Search

Recent research in scrape-off layer (SOL) and divertor physics is reviewed; new and existing data from a variety of experiments have been used to make cross-experiment comparisons with implications for further research and ITER. Studies of the region near the separatrix have addressed the relationship of profiles to turbulence as well as the scaling of the parallel power flow. Enhanced low-field side radial transport is implicated as driving parallel flows to the inboard side. The medium-n nature of edge localized modes (ELMs) has been elucidated and new measurements have determined that they carry ~10?20% of the ELM energy to the far SOL with implications for ITER limiters and the upper divertor. The predicted divertor power loads for ITER disruptions are reduced while those to main chamber plasma facing components (PFCs) increase. Disruption mitigation through massive gas puffing is successful at reducing PFC heat loads. New estimates of ITER tritium retention have shown tile sides to play a significant role; tritium cleanup may be necessary every few days to weeks. ITER's use of mixed materials gives rise to a reduction of surface melting temperatures and chemical sputtering. Advances in modelling of the ITER divertor and flows have enhanced the capability to match experimental data and predict ITER performance.

The latest results of B2-Eirene modelling of ITER divertor operation are presented. The operational window is further explored with an improved model of the neutral transport, the effect of gas leaks between the divertor cassettes is assessed and the sensitivity of the results to the features of the divertor geometry and to the gas puffing arrangement is analysed. The presence of neutral–neutral collisions in the model makes the results rather insensitive to the detail of the divertor shape. The analysis shows that the effect of the gas leak on the divertor performance in ITER is weak and therefore inter-cassette sealing may be not critical.

The 2D tokamak edge plasma simulation package B2-EIRENE has been upgraded with a model incorporating the effect of Lyman series photon trapping and ionization from photo-excited states. The numerical approach is based on Monte-Carlo sampling of photon trajectories and a collisional-radiative model for the effective ionization rate. A series of calculations with self-consistent solutions for the neutral gas, photons and plasma background was carried out for ITER edge plasma parameters. It was found that the extra ionization source in the model leads to a higher plasma density near the targets, and is balanced predominantly by enhanced volume recombination. Despite significant changes in the divertor dynamics due to photon trapping the main "engineering scalings" for ITER operational parameters (e.g. peak heat fluxes vs. divertor gas pressure) show no significant modification.

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.

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.

The effect of carbon re-erosion in a mixed-material environment on the results of ITER edge plasma modelling is presented. Divertor performance is investigated with varying erosion rate of re-deposited carbon and varying connection length. The results are unified as scaling relations in the edge parameters, allowing meaningful comparison as model parameters are varied and providing effective boundary conditions for core plasma modelling.

Integrated core–pedestal–SOL modelling is performed for ITER using the MMM95 model for energy transport, modified with additional magnetic shear stabilization to produce a pedestal. The additional stabilization term is calibrated against experiment. The operating space of ITER predicted by the model with neoclassical accumulation of intrinsic carbon impurity is determined. The relationship to an SOL-based density limit is investigated. A reasonable window for operation, with Q larger than 10 is found.

The constraints of divertor operation have been applied to ITER core plasma simulations by imposing boundary conditions on the calculation, implemented as a set of scaling relations derived from B2-Eirene modelling which describe the effect of the divertor. The core plasma simulations use the integrated core pedestal sol (ICPS) model, based on ITG transport for ions and RLW-like transport for electrons, which includes an increase of transport when the ballooning limit is attained in order to simulate the effect of ELMs in a time-averaged fashion. At the nominal average core density for ITER with the ICPS transport model stationary operation with a fusion power multiplier Q of 12–16 is obtained and a reasonable operating range exists for realistic pumping speeds and particle throughput at a peak divertor power load controlled to remain below 10 MW m−2. Fuelling of the plasma must be predominantly direct core fuelling, and the resulting required core-fuelling rates are realistic. Increased pumping speed and particle throughput are beneficial for maximizing Q. The relaxation to these stationary conditions is very slow, and transient values of fusion multiplier are appreciably higher.

