Investigation of Long-timescale Behavior of Positive DC Streamer Coronas

Abstract

Positive DC streamers are filamentary low-temperature discharges that are relevant to many applications, including sterilization, ionic wind generation, agriculture and atmospheric electricity. Even when excited by a DC voltage, streamers in atmospheric-pressure air typically self-pulsate with a frequency of several kilohertz. The generally-accepted explanation for DC streamer self-pulsation is that it is driven by recovery of the electric field near the tipped anode, due to electrostatic removal of ionic space charge from the inter-electrode gap over inter-pulse timescales. However, this theory has not been validated, either experimentally or numerically. Most prior works investigating DC streamers have focused on the streamer propagation phase (a few tens of nanoseconds) - few have investigated longer timescales, including the bridging of the electrode gap by the streamer and the subsequent current pulse (hundreds of nanoseconds) and the period in-between streamer pulses, leading up to initiation of the next streamer discharge (hundreds of microseconds). The work presented in this thesis focuses on investigation of the longer timescales of positive DC streamer development in a tip-to-plane geometry, in particular beyond the streamer propagation phase, through the current flow and inter-pulse phases. This begins with an experimental study to measure the long-timescale development of the electric field inside a streamer corona using the E-FISH laser diagnostic technique. This shows some surprising results, which do not seem to be consistent with the theory of DC streamer selfpulsation being driven by electric field recovery at the anode. The near-anode electric field is not observed to recover during the inter-pulse period - instead, the near anode behavior seems to be dominated by a persistent glow discharge and a curious wave-like feature is observed in the electric field, traveling towards the anode on ionic timescales. This is followed by the development of a 1.5D reduced-order numerical model of a DC streamer, which is optimized for solving over long timescales via a ‘triple-stack’ of transient solvers. The model is able to fully resolve the boundary sheath layers of the plasma and is able to capture detailed behavior of the cathode sheath development during bridging via the use of a kinetic flux boundary condition for the charged species. This model is firstly applied to modeling the bridging and current flow phases of streamer development, and its prediction shows a good qualitative match to the behavior of the experimental current pulse. Parameter sweeps show that the streamer current pulse is sensitive to the assumed radial behavior and the rate of electron-ion recombination, but insensitive to the applied boundary conditions or secondary emission. The final section describes an extension of the 1.5D streamer model to simulate the streamer inter-pulse phase and initiation of a second streamer. It is shown that initiation of a second streamer can be predicted by a fluid model and that radial expansion of positive ions plays an important role; however, it has proven difficult to integrate that effect into the 1.5D model. The model results are consistent with streamer self-pulsation being due to electric field recovery; however, comparison with the results of the E-FISH experiment suggest there may be different mechanisms driving positive DC streamer self-pulsation, depending on the presence or not of a glow discharge on the anode.