Table of contents

Externally applied magnetic fields are used on the Texas Experimental Tokamak (TEXT) to study the possibility of controlling the particle, impurity and heat fluxes at the plasma edge. Fields with toroidal mode number n = 2 or 3 and multiple poloidal mode numbers m (dominantly m = 7) are used, with a poloidally and toroidally averaged ratio of radial to toroidal field components 〈|b_r/B_ø〉 ≅0. 1%. Calculations show that it is possible to produce mixed islands and stochastic regions at the plasma edge (r/a ≥ 0.8) without affecting the interior. The expected magnetic field structure is described and experimental evidence of the existence of this structure is presented. The edge electron temperature decreases with increasing 〈|b_r/B_ø〉, while interior values are not significantly affected. The implied increase in edge electron thermal diffusivity is compared with theoretical expectations and is shown to agree with applicable theories to within a factor of three.

Neutral beam heating data from JET have been analysed in detail to determine what proportion of the current is driven non-inductively. It is found that in low density limiter discharges, currents of the order of 0.5 MA are driven, while in H-mode plasmas currents of the order of 0.7 MA are measured. These measured currents are found to be in reasonable agreement with theoretical predictions based on neoclassical models. In low density plasmas the beam driven current is large while the neoclassical bootstrap current dominates H-mode plasmas.

Impurities in JT-60 were studied by visible and VUV spectroscopy over two periods of operation. The two periods were distinguished by the use of different first wall materials: from April 1985 to March 1987, TiC coated molybdenum was used as limiter material; from June to October 1987, graphite was used. Quantitative spectroscopic measurements of Z_eff, impurity concentrations and radiated power losses were made for ohmically heated and neutral beam heated discharges with limiter and divertor configurations. In the first phase with metallic first wall material, oxygen, carbon and titanium were identified as the main plasma impurities. In neutral beam heated, diverted discharges, Z_eff was 1.6 at _e = 4 × 10¹⁹ m⁻³. The concentrations of oxygen, carbon and titanium were 1%, 0.1% and 0.006% of n_e, respectively. In the second phase with graphite material, the metallic impurities were reduced, and the contribution of metallic impurities to the radiated power loss was less than 1%. However, Z_eff increased up to 3 in neutral beam heated discharges. In limited plasmas, the concentrations of oxygen and carbon were 1% and 5%, respectively, at _e = 4 × 10¹⁹ m⁻³; in diverted plasmas, these concentrations were 2% and 0.4% at the same _e. The radiated power loss from the main plasma was 20–40% of the input power in neutral beam heated, limited discharges, and 7–25% in diverted discharges. The contributions of oxygen and carbon to the radiated power in limited discharges were comparable, and in diverted discharges the contribution of oxygen was dominant.

In general, Faraday screen elements in an ICRF antenna are not aligned precisely along the combined toroidal and poloidal magnetic fields. When plasma of density n > 2_∊0V/eg² ∼ 10⁹cm⁻³ (V being the voltage across the gap and g the gap spacing) is present in the gap between the elements, the electron response to the parallel electric field shorts out the electric field over most of the gap, leaving a narrow sheath of positive space charge and an intense electric field. This intense electric field accelerates ions up to an appreciable fraction of the gap voltage (∼1 kV), sufficient to cause physical sputtering of the screen material. Impurities so generated constitute the principal limitationon power density for ICRF antennas. Principles of ICRF antenna and Faraday screen design which minimize sputtering are discussed.

In the JET tokamak, ICRF driven fusion reactivity has been determined using measurements of 16.6 MeV γ-ray emission from d[³He, γ]⁵Li reactions during central RF heating in the (³He)d minority regime. Up to 1 MJ of fast minority ions in the plasma has been produced with the application of up to 15 MW of RF power. The maximum rate produced by d[³He, p]⁴He fusion reactions has been estimated as 2 × 10^l6 s⁻¹ (equivalent to 60 kW of fusion power in charged particle products). The reactivity increased strongly with coupled RF heating power (proportional to (P_RF)^5/3), with some evidence of a weakening of the dependence leading to a saturation in the energy gain Q at the highest coupled RF power levels (P_RF > 8−12 MW). The experimentally measured anisotropic fast ion energies and fusion reaction rates have been simulated using a radially dependent Stix model for a wide variety of discharges. Analysis of the radial profile of fusion reactivity shows that when the RF power density is maximized on the magnetic axis of the discharge, the fusion reactivity is peaked away from the axis. This effect is caused by the minority ions near the centre of the discharge being driven to energies beyond the maximum in the fusion cross-section.

