Table of contents

Three campaigns, prior to July 1994, attempted to increase the fusion power in DT plasmas on the Tokamak Fusion Test Reactor (TFTR). The first campaign was dedicated to obtaining >5 MW of fusion power while avoiding MHD events similar to the JET X-event. The second was aimed at producing maximum fusion power irrespective of proximity to MHD limits, and achieved 9 MW limited by a disruption. The third campaign increased the energy confinement time using lithium pellet conditioning while raising the ratio of alpha heating to beam heating.

A review of experiments with ECRH- and NBI-heated plasmas in W7-AS is given. Global results of W7-AS are summarized. Particular emphasis is put on electron cyclotron current drive, comparative analysis of electron heat transport derived from power balance and perturbative studies, and particle transport under combined ECRH- and NBI-heating conditions. The role of the plasma boundary field configuration as a necessary condition for the existence of the H-mode in W7-AS is discussed and first observations of coherent global Alfven eigenmodes and turbulent temperature fluctuations in the plasma core are reported.

JET was extensively modified in the 1992/93 shutdown. The new pumped divertor and many new systems were brought into operation early in 1994. Operations have progressed to 4 MA plasma current and, with substantial additional heating, H-mode confinement results confirm the expected scaling. The high power handling capability of the pumped divertor with sweeping is estimated at 20 MW for 20 s. H-mode plasmas have large Type I ELMs. With lower hybrid heating alone, 2 MA full current drive has been achieved with good efficiency, and with ICRF power, effective heating and direct electron heating have been demonstrated.

The question of anomalous transport of the energy of the electrons has been puzzling tokamak physicists for nearly as long as tokamaks exist (and internal disruptions is the other riddle which has reached its twentieth anniversary...). Magnetic turbulence has been suggested as a possible answer for a long time, but linear theory gave negative results, and experimental evidence could not be obtained with existing diagnostics until recently. In the past few years, much progress has been made, and although no final conclusion can be reached yet, it is useful to regroup and compare the latest theoretical and experimental results. As recent and exhaustive reviews already exist, we will not try to cover all the existing literature, but rather adopt a more personal and maybe biased point of view. In the first part, we discuss the possible theoretical approaches, and point out the directions along which progress should be made, especially the need to obtain a dynamical description of the magnetic structure, taking into account the feedback which the induced transport has on the instabilities. In the second part, we review and discuss the experimental evidence in favour of magnetic turbulence, and discuss the relevance of the observations to the explanation of the measured transport coefficients. Although we try to cover most of the available data, many illustrations will be taken from the Tore Supra tokamak, which is well endowed with diagnostics of the fine scale magnetic structure.

The growing field of applications of plasma as deposition, etching, surface modification and chemical conversion has stimulated a renewed interest in plasma science in the atomic physical chemistry regime. The necessity to optimize the various plasma processing techniques in terms of rates, and material properties has made it mandatory to take a new look at the various processes, as fragmentation, plasma and radical transport and plasma surface interaction with advanced diagnostics and with modelization. Many types of plasmas are used in the technology of plasma chemistry: varying from RF-glow, discharges, coronas to high density arcs. The physics involved will be illustrated at the hand of important examples: etching and modification of surfaces and the deposition of thin layers of amorphous (a-C:B, a-Si:H) and crystalline (diamond, graphite) layers. Besides dissociation and ionization in the plasma, wall association, recirculation, cluster formation and energy flows are important issues. These conclusions are also pertinent to the new field of waste destruction and are directly relevant to the economics of the plasma process.

The results of divertor studies on ASDEX Upgrade, at currents of up to 1.2 MA and heating powers up to 10 MW are described, with emphasis on the ELMy H-mode. The spatial and temporal characteristics of their heat load, and the simulation of ELMs by a time-dependent scrape-off layer code are described. High gas puff rates were found to lead to a large increase in divertor neutral pressure, at modest changes in n_e, and to a strong reduction in time-averaged power flow and complete detachment from both target plates in between ELMs. Using pre-programmed puffs of neon and argon, the radiative power losses could be raised to 75% of the heating power, in H-regime discharges, and the regime of enhanced divertor neutral pressure was found also to lead to an improved pumping of recycling impurities.

The history of low-aspect-ratio 'spherical' tokamak research is described and the advantages claimed for the concept are outlined and compared with experimental results from the START device and elsewhere. Further development of the spherical tokamak, and its potential for a materials test facility or ultimately a fusion plant, are discussed.

