The impact of field enhancements and charge injection on the pulsed breakdown strength of water

In any high voltage pulsed power system there is a need for a dielectric material to serve as a charge storage medium, switching medium, or insulator. Water, with its high dielectric constant, εr = 81, and favorable physical properties is an ideal candidate for use in compact pulsed power systems. Several research efforts have been conducted over the last several decades to investigate possible ways of increasing water’s dielectric strength [Joshi, Kun]. In the research documented here, experiments have been conducted in order gain further insight into the mechanisms that initiate the electrical breakdown of water.

With the application of a large enough pulsed electric field, any metal conductor will begin to emit electrons from its surface, especially from field enhancements on the electrode surface, into the dielectric region. This emission initiates the breakdown of water, though various theories of exactly what happens in the water prior to breakdown have been developed. Some suggest that the breakdown event is purely electronic where the emission enters the gap until a conductive channel is formed [1]. Others have suggested that the emission causes rapid heating within the water causing it to vaporize and a bubble to form [2,3,4]. Neither theory has been conclusively confirmed as the primary mechanism.

Several factors have been found to either increase or decrease the electric holdoff strength of water including the electrode surface roughness, the electrode material, the electrode surface area, and the water conductivity. Experiments have been conducted to test the effect each of these factors has on water’s dielectric strength and in each experiment, a water gap was tested under pulsed conditions with pulse widths of roughly 2 μs. Peak electric fields over 1 MV/cm and peak currents over 3.5 kA have been recorded across the gap. In all of the tests, electrodes machined with a Bruce profile were used on both the anode and cathode sides of the gap. Random and known surface roughness patterns were applied to the electrodes through mechanical sanding and etching processes. Surface roughnesses ranging from 0.26 ìm to 1.96 ìm and electrode surface areas ranging from 0.44 cm2 to 75 cm2 were tested. Electrodes constructed of various materials including Aluminum, Molybdenum, Copper, Tungsten, Nickel, Stainless Steel, and Zinc Oxide, all of which had a constant surface area of 5 cm2, were also tested. The conductivity of the water was varied from 1 µS/cm to 38.5 µS/cm. Additionally, shadowgraph images of a point plane geometry were taken to further understand the breakdown processes that occur.