Table of contents

Experiments have been performed on the JET tokamak with 2nd harmonic ion cyclotron resonance heating (ICRH) of hydrogen in deuterium plasmas to assess the role of finite Larmor radius (FLR) effects on the resonant ion distribution function. More specifically, the clamping of high-energy resonant particle distribution due to weak wave-particle interaction at high energy is studied. The distributions of ICRH heated hydrogen ions have been measured with a high-energy neutral particle analyser in the range of 0.29–1.1 MeV. By changing the electron density the energy E*, around which the wave-particle interaction becomes weak, is varied. The dependence of the ion distribution on E* is experimentally observed for a number of discharges and FLR effects are clearly seen to affect the high energy tail shape. Experiments have been analysed with the combination of ICRH modelling codes PION and FIDO, including FLR effects, and good agreement with measurements have been found.

In this paper results are presented on the changes induced by emissive electrode biasing in the ISTTOK edge transport. The boundary plasma is characterized with focus on the relation between E × B sheared flows and particle transport. We suggest that the distinct behaviour of the particle confinement for positive and negative bias observed in ISTTOK is related to the low E × B shear induced by positive bias in the core periphery region associated with the appearance of large amplitude fluctuations. In addition, the effect of the electrode bias on the edge turbulent transport has been investigated identifying the changes induced in the fluctuations frequency spectrum and probability density function. We have shown that negative electrode bias reduces the propagation of large-scale events, making the fluctuations distribution more Gaussian and resulting in low amplitude fluctuations across most of the edge plasma region. For positive bias, large amplitude broad spectrum fluctuations appear in the core periphery, which increase the cross-field diffusion and contribute to the observed asymmetry in particle transport with the bias polarity.

The effect of pre-ionization induced by depleted uranium (₉₂U²³⁸) around the insulator sleeve on the neutron emission of (1.8–3.3 kJ) plasma focus device is investigated by employing time resolved and time integrated neutron detectors. The maximum average neutron yield of 2.5 × 10⁸ is recorded at 3.5 mbar without pre-ionization, which increases up to 3.85 × 10⁸ with pre-ionization. It is found that the pre-ionization enhances neutron yield, about (50 ± 5)%, broadens the neutron emission pressure range and improves shot-to-shot reproducibility of plasma focus operation for neutron emission. The pinhole images of charged particles emitting zones indicate that the pre-ionization increases the high temperature plasma volume.

Transitions in confinement with characteristics of H-mode are observed in JET inner wall limited plasmas in experiments performed at a magnetic field of 0.8 T and at a current of 0.9 MA, using up to 7 MW of auxiliary heating power. These transitions in confinement are short lived, with durations of up to 20 ms, but the overall stored energy increases by up to 5% and edge density by up to 20%. The termination of the period of good confinement correlates with the observation of a burst of magnetic fluctuations, similar to those associated with edge localized modes (ELMs).

These transitions in confinement also correlate with a significant decrease in the density fluctuations measured by the microwave reflectometer and a small decrease in the magnetic fluctuations measured using 'Mirnov' probes, accompanied by a reduction of the D(alpha) emission. The comparison of the fluctuations in the limiter H-mode in JET with the divertor H-modes shows a similar behaviour, in particular the reduction of the fluctuation levels at high frequencies (f > 10 kHz) and an increase in the fluctuation levels at lower frequencies (f < 10 kHz). The strong electrostatic nature of the plasma edge turbulence is confirmed, where a reduction of up to ∼50% in the density fluctuation levels correlates with an increase of up to 50% in the confinement time during the transition to H-mode. However, the importance of instabilities with a dominant magnetic component (such as ELMs) is also highlighted, which cause significant increase in transport, including the loss of the good confinement phase in limiter H-modes.

