Session CI2: Plasma Technology: Plasma Thrusters, Plasma Discharges

Experimental and theoretical studies of cylindrical Hall thrusters

The Hall thruster is a mature electric propulsion device that holds considerable promise in terms of the propellant saving potential. The annular design of the conventional Hall thruster, however, does not naturally scale to low power. The efficiency tends to be lower, and the lifetime issues are more aggravated. Cylindrical geometry Hall thrusters have lower surface-to-volume ratio than conventional thrusters and, thus, seem to be more promising for scaling down [Y. Raitses and N.J. Fisch, \textit{Phys. Plasmas} \textbf{8}, 2579 (2001)]. The cylindrical Hall thruster (CHT) is fundamentally different from the conventional design in the way the electrons are confined and the ion space charge is neutralized. Both the large (9 cm channel diam., 600 -- 1000 W) and miniaturized (2.6 cm channel diam., 50 -- 300 W) CHTs exhibit performances comparable with those of the state-of-the-art conventional (annular) design Hall thrusters of similar sizes [A. Smirnov \textit{et al}., \textit{J. Appl. Phys}.\textbf{ 92}, 5673 (2002)]. A comprehensive experimental and theoretical study of the CHT physics has been conducted, addressing the questions of electron cross-field transport, propellant ionization, plasma-wall interaction, and formation of the electron distribution function. Probe measurements in the harsh plasma environment of the micro thruster were performed. Several interesting effects, such as the unusually high ionization efficiency and enhanced electron transport, were observed [A. Smirnov \textit{et al.}, \textit{IEEE Trans. on Plasm. Sci}. \textbf{34}, 132 (2006)]. Kinetic simulations suggest the existence of the strong fluctuation-enhanced electron diffusion and predict the non-Maxwellian shape of the electron distribution function [A. Smirnov \textit{et al.}, \textit{Phys. Plasmas} \textbf{11}, 4922 (2004)]. Through the acquired understanding of the new physics, ways for further optimization of this means for low-power space propulsion are suggested. Substantial flexibility in the magnetic field configuration of the CHT is the key tool in achieving the high-efficiency operation.   

Electron, Atomic, and Radiation Kinetics in Plasma Discharge Lighting: Advanced Models and Observations

Non-LTE discharges used in lighting sources provide an excellent testbed for understanding the interplay between plasma, atomic, and radiation physics. Standard models for the Hg fluorescent bulb include non-equilibrium kinetics for the species, but employ both a 0-D Boltzmann equation for the electron distribution function (EDF) and Holstein's probability-of-escape for radiation transport. These assumptions overlook some of the more interesting, and challenging, aspects of plasma lighting. The radial ambipolar potential requires the inclusion of the spatial gradient term in the inhomogeneous electron Boltzmann equation. The resulting EDF is found to depend on both electron energy and radial position [1]. Advanced radiation transport techniques account for non-local photo-pumping, line overlap within the Hg resonance lines, and partial frequency redistribution [2]. The results of our completely coupled model match the observed spatial distribution of Hg excited states and the line-of-sight intensity [3]. Due to environmental initiatives there is also recent interest in non-Hg discharges for high intensity lighting. One example is an RF electrodeless Mo-O-Ar plasma discharge bulb which operates by recycling the emitting Mo with an O catalyst. Based on atomic physics calculations for Mo [4], the kinetic pathways leading to visible emission can be identified [5] and explain the measured lighting efficiency of $\sim $40 lumens/watt of supplied power.\newline \newline[1] J. Appl. Phys., 94, p.62, 2003. \newline [2] Plasma Sources Sci. Tech., 14, p.236, 2005. \newline [3] J. Phys. D., 38, p.4180, 2005. \newline [4] New J. Physics, 6, p.145, 2004. \newline [5] J. Appl. Phys., 95, p.5284, 2004.   

Fundamentals and Applications of a Plasma Processing System Based on Electron Beam Ionization

Electron beam (e-beam) ionization has been shown to be both efficient at producing plasma and scalable to large area (square meters). NRL has developed a number of advanced research tools culminating in a ``Large Area Plasma Processing System'' (LAPPS) based on an e-beam sheet geometry. We have demonstrated that the beam ionization process is fairly independent of gas composition and capable of producing low temperature plasma electrons ($<$0.5 eV in molecular gases) in high densities (10$^{9}$-10$^{12}$ cm$^{-3})$. This system can offer increased control of plasma-to-surface fluxes and the ability to modify materials' surface properties uniformly over large areas. The systems to be discussed consist of continuous and pulsed planar plasma distributions generated by a magnetically collimated sheet of 2-3kV, $<$ 1 mA/cm$^{2}$ electrons injected into a neutral gas background (oxygen, nitrogen, sulfur hexafluoride, argon). Typical operating pressures range from 20-150 mTorr with beam-collimating magnetic fields (100-200 Gauss) for plasma localization. The attributes of beam-generated plasmas make them ideal for many materials applications. These systems have been investigated for a broad range of applications, including surface activation, line edge roughening, and anisotropic etching of polymers, electron-ion and ion-ion plasma etching, low-temperature metal nitriding and thin film deposition (reactive sputtering and plasma enhanced chemical vapor deposition). Details of some of these applications will be discussed in terms of the critical plasma physics and chemistry, with complementary time-resolved \textit{in situ} plasma diagnostics (Langmuir probes, microwave transmission, energy-resolved mass spectrometry and laser spectroscopy).   

