Table of contents

Recent 2D spatially and temporally resolved measurements of electron temperature fluctuations in the tokamak core have revealed new information on the dynamics of the sawtooth crash. Measures of poloidal localization of the reconnection zone are achieved through direct analysis of the 2D data and through an interpolated projection of the q = 1 region. An estimate of the toroidal coverage of the reconnection zone is achieved through analysis of shots exhibiting toroidal rotation in which the helically localized reconnection occurs both within and outside of the view of the measurement window. The localized trigger of the crash instability exhibits a dominant toroidal mode number of n ≈ 3, and a dominant poloidal mode number of m ≥ 10.

The effects of applied three-dimensional magnetic error fields on magnetic surfaces are measured in both the Joint European Torus (JET) and the Mega Ampère Spherical Tokamak (MAST). Static perturbation fields with various amplitudes have been applied in both devices using non-axisymmetric coils producing a dominantly n = 1 magnetic field to systematically perturb the plasma flux surfaces. The displacement of the plasma boundary is found to increase linearly with the strength of the applied field, and agrees well with predictive modelling which superimposes vacuum non-axisymmetric fields on a 2D equilibrium.

It is shown that relaxation oscillations associated with repetitive internal transport barrier (ITB) buildup and collapse in high-performance, overdriven tokamak plasmas, with ion-cyclotron resonance heating (ICRH), neutral-beam injection (NBI) and lower-hybrid current drive (LHCD), and with a dominant fraction of bootstrap current, can be overcome if the LHCD power is sufficiently high. This result has been obtained using a benchmarked, fully predictive transport model iterated with given ICRH profiles and self-consistently with NBI and LHCD modules, the stabilizing role of the E × B flow shear being combined with that of reversed magnetic shear in the simulation of ITB dynamics.

The experimental results of low pressure supersonic molecular beam injection (SMBI) fuelling on the HL-2A closed divertor indicate that during the period of pulsed SMBI the power density convected at the target plate surfaces was 0.4 times of that before or after the beam injection. An empirical scaling law used for the SMBI penetration depth for the HL-2A plasma was obtained. The cluster jet injection (CJI) is a new fuelling method which is based on and developed from the experiments of SMBI in the HL-1M tokamak. The hydrogen clusters are produced at liquid nitrogen temperature in a supersonic adiabatic expansion of moderate backing pressure gases into vacuum through a Laval nozzle and are measured by Rayleigh scattering. The measurement results have shown that the averaged cluster size of as large as hundreds of atoms was found at the backing pressures of more than 0.1 MPa. Multifold diagnostics gave coincidental evidence that when there was hydrogen CJI in the HL-2A plasma, a great deal of particles from the jet were deposited at a terminal area rather than uniformly ablated along the injecting path. SMB with clusters, which are like micro-pellets, will be of benefit for deeper fuelling, and its injection behaviour was somewhat similar to that of pellet injection. Both the particle penetration depth and the fuelling efficiency of the CJI were distinctly better than that of the normal SMBI under similar discharge operation. During hydrogen CJI or high-pressure SMBI, a combination of collision and radiative stopping forced the runaway electrons to cool down to thermal velocity due to such a massive fuelling.

Minimization of the external heating power to access self-ignition is advantageous to increase the reactor design flexibility and to reduce the capital and operating costs of the plasma heating device in a helical reactor. In this work we have discovered that a larger density limit leads to a smaller value of the required confinement enhancement factor, a lower density limit margin reduces the external heating power and over 300 s of the fusion power rise-up time makes it possible to reach a minimized heating power. While the fusion power rise-up time in a tokamak is limited by the OH transformer flux or the current drive capability, any fusion power rise-up time can be employed in a helical reactor for reducing the thermal stresses of the blanket and shields, because the confinement field is generated by the external helical coils.

