Table of contents

The paper discusses recent progress in fusion physics on the Joint European Torus (JET) in the fields of plasma confinement, magnetohydrodynamic stability, particle and power exhaust, and fusion performance and -particle effects. In this context results are presented from the three JET operating regimes: ELM-free H-mode, ELMy H-mode and optimized shear mode. These results together with those from other tokamaks form a sound basis for extrapolation to a next step device based on ELMy H-mode operation and point the way for the development of more advanced tokamak scenarios.

Few people are preoccupied with the energy issue. Indeed, inflation-corrected energy prices (in euros) are currently lower than before the first oil crisis of 1973; the annual growth rate of primary-energy use in the industrialized world has diminished considerably compared to before 1970, and oil and gas production is characterized by increased exploration activity and a wider geographical spread.

Nevertheless, there is a real energy issue. If the greenhouse effect turns out to be real, then mankind should at least slow down the consumption of fossil fuels. Given the fact that world energy consumption (especially by the developing countries) will rise in the future, and that nuclear fission power has become unpopular in the western world, the idea reigning in some circles to cope with this situation by total reliance on energy savings and renewable energy sources comes close to wishful thinking. A realistic analysis makes it clear that there will be a need for large workhorses for electricity generation to keep the overall electricity grid sufficiently robust.

From a global and long-term perspective, the logical conclusion is the following: because mankind cannot count on the continued use of fossil fuels (due to the finiteness of the resources combined with the possible climate change effects), our generation has the responsibility to develop alternative energy sources for the distant future. Many parallel lines of research and development therefore need be pursued; because of the uncertainties with other alternative sources, it would be irresponsible to kill some of these development lines. This holds for renewable sources, the nuclear fission breeder, and for nuclear fusion.

A major hurdle for the survival of long term energy research and development is the liberalization of the electricity market. Because of the revolutionary changes taking place, utilities concentrate on cost cutting and short-term survival. In addition, they are no longer supposed to take responsibility for the future strategic electric energy provision. Although they may be sympathetic to the further development of nuclear fusion research, they do not have any interest in financial support. According to utilities operating in a liberalized market, the research and development for energy technologies must be performed by the manufacturers; if these develop an interesting product, then utilities may buy it. Manufacturers in turn consider the payback time of fusion research and development investments too large to put much money into it. Public funding therefore remains the only option for the next few decades. But strangely enough, regardless of the requirements for a long-term energy policy, policy makers also concentrate on short-term returns. Everybody is blinded by the current cheapness of energy.

Utilities will only buy fusion plants if they are competitive. The initial investment cost should therefore be reasonable, the construction time limited, and the availability for operation sufficiently high. Present-day cost estimates for fusion power plants carry little weight; they merely serve to indicate the weak spots in present-day designs. However, there is no doubt that the future fusion reactor must become much simpler and more robust than present-day experiments. Future competitiveness of fusion plants will largely depend on the price of other energy sources.

Time works in the right direction for fusion: the other sources will become more expensive, and present-day sophisticated technologies characteristic for fusion (superconductivity, remote handling, etc) will by that time have become daily technologies at a reasonable cost.

Fusion may succeed in developing a good electricity generating product for the second half of the 21st century. The major challenge consists of finding sufficient financial funding for the continued development of fusion research.

Research in inertial fusion sciences and applications worldwide is making dramatic progress. The National Ignition Facility (NIF) in the US and the Laser MegaJoule (LMJ) in France are being built to achieve fusion ignition in the laboratory. Experiments that have been done on current Inertial Confinement Fusion (ICF) facilities in the US and around the world have demonstrated that the drive characteristics required for ignition are now well understood and a new plan for inertial fusion energy development has been put together by the community. Besides examining the conditions necessary for fusion ignition, targets were designed without fusion capsules. Equilibrium temperatures of hundreds of electron volts and megabar pressures were used to study astrophysical processes and measure equations of states at these extreme conditions. Recent studies of laser-matter interactions with femtosecond lasers have revealed some startling new phenomena due to the ability to achieve irradiances >1020 W cm^-2. This paper will review recent results in fusion and high energy density science achieved by high intensity lasers at LLNL and will look ahead to what may achieved on NIF.

