Table of contents

Injection of active species (scavengers) in divertor plasmas has been proposed as a possible technique to inhibit the formation of tritium-rich carbon films in hidden regions under carbon-based ITER scenarios. To date, molecular nitrogen injection is the best-documented example of this concept. In this paper, the results recently obtained in ASDEX-Upgrade of nitrogen injection in the divertor are presented. The deposition rates at the sub-divertor region are reduced by a factor of up to five in nitrogen-seeded plasmas, as measured by the quartz microbalance diagnostic. No perturbation in the divertor plasma parameters for the chosen injection conditions is observed. Mass spectra at several toroidal locations indicate that conversion of C₁ into C₂ hydrocarbons takes place near the injection port in the presence of nitrogen, in agreement with earlier observations made in laboratory plasmas.

Saturated internal kink modes have been observed in many of the highest toroidal β discharges of the National Spherical Torus Experiment (NSTX). These modes often cause rotation flattening in the plasma core, can degrade fast-ion confinement and in some cases contribute to the complete loss of plasma angular momentum and stored energy. Characteristics of the modes are measured using soft x-ray, kinetic profile and magnetic diagnostics. Toroidal flows approaching Alfvénic speeds, island pressure peaking and enhanced viscous and diamagnetic effects associated with high-β may contribute to mode nonlinear stabilization. These saturation mechanisms are investigated for NSTX parameters and compared with experimental data.

Active feedback control of resistive wall modes (RWMs) has been demonstrated in the EXTRAP T2R reversed-field pinch experiment. The control system includes a sensor consisting of an array of magnetic coils (measuring mode harmonics) and an actuator consisting of a saddle coil array (producing control harmonics). Closed-loop (feedback) experiments using a digital controller based on a real time Fourier transform of sensor data have been studied for cases where the feedback gain was constant and real for all harmonics (corresponding to an intelligent-shell) and cases where the feedback gain could be set for selected harmonics, with both real and complex values (targeted harmonics). The growth of the dominant RWMs can be reduced by feedback for both the intelligent-shell and targeted-harmonic control systems. Because the number of toroidal positions of the saddle coils in the array is half the number of the sensors, it is predicted and observed experimentally that the control harmonic spectrum has sidebands. Individual unstable harmonics can be controlled with real gains. However if there are two unstable mode harmonics coupled by the sideband effect, control is much less effective with real gains. According to the theory, complex gains give better results for (slowly) rotating RWMs, and experiments support this prediction. In addition, open loop experiments have been used to observe the effects of resonant field errors applied to unstable, marginally stable and robustly stable modes. The observed effects of field errors are consistent with the thin-wall model, where mode growth is proportional to the resonant field error amplitude and the wall penetration time for that mode harmonic.

High-Z plasmas have been produced with Ar and/or Ne gas fuelling to increase the ion temperature in Large Helical Device (LHD) plasmas heated with high-energy negative-ion-based neutral beam injection (NBI). Although the electron heating is dominant in the high-energy NBI heating, the direct ion heating power is significantly enhanced in low-density plasmas due to both an increase in the beam absorption (ionization) power and a reduction of the ion density in the high-Z plasmas. Intensive neon- and/or argon-glow discharge cleaning works well to suppress dilution of the high-Z plasmas with wall-absorbed hydrogen. As a result, the ion temperature increases with an increase in the ion heating power normalized by the ion density and reaches 10 keV. An increase in the ion temperature is also observed with the addition of centrally focused electron cyclotron resonance heating to a low-density and high-Z NBI plasma, suggesting improvement of the ion transport. The results obtained in the high-Z plasma experiments with high-energy NBI heating suggest that an increase in the direct ion heating power and improvement of the ion transport are essential to ion temperature rise, and that a high-ion temperature could be obtained as well in hydrogen plasmas with low-energy positive-NBI heating which is planned in the near future in the LHD.

