Search

Er2O3 sesquioxide thin films in the uncommon, monoclinic phase were produced for the first time by filtered cathodic arc deposition on Eurofer steel substrates. This was achieved by applying a sample bias of −250 V and deposition temperatures ⩽400 °C. X-ray diffraction texture analysis revealed a strong preferred orientation of the B-phase crystallites. Deposition at 600 °C without bias voltage resulted in thin films showing the regular cubic phase. A dense, columnar structure was found for both phases by scanning transmission electron microscopy.

In-vessel inspections of AUG reveal arc tracks at different locations. This paper presents investigation of arcing at the inner transition and retention module where arc tracks were found around the whole toroidal circumference. A representative tile was investigated with profilometry, SEM, EDX, RBS, and colourimetry. Type 2 arc tracks with an area fraction of 12% were found in a 10 mm wide region near to the leading edge. Here the tungsten coating (3–4 μm) is removed and a small amount of carbon (max. 1–2 μm) is eroded. Droplets are detected at the surface. The region of the inner divertor where the arcs are observed shows a local enhancement of plasma density, n e ≈ 10 20 m −3 , and neutral pressure, 0.1 mbar, favouring the arc ignition. A high sheath potential at the beginning of an ELM might trigger the arc. The arc pattern and the target load as measured with a fast IR-camera coincide. 2 g of tungsten are eroded in the inner divertor. If 10% of the material is eroded as droplets, 10 8 droplets (5 μm diameter) are ejected.

Finely dispersed Ti-doped isotropic graphites with 4 at.% Ti have been manufactured using synthetic mesophase pitch ‘AR’ as raw material. These new materials show a thermal conductivity at room temperature of ∼200 W/mK and flexural strength close to 100 MPa. Measurement of the total erosion yield by deuterium bombardment at ion energies and sample temperatures for which pure carbon shows maximum values, resulted in a reduction of at least a factor of 4, mainly due to dopant enrichment at the surface caused by preferential erosion of carbon. In addition, ITER relevant thermal shock loads were applied with an energetic electron beam at the JUDITH facility. The results demonstrated a significantly improved performance of Ti-doped graphite compared to pure graphite. Finally, Ti-doped graphite was successfully brazed to a CuCrZr block using a Mo interlayer. These results let assume that Ti-doped graphite can be a promising armour material for divertor plasma-facing components.

W fibre-reinforced Cu shows great promise to improve the mechanical performance at high-temperatures compared to conventional Cu-based alloys. Focus was placed on the optimization of the interface to achieve an enhanced adhesion between W fibre and Cu matrix. The interfacial properties were investigated through pull-out measurements of single matrix-coated fibres for different interfacial concepts. The interfacial adhesion of W and Cu is determined solely through mechanical interlocking. Interdiffusion and segregations experiment showed that there are no interface reactions between W and Cu at elevated temperatures. From the investigated interfacial concepts, a stepwise graded transition interface with additional heat treatment was found to achieve the highest interfacial shear strength. The thermal stability of the MMC in thermal cycling tests can be assured by depositing a stepwise graded transition between W fibre and Cu matrix.

The erosion behaviour of doped carbon materials was investigated in deuterium low pressure discharges. Therefore, amorphous carbon films on silicon substrates (film thickness ∼1 μm) with varying dopant species (V, Ti andW), dopant concentration (between 1 at% and 8.5 at%) and dopant distribution were exposed to the plasma. The dopant distribution is either atomically disperse or, by heating up to 1100 K, carbide grains in nm size are formed. All important plasma parameters are obtained by a combination of several diagnostic techniques like optical emission spectroscopy, energy resolved mass spectroscopy and Langmuir probe measurements. The morphology of the plasma treated surfaces was investigated afterwards by scanning electron microscopy and atomic force microscopy, the layer composition by ion beam analysis. Every investigated sample shows a strong decrease of the erosion yield with increasing fluence. The achieved erosion yields are clearly below the yield of undoped amorphous carbon and depend strongly on the dopant concentration. The lowest yields were achieved with the samples with atomically disperse dopant distribution due to metal enrichment at the surface. (© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)

Amorphous carbon layers doped with 8.5 at.% titanium (a-C:Ti) were produced by dual-source magnetron sputter deposition. To investigate the local atomic environment of Ti, EXAFS analysis was performed for as-deposited layers and after annealing to 700, 1100 and 1300 K, as well as for Ti and TiC bulk samples. After deposition of the a-C:Ti films, most Ti atoms are homogeneously distributed in an amorphous carbon matrix. Annealing to 700 K slightly increases the order in the Ti environment, but TiC-like crystallization is hardly existent. After annealing to 1100 K, clear correspondence to the TiC standard is observed. Annealing to 1300 K further increases the order and TiC crystallite size. Fitting of EXAFS data to a theoretical model was used to derive quantitative parameters for a-C:Ti layers, e.g., Ti–C and Ti–Ti distances and ordering parameters.