This paper gives an overview on the specific contributions from the tokamak TEXTOR/TEXTOR-94 to plasma-wall-interaction research, putting emphasis on the erosion and deposition properties (in particular graphite), hydrogen recycling mechanisms (molecules and atoms) and the characteristics of impurity radiation. The improved knowledge about these processes helps one to understand the mechanisms underlying these techniques, which have been pioneered on TEXTOR-94 for actively influencing plasma-wall interaction: impurity control by thin coatings, power exhaust on large areas with radiation from seeded impurities and the extension of the density limit by the control of radiative instabilities (MARFE). Also, the novel concept of the Dynamic Ergodic Divertor is presented as an innovative tool for controlling plasma-wall interaction.

In order to establish the physics basis for the sustainment of the high integrated performance required in the next-step experimental reactor, JT-60U has been optimizing the discharge control scenarios of high confinement plasmas and extending the operation regimes by utilizing a variety of heating and current drive systems, including the new Electron Cyclotron Heating/Electron Cyclotron Current Drive (ECH/ECCD) system, and the W-shaped pumped divertor. For the integration, current and pressure profile controls are essential and high triangularity is beneficial. The range of plasma current with high values of H-factor, _N and bootstrap fraction was extended to 1.5 MA with nearly full non-inductive current drive by Negative-ion-based Neutral Beam (N-NB) injection into the high-_p H-mode with Edge Localized Modes (ELMy H-mode). High current drive efficiency of 1.3 × 10¹⁹ A W^-1 m^-2 was demonstrated for N-NB. High triangularity (0.4-0.5) operation extended the long pulse (~3 s) high _N (2.5-2.7) region to low-q₉₅ (~3). For the reversed shear (RS) mode, feedback control of neutron production rate and stored energy enabled reproducible achievement of the DT equivalent fusion gain Q^eq_DT>1 and sustainment of Q^eq_DT~0.4-0.5 for ~1 s. Using the RS mode and the high high-_p ELMy H-mode, the high confinement region has been extended to higher T_e/T_i and also to higher density. Electron heating by Lower Hybrid Range of Frequency (LHRF) and N-NB extended the high confinement region to T_e/T_i>1. Argon gas puffing improved confinement in the high density regime. With the W-shaped pumped divertor, the threshold heating power for the L-H transition was reduced by 30% as compared with the open divertor. Using the pumped divertor, a multiple parameter feedback control including both core and divertor plasma parameters was demonstrated.

Steady-state discharges with improved core confinement and H-mode edge with edge localized modes (ELMs) are investigated. In plasmas with an upper triangularity _top close to zero an H-factor of H_ITER89-P = 2.7 and _N = 2.2 could be maintained for 1 s and H_ITER89-P = 2.4 and _N = 2.0 for 6 s, the latter corresponding to 40 confinement times or 2 1/2 resistive time scales for current redistribution, only limited by the duration of the possible discharge length. At a line averaged density of 4 × 10¹⁹ m^-3 the central temperatures reach values of T_i = 10 keV and T_e = 6.5 keV. The stationarity of the current profile is explained by magnetic reconnection driven by strong (m = 1, n = 1) fishbones, which, in the absence of sawteeth, also expel energy and impurities. Further increasing the pressure, is limited by neoclassical tearing modes. Raising the density by edge gas fuelling and the simultaneous increase of the neutral beam power, H_ITER89-P remained unchanged up to n_e = 5.5 × 10¹⁹ m^-3, accompanied by a substantial reduction of Z_eff. Increasing _top to 0.2, both confinement and -limit improved reaching values of H_ITER89-P = 3.0 and _N = 2.4 at densities above n_e = 5 × 10¹⁹ m^-3. This resulted in the highest fusion product of n_D,0T_i,0E = 0.9 × 10²⁰ keV s m^-3 so far observed in ASDEX Upgrade.