High densities exceeding the Greenwald limit by a factor of 1.7 have been obtained in L-mode discharges with high internal inductances of ℓ_i as high as 2.8 in JT-60U. The internal inductance is controlled by ramping down the plasma current. In addition to the extension of the operational regime limited by disruptions, confinement performance remains as good as an H89PL factor of 1.6 even above the Greenwald limit. While an earlier high ℓ_i study has indicated core confinement improvement due to enhancement of the poloidal field, the additional improvement of the tolerance against the high density turned out to be correlated with high edge temperature. The normalized density when the detachment occurs, characterized by a decrease in the Dα signal at the divertor, is even higher in the case with no disruption than in the case with a disruption. These comparisons have indicated that the improvement in thermal and particle transport does exist in the periphery and in the edge in high ℓ_i plasmas, and the shift of the density limit towards higher densities is observed coincidently. Although the high ℓ_i discharge studied here lies outside the usual parameter space for steady-state operation of a tokamak, demonstration of a stable discharge with good confinement beyond the Greenwald limit suggests that the magnetic shear at the edge is one key parameter to uncover the physical elements of the operational density limit.

Due to their discrete nature, a grid of active coils for the feedback control of plasma instabilities produces an infinite sequence of sideband harmonics in the magnetic field. If the magnetic sensors have the same periodicity as the coils, as in the intelligent shell scheme (Bishop 1989 Plasma Phys. Control. Fusion31 1179), the aliasing of the sidebands determines a systematic error on the Fourier analysis of the measurements. Since the intelligent shell scheme relies on raw measurements, it does not recognize the aliasing effect. This is a drawback for error field compensation and for the control of those perturbations that cannot be suppressed by the feedback, but only reduced in their saturation level, such as the non-linear tearing modes (TMs) involved in the dynamo process of reversed field pinch machines. We have derived analytical formulae, valid in cylindrical geometry, for the subtraction of the sideband effect, and we implemented them in a real-time correction algorithm of the Fourier analysis of the magnetic measurements. The Fourier harmonics so obtained are used as the feedback variable of a new control scheme named clean-mode-control (CMC). The first tests of CMC in RFX-mod have given some interesting results in the control of the saturated TMs: besides a reduction of the radial field at the edge and the consequent plasma surface distortion, a systematic rotation of these perturbations with frequencies of up to 100 Hz is seen for the first time. This brings a mitigation of the phase-locking and wall-locking phenomena giving the possibility of operating safely at high currents.

This contribution summarizes a number of aspects of the experimental and modelling programme at JET aimed at improving the characterization and understanding of edge localized mode (ELM) transport in the scrape-off layer (SOL). Divertor target energy deposition asymmetries favouring the inner target for the ion B × ∇B drift directed towards the X-point are observed with infra-red (IR) thermography. Similar trends are seen in the ELM resolved energy radiated in the divertor volume. Particle-in-cell kinetic calculations of the parallel ELM heat transport have been made for a range of ELM energies, revealing the detailed time response of target sheath heat transmission factors and indicating that electrons deposit ∼30% of the ELM energy. The simulation results are in good agreement with experimental measurements of the integral energy deposited at the outer target up to the peak in target heat loads. A transient model of ELM filament energy evolution has been developed at JET and is able to reproduce a number of experimental observations, including the high ion energies observed in the far SOL using an electrostatic retarding field electrostatic analyser (RFA) and estimates of ELM heat fluxes deposited on main chamber limiters. During the ELM, the RFA and a second, SOL turbulence probe, clearly show the presence of coherent spikes in the hot ion flux, the plasma flux and the electron temperature. Field aligned structures have also been seen for the first time on JET in the power deposition on main wall limiters and upper dump plate surfaces using a new wide angle IR camera system. The probe signals are interpreted as the arrival of interspaced plasma filaments, with successive filaments carrying less energy. They are also consistent with the ELM out flux entering the SOL primarily on the outboard side and launching a sound wave disturbance along field lines.

Predictions of the recently developed paleoclassical transport model are compared with data from many toroidal plasma experiments: electron heat diffusivity in DIII-D, C-Mod and NSTX ohmic and near-ohmic plasmas; transport modelling of DIII-D ohmic-level discharges and of the RTP ECH 'stair-step' experiments with electron internal transport barriers (eITBs) at low order rational surfaces; investigation of a strong eITB in JT-60U; H-mode T_e edge pedestal properties in DIII-D; and electron heat diffusivities in non-tokamak experiments (NSTX/ST, MST/RFP, SSPX/spheromak). The radial electron heat transport predicted by the paleoclassical model is found to be in reasonable agreement with a wide variety of ohmic-level experimental results and to set the lower limit (within a factor ≲2 in tokamaks) on the radial electron heat transport in most resistive, current-carrying toroidal plasmas—for where it is expected to be dominant over fluctuation-induced anomalous transport that scales with a gyro-Bohm diffusion coefficient.