Carbon layers doped with Ti, V, W, and Zr to about 15 at.% were produced by means of magnetron sputter deposition, which distributes metal atoms homogeneously in the amorphous carbon matrix. The effects of thermal annealing to temperatures of 500–1300 K on the phase, crystallinity and distribution of the dopant were determined. X-ray diffraction (XRD), MeV ion beam analysis (IBA), and extended X-ray absorption fine-structure spectroscopy (EXAFS) were used for characterization. The three techniques deliver complementary information on different length scales of the diffusion and crystallization. All four metal dopants are in carbidic state with crystallites on the nanometer scale after annealing to 1100 K. Dopant diffusion of no more than 20 nm, even after heating at 1300 K for 2 h, was observed.

Fe–W binary mixed layers were prepared as a model system for reduced-activation ferritic–martensitic (RAFM) steel for studying their dynamic erosion behavior resulting from energetic deuterium (D) irradiation. This investigation aims toward an assessment of RAFM steels as plasma-facing material. The surface composition of the model layers is modified by D irradiation. W is enriched at the surface with D irradiation fluence due to the preferential sputtering of Fe. It depends on the D impinging energy as well as the initial W fraction of the Fe–W layer. No significant development of surface topography was observed within the examined conditions. The erosion yield of a Fe–W layer is comparable to that of pure Fe in the low-fluence range and decreases with increasing D fluence. These results indicate that the dynamic change of erosion yield is significantly correlated with the surface W enrichment.

ASDEX Upgrade became a full tungsten experiment in 2007 by coating its graphite plasma facing components with tungsten. In 2013 a redesigned solid tungsten divertor, Div-III, was installed and came into operation in 2014. The redesign of the outer divertor geometry provided the opportunity to increase the pumping efficiency in the lower divertor by increasing the gap between divertor and vessel. In parallel, a by-pass was installed into the cryo-pump in the divertor region allowing adapting of the pumping speed to the required edge density.Safe divertor operation and heat removal becomes more and more significant for future fusion devices. This requires developing 'tools' for divertor heat load control and to optimize the divertor design. The new divertor manipulator, DIM-II, allows retracting a relevant part of the outer divertor into a target exchange box without venting ASDEX Upgrade. Different front-ends can be installed and exposed to the plasma. At present, front-ends for probe exposition, gas puffing, electrical probes and actively cooled prototype targets are under construction.The installation of solid tungsten, the control of the pumping speed and the flexibility for testing divertor modifications on a weekly base is a unique feature of ASDEX Upgrade and offers together with the extended set of diagnostics the possibility to investigate dedicated questions for a future divertor design.

We present the results of erosion measurements on actively cooled tungsten samples at quasi-constant surface temperature conditions performed in the high heat flux facility GLADIS. The samples were exposed to a H beam at a central power density of 10 MW/m 2 up to a fluence of 10 26 m −2 . We observe a weak temperature dependence of the erosion yield. The data are compared with similar data obtained from loading with a H beam with He admixture. Both datasets are analysed in a probabilistic approach. We obtain activation energies of 0.04 eV and 0.06 eV for the cases with and without He, respectively.

In order to provide a solid basis for the correlation of microstructure and hydrogen isotope retention in tungsten, reference samples with different microstructures were prepared from a single batch of polycrystalline tungsten by standardised polishing and heat treatment procedures. Representative samples were analysed by scanning electron microscopy and scanning transmission electron microscopy as well as by electron backscatter diffraction. We show that if the annealing temperature is increased from 1 200 to 1 500 K, practically only the density of dislocations and grain boundaries with very small misorientations of less than 2° is reduced, while for annealing at 1 700 and 2 000 K, also the density of high-angle grain boundaries is reduced due to grain growth. Furthermore, the dislocation density is reduced by nearly two orders of magnitude compared to tungsten annealed at 1 200 K. We also comment on two different textures on the front and rear side of the samples that were observed both by X-ray diffraction and EBSD.

