Table of contents

The spatial distribution of collisionless ions collected by a spherical object of radius much larger than the Debye length, in a flowing plasma, is calculated using a particle-in-cell code. The results provide the first rigorous theoretical calibration of a `Mach probe' in a plasma with negligible magnetic field. They are also applicable, for example, to spacecraft-plasma interactions. Ion to electron temperature ratios 0.1<T_i/ZT_e<10 are explored. Qualitative differences are observed between the results for low and high T_i/ZT_e, arising from ion dynamics in the wake. However, a practically universal form for the upstream to downstream current density ratio dependence on flow velocity is observed for T_i/ZT_e≲3.

A numerical study of neutral beam injection into spheromaks is presented. The beam evolution is calculated through a Monte-Carlo simulation and the plasma MHD equilibrium is determined self-consistently with the current produced by the beam. The exact equations of motion are used for the beam particles instead of the usual guiding-centre approximation. The guiding-centre trajectories are clearly different from the actual particle trajectories and this difference produces discrepancies in the total driven current and in the current profiles. A reduction of the plasma effective charge, Z_eff, does not result in an improvement in the current drive efficiency because the reduction of the stopping cross section is compensated by an increase in the electron cancelling current. The safety factor profile of the self-consistent equilibria shows a clear sensitivity to the driven current profile. The value at the magnetic axis (q₀) diminishes when the beam is injected at the magnetic axis and increases for injection above the axis. Power deposition profiles for simple injection configurations are also shown.

A neoclassical tearing mode (NTM) requires a finite size seed island to become unstable. Usually the local pressure gradient is relatively large at the β-values needed for these seed islands to destabilize the NTMs. Therefore, the island has a large growth rate at mode onset and grows rapidly to its saturated island width. This width is proportional to β as long as it is well above the marginal β-limit below which the mode is stable. The marginal β-limit is independent of the seed island trigger mechanism and provides detailed information on the stabilizing terms in the modified Rutherford equation, which are not unambiguously determined theoretically. It is shown that in JET the marginal normalized β-limit for the 3/2 mode, β_N,marg, is of the order of 0.5-1 for magnetic fields between 3.3 and 1 T, with q₉₅≈3.3, and near the H-L transition. Therefore, all H-modes with typical q-profiles (q₉₅≈3.3) are metastable in JET to 3/2 NTMs. In addition, the marginal island width is of the order of 2-4 cm and the stabilizing terms are such that they influence the saturated island width when it is smaller than 4-6 cm in these H-mode discharges. It is also shown that detailed analyses of the time evolution of the island width with slow β ramp-down suggest that the convective form of the stabilization term due to the `χ_⊥ model' is more appropriate and can explain the island decay between 4 and 6 cm to the marginal island width, while the polarization current model can explain the rapid stabilization when β<β_marg. The range of values of the different stabilizing terms are discussed in detail. In particular, it is shown that the mode is stabilized and has a large negative growth rate, when the effect of the stabilizing terms is such as to reduce the saturated width by a factor of 2.

Low plasma density discharges at high longitudinal electric fields (E₀⩾1 V m^-1) in the ISTTOK tokamak revealed characteristic features for the runaway (RA) generation process. The increase in soft x-ray emission was detected at the onset of the RA regime. The relaxations of the plasma parameters, a typical satellite effect of the instability driven by the RA electrons, were observed in experiments. In some discharges instability caused minor disruptions.

Simultaneous determination of the RA electron characteristics and reconstruction of the electron temperature temporal evolution has been performed using self-consistent 0-D calculations of the plasma transport on the basis of the measured plasma parameters with the inclusion of the RA generation process. The kinetic energy as well as the confinement time of RA electrons was evaluated. The modelling procedure developed was applied in the investigation of a possible RA generation process at the thermal quench in large-scale tokamak disruption.

Stability of alternative toroidal systems is theoretically studied. It is assumed that, on the one hand, these systems are stabilized by compressibility, and, on the other hand, `according to Mercier', they prove to be unstable. As shown, a remarkable property of `well-organized' alternative systems is the fact that, for such systems, in contrast to `classical' ones (like large-aspect-ratio tokamaks and stellarators), the Mercier modes are the relatively slow sound modes but not the rather fast Alfvén modes. As a result, the ideal instabilities, arising in violation of the Mercier stability criterion, have relatively small growth rates and, in principle, can be suppressed by effects lying beyond the scope of ideal magnetohydrodynamic. In this respect, alternative systems have an advantage over classical ones, for which the violation of the Mercier stability criterion should lead to catastrophic consequences. A new stability criterion, allowing one to distinguish `well-organized' from `badly organized' systems, is derived. To use this criterion one should know the so-called Glasser parameters, calculated by means of averaging certain equilibrium values over the magnetic surface of the system. It is explained that a `candidate' for `well-organized' systems should belong to the class of the finite-aspect-ratio systems. It is noted that, as such `candidates', one can consider, in particular, spherical tokamaks and compact stellarators.