Volume recombination of C⁴⁺ and e⁻ into C³⁺ is observed for the first time in detached divertor plasmas with an X-point MARFE. The recombination zone is located around the X-point, and the electron temperature and density are evaluated to be 6.3 eV and 7.8 × 10²⁰ m⁻³, respectively. In this zone, the volume recombination flux is larger by two orders of magnitude than the ionization flux of C³⁺. However, the radiation power due to the recombination process is only 2% of the total radiation power, measured by a bolometer. In contrast, the radiation power due to the excitation process from the ground state of C³⁺ by electron collision dominates the total radiation power.

The particle and energy transport in reversed field pinch experiments is affected by the locking in phase of the tearing modes, also dubbed dynamo modes, that sustain the magnetic configuration. In standard RFP pulses many m = 1 and m = 0 resonant modes have a relatively large amplitude (a spectrum dubbed MH for multiple helicity). The locking in phase of m = 1 tearing modes produces a helical deformation (locked mode (LM)) of the magnetic surfaces in a region of approximately 40 toroidal degrees. The region of the LM is characterized by a strong plasma–wall interaction and by high losses of energy and particles that account for a significant fraction of the input power and of the total particle outflux. The locking in phase of m = 0 modes modifies the plasma radius, shrinking and enlarging the plasma cross section in two wide toroidal regions of about 100°. The purpose of this paper is to investigate to what extent the locking in phase of m = 0 modes introduces toroidal asymmetries in the transport properties of the plasma. This study has been carried out investigating the shape of the density profile in the RFX-mod experiment. The analyses show that the profile exhibits a dependence on the toroidal angle, which is related to the deformation of the plasma column due to the locking in phase of m = 0 modes: the least steep density gradients at the edge are found in the region where the plasma column is shrunk, entailing that in this region the particle transport is enhanced. An analogous asymmetry also characterizes the density and magnetic fluctuations at the edge, which are enhanced in the same toroidal region where the particle transport also is enhanced. This result can be considered the first experimental evidence of an instability localized where the plasma column is shrunk.

Current decays after disruptions as well as after noble gas injections in tokamaks are examined. As is shown, the cooled plasmas at the stage of current decay are partially opaque for radiation in lines giving the main impact into total thermal losses. The thermal balance is supposed to be determined by Ohmic heating and radiative losses. A zero-dimensional model for radiation losses and temperature distribution over minor radius is used. Plasma current evolution is simulated with DIMRUN and DINA codes. Impurity distribution over ionization states is calculated from the time-dependent set of differential equations. The opacity effects are found to be most important for simulation of JET disruption experiments with beryllium- and carbon-seeded plasmas. The decay times calculated are in good agreement with the experimental values. Current decays in beryllium-, carbon-, argon- and neon-seeded plasmas for ITER parameters are simulated. The temperatures after thermal quench are shown to be significantly higher in comparison with the model of transparent plasmas. Opacity effects are found to be most important for Be- and C-seeded plasmas. Runaway electron currents are damped significantly if opacity effects are taken into account in any case examined.

Since 1998 ASDEX Upgrade has developed stationary H-modes that routinely obtain confinement enhancement factors H_98(y,2) > 1 and normalized beta, β_N = 2–3. These discharges are characterized by a q-profile with low magnetic shear in the centre and q(0) ∼ 1. New results presented here concentrate on extending the operational range of these improved H-modes at ASDEX Upgrade and extrapolating the results to ITER. Discharges are obtained at high density, over a wide range of plasma collisionality and with a first wall predominantly covered by tungsten coated carbon tiles. The performance is optimized for q₉₅ ranging from 3 to 5. At q₉₅ ∼ 3 real time control of β_N is used and in some cases ECCD to suppress NTM activity at low β_N ∼ 2. For the extrapolation to ITER, the fusion power is calculated using the same thermal beta (β_N,th) and kinetic profile shapes as obtained in ASDEX Upgrade and setting ⟨n_e⟩/n_GW = 0.85. The fusion gain that could be obtained is evaluated using different confinement scaling expressions. The results indicate that improved H-modes are a candidate for an ITER hybrid scenario or could extend ITER operation beyond what is currently foreseen using standard H-modes.

