52.80.-s Electric discharges (see also 51.50.+v Electrical properties of gases; for plasma reactions including flowing afterglow and electric discharges, see 82.33.Xj in physical chemistry and chemical physics)
Refine
Document Type
- Doctoral Thesis (3)
Has Fulltext
- yes (3)
Is part of the Bibliography
- no (3)
Keywords
- Atmosphärendruckentladung (1)
- Barrierenentladung (1)
- Brennfleck (1)
- Emission (1)
- Gasaufzehrung (1)
- Gasentladung (1)
- Gastemperatur (1)
- Kathode (1)
- Laserinduzierte Fluoreszenz (1)
- Niederdruckentladung (1)
Institute
This thesis investigated dielectric barrier discharges (DBDs) in N2-O2 gas mixtures at atmospheric pressure, with a focus on the gas discharge physics. The main goal was to evaluate whether possible control mechanisms exist that can manipulate the breakdown and the development of DBDs, especially for pulsed operation. To examine the pre-breakdown phase, the actual breakdown and the main DBD development, DBDs in a double-sided, single filament arrangement with a 1 mm discharge gap were investigated by means of electrical and optical diagnostics with high resolutions. Spectrally- and temporally-resolved iCCD pictures (2D in space), spectrally- and spatio-temporally-resolved streak camera and CCS images (1D in space) were simultaneously recorded accompanied by a full electrical characterisation with fast voltage and current probes. Sinusoidal- and pulsed-driven DBDs were found to have a qualitatively similar spatio-temporal development, i.e. a cathode-directed ionisation front (v ~ 10^6 m/s, positive streamer mechanism), followed by a transient glow-like phase in the gap. For sinusoidal operation, the slope of the applied voltage is flat (dU/dt ~ 1 V/ns) compared to pulsed operation (dU/dt ~ 100 V/ns). Thus, during the longer pre-phase of the sine-driven DBD, many more charge carriers were generated, in contrast to the pulsed-driven DBDs, where the pre-phase is limited by the short voltage rise time. Consequently, just before the breakdown occurs, the charge carrier density is higher for sine-driven DBDs, i.e. the positive streamer starts in a highly pre-ionised environment, which leads to a lower propagation velocity. In addition to limiting the pre-phase (lower pre-ionisation), the steep voltage slope of the pulsed DBD amplifies the streamer breakdown because the applied voltage rises significantly during its propagation. Therefore, the transferred electrical charge and the electrical power of a single DBD can be controlled by the applied voltage amplitude, but only in pulsed operation. In addition to the effects of different voltage slope steepness, the pulse width is an excellent parameter in the pulsed operation to set the pre-ionisation, by shifting the DBDs into the after-glow of the previous discharge using asymmetrical HV pulse waveforms. The subsequent DBDs ignite in different pre-ionised conditions, defined by the residual charge carrier densities in the gap that originated from the previous DBD. The breakdown characteristics of these DBDs could be controlled down to the fundamental level. This thesis has described for the first time four different breakdown regimes in single filament DBDs for 0.1 vol% N2 in O2 and connected them to the processes during their pre-phases. The “classic” DBD development (a cathode-directed streamer followed by a transient glow discharge) could be controlled in a certain range, followed by a transition first to a breakdown regime featuring a simultaneous propagation of a cathode- and an anode-directed streamer, and finally to a reignition of the previous DBDs without any propagation, just by reducing the pulse width (time between two subsequent DBDs), i.e. increasing the pre-ionisation level. All differences between the DBDs at rising and falling slopes could be explained by the different pre-conditions in the gap. The O2 concentration in the N2-O2 gas mixtures offers another way of controlling the pre-ionisation. Due to the electron attachment as a consequence of the electronegativity of oxygen, the electron density decreases for higher O2 admixtures. Furthermore, the differences in the first Townsend ionisation coefficient and in the photo-ionisation between N2 and O2 influence the DBD behaviour as well. To some extent, some of the reported effects achieved by varying the pulse width at a fixed O2/N2 ratio were also observed for a fixed pulse width and changing O2 concentration. Hence, the response of the DBD properties to changing pre-ionisation levels seems to be a general principle of DBD control. Additional effects of the O2/N2 ratio, such as an increasing DBD inception jitter or higher streamer velocities, were also reported. Finally, a reverse of the effects induced by the O2 admixture such as DBD emission duration or DBD inception delay, was observed for O2 concentrations below 0.01 vol%, and were especially pronounced at a pressure of 0.5 bar. For 0.1 vol% O2 in N2, a minimal electron recombination rate was found, which can be explained by the different decay and recombination rates of positive nitrogen and oxygen ions. These different rates effect the charge carrier dynamics and consequently, the pre-ionisation in the gap. In conclusion, this investigation has highlighted the importance of volume memory processes on the breakdown and development of single filament DBDs at elevated pressures.