The effort to produce plasmas that can approach the conditions where ignition by fusion reactions can take place has generated a series of new experiments and theories on high energy plasmas whose results are relevant to the understanding of important phenomena in astrophysics. In particular, magnetic reconnection processes that occur in collisionless regimes where the effects of finite electrical resistivity are no longer significant are observed in magnetically confined plasmas where the peak electron temperatures approach or exceed 10 keV. In addition, means have been devised to control the onset of reconnection, for instance by injecting or producing high energy particle populations in regions of the plasma column where the factor which drives reconnection-the pressure gradient-resides. The relationship of these results to space physics regimes such as those characteristic of the Solar Corona or of planetary magnetospheres is discussed. Another important process is that of the so-called anomalous viscosity: the rate of angular momentum transfer which is higher than that expected from the effects of (discrete) ion-ion collisions. To explain the observation of accretion on compact objects (e.g. Neutron stars or black holes) it is, in fact, necessary to assume an effective viscosity with a rather specific spatial profile. Toroidal laboratory experiments where plasma rotation is induced by the injection of energetic neutral beams show definite evidence of anomalous viscosity which theoretical models can explain. As for the spatial profile, the support laboratory experiments have provided to the so-called "Principle of Profile Consistency" of the electron temperature, that translates into a specific class of relevant transport coefficients, can be considered as a starting basis to deal with the problem of the viscosity characterizing accretion disks. Finally, the scattering of high energy particle populations produced within thermal plasmas by collective modes such as lower hybrid modes (in the so-called slide-away regime), magnetosonic and whistler modes (as observed in deuterium-tritium plasmas where 3.5 MeV alpha -particles are produced) and bursting processes associated, for instance, with the so-called "fishbone" modes are described. The applicability is pointed out to such diverse issues in space physics as the observation of ion-conics in the Earth's magnetosphere and the process by which the transverse energy level of electrons emitting synchrotron radiation are repopulated.

The 1993-94 experimental campaign has seen a significant increase of the application of waves in the ion cyclotron range of frequency on Tore Supra. Ion cyclotron resonant heating has been applied to plasmas with strongly radiating layers, pellet-enhanced performance and lower hybrid enhanced performance. Initial experiments on direct coupling of the fast magnetosonic wave to the electrons have demonstrated coupling up to 5 MW to the plasma and have demonstrated an efficient electron heating in these discharges. A significant bootstrap current fraction is observed. This heating scheme leads to a stationary improvement in energy confinement, with implications for the Rebut-Lallia-Watkins local transport model. Finally, preliminary experiments on antenna phasing have shown clear evidence of fast wave current drive.

High resolution Thomson scattering measurements of electron temperature and density in RTP show several hot filaments in the plasma centre and sharp gradients near the sawtooth inversion radius under central ECH, and structures outside the sawtooth region both in ohmic and ECH plasmas. These measurements give evidence of a complex magnetic topology. A numerical scheme is proposed which allows computation of the magnetic equilibrium with broken surfaces. Some properties of transport in a inhomogeneous medium are discussed.

Early results from the Alcator C-Mod tokamak indicate that ohmic energy confinement does not obey the Neo-Alcator scaling. Instead 'Ohmic L-mode' confinement, close to ITER89-P, is obtained, which substantially exceeds Neo-Alcator under some circumstances. ICRF-heated plasmas follow the same scaling. H-modes are observed both in ohmic and ICRF heated plasmas, with confinement enhancements of about 2 for ELM-free conditions. Divertor operation into the closed chamber shows different regimes, depending on the density. Of particular interest is the progressive detachment at moderate to high densities. Power to the divertor plates is generally low, reflecting the influence of strong radiation in the divertor. The scaling of parameters from Alcator's growing edge database is briefly summarized.

The anomalous diffusion coefficient in a turbulent plasma with destroyed magnetic surfaces can be defined in two independent ways. The first one requires the evaluation of the mean square displacement of a particle described by a stochastic Langevin equation. The second one starts from the correlation between the magnetic fluctuation and the fluctuating distribution function, to be obtained from a kinetic equation. The two approaches are critically compared.