A strong effect of the shape of the divertor floor near the separatrix strike point on power loading of the divertor targets in ITER–FEAT is found in the modelling studies. A similar effect has been demonstrated in JET experiments with Mark I and Mark II divertors (A. Loarte, Nucl. Fusion 38 (1998) 587; R. Monk, et al., in: Proceedings of the 24th EPS Conference on Controlled Fusion and Plasma Physics, Berchtesgaden, 1997, vol. 21A, p. 117) thus providing experimental verification of the modelling results. An operational window in six-dimensional phase space is found and different controls, like pumping speed or impurity seeding, are considered for ITER–FEAT.

The operating parameters of the ITER divertor are varied around the partially attached operating point of Refs. [A.S. Kukushkin et al., Berchtesgaden, 1997; A.S. Kukushkin et al., Contributions to Plasma Physics 38 (1998) 20], to obtain an operating window with low heat load, acceptable He pumping and high divertor radiation. Lower heat diffusivities lead to higher heat loads which can be compensated by somewhat lower input power. The divertor geometry is varied: longer domes lead to higher heat loads, whereas the case without dome exhibits poorer helium pumping. Shorter divertor length cases show higher heat load and little change in helium pumping.

The B2-EIRENE code has been compared with plasma edge measurements from Alcator C-MOD, DIII-D, JET, JT-60U in L-mode and H-mode regimes. Similarities in L-mode discharges between JET and JT-60U are well reproduced by the code as well as the sudden approach to detachment seen in JT-60U when a 1% chemical sputtering yield for carbon is assumed. Comparisons between similar L-mode regimes in JET, JT-60U and Alcator C-MOD reveal the larger importance of anomalous energy convection in Alcator C-MOD with regard to larger tokamaks. H-mode discharges of DIII-D and JET can be satisfactorily modelled with similar values of the transport coefficients (approximately an order magnitude smaller than in L-mode) when the power flow into the scrape-off layer (SOL) is dominated by the ion channel, as previous EDGE2D simulations for JET suggested.

Time-dependent theoretical models have been implemented in a 1.5-D transport code to describe the operational boundaries (L–H transition, ballooning limit) in terms of edge parameters. The width of the H-mode transport barrier (F. Wagner et al., Phys. Rev. Lett. 49 (1982) 1408) is self-consistently calculated and is related to α crit . Access to the second stable branch of the ballooning modes is possible at high triangularity. Time dependent trajectories are calculated for several machines and the L- to H-mode transition conditions are compared with global scalings. Calibrated to one experiment (ASDEX-UP), the model yields good agreement with global scaling laws and experimental pedestal electron temperatures, and is then applied to ITER.

Carbon Fiber Composites (CFCs) are one of the candidate armour materials for the plasma facing components of the International Thermonuclear Experimental Reactor (ITER). For the present reference design, CFC has been selected as armour for the divertor target near the plasma strike point mainly because of unique resistance to high normal and off-normal heat loads. It does not melt under disruptions and might have higher erosion lifetime in comparison with other possible armour materials. Issues related to CFC application in ITER are described in this paper. They include erosion lifetime, tritium codeposition with eroded material and possible methods for the removal of the codeposited layers, neutron irradiation effect, development of joining technologies with heat sink materials, and thermomechanical performance. The status of the development of new advanced CFCs for ITER application is also described. Finally, the remaining R&D needs are critically discussed.

We propose a set of unified models to describe H-mode operation of the core plasma, the edge plasma including the transport barrier, and the scrape-off layer plasma to extrapolate to ITER. These models are being incorporated into several tokamak transport codes for use in extrapolating to ITER. Edge pedestal temperatures of 3.5-4 keV for edge densities between 6 and are predicted for ITER, which is the pedestal temperature range required for good confinement with ITG-based core models. The edge operational space predicted for ITER is compatible with both the required core density and core confinement and exhibits a large window between H-mode threshold and the ballooning limit.