A microscopic theory of electron cyclotron frequency (ECE) intensity fluctuations caused by a weak low-frequency plasma turbulence is presented. The emission is treated as a collective process and the ECE signal fluctuations are expressed in terms of the correlation tensor of random source-currents of bare electrons modified by the turbulence. A contribution of the background radiation modulation due to turbulence induced fluctuations of the wave damping rate is also taken into account. The plasma electron excursions resulting from drift in the turbulent electric field in the presence of the temperature gradient are found to be the principal mechanism of the ECE fluctuations. Opportunities for ECE fluctuation measurements using the ordinary and extraordinary wave modes radiation at various cyclotron harmonics are analysed.

The data in figure 15 were incorrectly plotted and the graph should be replaced with the following corrected version.

Figure 15. Ion pressure profile peaking factor plotted against β_N at the disruption of selected high performance ITB plasmas.

This special issue highlights the continued progress and new insights gained with the JET facility, which is now being operated under EFDA (European Fusion Development Agreement), a structure in which all European Associations participate actively and work together for a common goal. The current set of papers intends to give a broad overview of the work completed in the period 2000-2001 in the framework of the so-called Task Force Scenario 1. This Task Force has the mission of studying the physics of the ELMy H-Mode, the reference scenario for ITER.

Plasmas with improved performance at high density have been demonstrated making use of strong shaping, impurity seeding and pellet injection. The results constitute an important extension of the H-mode database at high density and high energy confinement. Special attention is given to the characterization and mitigation of Edge Localized Modes (ELMs) which can present a large instantaneous heat flux to the divertor. Various other aspects of H-mode plasmas are addressed, in particular the physics and implications of spontaneous peaking of the density profiles and the characterization of the H-mode pedestal for the new configurations explored.

In finalizing this cluster of articles, thanks are due to all colleagues involved in preparing and executing the JET programme under EFDA: the participants in the programme of the Task Force S1 and all other European and non-European scientists who contributed to the JET scientific programme, the Operating Team of the JET facility and the colleagues of the Close Support Unit (CSU). Thanks are also due to the Editors, Editorial Board and Referees of Plasma Physics and Controlled Fusion together with the Publishing Staff of IOPP who have supported and contributed substantially to this initiative.

We present the results of experiments in JET to study the effect of plasma shape on high density ELMy H-modes, with geometry of the magnetic boundary similar to that envisaged for the standard Q = 10 operation in ITER. The experiments described are single lower null plasmas, with standard q profile, neutral beam heating and gas fuelling, with average plasma triangularity δ calculated at the separatrix ~0.45-0.5 and elongation κ~1.75. In agreement with the previous results obtained in JET and other divertor Tokamaks, the thermal energy confinement time and the maximum density achievable in steady state for a given confinement enhancement factor increase with δ. The new experiments have confirmed and extended the earlier results, achieving a maximum line average density n_e~1.1n_GR for H₉₈~0.96. In this plasma configuration, at 2.5 MA/2.7 T (q₉₅~2.8), a line average density ~95% n_GR with H₉₈ = 1 and β_N~2 are obtained, with plasma thermal stored energy content W_th being approximately constant with increasing density, as long as the discharge maintains Type I ELMs, up to n_ped~n_GR (and n_e~1.1n_GR).

A change in the Type I ELMs behaviour is observed for pedestal densities n_ped≳70% n_GR, with their frequency decreasing with density (at constant P_sep), enhanced divertor D_α emission and increased inter-ELM losses. We show that this change in the ELM character at high pedestal density is due to a change in transport and/or stability in the pedestal region, with the ELMs changing from Type I to mixed Type I and Type II. The similarity of these observations with those in the Type II ELM regime in ASDEX Upgrade and with other small ELM regimes in DIII-D, JT-60U and Alcator C-MOD is discussed.

Finally, we present the first results of experiments by studying in more detail the effects of the plasma boundary geometry, in particular by investigating separately the effect of the upper and lower triangularity, at high average δ. We show that the changes to the lower δ (or of the radial position of the x-point) affect the pedestal parameters, the size of ELM energy losses as well as the global energy confinement of the plasma.

This paper discusses how the proximity to the L-H threshold affects the confinement of ELMy H-modes at high density. The largest reduction in confinement at high density is observed at the transition from the Type I to the Type III ELMy regime. At medium plasma triangularity, δ≅0.3 (where δ is the average triangularity at the separatrix), JET experiments show that by increasing the margin above the L-H threshold power and maintaining the edge temperature above the critical temperature for the transition to Type III ELMs, it is possible to avoid the degradation of the pedestal pressure with density, normally observed at lower power. As a result, the range of achievable densities (both in the core and in the pedestal) is increased. At high power above the L-H threshold power the core density was equal to the Greenwald limit with H97≅0.9. There is evidence that a mixed regime of Type I and Type II ELMs has been obtained at this intermediate triangularity, possibly as a result of this increase in density. At higher triangularity, δ≅0.5, the power required to achieve similar results is lower.