Carbon films doped with Ti, V, W, Zr, Cr, and Cu were produced by magnetron sputtering. To predict the film composition, the deposition rates were systematically studied as a function of discharge power and working pressure. The achieved dopant concentrations range from 20 down to 1 at.%. The films are laterally homogeneously doped and show columnar growth. The dopant distribution is not thermally stable. After heating at 1100 K, the carbides TiC, VC, WC, ZrC, and Cr 3 C 2 are definitely present and their grain size is on the nanometre scale. Cu segregates out. There are strong indications of the formation of carbides already during deposition.

Significant progress has been made on ASDEX Upgrade during the last two years in the basic understanding of transport, in the extension of the improved H-mode in parameter space and towards an integrated operating scenario and in the development of control methods for major performance limiting instabilities. The important features were the understanding of particle transport and the control of impurity accumulation based on it, the satisfactory operation with predominantly tungsten-clad walls, the improved H-mode operation over density ranges and for temperature ratios covering (non-simultaneously) the ITER requirements on ν*, n/nGW and Te/Ti, the ELM frequency control by pellet injection and the optimization of NTM suppression by DC-ECCD through variation of the launching angle. From these experiments an integrated scenario has emerged which extrapolates to a 50% improvement in n T τ or a 30% reduction of the required current when compared with the ITER base-line assumptions, with moderately peaked electron and controllable high-Z density profiles.

The oxidative erosion of seven types of graphite has been investigated by heating in air at temperatures between 600 and 1000 K. The specimens include pyrolytic graphite, fine-grain graphites, carbon-fibre composites (CFC), and graphites doped with Si and Ti. The weight loss was measured using a microbalance, the surface morphology by scanning electron microscopy, and the composition of the surface layer by MeV ion beam techniques. Pyrolytic graphite is least affected by erosion, while pure and Si-doped CFCs erode particularly fast. Typical erosion rates for specimens with a surface area of ≃4 cm 2 are below 0.2 μg/m 2 s at 600 K for all graphite types, and at 900 K range from 0.34 mg/m 2 s for pyrolytic graphite to about 9 mg/m 2 s for the strongest eroding types. The temperature dependence of the erosion rate of all types of graphite studied is well described by an activation energy of 1.7 eV. The erosion rates of these graphites are by far lower than the removal rates for deposited amorphous hydrocarbon layers. In contrast to all other types, the Ti-doped graphite absorbs a significant amount of oxygen reaching up to ≃5% of its original mass. Once the oxygen uptake is saturated, it erodes with rates similar to those of the strongest eroding types.

Magnetron-sputtered layers consisting of carbon and metal (W, Ti, V) were produced with 0–20 at.% metal concentration. The mixed layers were characterised by RBS, SEM, XRD and XPS, and exposed to D 3 + ions of 30 eV/D at temperatures between 77 and 1100 K. The chemical erosion yield was investigated by mass spectrometry and RBS. Above RT (∼300 K), the CD4 production yield for pure C layers exhibits a maximum around 750 K, which decreases with increasing metal concentration. For more than ∼3 at.% W, ∼6 at.% V and ∼7 at.% Ti, the maximum vanishes and the CD4 yield continuously diminishes with temperature. A decrease of the activation energy for ion-induced hydrogen release by adding dopants is responsible for the decrease. The CD4 yield at RT increases depending on metal and concentration, because the distribution of the erosion products is changed. Enrichment of metal on the surface with ion fluence is observed.

Numerous traces of arcs have been found on W-covered graphite tiles of ASDEX Upgrade after exposure. The distributions of number density, lengths and orientation are calculated and compared to pure graphite tiles at comparable locations. It was established that arcs perforate a 1 μm tungsten layer down to the carbon substrate. The amount of removal should rise with arc current, but a surface fraction of about 8% is eroded at 10 A already. At tiles of the divertor baffle the layer is continuously removed along the entire track pointing to higher currents. The carbon of the stripped parts is subject to subsequent erosion processes. The distribution of materials in and around arc tracks was investigated by sputter depth profiling (SIMS and AES) and the characteristic geometry was studied using an electron microscope. Observations are interpreted using results from laboratory vacuum arcs on the same material.

