Experimental study on hard radiation from long laboratory spark discharges in air

Abstract

The detection of hard radiation emanating from an electrical discharge in air is still a mysterious phenomenon. This thesis focuses on collecting experimental data around spark condition that could lead to the production of energetic photon bursts. Long spark discharges with positive and negative polarity in air are studied. We accurately measure the electrical currents on both electrodes during the formation of the discharge. The Xrays are detected with scintillation detectors, time synchronized with the electrical parameters. Bursts up to several 100 keV photons are observed. The advantage of ‘laboratory lightning’ is the controlled environment that allows to study the distribution of the X-rays in space and time. The experiments are performed in the high voltage laboratory at Eindhoven University of Technology. A 2 MV twelve stage Marx generator, with a standardized lightning impulse with 1.2/50 µs rise/decay time to half-maximum when unloaded, delivers the high voltage air breakdown. A 9 m tall 1:2000 capacitive high voltage divider (part of the waveshaping circuit) is used to monitor the voltage waveform produced by the Marx generator. The generator is connected to a spark gap with two conical electrodes at distances varying between 0.76 and 1.46 m. The current at the grounded electrode is measured by a Pearson current probe. An identical probe around the high voltage electrode was connected through a fiber optical data transmission system for electrical isolation. Electromagnetic disturbance from the discharge itself was reduced to a negligible level in the measurements by proper design of the cables and protection equipment. Fast X-ray detectors with good energy resolution are imperative for reliable X-ray registrations. We use conventional NaI(Tl), nanosecond fast BaF2 and two newly developed LaBr3(Ce+) scintillation detectors, all with suitable photomultiplier integrated. Later the DTU National Space Institute assisted in the experiments with a test Cadmium Zinc Telluride (CZT) semiconductor detector intended for the Atmosphere-Space Interactions Monitor (ASIM) project. In early measurements partial discharges at unexpected positions occurred that could also produce X-rays. This effect could be controlled by covering sharp protrusions with conducting foil. The currents measured through both electrodes differ substantially during the firsts few microseconds. This is caused by the Ramo-Shockley effect. In the development phase of the discharge a charge cloud developed around the high voltage electrode and most of the associated electric field lines end in the environment of the grounded electrode, but not on the electrode. Through this current difference between both electrodes it was possible to tell where burst of X-rays are formed. For both polarities of the high voltage, the bursts of X-rays are associated with the negative streamer formation at the cathode. For positive polarity surges X-ray bursts detected coincide with the onset of the upward negative streamer prior to the bridging of the electrode distance. In the case of negative polarity surges X-ray bursts coincide with the negative streamers immediately at the onset of the spark formation. No X-rays have been detected during the large current of the gap breakdown. At gap breakdown both currents become equal. In a parallel investigation we confirmed that the LaBr3(Ce+) scintillation detector suited best for our experimental study because of the short decay time compared with NaI(Tl) and the better energy resolution compared with BaF2. Still, we found that pile-up of multiple photons and/or electrons can occur in the LaBr3(Ce+) detectors in the 23 ns of the 1/e decay. The LaBr3(Ce+) detector has been calibrated and tested for its linearity at photon energies between 59.5 and 2505 keV, employing photomultiplier bias voltages from 568 up to 1000 V. Preliminary X-ray measurements with a small CZT semiconductor detector gave no conclusive results due to the poor detection events (6 out of 100) obtained. Additional experiments with a larger detector for higher detection rates are recommended for a better understanding of the particle distributions involved. An additional experiment was carried out to confirm the emission of X-ray during the streamer phase of an electrical discharge in air. Streamer filaments were produced in a small streamer-corona reactor with nanosecond high voltage pulses up to 65 kV with an optional 20 kVdc offset. The short high voltage pulse period prevents the streamers to develop into an air breakdown. The LaBr3(Ce+) scintillation detector recorded X-rays from the streamer filaments with very consistent timing of occurrences. This proves that processes near the streamers heads are able to produce X-rays. Various results were presented at a number of international conferences and workshops. Chapter 3 and Chapter 4 have been published in Journal of Physics D: Applied Physics.