• Keine Ergebnisse gefunden

Association EURATOM FZJ:  Annual Progress Report 2007

N/A
N/A
Protected

Academic year: 2022

Aktie "Association EURATOM FZJ:  Annual Progress Report 2007"

Copied!
280
0
0

Wird geladen.... (Jetzt Volltext ansehen)

Volltext

(1)

Nuclear Fusion Project: SC-FZJ 80(08)/4.1.2

Association EURATOM FZJ:  Annual Progress Report 2007

final version August 5th 2008

Member of the Helmholt

(2)

edited by Ralph P. Schorn

(3)

A. The FZJ Nuclear Fusion Programme (executive summary)... 5

B. Scientific and Technological Programme ... 15

B.1. Plasma-Wall Interaction... 15

B.2. Tokamak Physics ... 45

B.3. Technology ... 99

B.4. Diagnostics and Heating... 119

B.5. Contributions to ITER ... 157

B.6. Contributions to Wendelstein 7-X ... 168

B.7. Materials under high Heat Loads... 172

B.8. Theory and Modelling ... 201

C. Specific Contributions of the Partners within the IEA Implementing Agreement... 224

C.1. Japan... 224

C.2. United States of America... 230

D. Summary on results of the main projects in the framework of "Projects for ... 234

enhancing the mutual co-operation between Associations" E. Structure of the Fusion Programme and related Figures ... 244

F. List of scientific Publications, Talks and Posters... 247

(4)
(5)

Nuclear Fusion Progress Report 2007 SC-FZJ 80(08)/4.1.2 A. The FZJ Nuclear Fusion Programme – Executive Summary

Ulrich Samm, Detlev Reiter (IEF-4 Plasma Physics, u.samm@fz-juelich.de)

Introduction

Forschungszentrum Jülich (FZJ) as a EURATOM Association coordinates its fusion research activities within the Project Nuclear Fusion (KFS). The programme is based on several insti- tutes and is well embedded in the European fusion research structure. The largest part of the Jülich research activities is located within the Institute of Energy Research (IEF). The former Institute for Plasma Physics (now IEF-4) has by far the largest share of scientific staff in phys- ics and technology for fusion and operates the tokamak TEXTOR. The IEF-2 (Microstructure and Properties of Material) operates the electron beam facilities JUDITH 1/2 located inside the hot cells and represents the materials science expertise within the fusion programme. The Cen- tral Technology Division (ZAT) provides engineering expertise and specialized workshop ca- pacities. The Jülich Supercomputing Centre is now integrated as a division into the newly founded Institute for Advanced Simulation (IAS).

The association EURATOM-FZJ has very close contacts to the neighbouring EURATOM as- sociations in Belgium and The Netherlands. In 1996 they have founded the Trilateral Euregio Cluster (TEC), which provides a clustering of resources in order to perform a coordinated R&D programme, to operate or construct large facilities (TEXTOR, MAGNUM-PSI), to com- bine different expertises and to allow the forming of a strong partnership as a consortium within the ITER construction phase.

Co-operations beyond Europe are strongly supported by an IEA Implementing Agreement on

“Plasma-Wall Interaction in TEXTOR” (with Japan, USA, Canada), which meanwhile also serves as a basis for the exchange of scientists to other devices than TEXTOR.

Highlights 2007

Suppression of the avalanche multiplication of runaway electrons by Resonant Magnetic Perturbations (RMPs)

High energy runaway electrons produced during disruptions pose a severe threat to plasma facing components of fusion reactors. In 2007 it could be shown that the avalanche like ampli- fication of the runaway beam can be suppressed in TEXTOR by the application of RMPs with toroidal mode numbers n = 1 and n = 2 imposed by the Dynamic Ergodic Divertor above a certain threshold of the perturbation current. This finding has been related to the enhanced dif- fusion coefficient for the trajectories of the fast electrons in the stochastic magnetic field. As a consequence, the duration of the current plateau caused by the runaways is considerably re-

(6)

duced and the high energetic runaways are suppressed, indicating that RMPs are a potential method for runaway suppression as an alternative to the use of massive gas injection which is studied in TEXTOR in collaboration with JET, DIII-D and AUG.

ELM- induced erosion in the JET divertor

In 2007 the impact of pulsed heat loads (large type-I ELMs) has been investigated in ITER-like divertor configurations in JET. Two new diagnostics (quartz- micro balance, high resolution bolometry) provided under the leadership from FZJ helped to identify the role these ELMs on erosion and deposition of carbon materials. These results are of high relevance for the life time of wall components and tritium retention rates in ITER.

Thermal disintegration of amorphous carbon-deuterium layers from the inner divertor has been observed after large ELMs with energies > 100 kJ. This leads to the deposition of carbon layers on remote areas. For ELM energies larger than 700 kJ, non-linearly enhanced radiation in the plasma has been detected, which is related to the ablation of deposited layers on the inner tar- get and a subsequent massive re-distribution of the eroded material.

Objectives and incorporation into the research area

Today fusion research at Forschungszentrum Jülich is largely organized along topical groups (plasma-wall interaction, tokamak, diagnostics, theory and modelling, technology). These groups use a variety of different facilities. Among these the most important machine is JET where scientists from Jülich are strongly involved, in particular in the scientific exploitation and also in the technical preparation of the new ITER-like wall project as well as experiments on ELM-mitigation. Other facilities outside of Jülich with participation from FZJ are DIII-D, PISCES-B, AUG, TS, LHD and MAST.

IEF-4 operates the tokamak TEXTOR as a local facility in Jülich (IP, max = 0.8 MA, BT, max = 3.0 T, R = 1.75 m, a = 0.46 m, plasma volume 7 m3, circular cross section, toroidal graphite belt- limiter (pumped), 16 TF coils, pulse length 12 s; auxiliary heating power: NBI co 2 MW, NBI ctr 2 MW, ICRH 4 MW and ECRH 1 MW).

The Dynamic Ergodic Divertor (DED) on TEXTOR provides unique means for resonant mag- netic perturbations: 16 helical in-vessel RMP coils; base modes: 12/4, 6/2, 3/1, Imax = 15 kA, DC and rotating field up to 10 kHz. Based on these means the programme participates in ELM- mitigation studies (joint experiments) and in the investigation of power exhaust in helical di- vertor structures in preparation of long pulse and steady-state operation in stellarators.

For Plasma-Wall Interaction (PWI) studies a powerful PWI test facility is available on TEX- TOR: two air-lock systems to expose movable and easily exchangeable larger scale wall com- ponents with gas feed, external heating and active cooling under ITER-relevant parallel heat and particle flux densities. The system is equipped with a comprehensive in-situ set of PWI diagnostics.

(7)

In addition the programme is supported by a variety of smaller laboratory devices: Tandem accelerator device for the quantitative determination of surface material compositions (NRA, RBS), dedicated laboratory devices for in-situ PWI simulation and analysis (TOF-SIMS) and various devices for the plasma assisted preparation of fusion relevant layers and coatings, and a

“mirror laboratory” for the characterisation and analysis of experiments with plasma facing mirrors in tokamaks.

The special expertise of IEF-4 in fusion technology is manifested by major engineering pro- jects: concept development, design, construction and installation of the TEXTOR tokamak and various upgrades, and most recently design, layout, manufacturing and assembly of the super- conducting bus-bar-system for Wendelstein 7-X and design and procurement for a bulk tung- sten plasma facing component for the new JET divertor.

