The edge and scrape-off-layer (SOL) of a tokamak plasma, while comprising only a few percent of the minor radius, is of great importance in understanding global tokamak confinement and heat deposition on the divertor plates. In this talk, a tutorial introduction to some of the relevant physics governing boundary plasma instabilities and turbulence will be presented, including the use of the collisional fluid Braginskii equations, the role of X-point geometry on wave physics, and the types of modes and instability drives thought to be important. Then, the results of recent work will be presented in which we have identified the curvature-driven resistive X-point mode as the dominant instability of a characteristic L-phase discharge in the DIII-D tokamak. This mode, expected to be a generic L-mode edge/SOL instability, is electromagnetic in the "bad curvature" region, but transitions to an electrostatic mode near the X-point due to the combined effects of resistivity and X-point magnetic shear. Motivated by observations of elevated electron temperature near the X-point, we investigate heating and parallel energy flow induced by the resistive X-point mode. We speculate that energy in the unstable waves flows to and dissipates in the X-point region, heating the electrons. Progress in understanding the nonlinear saturation levels and implications of this mode for perpendicular wave-induced transport will be discussed.
Landau's prescription for determining the dispersion relation of waves in a collisionless plasma with a given velocity distribution for each species lies at the heart of kinetic plasma theory. However the treatment of damped modes requires the analytic continuation of the distribution functions for complex velocities.
In practical application, velocity distributions are often determined experimentally, or are numerically generated by computer simulations. These general distribution functions cannot be assumed analytic, nor can they necessarily be well-modeled by a superposition of analytic functions (e.g., Maxwellians).
An alternative perspective on the problem of determining the dispersive properties of damped plasma waves will be presented. This method draws a connection between the plasma susceptibility and the convolution (or deconvolution) of the distribution function with simple Lorentzians. This approach lends itself to the numerical determination of the plasma susceptibility for piecewise-analytic distributions that have been subjected to Gaussian smoothing.
The Penning trap confines plasma of electrons (or ions) by using an electrostatic potential to prevent the escape of particles along magnetic field lines. A modified Penning trap has been constructed which has an azimuthal magnetic field made by wires along the axis. The resulting field is helical and particles execute bounce orbits along helical field lines. Drifts cause a particle path in the +z direction to be at a slightly larger radius than in the -z direction so that the orbit is oval, like a rubber band. Collisions cause a step-wise change in position with a characteristic length determined by the width of the oval rather than the the Larmor radius. Experiments in the new trap allow aspects of transport theory for toroidal devices to be isolated and tested. The first experiment shows that the transport due to electric mobility scales with the width of the oval drift orbit (the "banana width") and not with the Larmor radius.
Recent advances in available on-board electrical power in satellites has moved plasma propulsion devices from the laboratory to space-based application. In the past year alone more than 80 spacecraft employing some type of plasma propulsion system were flown on commercial vehicles. This talk will present a brief introduction to plasma thrusters including Arcjets, Resisojets, Electrostatic Ion Thrusters, and Magnetoplasmadynamic Thrusters before entering a more detailed analysis of the plasma transport properties in the plume of closed-drift Hall-effect thrusters. The Hall thruster utilizes static crossed electric and magnetic fields to accelerate ionized xenon producing a reactive thrust. The experimental investigation reported utilized in-situ probes to quantify the transport of charged and neutral species in the flowing exhaust plume as well as the construction of a molecular beam mass spectrometer to provide species-dependent measurements of the heavy-particle energy distributions. Among the phenomena discovered was the existence of an anomalous population of ions having apparent acceleration voltages of up to three times that applied to the discharge electrodes. Discussion of such phenomena in the context of charge- and momentum-transfer collisions will be presented.
