Plasma breakdown of low-pressure gas discharges

Natural gas discharges like lightning and polar light are spectacular phenomena that have impressed and fascinated people for a long time. During the last two centuries, people have learned how to create their own gas discharges and how to make use of them. Nowadays, man-made gas discharges are commonly used in many applications. Well-known examples include fluorescent lamps, plasma televisions and high intensity lamps used in data projectors. Also in industry, gas discharges are often part of the production process, for instance in the fabrication of computer chips, cleaning of exhaust gases, coating of fabrics and production of solar cells. A gas discharge, or plasma, consists of a variety of different particles like electrons, ions, atoms and molecules in various excited states. These particles not only interact with each other, but also with electric and magnetic fields and with electrode and wall surfaces in the discharge volume. This wide variety of particles and interactions makes a gas discharge a complex system, which is difficult to understand completely and control accurately. To create a gas discharge one generally needs to apply a sufficiently large voltage across a volume of neutral gas. There are many options how to do this, resulting in many different types of gas discharges. Nevertheless, the basic principles involved are the same for all these situations. The electrically insulating, neutral gas is transformed into a conducting, (partially) ionized state by the applied voltage. This evolution from a neutral gas to a selfsustaining discharge is known as plasma breakdown, or plasma ignition, and is the subject of this thesis. Plasma breakdown is a fundamental process in gas discharges; it is a highly transient process that involves particles drifting in electric fields, charge multiplication in electron avalanches and moving ionization fronts. The driving force for these processes is the electric field in the discharge volume. The research in this thesis was aimed at obtaining a better understanding of the fundamental processes involved in plasma breakdown in low-pressure discharges by experimental investigations. Two types of discharges were studied; a pulsed discharge between parabolic, metal electrodes and a parallel-plate, low-pressure dielectric barrier discharge. The breakdown 155 phases of these discharges were investigated using various experimental techniques. Breakdown processes in the low-pressure dielectric barrier discharge were investigated by studying the light emission from the discharge in a spatially, temporally and spectrally resolved way. Additionally, electrode voltages and discharge currents were measured. These investigations, together with the results from a two-dimensional fluid model, showed that the breakdown process in this discharge followed a Townsend-like mechanism in which the effects of the dielectric plates were limited. The pulsed discharge between parabolic, metal electrodes, was firstly studied in a lowpressure argon environment by light emission imaging with an intensified charge-coupled device (ICCD) camera. This relatively simple diagnostic provided time- and space-resolved information on the characteristic features of the breakdown process. Different phases in the breakdown process were identified. Firstly, the build-up of a light emission region in the discharge gap in front of the anode, followed by a light front crossing the electrode gap from anode to cathode and finally, a stable discharge covering the cathode surface. These features were in qualitative agreement with the breakdown process observed in parallelplate discharges at low pressure, which are accurately described by Townsend’s breakdown theory. The ICCD imaging experiments also showed that before the main breakdown process started, a weak flash of light could be observed around the anode. This pre-breakdown light emission occurred during the rise of the applied voltage, but before the breakdown voltage was reached. The origin of this feature was found to be electron avalanches seeded by volume charges left over from previous discharges in combination with the specific discharge geometry used in our experiments. It was concluded that the initial conditions of the discharge influenced the breakdown process. Although the qualitative behaviour of the main breakdown phase did not change for a wide variety of discharge conditions, the details of the process, especially the timing of the different phases were strongly influenced by the initial conditions. Finally, a new diagnostic was developed to measure electric field distributions during the breakdown phase of a discharge. With this diagnostic, electric field strengths were determined by measuring Stark effects in xenon atoms using laser-induced fluorescence-dip spectroscopy. Stark shifts of up to 4.8 cm-1 (160 pm) were observed for ns and nd Rydberg states, with principal quantum numbers ranging from 12 to 18, as a result of electric fields between 250 and 4000 V/cm. Additionally, a theoretical calculation method, based on solving the Schr¨odinger equation for a perturbed Hamiltonian by matrix diagonalization, proved to be very accurate for describing the observed Stark effects in nd Rydberg levels. With this diagnostic we performed measurements of the electric field distribution during the breakdown phase of the discharge between parabolic electrodes in xenon. For the first 156 time, quantitative, direct measurements of the evolution of electric field during breakdown were obtained. Electric fields between 0 and 1600 V/cm were measured with a resolution of 200–400 V/cm, depending on the magnitude of the electric field. These experiments showed that the ionization front, already observed in the ICCD imaging experiments, is sustained by a spatially narrow, rapidly moving region of strong electric field. Additionally, this ionization front did not completely modify the potential distribution in the discharge gap; the discharge continued developing towards a steady-state after the ionization front crossed the gap. In conclusion, the investigations in this thesis show that the fundamental processes involved in plasma breakdown are well understood. However, the details of the process, especially timing of events, depend strongly on the specific discharge geometry and the initial discharge conditions. The electric field diagnostic has proven to be very useful for breakdown studies, identifying a narrow, moving region of high electric field as the cause of the moving ionization fronts . This diagnostic offers the possibility to obtain quantitative, direct information on local electric field strengths with both spatial and temporal resolution. In future research, such electric field measurements can be very valuable not only for investigations on low-pressure breakdown but also for studies on breakdown processes at higher pressures, which are only poorly understood at the moment