Table of contents

The effect on the scrape-off layer (SOL) of changing the position of a tokamak limiter from the low field side (LFS) to the high field side (HFS) of the plasma is considered. Conservation of magnetic flux in the SOL requires that the area, A_||, for flow of particles and energy parallel to the magnetic field, be smaller on the HFS. The effect that this reduction in A_|| has on the SOL is that, when the limiter is on the HFS, as compared with the LFS, then the characteristic scrape-off thickness is increased and the plasma particle and energy densities are higher. Such increases are described. They were observed in an experiment in JET in which the plasma was limited first on the LFS and then, during the same discharge, on the HFS. The magnitude of the effect was larger than expected, with the line integrated density of the SOL, nλ_n, having an average increase by a factor of 5.8 ± 1.1 and an increase by a factor of 7 at the respective limiter. Moreover, there was an increase in the characteristic scrape-off thickness for power flow, λ_p, by a factor of 2.5 ± 0.4 on average, and a corresponding increase by a factor of 3.0 at the respective limiter. Such an increase in n λ_n can substantially improve the screening of recycled particles and can increase wall pumping, although not strongly in the experiments described, which were in helium to avoid these effects. The large increase in λ_p measured in JET is useful in spreading power over a larger area of a limiter on the HFS. Evidence is presented which suggests that either the observed enhancement of the effect on the SOL above expectations is due to a reduction in the Mach number at the limiter sheath when on the HFS or there are fundamental problems with the modelling procedure usually used for the SOL. Related experiments are discussed

The non-inductive current drive from directional fast Alfven and electron cyclotron waves was measured in the DIII-D tokamak in order to demonstrate these forms of radiofrequency (RF) current drive and to compare the measured efficiencies with theoretical expectations. The fast wave frequency was 8 times the deuterium cyclotron frequency at the plasma centre, while the electron cyclotron wave was at twice the electron cyclotron frequency. Complete non-inductive current drive was achieved using a combination of fast wave current drive (FWCD) and electron cyclotron current drive (ECCD) in discharges for which the total plasma current was inductively ramped down from 400 to 170 kA. For steady current discharges, an analysis of the loop voltage revealed up to 195 kA of non-inductive current (out of 310 kA) during combined electron cyclotron and fast wave injection, with a maximum of 110 kA of FWCD and 80 kA of ECCD achieved (not simultaneously). The peakedness of the current profile increased with RF current drive, indicating that the driven current was centrally localized. The FWCD efficiency increased linearly with the central electron temperature as expected; however, the FWCD was severely degraded in low current discharges owing to incomplete fast wave absorption. The measured FWCD agreed with the predictions of a ray tracing code only when a parasitic loss of 4% per pass was included in the modelling along with multiple pass absorption. Enhancement of the second harmonic ECCD efficiency by the toroidal electric field was observed experimentally. The measured ECCD was in good agreement with Fokker-Planck code predictions

An argon injection experiment on the Frascati Tokamak Upgrade (FTU) plasma has been utilized to clarify the physical mechanism responsible for impurity production. A simple 0-D model for the scrape-off layer (SOL), coupled with sputtering by the main deuteron flux and self-sputtering, can reproduce the variation of impurity concentration, nickel in the case of an Inconel limiter in FTU, when argon is puffed into the plasma. The increase of radiation power losses leads to a decrease of the SOL temperature, and therefore to a decrease of the sputtering yield. Z_eff remains essentially unchanged by argon. All the other plasma parameters (temperature and density profiles, ohmic power and energy confinement time) also show no substantial changes. Radiation losses increase, up to 80% of the input power, at relatively high densities, with a corresponding decrease of the heat load on the limiter surface

In some experiments carried out in the T-10 tokamak, injection of macroparticles resulted in a particle flux directed inwards in an inner part of the plasma column. It is proposed to take into account the toroidal Lorentz electric field, induced by the radial transport plasma flow, which is derived from the experimental data on the particle balance, in Ohm's law. This approach allows an explanation of the loop voltage evolution in the pellet injection experiments in T-10

The behaviour of intrinsic carbon (the dominating impurity) and laser blow-off injected nickel has been studied in Tore Supra during ergodic divertor (ED) activation. The carbon content is reduced in the plasma core as a result of the screening effect of the ergodic layer, which is due to increased transport in the layer, to increased recycling flux and to a modified impurity source term. Simulations of the C VI and C V intensity ratios require in the ergodic edge both an increased diffusion and a large (in the 10¹⁶-10¹⁷ m^-3 range) neutral hydrogen isotope density. The nickel burst is not 'screened' and penetrates into the core plasma. The confinement time tau _p of the injected elements is always increased when the ED is activated, but there is no modification of the core plasma transport (diffusion coefficients, convection velocities, central reduced transport region extension). To simulate the increased central tau _p values, a peripheral transport barrier has to be introduced. Satisfactory nickel brightness simulations can be obtained by modifying the inward convection velocity and/or the diffusion coefficient. If the barrier is taken to be purely diffusive, satisfactory carbon line ratios can be recovered, but the simulated injected impurity confinement time tau _p is too short. For a satisfactory nickel brightness simulation it is necessary to add to the barrier some inward convection, pushing also the intrinsic carbon ions inwards, but this hinders correct line ratio evaluations

