to approximately 90Torr, and this paper reports one such study at 2.85GHz and MW∕cm2 pulsed power densities. Due to the high (∼600GHz at standard temperature and pressure) elastic collision frequencies of the electrons with the neutral gas molecules and added energy-loss channels through molecule excitation, proven concepts of vacuum flashover, such as multipactoring electrons, have to be abandoned. The observed flashover field is roughly a factor 3 higher in SF6 compared to air, which is consistent with unipolar volume breakdown data. Quantitative comparisons of HPM flashover data with results from a recently developed computer code are given.

Pulsed dielectric surface flashover in nitrogen at atmospheric conditions

Dielectric flashover along insulators in vacuum has been comprehensively researched in the past. Less studied, but of similar importance, is surface flashover at atmospheric pressures and the impact of an atypical electrode geometry, humidity, and ultraviolet (UV) illumination. Previous research has shown distinct discharge behavior in air and nitrogen environments for an electrode geometry in which the applied electric field lines curve above the dielectric surface. It was concluded that the discharge development path, whether along the electric field lines or the surface of the dielectric, is related to the oxygen content in the atmospheric background. It is believed that this dependence is due to the discharges production of UV radiation in an oxygen rich environment. Thus, experiments were conducted in a nitrogen environment employing UV surface illumination in order to observe the affects on the flashover spark behavior. From the experimental data, it can be ascertained that UV illumination and intensity play a significant role in the discharge development path. Based on these results an explanation of the physical mechanisms primarily involved in unipolar surface flashover will be presented. Additional experiments regarding the effects of humidity on the discharge behavior will be discussed as well

Modeling statistical variations in high power microwave breakdown

Flashover of high power microwave (HPM) vacuum isolation windows presents a serious design limitation of megawatt class HPM systems. The delay time from HPM radiation incident on the window to flashover development on the atmospheric side is critical. Previously developed modeling efforts have yielded reasonable correlation with experimentally observed average delay times while failing to capture any statistical variations. Simply preseeding the volume with an initial electron density is identified as inadequate to describe the source of initiatory electrons. The process of field assisted electron detachment is examined and shown to be a probable candidate for the initiatory electron generation.

DC flashover of a dielectric surface in atmospheric conditions

Surface flashover is a major consideration in a wide variety of high-voltage applications, and yet has not been studied in great detail for atmospheric conditions, with modern diagnostic tools. Environmental conditions to be considered include pressure, humidity, and gas present in the volume surrounding the dielectric. In order to gain knowledge into the underlying process involved in dielectric surface flashover, a setup has been created to produce and closely monitor the flashover event. Within the setup parameters such as geometry, material, and temporal characteristics of the applied voltage can be altered. Current, voltage, luminosity, and optical emission spectra are measured with nanosecond to subnanosecond resolution. Spatially and temporally resolved light emission data is also gathered along the arc channel. Our fast imaging data show a distinct trend for the spark in air to closely follow the surface even if an electrical field with a strong normal component is present. This tendency is lacking in the presence of gases such as nitrogen, where the spark follows more closely the electric field lines and develops away from the surface. Further, the breakdown voltage in all measured gases decreases with increasing humidity, in some cases as much as 50% with an increase from 10% relative humidity to 90% relative humidity.

Rapid formation of dielectric surface flashover due to pulsed high power microwave excitation