This paper reports the results of the progress in stationary discharges with a large bootstrap current fraction in JT-60U towards steady-state tokamak operation. In the weak shear plasma regime, high-β_p ELMy H-mode discharges have been optimized under nearly full non-inductive current drive conditions by the large bootstrap current fraction (f_BS ∼ 45%) and the beam driven current fraction (f_BD ∼ 50%), which was sustained for 5.8 s in the stationary condition. This duration corresponds to ∼26τ_E and ∼2.8τ_R, which was limited by the pulse length of negative-ion-based neutral beams. The high confinement enhancement factor H₈₉ ∼ 2.2 (HH_98y2 ∼ 1.0) was obtained and the profiles of current and pressure reached the stationary condition. In the reversed shear plasma regime, a large bootstrap current fraction (f_BS ∼ 75%) has been sustained for 7.4 s under nearly full non-inductive current drive conditions. This duration corresponds to ∼16τ_E and ∼2.7τ_R. The high confinement enhancement factor H₈₉ ∼ 3.0 (HH_98y2 ∼ 1.7) was also sustained, and the profiles of current and pressure reached the stationary condition. The large bootstrap current and the off-axis beam driven current sustained this reversed q profile. This duration was limited only by the duration of the neutral beam injection.

The hypothesis that improved H-modes result from reduction of transport where low-order rational surfaces are sparse is investigated with the integrated core pedestal SOL model. A function which expresses this sparseness is defined and a strength of reduction is determined which reproduces the improvement in confinement observed in Asdex-UG. The same dependence then agrees approximately with JET results. In application to ITER, the improved confinement gives improved fusion performance only if additional neoclassical impurity accumulation can be reduced or eliminated.

Magnetic islands with the toroidal mode number n = 1, e.g. m/n = 1/1 and 2/1 islands, in a Large Helical Device (LHD) equilibrium are studied using the three-dimensional MHD equilibrium code, HINT. In order to accomplish this purpose, the HINT code has been improved. The equilibrium analysis, in particular an analysis of the LHD equilibrium with an m/n = 1/1 island, is required for the local island divertor experiment, in order to understand the magnetic structures of field lines, i.e. flux surfaces, islands and ergodic field lines. We find that the m/n = 2/1 island can be healed for a finite equilibrium beta, while the m/n = 1/1 island is not healed and is surrounded with ergodic field lines for finite-β. From the latter result, we can conjecture that the island divertor concept is effective even for finite equilibrium beta-values, but the performance of the island divertor is deteriorated for finite-β because of the existence of the ergodic zone between the closed surfaces (i.e. the core region) and the m/n = 1/1 island. We also find that the width of the m/n = 1/1 island depends on the equilibrium beta value and that the island located at the inside of the torus has the advantage of retaining its width.

Large sub-millisecond heat pulses due to Type-I edge localized modes (ELMs) have been eliminated reproducibly in DIII-D for periods approaching nine energy confinement times (τ_E) with small dc currents driven in a simple magnetic perturbation coil. The current required to eliminate all but a few isolated Type-I ELM impulses during a coil pulse is less than 0.4% of plasma current. Based on magnetic field line modelling, the perturbation fields resonate with plasma flux surfaces across most of the pedestal region (0.9 ⩽ ψ_N ⩽ 1.0) when q₉₅ = 3.7 ± 0.2, creating small remnant magnetic islands surrounded by weakly stochastic field lines. The stored energy, β_N, H-mode quality factor and global energy confinement time are unaltered by the magnetic perturbation. Although some isolated ELMs occur during the coil pulse, long periods free of large Type-I ELMs (Δt > 4–6 τ_E) have been reproduced numerous times, on multiple experimental run days in high and intermediate triangularity plasmas, including cases matching the baseline ITER scenario 2 flux surface shape. In low triangularity, lower single null plasmas, with collisionalities near that expected in ITER, Type-I ELMs are replaced by small amplitude, high frequency Type-II-like ELMs and are often accompanied by one or more ELM-free periods approaching 1–2 τ_E. Large Type-I ELM impulses represent a severe constraint on the survivability of the divertor target plates in future burning plasma devices. Results presented in this paper demonstrate that non-axisymmetric edge magnetic perturbations provide a very attractive development path for active ELM control in future tokamaks such as ITER.

This paper describes the present state of the development of the computational model of the ITER edge plasma. Neutral–neutral collisions and molecular dynamics are introduced into the self-consistent scheme. First results of ITER modelling including these effects indicate that the operational window for the ITER divertor shifts towards higher neutral pressure in the private-flux region, retaining the operational flexibility determined in the previous analyses.