An investigation has been made of the behaviour of fast electrons in toroidal discharges. The kinetic equation, describing the evolution of the fast particle distribution has been derived and analysed. A semi-analytical solution of the kinetic equation has been obtained in the superthermal energy region. The external electric field, turbulence, the non-uniformity of the mean magnetic field and collisions are shown to be important factors affecting the distribution function. The role of the ambipolar electric field has been established and identified as an essential factor in the process of fast electron diffusion. The strong influence of the density profile on the diffusion of fast particles is clearly demonstrated. A comparison has been made with experimental data obtained on the ZT-40M device. Agreement with the results of these experiments is observed

Self-consistent, coupled transport/MHD/scrape-off layer (SOL) divertor calculations indicate that a stable, radiating mantle can be maintained just inside the separatrix by iron impurity injection into the SOL in an ITER EDA model. The power flux to the divertor plate is reduced thereby by an order of magnitude or more. This reduction of the power exhausted to the divertor is predicted to be achieved without significant deleterious effect on the core power balance, without producing significant changes in the current profile that might trigger a disruptive collapse of the current channel and without inducing an H to L mode transition. These results suggest that an impurity seeded radiative mantle should be considered as part of the solution to the divertor heat load problem in ITER and other future tokamaks

In order to understand the particle confinement properties of the hot ion H mode plasmas of JT-60U, the particle confinement time was estimated through analyses of the neutral particle behaviour based on measurements of the line emission intensity from the neutral particles combined with neutral particle transport simulations. The relationship between the properties of the particle confinement and the energy confinement was then examined. It was found that: (a) the particle confinement time depended strongly on the averaged electron density, (b) the ratio of the particle confinement time to the energy confinement time was in the range of 1-2 and (c) the particle confinement time was well correlated with the edge ion temperature. Furthermore, in order to clarify the roles of the particle confinement in the edge ion power balance, the charge exchange and convective losses were quantitatively evaluated using the distribution of the neutral particle density and the charged particle outflux, both of which were obtained through an analysis of the neutral particle behaviour. The conductive loss was estimated on the basis of the charge exchange and convective losses and the input power, and then the thermal diffusivity was estimated. From these estimates, it was found that the contributions of the charge exchange and convective losses in the edge power balance were small, and the increase of the edge ion temperature was caused by a decrease of the thermal diffusivity at the plasma edge

Analyses of ideal MHD stabilities of high β_p plasmas show that the occurrence of β_p collapse in JT-60U is consistent with the violation of the low n kink mode stability boundary. It is demonstrated theoretically that the pressure profile, internal inductance, l_i, q_s/q₀ and β_p influence the ideal stability limit, g, imposed by the low n kink mode and the infinite n ballooning mode; here, g = β_t/(I/aB_t). Peaking of the pressure profile decreases the stability limit g to low values (2-1.5) in both high l_i (≈ 1.2) and low l_i (≈ 0.8) plasmas, though the maximum g value does not strongly depend on β_p and q_s/q₀. In a high l_i plasma with peaked pressure, the stability limit is determined by internal low n kink modes, especially by the infernal mode in the low q₀ region, while, in the low l_i plasma, it is determined by the infinite n ballooning mode. Broadening of the pressure profile combined with the high l_i (≈ 1.2) significantly increases the stability limit g (≈ 5) and makes the displacement more global, though the g value decreases with β_p for fixed values of q_s/q₀. Experimentally, by the broadening of the pressure profile with high l_i (>1.2), high g (≈ 4) plasmas are obtained and no β_p collapse is observed, which is consistent with non-violation of the ideal MHD stability boundary

The conditions are examined under which optical thickness (τ) corrections to electron cyclotron emission (ECE) measurements of electron temperature (T_e) can be neglected. By means of simple algebra it is demonstrated that for measurements of T_e transients the ECE radiation temperature (T_rad) can be directly interpreted as T_e to a 15% inaccuracy for τ > 0.7. This method is a solution to the problem identified by Janicki in 1993, namely that to estimate the variation of T_e relative to the time averaged T_e (_e/⟨T_e⟩) by _rad/⟨T_rad⟩ with the same accuracy requires τ > 3. This result implies that optical thickness effects have considerably less impact on the interpretation of ECE transients than suggested by Janicki and that they do not pose a serious problem to perturbative transport studies in small toroidal devices

The power efficiency of neutron production in the Beam Plasma Neutron Source (BPNS), proposed as a volumetric 14 MeV neutron source, is maximized at an energy of 110-120 keV for positive ion based (deuteron) neutral beams. Deuterons injected into a triton plasma and tritons into a deuteron plasma are shown to give equal results at the same injection energy per nucleon. For deuterons, this is the energy used in long pulse neutral beams on TFTR; thus, the required DC beams could be developed from the existing systems

In order to reduce or to avoid ion cyclotron damping, the use of frequencies below the ion cyclotron frequency of minority ion species or the second harmonic of majority ion species has been proposed for fast wave current drive based on direct electron absorption. For these scenarios, the Alfven or ion-ion hybrid resonance can appear on the high field side of a tokamak. The presence of these resonances causes parasitic absorption, competing with the electron Landau damping and transit time magnetic pumping responsible for the fast wave current drive. In the present study, neglecting effects from toroidicity, the mode conversion at the Alfven resonance is shown to be of the order of 5 to 10% in the current drive scenarios for the planned ITER experiment. If the single pass absorption in the centre can be made sufficiently high, the conversion at the Alfven resonance becomes negligible