Neoclassical bootstrap currents in the banana and plateau regimes are evaluated for a large helical system with L = 2 and with a major radius of 5 m. Various vacuum magnetic field configurations are studied with a view to optimizing the bootstrap current. In the banana regime, shifting of the magnetic axis and shaping of magnetic surfaces have a remarkable influence on the bootstrap current. It has been found that a small outward shift of the magnetic axis and vertically elongated magnetic surfaces are favourable for a reduction of the bootstrap current. In contrast, in the plateau regime the bootstrap current depends strongly on the toroidal pitch number of the helical winding coils, but shifting of the axis and shaping of magnetic surfaces are not effective. The magnitude of the bootstrap current is estimated for plasmas in a large helical system with parameters given by a one-dimensional radial transport code.

The marginal pressure profile that is imposed either by ideal ballooning modes or by interchange modes on every nested flux surface of a three-dimensional equilibrium modelling the ATF device is determined for conditions of vanishing net toroidal plasma current. The corresponding limiting values are β = 3.75% and β(0) = 9.89%. Mercier modes dominate in regions of strong stellarator-like magnetic shear (ι'/ι > 0), and non-symmetric ballooning modes (non-symmetric eigenstructures in the extended poloidal angle domain) dominate in regions with strong tokamaklike magnetic shear (ι'/ι < 0).

The results of extensive investigations of the low-n ideal magnetohydrodynamic stability of the TCV, NET and DIII-D tokamaks are compared in order to identify the general qualitative effects of cross-sectional shaping on the low-n stability of elongated tokamaks. The results consistently indicate a breakdown of the simple stability picture for circular and moderately elongated cross-sections when the cross-section deviates strongly from circular. Simplified extrapolations of the standard stability picture to high elongation and triangularity and to low aspect ratio are unreliable; at high elongation, the scaling of both the current limit and the beta limit is strongly dependent on the cross-section shape. Several implications for next generation tokamaks can be drawn from the conclusions, and these are discussed briefly.

In JET, both high density and low-q operation are limited by disruptions. The density limit disruptions are caused initially by impurity radiation. This causes a contraction of the plasma temperature profile and leads to an MHD unstable configuration. There is evidence of magnetic island formation resulting in minor disruptions. After several minor disruptions, a major disruption with a rapid energy quench occurs. This event takes place in two stages. In the first stage there is a loss of energy from the central region. In the second stage there is a more rapid drop to a very low temperature, apparently due to a dramatic increase in impurity radiation. The final current decay takes place in the resulting cold plasma. During the growth of the MHD instability the initially rotating mode is brought to rest. This mode locking is believed to be due to an electromagnetic interaction with the vacuum vessel and external magnetic field asymmetries. The low-q disruptions are remarkable for the precision with which they occur at q_ψ = 2. These disruptions do not have extended precursors or minor disruptions. The instability grows and locks rapidly. The energy quench and current decay are generally similar to those of the density limit.

The stellarator equilibrium with a steady flow is investigated. A set of averaged, axially symmetric scalar equations is obtained, equations for the shift of the magnetic surfaces and the surfaces of equal density, ρ = const, are derived, and the changes in the topology of a given vacuum magnetic configuration are determined. It is shown that the topology of the ρ = const surfaces both in the cases of tokamak and stellarator can change while the topology of the magnetic surfaces does not.

An improved triton burnup code, using the measured shot-to-shot plasma parameters and profiles, is used to recalculate the ratio of 14 MeV to 2.5 MeV neutron fluences under the assumption of neoclassical slowing down. This neutron fluence ratio is compared with the experimental fluence ratio from the FT tokamak for various toroidal magnetic fields and plasma currents. The accuracy achieved in the calculations permits a better evaluation of the confinement of fusion produced tritons during slowing down in FT Ohmic plsma. Assuming a current distribution J(r) ∼ T_e^3/2, it is found that the confinement is neoclassical, within the accurary given by the error bars, but a deterioration is observed for decreasing q(a). When the Kadomtsev sawtooth model is used for the current distribution, such a deterioration does not seem to occur.

The plugging regime in which there is no outstreaming neutral flux is studied for a closed configuration pump limiter (with a throat). The plugging length and the neutral density buildup at the neutralizer plate are derived by computations. The analytical expressions are supported by numerical evidence. Owing to the throat effect, an improved pump limiter 'efficiency' is found. This effect is mainly due to neutral-side wall interactions.

Gas cells which are used to neutralize the negative ion beam of a high energy neutral beam injector can have an efficiency of ≈55%. It is made plausible that the application of direct recovery of the energy of the non-neutralized beam fraction is beneficial for a negative ion based system. In first order, the gas cell efficiency increases to 75%. However, to have a constant neutral particle flux, the source area has to be increased by about 10%.