Table of contents

Biasing the scrape-off layer of a tokamak is discussed as a means of equalizing the power loads to and the recycling regimes in front of the target plates of a single null divertor. The scheme would be optimal with the grad-B drift directed towards the targets, thus ensuring a low H-mode threshold, and it might, at the same time, allow an increase in the wetted plate area. The electrical requirements for and development needs towards its application under ITER-like conditions are also discussed.

An analytical procedure is presented which enables a fast estimate of collision-energy-dependent cross-section effects on thermal charge exchange spectra. The model is based both on experimental evidence and numerical simulations showing that the observed charge exchange (CX) spectra are essentially Gaussian in their shape. The collision-energy-dependent emission rate leads effectively to a lineshift (apparent velocity), usually to a reduction in linewidth (apparent temperature), and to a change in the effective emission rate averaged over the entire thermal velocity distribution function. It is demonstrated that the cross-section effect can be treated analytically introducing an approximated emission rate factor which retains the characteristics of a Maxwellian velocity distribution function using an exponential expression with only linear and quadratic velocity terms in its exponent. An algebraic deconvolution procedure is described, which enables the reconstruction of true temperature, velocity and intensities from measured CX spectra. Examples taken from a recent JET experimental campaign are used to illustrate the cross-section effects on low-Z impurity CX spectra for a comprehensive variety of neutral beams (deuterium, tritium or helium), target densities, temperatures and toroidal rotation speeds. An overview is given of representative correction factors established for high-power, high-temperature plasmas, as well as for plasmas with combined neutral beam and radiofrequency heating, and for the case of locked modes.

The eigenmode equation describing ballooning collisional drift instabilities is analysed in the presence of a radially varying poloidal velocity. The instability is found to be stabilized by an H-mode-type velocity profile. The radial mode structure is found to be more extended than its conventional localized Gaussian form. Electron Landau damping (contrary to the usual ion Landau damping of the corresponding slab problem) is conjectured as the possible sink of the outgoing energy.

The characteristic time of (incompressible) resistive modes in the neighbourhood of an X-point in the absence of external forcing is found to scale, when normalized on the Alfven time, as ( epsilon mod ln epsilon mod )^-1 with epsilon the ratio between the Alfven and the resistive time. In X-point configurations with oblique separatrices, the mode time-scale becomes faster as the angle between the separatrices decreases. When this angle becomes of the order of mod ln epsilon mod ^-1, the ( epsilon mod ln epsilon mod )^-1 scaling breaks down. Similar to the case of a resonant magnetic line configuration, resistive modes can be driven unstable by the inhomogeneity of the current density.

The biorthogonal decomposition (also referred to as the singular value decomposition) has recently emerged as a powerful tool for analysing and processing multichannel data. It is shown how this method can be used to enhance the signal-to-noise ratio of such data in a self-consistent way. Three applications to multichannel diagnostics that are routinely used on plasma devices, are presented: a soft X-ray camera, an XUV spectrometer and a microwave reflectometer. In each case, a physical interpretation of the enhancement is given. Furthermore, a robust criterion is presented, allowing data to be processed without making restrictive assumptions about their noise properties.

A global particle transport calculation for field-reversed configurations with possible anisotropy in the kinetic stress tensor is presented. The resulting particle confinement times may be greater or smaller than in the scalar pressure case, depending on the amount and behaviour of anisotropy. The calculation is restricted to a special quasi-equilibrium model we named 'anisotropic' Hill's vortex; however, similar effects should also arise for other anisotropic models. Comparison with experimental results is of limited meaning due to lack of measurements of any eventual anisotropy.

A highly coherent drift-like instability in cylindrical plasma geometry is experimentally investigated using Langmuir probes. Phase-sensitive recorded probe characteristics permit the simultaneous measurement of all relevant plasma parameters (potential, density and electron temperature). Strong electron temperature fluctuations are observable whenever a significant radial electron temperature gradient is present. These temperature fluctuations are not damped by axial electron heat fluxes, since the axial wavelength is very large. A thorough investigation of the axial fluctuation structure and of the effects of the radial electric field on the instability suggest a drift-like instability as an interpretation of the observed fluctuations. The radial fluctuation-induced heat and particle fluxes can be determined from the temporal evolution of the plasma parameters.

Kinetic and fluid treatments of nonlinear wave interactions (wave mixing and electromagnetic wave scattering) in plasmas are compared and the difference between the two approaches is discussed. The nonlinear kinetic equation for cold plasmas is shown to yield the same expression for the current that induces scattered waves as the one obtained in the fluid approximation. Thus, Bindslev's conclusion (ibid., vol. 35, p. 1615 (1993)) that the fluid approach is inapplicable to the description of electromagnetic wave scattering in plasmas turns out to be erroneous.

The use of the cold fluid (CF) model for describing scattering from thermal fluctuations was initially suggested by Sitenko (1967) and has since been widespread. This use of the CF model was shown to be inappropriate by Bindslev (ibid., vol. 35, p. 1615 (1993)) where the low-temperature kinetic (LTK) model was proposed instead. Sitenko in his present comment (ibid., vol. 37, p. 163 (1994)) appears to reject the findings of Bindslev but only deals with scattering from cold collective fluctuations, not thermal fluctuations. Here we confirm the findings by Bindslev that the CF model is inappropriate for describing the bilinear interaction of thermal waves and, in particular, scattering from thermal fluctuations. We note that although both the LTK and the CF models describe scattering in collisionless plasmas and both models involve only fields and zeroth- and first-order moments of distribution functions, they are nonetheless fundamentally different: within the limits of the collisionless plasma model, the LTK model makes no assumptions about the nature of the interacting waves and fluctuations. The CF model, on the other hand, has only been shown to be correct when the cold fluid relations apply to both interacting waves and thus is only guaranteed to correctly describe the interaction of cold collective electron oscillations in a collisionless plasma. We demonstrate that the use of the CF model in present microwave scattering experiments on fusion plasmas can lead to significant errors, and finally note that there is no computational advantage in using the CF model rather than the LTK model.

The idea of the generation of eigenmodes in plasma caused by motion of particles along the field lines, like a curvature radiation of the particles, was first considered by Blandford (1975) and then by other authors. The result was simple: if the particle drift is neglected, no wave generation takes place. Beskin, Gurevich and Istomin (BGI) revived this problem in their series of papers. It is proved in their papers that the weak curvature of the field lines provokes the appearance of the new eigenmodes and their generation. In the paper by Machabeli (1991) it was clearly shown that the statement of BGI is wrong. The same point of view was previously outlined in the papers by Larroche and Peliat (1987) and Nambu (1989). However, Istomin (1994) criticized the papers noted above and supports the reasoning of BGI. Istomin stresses that they represented a new correct method of taking into account the effect of the curvature of the magnetic field lines. The author has, in an earlier article, shown that this reasoning is incorrect and reiterates his position in this comment.