The institute IEF-2 operates the high heat flux test facilities JUDITH 1 and JUDITH 2. These electron beam facilities are capable for ITER- and DEMO-relevant quasi-stationary and tran- sient high heat flux tests in the sub-millisecond range with loaded areas of up to 50 x 50 cm2. A unique feature is its operation inside of hot cells which allows testing of neutron irradiated and toxic materials (Be, T-containing samples).

ZAT is developing and manufacturing experimental devices and techniques which are not available on the market for a wide range of scientific applications. The central facility of FZJ provides engineering expertise in the fields of project engineering, joining and testing technol- ogy, and prototype manufacturing using special tools and techniques.

The Jülich Supercomputing Centre is now integrated as a division into the newly founded Insti- tute for Advanced Simulation (IAS). It operates a dual super computing system (both: general purpose and massive parallel architectures). It is foreseen to host the first dedicated European Supercomputer for Fusion (100 Tera Flop). This HPC facility is embedded into the European theory and modelling activities, such as the EU-ITM task force, and it also serves as a training platform for the Petaflop Computer for ITER foreseen within the Broader Approach agreement between EU and Japan, from 2012.

The Helmholtz Association's fusion activities are based on the European fusion research pro- gramme. The following Helmholtz Centres are involved: Max Planck Institute of Plasma Phys- ics (IPP, Garching and Greifswald), Forschungszentrum Karlsruhe (FZK), and Forschungszen- trum Jülich (FZJ). The research is organized along the topics: a) stellarator research, b) toka- mak physics – ITER and beyond, c) fusion technology for ITER, d) fusion technology after ITER, e) plasma-wall interaction, and f) plasma theory. This report presents results having been achieved by FZJ in these topics.

Programme results

Stellarator research

FZJ is responsible for design and fabrication of the superconducting bus-bar system and for some plasma diagnostics.

(8)

Superconducting bus bar system

The superconducting bus-bar system consists of a geometrically complex mesh of conductors being exposed to strong forces. It provides the electrical connections to the stellarator's mag- netic field coils. After the construction of the production line and the qualification of the fabri- cation and assessment process, the first set of six bus conductors was manufactured and suc- cessfully tested. In 2007 all busbars required for module 5 – which is assembled first – have been delivered to Greifswald.

The construction and fabrication of supporting elements are currently being worked on. Series production was optimised concerning the joints. A welding procedure was developed and tested for repeated mountability. The design of the support structure is based on different ad- justable sub-modules which are able to compensate fabrication tolerances in all directions and to facilitate the assembly on site. Design and stress calculations are taking into account these effects.

Approximately 230 low-resistance joints are required for electrical and hydraulic interconnec- tions between superconductors at the coil terminals and between five adjacent modules. Based on a conceptual design for a pressure of 30 bars and a current of 18 kA a demountable joint for 200 bar and 20 kA have been redesigned. After design review three joints have been manufac- tured and tested under pressure. The manufacturing of 230 joints including inner clamping parts is ongoing at FZJ.

Diagnostics for Wendelstein 7-X

The new High Efficiency eXtreme ultraviolet Overview Spectrometer system (HEXOS) for the stellarator Wendelstein 7-X is now mounted for testing and adjustment on TEXTOR. One part of the testing phase was the intensity calibration of the two double spectrometers which in total cover a spectral wavelength range from 2.5 nm to 160.0 nm with overlap.

A high energetic hydrogen beam (RUDIX) for diagnostic applications is foreseen for the measurement of ion temperatures and impurity density profiles using charge exchange recom- bination spectroscopy. A diagnostic injector will be developed, the beam of which provides an equivalent current of more than 5 A at 60 keV energy. The Budker Institute of Nuclear Physics (BINP Novosibirsk) has continued the construction of the 60 kV, 10 A high voltage power supply. The final acceptance tests are planned for 2008 in Novosibirsk. The tests of RF and arc ion sources at the diagnostic beam at TEXTOR have been finished, favouring the RF source for RuDIX.

The installation of a 14 channel dispersion interferometer (DI), developed by the Budker Insti- tute of Nuclear Physics (BINP), has made significant progress in 2007. To proof the basic prin- ciple of the DI and to develop the final design of the interferometer set-up, measurements with the prototype DI channel installed at TEXTOR have been completed in the first half of the year. As a highlight of the measurements using the prototype DI, it was used for real-time plasma density control at TEXTOR instead of the HCN interferometer.

(9)

The imaging Bragg X-Ray spectrometer was installed to TEXTOR in 2007. The development of a detector is almost completed and the measurement of first spectra from TEXTOR is scheduled for 2008.

Tokamak physics

The scope of this topic includes all the physics being related to a tokamak plasma with a spe- cial emphasis on Resonant Magnetic Perturbations (RMP).

Main focus in 2007 with respect to control of transient power and particle fluxes by RMP were the experiments on ELM mitigation at JET and the collaborative experiments on ELM suppres- sion at DIII-D. Complementary investigations on the impact of RMP on the limiter H-mode at TEXTOR were continued.

Generic transport properties in partly/fully ergodised edge plasmas by RMPs are further inves- tigated. This topic is related to the ELM mitigation scenarios by RMP as well as to stochastic plasma structures present in stellarators/heliotrons. The main focus in 2007 was on the impact of RMP on edge turbulence/intermittent transport and on particle confinement.

With respect to plasma stability two main topics were in the focus of the 2007 activities: exci- tation and suppression tearing modes with RMP and ECRH/ECCD and disruption mitigation by massive gas injection (ITPA task MDC-1). The experiment on tearing modes aimed at the study of the effects of early ECRH and ECCD on the penetration of RMP and on the transport properties of large magnetic islands. The focus of the disruption mitigation experiments has been on the runaway suppression during TEXTOR disruptions by MGI and RMP.

Related to turbulent transport the main aim in 2007 was the further characterization of the geo- desic acoustic modes (GAMs) and the investigation of the related velocity fluctuations. These measurements were performed using the versatile correlation reflectometry system at TEX- TOR.

Collective Thomson scattering (CTS) diagnostic is the main tool for physics studies of fast ion interaction with MHD oscillations such as sawtooth oscillations, and also experiments studying the dynamics of the confined fast ions during RMPs. With the help of a new scintillator probe, the energy spectrum of runaway electrons in flat-top plasmas as well as in disruptions was measured for the first time.

The helical structure of the dynamic ergodic divertor (DED) allows studies of the divertor physics of helical divertors as implemented or planned for stellarators and heliotrons. Studies addressing particle recycling and divertor regimes were started in 2007 using new spectro- scopic observations of the target plates. These studies include the benchmarking of the 3D transport code EMC3.

(10)

Fusion technology for and beyond ITER

ITER Diagnostics and Port Plug Engineering

In May 2006, a European cluster of associations has been formed to work on the development of the ITER core Charge Exchange Recombination Spectroscopy (CXRS) diagnostic. The clus- ter is jointly led by the TEC partners FZJ and FOM, with participation from the EURATOM associations UKAEA, HAS, IPP, CEA and ENEA. The Dutch organisations TNO and NTG entered into this cluster in the frame of the new ITER-NL organisation with substantial contri- butions in the fields of project engineering, optical design and neutronics.

