Energetic wave-trapped electrons of ϳ20 eV, moving at the 13.56 MHz helicon wave phase velocity, are directly measured with a 3 ns response-time retarding-potential analyzer. The rf axial wavelength is measured with B-dot probes scanned axially. Changes in the axial magnetic field or the rf power cause order of magnitude changes in the energetic electron current. The current correlates with either a transition from partially propagating to nearly pure standing waves or quantized jumps in the number of half wavelengths between the two azimuthal antenna straps. [S0031-9007(97)03548-5] PACS numbers: 52.50. Dg, 52.40.Db, 52.70.Ds, 52.70.Nc Pulses of argon line radiation have been reported [1] at the heating frequency in a helicon plasma source: the pulses originated in the near field of the antenna and propagated at the helicon wave phase velocity, consistent with trapping and accelerating pulses of electrons to 30 eV in the radio-frequency (rf) fields under the antenna. We report on strong variations in the current of energetic, wavetrapped electron pulses with changes in the magnitude of the axial magnetic field or rf power. The electrons are directly measured with a retarding-potential analyzer, having sufficient time response to resolve 15 ± of phase angle of the 13.56 MHz rf wave. We have previously shown that the energetic electrons are wave-trapped within a 75 ± FWHM phase angle each rf cycle [2]. This supports the picture that electrons "surf" on the helicon wave [1]; although, for our parameters, a given electron is expected to remain trapped for only a mean-free path of ϳ4 cm (limited by electron-neutral collisions) which is less than the helicon wavelength of ϳ20 cm. Orbit code calculations are used to assess this effect.While Langmuir probes are a more common method of measuring the electron energy distribution function, the proper technique in rf fields continues to be controversial [3][4][5]. Retarding-potential gridded-energy analyzers (GEA) avoid these uncertainties; however, only electrons with axial energy exceeding the sheath potential drop can be detected. Sheath potential oscillations at the heating frequency are small. The primary diagnostic for this paper is a 1.9 cm diameter GEA [2,6] with grid voltages swept to measure electron and ion distribution functions. The distance between the last grounded entrance grid and the collector is 0.20 cm, providing 1 ns time response for 20 eV electrons, limited to 3 ns (corresponding to 15 ± of phase) by the 150 MHz analog bandwidth recorder. For 5 eV argon ions, the response time of 400 ns averages the ion distribution over 6 wave periods, so plasma potential measurements are not phase resolved. The GEA and other probes are downstream of the antenna region where hot electrons are produced, so should not perturb these data.Helicon sources are attractive for plasma processing and a number of other applications because they provide high plasma density in low magnetic fields, with a high ion-ization efficiency, and a low, ϳ3 eV, electron temperature [7-9]....

Gas desorption and electron emission coefficients were measured for 1 MeV potassium ions incident on stainless steel at grazing angles (between 80 and 88 from normal incidence) using a new gaselectron source diagnostic (GESD). Issues addressed in design and commissioning of the GESD include effects from backscattering of ions at the surface, space-charge limited emission current, and reproducibility of desorption measurements. We find that electron emission coefficients e scale as 1= cos up to angles of 86 , where e 90. Nearer grazing incidence, e is reduced below the 1= cos scaling by nuclear scattering of ions through large angles, reaching e 135 at 88. Electrons were emitted with a measured temperature of 30 eV. Gas desorption coefficients 0 were much larger, of order 0 10 4. They also varied with angle, but much more slowly than 1= cos. From this we conclude that the desorption was not entirely from adsorbed layers of gas on the surface. Two mitigation techniques were investigated: rough surfaces reduced electron emission by a factor of 10 and gas desorption by a factor of 2; a mild bake to 220 had no effect on electron emission, but decreased gas desorption by 15% near grazing incidence. We propose that gas desorption is due to electronic sputtering.