Recent tokamakdivertorexperiments with “detached” and “partially detached” divertor regimes show that a major portion of the heating power can be radiated to the walls to reduce the peak heat loads on divertor plates and that a combination of fueling and pumping can control the plasma density and exhaust impurities such as He. Tokamaktransport codes and divertor modeling codes have been developed that can reproduce and predict the major features of these experiments. The combination of these recent experiments and the use of the models to extrapolate to ITER have led to the development of a concept for the International Thermonuclear Experimental Reactor [R. Aymar, Fusion Technol. 30, 397 (1996)] power and particle control system, which relies upon a single null “vertical target” divertor with gas and pellet fueling and active pumping. The experiments indicate that energy confinement can be adversely affected if the recycling neutrals and impurities are not well confined in the divertor, so that efficient confinement of neutrals in the divertor is a key feature of the divertor design.

This paper describes models that address both power and particle control with the ITER divertor. For power control, the characteristics of the transition from attached to detached divertor operation are studied with an extension of our previous simple model [Y. Igitkhanov et al., 22nd EPS, Bournemouth (1995) p. IV-317]. This new model has been applied to JET experiments to show that the model can reproduce global features of the transition. The model has been applied to ITER to identify the range for detached operation in terms of the heating power, upstream density, and impurity level, and the relation between divertor control and core plasma conditions. For particle control, we use the PRETOR transport code together with a two point divertor model [K. Borrass and G. Janeschitz, Nucl. Fusion 34 (1994) 1203] for the edge boundary conditions to study the sensitivity of the central helium concentration to the pumping speed and helium enrichment in the divertor.

This paper reviews the status of the design of the divertor and first-wall/shield, the main in-vessel components for ITER. Under nominal ignited conditions, 300 MW of alpha power will be produced and must be removed from the divertor and first-wall. Additional power from auxiliary sources up to the level of 100 MW must also be removed in the case of driven burns. In the ignited case, about 100 MW will be radiated to the first wall as bremsstrahlung. Allowing the remaining power to be conducted to the divertor target plates would result in excessive heat fluxes. The power handling strategy is to radiate an additional 100–150 MW in the SOL and the divertor channel via a combination of radiation from hydrogen, and intrinsic and seeded impurities. Vertical targets have been adopted for the baseline divertor configuration. This geometry promotes partial detachment, as found in present experiments and in the results of modelling runs for ITER conditions, and power densities on the target plates can be ≤ 5 MW / m 2 . Such regimes promote relatively high pressure (> 1 Pa ) in the divertor and even with a low helium enrichment factor of 0.2, the required pumping speed to pump helium is ≤ 50 m 3 / s . An important physics question is the quality of core confinement in these attractive divertor regimes. In addition to power and particle handling issues, the effects of disruptions play a major role in the design and performance of in-vessel components. Both centered disruptions and VDE's produce stresses in the first-wall/shield modules, backplate and the divertor wings and cassettes that are near or even somewhat in excess of allowables for normal operation. Also plasma-wall contact from disruptions, including at the divertor target, together with material properties are major factors determining component lifetime. Considering the potential for impurity contamination and minimizing tritium inventory as well as thermomechanical performance, the present material selection calls for carbon divertor targets near the strike point, tungsten on the rest of the target and on the baffle where the charge-exchange flux could be high, and beryllium elsewhere. All three materials and relevant joining techniques are being developed in the R&D program and the final selection for the first assembly will be made at the end of the EDA.

For ITER divertor operation in the detached regime, erosion of divertor plates made of beryllium, carbon fibre composites and tungsten-rhenium alloys is evaluated. Erosion due to (a) physical and, for CFC, chemical sputtering (including surface temperature variation over the lifetime because of neutron irradiation), (b) disruption thermal quench using published data and (c) evaporation and melt layer loss (metals) for infrequent high-power transients (20 MW/m 2 , up to 10 s) is evaluated both with and without the reduction of net incident power by radiation from the evaporated impurities (‘low density vapour shield’). The composite lifetime is calculated for normal operation (detached, 10% transients, 10% disruptions); the effect of an increase in transient frequency and in incident power (semi-attached operation) is determined. The resulting erosion lifetime excludes beryllium, whereas both CFC and W-3%Re yield acceptable lifetimes.

