Table of contents

A novel gas target concept for recovering both energy and particles from a high-energy plasma stream is presented. This concept includes the maintenance of a pressure discontinuity by a normal shock and a very high mass flow rate in a relatively small system. The pressure discontinuity allows the exhaust plasma stream to minimize backflow into the plasma, by interacting with the target in a low-pressure region; the high mass flow rate allows exit temperatures that are reasonable from a materials viewpoint and suitable for energy recovery.

A method for determining plasma shape and global properties of high-beta non-circular tokamak equilibria (β_p, β, q and ℓ_i) from magnetic sensor data is described. The technique uses a least-squares fitting procedure to find the plasma boundary and a global force balance analysis (as opposed to a complete solution of the MHD equilibrium equation) to determine plasma pressure. Estimates of the uncertainties in the computed quantities are also obtained, and the method is fast enough that it can be used to provide a complete time history (up to 100 time points) of ISX-B high-beta discharges within two to three minutes of each shot. Values obtained using this method are in excellent agreement with more detailed measurements and with free-boundary MHD equilibrium computations.

The kinetic theory of high-n electromagnetic low-frequency modes in a finite-β toroidal plasma is investigated in the collisionless limit. The high-n electromagnetic ballooning mode is identified and compared with the MHD ballooning mode. This mode is found to be always unstable, with the growth rate of the order of the drift frequency ω_*. As the β-value increases and approaches the critical β predicted by the MHD theory, the growth rate becomes large and a real-frequency downshift appears. The toroidal shift of the magnetic surface and the magnetic well have a stabilizing effect, but the mode is unstable, because of the wave-particle interactions taking place even in the 'second stability region' of the MHD theory. The growth rate normalized to ω_* has a peak in the MHD-unstable parameter region, but the peak is low. In the zero-β limit, this mode turns out to be the electrostatic ballooning mode. The drift branch and the drift-Alfvén branch are also identified in a toroidal plasma but remain stable, because of magnetic shear.

Before a complete set of impurity control techniques were implemented in the Doublet III tokamak, two distinct discharge types were obtained, with nearly equal probability at high power densities with the same external control parameters. In the normal discharge (type S), sawtooth activity develops quickly and no impurity peaking on axis occurs thereafter. The other type of discharge (type O) initiates with more plasma/wall interaction, proceeds through the current-ramp phase with no central MHD activity, and impurity density profiles are found to be peaked on axis. Later in the current flat-top phase, type-O discharges experience the sudden onset of a large oscillation with m = 1, n = 1, and a concurrent decrease of central impurity densities. The (sinusoidal) oscillation eventually evolves into sawteeth, and type-O discharges attain a final state identical to type-S discharges. Measurements of absolute impurity densities show a direct correlation of impurity confinement with the presence or absence of tearing-mode activity, which suggests that the impurity transport is modified by the influence of the m = 1 magnetic island.

The usual ripple transport calculations lead to a ν⁻¹ scaling of the transport coefficients with collision frequency ν. The paper extends and clarifies this scaling by taking into account the fact that the dominant contributions to transport come from particles in the high-energy tail of the distribution function. This means that particles with energy E ∼(4–6)T dominate, and this restricts the ν⁻¹ scaling range to ν(T) > (E/T)^5/2 δω_d(T) ∼ 100 δω_d(T), where T is the thermal energy, δ is the ripple-well depth, and is the time to drift over the region where the localized ripple well exists. In addition, transport coefficients are derived in weak-ripple (α ≡ /N_qδ ≫ 1, with the inverse aspect ratio, N the number of toroidal coils, q the safety factor), low-collisionality regimes where the radial step size is determined by the distance a particle drifts before it 'collisionlessly detraps' from a ripple well. The maximum transport rate is found to be about two orders of magnitude smaller than usually assumed, primarily because of the restriction of the ν⁻¹ scaling regime to the much higher collision frequencies.

Instabilities are governed by transport-determined plasma profiles and transport is in turn affected by instabilities. A one-dimensional code has been developed which models this relationship. The dependence of magnetic island sizes on total current is described. Above a critical current, there is a non-linea interaction between the m = 1 and m = 2 modes which takes place through their effect on the current profile. This leads to a disruptive loss of confinement. Inclusion of the effect of impurity radiation leads to repeated soft disruptions at a lower current.