In 2007 activities of the cluster partners have started to form a consortium that shall act as the contractor for the future design and manufacturing of ITER CXRS. A detailed project plan for the full development and implementation of the ITER core CXRS diagnostic has been devel- oped.

Because of several similarities of the CXRS diagnostic with the LIDAR system being prepared by the UK, a collaboration has been established. As far as engineering is concerned FZJ is con- tributing to the LIDAR project with the development of shutter options.

ITER upper ports are available for the installation of both diagnostic systems and Electron Cy- clotron Heating (ECH) launchers. FZK has designed a port plug shell for an upper ECH launcher. It consists of a closed shell with a cover bolted on the top. FZJ has investigated the ability of this shell to withstand electromagnetic loads caused by transient fields during a plasma disruption. Recommendations for design improvements have been derived from the simulation results.

Materials for components in contact with fusion plasma

The development of plasma-interactive components for next step devices (ITER) and future electricity generating fusion reactors such as DEMO are an essential material related issue.

Hence, a major aim of the R&D activity at FZJ is the characterization of plasma facing materi- als and actively cooled components and their assessment with respect to their thermo- mechanical behaviour. Central facilities for these investigations are the electron beam devices JUDITH 1 and 2 located inside the hot cells.

In 2007, apart from the determination of mechanical and thermophysical material characteris- tics, the investigation of the thermal-shock and fatigue behaviour of JET and ITER related tungsten, beryllium and CFC materials and components under cyclic thermal loading was of particular interest. Hereby especially wall erosion as a result of plasma-wall interaction proc- esses, i.e. ELMs, has been addressed also in combination with neutron induced degradation of the materials with respect to their thermal and mechanical properties. These are critical issues that have significant impact on the lifetime of plasma facing components and on the contami- nation of the fusion plasma.

With regard to DEMO, where refractory metals such as tungsten or its alloys will play an even more important role and He-cooling might be a viable technique, new designs have been quali-

(11)

fied and damage mechanisms addressed. Furthermore by participation in the European IP- project ExtreMat, among others, knowledge on future high temperature heat sink materials, e.g.

fibre reinforced metal matrix composites, have been gained offering alternative solutions for the DEMO first wall and divertor design.

Plasma-wall interaction

Controlling the interaction between the fusion plasma and the walls of the burn chamber is becoming increasingly important for ITER's operating range and for further developments leading to a steady-state burning fusion plasma. The aim here is to attain conditions with very little erosion of the wall material and at the same time with a very low fuel retention. In addi- tion, these conditions must be compatible with the requirements for energy confinement in ITER. Reaching this goal is strongly influenced by the choice of wall materials: For the first operational phase of ITER a combination of carbon resp. graphite, beryllium and tungsten is foreseen.

This Topic Group is fully aligned with the goals of the European Task-Force on PWI. ITER- relevant subjects are thus treated with priority.

In line with the preparation work for ITER is the strong contribution of FZJ to the currently largest tokamak JET within the associated Task Force Exhaust. At JET, ITER like wall mate- rial combinations will be tested in the frame of the ITER-like wall project acting as a test bed for the currently foreseen material options in ITER.

In parallel, the basic understanding of plasma-surface interaction and related plasma processes near plasma-facing components is still to be improved. Modelling of material migration, thus erosion, transport and deposition in JET and TEXTOR, as well as simulation of general plasma-wall interaction processes such as material mixing at the surface are essential part of the work. The applied modelling has to be benchmarked with experiments and applied to pro- vide reliable predictions for ITER operation.

Erosion and deposition, carbon migration and tritium retention

In 2007 the impact of large type-I ELMs on carbon erosion has been investigated in ITER-like divertor configurations in JET. Detailed measurements with a quartz- micro balance show the deposition of carbon layers on remote areas following the thermal disintegration of amorphous carbon- deuterium layers from the inner divertor after large ELMs with energies > 100 kJ. For ELM energies larger than 700 kJ, non-linearly enhanced radiation in the plasma has been de- tected with the upgraded divertor bolometry in JET. This radiation, which can contribute with up to 50% to the drop of stored energy during a giant ELM in JET (the energy limit for ELMs in ITER has been defined as 1 MJ), is related to ablation of deposited layers on the inner target and a subsequent massive re-distribution of the eroded material.

Studies to remove deposited carbon layers with special ICRH discharges in different reactive gases were ongoing. The magnetic field configuration plays a crucial role for the homogeneity of the cleaning process.

(12)

A new method for the in-situ determination of the hydrogen content of carbon layers using laser desorption and spectroscopy has been further developed. It has been demonstrated on TEXTOR that the hydrogen content in co-deposited carbon layers can be measured during the tokamak discharge with very high sensitivity. This allows even to monitor the build-up rate of hydrogen containing layers. Design work for a system suitable for ITER is in preparation.

The melt layer behaviour of tungsten with brush-limiters has been investigated under extreme heat loads in TEXTOR.

Modelling

The ERO code has been significantly advanced. The versions for various machine geometries (divertor, limiter, linear devices) have been merged into one common source code. The calcula- tion of collisions between neutrals has been further improved.

The ERO Code has been used e.g. together with Be-experiments on PISCES-B, for tritium re- tention estimates for ITER and for the determination of the deposition behaviour in castellated structures.

Benchmarking of the code has been done by a comparison with experimental data from local deposition effects obtained with methane injection. Differences for the deposition probabilities with tungsten and carbon surfaces have been observed. Also the surface roughness plays an important role.

JET ITER-like wall

The ITER like wall project (ILW) has been launched in JET to study the compatibility of the ITER operating scenarios with metallic plasma-facing components (Be and W) and to address a number of dedicated outstanding PWI questions such as to the long term Tritium retention under ITER material conditions. FZJ provides the leading Project Scientist (V. Philipps).

FZJ is responsible for the development of bulk tungsten divertor modules for the highly loaded divertor plates. A row of bulk tungsten tiles for the outer strike point in the divertor of JET is currently in the procurement phase. Extensive calculations of temperature distributions in the module, of corresponding thermo-mechanical stresses, of electromagnetic forces and of overall structural mechanics have been performed.

Heat load cycle tests of sample tiles have been performed in JUDITH-2 showing no damage up to 80 MJ/m2 and temperatures around 1860 °C. Further tests are planned in parallel to the manufacturing process.

Magnum-PSI

As part of the TEC collaboration the FOM-Institute for Plasma Physics Rijnhuizen/The Neth- erlands is building a new machine to study plasma-wall interactions. The steady-state high par- ticle flux density and a magnetic field of 3 T, and the large beam diameter will bring the rele-

(13)

vant parameters typically an order of magnitude beyond what is presently available in linear plasma devices, and into the realm of the ITER divertor.

The contributions of FZJ to this project are the concept, design and procurement of the target chamber which allows replacing samples without breaking the vacuum. A significant part of the definition and design phase of the Magnum-PSI project has now been completed. In FZJ, the definition and design activities of the target station and the manipulator have been final- ized. The procurement of the system has started. The assembly and installation activities will commence in 2008.

Theory and modelling

The main focus at FZJ in this area is development and application of computational tools to quantify PWI related aspects in fusion edge plasmas, such as those resulting from wall released impurities, divertor chemistry and SOL turbulent transport. The new European fusion HPC to be hosted at FZJ, with computational edge plasma science as one of its foreseen key applica- tion areas, has lead to a further strengthening and focusing of this programme.