A stationary regime with improved confinement (H_{98(y, 2)} > 1) and, simultaneously, improved stability (β_N > 2.5) compared to standard H-mode has been investigated on ASDEX Upgrade for many years. This so-called 'improved H-mode' is characterized by a q-profile with low central magnetic shear and q₀ ⩾ 1 that is obtained by early heating during the current ramp. Studies of this discharge scenario have been continued. New results are presented concerning the existence domain in the q₉₅ range, the dependence on the normalized Larmor radius and initial experiments showing that high performance improved H-modes can be obtained with strong central ion cyclotron resonance heating (ICRH). In addition, the present status of understanding of the improved H-mode is reviewed. Improved H-mode plasmas are documented for 3.2 < q₉₅ < 4.5 and n_e/n_GW up to 0.85. At low densities the normalized collisionality ν^* is close to the ITER value and within the range of , accessible in ASDEX Upgrade, no dependence of the performance on this parameter was found. When compared with standard H-mode, core transport is still governed by drift-wave turbulence. Impurity control is achieved by central wave heating. The improved stability is due to the q-profile: by q₀ > 1 sawteeth as the main trigger of large amplitude neoclassical tearing modes are avoided. The stability is eventually limited by the occurrence of a mode with poloidal mode number m = 2 and toroidal mode number n = 1 at typically β_N ∼ 3. As far as the reactor relevance of this regime is concerned, its compatibility with significant central electron heating by ICRH, with high edge densities and low amplitude edge localized modes is of importance. The improved H-mode is, therefore, seen as a candidate for a long pulse ITER 'hybrid' operation.

In 2003, the performance of the 'hybrid' regime was successfully validated in JET experiments up to β_N = 2.8 at low toroidal field (1.7 T), with plasma triangularity and normalized Larmor radius (ρ^*) corresponding to identical ASDEX Upgrade discharges. Stationary conditions have been achieved with the fusion figure of merit ( ) reaching 0.42 at q₉₅ = 3.9. The JET discharges show similar MHD, edge and current profile behaviour, when compared with the ASDEX Upgrade. In addition, the JET experiments have extended the hybrid scenario operation at higher toroidal field of 2.4 T and lower ρ^* towards the projected ITER values. Using this database, transport and confinement properties are characterized with respect to the standard H-mode regime. Moreover, trace tritium has been injected to assess the diffusion and convective coefficients of the fusion fuel. The maximization of confinement and stability properties provides, to this scenario, a good probability of achieving a high fusion gain at reduced plasma current for durations of up to 2000 s in ITER.

The primary mission of the international fusion material irradiation facility (IFMIF) is to generate a material database to be used for the design of various components, for the licensing and for the assessment of the safe operation of a demonstration fusion reactor. IFMIF is an accelerator-based high-energy neutron source whose irradiation volume is quite limited (0.5 litre for the high fluence volume). This requires the use of small specimens to measure the irradiation-induced changes in the physical and mechanical properties of materials. In this paper, we have developed finite element models to better analyse the results obtained with two different small specimen test techniques applied to the tempered martensitic steel F82H-mod. First, one model has been used to reconstruct the load-deflection curves of small ball punch tests (SPTs), which are usually used to extract standard tensile parameters. It has been shown that a reasonable assessment of the overall plastic flow can be done with SPTs. Second, we have investigated the stress field sensitivity at a crack tip to the constitutive behaviour, for a crack modelled in plane strain, small-scale yielding and fracture mode I conditions. Based upon a local criterion for cleavage, which appears to be the basis to account for the size and geometry effects on fracture toughness, we have shown that the details of the constitutive properties play a key role in modelling the irradiation-induced fracture toughness changes. Consequently, we suggest that much more attention has to be paid and efforts made to investigate the post-yield behaviour of the irradiated specimens and, in order to reach this goal, we recommend the use of not only tensile specimens but also of compression ones in the IFMIF irradiation matrices.