Recent experiments on the Type I ELMy H-mode regime performed at JET with improved diagnostics have expanded the range of parameters for the study of Type I ELM energy and particle losses. Deviations from the standard behaviour of such losses in some areas of the Type I ELMy H-mode operating space have revealed that the ELM losses are correlated with the parameters (density and temperature) of the pedestal plasma before the ELM crash, while other global ELM characteristics (such as ELM frequency) are a consequence of the ELM-driven energy and particle flux and of the in-between ELM energy and particle confinement. The relative Type I ELM plasma energy loss (to the pedestal energy) is found to correlate well with the collisionality of the pedestal plasma, showing a weak dependence on the method used to achieve those pedestal plasma parameters: plasma shaping, heating, pellet injection and impurity seeding. Effects of edge plasma collisionality and transport along the magnetic field on the Type I ELM particle and energy fluxes onto the divertor target have also been observed. Two possible physical mechanisms that may give rise to the observed collisionality dependence of ELM energy losses are proposed and their consistency with the experimental measurements investigated: collisionality dependence of the edge bootstrap current with its associated influence on the ELM MHD origin and the limitation of the ELM energy loss by the impedance of the divertor target sheath to energy flow during the ELM event.

The design value for ITER is based on operation at n/n_GW = 0.85, β_n = 1.8 and H98(y,2) = 1. These values have been routinely achieved in JET in argon seeded ELMy H-mode discharges in different divertor configurations and with different triangularities. Two main scenarios are emerging from the experiments.

First, low triangularity (δ_u = 0.19) in septum configuration. In this case large D₂ fuelling rates lead to confinement degradation towards L-mode. The seeding of Ar during the D₂ fuelling phase gives rise to a density close to the Greenwald value. After the switch-off of the D₂ gas fuelling (`afterpuff' phase), the confinement recovers to H-mode quality whereas the density stays near the value reached at the end of the main fuelling phase and Z<sub>eff</sub> stays close to or below 2. Acting on the refuelling of Ar and D<sub>2</sub> in the `afterpuff' phase allows us to improve the stationarity of the high performance phase while maintaining up to the end of the heating phase the good confinement, density and radiation level.

Second, high triangularity (δ_u = 0.45) in vertical target configuration. In this case large fuelling rates do not lead to strong confinement degradation and the D₂ fuelling is applied continuously throughout the discharge. A radiated power fraction of up to 70 %, H98(y,2) = 0.9 at β_n = 2.1 and n = 1.15n_GW - together with the formation of a radiating mantle and moderate Z_eff - are achieved in this scenario. Furthermore, there are indications of significantly reduced heat load on the divertor target plates.

This paper addresses the issues of impurity behaviour during the argon seeding experiments in JET, in which argon and D₂ have been simultaneously puffed in ELMy H-mode discharges to reach high density regimes, maintaining good confinement properties throughout the plasma discharge. The analysis is based mainly on a 1-D impurity diffusion model, which evaluates the impurity transport coefficients in three experimental scenarios.

The available experimental data include, as a function of the time, the brightnesses of the soft x-rays and of several impurity lines, the effective charge Z_eff, the profiles of the radiated power and the fully stripped Ar and C ion densities from charge-exchange recombination. From the simulation of these data it is possible to characterize the considered scenarios in terms of high-Z ion increase in the plasma centre and high radiation from the edge. In particular, it has been found that the configuration at high triangularity with continuous D₂ puffing, characterized by an outward impurity pinch velocity, features flat or slightly hollow impurity ion profiles and high radiation from a narrow region at the edge, possibly associated to the presence of a significant neutral density. In this case an edge radiative mantle is established without high radiation from the plasma centre.

ELMy H-mode is one of the foreseen operational scenarios for ITER-FEAT. Good confinement and density is usually associated with Type I ELMs, which, however, will have a severe impact on the divertor target plates in ITER. At the JET-tokamak, impurities have been seeded into a variety of discharge configurations, by which the radiation level can be risen by a factor of 2. This delays the appearance of ELMs and reduces the power load onto the divertor target plates. In this paper, the influence of the impurities on the ELM behaviour and the plasma edge properties will be discussed. Several divertor diagnostics, such as electrical probes and infrared camera indicate a reduction of the heat power load at the inner divertor target tiles and a shrinking of the impact zone at the outer divertor.