The tungsten programme in ASDEX Upgrade is pursued towards a full high-Z device. The spectroscopic diagnostic of W has been extended and refined and the cooling factor of W has been re-evaluated. The W coated surfaces now represent a fraction of 65% of all plasma facing components (24.8 m(2)). The only two major components that are not yet coated are the strikepoint region of the lower divertor as well as the limiters at the low field side. While extending the W surfaces, the W concentration and the discharge behaviour have changed gradually pointing to critical issues when operating with a W wall: anomalous transport in the plasma centre should not be too low, otherwise neoclassical accumulation can occur. One very successful remedy is the addition of central RF heating at the 20-30% level. Regimes with low ELM activity show increased impurity concentration over the whole plasma radius. These discharges can be cured by increasing the ELM frequency through pellet ELM pacemaking or by higher heating power. Moderate gas puffing also mitigates the impurity influx and penetration, however, at the expense of lower confinement. The erosion yield at the low field side guard limiter can be as high as 10(-3) and fast particle losses from NBI were identified to contribute a significant part to the W sputtering. Discharges run in the upper W coated divertor do not show higher W concentrations than comparable discharges in the lower C based divertor. According to impurity transport calculations no strong high-Z accumulation is expected for the ITER standard scenario as long as the anomalous transport is at least as high as the neoclassical one.

The process development and characterization of titanium nitride (TiN) as a diffusion barrier for silver (Ag) metallizations were discussed. The structure of Ag/TiN/SiO"2/Si and Ag/TiN/Si metallizations under N"2/H"2 thermal annealing at 300-650 ^oC were investigated with sheet resistance measurement, X-ray diffraction (XRD), focused ion beam-scanning electron microscopy (FIB-SEM), atomic force microscope (AFM) and X-ray photoelectron spectroscopy (XPS) and compared to as-deposited multilayers. For 12 nm TiN barrier, no change of sheet resistance was observed after annealing at 600 ^oC, but an abrupt rise appeared at 650 ^oC annealing. FIB-SEM revealed that this is due to the surface morphology change after annealing of the samples at different temperatures and the destruction of the TiN film.

To minimize radiation losses of the plasma during long pulse operation, the first wall protection of WENDELSTEIN 7-X demands low-Z plasma facing materials. In addition to carbon materials on high heat flux loaded components, 300–500 μm thick vacuum plasma-sprayed (VPS) layers of boron carbide (B4C) are considered as coatings on large actively cooled stainless steel panels. In order to evaluate the behaviour relevant to the expected plasma wall interactions, an extensive material characterization and test programme was executed. Also included is the examination of the industrial manufacturing and the suitability of these coatings as 70 m2 first wall plasma facing material for W7-X. To improve the adhesion of thick B4C layers on stainless steel, different interlayers have been investigated. Based on the results, a VPS-based coating technique was identified which is suitable to manufacture the B4C protection layers on stainless steel wall panels of W7-X.

Surface roughening of polycrystalline stainless steel mirrors due to hydrogen bombardment was studied as a function of ion fluence and energy (1–4.3 × 10 24 H/m 2 , 300–1500 eV/H). A strong micro-relief (100 nm scale) on single grains and a strong variation of the erosion depth between different grains were observed by scanning electron microscopy (SEM). A broad variation of the in-grain micro-relief was found to be independent of impact energy. The sputtering yield (Y) depends on the grain orientation, varying by a factor of about 2 for all impact energies. The correlation between surface morphology and orientation of the single grains was investigated using electron back-scattering diffraction (EBSD). Grains with nearly (1 1 1) surface orientation do not show any significant in-grain micro-relief, although they have nearly the highest Y. For other orientations close to low indexed surfaces, e.g., (3 1 1), no correlation was found between Y and roughness.

Using a filtered vacuum arc deposition device, stoichiometric aluminum oxide (Al2O3) films, with thickness ranging from 20 nm to several microns, were produced under various substrate bias voltages and temperatures. Analysis of the resulting alumina crystal structures was performed with transmission electron microscopy, Fourier transform infrared spectroscopy and X-ray diffraction. Depending on the negative substrate bias voltage, the deposition temperature required to form α-Al2O3 could be reduced. A crystal phase diagram showing the effect of bias and temperature is presented. Also, preliminary hydrogen permeation measurements of these coatings deposited on thin palladium foil show a good barrier performance as compared with uncoated samples.

Plasma sprayed B4C-layers considered as wall coatings for the W7X stellarator have been studied during and after exposure to TEXTOR and after arcing experiments in vacuum. Arcing through the B4C layer occurred favoured by high power fluxes and not restricted to less stable phases. But this arcing implies an especially noisy scrape-off layer (SOL). Instead of moving retrograde in the external magnetic field, the arc spot on the B4C-layer sticks to the same location for its whole lifetime. Consequently, the arc erodes the entire B4C layer, finally burning down to the Cu substrate. In the neighbourhood of craters the surface contains Cu originating from those craters. This material, hauled to the surface by the arc, is subject to subsequent erosion, transport, and redeposition by the SOL-plasma. The behaviour of arcs on B4C is most probably caused by the peculiar temperature dependences of the electrical and heat conductivity of B4C. 2003 Elsevier Science B.V. All rights reserved.