The thermal performance of divertor plates armoured with beryllium (thickness 2–30 mm), carbon fibre composite (CFC, 2–60 mm) or tungsten (W3Re, 2–60 mm) was modelled for transients of 10 s at 20 MW/m 2 — with and without vapour shielding. Due to the high resulting temperatures special attention was paid to the material properties up to and above melting or sublimation. Pure copper was chosen for the flat heat sink of Be and W3Re, the coolant tube material was DS-Cu in all cases. The melt layer depth, the surface temperature and peak power densities to the coolant were obtained in full 2D geometry using finite element modelling. The code was modified to include melting. Evaporation and vapour shielding at elevated temperatures improve the cooling efficiency and reduce the incident power (converting it to radiation, some of which is radiated elsewhere) and thus reduce the depth of melting. For Be armour 10 mm thick, slow plasma transients at 20 MW/m 2 after 10 s will induce melting to a depth of ∼ 2 mm without vapour shielding, and ∼ 0.15 mm with vapour shielding. For up to 60 mm of CFC the peak temperatures attained stay below sublimation; for up to 20 mm of W3Re the surface temperature stay below melting. The lifetime of the divertor plates is determined from the reduction of thickness due to these transients (10% of shots, 10 s), as well as from disruptions (10% of shots) and sputtering. When the maximum or initial armour thickness is given either by the maximum permitted surface temperature during steady state operation at 5 MW/m 2 (which is 1050 K for Be, permitting 11 mm armour thickness, 1780 K for carbon, permitting 40 mm of CFC after neutron irradiation), or in the case of W3Re by the requirement to stay below melting after 10 s at 20 MW/m 2 (permitting 20 mm), and assuming a minimum remaining thickness of 2 mm for all three armour materials, the predicted lifetimes without vapour shielding during transients are as follows: 120–225 shots for Be, 5840–8170 shots for CFC with moderate chemical sputtering, and 2370–7740 shots for W3Re.

The ITER divertor must exhaust most of the alpha particle power and the He ash at acceptable erosion rates. The high recycling regime of the ITER-CDA for present parameters would yield high power loads and erosion rates on conventional targets. Improvement by radiation in the SOL at constant pressure is limited in principle. To permit a higher radiation fraction, the plasma pressure along the field must be reduced by more than a factor 10, reducing also the target ion flux. This pressure reduction can be obtained by strong plasma-neutral interaction below the X-point. Under these conditions Te in the divertor can be reduced to

From 2-D modelling of the scrape-off plasma of a next step device, the scaling of key divertor parameters with density, power, transport coefficients and geometry (pitch, size) is derived. In high-recycling conditions, both power density and electron temperature at the divertor plate decrease strongly with increasing upstream density and decreasing input power. When the upstream density is adjusted to obtain constant, low electron temperature at peak power load, the peak power per unit area is proportional to the input power, and the scaling with connection length and transport coefficients is relatively benign. When divertor depth increases, the temperature still increases strongly with decreasing upstream density and both electron temperature and peak power per unit area decrease. Only for a very short divertor, a thermally unstable region is found (i.e. the upstream density is a multivalued function of plasma temperature at the divertor plate).