Numerical, as opposed to symbolic, differential algebras have a growing number of uses in computational science. These differential algebras can be used for simultaneous calculation of some quantity plus a certain number of its derivatives up to some order in some number of variables. Solving a differential equation for a differential algebra element gives the Taylor series expansion for the transfer map - the solution for initial conditions near some particular initial condition. In recent years, it has been recognized that numerical differential algebras are most easily treated by Object Oriented Programming methods in C++, a language that allows operator overloading. In this talk we will review numerical differential algebras and object oriented programming methods. We then discuss the implementation of numerical differential algebras within C++ and show that by judicious choice of data layout, the effiency of the numerical calculations can be increased by an order of magnitude.
The effect of poloidally mode coupled, ballooning type electrostatic drift waves on a magnetic island has been studied both analytically and numerically. It has been shown quantitatively that particle orbits become stochastic and their behavior can be a possible candidate for the radial plasma transport across a magnetic island of a tokamak. The transport is significant in that it takes place even when the flux surface is not destroyed. The mechanism of the stochasticity generation is understood as an overlapping of secondary islands caused by resonance between periodic particle motions in the magnetic island and Fourier modes of E x B drift due to the electrostatic drift waves. The diffusion process perpendicular to the island magnetic surface has been shown to follow the Gaussian type and can be influential for the deterioration of the plasma confinement. In addition, local diffusion process in the vicinity of Kolmogorov, Arnold and Moser (KAM) surfaces is discussed.
The spatiotemporal evolution of stimulated Brillouin scattering (SBS) in homogeneous plasmas and some aspects of the influence that nonlinear and kinetic effects have on the evolution of SBS were studied.
A one-dimensional analytical linear model based on a fluid description of the plasma was initially developed. It was found that the threshold intensity of the absolute instability and the steady-state spatial growth rate of the convective instability are both independent of the scattering angle. However, the saturation time of the convective instability exhibits a strong inverse dependence on the scattering angle.
The basic model was improved by extending the one-dimensional analysis to include two spatial dimensions and time. In order to assess the effects that the finite size of the laser beam has on SBS, wide and narrow laser-beam geometries were considered. Detailed comparison were made between the predictions of a reduced 1D and 2d models, which can be solved and analytically, and the results of 2D numerical simulations.
The influence that nonlinear and kinetic effects have on SBS was investigated by performing particle-in-cell (PIC) simulations. The results of these PIC simulations were compared against fluid simulations, and good agreement was obtained for sufficiently weak laser intensities. When the laser intensity is sufficiently strong for ion trapping to be significant, PIC and fluid simulations differ substantially. The SBS reflectivity is shown to depend sensitively on the frequency mismatch between the light wave used to seed the instability and the incident laser.
Plasmas, the ionized states of matter, are usually hot and gaseous. However, a sufficiently cold or dense plasma can be liquid or solid. We trap beryllium ions in a Penning trap, and utilize laser cooling to reduce the ion temperature to less than 5 mK. By applying an asymmetric electric field that rotates in the same sense as the ions, we are able to phase-lock the rotation of these plasmas, therefore enabling precise control of the plasma density and shape. The ions freeze into Coulomb crystals, which we have studied through both Bragg diffraction and spatial imaging. The crystals have a rich phase structure, whose features are shared with such diverse systems as neutron star crusts, hard spheres, colloidal suspensions and semiconductor electron bilayers. Plasma modes can be excited with potentials applied to the trap electrodes, and directly imaged by changes in the ion resonance fluorescence produced by Doppler shifts from the coherent velocities of the mode. Enhanced radial transport is observed where modes are resonant with static external perturbations; similarly, the plasma angular momentum can be usefully changed through the deliberate excitation of azimuthally asymmetric modes. Precise control of the plasma's angular momentum and structure is important for possible applications such as frequency standards, quantum computing and antihydogen production.
I will briefly review the physics of the saturation of kinetic instabilities due to wave trapping, and the role of collisions in wave-particle resonant interaction. I will provide a general introduction to the delta-f method, then discuss some of the difficulties in extending delta-f method to cases with collisions, and show how the introduction of a convenient tool (the marker distribution in extended phase space) overcomes such difficulties. I will then try to apply the new method to the simulation of a recent TFTR experiment and discuss the result.