Impact of the electron cyclotron range of frequency wave (ECRF) on the internal transport barriers (ITBs) in a weak shear (WS) plasma has been investigated in JT-60U. The fundamental O-mode ECRF of 110 GHz injected obliquely (co-current drive) from the low field side is used. It is observed that the ion temperature (T_i) ITB in a WS plasma can be degraded by ECRF. It is clarified for the first time that the degradation depends increasingly on the EC power (P_EC) but decreasingly on the plasma current (I_p). Moreover it is confirmed that ECRF affects the toroidal rotation (V_t) indirectly and results in the flattening of V_t(ρ) and therefore the radial electric field (E_r) profiles regardless of the direction of the target V_t(ρ), peaking co or counter direction (relative to the I_p direction). Furthermore, it is recently found that T_i and V_t in the whole ITB region are affected with almost no delay from the EC onset even with off-axis EC deposition. These results indicate that EC injection unveiled a semi-global structure that characterizes T_i ITB in a WS plasma.

Controllability of plasmas with a large bootstrap current fraction (f_BS) has been investigated in JT-60U. Real time control logic for avoidance of collapses by pressure profile control through the toroidal rotation has been newly installed for the long sustainment of reversed magnetic shear plasmas. The new real time control logic can control the injection timing of the neutral beam based on the real time detection of the minimum value of safety factor (q_min) using the motional Stark effect diagnostic. Using the real time control logic, the weak reversed magnetic shear plasma with f_BS ∼ 70% is sustained for ∼8 s. In such a plasma, dynamic change in the current profile, especially q_min, which was induced by the change in the pressure profile at the internal transport barrier (ITB) through the rotation control, was observed. The response of q_min to the change in the ion temperature gradient at the ITB for large f_BS plasmas is stronger than that for smaller f_BS plasmas.

The design of the modification of JT-60U, JT-60SA has been optimized from the viewpoint of plasma performance, and operation regimes have been evaluated with the latest design. Upper and lower divertors with different geometries will be prepared for flexibility of the plasma shape, which will enable both low aspect ratio (A ∼ 2.65) and ITER shape (A = 3.1) configurations. The beam lines of negative-ion neutral beam injection will be shifted downwards by ∼0.6 m for the off-axis current drive (CD), in order to obtain a weak/reversed shear plasma, as well as having the capability of heating the central region. The feedback control coils along the openings in the stabilizing plate are found effective in suppressing the resistive wall mode and sustaining high β_N close to the ideal wall limit. Sustainment of plasma current of 3–3.5 MA for 100 s will be possible in ELMy H-mode plasmas with moderate heating power, β_N, and density within an available flux swing. It is also expected that higher β_N, high-density ELMy H-mode plasmas will be maintained for 100 s with higher heating power. The expected regime of full CD operation has been extended with upgraded heating and CD power. Full CD operation for 100 s with reactor-relevant high values of normalized beta and bootstrap current fraction (I_p = 2.4 MA, β_N = 4.3, f_BS = 0.69, , H_H98y2 = 1.3) is expected in a highly-shaped low-aspect-ratio configuration (A = 2.65).

The European fusion programme is 'reactor oriented' and it is aimed at the successive demonstration of the scientific, the technological and the economic feasibility of fusion power. The European Power Plant Conceptual Study (PPCS) has been a study of conceptual designs of five commercial fusion power plants and the main emphasis was on system integration. It focused on five power plant models which are illustrative of a wider spectrum of possibilities. They are all based on the tokamak concept and they have approximately the same net electrical power output, 1500 MWe. These span a range from relatively near-term, based on limited technology and plasma physics extrapolations, to an advanced conception.

The PPCS allows one to clarify the concept of DEMO, the device that will bridge the gap between ITER and the first-of-a-kind fusion power plant. An assessment of the PPCS models with limited extrapolations highlighted a number of issues that must be addressed to establish the DEMO physics and technological basis.