In ASDEX Upgrade, tungsten was employed as plasma facing material at the central column heat shield in the plasma main chamber. The campaign averaged tungsten erosion flux was determined by measuring the difference of the W-layer thickness before and after the experimental campaign using ion beam analysis methods. The observed lateral variation and the total amount of eroded tungsten are attributed to erosion by impact of ions from the scrape-off layer plasma. Migration and redeposition of eroded tungsten were investigated by quantitative analysis of deposited tungsten on collector probes and wall samples. The obtained results, as well as the spectroscopically observed low tungsten plasma penetration probability, indicate that a major fraction of the eroded tungsten migrates predominantly through direct transport channels in the outer plasma scrape-off layer without entering the confined plasma.

The possibilities for optimization of doped fine grain graphites with high thermal conductivity and high thermal shock resistance are demonstrated at laboratory scale. A mixture of MCMB powder and different carbides (B4C, TiC, VC, ZrC and WC) was used as starting material. VC acts as catalyst of the graphitization at the lowest temperature, and ZrC is the most effective catalyst of all investigated carbides. A direct proportionality between the mean crystallite height, Lc, and the thermal conductivity at room temperature was found for all materials except for the B4C- and the ZrC-doped graphites. With increasing graphitization temperature the open porosity of all doped materials becomes gradually closed, suggesting the existence of a diffusion mechanism responsible for both the catalytic effect and the closing of the open porosity. The addition of carbides does not strongly influence the mechanical properties of pure graphite. A high ratio flexural strength to Young’s modulus was achieved.

In the tokamak experiment ASDEX Upgrade, the investigation of tungsten as a first wall material is an ongoing research project. In a step-by-step strategy, the tungsten covered surface area is increased from campaign to campaign. For this purpose an industrial-scale method for coating graphite with micrometer tungsten films had to be identified. Test coatings deposited by magnetron sputtering and by plasma-arc deposition were compared. By X-ray analysis it was found that sputter-deposited coatings suffer from high compressive stress (1.7 GPa). This leads to delamination when a film thickness of about 3 μm is exceeded. For arc-deposited coatings, a compressive stress value of 0.5 GPa was determined and no delamination occurred up to the maximum film thicknesses investigated, i.e. 10 μm. Upon thermal loading, none of the arc-deposited coatings failed up to the melting condition, while one sputter-coating delaminated. First results on similar investigations employing CFC substrates are presented.

Up to now, blister formation on rough or technical tungsten surfaces exposed to hydrogen isotope plasma was believed to be completely suppressed. The few dedicated experiments on this issue that can be found in literature appear to support that claim. Using a novel technique of 3D difference imaging of tungsten surfaces, we now demonstrate that roughness introduced by chemical etching, i.e. without the associated mechanical deformation layer introduced by grinding, only moderately reduces blistering. A technical surface with comparable roughness produced by precision grinding (R a ≤ 1.6 µm) led to a strong reduction in blister size and density, but blisters were found nevertheless. In this article we give a detailed description of the investigated rough W surfaces and present a statistical evaluation of blistering on these surfaces after exposure to a low-temperature deuterium plasma.

We performed investigations to determine the lifetime of tungsten coatings on carbon under high heat flux with respect to tungsten carbide formation. Tungsten coatings with a molybdenum interlayer were deposited by combined magnetron sputtering and ion implantation. These coatings were heat treated in a vacuum furnace. Subsequently the fraction of the coating thickness which had converted into carbide, was determined by focussed ion beam preparation and electron microscopy. This thickness was found to extend from 10 μm to 18 μm for heat treatments between 2 and 20 h at 1350 °C. Such coatings were then exposed to high heat flux pulses of 16.5 MW/m 2 with pulse durations of 1.5 s. A clear threshold behaviour was found: While after 2 h of heat treatment no coating failure was found, the delaminated surface fraction was up to 5% after treatment for 5 h and up to 25% after treatment for 20 h at 1350 °C.

We discuss the surface morphology modification of W samples observed after simultaneous heat and particle loading using a mixed H/He particle beam with a He concentration of 1 at.%. The applied heat flux of 10 MW/m 2 is representative for the normal operation of the divertor of DEMO or a power plant. The long pulse high heat flux experiments on actively water-cooled W samples were performed in the GLADIS facility at surface temperatures between 600 °C and 2000 °C. This allows together with the applied total fluences between 1 × 10 24 m −2 and 1 × 10 26 m −2 the temperature- and fluence dependent study of the growing nano-structures. We analyse in detail the surface modifications up to a depth of several μm by scanning electron microscopy combined with focussed ion beam preparation. The hydrogen and helium release of the samples is analysed by long term thermal desorption spectroscopy and compared with the prediction of a diffusion trapping model.