Integrated edge plasma modelling

Within the ongoing long term collaboration with the ITER edge modelling team a new version of the 2D B2-EIRENE edge modelling code has been developed and released (SOLPS4.3). A distinguishing feature is the realization of Message-Passing (MPI) parallelization in the Monte- Carlo (EIRENE) part of the code. Full backward comparability with the ITER divertor design code SOLSP4.2 has been proved. The upgraded code is in use jointly with the ITER team in particular for assessment of high divertor density conditions (ITER and DEMO), which have previously not been accessible due to run-time limitations.

A quantitative 3D analysis (EMC3-EIRENE) of integrated edge transport (plasma, radiation and neutral gas) in helical divertor configurations and under the influence of resonant magnetic perturbations (RMPs), as foreseen now also for ITER, has been carried out for TEXTOR-DED and D-IIID limiter edge plasma conditions. This provides a first direct experimental validation of 3D computational edge models also for the future divertor in W7X. Newly developed geo- metrical options now also allow application to poloidal divertor tokamak configurations (e.g.

DIII-D, JET, MAST, ASDEX-U) and hence enabling for the first time direct code-code com- parison between the recent 3D (EMC3-EIRENE) and the far better established 2D edge codes (such as B2-EIRENE or EDGE2D-EIRENE).

Impurity transport

The physical origin of plasma confinement modifications due to the presence of impurities is investigated using the RITM code for JET in order to identify if the same causes as identified earlier for the TEXTOR RI-mode transition also are at work at other machines (density peaking and increased Zeff as main causes of the suppression of the ITG turbulence and of the peaking of the central ion temperature). Applications to JET conditions indicate that there (distinct from TEXTOR) the anomalous convection, not included in the code, might be an important factor.

(14)

European Transport Solver within the EFDA ITM Task Force

The main objective of this activity is to provide the computational basis for a modular transport code, the European transport Solver (ETS), for self-consistent modelling of plasma parameters in the core, the pedestal and scrape-off layer – ultimately to simulate complete discharge sce- narios, e.g. for ITER. For this transport codes used in European Associations have been re- viewed, the physics content and numerical concept of the ETS have been defined and standard- ized data structure for 1-D transport computations have been formulated. In particular new nu- merical approaches to integrate the highly non-linear transport equations in the codes RITM and ETS have been further developed.

3D fluid turbulence in edge and SOL

Formulation of plasma turbulence models and development of numerical codes capable to deal self-consistently with open and closed field lines, i.e. to solve for turbulent transport across a separatrix, is still one of the major challenges in fusion plasma theory and modelling.

For this the 3D drift fluid turbulence code ATTEMPT (FZJ) has been extended to open field line geometries. Intermittent (“blobby”) particle transport studies for TEXTOR conditions have been successfully validated against experimental data. They show in particular that RMP per- turbations (as e.g. produced with the DED in TEXTOR in a very controllable way) might cause a strong suppression of the intermittent transport. This is a key finding possibly also with re- spect to the now foreseen RMP coils in ITER (a result from the design review) and with poten- tially important consequences for main chamber recycling, diagnostic (mirror) design, etc., for ITER.

(15)

Nuclear Fusion Progress Report 2007 SC-FZJ 80(08)/4.1.2 B.1. Plasma-Wall Interaction

Sebastijan Brezinsek (IEF-4 Plasma Physics, s.brezinsek@fz-juelich.de)

Introduction

Solutions of remaining physical difficulties in operation of the next-step fusion device ITER, especially those related to plasma-wall interaction, are urgently needed in the current phase in order to make the best engineering choices to ensure a safe and reliable plasma operation. Op- erational experiences must be gained in parallel to the ITER construction for a number of addi- tional specific topics – in order to support a safe ITER operation. The Main Topic Group on Plasma-Wall Interaction (PWI) remains organised in a programmatic oriented fashion and, with the motivation of developing a viable solution to the questions of the ITER first wall and divertor materials as well as to possible operation scenarios to reduce power and head loads to the first wall and in particular the divertor. In this sense, the Main Topic Group is fully aligned with the goals of the European Task-Force on PW).

ITER-relevant subjects are thus treated with priority.

In line with the preparation work for ITER is the strong contribution of the Main Topic Group to the currently largest tokamak JET within the associated Task Force Exhaust (see binations will be tested in the frame of the ITER-like wall project acting as a test bed for the currently foreseen material options in ITER. In parallel, the basic understanding of plasma- surface interaction and related plasma processes near plasma-facing components is still to be improved. Modelling of material migration, thus erosion, transport and deposition in JET and TEXTOR, as well as simulation of general plasma-wall interaction processes such as material mixing at the surface are essential part of the work of the Main-Topic Group. The applied modelling is benchmarked with experiments and applied to provide predictions for ITER for the key PWI questions for ITER operation.

The two categories of open questions for PWI on ITER are clearly: (i) safety issues (operation) with respect to tritium inventory and (ii) the lifetime of the plasma-facing components (PFC) such as divertor target plates. (i) includes the measurement and the prediction of fuel retention via approved modelling, as well as techniques to clean up the plasma-facing components and control the dust inventory. (ii) is focused on the modification of plasma-facing material, and here especially of the target and other highly exposed elements, with long-time operation, thus mainly related to erosion, sublimation and melting. The need for qualification of high-Z mate- rials such as W (or others) which are suitable for plasma-facing components has turned to the focus of the ITER PWI questions in the last year.

(16)

The Main Topic on Plasma-Wall Interaction therefore in first instance deals with the following fields:

(a) Carbon-based PFCs: Erosion, transport and deposition, thus, material migration and tritium retention for ITER components conceived on a carbon basis. The de- velopment of deposition mitigation techniques and of removal techniques for hy- drogen isotopes belongs to this field. Work concentrates on erosion at ITER like divertor conditions, transport of carbon along surfaces and to gaps and qualifica- tion of fuel removal techniques under the presence of magnetic fields.

(b) High-Z PFCs: W is foreseen for the baffles or – most likely foreseen for the acti- vated phase in ITER, for the divertor plates. Research in this field concentrates on high temperature behaviour of W close and at melting and on retention of hydrogen in the bulk material.

(c) Material mixing: The ITER first wall material mix may lead to material mixtures on pfc materials and to the appearance of alloys, carbides etc. Their appearance and physical properties are of vital importance with respect to material properties and fuel retention. Research concentrates on the mixing of W with C by codeposition and implantation.

(d) Plasma operation under detached conditions: The operational divertor regime in ITER requires a divertor detachment to avoid damage of the divertor target plates.

Here, high density operation is close to the appearance of radiative instabilities such as MARFEs which finally can lead to disruptions.

(e) Qualification of atomic and molecular data: The study of plasma-wall interac- tion and the associated modelling requires a detailed diagnostic investigation of the plasma boundary layer as well as a good atomic and molecular database for the in- terpretation.

The Plasma-Wall Interaction group works on a number of fusion and other facilities in order to match best the needs for the specific topic. The majority of the work is done at TEXTOR and JET with additional contributions from AUG (ASDEX-Upgrade), DIII-D, Tore Supra, PI- SCES, Pilot-PSI, and others. The full research programme is organised within the Trilateral Euregio Cluster, which encompasses the Belgian (ERM/KMS) and Dutch (FOM) partners in addition to the German Institute in Jülich. The IEA partners (Japan, USA, Canada) are closely linked to this research programme, as shown in this report. TEXTOR serves as the central fu- sion facility for the TEC partners, without prejudice to resorting to any other device if better suited. Joint experiments are performed at different experiments in the frame of different ITPA Divertor Scrape-Off Layer working groups. These working groups directly address requests from the ITER PWI science community (EFDA, ITER Organisation, F4E, etc).