Impurity seeding in both the Joint European Torus (JET) and DIII-D tokamaks has produced L-mode discharges with confinement enhancements comparable to H-mode and a near doubling of the core ion temperature when compared to similar unseeded discharges. Although Z_eff increases with the neon injection, the total neutron rate is as high, or higher, than reference discharges and the calculated thermal neutron rate increases dramatically in both devices. Modelling with the gyrokinetic simulation code shows a reduction in low k turbulence growth rates with neon injection decreasing to less than the E×Bshearing rate, consistent with stabilization of ion temperature gradient modes in both JET and DIII-D. Reductions in ion thermal diffusivity are also observed with impurity seeding. Neoclassical m/n = 3/2 tearing modes limit the duration of best performance in DIII-D with neon injection, while n = 1 and n = 2 magnetohydrodynamic modes limit the performance in JET.

Seeding of impurities into JET resulted in a pronounced modification of plasma confinement properties. In order to analyse the underlying mechanisms a predictive transport modelling has been performed by means of two approaches. (i) The Chalmers advanced fluid model for temperature gradient driven instabilities is applied to determine all transport coefficients necessary for transport modelling. In the present code computations the radial profiles of the particle densities and temperatures are computed using experimentally measured sources. The computations give good agreement of the temperature profiles with the measured ones. (ii) The Jülich transport model, used earlier to simulate the plasma evolution in TEXTOR radiative improved mode, has been amended by a proper description of trapped electron transport, which is important at low collisionality typical for JET. This model incorporated into the transport code RITM, which self-consistently computes particle sources due to plasma recycling and impurity ionization, reproduces well the peaking in density and pressure profiles caused by impurity seeding.

Long timescale density peaking has been observed in JET plasmas leading to densities exceeding the Greenwald value. These neutral beam heated discharges are characterized by type-I ELMs and good energy confinement. Central density is limited by neoclassical tearing modes or by termination of H-mode preceded by loss of sawteeth. When these limiting factors are avoided (i.e. at intermediate power level and with optimization of gas puffing) quasi-stationary high density plasmas with peaked density profile are obtained.

Several approaches are used at JET to achieve operation at high density with good energy confinement. One of them is the injection of solid fuel pellets to realize efficient particle refuelling by deposition deep inside the plasma column. The new pellet launch system capable of launching from the torus magnetic high field side was investigated for its capability to fulfil this task in conventional ELMy H-mode discharges. Optimized pellet scenarios were developed for plasma configurations with I_P = 2.5 MA, B_t = 2.4 T, averaged triangularity ⟨δ⟩≈0.34 and mainly neutral beam heating at a level of approximately 17 MW. The accessible operational range was extended by the pellet tool with respect to gas puff refuelling. For example, H-mode conditions could be maintained at densities beyond the Greenwald level. Plasma energy confinement was observed to become density independent at high densities. When avoiding confinement deterioration due to pellet triggered MHD activity or parasitic pellet born gas in appropriate pulse schedules, enhanced particle inventory with more peaked density profiles was achieved while maintaining the plasma pressure profile. First attempts indicate that pellet advantages can be combined with the benefits achieved by higher triangularity. Very high particle refuelling efficiencies were found for pellets injected into `ITER-like' discharges with an upper triangularity ⟨δ^u⟩ of 0.53.

It is shown statistically that, divertor closure, plasma shaping and density peaking improve the energy confinement of ELMy H-modes, whilst the confinement degrades as the Greenwald density limit is approached. A prediction of the influence of these effects on the next step device ITER is given.

A high-density, ELM-free, `enhanced Dα' (EDA) H-mode has been demonstrated with Ohmic or RF heating on the Alcator C-Mod device. An allied `low particle confinement' H-regime was also observed previously in `Mk0' divertor experiments on JET. Initial attempts to recover similar states in the JET `MkIIGB' divertor configuration are reported, employing appropriately shaped magnetic equilibria. Comparisons between cases with and without divertor cryopumping show enhanced edge sources tend to increase ELM frequency and reduce their size, without degrading confinement. However, torus pressure remains below that in C-Mod owing to lower wall recycling in JET. With cryopumping on, ion cyclotron resonance heated (ICRH)-dominated combined heating may be adapted for maximum magnetohydrodynamic stability, leading to recurrent ELM-free H-mode phases. Core confinement peaks as density and edge radiation rise in each ELM-free phase, while only small compound ELMs occur between them. Moreover magnetic and density fluctuations in each quiescent period are concentrated around two frequencies, although these modes appear to rotate in the opposite direction to coherent modes seen during EDA on C-Mod. Pedestal collisionality also remains much lower in both latest and even earlier LPCH JET cases. Finally an effect of heating scheme itself on edge properties is indicated by alternative H-modes at the same density and power, with lower pedestal electron temperature and higher ELM frequency occurring for ICRH than for NBH. Closer matches in dimensionless plasma variables would clarify scalability of EDA-mode and its potential as an integrated next-step regime.