ECR (electron cyclotron resonance) heated plasmas have been studied in the low magnetic shear TJ-II stellarator (R = 1.5 m, a<0.22 m, B = 1 T, f = 53.2 GHz, P_ECRH = 300 kW, power density = 1-25 W cm^-3). Recent experiments have explored the flexibility of the TJ-II across a wide range of plasma volumes with different rotational transforms and rational surface densities. In this paper, the main results of this campaign are presented and, in particular, the influence of iota and rational surfaces on plasma profiles is discussed.

Localized non-inductive currents due to electron cyclotron wave absorption have been measured on the DIII-D tokamak. Clear evidence of non-inductive currents is seen on the internal magnetic field measurements by motional Stark effect spectroscopy. The magnitude and location of the non-inductive current is evaluated by comparing the total and ohmic current profiles of discharges with and without electron cyclotron wave power. The measured current agrees with Fokker-Planck calculations near the magnetic axis, but exceeds the predicted value as the location of the current drive is moved to the half-radius.

Although ergodic divertors are primarily designed to control particle and heat fluxes at the plasma edge, they also happen to affect the MHD stability of tokamak discharges. On Tore Supra, the ergodic divertor has long been known to stabilize the m/n = 2/1 tearing mode induced, for instance, by edge radiation and detachment processes, thus allowing safe high-current and high-density operations. More recently, though, in discharges where ergodic divertor operations were optimized relative to the control of the edge-plasma (i.e. with large divertor perturbation), a detrimental increase in the disruptivity has been observed. The action that the ergodic divertor has on the MHD activity is interpreted in terms of a redistribution of the current profile. The latter results from a large increase in the edge resistivity, primarily induced by the degradation of the electron energy confinement in the ergodic layer. The possibility that a transport barrier develops in the vicinity of the separatrix strongly affects the considered modelling.

A description of the status of reversed field pinch (RFP) research, as resulting from recent experiments on RFX and other devices, is presented. New physics understanding is illustrated leading to a reconsideration of critical issues in the field. For example, the role of the closely fitting shell on the magnetohydrodynamic properties of the configuration seems presently less critical while its impact on plasma-wall interaction has become clear. A novel method to overcome the problem of localized plasma-wall interaction is discussed. It relies on the use of an m = 0, n = 1 external field to induce toroidal rotation of the magnetic structure. The nonlinear nature of tearing mode interaction is highlighted in these experiments, which have clearly shown phase conjugation and harmonic generation phenomena. The possibility to interact with the dynamo mechanism, demonstrated in the experiments, opens new possibilities to improve the configuration with respect to its transport properties both reducing the amplitude of magnetic turbulence and its spectrum. The impact of sheared flow on the properties of the edge layer in a RFP is also addressed.

Neutral beam injection (NBI) heating started in the second experimental campaign of the Large Helical Device (September to December 1998) by two tangential beam lines. With 100 keV hydrogen, the beam port through power of up to 3.7 MW was injected for 1 s typically. The energy confinement was systematically better than that predicted by the International Stellerator Scaling 95 up to a factor of 1.5. The temperature pedestal observed contributes to this enhancement. We have also demonstrated a long pulse discharge by NBI in the LHD. By injecting 0.7 MW of beam, a plasma with a density of 0.3 × 10¹⁹ m^-3 was sustained for 22 s. A unique oscillating phenomenon of plasma quantities with a long time scale was observed in the long pulse discharge. One of the topics of NB discharge is that the plasma can be started up by NB alone. This technique is unique for helical systems that have a vacuum magnetic field confining high energy ions, and it is useful for helical systems to be free from the constraint of magnetic field strength that must coincide with the frequency required by electron cyclotron resonance heating (ECH).