Superthermal ion cyclotron emission (ICE) is observed in both fusion and space plasmas. Typical spectra display strong peaks at sequential multiple ion cyclotron harmonics, and distinct energetic ion populations are present in the emitting regions. In JET and TFTR, for example, ICE appears to be driven by fusion products or by injected beam ions in the outer mid plane; and in the Earth's ring current, radiation belts, and bow shock, ICE has been observed by the spacecraft OGO 3, GEOS 1 and 2 and AMPTE/IRM, often in conjunction with highly non-Maxwellian proton populations. Common emission mechanisms, arising from collective relaxation of energetic ion populations, appear to operate in both the fusion and space plasma environments. These are reviewed, and the potential role of ICE as a diagnostic of energetic ion populations is also examined.

The hypothesis that fluctuation levels and anomalous transport in tokamaks are suppressed by radial shear in E*B rotation is reviewed, with emphasis on the L- to H-mode transition. At present there is no unambiguous experimental support for the hypothesis and further work with better space and time resolution is required. Mechanisms which rely on ion loss at the plasma edge to provide a radial electric field appear unlikely.

Recent progress in the investigations on improved confinement modes and plasma current profile control on JT-60U is described. The confinement improvement in the high- beta p mode has been found to come from a formation of an internal transport barrier at the q(r) approximately 3 flux surface. Inside of the transport barrier, ion thermal diffusivity is reduced significantly. In contrast to the ion thermal diffusivity, reduction in electron thermal diffusivity is not evident. Preliminary result of a new approach in JT-60U to evaluate local diffusivity with a density fluctuation analysis shows a qualitatively good correlation between the evaluated local diffusivity and the confinement modes: the H-mode and the L-mode, From the viewpoint of the particle control in a tokamak reactor, particle confinement and helium transport in improved confinement modes have been investigated intensively. Particle confinement time of the main plasma in H-modes is found to be less than twice the value of the energy confinement time; the value of particle confinement time is in the acceptable range. Helium transport study shows that the helium concentration into the core plasma will not be so serious in ELMy H mode discharges, because the value of the diffusivity is about 1 m²/s with a peaking parameter of 1 approximately 1.5.

The existence and dynamics of vortical structure in both homogeneous and inhomogeneous systems will be discussed. In particular the dynamics of monopolar and dipolar vortices in a plasma with nonuniform density and in a rotating fluid with varying Coriolis force is described. The role of vortical structures in connection with electrostatic plasma turbulence and the associated cross-field plasma transport will also be addressed.

A short presentation of the main characteristics and peculiar aspects of the RFP configuration is given. Confinement data from a number of experiments, which have been operated after 1980, are recalled; the main emphasis is on results from MST and RFX, the two largest existing devices, RFX, with a current capability of 2 MA, in the first two years has been operated at low currents (350+700) kA, mainly to compare different modes of creating the RFP configuration, to study field error effects and plasma wall interaction, to improve and optimize, in particular by current and field control, the discharge parameters. The reported results are in line with expectations.

Experimental and theoretical results from around the world point to the possibility of high confinement, high- beta , and high-bootstrap-fraction steady-state tokamak operating modes. These modes of operation, if fully developed and extended to steady-state, could lead to much less expensive tokamak demonstration power reactors and to a significantly reduced cost-of-electricity from fusion, as compared to projections based on low- beta _N, pulsed operating modes. Present results have clear implications in the areas of particle control, plasma shaping, and current-profile control. Thus they have strongly influenced the design of the steady-state advanced tokamak TPX, which has the mission to combine the best results from present experiments and extend them to steady-state. These results also have important implications for follow-up tests in ITER, which have the goal of studying advanced-tokamak operation in an ignited plasma, as well as for the eventual configuration of an advanced-tokamak fusion reactor.

Simultaneous achievement of high energy confinement, tau _E, and high plasma beta, beta , leads to an economically attractive compact tokamak fusion reactor. High confinement enhancement, H= tau _E/ tau _E-ITER89P=4, and high normalized beta beta _N beta /(I/aB)=6%-m-T/MA have been obtained in DIII-D experimental discharges. These improved confinement and/or improved stability limits are observed in several DIII-D high performance operational regimes: VH-mode, high l_i H-mode, second stable core, and high beta poloidal. We have identified several important features of the improved performance in these discharges: details of the plasma shape, toroidal rotation or E*B flow profile, q profile and current density profile, and pressure profile. From our improved physics understanding of these enhanced performance regimes, we have developed operational scenarios which maintain the essential features of the improved confinement and which increase the stability limits using localized current profile control. The stability limit is increased by modifying the interior safety factor profile to be nonmonotonic with high central q, while maintaining the edge current density consistent with the improved transport regimes and the high edge bootstrap current. We have calculated high beta equilibria with beta _N=6.5, stable to ideal n=1 kinks and stable to ideal ballooning modes. The safety factor at the 95% flux surface is 6, the central q value is 3.9 and the minimum in q is 2.6.