A dual functional lithium–lead (DFLL) test blanket module (TBM) concept has been proposed for testing in the International Thermonuclear Experimental Reactor (ITER) and the Experimental Advanced Superconducting Tokamak (EAST) in China to demonstrate the technologies of the liquid lithium–lead breeder blankets with emphasis on the balance between the risks and the potential attractiveness of blanket technology development. The design of DFLL-TBM concept has the flexibility of testing both the helium-cooled quasi-static lithium–lead (SLL) blanket concept and the He/PbLi dual-cooled lithium–lead (DLL) blanket concept. This paper presents an effective testing strategy proposed to achieve the testing target of SLL and DLL DEMO blankets relevant conditions, which includes three parts: materials R&D and small-scale out-of-pile mockups testing in loops, middle-scale TBMs pre-testing in EAST and full-scale consecutive TBMs testing corresponding to different operation phases of ITER during the first 10 years. The design of the DFLL-TBM concept and the testing strategy ability to test TBMs for both blanket concepts in sequence and or in parallel for both ITER and EAST are discussed.

Micro-turbulence and macro-magnetohydrodynamic (macro-MHD) instabilities can appear in plasma at the same time and interact with each other in a plasma confinement. The multi-scale-nonlinear interaction among micro-turbulence, double tearing instability and zonal flow is investigated by numerically solving a reduced set of two-fluid equations. It is found that the double tearing instability, which is a macro-MHD instability, appears in an equilibrium formed by a balance between micro-turbulence and zonal flow when the double tearing mode is unstable. The roles of the nonlinear and linear terms of the equations in driving the zonal flow and coherent convective cell flow of the double tearing mode are examined. The Reynolds stress drives zonal flow and coherent convective cell flow, while the ion diamagnetic term and Maxwell stress oppose the Reynolds stress drive. When the double tearing mode grows, linear terms in the equations are dominant and they effectively release the free energy of the equilibrium current gradient.

This paper reports on the first demonstration of electron Bernstein wave heating (EBWH) by double mode conversion from ordinary (O-) to Bernstein (B-) via the extraordinary (X-) mode in an over-dense tokamak plasma, using low field side launch, achieved in the TCV tokamak H-mode, making use of its naturally generated steep density gradient. This technique offers the possibility of overcoming the upper density limit of conventional EC microwave heating. The sensitive dependence of the O-X mode conversion on the microwave launching direction has been verified experimentally. Localized power deposition, consistent with theoretical predictions, has been observed at densities well above the conventional cut-off. Central heating has been achieved, at powers up to two megawatts. This demonstrates the potential of EBW in tokamak H-modes, the intended mode of operation for a reactor such as ITER.

A disruption prediction system, based on neural networks, is presented in this paper. The system is ideally suitable for on-line application in the disruption avoidance and/or mitigation scheme at the JET tokamak.

A multi-layer perceptron (MLP) predictor module has been trained on nine plasma diagnostic signals extracted from 86 disruptive pulses, selected from four years of JET experiments in the pulse range 47830–57346 (from 1999 to 2002).

The disruption class of the disruptive pulses is available. In particular, the selected pulses belong to four classes (density limit/high radiated power, internal transport barrier, mode lock and h-mode/l-mode).

A self-organizing map has been used to select the samples of the pulses to train the MLP predictor module and to determine its target, increasing the prediction capability of the system.

The prediction performance has been tested over 86 disruptive and 102 non-disruptive pulses. The test has been performed presenting to the network all the samples of each pulse sampled every 20 ms. The missed alarm rate and the false alarm rate of the predictor, up to 100 ms prior to the disruption time, are 23% and 1%, respectively.

Recent plasma configurations might present features different from those observed in the experiments used in the training set. This 'novelty' can lead to incorrect behaviour of the predictor. To improve the robustness and reliability of the system, a novelty detection module has been integrated in the prediction system, increasing the system performance and resulting in a missed alarm rate reduced to 7% and a false alarm rate reduced to 0%.