1. Erosion and deposition, carbon migration and tritium retention

1.A. Chemical erosion, hydrocarbon catabolism and spectroscopy

Carbon-based materials such as carbon-fibre components have been deployed successfully as plasma-facing materials at the locations of highest heat flux, such as the divertor target plates

(17)

in present-day fusion devices. Therefore, the current ITER-material choice foresees CFC at the divertor areas with highest heat flux for the non-activated operation phase. However, chemical sputtering of carbon and the appearance of fuel co-deposition are critical issues for the usage of carbon with respect to machine reliability. In the case of tritium usage – such as in ITER – this also addresses machine safety. Note that carbon erosion is no longer envisaged as a lifetime issue due to the short planned installation time of the full set of divertor modules and the re- cently reduced allowed maximum loss of energy per ELM.

The main question is: Is the remaining CFC acceptable in the activated phase of ITER opera- tion or not? This topic is closely related to the determination of chemical erosion yields under ITER-relevant plasma conditions because a substantial part of the tritium inventory is found in the carbon-based re-deposited layers. Quantification of CFC erosion yields and their estimation with erosion and deposition simulation codes, e.g. ERO (see below, sect. 1.H.), is topic of the ITPA Divertor and SOL Group (D-SOL2) and of the SEWG (Special Expert Working Group) on Chemical Erosion in the EU Task Force on Plasma-Wall Interaction. Experiments under the scientific leadership of FZJ have been performed in 2007 at JET and TEXTOR, as well as in the linear machines PSI-2 and Pilot-PSI (data from linear machines are in the validation process).

Experiments focused mainly on

a) the database (HYDKIN) and code validation (ERO-code), and

b) the measurement of erosion yields in situ, in dependence on the energy and flux of incident particles as well as on surface conditions (validation of the so-called "Roth" formula) with the aid of optical spectroscopy on the CD radical and local gas injection of hydrocarbon species.

Comparison of measured and modelled effective D/XB values from hydrocarbon injection experiments

At TEXTOR, a series of experiments has been carried out to determine the inverse photon effi- ciencies, in particular of the pair CH/Gerö band, thus, to verify the HYDKIN database and the plasma-related part of the ERO code. These experiments were performed with gas injection modules made of metal and minimised contact surface area to exclude any surface effects. Ex- perimental results, i.e. photon efficiencies for typical ohmic TEXTOR plasmas with electron temperatures of about 50 eV at the LCFS as well as HYDKIN calculations were presented last year.

The experimental data is in a fair agreement with modelling results obtained from the ERO code considering the uncertainties of the local plasma parameters at the location of the gas in- jection to be about 30%. The comparison of effective D/XB values from the pair C2/Swan band from the injection experiment and a new set of atomic and molecular data for hydrocar- bons form the HYDKIN database is an ongoing task and first modelling results are expected for the year 2008.

(18)

Fig. 1: 2D distribution of CD A-X light emission from the side and top view on the gas injection location. The distribution modelled with the ERO code for a local methane injection with a

comparable puff-in rate.

Reduction of the hydrocarbon flux in the outer divertor of JET during divertor detachment

High density L-mode discharges in JET with outer divertor detachment have shown a reduction of the CD A-X band light which is an indicator of chemical erosion of carbon-based plasma- facing materials. In order to distinguish if the reduction is attributed to a change of the meas- ured photon flux or the deduced particle flux an in-situ calibration with hydrocarbons was per- formed. CD4 was injected through GIM 14, a single valve in the LBSRP, into the near scrape- off layer of the outer divertor. The target plasma was a high density L-mode plasma with at- tachment-detachment oscillation. The injected amount of molecules leads to a comparable emission of CD photons in both phases indicating comparable photon efficiencies in recombin- ing and ionising plasmas. This gives evidence to a clear reduction of the chemical erosion of CFC in the phase with (semi-) detached outer divertor plasma. Exact quantification is subject of calibration of the line-of-sight of KS3 with view on the single gas inlet location.

However, the strong reduction is partially compensated by the reduction of the impinging ion flux when the divertor detaches. The erosion yield itself is only reduced if the particle flux of neutral atoms to the target is taken into account. These experimental results confirm the Roth formula for detached plasma conditions if both impinging fluxes, the atomic flux and ion flux, are considered.

The consequences for ITER can be described in the following way:

i) Detachment of the outer divertor actually leads to a reduction of the hydrocarbon flux.

ii) The overall erosion of graphite at the outer target plate is reduced.

(19)

KS3A view on local gas inlet (GIM14)

#70579 detached

CII

CD

attached D

de- D

C

CII

CD D

attached

extrinsic intrinsic

Fig. 2: Optical spectra and time traces of CII, CD and Dline intensities prior and during local injection of deuterated methane into the near scrape-off layer in high density L-mode plasmas.

Soft-layer disintegration in the inner divertor of JET

A number of new experiments in JET were performed to clarify the last step of the carbon mi- gration in the inner divertor leg. The appearance and disintegration of a-CH layers at entrance of the inner divertor pump duct – as found in post-mortem analysis – can be described as func- tion of primarily the magnetic configuration, the loss of energy per ELM and the history of previously performed shots – labelled "history effect.". So far disintegration of the topmost (soft) layer could be observed in the normal parameter range of JET; ablation of the harder and thicker layer below, which is build up during an experimental campaign, could only be ob- served in the case of ELMs close to the new limit for ITER (~ 1 MJ) (see below). QMB (Quartz Micro-Balance) measurements and hydrocarbon spectroscopy were applied as diagnos- tics to monitor the gross and net-erosion and the inner strike point was utilised as a tool to de- tect the location of the layer. The location of loosely bound layers in the inner divertor was found in most cases close to the corner region. Discharges in H-mode with type I ELMs and fast strike-point sweeps (4 cm) over a few centimetres on the horizontal base plate were ap- plied to disintegrate this soft-type layer. Figure 3 a shows typical spectrum recorded of a nar- row band cord observing the corner region of the inner during the disintegration of the layer.

QMB results have been reanalysed and sorted with respect to the ELM strength. The modelling of the local material transport – without ELMs – in the inner divertor leg with the ERO code is an ongoing task and relies on the availability of suitable background plasma reconstruction.

ERO results are expected for 2008.

(20)

Fig. 3: Spectroscopic footprint of a decomposed layer in JET. Decomposition occurs due to the large incident particle flux at the inner strike point which has been utilised to heat up the layer.

1.B. Quartz Micro-Balances in JET and TEXTOR QMB operation in JET

In the JET shut down starting April 2007 all the carriers of module02 have been removed re- motely from the machine and transported into the Beryllium Handling Facility (BeHF) to ex- change carbon protection tiles and to carry out operational checks of the Quartz Microbalance Systems attached. Required repair actions carried out if required and possible are listed below.

 The QMB 2 system (inner heated) on inner divertor carrier: This system was never op- erational since its installation in the shutdown 2004. Faulty cabling was found outside the JET vessel leading to the destruction of the ASIC (electronic chip) on the electronic board. This was confirmed by the tests in the BeHf. This system was replaced by a spare system.