This paper focusses on JET's contribution through experimental and computational studies to solving the ITER divertor problem. Results from detached plasmas in the "Old" JET configuration of 1991 and the new JET pumped divertor of 1994 are presented and compared with numerical simulations based on the EDGE2D/NIMBUS suite of codes. A computational study of detached plasmas is undertaken in the context of the different divertor configurations planned in the JET programme.

Experiments with different dominant impurities (C, Si, Ni, Mo) have been carried out on the FTU tokamak. Density limit, marfe, Z_eff and radiation losses are compared and discussed. An analytical model that couples in a self consistent way the parameters of the SOL and the transport of the impurities generated at the limiter to the bulk plasma parameters has been used to reproduce the experimental data.

Coating the inner wall areas of tokamaks and stellarators with lithium, beryllium, boron or silicon has often led to very high plasma performance and high confinement modes of operation. The wall coating processes are reviewed with emphasis on plasma-surface-interaction mechanisms related to high performance operation.

During the first year of operation, the TCV tokamak has produced a large variety of plasma shapes and magnetic configurations, with 1.0<or=B_tor<or=1.46 T, I_p<or=800 kA, kappa <or=2.05, -0.7<or= delta <or=0.7. A new shape control algorithm, based on finite element reconstruction of the plasma current in real time, has been implemented. Vertical growth rates of 800 sec^-1 corresponding to a stability margin f=1.15, have been stabilized. Ohmic H-modes, with energy confinement times reaching 80 ms, normalized beta ( beta _toraB/I_p) of 1.9 and tau _E/ITER89-P of 2.4 have been obtained in single-null X-point deuterium discharges with the ion grad B drift towards the X-point. Limiter H-modes with maximum line averaged electron densities of 1.7*10²⁰m^-3 have been observed in D-shaped plasmas with 360 kA<or=I_p<or=600 kA.

The spontaneous transition from Ohmically heated limiter discharges into the regime with improved confinement termed as "Ohmic H-mode" has been investigated in "TUMAN-3". The typical signatures of H-mode in tokamaks with powerful auxiliary heating have been observed: sharp drop of D_alpha radiation with simultaneous increase in the electron density and stored energy, suppression of the density fluctuations and establishing the steep gradient near the periphery. The crucial role of the radial electric field in the L-H transition was found in the experiments with boundary biasing. The possibility of initiating the H-mode using single pellet injection was demonstrated. In Ohmic H-mode strong dependencies of tau _E on plasma current and on input power and weak dependence on density were found. Thermal energy confinement time enhanced by a factor of 10 compared to predictions of Neo-Alcator scaling. Longest energy confinement time (30 ms) was obtained in the small tokamak TUMAN-3. Absolute values of the energy confinement time are in agreement with scaling proposed for description of the ELM-free H-modes in devices with powerful auxiliary heating ("DIII-D/JET H-mode" scaling).

The economic viability of Tokamak fusion reactors would be greatly improved if the capital cost were significantly reduced and continuous operation were possible. Designs which realise these goals rely on reduction by a factor 2-3 in plasma current, from the 25 MA needed for a good ignition margin in ITER. A substantial proportion of bootstrap current is needed, in order to keep the recirculating power to a small proportion of the fusion output. These requirements define an operating point which has to be at large beta _p ( approximately 2.5-3) and high beta _t ( beta _N approximately 3-5). Also the energy confinement must be well above H-mode. Steps have already been taken in this direction at JET and a programme has evolved which emphasises coherent development of steady state plasmas. JET has large heating and current drive systems, long pulse capability, He neutral beam injection for ash transport studies and the recently installed, pumped divertor. The results which have already been obtained and the future programme are described. Attention is drawn to the potential application of concept improvement techniques to JET's high performance plasmas so that the fusion yield and Q may be improved.