Table of contents

It has been found that values of n_eu and T_eu in the SOL of single null, high recycling divertor discharges in JET and Alcator C-Mod are related approximately as T_eu ∝ n_eu^1/2, where (T_eu,n_eu) are the values of plasma temperature and density on any given flux tube at the `upstream' end, for example, halfway between the two targets. Using the standard two point model for the high recycling (conduction limited) divertor, this result is shown to be directly related to criteria based on plasma collisionality - which have to be satisfied if the flux tube is to be in the conduction limited (high recycling) regime rather than the sheath limited (low recycling) regime or the detached regime: the (T_eu, n_eu) points for conduction limited flux tubes are predicted to fill a (somewhat narrow) band in the (T_eu, n_eu) plane. Secondly, a tendency for the collisionality, as calculated for upstream conditions, to remain constant across the SOL is noted and an explanation is given. The latter then implies a tendency for T_eu ∝ n_eu^1/2, to hold, not just at the separatrix but across the SOL. Thirdly, for the JET data a further re-enforcing factor tending to cause a rather strong T_eu ∝ n_eu^1/2 correlation is identified in the dependence of χ^SOL_⊥ on the plasma parameters. The values of χ^SOL_⊥ were extracted from the JET data using an onion skin method analysis. The first effect is expected to be general to tokamaks operated in the high recycling/conduction limited regime, although it only states that (T_eu, n_eu) values will fall within a certain band, and a decreasing value of T_eu with increasing n_eu is not ruled out. It is not known at this time how general the second two effects are.

In Part I it was shown that values of T_eu and n_eu in the SOL of single null, high recycling divertor discharges in JET and Alcator C-Mod are related approximately as T_eu ∝ n^1/2_eu, where (T_eu, n_eu) are the values of plasma temperature and density on any given flux tube at the `upstream' end, for example, halfway between the two targets. This implies that the upstream radial decay length can be expected to approximately satisfy λ_n/λ_T ≈ 1/2, in the conduction limited regime. Other simple correlations, such as between target ion saturation current density, J_sat,t, and n_eu are also implied, and are found to hold for the database considered. The variation of n_eu with ⟨n_e⟩ is established for the JET database and this is used to express the region in (⟨n_e ⟩, P_SOL) space for conduction limited operation, i.e. the respective transitions to sheath limited and detached divertor regimes.

The first simultaneous measurements of the toroidal rotation velocities of main (D⁺) and heavy impurity (Cr²²⁺) ions in the core of ohmic plasmas are presented. For the main ions a system of five neutral particle (charge exchange) analysers is used, one having a near tangential line of sight. The velocities of the Cr²²⁺ ions are obtained from the Doppler shift of X ray lines observed with a high resolution crystal spectrometer. Both ion species move in the direction opposite to the plasma current. The velocities are in the range 20-40 km/s. There is a strong indication that the toroidal motion is driven by the radial electric field created by ripple losses, as expected from neoclassical transport theory.

Vertical displacement events (VDEs) and their attendant `halo currents' flowing in the plasma facing components pose a potentially serious hazard to operation in large fusion devices. Experiments on the COMPASS-D tokamak indicate that, on this machine, peak halo currents I_θ exhibit a scaling I_θ ∝ I_p0/q₉₅, with all data falling within the boundary I_θ ⩽ 1.2 I_p0/q₉₅. Toroidally asymmetric halo currents peak along with the symmetric halo currents, and are nearly fixed in toroidal phase, according to magnetic coil and resistive shunt data. The toroidal peaking factor (peak to average ratio) of the halo currents decreases with increasing I_θ/I_p0, but exhibits significant scatter. Here, the experimental arrangement is described, and the above results are discussed. A model for the observed boundary in the peak halo current data is introduced, employing near perfect imaging of the plasma current by a halo that becomes highly conducting during a VDE. The model reproduces both the above limit observed on COMPASS-D and also the higher boundary observed on Alcator C-Mod. In addition, the model results are consistent with reports from other machines of the effectiveness of schemes to reduce halo currents such as gas puffing and killer pellet injection.

Thermographic measurements using an IR scanner have been performed at the pump limiter ALT-II of TEXTOR-94 during RI mode discharges and during disruptions. The measurements on the RI mode discharges were done to complete the TEXTOR database which had shown a structured decay pattern of the deposited power. It was found that the underlying radial heat flux can be described by two exponential decay functions. This structure, which generates an unexpected heat component close to the tangent line, has been observed in all discharge conditions including the RI mode. During disruptions, the heat is released in short pulses with a typical duration of 0.01-0.1 ms. The radial decay length of these pulses has a similar shape to the heat flux during normal discharges: it consists again of a strong component close to the tangent line with a radial decay length of 2-5 mm and probably one with a decay length of the order of 1 cm. The heat is released at the time when the edge electron temperature of the plasma drops, when intense hydrogen and carbon fluxes occur near the walls, and when electrical currents in the limiter blades are excited. In a tentative interpretation, the temporal and spatial structure of the heat pulse is attributed to the presence and growth of a laminar zone at the plasma edge, which is connected with the ergodization of the plasma edge during a disruption.