 The QMB 3 (cooled system) on inner divertor carrier: This system showed always non reliable frequency behaviour from the very beginning after installation in the shutdown 2004. The reason is still not clear and further under investigation. The system could not be replaced.

 The QMB4 system (under wedge) was replaced as planned. It failed already due to over- heating, the system has on shutter, during the commissioning phase before campaign C15. The QMB was replaced by a spare system supplied by UKAEA. An additional shielding with an aperture reduced to 8mm in diameter was mounted in front of the lid of the DU.

 The QMB5 was replaced with a system calibrated at FZJ.

QMB benchmark system in TEXTOR

A new holder was developed and successfully tested to mount QMB-Systems to the limiter locks in TEXTOR. It is equipped with a gas inlet next to the quartz measuring the mass

(21)

W(Ra=10nm) W(Ra=180nm) W(Ra=20nm) Gr(Ra=350nm) Gr(Ra=70nm) Gr(Ra=700nm)

1 2 3 4 5 6

changes of the layers. Furthermore samples can be exposed for various post mortem surface analysis. Such the influence of the various chemistry of different gases can be studied for dif- ferent plasmas and various radial positions in the Scrape-Off layer of TEXTOR.

Fig. 4: View onto the coated front side of the QMB holder after exposure to CD4 gas injection in the SOL of TEXTOR.

1.C. Influence of surface morphology on carbon deposition

Carbon deposition on graphite and tungsten substrates with different surface morphology con- ditions was studied in a series of experiments in TEXTOR. Stripe-shaped samples of both ma- terials with different surface roughness (figure 5a) as well as tungsten samples with different pre-irradiation conditions (figure 5b) were installed on a roof-like limiter and exposed to the scrape-off layer plasma. It turned out that for both graphite and tungsten the net-deposition zone extends with increasing surface roughness. For tungsten probes pre-irradiated by carbon the amount of deposition increases with carbon concentration in the sample. The nano- structured surface layer on tungsten obtained by pre-irradiation by helium at elevated tempera- ture was removed during the exposure in TEXTOR.

Fig. 5: (a) Stripes of tungsten and graphite installed on a roof-like limiter after exposure in TEXTOR.

(b) Tungsten stripes with different pre-irradiation conditions after exposure in TEXTOR.

1 – 10%-20% of C, 2 – 20%-40% of C, 3 – 50%-70% of C, 4 – reference sample, 5 – He-irradiated at 1300 C, 6 – He-irradiated at 1000 C.

(22)

0 100 200 300 400 500 ELM energyDWELM[kJ]

1013 1014 1016

Carbon deposition per ELMDCELM[atoms/cm ]

2

Uncorrected data Corrected data Sputtering

1015

Arrhenius Sputtering +Arrhenius

Fig. 6: Six stripes of different carbon based materials installed on a roof-like test limiter.

1.D. Retention in Carbon-Fibre Composites (CFC) compared to graphite

The study of fuel retention in different carbon fibre reinforced composites (CFCs) was ex- tended in 2007. Five different materials were simultaneously exposed to the TEXTOR scrape- off layer plasma: the new ITER-grade CFC NB41, its predecessor NB31, JET CFC DMS780, Tore Supra CFC N11 and rough and smooth samples of fine-grain graphite EK98 (Figure 6).

The material samples were shaped as stripes and installed on a roof-like test limiter, allowing thus a range of deuterium fluences along the stripes in the same exposure. In contrast to 2006 experiments, the maximum fluence range was extended to about 51025 D/m2. The post- mortem analysis of the samples is underway.

Fig. 7: Illustrating the ELM-induced enhanced erosion. Amount of carbon deposited on the inner louvre QMB per single ELM as function of ELM stored energy drop. (□) Original uncorrected data;

() Data corrected for the influence of smaller ELMs; () Self-consistent fit function comprising (· · · ·) linear term for physical sputtering and (- - -) Arrhenius term for thermal decomposition.

(23)

Ip, Bt

Poloidal gap Toroidal

gap

69.6%

2.3%

36%

2.5%

41.6%

3.0%

27.3%

3.5%

b

c 20

a

1.E. ELM-induced enhanced erosion in the JET divertor

The impact of Edge Localized Modes (ELMs) carrying energies of up to 450 kJ on carbon ero- sion in the JET inner divertor is assessed by means of time resolved measurements using an in- situ quartz microbalance diagnostic. The inner target erosion is strongly non-linearly dependent on the ELM energy: A single 400 kJ ELM produces the same carbon erosion as ten 150 kJ events. Even when divertor target conditions are far from the threshold for carbon sublimation, ELMs enhance erosion of co-deposited amorphous carbon-deuterium layers in a non-linear manner, provoking much faster material re-dislocation. The observed Arrhenius behaviour of this erosion points to a thermally assisted process, strongly suggesting that large ELMs induce thermal decomposition of deposited layers on the inner divertor target surfaces.

1.F. Studies of carbon deposition and fuel accumulation in gaps of ITER-like castellated structures

Investigations of ITER-like castellated structures were continued exploring the impurity depo- sition and fuel accumulation in the gaps. Recently, the castellated structures of two shapes of cells: rectangular and roof-like were exposed in the SOL of TEXTOR. The shaped roof-like cells (Fig. 8a) were designed to reduce the carbon migration and corresponding fuel accumula- tion in the gaps. The test limiter with two shapes of castellation (Fig. 8b) was exposed for 16 repetitive identical NBI-heated discharges with a total plasma duration of 112 seconds. Post- exposure measurements were made using ion-beam and electron-beam analyses as well as sty- lus profiling.

Fig. 8: a) A view of a castellated limiter during exposure in TEXTOR; b) An example of the quantifica- tion of metal intermixing in the deposit (poloidal gap of rectangular cell is shown).

It was found that the amount of fuel deposited in the gaps parallel to the poloidal field direction (so-called poloidal gaps) was a factor of 2 less than that of conventional geometry. However, the detailed quantification of the carbon amount deposited in the gaps revealed only marginal advantages of the cells with new geometry which outlines a need in further shape optimization.

The same time, a strong metal intermixing was detected in the deposits on the poloidal gaps:

(24)

В

Periscope system (partially in

the port) TEXTOR vessel wall Upper port

Up to 70% atomic percent of tungsten (W) was found intermixed in the carbonaceous deposit (Fig. 8c). Metal intermixing in the deposited layers may significantly degrade the efficiency of cleaning techniques in ITER.

The investigations are being made in the framework of IEA-ITPA multi-machine joint experi- ments (Task DSOL-13) of the ITPA TG on Divertor Physics and SOL.

1.G. Protection of carbon deposition on mirrors for ITER diagnostics and active mitigation of carbon transport in remote areas

Mirrors will be used as a plasma-viewing components in all optical and laser diagnostics of ITER. Operating in the harsh ITER environment, mirrors may become deposited with the im- purities originating from eroded in-vessel components. The research and development of the techniques for deposition mitigation and cleaning of the mirrors are presently of paramount importance. The new concept of an active prevention from carbon deposition was recently ap- plied in TEXTOR. The Periscope-Upgrade mirror system, a prototype of the diagnostic duct was recently equipped with gas-feeding system. The Periscope Upgrade was installed in the Limiter Lock III transport system and exposed for the series of identical repetitive NBI-heated discharges. In total two exposures were made resulting in 42 discharges with a total duration of

~ 210 plasma seconds. During exposure the entire system was kept at a temperature of ~ 450

°C to 520 °C to ease the processes of chemical erosion of carbon deposits (if any) by deuterium atoms and ions. D2 was fed into the volume of a Periscope Upgrade during the discharges. This experiment has to be compared with one from the last year, where similar plasma conditions were applied but no gas feeding was organized. In both experiments molybdenum mirrors were used whose optical properties were characterized before and after exposure in TEXTOR.