Dielectric Surface Flashover at Atmospheric Conditions under High Power Microwave Excitation

Due to recent advances in the peak output power densities and pulsewidths of high power microwave (HPM) devices, the ability to radiate this power into the atmosphere is limited by surface plasma formation at the vacuum-air interface. In applications such as high power radar or electronic warfare, the 'air' side of the window may rest at pressures from atmospheric down to approximately 90 Torr. Very little is known about window breakdown under HPM environments and this paper reports one such study. Due to the high (THz) elastic collision frequencies of the electrons with the neutral gas molecules and added energy loss channels through molecule excitation, the well-established lessons on VACUUM-surface flashover cannot be directly applied to the dielectric-AIR flashover. For example, dielectric flashover in an atmospheric environment has a lower threshold than flashover in vacuum. This paper further discusses that a typical transition point from vacuum to atmospheric flashover occurs at $\sim $ 1 Torr, where proven concepts of vacuum flashover, such as multipactoring electrons, have to be abandoned. It will describe new experimental and modeling results of MW/cm$^{2}$ pulsed power densities at 2.85 GHz and 110 GHz. Spectroscopic measurements confirm that the N$_{2}$ molecular vibrational temperature is $\sim $ 2,700 K, while the rotational temperature is $\sim $ 300 K, thus indicating the non-thermalized nature of the flashover plasma kinetics. A universal scaling law has been shown to exist between the flashover electric field strength divided by pressure, $E$/$p$, and the gas pressure times the flashover delay time, \textit{p $\tau $} over an extremely large parameter space. The impact of UV radiation on the flashover path is determined. Quantitative comparisons of data with most recently developed theoretical models and computer codes are given.   

Fermi Acceleration -- From Cosmic Rays to Discharge Heating

Low pressure, radio frequency (rf) driven discharges are widely used for materials processing in the microelectronics industry. Electrons in these discharges can be heated ``collisionlessly'' by repeated interaction with the fields near the plasma skin. The physical description of this ``collisionless'' heating harks back to two seminal ideas originating over fifty years ago in the disparate fields of astrophysics and condensed matter physics. The motion of a ball bouncing between a fixed and an oscillating wall was originally proposed by E.Fermi [{\it Phys. Rev.} {\bf 75} 1169 (1949)] in April 1949 as a model for cosmic ray acceleration. Expectations that the ball could be heated to very high energies gave way to the realization that while the motion is chaotic at low energies, the phase space has an intricate fractal structure and there is an adiabatic limit to the heating. Also in April 1949, A.B. Pippard [{\it Physica} {\bf 15} 45 (1949)] proposed an explanation for the anomalous high frequency skin resistance of metals at low temperatures, in which he divided electrons into two classes that ``interacted with'' and ``did not interact with'' the skin layer fields. The application of these ideas to collisionless electron heating in discharges has been fruitful [M.A. Lieberman and V.A. Godyak, {\it IEEE Trans. Plasma Sci.} {\bf 26} 955 (1998); E. Kawamura, M.A. Lieberman, and A.J. Lichtenberg, {\it Phys. Plasmas} {\bf 13} 053506 (2006)]. In this talk, Fermi's proposal for the origin of cosmic rays is reviewed. The Fermi acceleration model is used to describe collisionless heating in radio frequency-driven discharges, with illustrations drawn from experiments, computer simulations and analysis. The re-discovery of Pippard's model of the anomalous skin effect in metals, in the context of collisionless heating in discharges, is described.   

FDTD simulation of fusion plasmas at RF time scales.

Simulation of dense plasmas in the RF frequency range are typically performed in the frequency domain, i.e., by solving Laplace-transformed Maxwell's equations. This technique is well-suited for the study of linear heating and quasi-linear evolution, but does not generalize well to the study of non-linear phenomenon. Conversely, time domain simulation in this range is difficult because the time-scale is long compared to the electron plasma frequency, yet the waves still have appreciable electromagnetic character. Thus, a full set of Maxwell equations is needed, but one cannot afford a time-step small enough to resolve the full physics of the plasma. A long time step, coupled with the various cutoff and resonance behaviors within the plasma, insure that an explicit finite-difference scheme would be numerically unstable. To resolve this dilemma, we introduce a locally implicit method to treat the plasma current, while preserving the performance of explicit finite-difference for the Maxwell terms. Careful time-centering provides an energy conserving algorithm that will faithfully reproduce all CMA-diagram dispersion behavior, at the available temporal and spatial resolution, despite the fact that the simulation time-step may exceed the electron gyro and electron plasma time scales by orders of magnitude. The remaining stability criterion is the vacuum speed-of-light Courant condition. We plan to couple this implicit algorithm, as a noise-free background plasma, to a particle-in-cell method, in order to provide more effective study of kinetic and non-linear effects. We present 3-D demonstrations of the method for several classical benchmarks, including mode-conversion to ICW (ion cyclotron wave), cyclotron resonance, propagation into a plasma-wave cutoff, and tunneling through low density edge plasma.