In specific cases, the critical beta (β_c) for stability against internal magnetohydrodynamic (MHD) ballooning modes can be improved significantly by taking into account the stabilizing influence of the ion diamagnetic drifts. This kinetic modification to the ideal MHD analysis and radial corrections to the local theory are included in a calculation of β_c as a function of toroidal mode number (n) for a particular tokamak equilibrium sequence of interest.

Tomography based on poloidal harmonic expansion is modified to be applied to an elliptical source in order to investigate the internal mode structures due to tearing modes in the ℓ = 2 stellarator configuration. This method was applied to the soft X-ray signals of the JIPP T-II stellarator and the image of the m = 2 magnetic island of the plasma was obtained.

Experimental studies of electron cyclotron heating in a stellarator plasma have been carried out by injecting 40 kW of 35.5 GHz microwave power. Electron cyclotron heating with the ordinary and the extraordinary modes injected from the low-field side show almost the same heating efficiency of 2.2 × 10¹³ eV·cm⁻³·kW⁻¹ at the average electron density of 6 × 10¹² cm⁻³. Efficient heating by the extraordinary mode in the presence of the cyclotron cut-off is interpreted to be achieved by the penetration of scattered waves from high- and low-field sides, polarization being disturbed by microwave reflection from the vacuum chamber wall.

Increased fluxes of runaway electrons at the PLT limiter are observed in the few milliseconds following internal disruptions. These fluxes have an inverted (outside) sawtooth character. The time for the flux to reach a maximum after the disruptions has been studied as a function of the plasma parameters for thousands of PLT discharges. One interpretation is that this delay represents the time for a perturbation to the runaway electron population to travel from the q = 1 region to the plasma boundary. These times are ∼ 10⁻¹ of the electron thermal confinement times and increase with the plasma electron density.

Non-circular tokamak plasmas can be unstable to ideal-MHD axisymmetric instabilities. Passive conductors with finite resistivity will at best slow down these instabilities to the resistive (L/R) time of the conductors. It is demonstrated that a simple active feedback system far from the plasma which responds on this resistive time can stabilize the system, provided its mutual inductance with the passive coils is small enough.

Electron temperature and density measurements by Thomson scattering were performed for the first time for the whole duration of a tokamak discharge. A Nd:YAG laser of 60 pulses per second at 1.06 μm was used in ASDEX in combination with silicon avalanche photodiode detectors. Density calibration was done by rotational anti-Stokes Raman scattering from hydrogen. The system is used for measurements at electron densities as low as 3 × 10¹² cm⁻³.

A 16-kV, 4-A hydrogen beam is co-injected into hydrogen plasmas and coupled to an RF of 800 MHz, up to 150 kW. It is observed, from the measurement of parallel charge-exchange neutrals, that fast ions above 16 keV are enhanced by interaction with lower hybrid waves. This enhancement is critically reduced when the line-averaged electron density is higher than 1.6 × 10¹³cm⁻³, in the toroidal fields of both 22 and 16 kG. For the 22-kG case, the cause of the reduction may be related to the location of the turning point of the incident waves, i.e. the turning point, in higher-density regions, is located outside the region where the fast ions produced by NBI are relatively abundant. For the 16-kG case, however, this interpretation is only valid when a portion of the incident wave energy is transferred to waves of shorter wavelength. The ion temperature increase for simultaneous heating agrees with the sum of those in the individual heatings even when the beam-RF interaction is rather strong.

Heat deposition on the first wall due to ripple-trapped loss particles produced in the slowing-down process of injected beam ions is investigated in a reactor-grade tokamak. Ripple-trapped particles enter specific narrow areas of the first wall in the course of their toroidal drift and cause a significantly large heat loading, which will become one of the important problems in the design of the tokamak first wall. Localized heat loads due to ripple-trapped loss ions may reach a magnitude of 1 MW·m⁻² for the next-generation tokamak.

During the past decade, progress in the understanding and control of plasma surface interactions has led to major advances in the production of better confined, higher-density, cleaner, and hotter plasmas. Examples of these achievements include effective discharge cleaning, gettering, low-Z limiters, and beam duct and RF antenna and waveguide preparation. Continued interest in improving both the operation of present plasma devices and the design of future reactors brought together 198 scientists from 15 countries at the Fifth International Conference on Plasma Surface Interactions in Controlled Fusion Devices. The Conference was sponsored by Oak Ridge National Laboratory and the US Department of Energy.