Fig. 9: a) Schematic of the exposure of Periscope-Upgrade system in TEXTOR;

b) Photograph of the system.

The view of the first mirror exposed without gas feeding is presented in figure 10a. Photo- graphs of the first mirror exposed with gas feeding before and after exposures are provided in figure 10b and 10c respectively. In this case, no visible deposits were noticed on the mirror surface. Ion beam analyses (SIMS) made on these mirrors confirmed the complete suppression

(25)

a b c

Total Reflectivity, %

0 500 1000 1500 2000 2500 50

60 70 80 90 100

Before

After Wavelength, nm

of carbon deposition on the surface of the first mirror. Only negligible decrease in optical re- flectivity of the mirror exposed with gas feeding was noticed as illustrated in Fig. 10. It was proven therefore, that the gas feeding is an attractive option to protect the mirrors in ITER.

Fig. 10: Photographs of first mirrors of Periscope-Upgrade mirror system: a) the mirror exposed in 2006 without gas feeding; b) the mirror for experiment 2007, before exposure;

c) mirror for the experiment 2007 exposed with a gas feed.

Further investigations will be devoted to the optimization of the flow rates and an assessment of the cleaning efficiency of the gas fed into the Periscope volume. The investigations are be- ing made in the framework of IEA-ITPA multi-machine joint experiments (Task DIAG-2) of the ITPA TG on Diagnostics as a collaboration effort between FZJ and University of Basel.

Fig. 11: Total reflectivity of the first mirror of Periscope-Upgrade mirror system before and after exposure in TEXTOR.

1.H. ERO Modelling

The ERO code has been significantly advanced. The versions for various machine geometries (divertor, limiter, linear devices) have been merged into one common source code. Within the divertor version the reading of plasma parameters (which is an output of plasma codes such as B2-EIRENE and Edge-2D and therefore ordered according to flux surfaces) has been opti- mized. The output data from the plasma codes are now transferred to a rectangular grid, which can be handled by ERO in a much more effective manner leading to a reduction of computa- tional time by a factor of about 20.

(26)

-150 -100 -50 0 50 100 150 0

50 100 150 200 250

x [mm]

z (along the plasma column) [mm]

Emiss. int. [Ph/(sr*s*cm2)] of BeI. Total 3.45e+016.

0.5 1 1.5 2 2.5 3 3.5 4 4.5

x 1013 00 50 100 150 200

0.2 0.4 0.6 0.8 1

Profile normalized to its maximum

z (along plasma column) [mm]

Exp. BeI ERO BeI

ERO (no el.c.) BeI

The calculation of collisions between neutrals has been further improved. For the PISCES-B calculations this leads to a much better agreement between measured and modelled BeI and BeII emission patterns. From this it has been concluded that the major disagreements between observed impurity transport and modelling have been solved.

Fig. 12: Left: 2D pattern of the modelled BeI emission in PISCES-B with the Be originating from seed- ing into the plasma. The seeding location (oven) is indicated with red lines. Right: Modelled profiles (with and w/o elastic neutral collisions) of BeI emission along the axis of PISCES-B.

For comparison also the observed profile is shown.

Figure 12 shows an example of modelled axial BeI profiles with and without neutral collisions in comparison with the experimental observation. The Be is launched into the plasma at x = 175 mm and z = 150 mm simulating the evaporation from the oven. It can be concluded that the Be transport in the plasma is now described satisfactorily. The coupling of ERO with the 3D Monte Carlo code SDTrimSP, which describes the transport of ions in matter, has been finished. A technical user manual for the coupled code package has been written. In the frame of ITER modelling, the erosion and redeposition of carbon, beryllium and tungsten has been analysed for the outer divertor plate. At first it has been assumed that the impurity influx from the main plasma can be neglected. The following table summarises the total gross erosion /w/o redeposition) and the amount of redeposition.

Target Material Total gross erosion Redeposition

C 7.7  1022 C/s 98%

Be 3.3  1022 Be/s 90%

W 2.8  1016 W/s 95%

It is seen that tungsten suffers from the lowest erosion whereas the total gross erosion of car- bon and beryllium is about 6 orders of magnitude larger. In case of carbon chemical sputtering is an additional erosion mechanism to physical sputtering resulting in about a factor of 2 larger gross erosion of carbon compared to beryllium. Taking into account redeposition, the carbon net erosion is about 150 nm/s in the maximum However, including a constant beryllium con-

(27)

0 2 4 6 8 10 12 14 16 18 20 0

2x103 4x103 6x103 8x103 1x104

deposited particles

distance s along gap (mm) R = 0 R = 0.1 R = 0.5 R = 0.9

centration in the incoming plasma flux of 0.1% this value is reduced by a factor of about 3.

Further modelling is planned to take into account the most recent profiles of beryllium in the background plasma (instead of a constant value assumed so far). Also a TriDyn-based surface model will be used to calculate the formation of mixed layers. Improved estimations of target lifetime and tritium retention (both for inner and outer divertor plate) will be done.

First modelling of deposition profiles inside gaps of castellated surface structures has been per- formed. For benchmarking, the experimental results from a castellated test limiter, which has been exposed to the edge plasma in TEXTOR, have been used. Similar to the ITER design, the depth of the gap is chosen to 10 mm and the width to 0.5 mm. In the modelling the (one di- mensional) gap surface is divided in total into 41 cells of 0.5 mm length. A source of neutral particles (e.g. generated by erosion due to the background plasma), which can enter the gap is assumed to be located at the upper right edge of the gap. The particles start at the location of their origin with a cosine distribution relative to the surface normal. Only a certain part of the starting particles is able to enter the gap. As a first attempt particles entering the gap are as- sumed to move along straight lines without any collisions with neutrals. If a particle hits the inside wall of the gap it is reflected with a probability according to a reflection coefficient R, which is given as input parameter of the modelling. Whether a specific test particle is reflected is decided via a random number Nran (between 0 and 1), which is compared with the reflection coefficient R. If R is greater than the random number Nran reflection takes place, if not the par- ticle is deposited. Reflected particles are re-launched again with a cosine distribution.

Fig. 13: Modelled deposition profiles inside the gap assuming various reflection coefficients.

Figure 13 shows modelled deposition profiles along the inner surface of the gap for different reflection coefficients. The s-coordinate starts at the upper left corner of the gap. Altogether 30.000 particles are launched at three different positions (10.000 particles at s = 20 mm, 20.25 mm and 20.5 mm) of the last surface cell. With R = 0, deposition inside the gap just takes place at the vertical side in opposite of the source. Particles not deposited (34%) leave the gap di- rectly without hitting a gap surface. With reflection coefficients between 0.1 and 0.9 the depo- sition profile becomes broader and a certain amount of particles is also deposited on the gap side where the source is located. However, due to the large ratio of the length of the side wall to the bottom of the gap the probability of deposition at the bottom of the gap is negligible. A first comparison with the observations at the TEXTOR test limiter shows that a reasonable match with (normalised) modelled deposition profiles is obtained if a reflection coefficient of

(28)

R = 0.5 is assumed in the modelling. Further development of the gap modelling is planned (e.g.

including neutral collisions, angle-dependent reflection coefficients).