This paper reviews several aspects of magnetic helicity, including its history from Gauss to the present, its relation to field structure, its role in Taylor relaxation, and how it is defined for sub-volumes of space. Also, its importance in solar physics will be discussed.

Magnetic helicity quantifies various aspects of magnetic field structure. Examples of fields possessing helicity include twisted, kinked, knotted, or linked magnetic flux tubes, sheared layers of magnetic flux, and force-free fields. Helicity thus allows us to compare models of fields in different geometries, avoiding the use of parameters specific to one model. Magnetic helicity is conserved in ideal magnetohydrodynamics and approximately conserved during reconnection.

Often, physical systems are described in terms of interacting parts: for example one might separate the solar magnetic field into an interior field and an atmospheric (coronal) field. We can obtain insight into how the parts of a magnetic system interact by describing how magnetic helicity is transferred from one part to another. This transfer is governed by an equation similar to Poynting's theorem for the transfer of energy through boundaries.

In a confined volume, widespread reconnection may reduce the magnetic energy of a field while approximately conserving its magnetic helicity. As a result, the field relaxes to a minimum energy state, often called the Taylor state, where the current is parallel to the field. Such relaxation processes are important to both fusion and astrophysical plasmas.

Recent observations show that structures in the northern hemisphere of the sun have predominantly negative helicity, and structures in the south have predominantly positive helicity. Helicity injection by differential rotation may explain this dependence.

Operation in ASDEX Upgrade and JET with different divertor geometries is compared with emphasis upon energy and particle exhaust and the impact of the divertor on core plasma performance. Closing the divertor increases the ratio of the neutral flux density in the divertor to that in the main chamber and facilitates improved pumping. The influence of closure on H-mode performance turns out to be weak, the main chamber physics dominates. The electron density/neutral hydrogen flux in the scrape-off layer is found to be a key parameter for particle and energy confinement as well as for the radiative losses. The density in the outer scrape-off layer is related to main chamber recycling, while the separatrix density is primarily affected by divertor geometry. Reduced target plate power load due to an increased divertor radiation level is obtained in ASDEX Upgrade Div II. The radiated power fraction in JET is systematically lower, which is partly explained by the lower edge densities. Core plasma Z_eff and carbon concentrations are similar in both machines and do not depend on divertor geometry.

The spherical tokamak (ST) provides a unique environment in which to perform complementary and exacting tests of the tokamak physics required for a burning plasma experiment of any aspect ratio, while also having the potential for long-term fusion applications in its own right. New experiments are coming on-line in the UK (MAST), USA (NSTX, Pegasus), Russia (Globus-M), Brazil (ETE) and elsewhere, and the status of these devices will be reported, along with newly-analysed data from START. Those physics issues where the ST provides an opportunity to remove degeneracy in the databases or clarify one's understanding will be emphasized.

The generation of radial electric fields through radiofrequency has been investigated in a toroidal magnetized plasma without rotational transform. By launching different kinds of electrostatic waves the plasma profiles were modified and sheared electric fields were generated. Direct observation of sheared E × B flows driven by fluctuations via Reynolds stress at the resonance layer of ion Bernstein is reported. These experiments provide a new bridge between experimental measurements and theoretical models in understanding the underlying physics of improved confinement regimes using RF in magnetically confined plasmas.

The fast-ignition scheme in ICF has instigated research in a new area of laser-plasma interaction physics. This is due to the high laser-pulse intensities involved. At power densities of 10¹⁸ W cm^-2, the physics becomes essentially kinetic in nature and highly nonlinear and hence requires special tools for its adequate description. Particle-in-cell (PIC) codes meet this challenge. Simulations obtained with the first three-dimensional PIC code VLPL (virtual laser plasma laboratory) developed at MPQ are presented for a broad variety of irradiation and plasma conditions. Particular attention is paid to a detailed comparison of numerical and experimental results and to the mechanism which accelerates the electrons to relativistic energies.

