No abstract

A multichannel far-infrared interferometer used on the Joint European Torus (JET) is described. The light source is a 195-μm DCN laser. The instrument is of the Mach–Zehnder type, with a heterodyne detection system. The modulation frequency (100 kHz) is produced by diffraction from a rotating grating. There are six vertical and two oblique channels. The latter rely on retroreflection from mirrors mounted on the vessel wall. Their vibration is compensated by a second wavelength interferometer at 118.8 μm. The various subsystems are described, with emphasis on features necessitated by (a) large path lengths, (b) remote handling requirements, (c) fluctuations in atmospheric humidity, and (d) unmanned automatic operation. Typical measurements, along with real-time and off-line data analysis, are presented. The phase-shift measurement is made with an accuracy of (1)/(20) of a fringe, corresponding to a line-integrated electron density of 5×1017 m−2. Comparison with other electron density diagnostics are shown. The introduction of additional optics allows measurements of the Faraday effect and a determination of the poloidal magnetic field distribution. The signal processing and data analysis are described. Errors introduced by the calibration procedure, birefringence of the probing beams, toroidal field pickup, the flux geometry, and the density profile are considered. The Faraday angle is measured with an accuracy of 5% and a time resolution of 1–10 ms. The poloidal magnetic field is deduced with an accuracy of ±15%.

Results are presented of transport studies conducted on trace impurities injected with the laser blow-off technique in a variety of JET L mode pulses. In the core of the discharge, the transport is much slower, the impurity diffusion coefficient D is more than an order of magnitude below the values it assumes further out, but still above the neoclassical predictions. The extent of the slow transport core region varies with the magnetic field and with the total plasma current and is broadly correlated with the edge value of the safety factor. Closer analysis reveals that the current profile is essential in determining the radial dependence of D. This parameter appears to undergo a rapid transition to highly anomalous levels in the vicinity of the radial position where the dimensionless shear parameter is equal to 0.5. Within that region D stays moderate even when the electron temperature gradient is high. A marked increase of D in the outer region of the discharge is observed when the power per particle is raised or, alternatively, when the temperature and its gradient grow in that region, but no clear dependence of D on plasma density is found when the electron temperature profile is kept constant. Transport modelling based on the critical Del Tc assumption leads to D profiles that are similar, although not in detailed quantitative agreement, to the experimental ones when the temperature profiles are flat in the centre; when the temperature profiles are peaked in the centre, even the radial dependence of the predicted diffusion profiles is very different from the one observed. Recent theoretical attempts to analyse the radial structure of the microturbulent fluctuations predict a strong positive dependence of anomalous diffusion on the magnetic shear as observed in our experiments

Impurities are injected into JET and Tore Supra by the laser blow-off ablation technique for a variety of experimental situations. The impurity confinement times τp, which reflect the impurity transport, have been measured by fitting exponential curves to the decays of the central brightnesses. Two databases are built up, including ohmic and L mode discharges, with the aim of determining a τp scaling law common to both devices. Different monomial scaling laws are tested and the best one is chosen on the basis of minimization of the standard deviation of the individual regressions. Moreover, the impurity confinement times are compared with the energy confinement times τE for the same data. The energy confinement times are larger than the impurity confinement times, the average ratio tau E/τp being approximately 2.5