WINDMI is a nonlinear dynamical model for the coupled Solar-Wind--Magnetosphere--Ionosphere system. The model couples the four basic energy components of the night-side magnetotail (lobe magnetic energy, current sheet ExB energy, parallel kinetic energy, and thermal energy) to the ionosphere by the nightside region 1 currents (substorm current wedge). It includes the large ion gyroradius kinetic physics in the quasineutral sheet that converts ExB convection to thermal energy through the chaotic conductivity, and the finite parallel heat flux neglected by MHD. The model predicts both the state of the geotail and the ionospheric westward electrojet index from the value of the solar wind speed. In the absence of solar wind driving and ionospheric damping the model conserves energy and is Hamiltonian. With solar wind driving and ionospheric damping the system is consistent with Kirchhoff's rules expressing the conservation of charge and energy. This ensures that the solar wind in! put power is divided into physically realizable sub power components, a property not shared by signal processing filters for instance. WINDMI provides a consistent mathematical framework that can be used to investigate different possible scenarios for the evolution of the driven Wind-Magnetosphere-Ionosphere system. We will report on three studies of such models: (1) lobe stored magnetic energy as a function of the IMF and solar wind dynamic pressure, (2) the role of the nonlinear ionospheric conductivities and (3) the change in dynamics with the nature of the unloading mechanism, as for example the near Earth neutral line.
The most popular theory of solar system formation states that the nine planets accumlated from a disk of gas nd dust orbiting the Sun. N-body simulations of the final stages of planet formation from several hundred planetary embryos yield satisfactory results in the inner solar system, but completely fail to form massive planets in the outer solar system. The problem is that bodies in the outer solar system are weakly bound to the Sun and widely dispersed. Mutual gravitational interactions in the outer solar system produce large velocity dispersions, small collision cross sections, and stalled planetary growth.
A possible solution to this problem is the addition of collective gravitational wave dynamics in the disk. Planetary embryos tend to excite waves in the disk which can strongly damp the velocity dispersion of the larger bodies, leading to larger collision cross sections and more rapid planetary growth. Although the theory of wave-planet interactions is fairly mature, practical techniques for including wave-planet interactions in N-body simulations remain to be developed. I will discuss a reduced description of wave-planet interactions that is somewhat similar to reduced descriptions of wave-particle interactions that are used in plasma physics simulations. The goal of this talk is to draw useful analogies between the methods of plasma physics and the methods of solar system dynamics and to facillitate the exchange of ideas between these two disciplines of classical physics.
The talk will present experiences with parallelizing and running a 3D gyrokinetic flux tube code on the Origin 2000 at ACL, LANL. The parallellization and performance of the code will be discussed. As introduction, a brief overview of parallel computing will be presented. This will cover some terminology, basic architectures of massively parallel machines, parallel schemes and philosophies, and some MPI.
I will discuss the derivation and results for a one-dimensional gyrofluid model of Alfven waves in the Magnetosphere. This model is derived from the drift-kinetic equation and it accounts for the effect of a dipole field beta variation. Interesting results include a realistic wave period, and significant electric field fluctuations at the ionosphere-magnetosphere boundary which could contribute to acceleration of auroral electrons.
SOHO has been the flagship solar observatory of the European Space Agency and NASA since it began service in April, 1996. Its remarkable discoveries and nearly flawless performance for two years induced agencies to grant an extended mission beginning after its first two years. In late June, 1998 within the span of a few hours, the one billion dollar mission started spinning out of control and contact was lost for weeks. It has since been found, communications are restablished, and a painstaking "healing" process is underway that may restore much of the observatory's function.
Substorms are always observed during the main phase of magnetic storms. In the expansion phase of each substorm ions and electrons suddenly appear at synchronous orbit, accelerated by the collapse of the tail-like field. This correlation has led many researchers to believe that storms are a consequence of the ring current produced by the drift of these particles. However, studies of the Dst index which measures the strength of the ring current show that injection occurs before the expansion phase, and that Dst actually decreases in strength at expansion onset. What then is the role of substorms in producing magnetic storms?