Scaling laws for the divertor conditions to be expected in a next step device, ITER, are derived from 2D modelling. Model calculations are carried out for varying density, power, transport coefficients, pitch, size and particle throughput. In the low temperature, high recycling domain, the peak power, the electron temperature, and the particle density at the plate are found to be extremely sensitive to upstream parameters, especially to upstream density. Low density, high power operation therefore remains very demanding for the divertor. For operation at a critical density, for which the electron temperature at the point of peak power load is a fixed low value, the peak power load scales more gently, proportional to the power density flowing into the scrape-off layer. Precise control of the density is required to maintain this temperature

Divertor power loads and electron temperatures, obtained numerically using the Braams Plasma Edge Code, B2, are reported and interpreted for the parameter range of a Next Step device. A guideline for terminating a B2 run is proposed: the power balance should follow a ‘strict’ exponential decay when the plasma is approaching its final state. Evidence is presented that erroneous solutions can be obtained if this is not the case. A comparison of results for imposed core density (at fixed plate porosity) and imposed core particle flux (with plate porosity adjusted to match this flux) is made. It is found that external pumping has no effect as long as the flux removed at the divertor plate stays in the range of a few tenths of a percent. We document the effect of an adhoc fluxlimiter. It is shown that similar ‘low’ plasma temperatures can be obtained without a flux limiter, by a small increase of the upstream density, and that the scaling relations are unaffected.

Tokamak ion temperatures measured using the Doppler‐broadened impurity spectral line technique are usually underestimated because the line intensity is integrated along chordal lines of sight and is assumed to be the intensity at the chordal distance from the center of the plasma (radial distance). An analytical correction of the ion temperature radial profile during a series of reproducible discharges on the Tokamak de Varennes reveals a 10%–20% underestimation at the plasma center. The lines used are Ne IX (1248.2 A) and O VII (1623.7 A). Reconstruction of spectral line intensities and shapes using measured electron temperature and density profiles, ion density radial profiles simulated by a plasma impurity transport code, and published atomic data can verify and predict this correction.

Experimental results obtained on the Tokamak de Varennes during a series of reproducible ohmic discharges are presented. Profiles of basic plasma parameters are constructed and compared with theoretical predictions. In particular, the measured plasma resistivity agrees with the neoclassical scaling rather than with Spitzer resistivity. The study of electron-density fluctuations indicates a linear dispersion relation with a propagation velocity of 3.0 × 104 cm s−1. Particle transport investigations are initiated, giving experimental diffusion and convection coefficients across the plasma radius for electrons and impurity ions.

Simulation results obtained with the axisymmetric Princeton Tokamak Simulation Code TSC are compared with experimental results for the time evolution of the Tokamak de Varennes plasma. The code correctly predicts experimental conditions; of special interest are the favourable comparisons of the simulation with experiments during fast current ramp-down (from 250 kA in 40 ms, three times faster than the L/R time).

Plasma contamination due to the generation of impurity molecules has been studied by mass spectrometry and by visible emission spectroscopy in the Tokamak de Varennes. The dominant effects are carbon monoxide formation, which is correlated with the residual water vapor pressure in the vacuum chamber, and the formation of C{sub 1}, C{sub 2}, and C{sub 3} hydrocarbons. The measured molecular fluxes are sufficient to account for a large part of the plasma impurity content. Visible spectroscopy indicates that the plasma is significantly affected by these chemical impurity sources. The molecules appear to originate mainly from the stainless steel walls rather than from the graphite limiters.

A system consisting of magnetically shielded ion gauges and a differentially pumped quadrupole mass spectrometer is described. It provides time‐resolved measurements of the neutral gas fluxes during discharges in the Tokamak de Varennes. A response time of 15 ms for hydrogen and ≊50 ms for impurity gases was obtained by a judicious choice of distance (to avoid magnetic disturbances), conductance, and pumping speed. An interface between the spectrometer and the data acquisition system of the tokamak allows fast (ms) selection of the mass and real‐time (≊10 ms) adjustment of the gain to optimize the signal‐to‐background ratio. The usefulness of this diagnostic is illustrated by its use in three experiments: propagation of a plasma density perturbation following pulsed (5 ms) gas puffing, exchange of the discharge gas from H2 to He, and chemical impurity (CO, CO2 ) production during discharges at the density limit and qa≊2. The correlation of low impurity fluxes with absence of disruptions in these last shot...