Two reproducible regimes of tokamak operation, with excitation or suppression of MHD activity can be obtained using a voltage-biased electrode inside the edge of the TCABR tokamak. The experiment was carried out adjusting the tokamak parameters to obtain two types of discharges: with strong or weak MHD activity, without biasing in both cases. The plasma current was adjusted to cover a range of safety factor from 2.9 up to 3.5, so that when biasing was applied the magnetic island (3,1) could interact with the edge barrier. The application of biasing in subsequent discharges of each type resulted in excitation or suppression of the MHD activity. The results show that the dominant modes are m = 2, n = 1 and m = 3, n = 1 for excitation and partial suppression, respectively. In both regimes a strong decrease in the radial electric field is detected with destruction of the transport barrier and of the improved confinement caused by different mechanisms. The measurements include temporal behaviour of edge transport, turbulence, poloidal electric and magnetic fields, edge density, radial electric fields and radial profile of H_α line intensity. The explanation of the excitation and suppression processes is discussed in the paper.

In JT-60U, erosion/deposition analyses of the plasma facing wall have shown that local carbon transport in the inboard direction was appreciable in addition to long-range transport. The total deposition and erosion rates in the divertor region were ∼1 × 10²¹ C atoms s⁻¹ and ∼−6 × 10²⁰ C atoms s⁻¹, respectively. About 40% of the deposition in the divertor region likely originates from the main chamber wall. At the plasma facing surfaces of the divertor region, the highest hydrogen concentration in the (H + D)/C ratio and the retention amount were found to be ∼0.13 and ∼1 × 10²³ atoms m⁻², respectively. In the plasma-shadowed area underneath the divertor region with a vacuum vessel baking temperature of 420 K, redeposited layers of ∼2 µm thickness were found with a high hydrogen concentration of ∼0.75 in (H + D)/C, which was nearly the same level as that observed in JET. Large deuterium retention was also observed at the main chamber wall covered with boron layers. Their H + D retention and (H + D)/C were ∼1 × 10²³ atoms m⁻² and ∼0.16, respectively, for the vacuum vessel temperature of 570 K. Such a high deuterium retention is most likely caused by D retained in the boron layers. Nevertheless, the integration of this retention over the whole main chamber wall results in significant inventory and needs further investigation.

The electromagnetic torque on the toroidal plasma is calculated assuming a linear plasma response to the applied perturbation, which may be the error field or the field created by the correction coils, or both. The result is compared with recently published expressions for the error field induced torque (Zheng et al2006 Nucl. Fusion46 L9, Zheng and Kotschenreuther 2006 Phys. Rev. Lett.97 165001), and the conclusions of these papers are revised. We resolve the problem of the torque resonance raised there. It is shown that the strong increase in the torque due to the static error field must occur at the resistive wall mode stability limit and not at the no-wall stability limit.

We discuss the processes underlying the excitation of fishbone-like internal kink instabilities driven by supra-thermal electrons generated experimentally by different means: electron cyclotron resonance heating (ECRH) and by lower hybrid (LH) power injection. The peculiarity and interest of exciting these electron fishbones by ECRH only or by LH only is also analysed. Not only is the mode stability explained, but also the transition between steady state nonlinear oscillations to bursting (almost regular) pulsations, as observed in FTU, is interpreted in terms of the LH power input. These results are directly relevant to the investigation of trapped alpha particle interactions with low-frequency MHD modes in burning plasmas: in fact, alpha particles in reactor relevant conditions are characterized by small dimensionless orbits, similarly to electrons; the trapped particle bounce averaged dynamics, meanwhile, depends on energy and not mass.

Deuterium pellets have been injected into plasmas in the DIII-D tokamak from the inner wall, top, and outer midplane port locations to investigate fuelling efficiency, mass deposition and interaction with edge localized modes (ELMs). Pellets injected from the outer midplane port (low field side (LFS)) show a large discrepancy in the mass deposition profile and fuelling efficiency from conventional pellet ablation theory, while the penetration depth compares favourably with theory. The mass deposition from pellets injected from inner wall and top locations is deeper than expected from ablation theory. The profile measurements indicate that pellet mass is deposited inside the measured penetration radius, thus verifying that a drift of the pellet ablatant is occurring in the major radius direction during the toroidal symmetrization process. The scaling of the measured drift magnitude in DIII-D is found to depend strongly on the pellet size and plasma pedestal temperature. Extrapolation to a burning plasma configuration on ITER is favourable for inner wall pellet fuel deposition depth well beyond the separatrix. Pellets injected into H-mode plasmas from all locations trigger ELMs with much larger ELM events induced by the outside midplane injected pellets. This suggests that the LFS is more sensitive to ELM triggering and may be the preferred location to inject very small pellets to trigger frequent small ELMs and thus minimize ELM induced damage to the divertor material surfaces.