Making axisymmetric mirrors magnetohydrodynamically (MHD) stable opens up exciting opportunities for using mirror devices as neutron sources, fusion-fission hybrids, and pure-fusion reactors. This is also of interest from a general physics standpoint (as it seemingly contradicts well-established criteria of curvature-driven instabilities). The axial symmetry allows for much simpler and more reliable designs of mirror-based fusion facilities than the well-known quadrupole mirror configurations. In this tutorial, after a summary of classical results, several techniques for achieving MHD stabilization of the axisymmetric mirrors are considered, in particular: (1) employing the favorable field-line curvature in the end tanks; (2) using the line-tying effect; (3) controlling the radial potential distribution; (4) imposing a divertor configuration on the solenoidal magnetic field; and (5) affecting the plasma dynamics by the ponderomotive force. Some illuminative theoretical approaches for understanding axisymmetric mirror stability are described. The applicability of the various stabilization techniques to axisymmetric mirrors as neutron sources, hybrids, and pure-fusion reactors are discussed; and the constraints on the plasma parameters are formulated.

We report results of plasma confinement experiments with an auxiliary warm-plasma component flowing along magnetic field lines to suppress ion-cyclotron instabilities. The reduced plasma losses, with the lower fluctuation amplitude, permits neutral-beam buildup of a 13-keV deuterium plasma to densities as high as 4 x 10 13 cm" 3 corresponding to peak beta values of 0.4. Variation of the beam energy demonstrates that longer confinement times are achieved at higher ion energies.

A tandem-mirror plasma has been sustained and heated by rf alone without the need for neutral beams. End plug density and energy are maintained by ion cyclotron-resonance heating which traps and heats a fraction of the central-cell loss stream. The central-cell plasma is maintained by gas fueling and rf heating. Magnetohydrodynamic stability limits the ratio of the central-cell to plug plasma pressure, and the central-cell electron temperature must be kept high enough for ionization. A quasi steady state is achieved that lasts much longer than the decay times of the plugs and central cell.PACS numbers: 52.55. Mg, 52.50.Gj, 52.55.Ke The tandem-mirror approach to the development of a fusion reactor promises to have many advantages over simple mirrors as well as other magnetic confinement devices. 1 In a tandem mirror, a central-cell plasma in a solenoid is bounded by "plug" plasmas in mirror cells. The TMX experiment 2 demonstrated that electrostatic confinement of central-cell ions by the plug plasmas significantly reduced end losses. Calculations indicate that the overall Q (power gain) of a tandem-mirror fusion reactor can be quite high (~5-10) if the volume of the central cell greatly exceeds that of the plugs. 3 In previous experiments, plug plasmas were fueled and heated by energetic neutral beams. However, technological constraints on neutral-beam heating when scaled to a reactor could be eased or eliminated by supplementing or replacing neutral-beam power with rf power. We show in the Phaedrus experiment that a tandem-mirror plasma can be sustained by rf alone, without the application of neutral beams.The operation of a tandem mirror can be thought of as taking place in two stages, a transient stage followed by a quasi-state stage. In the transient stage, plasma is injected into the machine by stream guns. In fact, earlier experiments with Phaedrus 4 and Gamma 6 5 were restricted to the transient stage. Such plasmas side-stepped many of the issues of tandem-mirror physics because plasma characteristics were dominated by the external plasma between the plugs and the stream guns. The high-density external plasma line-tied the plasma to the stream guns, significantly reducing both magnetohydrodynamic (MHD) instabilities and microinstabilities and fixing the electron temperature. When the stream guns were turned off in Phaedrus, the plasma decayed away in approximately 150 jus in the plugs and 400 JUS in the central cell. Quasi-steady-state operation after stream guns are turned off requires sources of particles and energy for both the plug and central-cell plasmas and also adds constraints that must be satisfied in order to maintain microstability and MHD stability. In TMX, fueling and heating were provided by neutral beams in the plugs and gas puffing in the central cell. In the Phaedrus experiment, fueling is provided only by gas puffing in the central cell while rf provides heating. In both experiments, stability was provided by high plug energy density and the presence of the central-cell loss flow.The ...