Wide field-of-view, high transparency birefringent filters (essentially fixed delay interferometers) are proposed for incoherent Thomson scattering measurements of the temperature and density of plasma free electrons. For thermal electrons, the optical coherence of the Thomson scattered light at an appropriately chosen optical path delay, is a unique function of the electron temperature and density. The detection system utilizes a single filter combined with imaging optics and dual detector arrays to simultaneously observe both dark and bright scattered light interference fringes. The system delivers two signals that allow the recovery of the two unknowns—temperature and density. It is shown that the normalized intensity difference between the bright and dark interference fringes gives a direct measure of the electron temperature, even for strongly blue shifted high temperature spectra. As usual, the total scattered light flux (the sum of bright and dark signals) is proportional to the number of illuminated electrons. For multi-pulse systems, an electronically switchable ferroelectric liquid crystal delay plate synchronized with the laser repetition rate can allow density and temperature to be obtained using a single detector array. The use of a time-multiplex approach both simplifies relative channel calibration issues and opens the possibility for 2D temperature imaging. This paper describes the measurement principle and presents the results of numerical simulations for both low and high temperature scenarios.

The dependence of the Shafranov shift on magnetic field configuration, toroidicity and central rotational transform ι(0) in neutral beam heated plasma has been experimentally investigated in the Large Helical Device. The toroidicity of the plasma is controlled by the quadrupole field, while the central ι(0) is controlled by changing the distance of the current centre of the helical coil to the plasma. It is experimentally confirmed that both the lower toroidicity and the higher ι(0) contribute to the reduction of the Shafranov shift as predicted by the three-dimensional equilibrium code, VMEC.

Dependence of the energy confinement in ELMy H-mode tokamak on the beta has been investigated for a long time, but a common conclusion has not been obtained so far. Recent non-dimensional transport experiments in JT-60U demonstrated clearly the beta degradation. A database for JT-60U ELMy H-mode confinement was assembled. Analysis of this database is carried out, and the strong beta degradation consistent with the above experiments is confirmed. Two subsets of ASDEX Upgrade and JET data in the ITPA H-mode confinement database are analysed to find the origin of the various beta dependences. The shaping of the plasma cross section, as well as the fuelling condition, affects the confinement performance. The beta dependence is not identical for different devices and conditions. The shaping effect, as well as the fuelling effect, is a possible candidate for causing the variation of beta dependence.

Oxidation experiments by ion cyclotron resonance discharge in a gas mix of oxygen and helium (He/O-ICR) have been performed in a HT-7 in the presence of a permanent magnetic field of 1.5–2.0 T at wall temperatures of 400 to 470 K. Two kinds of gas mixture ratios of 4 : 1 and 1 : 1 (helium to oxygen) were used. With the same filling rate of oxygen, a higher pressure of He in the He/O-ICR plasma is beneficial for removal of co-deposition and reduces oxygen retention. For the same filling pressure, both the oxygen retention rate and the removal rate of H and C atoms during the He/O-ICR experiment were lower than that in the pure O-ICR experiment. The influence of ICR power and filling pressure on hydrogen and carbon removal rates was analysed. The highest removal rates of H and C atoms up to 5.4 × 10²¹ atoms/h and 7.2 × 10²¹ atoms/h, respectively, were obtained in 40 kW He/O-ICR cleaning with a ratio of He/O of 4 : 1 at 9.8 × 10⁻² Pa. By He-ICR cleanings and baking in helium gas, most oxygen retained on the wall was sufficiently removed before the subsequent plasma discharge. High power and high pressure He-ICRF cleanings are effective in removing the oxygen retained in the walls. Plasma discharges could be recovered after a few tens of disruptive plasma discharges.

The paper is an overview of recent results of Li limiter testing in T-11M tokamak. The lithium limiter is based on the capillary-pore system (CPS) concept. The Li erosion process and deuterium (D₂) and helium (He) sorption by Li first wall were investigated. The ability of capillary forces to confine the liquid Li in the CPS limiter during disruption was demonstrated. The idea of combined lithium limiter with thin (0.6 mm) CPS coating as a solution of the heat removal problem was realized. As a result the quasi steady-state tokamak regime with duration up to 0.3 s and clean (Z_eff = 1) deuterium plasma has been achieved. The temporal evolution of the lithium surface temperature during discharge was measured by a IR radiometer and then was recalculated to the surface power load. For the estimation of the Li limiter erosion the Li neutral and ions spectral line emission were observed. The increase in lithium erosion as a result of limiter heating was discovered. The radial distribution of plasma column radiation measurements showed up to 90% of the total radiation losses in a relatively thin (5 cm) boundary layer and only 10% in a plasma centre during discharges with high Li influx. Oscillations of Li emission and saw-tooth-like oscillations of the limiter surface temperature have been detected in discharge regimes with highest Li limiter temperature (>600 °C). A version of Li CPS first wall of DEMO reactor and Li CPS limiter experiment in the International Thermonuclear Energy Reactor are suggested.