A calculation method for fusion product momentum deposition in dense plasma spheres has been developed and applied to laser imploded DT targets. The net radial momentum deposition from αparticles accelerates the expansion of the plasma sphere and thus reduces the rate of thermonuclear reactions. However, this effect is not so significant as previously expected because the rate of momentum deposition by the α particles is much less than the local pressure gradient around the burn front. The momentum deposition from neutrons is negligibly small because of its long mean free path.

Electron transport coefficients have been determined in the W7-AS stellarator. The diffusion and convection coefficients were obtained by modulating the gas feed into the plasma and by measuring the propagation of the density perturbation. The experiments were carried out at a variety of plasma densities, heating powers and magnetic fields. The results are summarized in the form of a scaling expression for the diffusion coefficient. Transient inward convection was found in the boundary plasma. This convection plays a minor role in the core plasma except with higher heating power, where outward convection was observed. The results are compared with earlier W7-AS transport results.

The effects of increased geometrical closure on the behaviour of the recycling and intrinsic impurities are investigated in JET Mark I, Mark IIA and Mark IIGB pumped divertors. Increasing the divertor closure leads to a significant improvement in exhaust for both deuterium and recycling impurities. However, the impurity enrichment in the exhaust gases remains unchanged due to a simultaneous increase in deuterium and impurity compression in the divertor. A comparison is made for helium, neon and argon under different plasma conditions. In addition, the operation of the Mark II and Mark IIGB divertors has shown that Z_eff is reduced with the improved divertor closure in the L mode discharges, although no obvious changes in the Z_eff values have been observed in the ELMy H modes. The divertor target surface temperature has a strong influence on intrinsic carbon production. The carbon source in the Mark II and Mark IIGB divertors is significantly higher than that in the Mark I divertor, which is attributed to enhanced chemical sputtering at the increased divertor tile temperature of the Mark II and Mark IIGB divertors (related to the divertor cooling system), as opposed to the increased closure. The consequences of this elevated yield for plasmas under different operation conditions are discussed, and further evidence, obtained from a specific wall/divertor temperature reduction experiment, is presented. The effect of the divertor screening on the chemically produced impurities is investigated using the EDGE2D/NIMBUS/DIVIMP codes for the different recycling regimes and comparisons are made with experimental observations from the Mark I, Mark IIA and Mark IIAP divertors taking into account the change in chemical sputtering yield due to the different tile temperatures of these divertors.

Losses and trapping regions in real and velocity space have been studied in the TJ-II flexible heliac for collisionless ions of low and intermediate energy (0.1-1 keV, roughly). The complexity of the magnetic field structure is described with emphasis on the ripple, Fourier mode composition, angular dependences and drift velocity behaviour. The resulting strong asymmetries of the loss distributions are discussed and explained. The influence of pitch and position is analysed by means of different complementary diagrams and a strong sensitivity to the angular position of trapping regions is shown. Electric field effects have not been included, this will be done in a companion article, with particular stress on the influence of resonances.

Previous studies are extended of losses and trapping for collisionless ions of low and intermediate energy (0.1-1 keV roughly), done for the TJ-II helical axis stellarator, to the case where radial electric fields are present. The modifications produced by the electric field on the radial profile of losses, the angular distributions at the plasma exit and the impact patterns on the TJ-II vacuum vessel are analysed showing a progressive peripheral concentration and a slow rotation of the distributions of lost particles at the plasma exit and on the vacuum vessel in the direction of the poloidal rotation induced by the electric field. Changes in radial and angular trapping distributions are discussed, showing the relative insensitivity to the electric field of the angular positions where the maxima and minima of trapping occur. The effect of the electric field resonances is analysed and there is a discussion of the order of appearance of the different TJ-II magnetic configurations and the influence on trapping and losses. The modifications produced by the electric field and the resonances on the different loss cone diagrams are analysed and explained.

Correction to Nuclear Fusion 39 (1999) p2471. We regret that an error caused the omission of S. Zweben's name from the list of members of the ITER Physics Expert Group on Energetic Particles, Heating and Current Drive given in a footnote.