The onset conditions for neoclassical tearing modes are shown to be over a wide range of experiments in good agreement with predictions of the polarization current model. In the presence of sufficiently large seed islands, this model extrapolates unfavourably to reactor-grade devices operating with peaked current profiles. The expected decrease of the seed islands produced by sawteeth at the (3, 2) and (2, 1) resonant surfaces may counteract this trend, but extrapolation of this effect has large uncertainties. Active control of this mode on ASDEX Upgrade and COMPASS-D through modulated or unmodulated current drive at the resonant surface has been demonstrated to stabilize this mode, in the electron cyclotron current drive (ECCD) case with a small fraction of driven current (2% of I_p). In advanced scenarios, magnetohydrodynamic (MHD) modes can contribute to the stationarity of the current profile. Fishbones can clamp the current profile development near the q = 1 or q = 2 surface, without significantly deteriorating energy confinement, whereas double-tearing modes, acting in a similar form, lead to substantial confinement losses. First experiments on ASDEX Upgrade with application of central electron cyclotron resonance (ECR)-heating and current drive to discharges with an internal transport barrier show a substantial effect on MHD stability, affecting the passage of the q-profile through q_min = 2, and degrading or prolonging the core reversed shear phase, depending on the current drive direction.

The present paper addresses the issue of particle recirculation in discharges where low-energy flux to ergodic divertor target plates is achieved in highly-radiating detached ohmic plasmas. Plasma temperature and particle flux are measured by flush-mounted probes in the divertor plates and by an upstream fast scanning Mach probe. The scalings with core density of the ion flux and electron temperature are well described by the simple two-point model used in axisymmetric poloidal divertors. The detachment signature is a pressure drop that occurs when the edge temperature falls below 10 eV. The parallel ion flux gradient is always positive, indicating that recombination is unlikely to play an important role in detachment. Visible spectroscopy of a neutralizer plate shows that attainment of cold detached plasmas near the density limit coincides with an abrupt increase of fuelling efficiency for both deuterium and impurities. A feedback algorithm based on real-time Langmuir probe measurements has been developed to monitor detachment and avoid disruptions.

Low-temperature, non-equilibrium plasmas form the basis of a growing variety of plasma-related processes. The demand for high-density plasmas over a wide pressure range has stimulated the development and use of microwave plasma sources in the last few years. Depending on the specific application, quite different and specialized sources have emerged. Other than empirical trial and error methods, computer simulations drastically reduce the time and effort needed to optimize the power coupling and distribution into a given gas.

The computation of electromagnetic fields in plasma sources, including the plasma as a lossy dielectric, is a practical (though not self-consistent) approach yielding valuable insight on a short time-scale. Finite integral methods (FIMs) have proven to be powerful tools because they may be interpreted as a discrete analogue representation of Maxwell's equations in the computational grid.

We have already developed and optimized a whole family of slot antenna microwave plasma sources (SLANs) based on such an approach. Our work included three-dimensional numerical simulations of the coupling structures and impedance matching in the time and frequency domains. For the smallest source µSLAN geometry-dependent resonances were also identified, suppressed or enhanced to improve plasma ignition and power coupling. In that case the driving force was to use these sources more efficiently for time-modulated power flow, which is becoming very attractive for advanced plasma-based materials processing.

The insight gained from our smallest source type µSLAN has also been successfully applied to model larger plasma sources with diameters of up to more than 0.5 m and predict their performance at realistic working parameters before construction. Examples of these approaches as well as performance data will be given.