The physics of ITER (International Thermonuclear Experimental Reactor) is based on demonstrated tokamak physics and credible extrapolations of that physics. Assessment of the energy confinement and MHD stability requirements led to the choice of the major plasma parameters of 22 MA for the plasma current, a toroidal field of approximately 5 T, an aspect ratio of approximately 3, and an elongation of approximately 2. Among the major accomplishments of the physics group has been the development of a database and an empirical scaling for L-mode energy confinement and the facilitation of an H-mode database and scaling. The divertor heat loads have been estimated by using experimentally validated models. The thermal and mechanical loads due to off-normal events such as disruptions have been based on analysis of the data from other tokamaks. To achieve the required availability of 10%, the pulse length has been extended by the use of current drive using 75 MW 1.3 MeV neutral beam and 45 MW lower hybrid systems. A relatively complete set of plasma diagnostics is planned for ITER. Finally, a physics R and D program has been developed. >

The requirements for the divertor diagnostics have originally been determined by the examination of the diagnostic experience on present large divertor tokamaks. In the past years the design of most diagnostic systems inclusive of those for the divertor has progressed beyond the conceptual stage so that their expected performance can now be predicted using this design information and the results of 2D modelling of the divertor plasma in a regime close to the desired operating point. The profiles of plasma parameters and the line integrals of spectral emission, which would be measured along the chords of the planned diagnostics, have been calculated for 3 cases for which the 2D model predicts sustainable heat loads on the divertor target. The results are assessed according to their capability to fulfil the requirements for protection, plasma control and physics understanding.

In the last Diagnostic Workshop, which took place at the mid-point of the ITER EDA, the ITER machine design and physics performance was presented [1] based on the Interim Design Report (IDR). Whereas the design of several major machine components (e.g. TF-coils) were changed considerably compared to the earlier Outline Design (OD), it was pointed out at the time that the overall objectives and the plasma performance remained in line with the requirements stated by the Special Working Group(SWG 1).

Previously obtained data on the Wega tokamak by lower hybrid heating (f = 500 MHz; PHF = 130 kW) are re-visited in the light of recent theories on ion stochastic heating and quasi-linear electron Landau damping. It is possible, with the help of these theories, to correctly estimate the fast-ion mean energy, the HF power density coupled to the ions and that coupled to the electrons. The values of the parallel index of refraction, N||, which are necessary to obtain a good quantitative agreement between experiment and theoretical estimates, are the same for ions and electrons, although higher than expected.

Observations of marginally unstable resistive drift waves in a hydrogen plasma immersed in the sheared magnetic field of the Princeton spherator are reported. A theoretical treatment based on the non-local normal mode is proposed, which predicts the wave parameters of the most unstable mode, and provides a new stability criterion. The experimental wave parameters and observed stabilization of the resistive drift waves are found to agree well with the predicted values. 11 refs., 2 figs., 1 tab.

A tentative plasma power balance in the presence of H.F. is presented. It is shown that on the application of 125 kW H.F. input power, 15 to 201 are coupled to the ions and 30 to 50 kW to the electrons, increasing the plasma energy content by a factor larger than 2.

The previously obtained results of lower hybrid heating in the WEGA Tokamak (f = 500 MHZ, PHF ≤ 130 kW) are reconsidered in the light of recent theories on ion stochastic heating and electron Landau damping. The parametric dependences of the observed hot ion mean energy and the HF power coupled to the ions and electrons are found to correspond to the theore -tical dependences. Under reasonable hypotheses, good quantitative agreement is found. In particular, ion and electron heating may be attributed to the same coupled HF spectrum.

A modelling study of the ITER divertor is done using the B2-Eirene code package with full multi-species impurity treatment and volumetric recombination of deuterium ions. A ‘virtual target regime’ is identified, for which strong recombination and ionization form a double layer which acts like a target surface inside the divertor plasma. The power loads, erosion, and helium removal seem to be within the acceptable range for the ITER design, making this regime, together with operational points close to it, a reasonable starting point for the optimization of the ITER divertor.