The absolute density of the metastable N2(A,v=0) molecule was extensively studied in nitrogen barrier discharges at 500 mbar. For the detection of the metastables laser-induced fluorescence spectroscopy (LIF) was used, at which for the calibration of the absoute metastables density a comparison with Rayleigh scattering was performed. To get the ratio of the LIF signal to the Rayleigh signal it is shown that the LIF signal is the convolution of the Rayleigh signal with an exponential decay. Besides, the different cross sections are calculated and the ratio of the detection sensitivities at the laser and fluorescence wavelength is determined. As a first step on the way to atmospheric pressure barrier discharges, the laser-induced fluorescence spectroscopy was implemented in low pressure capacitively coupled radio-frequency discharges. The determined metastables density in the capacitively coupled radio-frequency discharge is somewhat below 10^12 cm^(-3) at 40 Pa and somewhat below 10^13 cm^(-3) at 1000 Pa. The axial density profiles show a nearly symmetric shape due to the long lifetime of the metastable state. At a pressure of 500 mbar the two discharge modes of the barrier discharge, the filamentary and the diffuse mode, were analysed. The filamentary mode was mainly investigated in an asymmetric discharge configuration. Typical densities in the detection volume are in the range of 10^13 cm^(-3), resulting in maximal densities of up to 10^15 cm^(-3) in the microdischarge channel. Such large densities are in agreement with the fast decay by the pooling reaction after the maximum of the metastables density in the afterglow of the discharge pulse. The time dependent measurements in the afterglow of single microdischarges offer a delay of the metastables production with respect to the discharge current. This delay indicates that the metastables production takes place mostly by cascades from higher triplet states, which are in turn excited by electron impact. The axial density profiles show a maximum in metastables density in front of the anode in agreement with optical emission spectroscopy, but which cannot be clearly identified because of the asymmetric discharge configuration. The measurements for the diffuse discharge mode were performed in a symmetric discharge configuration. The metastables density is in the range of 10^13 cm^(-3). It increases during the current pulse of the discharge and decays afterwards. The maximum of the metastables density is delayed with respect to the maximum of the discharge current. The depletion of metastables in the early discharge afterglow is dominated by the pooling reaction, afterwards quenching by nitrogen atoms becomes important assuming a nitrogen atom density in the order of 10^14 cm^(-3). As for the filamentary mode, the losses by diffusion are negligible for the measurement positions. The measured axial density profiles show an accumulation of metastables in front of the anode, whereas the density in front of the cathode is below the detection limit. To calculate the metastables current density to the dielectrics after the discharge pulse a simulation is developed including the dominant volume processes for the depletion of metastables and the axial diffusion. Starting point for the simulation is the axial metastables density distribution at the end of the discharge pulse. The calculated metastables current density at the dielectrics is in the range of 10^14 cm^(-2)s^(-1). With the use of recently calculated secondary electron emission coefficients a comparison of the secondary electron emission by metastables with the discharge current is done. It is figured out that the secondary electron emission current is large enough to be important during the discharge ignition. To expand the simulation to the whole voltage cycle, the excitation of metastables is assumed to be proportional to the discharge current and electron density. Using this model, the measured time dependences of the metastables density are well reproduced for the investigated parameter variations. This is not the case for the axial profiles, where a metastables loss process is missed to explain the formation of a density plateau in front of the anode during the discharge pulse. The intended calculation of the metastables current density shows that the delay of the metastables production with respect to the discharge current might be responsible for the ignition of microdischarges at the beginning of the discharge pulse.
In der vorliegenden Arbeit wurde die Katodenregion einer quecksilberfreien Helium-Xenon Niederdruckentladung im Brennfleckbetrieb experimentell untersucht. Diese Region ist von besonderem Interesse, da sich hier die Elektronenemission, die Erzeugung von Ionen und metastabilen Atomen sowie lebensdauerbegrenzende Prozesse abspielen. Um die Entladung im Brennfleckbetrieb zu realisieren, kam als Katode eine im Rahmen dieser Arbeit entwickelte neuartige planare Geflechtelektrode zum Einsatz. Mit der Methode der ortsaufgelösten Laser-Atom-Absorptionsspektroskopie (LAAS) wurden die absoluten Teilchendichten der zwei untersten angeregten Xe-Atome und die Gastemperatur in der Katodenregion bestimmt. Die Inhomogenität des Spot-Plasmas fand dabei besondere Berücksichtigung. Sowohl die Teilchendichten der zwei untersten angeregten Xe-Atome als auch die Gastemperatur sind unmittelbar vor dem Brennfleck maximal und fallen in axiale und radiale Richtung stark ab. Insbesondere die Gastemperatur beträgt in einem Abstand von 1 mm vor dem Brennfleck circa 650 K und liegt damit deutlich über Raumtemperatur. Des Weiteren ließ sich die Temperatur im Brennfleck auf der Katodenoberfläche mittels optischer Emissionsspektroskopie ermitteln. Dies geschah durch Anpassung des aufgenommenen Spektrums an die Plancksche Strahlungsgleichung. Die Brennflecktemperaturverteilung weißt ein ausgeprägtes Maximum auf, das je nach Entladungsstromstärke maximale Werte zwischen 1414 K bei 40 mA und 1524 K bei 80 mA annimmt. Von diesem Maximum aus wurde ein starker in alle Richtungen nahezu symmetrischer Temperaturabfall festgestellt. Ein technologisch wichtiger Aspekt hinsichtlich der Lebensdauer einer auf Xenon basierenden quecksilberfreien Lampe ist der negative Effekt der Xe-Gasaufzehrung. In dieser Arbeit wird gezeigt, dass die Gasaufzehrung unter Verwendung der planaren Geflechtelektrode im deutlichen Gegensatz zur industriell gefertigten Becherelektrode, wie sie vielfach in Lampen für Lichtwerbung vorkommt, vernachlässigbar klein ist. Dies wird auf die Ausbildung eines heißen Brennflecks und die damit verbundene hohe Katodentemperatur und niedrige Katodenfallspannung zurückgeführt.