The deuterium inventory in ASDEX Upgrade was determined by quantitative ion beam analysis techniques and SIMS for different discharge campaigns between the years 2002 and 2005. ASDEX Upgrade was a carbon dominated machine during this phase. Full poloidal sections of the lower and upper divertor tile surfaces, limiter tiles, gaps between divertor tiles, gaps between inner heat shield tiles and samples from remote areas below the roof baffle and in pump ducts were analysed, thus offering an exhaustive survey of all relevant areas in ASDEX Upgrade. Deuterium is mainly trapped on plasma-exposed surfaces of inner divertor tiles, where about 70% of the retained deuterium inventory is found. About 20% of the inventory is retained at or below the divertor roof baffle, and about 10% is observed in other areas, such as the outer divertor and in gaps between tiles. The long term deuterium retention is 3–4% of the total deuterium input. The obtained results are compared with gas balance measurements, and conclusions about tritium retention in ITER are made.

Parametric scalings of the intrinsic (spontaneous, with no external momentum input) toroidal rotation observed on a large number of tokamaks have been combined with an eye towards revealing the underlying mechanism(s) and extrapolation to future devices. The intrinsic rotation velocity has been found to increase with plasma stored energy or pressure in JET, Alcator C-Mod, Tore Supra, DIII-D, JT-60U and TCV, and to decrease with increasing plasma current in some of these cases. Use of dimensionless parameters has led to a roughly unified scaling with M_A ∝ β_N, although a variety of Mach numbers works fairly well; scalings of the intrinsic rotation velocity with normalized gyro-radius or collisionality show no correlation. Whether this suggests the predominant role of MHD phenomena such as ballooning transport over turbulent processes in driving the rotation remains an open question. For an ITER discharge with β_N = 2.6, an intrinsic rotation Alfven Mach number of M_A ≃ 0.02 may be expected from the above deduced scaling, possibly high enough to stabilize resistive wall modes without external momentum input.

Argon and neon seeded ITER discharges are numerically simulated by coupling a 1D multifluid model for the plasma core with a 2D model for the SOL-divertor region. The model is fully self-consistent with respect to both the effects of impurities on the α-power level and the interaction between seeded and intrinsic impurities. This interaction leads to a significant change in the intrinsic impurity fluxes, and it is found to be essential for a correct evaluation of the average power to the target plates. Even though carbon and tungsten are the real candidates for the ITER target plates, we have compared carbon and molybdenum plates for two ITER inductive scenarios, due to the uncertainties in the tungsten atomic data. In general, the integrated edge-core scenarios with impurity seeding are more favourable in the case of carbon than of molybdenum plates. However, a high density/low confinement scenario with Ne seeding is found for which the requirements in terms of power to the plate and of power through the separatrix are also fulfilled in the case of Mo. The interplay mechanism among different impurities also holds for He ash resulting in a rather weak dependence of the power amplification factor on He confinement.

The radial transport of tungsten ions in a fusion plasma of the HELIAS stellarator with five magnetic field periods is studied by means of a new numerical code. The code solves guiding center equations for test particles (tungsten ions) with the use of a Runge–Kutta integrating scheme. Coulomb scattering of the tungsten ions on the background plasma particles (electrons, deuterons and tritons) is simulated by means of a discretized collision operator based on the binomial distribution and presented in terms of pitch-angle scattering and energy slowing down and scattering. The coronal model is used to determine the mean charge state of the tungsten ion ensemble ⟨Z(T_e, n_e)⟩ as a function of background electron temperature and density. Two plasma configurations with and without ergodic confinement regions and both with finite plasma pressure of β = 3% are considered. The nonergodic configuration possesses closed nested magnetic surfaces throughout the entire confinement volume. The ergodic magnetic field configuration is represented through additional magnetic field perturbations. Comparative analysis of the radial transport is performed for a time interval greater by a factor of 15 than the energy confinement time τ_E = 1.62 s required for the HELIAS reactor. In spite of the fact that the tendency of impurities to penetrate towards the plasma core is observed in both cases, the stochastic scenario exhibits reduced inward impurity flux.