A low density H-mode plasma has been selected for detailed inter-ELM modelling by the SOLPS code package, with the coupled treatment of its plasma (fluid code B2) and neutral (Monte-Carlo code Eirene) parts. Good quality measured midplane density and temperature profiles, covering the pedestal region and stretching far into the SOL, as well as several other parameters and profiles measured in the divertor, have enabled testing the consistency of code solutions with experiment. Once the upstream, midplane profiles have been fitted, and the global parameters (e.g. input power into the computational grid, radiated power) matched, the code reproduces experimental profiles and control parameters in the divertor with an accuracy within a factor of 2. Deviations of modelled parameters from the experiment were found around the strike point position where most of the power was deposited on the target. The deviations are consistent among themselves and all point to one common problem with the modelling: the predicted divertor electron temperature is too low and the density too high, compared with the experiment. The largest inconsistency between the code and experiment was in the magnitude of the peak H_α radiation in the outer divertor, which was larger by a factor of 2 in the code simulations. In addition, the code predicts a somewhat higher sub-divertor neutral flux but lower carbon impurity content in the edge plasma than in the experiment, as well as lower CIII emission. The discrepancy between H_α profiles can to a large degree be attributed to profile effects: the simulated H_α emission profiles are narrower than in the experiment, reflecting the tendency of the neutral–plasma mix to congregate excessively around the strike point in the modelling. At the same time, the integrated H_α emission matches very well with the experiment.

Extensive sensitivity studies of the influence of variations in input parameters and assumptions of the code on the modelled divertor conditions have been conducted. They have not resulted in an identification of any SOLPS input/control parameters capable of removing the main disagreement between the code output and experiment. A possibility of parallel transport effects related to low collisionality to increase the effective plasma temperature near the strike point position or of increased perpendicular transport by neutrals (due to some missing reactions in Eirene) to widen the target profiles, will be explored in the future.

Peculiarities of the ordinary to extraordinary wave conversion in the electron cyclotron frequency range near the cut-off surfaces in magnetized plasmas are analysed in a two-dimensionally inhomogeneous simplified tokamak-like geometry. It is demonstrated that the mode conversion may be of essentially two-dimensional nature even neglecting the curvature of the cut-off surfaces; for the latter case a set of reduced wave equations in the transformation region is derived and solved analytically. Transformation coefficients for quasi-optical beams with a finite transverse distribution of rf field are obtained as well as a recipe for synthesis of the optimal (perfectly converted) beam. Applicability limits of a conventional one-dimensional theory are finally discussed.

The stability of the trapped electron mode is studied in conditions characteristic of internal transport barriers, namely steep density and temperature gradients. An analytic model allows a unified treatment of all collisionality regimes, from the dissipative limit to the weakly collisional regime (when the velocity space boundary layer between passing and trapped populations of electrons plays a role). Furthermore, it reveals the key parametric dependences on wavelength, collisionality and η_e = d(lnT_e)/d(lnn_e). The roles of shear damping and Landau-drift resonance are also discussed. The main outcome is that below a critical collisionality, defined by the parameter (where ν_the is the thermal electron collision frequency, L_n the density scale length and v_thi the ion thermal speed), there is strong stabilization of long wavelength modes, so the unstable spectrum may be restricted to shorter wavelengths as the collisionality falls and the density profile steepens. The predicted critical value of is experimentally relevant and this theory suggests a mechanism for barrier formation.