The effect of surface roughness on the long-term erosion patterns of tungsten coatings was investigated in the outer strike-point region of ASDEX Upgrade during its 2010–11 plasma operations. The net erosion rates of rough coatings (Ra = 5–6 μm) were three to seven times smaller than those of smooth coatings (Ra = 0.4–0.8 μm). This is because rough surfaces are largely modified and damaged in the microscopic scale but the material is re-deposited together with boron, deuterium and carbon on the shadowed sides of the most prominent surface features. In addition, we observed that W coatings were eroded on average at a rate of 0.03 nm s−1, which was three to four times smaller than the value for Cr, simulating here steel.

Blistering of tungsten (W) surfaces due to deuterium (D) implantation was investigated by a sequence of exposures at two different temperatures—230 and 450 K—and by the reversed sequence. A single exposure at 230 K leads to a high areal density of small dome-shaped blisters (up to 3 µm in diameter) together with much smaller flat-topped structures, while a single 450-K exposure produces large dome-shaped blisters up to 40 µm in diameter without the flat-topped structures. Most of the small dome-shaped blisters from 230 K exposure disappeared after annealing at 450 K for 17 h, but survived and even grew in size if the surface was exposed to D plasma during annealing. Sequential exposure at the two temperatures reveals a non-commutative behaviour: after a first exposure at 450 K the second exposure at 230 K leads to superposition of the observed blister structures without changing the large blisters from the first exposure. By contrast, a first exposure at 230 K almost completely suppresses the formation of large blisters during a second exposure at 450 K. Obviously, the presence of the small blisters strongly influences the penetration of D into the W bulk.

In fusion machines (like e.g. ASDEX Upgrade and JET) equipped with W-coated plasma facing components (PFCs) arc tracks are observed post-mortem and W particles are identified as a component of the dust. Results from laboratory arcing of W-coated PFC-material concerning the size, velocity and direction of macro-particles are combined with model calculations of the heating, cooling and evaporation of W particles flying through a simplified scrape-off layer (SOL) plasma to assess the role of arc-produced particles as a source of W impurities. W atoms or ions eroded from a macro-particle flying across the SOL plasma are not subject to prompt re-deposition onto a PFC. Therefore, macro-particles constitute an essential source of W in the SOL and may even result in an impurity input into the region of confined plasma if the particle succeeds in pathing the SOL and crosses the last closed flux surface. The main result of the modelling is that a W particle from the large-size end of the distribution obtained in laboratory arcing and having a typical velocity (also known from laboratory experiment) is able to pass a SOL of 10?cm thickness filled with a realistic plasma and, finally, inject a certain amount of W-material into the confined plasma region.

The surface morphology and deuterium retention were investigated of polycrystalline tungsten targets that were exposed to deuterium plasmas at widely varying conditions. By changing only one parameter at a time, the isolated effects of flux, time and pre-damaging on surface modifications and deuterium retention were studied. The sample exposed to low-flux plasma (10 20 m −2 s −1 ) is mostly smooth with only a few areas containing very large blisters(50-500 µ m). The samples exposed to high-flux plasmas (10 24 m −2 s −1 ) show large numbers of smaller blisters(1-10 µ m) and in addition even smaller protrusions (

Five tungsten (W) grades were simultaneously exposed to deuterium (D) plasma with 10 20 D/(m 2 s) of 38 eV/D up to 10 26 D/m 2 at 500 K specimen temperature. The D inventories and their depth profiles within the topmost 12 μm were determined by nuclear reaction analysis (D( 3 He, p)α). Morphological modifications at and below the surface were analysed by confocal laser scanning microscopy and scanning electron microscopy assisted by focused ion beam cross-sectioning. The observed variation of the D inventory by more than one order of magnitude (0.5–15 × 10 20 D/m 2 ) is attributed only to the different properties of each W grade. Spherical blisters and stepped flat-topped extrusions are observed depending on the W grade. These modifications are interpreted as an indication for hydrogen loading-induced damaging. The exposure conditions and W grades were chosen to allow a comparison between published data sets.