Key physics issues in the design of a high- quasi-axisymmetric stellarator configuration are discussed. The goal of the design study is a compact stellarator configuration with aspect ratio comparable to that of tokamaks and good transport and stability properties. Quasi-axisymmetry has been used to provide good drift trajectories. Ballooning stabilization has been accomplished by strong axisymmetric shaping, yielding a stellarator configuration whose core is in the second stability regime for ballooning modes. A combination of externally generated shear and non-axisymmetric corrugation of the plasma boundary provides stability to external kink modes even in the absence of a conducting wall. The resulting configuration is also found to be robustly stable to vertical modes, increasing the freedom to perform axisymmetric shaping. Stability to neoclassical tearing modes is conferred by a monotonically increasing profile. A gyrokinetic f code has been used to confirm the adequacy of the neoclassical confinement. Neutral beam losses have been evaluated with Monte Carlo codes.

Core, edge and scrape-off-layer plasma behaviour is studied principally under conditions of an _a = 5/9 boundary island configuration - which is relevant for the upcoming W7-AS divertor campaign - but for now with ten inboard sector limiters. The major focus is on compatibility between good core confinement and attainment of high recycling at the limiter. At low input power P_in0.4 MW, operation at densities necessary to attain effective divertor action in the future invariably leads to a transition to the ELM-free H-mode accompanied by lower edge densities and increased core radiation until radiation collapse ensues. Thereby, enhancement factors in _E of nearly two above the international stellarator confinement scaling are transiently achieved. The threshold density _e^thr, necessary to attain the H-mode increases with heating power, such that at 2 MW NBI heating power the H-mode is completely suppressed and peak densities at the limiter exceeding 1.5 × 10²⁰ m^-3 are realized. The efficacy of newly-installed control coils designed to manipulate the island geometry is tested. Their influence on the core plasma is verified. Due to geometrical effects associated with the mutual shadowing of the inboard limiters, statements regarding the influence on island physics must await the divertor configuration.

Active space experiments involving the controlled injection of electron beams and the formation of artificially generated currents can provide in many cases a calibration of natural phenomena connected with the dynamic interaction of charged particles with fields. They have a long history beginning from the launches of small rockets with electron guns in order to map magnetic fields lines in the Earth's magnetosphere or to excite artificial auroras. Moreover, natural beams of charged particles exist in many space and astrophysical plasmas and were identified in situ by several satellites; a few examples are beams connected with solar bursts, planetary foreshocks or suprathermal fluxes traveling in planetary magnetospheres.

Many experimental and theoretical works have been performed in order to interpret or plan space experiments involving beam injection as well as to understand the physics of wave-particle interaction, as wave radiation, beam dynamics and background plasma modification. Recently, theoretical studies of the nonlinear evolution of a thin monoenergetic electron beam injected in a magnetized plasma and interacting with a whistler wave packet have led to new results. The influence of an effective dissipation process connected with whistler wave field leakage out of the beam volume to infinity (that is, effective radiation outside the beam) on the nonlinear evolution of beam electrons distribution in phase space has been studied under conditions relevant to active space experiments and related laboratory modelling. The beam-waves system's evolution reveals the formation of stable nonlinear structures continuously decelerated due to the effective friction imposed by the strongly dissipated waves. The nonlinear interaction between the electron bunches and the wave packet are discussed in terms of dynamic energy exchange, particle trapping, slowing down of the beam, wave dissipation and quasi-linear diffusion.

The characteristics of the toroidal ion temperature gradient (ITG) instability, considered as the main source of anomalous transport in the low (L) confinement mode of tokamaks, are analysed for the conditions of the radiatively improved (RI) mode triggered by seeding of impurities. Based on experimental profiles from TEXTOR-94 we deduce that the ITG-induced turbulence is quenched in the RI-mode in a large part of the plasma radius. Under those conditions the dissipative trapped electron (DTE) instability dominates the transport. The peaking of the electron density, one of the most remarkable features of the L-RI transition, is explained by the difference in contributions of ITG and DTE unstable modes to the diffusive and convective components of the electron flow. The shearing of the plasma toroidal velocity observed in the RI-modes with unbalanced neutral injection is successfully modelled by taking into account both the anomalous plasma viscosity due to ITG-turbulence and predictions of the revisited neoclassical theory.

