Table of contents

A major problem in astrophysical plasma physics is to explain how the outer atmosphere, the corona, of the Sun is heated to temperatures of millions of degrees Kelvin. It is accepted that the heating mechanism is magnetic, with the energy source being turbulent motions below the solar surface. Two classes of theory are proposed, according to the timescale of the driving motions in relation to the Alfven timescale of the coronal plasma. Fast motions generate MHD waves which propagate up into the corona carrying energy and can heat the corona if the waves are damped. Slow motions move the footpoints of the coronal field, generating field-aligned currents which may dissipate to provide heat. In each case, the main difficulty is in finding an adequate means of dissipation in the highly-conducting coronal plasma. Some proposed heating mechanisms are outlined, which present a number of interesting plasma physics problems closely related to those arising for fusion plasma; in particular, Alfven wave propagation and absorption in a nonuniform medium, and anomalous heating by reconnection, turbulence and relaxation.

An average invariant which describes average magnetic surfaces for a system without symmetry is obtained. The system is a tokamak toroidal equilibrium perturbed by resonant helical windings. The magnetic field is a superposition of the magnetic fields of the equilibrium and the helical windings, and the corresponding vector potential is determined. An average vector potential is defined to obtain the average invariant. Analysis of the average surfaces showed that the magnetic islands move towards the plasma centre and decrease in width as the pressure increases.

Ice pellet injection experiments were carried out in the JIPP T-IIU tokamak in order to study thermal (cooling) transport just after injection. The cut-off problem of ECE signals due to the rise in density has been resolved by careful measurements of the temperature profile at a high time resolution during its decay phase. The phenomenon of ultra-fast cooling (so-called precooling) has been identified using the two different methods of ECE and soft X-ray measurements. In the outer region of the plasma the cooling propagation velocity is comparable to or slightly greater than the pellet velocity, while in the central region the propagation velocity is significantly greater than the pellet velocity. Ice pellets were injected into various kinds of JIPP T-IIU plasmas, the current and sawtooth phase of which had different values, including a no-sawtooth plasma. The existence of the q=1 surface and arrival of a pellet near the q=1 surface have turned out to be necessary conditions for precooling, and even just after the sawtooth crash the precooling starts around the q=1 surface, not at the plasma center. Simultaneous measurements of electron temperature and density profiles indicate that the central temperature always decreases before the central density increases. Some anomalous transport might be induced by pellet injection at the central region.

Angular momentum transport studies were conducted for nearly stationary situations of about 50 neutral-beam-heated ASDEX discharges under a variety of experimental conditions. Comparison of the confinement times of angular momentum and energy reveals close similarities between thermal and perpendicular momentum transport. Scaling laws are derived for the dependence of the central rotation speed, the momentum confinement time and the radially-averaged momentum diffusivity from the main experimental parameters of L-mode discharges. A well-developed isotope effect, significant power degradation and a favourable current scaling are found to be characteristic features of momentum confinement. The L-mode results are compared with the momentum transport behaviour during the improved confinement phase of H-mode, pellet-fuelled and counter-NI discharges. For the H-mode, the confinement improvement is comparable for momentum and energy. The peaked density profile scenarios exhibit a considerably more pronounced increase in momentum confinement than in energy confinement.

A model is developed to calculate the impurity influx caused by induced radio-frequency (RF) sheaths near ion-cyclotron range of frequencies (ICRF) antennas, which accelerate ions to energies sufficient to sputter the Faraday screen and other adjacent material surfaces. The existence of several types of RF sheaths is discussed, and detailed models of Faraday-screen-gap and front-face sheaths are described which are relevant to the antenna geometry in the Joint European Torus (JET) experiment. Calculations of the local influx of metal impurities from the Faraday screen show good qualitative agreement with the observed dependence of the impurity influx in JET on phasing, magnetic field angle, and screen coatings, and the model provides a useful framework for understanding other aspects of the data. It is also shown that a great increase in the metal influx due to self-sputtering of the screen ('avalanche') can occur under certain conditions. However, the model calculations also demonstrate that proper antenna design can virtually eliminate the impurity influx from the screens.

Parametric coupling of lower-hybrid pump waves with purely-damped convective cell modes generates enhanced electrostatic fluctuations in a magnetized plasma. The threshold electric field for exciting a purely-growing instability is found. In order to consider the saturation of the parametric instability, the authors have invoked a nonlinear mechanism which involves scattering of charged particles by suprathermal fluctuations. An expression for the effective collision frequency is also obtained, which characterizes the HF power dissipation in the plasma as well as the heating. Attempts have been made to correlate the results with some experimental data.

The neutral Monte Carlo code NIMBUS was used in conjunction with measured plasma temperature and density profiles for JET ohmically-heated discharges to analyze various aspects of hydrogenic edge re-cycle fueling at carbon limiters.

It is necessary to operate a divertor with a sufficiently high plasma density at the symmetry point n_s, i.e. adjacent to the main plasma, to satisfy the Lawson criterion and, at the same time, a low enough value of the plasma temperature at the target plate T_t to minimize sputtering. The possibilities for realizing these two objectives simultaneously are examined. It is found that the combination of high n_s and low T_t can require that the plasma flow be supersonic before reaching the divertor sheath. It is not known if such plasma conditions are stable.

The turbulent diffusion coefficient is calculated using two different approaches, one involving test particles and the other involving the dynamically-passive convected field. The particle approach is Lagrangian in character. The version the author offers is a refinement of the classical version employing new types of independence hypothesis, Gaussianity-implying hypothesis and diffusion hypothesis. The field approach is Eulerian in character. The version offered is a slight modification of the classical version involving transient effects. When the transient effects are no longer important both approaches give the same results, and rely on the same set of restrictions. The fluctuation amplitudes and Lagrangian correlation times that are not necessarily very small are considered. The problems associated with the zero micro-diffusivity limit are addressed.

Power and particle balance studies on the Advanced Toroidal Facility (ATF) torsatron are carried out using a rail limiter system. Both top and bottom limiters are made of graphite tile arrays, and these tiles are instrumented with thermocouples and Langmuir probes for calorimetric and particle flux measurements. Initial experimental results indicate that the limiter power loss accounts for about 12% of the total and the radiation loss for about 30% of the total; the rest of the plasma heating power appears to be going to the vessel wall. The particle flux to the limiters is also about 18%. The fractions of power and particle flux to the limiters are relatively lower than in Tokamaks because of the low edge safety factor, q approximately 1 rather than q approximately 3 as in a typical Tokamak, at the natural boundary of the ATF plasma (which results from the magnetic stellarator configuration of this currentless device). Therefore, for limiters of the same size, these fractions are about a factor of q lower in ATF than in a comparable Tokamak device.