Samples prepared from steels F82H and EUROFER97 were irradiated with 20 MeV W ions at 300 K to 0.54 displacements per atom at the damage peak. Damaged and undamaged samples were exposed at elevated temperatures both to deuterium plasma at ion energies of 60 and 200 eV to a fluence of ≈1026 D m−2 and to D2 gas at a pressure of 100 kPa. The surface modification after plasma exposure was examined by scanning electron microscopy and Rutherford backscattering spectroscopy. Deuterium depth profiles were determined by the D(3He, p)4He nuclear reaction. In damaged steels loaded with deuterium, deuterium decorates the damage profile and the D concentration decreases with increasing temperature. After exposure of the F82H steel to the D plasma W-enriched near-surface layers are formed. The effective concentration of W in the near-surface steel layer depends on plasma exposure conditions.

Doping of carbon leads to a reduction of chemical erosion. This reduction is more effective if dopants are in the form of sub-μm precipitates with a very homogeneous distribution. However, dopants should not degrade but rather improve the thermal conductivity of graphite. Taking into account the catalytic effect of some metals and carbides on the graphitization, the potential for a development of improved doped carbon materials with reduced chemical erosion and optimized thermomechanical properties becomes evident. First results on the development of doped carbon materials starting from mesophase carbon powder and with different fine grain carbides as dopants show that VC acts as an effective catalyst for the graphitization. The materials obtained show high densities, a relatively low open porosity and good mechanical properties.

Many plasma-facing components in today's experiments are inertially cooled. Graphite tiles are mounted on a water-cooled steel plate using a spring and bolt mechanism. When exposed to high heat flux, such a shield can take loads only for a few seconds without becoming overheated. To operate in high intensity long pulsed/steady-state conditions, new actively cooled heat shields of brazed and non-brazed types are under development. With clamped (form-locked) monoblock tiles a prototype of non-brazed heat shield structure underwent its first high heat flux tests for divertor applications in the Julich MARION ion source [1]. The test results are encouraging, and validated a finite element model as discussed in this paper.

The composition of co-deposited hydrocarbon-silicon layers (a-C:Si:D) with varying Si concentrations and their removal by heating in air were investigated using MeV ion beam techniques. The amount of trapped D per re-deposited target atom depends weakly on the Si concentration. For pure C and Si, the D concentrations are about 0.45 and 0.5 D atoms per re-deposited target atom at room temperature, respectively. A maximum of about 0.7 D/(Si+C) was found at Si/C≈1. For increasing deposition temperature the D concentration does not decrease significantly until about 600 K. At about 1000 K the D concentration for pure C layers is still about 30% of the concentration at room temperature. The removal rates of D and C by heating in air increase strongly at temperatures around 550 K for a-C:D layers. With increasing Si content, these temperatures rise to above 650 K for layers with Si concentrations higher than 0.2 Si/(Si+C). The C removal rate is always lower than the D removal rate. Si is not removed by this method. For comparison, the composition of co-deposited stainless steel layers and Ti–C mixtures were investigated.

Tungsten coatings on fine grain graphite with film thicknesses in the micrometer range were prepared by three different methods: electron beam evaporation, magnetron sputtering, and arc deposition. They were compared with respect to properties relevant for the application of such systems in the nuclear fusion experiment ASDEX Upgrade. The content of light impurities was determined by sputter XPS. A deposition parameter study for magnetron sputtered coatings is presented; the formation of compressive stress is identified to be the major drawback of this method. Furthermore, experiments with thermal loading by an electron beam facility are presented.

Several carbide-doped (SiC, TiC, V 8 C 7 , WC, ZrC) graphites have been produced. The erosion of these materials at low-energy (eV) hydrogen ion bombardment has been investigated using the weight-loss method, mass spectroscopy, ion beam analysis, and scanning electron microscopy (SEM). The erosion yields of the WC- and V 8 C 7 -doped graphites are reduced by a factor of 2 for 30 eV D at 300 K compared to pure graphite. This observed reduction is partly attributed to surface enrichment of carbide due to preferential C erosion. The other part is assigned to changes in the chemical erosion process (Y surf ) as well as at elevated temperatures in the thermal activated process (Y therm ). The reduction of both erosion processes is determined for all dopants to be more than 25% of the erosion yield of the undoped graphite.

Doping of carbon generally results in a reduction of its chemical reactivity during hydrogen ion bombardment. Due to preferential sputtering of carbon, the dopants may be enriched at the surface, resulting in an additional reduction of the erosion yield. Dopants in the form of sub-μm precipitates with a very homogeneous distribution lead to in a more effective reduction of both chemical erosion processes, Ytherm and Ysurf. However, dopants may degrade the thermal conductivity of graphite, which has to be avoided. First results on the development of carbon materials doped with different carbides and with optimized microstructure and thermomechanical properties show that VC acts as an effective catalyst for graphitization causing an improvement of the thermal conductivity. It leads further to a reduction of both, the Ytherm and the Ysurf chemical erosion processes, which is partly attributed to surface enrichment but is also the result of a chemical influence.