High power microwave (HPM) dielectric surface flashover can be rapidly induced by providing breakdown initiating electrons in the high field region. An experimental setup utilizing a 2.85 GHz HPM source to produce a 4.5 MW, 3 μs pulse is used for studying HPM surface flashover in various atmospheric conditions. If flashover is to occur rapidly in an HPM system, it is desirable to provide a readily available source of electrons while keeping insertion loss at a minimum. The experimental results presented in this paper utilize a continuous UV source (up to 0.3 mW/cm2) to provide photo-emitted seed electrons from the dielectric surface. Similarly, electrons were provided through the process of field emission by using metallic points deposited on the surface. Initial experiments utilizing 0.2 mm2 aluminum points with a spatial density of 25/cm2 have increased the apparent effective electric field by a factor of ~1.5 while keeping the insertion loss low (<;0.01 dB). The field enhancements have sharply reduced the delay time for surface flashover. For an environment consisting of air at 2.07×104 Pa (155 Torr), for instance, the delay time is reduced from 455 ns to 101 ns. Two radioactive sources were also used in an attempt to provide seed electrons in the high field regions. Presented in this paper is a comparison of various field-enhancing geometries and how they relate to flashover development along with an analysis of time resolved imaging and an explanation of experimental results with radioactive materials.

Effect of Dielectric Photoemission on Surface Breakdown: An LDRD Report

The research discussed in this report was conceived during our earlier attempts to simulate breakdown across a dielectric surface using a Monte Carlo approach. While cataloguing the various ways that a dielectric surface could affect the breakdown process, we found that one obvious effect--photoemission from the surface--had been ignored. Initially, we felt that inclusion of this effect could have a major impact on how an ionization front propagates across a surface because of the following argument chain: (1) The photon energy required to release electrons from a surface via photoemission is less than the photon energy required to ionize gas molecules directly. (2) The mean free path of a photon in gas is longer for low-energy photons than for high-energy photons. (3) Photoionization is a major effect in advancing the ionization front for breakdown in gas without a surface, therefore, we know that even high-energy photons can be released from the head of a streamer and propagate some distance through the gas. Our hypothesis, therefore, was that photons with energies near the threshold of photoemission could travel further in front of the streamer before being absorbed than higher-energy photons needed for photoionization, yet the lower-energy photons, with the help of the surface, could still create seed electrons for new avalanches. Thus, the streamer would advance more rapidly next to a surface than in gas alone. Additionally, the photoemission from the surface would add to the electrons in the avalanche and cause the avalanche to grow faster. After some study, however, we are forced to conclude that although photoemission does contribute to avalanche growth at fields near breakdown threshold, secondary electron emission causes electrons to stick to the surface and cancels out the growth due to photoemission. This conclusion assumes a discharge that occurs over a short period of time so that charging of the surface, which could alter its secondary electron emission characteristics, does not occur. This report documents the numerical work we did on investigating this effect and the experimental work we did on pre-breakdown phenomena in gas.

Effects of UV Illumination on Surface Flashover Under Pulsed Excitation

Undesirable surface flashover of high voltage support structures can severely limit the compactness of open air high voltage systems. Only recently, increased effort has been invested in characterizing and quantifying the physical processes involved in surface flashover occurring under atmospheric conditions and under the influence of UV illumination. In this paper, a UV flash lamp and a solid-state UV source, with its much faster turn-off time, were utilized in conjunction with a high temporal resolution testing apparatus. The UV pulse, excitation voltage, discharge current, and flashover self-luminosity were measured with high temporal precision. We relate recent experiments to our experimental findings of surface flashover under atmospheric conditions gained over the past five years. A simple model that describes the observed behavior will be presented. In addition, a more advanced Monte Carlo-type code for electron collision dynamics will be utilized to further analyze the role of UV in surface flashover under atmospheric conditions.

Contributing Factors to Window Flashover under Pulsed High Power Microwave Excitation at High Altitude

One of the major limiting factors for the transmission of high power microwave (HPM) radiation is the interface between dielectric-vacuum, or even more severely, between dielectric-air if HPM is to be radiated into the atmosphere. Surface flashover phenomena which occur at these transitions severely limit the power levels which can be transmitted. It is of major technical importance to predict surface flashover events for a given window geometry, material and power level. When considering an aircraft based high power microwave platform, the effects on flashover formation due to variances in the operational environment corresponding to altitudes from sea level to 50,000 feet (760 Torr to 90 Torr) are of primary interest. The test setup is carefully designed to study the influence of each atmospheric variable without the influence of high field enhancement or electron injecting metallic electrodes.