1.J. ICRF Wall Conditioning in TEXTOR

Ion Cyclotron Wall Conditioning (ICWC) is a promising conditioning technique for present and next generation superconducting fusion machines in the presence of permanent high mag- netic field. The ICWC technique was recently approved for integration into the ITER baseline and this scenario was added to "Functional Requirements" of the ITER ICRF system. There- fore, further development of the ITER relevant ICWC scenarios is an important and urgent task.

The present studies of ICWC have been performed in TEXTOR in the presence of toroidal (0.22.35 T) and vertical (00.04 T) magnetic fields using the conventional ICRF antennas without modifications in hardware. The conditioning ICRF plasmas (ne  21017 m-3, Te  5 eV) were reliably produced in He and N2/H2-mixtures in the pressure range (24)10-2 Pa and coupled RF power 10100 kW from two ICRF antennas powered in a single-pulse mode (RF-tot  7 s). The ICWC effect was studied by mass spectroscopy for the removal rate several marker masses. Prior to each ICRF conditioning shot, a standard glow discharge procedure was used for wall cleaning (in He) and wall pre-loading with a reproducible amount of D2/Ar- mixture.

ICWC in inert (He) gas

Several factors derived from theory of plasma wave propagation/absorption and antenna- plasma coupling in low density/temperature plasmas were identified which could have a crucial impact on the conditioning efficiency, e.g. the removal rate:

 RF power coupled to the plasmas;

 ICWC at higher cyclotron harmonics ( = nci, n >> 1) (Fig. 14);

 ICWC with the additional steady state or triangular-modulated vertical magnetic field superimposed on the toroidal field (BV << BT) (Fig. 15).

All the observed effects will be a subject for further analysis and optimization.

ICWC in reactive gas mixtures (N2+H2 and He+NH3)

The gas mixtures N2/(N2+ H2) = 0 – 0.8 were produced by N2/H2 continuously flows and an additional N2 puffing up to 30 mbar·l during RF pulse. Total pressure drops during ICRF dis- charge (Fig. 15). N2 and H2 are mainly consumed. HD and D2 partial pressures increase tempo- rary after ICRF plasma. Enhanced outgasing of HD and D2 was observed with additional N2

puffing. Total gas balance was negative in ICWC with He/NH3 gas mixture when a small amount of NH3 (8 mbar·l) was puffed during ICRF plasma (PRF-G ~ 100 kW). Strong increase of partial pressure for mass m = 18 after the ICRF discharge is probably due to NH2D outgas-

(29)

Partial pressure [a.u.]

m=40

RF plasma

-20 0 20 40 60 80 100 120 140 160

0 1 2 3 4 5 6

7 RF plasma

Total pressure [10-4 mbar]

Time [s]

Bt=2.3T,Bv=0, #102948 Bt=2.3T, Bv=0, #102962

ing. Partial pressure of HCN molecules is small that indicates low carbon removal rate. Further optimization with respect to removal rate of carbon is required.

Fig. 14: Time evolution of the Ar partial pressure during and after RF conditioning pulse as a function of toroidal and vertical magnetic fields for constant RF power at generators

(PRF-G  130 kW, f = 29 MHz).

Fig. 15: Time evolution total pressure during and after RF conditioning pulse for two conditions:

#102948 PRF-G ~ 40 kW, continuous H2 and N2 flow #102962 PRF-G ~ 80 kW, continuous H2 and N2 flow and pulse puffing of N2 15 mbar•l.

2. High-Z materials for plasma-facing components

2.A. Atomic and molecular data: the case of atomic tungsten

To convert measured spectroscopic line intensities (Itot) to fluxes (p) and densities according to p = S/XB  Itot, the conversion factor S/XB (ionisations per photon) has to be known. Vari- ous theoretical codes exist to determine these values – like GKU, R-matrix, databases (ADAS) etc. TEXTOR offers an excellent possibility to compare those theoretical data with experimen- tal results.

For tungsten the values for energies and radiative transitions have been compiled from a recent version of the NIST data base. Unfortunately the level identifications are given for the three

(30)

lowest configurations (not all terms) only. The ionisation rate coefficients were recalculated by ATOM for the lowest configurations 5d46s2 and 5d5 6s. Calculations of the excitation cross sections meet problems due to a complicated coupling scheme and configuration mixing. For many levels the identification is unknown and, therefore, the semi-empirical Regemorter for- mula was used with the coronal approximation for excitation only from the group of "ground"

levels. Finally, S/XB values for several W I lines in the visible and UV could be calculated and compared with experimental values from TEXTOR and PSI-2 (MPI Berlin). Because of the definition above the inverse photon fluxes should depend linearly on the respective S/XBs for one experimental condition. Fig. 16 shows this behaviour for a number of lines observed. The best linear dependence was observed for plasmas with Te = 1 eV and a ground state population represented by a Maxwellian distribution with TW = 0.3 eV. Whereas for PSI-2 this plasma temperature is a reasonable assumption a value of 40 eV is much more reasonable. The reason for this discrepancy is not yet clear.

Fig. 16: Spectroscopic W I data from TEXTOR (red, blue) and PSI-2 (magenta).

2.B. Melting and melt layer propagation

In order to perform experiments with tungsten at high particle fluxes and intense heat loads, a brush limiter had been constructed and built at the Efremov Institute (St. Petersburg, Russia), which consists of cut tungsten slices on a copper base target. The advantage of this limiter is that it can be disassembled in slices along one (here toroidal) direction for further investigation of deposition in the gaps in the other (here poloidal) direction. Experiments have been carried out with a fixed width of 0.5 mm on one half and a variable width from 0.3 mm to 0.8 mm on the other half.

Detailed surface studies performed after two experiments with a tungsten limiter have provided a deep insight into the deposition, fuel inventory and material mixing occurring in the gaps of castellation. It has been shown that the deposition characterised by short decay length (~ 1.5

Referenzen

ÄHNLICHE DOKUMENTE

Open Access This article is licensed under a Creative Com- mons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in

• All the Procurement Arrangements for the ITER Vacuum Vessel and In-Vessel components have been signed and Procurement Specifications for materials have been prepared and agreed

Hydrogenic retention studies in the all graphite first wall DIII-D tokamak during ELM-y and resonant magnetic perturbation ELM-supressed H-mode discharges 19th

During 2013, a development has been started at FZJ to investigate synergistic effects which govern the performance of plasma facing materials under fusion relevant loading

The optimization of the plasma source for JULE-PSI is currently being carried out on the pilot- experiment PSI-2 outside the controlled area of HML. For this linear device, also

Modelling of deposition at the bottom of gaps in TEXTOR experiments (2011) 13th International Workshop on Plasma-Facing Materials and Components for Fusion

Board Member: Thernlunds AB, UN Foundation and the Whitaker Peace and Development Initiative. Principal work experience and other information: President and CEO

In Japan, company data in their primary form are mainly available in four types: uncon- solidated annual accounts according to the Commercial Code, reports according to the