The importance of electron cyclotron (EC) wave emission to the local electron power balance is analysed for various ITER operation regimes and, for comparison, for typical working conditions of FIRE, IGNITOR and the reactor-grade ITER concept as considered during the Engineering Design Phase (ITER-EDA). To cover the non-local effects in EC wave emission as well, the CYTRAN routine along with the ASTRA transport code is used. As a result, EC wave emission is shown to be a significant contributor to core electron cooling if the core electron temperature is about 35 keV or higher, as expected for ITER and tokamak reactor steady-state operation; in fact, it becomes the dominant core electron cooling mechanism at temperatures exceeding 40 keV, as such affecting the core plasma power balance in an important way.

The development of first wall, blanket and divertor materials that are capable of withstanding for many years high neutron and heat fluxes, is a critical path to fusion power. Therefore, the timely availability of a sound materials database has become an indispensable element in international fusion road maps. In order to provide materials design data for the short-term needs of ITER test blanket modules (TBMs) and for a DEMOnstration fusion reactor, a wealth of R&D results on the European reduced activation ferritic–martensitic steel EUROFER, and on oxide dispersion strengthened variants are being characterized, mainly in the temperature window 250–650°C. The characterization includes irradiations up to 15 dpa in the mixed spectrum reactor HFR and up to 75 dpa in the fast breeder reactor BOR60. Industrial EUROFER-batches of 3.5 and 7.5 tons have been produced with a variety of semi-finished, quality-assured product forms. To increase the thermal efficiency of blankets, high temperature resistant SiC_f/SiC channel inserts for liquid metal coolant tubes are also developed. Regarding radiation damage resistance, broad based reactor irradiation programmes deal with several ranges from ≤ 5 dpa (ITER TBMs) up to 75 dpa (DEMO). For the European divertor designers, a materials database is currently being set up for pure W and its alloys, and related reactor irradiations are foreseen with temperatures from 650°C to 1000°C.

In order to experimentally evaluate the tungsten armour impact on tritium production of the solid breeding blanket being developed by JAERI for tokamak-type DEMO reactors, neutronics integral experiments have been performed using DT neutrons at the Fusion Neutron Source facility of JAERI. Solid breeding blanket mockups relevant to the DEMO blanket have been applied in this study. The mockups are made of a set of layers consisting of 0–25.2 mm thick tungsten, 16 mm thick F82H, 12 mm thick Li₂TiO₃ and 100–200 mm thick beryllium with a cross-section of 660 × 660 mm in maximum. Pellets of Li₂CO₃ are embedded in the Li₂TiO₃ layers to measure the tritium production rate. By installing the 5 mm, 12.6 mm and 25.2 mm thick tungsten armours, the sum of the integrated tritium productions at the pellets are reduced by about 2.1%, 2.5% and 6.1% relative to the case without the armour, respectively. Numerical calculations have been conducted using the Monte Carlo code. In the case of the mockups with the tungsten armour, calculation results for the sum of the integrated tritium productions agree well with the experimental data within 4% and 19% in the experiments without and with a neutron reflector, respectively.

Comprehensive neutronic analyses using the MCNP Monte Carlo code have been performed to predict the neutronics performance of the EU water-cooled lithium–lead test blanket module (TBM) integrated in ITER-FEAT. The analyses have been performed using models representing the complex ITER and TBM structure close to reality. Tritium generation and parameters relevant to the lifetime performance of the TBM such as helium, hydrogen and atomic displacement production have been calculated. Assuming 90% ⁶Li enrichment and 22% duty cycle, the estimated tritium production amounts to 3.8738 × 10¹⁶ T atoms s⁻¹ (16.62 mg d⁻¹). The maximum radiation damage obtained, based on 0.3 MWa m⁻² neutron fluence, is in the first wall first steel layer of the breeder container totalling 2.77 dpa of Fe in Euro steel, 19.3 appm He production and 103 appm H generation. The calculated helium production in the demountable hydraulic connections of TBM is below the required limit for reweldability of steel structure.

