Results from the tandem mirror experiment are described. The configuration of axial density and potential profiles are created and sustained by neutral-beam injection and gas fueling. Plasma confinement in the center cell is shown to be improved by the end plugs by as much as a factor of 9. The electron temperature is higher than that achieved in our earlier 2XIIB single-cell mirror experiment.PACS numbers: 52.55. Mg, 52.55.Ke This Letter reports the first results obtained from the tandem mirror experiment (TMX) at the Lawrence Livermore Laboratory. Steady-state tandem-mirror plasmas have been produced and an electrostatic barrier that improves plasma confinement has been observed. The tandem-mirror configuration 1 ' 2 can enhance the performance of a magnetic-mirror thermonuclear reactor. Such a reactor would produce power in a cylindrical, high-/3, magnetic solenoid. End losses from this center cell are reduced by electrostatic endplug barriers of positive potential, which turn back those low-energy ions which escape through the magnetic mirror. These potential barriers are established on both ends of the center cell by high-density, high-temperature, mirror-confined plasmas, which have a larger ambipolar potential than does the center-cell plasma.Earlier tandem-mirror experiments, 3 in which plasma guns were used to establish end-plug densities larger than those in the center cell, have produced potential wells. Langmuir-probe measurements indicated that the magnitude and scaling of the potential-well depth is consistent with theoretical predictions. Our results demonstrate that we can produce and sustain a tandem-mirror plasma configuration by use of neutral beams to fuel the end plugs and gas to fuel the center cell. This method can be extrapolated to continuously operated systems. Our experiments further demonCee coil Baseball coilSolenoid coils Octupole coil -Plasma flux tube 1132 Neutral beam injectors Startup plasma guns FIG. 1. Schematic diagram of TMX magnet and neutral-beam system.

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.

Data on field-reversal experiments in the neutral-beam-injected 2XIIB mirror machine are reported. The best result is an estimated field-reversal parameter ζ = ΔB/Bvac = 0.9 ± 0.2 with vacuum field strength Bvac = 4.35 kG. Experiments at higher field strength Bvac = 6.7 kG achieved ζ = 0.6 ± 0.1. Ion energy confinement ⟨nτEi⟩ for the Bvac = 6.7 kG experiment is less than that predicted by classical Spitzer electron drag. Ion-cyclotron oscillations increasing with injected neutral-beam current suggest that ion-cyclotron losses are present and that ΔB/Bvac could be increased by improving stabilization of the ion-cyclotron oscillations.

We have observed and confirmed significant ion heating by the electrostatic ion-cyclotron instability in a barium plasma. We present spectroscopic evidence showing that this mechanism drastically alters the velocity distribution, demonstrating at least a highly nonlinear process and giving strong evidence for randomization of the particle motion. Experimental corroboration of a theory of Drummond and Rosenbluth is presentedWe report the direct observation of ion heating in a fully ionized barium plasma, as indicated by an order-of-magnitude increase in ion temperature (from 2500°Kto greater than 25 000°K) due to the current-driven electrostatic ion-cyclotron instability. 1 The instability was generated by a technique first used by Motley and D'Angelo in the Princeton University Plasma Physics Laboratory Q device, 2 the theory for which was subsequently developed by Drummond and Rosenbluth. 1 The experiment was performed in the Q machine at the University of California at Irvine. Perpendicular temperature measurements were made by spectroscopically observing the Doppler line broadening of light scattered by singly ionized barium after the techniques developed by Hinnov et al 3 The significance of these measurements is that this is the first demonstration of ion heating by this current-driven instability. 4 Historically, this series of measurements was undertaken in an attempt to explain the previous results of Hinnov et al. 3 They measured and reported anomalously high ion temperatures in a barium plasma but the heating mechanism was not understood. We repeated the earlier experiments and recovered some of their results. However, we noticed that high temperatures occured when the rhenium coating on the hot tungsten end plates, used to ionize the barium, 5 was uneven so that patches of tungsten appeared through the rhenium coating. As a consequence small areas of different work function were formed on the plates and we speculated that the localized emfs thus formed produced axial filaments of current (easily visible since the barium ions emit visible light). To test this hypothesis, and to identify the heating mechanism (which we were unable to do initially) we went to a single-ended Q machine with insulated buttons, as indicated schematically in Fig. 1(a). In this configuration the electrostatic ion-cyclotron instability was observed, easily identified from its spectrum, from its dependence on the magnetic field, and by recovering the results of Motley and D'Angelo, 2 most notably that the observed frequency is slightly higher than the ion-cyclotron frequency (typical values at onset: / ci = 5 kHz, / = 61 kHz, linewidth approximately 1 kHz). Several configurations were tried, varying from a single button to an BUTTON PROBE BARIUM OVEN TO FABRY-PEROT •5 FIG. 1. (a) Schematic of the experimental arrangement. The length L was 123 cm; the diameter, 5 cm, button diameter, 0.6 cm. The Fabry-Perot looked into a light dump also, (b) Typical low-temperature (T f ) Doppler broadening scan. a 0 is the 4934 A line of Ba n, ...

The temporal evolution of the current-driven electrostatic ion cyclotron instability was investigated experimentally. The critical destabilizing electron drift velocity for different values of mode phase velocity was measured. The phase velocity was changed by varying the effective plasma column length and hence the parallel wavelength. Ion cyclotron damping was observed to dominate over electron Landau damping at low phase velocities. The temporal growth rate was experimentally determined for several values of electron drift and found to agree very well with linear theory. The observed saturated mode amplitudes compare favorably with the saturation levels predicted by a theory of nonlinear stabilization due to wave induced collisions.