The performance of ELMy H-mode operation in ASDEX Upgrade and JET is compared. Special attention is paid to variations (usually reductions) in this performance near the operational limits which will need to be approached in a next-step device. In JET it is found that input powers substantially above the H-mode threshold power are required to obtain discharges with energy confinement enhancement factors at or above the usual ELMy H-mode scalings. Such a margin (as much as a factor of two in JET) is not observed in ASDEX Upgrade. It is proposed that this difference may be due to the higher edge collisionality in ASDEX and the results are compared to a recent theory based on interchange instabilities and magnetic flutter. In ASDEX Upgrade, the confinement in type I ELMy discharges degrades as the density is raised due to a stiffness of the temperature profiles which leads to a degradation of the core confinement. This type of stiffness is observed in JET only at relatively high edge densities. In JET, the edge confinement degrades as the density is increased by external gas fuelling, consistent with a constant edge pressure gradient and an edge barrier width which reduces in proportion to the edge ion poloidal Larmor radius. In both machines, H-mode performance is limited at high density by a transition first to the type III ELM regime and then to the L-mode. The confinement penalty, relative to good type I ELM discharges, of operating with type III ELMs is about 25-30%. The maximum densities for operation with type I or type III ELMs can be substantially increased by increasing the plasma triangularity in both machines.

We review the problem of plasma penetration of astrophysical magnetospheres, focusing on close binary stellar systems in which the Roche lobe overflow occurs. We point out that the accretion stream can penetrate the magnetosphere, even if its ram pressure is smaller than the magnetic pressure. This can occur without any turbulent mixing postulated in the past. The particular mechanism of cross-field penetration is an E × B drift caused by an electric field arising from polarization currents in the accretion stream. We determine the criteria for deep magnetosphere penetration; in particular, we find that deep penetration results for narrow streams (such as suggested by the previous theoretical analyses), and for reasonably low ambient plasma densities and temperatures.

Magnetohydrodynamic (MHD) activity and energy transport at rational-q surfaces is analysed on the basis of experimental results on current density profile control obtained with localized electron cyclotron resonance heating (ECRH) on FTU tokamak. The MHD response, in particular 2/1 and 1/1 modes, to ECRH is in agreement with expectations from a theoretical model including resistive wall braking and toroidal mode coupling. It is also shown that the magnetic shear at r_{q = 1} could control m = 1 mode saturation and magnetic reconnection. Heating results with ECRH at steady state indicate that transport enhancement is the dominant effect on confinement at the q = 2 surface, and suggest that conduction and convection inside the asymmetric m = 1 island should both be taken into account for a proper description of the thermal response to localized ECRH.

Many tokamak plasma regimes have been developed which exhibit improved plasma stability and energy confinement compared with the ELMy H-mode which has been used for some time as the baseline for reactor design. In particular, transient high performance has been achieved in experiments with low or reversed magnetic shear (s = r/q(dq/sr)) and substantial progress is now being made to extend these plasmas towards steady state. The achievements with regime in several tokamak devices are presented and outstanding issues for the further exploitation of this scenario are discussed.

The ITER design activities are now focused on the development of a design for a tokamak burning plasma experiment which will achieve Q = 10 in long-pulse, inductively driven plasmas and which aims to demonstrate steady-state operation at Q = 5 using non-inductive current drive. The device is designed to allow operational flexibility in plasma equilibrium and operating scenarios, and it is intended that the possibility of ignited operation under favourable plasma scenarios should not be excluded. A final design and parameter set has not yet been selected, but considerable progress has been made in developing a series of candidate designs which allow the critical physics and engineering issues to be explored. Here the physics underlying the design options and performance projections, as well as the physics issues to be addressed in the proposed experiment, are analysed.