Low-energy deuterium (D) plasma exposure on tungsten (W), which is an important material for ITER, results in blister formation on the surface. Blister formation increases both micron-sized dust production and D retention. Blister formation depends greatly on surface pre-treatment. Deuterium plasma exposure on mirror-finished powder metallurgy W at 500 K for 3 h forms a blister with a diameter of a few hundred micrometres on the surface. Blister formations on the mechanically-polished and helium-pre-exposed surfaces are drastically suppressed. Deuterium retention is also reduced on both the mechanically-polished surface and the helium-pre-exposed surface compared with that on the mirror-finished surface. The suppressive effect of blister formation on the mechanically-polished surface is maintained for 50 h. The size of blisters and D retention on mirror-finished surface exposed for 50 h increases by some degrees (≲500 µm, 7 × 10²⁰ m⁻²) compared with that on mirror-finished surface exposed for 3 h (≲200 µm, 5 × 10²⁰ m⁻²), but is not proportional to the exposure time. The saturation level of D retention even on blister-rich surfaces seems to be lower than an order of 10²¹ D₂ m⁻².

A Green's function method is developed to evaluate the currents induced during startup in the vacuum vessel of ETE (Experimento Tokamak Esférico). The non-homogeneous integral equation for the axisymmetric eddy currents distribution is determined using a thin shell approximation for the vacuum vessel and local curvilinear coordinates. This problem is reduced to a circuit model by adopting spectral representations both for the centreline of the vacuum vessel and the surface current density. Results of this model are compared with the distribution of eddy currents measured in ETE.

Boronization of FTU tokamak strongly affects its gas recycling, impurity concentration and its temporal evolution. Two main phases are distinguished. In the first one, lasting about 60 discharges, with the machine fully boronized, the radiation losses decrease from 80% to 40% and the impurity content drops from Z_eff = 6 to Z_eff ≈ 1.5 at in ohmic conditions. In the second phase, lasting >500 discharges, with the B film eroded from the TZM toroidal limiter surface but not from the SS walls, the reduction of radiated power (≈60%) and impurities (Z_eff ≈ 2.0) though smaller, is still important with respect to pre-boronization, allowing for optimization of plasma characteristics up to the maximum RF power so far injected (2.6 MW ≈ 1.6 MW m⁻³). The recycling properties of the B film are fully stabilized, making the plasma operations much more reliable, while the gettering properties are preserved and they maintain the oxygen below n_O ⩽ 0.5%.

Electron cyclotron heating (ECH)-assisted startup experiments have been performed in JT-60U. The breakdown loop voltage was successfully reduced from 25 to 4 V (=0.26 V m⁻¹) by 200 kW ECH. This is lower than the 0.3 V m⁻¹, which corresponds to the maximum electric field in ITER. Parameter scans of ECH power, prefill pressure, resonance position and polarization were carried out. The sensitivity of the breakdown to polarization and resonance position was observed. A prefilling gas pressure scan showed that the initial breakdown density increases with prefill pressure when it is is lower than 8 × 10⁻⁵ Torr. Higher harmonic ECH was also attempted. The second harmonic ECH-assisted startup was possible with higher ECH power injection. However, the third harmonic ECH-assisted startup was not successful.

Fast wave current drive (FWCD) experiments have been performed in JET plasmas with electron internal transport barriers produced with LHCD. Because of a large fraction of parasitic absorption, owing to weak single pass damping, the inductive nature of the plasma current and the interplay between the RF-driven current and the bootstrap current only small changes are seen in the central current profiles. The measured difference in the central current density for co- and counter-current drive is larger than the response expected from current diffusion calculations, but smaller than the driven currents, suggesting a faster current diffusion than that given by neo-classical resistivity. A large fraction of the power is absorbed by cyclotron damping on residual ³He ions while a significant fraction appears not to have been deposited in the plasma. The strong degradation of heating and current drive occurs simultaneously with strong increases in the Be II and C IV line intensities in the divertor. The degradation depends on the phasing of the antennas and increases with reduced single pass damping which is consistent with RF-power being lost by dissipation of rectified RF-sheath potentials at the antennas and walls. Asymmetries in direct electron heating, lost power and production of impurities, fast ions and gamma-rays are seen for co- and counter-current drive. These differences are consistent with the differences in the absorption on residual ³He ions owing to the RF-induced pinch. Effective direct electron heating, comparable to the indirect electron heating with H-minority heating, occurs for dipole phasing of the antennas without producing a significant fast ion pressure and with low impurity content in the divertor plasma.