In thermal-barrier experiments in the tandem mirror experiment upgrade, axial confinement times of 50 to 100 ms have been achieved. During enhanced confinement we measured the thermal-barrier potential profile using a neutral-particle-beam probe. The experimental data agree qualitatively and quantitatively with the theory of thermal-barrier formation in a tandem mirror.

A neutral-beam-sustained plasma in the 2XIIB mirror machine is stabilized by hydrogen gas ionized at one mirror. _Peak plasma energy densities that exceed the energy den-sity_of the vacuum field (P=Sft D E r> /B VSiC 2 ^1) are achieved with deuterons of average energy £ D « 9 keV, peak density 6 D^1 .5xl0 14 cm" 3 , in a magnetic field B vac =0.56 T. These high /3's are sustained for the duration of gas and neutral-beam injection, with no evidence of a £ limit.We report here recent substantial increases in the plasma j3 in the 2XIIB mirror machine, due to stabilization of the neutral-beam-sustained plasma by gas injected in a mirror throat. /3 is defined by p = &nn D E D /B vac 2 , where n D is the central deuterium energy, andB vac is the applied central magnetic field. Previous experiments 1 in the 2XIIB device employed deuterium-loaded titanium-washer guns as sources of cold streaming plasma to suppress ion cyclotron fluctuations caused by the loss-cone nature of the confined ion distribution. 2 A beam-injected plasma j3 =0.4 was achieved with a single gun, limited by periodic bursts of ion cyclotron fluctuations rather than by the available neutral-beam current. 3 Increasing the amount of stabilizing plasma stream with three guns raised j3 to 0.6; again $ was limited by rf bursts. The amount of cold plasma has recently been increased still further by injecting hydrogen gas in a mirror throat. This has resulted in central hot-plasma /3 's exceeding unity, 0>1. This result has significance for mirror fusion reactors, since higher /3's will allow higher fusion power densities for a given field, and may possibly lead to a field-reversed configuration that would enhance the fusion energy balance. 4 The gas-feed system is shown in Fig. 1. A ceramic box is located just beyond one mirror throat of the minimum-B field, with its elliptical apertures conforming to the flux tube passing through a 12-cm-diam circle at the central midplane. Hydrogen or deuterium gas is injected into the box above and below the plasma fan by four pulsed gas valves. The gas is injected with a rise time of ^1 ms, and injection continues for the duration of beam injection (^5 ms) at a rate controlled by orifices at the valve openings inside the box. To start ionization of injected gas, a hot plasma of ft D~2 xl0 13 cm" 3 is initially created in a steady mirror field by beam injection into a streaming plasma target provided by guns located beyond the opposite mirror. 3 Gas neutrals are ionized in the box by electrons (T e »60 eV) coducted along the field from the beam-injected central plasma. The cold ions formed are presumably driven out of the box at the sound speed [v s = (T e /m i ) 1/2 ] by ambipolar electric fields. With sufficient density of cold plasma at the mirror, a small fraction can penetrate the hot central plasma to provide the stabilizing stream after the plasma guns are turned off. 5 During neutral-beam and gas injection, measurements with a microwave interferometer and a Langmuir probe indicate that a dense, cold plasma (n...

Contaminating clouds of electrons are a concern for most accelerators of postive-charged particles, but there are some unique aspects of heavy-ion accelerators for fusion and high-energy density physics which make modeling such clouds especially challenging. In particular, self-consistent electron and ion simulation is required, including a particle advance scheme which can follow electrons in regions where electrons are strongly-, wealdy-, and un-magnetized. We describe o w approach to such self-consistency, and in particular a scheme for interpolating between full-orbit (Boris) and dr&kinetic particle pushes that enables electron time steps long compared to the typical gyro period in the magnets. We present tests and applications: simulation of electron clouds produced by t h e e Werent lcinds of sowces indicates the sensitivity of the cloud shape to the nature of the source; first-of-a-kind self-consistent simulation of electron-cloud experiments on the High-Current Experiment (HCX)