The stationary edge localized mode (ELM)-free 'Quiescent H-mode' (QH-mode) regime, obtained with counter neutral beam injection, is studied in ASDEX Upgrade and Joint European Torus. QH-mode plasmas have high pedestal and core ion temperatures together with good High-confinement mode (H-mode). ELMs are replaced by continuous MHD oscillations, the 'edge harmonic oscillation' (EHO) and the 'high frequency oscillation'. Stationarity of particle and impurity densities is linked to the occurrence of these MHD modes. The EHO location in the steep-gradient region and its appearance with increasing edge pressure points towards the edge pressure or pressure gradient as possible drivers for the EHO. Injection of small cryogenic pellets can raise the plasma density to about 40% of the Greenwald density limit without triggering ELMs. Orbit following calculations of the slowing-down distribution show the presence of an enhanced fast particle density in the H-mode barrier region despite the large loss currents with counter-injection. The radial electrical field in the edge barrier region, about twice as large in QH-mode as in ELMy H-mode, is large enough to reverse the precession drift direction of injected beam ions, leading to a resonance of the EHO with the drift precession frequency.

The MFBE procedure developed by Strumberger (1997 Nucl. Fusion37 19) is used to provide an improved starting point for free boundary equilibrium computations in the case of W7-X (Nührenberg and Zille 1986 Phys. Lett. A 114 129) using the Princeton iterative equilibrium solver (PIES) code (Reiman and Greenside 1986 Comput. Phys. Commun.43 157). Transferring the consistent field found by the variational moments equilibrium code (VMEC) (Hirshmann and Whitson 1983 Phys. Fluids26 3553) to an extended coordinate system using the VMORPH code, a safe margin between plasma boundary and PIES domain is established. The new EXTENDER_P code implements a generalization of the virtual casing principle, which allows field extension both for VMEC and PIES equilibria. This facilitates analysis of the 5/5 islands of the W7-X standard case without including them in the original PIES computation.

A continuum global gyrokinetic code GYRO has been developed to comprehensively simulate core turbulent transport in actual experimental profiles and enable direct quantitative comparisons to the experimental transport flows. GYRO treats not only the now standard ion temperature gradient (ITG) mode turbulence but also trapped and passing electrons with collisions and finite β, equilibrium E × B shear stabilization and all in real tokamak geometry. Most importantly, the code operates at finite relative gyroradius (ρ_*) so as to treat the profile shear stabilization and nonlocal effects which can break gyroBohm scaling. The code operates in either a cyclic flux-tube limit (which allows only gyroBohm scaling) or globally with a physical profile variation. Bohm scaling of DIII-D L-mode has been simulated with power flows matching experiment within error bars on the ITG. Mechanisms for broken gyroBohm scaling, neoclassical ion flows embedded in turbulence, turbulent dynamos and profile corrugations, plasma pinches and impurity flow and simulations at fixed flow rather than fixed gradient are illustrated and discussed.

This paper deals with the problem of helium ash removal from stellarator-reactor. The lower hybrid (LH) heating of ash ions is proposed to solve this problem. The theory of ion stochastic heating, developed earlier by Karney, is generalized in the case of heating in stellarators. The features of the LH waves propagation and the ions motion in the stellarator confining field are taken into account. With a proper choice of wave parameters (such as frequency, antenna position and initial spectrum of longitudinal refractive index) the slow mode of LH waves penetrates from the launching system to the plasma core (and back) without conversion to kinetic plasma mode or to fast mode. With all this going on, the LH wave is absorbed by alpha particles only. The electron Landau damping is negligibly small, and there is no bulk ions stochastic heating. The motion of high-energy (>100 keV) ions in the LHD heliotron with inwardly shifted magnetic axis, as an example of stellarator type device, is calculated numerically using the single particle simulation code which couples modified Karney's ion stochastic heating theory. The effect of collisions was taken into account through the Monte Carlo equivalent of the Lorentz collision operator. It is shown, that due to interaction with LH wave, initially well-confined alpha particles are expelled from the plasma volume during the time period which is shorter than the collision time. At the same time, the low hybrid heating does not remove the ions having an energy higher than 500 keV. Therefore, it is possible to use this method of RF heating for helium ash removal in stellarator-reactor. The required LH power is estimated to be of the order of 10 MW.