The erosion yield of SiC due to D bombardment has been determined by the weight-loss method as a function of temperature and energy in the range of 20–300 eV up to 1100 K. A temperature dependence exists clearly below 100 eV. For 20 eV D, the yield is 0.5% at 300 K and 1.5% in the maximum at 500 K. Contrary to Si erosion, no formation of silane was observed with mass spectrometry. Hydrocarbons, predominantly CD 4 molecules, are only found at 20 eV. From the erosion of a thin SiC layer investigated using MeV ion beam analysis, segregation and preferential erosion of carbon are ruled out, because the layer thickness decreases while the composition stays constant. As no volatile silane production is detected, it is assumed that sputtering of Si or silane precursors with low binding energy occurs at low ion energies. The mechanism of the chemical erosion of Si in presence of C and absence of C is different.

In this paper, reduced-activation ferritic/martensitic (RAFM) steels including Eurofer (9Cr) and oxide dispersion strengthening (ODS) steels by the addition of Y2O3 particles with different amounts of Cr, namely, (9-16)Cr were exposed to low energy deuterium (D) plasma (~20–200 eV per D) up to a fluence of 2.9 × 1025 D m−2 in the temperature range from 290 K to 700 K. The depth profile of D in steels was measured up to 8 µm depth by nuclear reaction analysis (NRA) and the total retained amount of D in those materials was determined by thermal desorption spectroscopy (TDS). It was found that the D retention in ODS steels is higher compared to Eurofer due to the much higher density of fine dispersoids and finer grain size. This work shows that in addition to the sintering temperature and time, the type, size and concentration of the doping particles have an enormous effect on the increase in the D retention. The D retention in undamaged ODS steels strongly depends on the Cr content: ODS with 12Cr has a minimum and the D retention in the case of ODS with (14-16)Cr is higher compared to (9-12)Cr. The replacing of Ti by Al in ODS-14Cr steels reduces the D retention. The formation of nano-structure surface roughness enriched in W or Ta due to combination of preferential sputtering of light elements and radiation-induced segregation was observed at incident D ion energy of 200 eV for both Eurofer and ODS steels. Both the surface roughness and the eroded layer enhance with increasing the temperature. The surface modifications result in a reduction of the D retention near the surface due to increasing the desorption flux and can reduce the overall D retention.

Fundamental understanding of hydrogen–metal interactions is challenging due to a lack of knowledge on defect production and/or evolution upon hydrogen ingression, especially for metals undergoing hydrogen irradiation with ion energy below the displacement thresholds reported in literature. Here, applying a novel low-energy argon-sputter depth profiling method with significantly improved depth resolution for tungsten (W) surfaces exposed to deuterium (D) plasma at 300 K, we show the existence of a 10 nm thick D-supersaturated surface layer (DSSL) with an unexpectedly high D concentration of ~10 at.% after irradiation with ion energy of 215 eV. Electron back-scatter diffraction reveals that the W lattice within this DSSL is highly distorted, thus strongly blurring the Kikuchi pattern. We explain this strong damage by the synergistic interaction of energetic D ions and solute D atoms with the W lattice. Solute D atoms prevent the recombination of vacancies with interstitial W atoms, which are produced by collisions of energetic D ions with W lattice atoms (Frenkel pairs). This proposed damaging mechanism could also be active on other hydrogen-irradiated metal surfaces. The present work provides deep insight into hydrogen-induced lattice distortion at plasma–metal interfaces and sheds light on its modelling work.

The composition of co-deposited carbon–silicon layers (a-C:Si:D) with varying Si concentrations and their removal by heating in air were investigated using MeV ion beam techniques. The ion-induced release of D due to the analysing beam (1.2 MeV 3He) was determined. The removal rates of D and C by heating in air increase strongly at temperatures around 550 K for a-C:D layers. With increasing Si content, these temperatures rise to above 650 K for layers with Si concentrations larger than 0.2 Si/(Si+C). The C removal rate is always lower than the D removal rate. Si is not removed by this method. The observed properties of the layers are compared with those of hard and soft a-C:D films.

Erosion yields of Si, C and SiC due to D ion impact were determined between 20 and 300 eV up to 1100 K. A temperature dependence of the erosion yield has been observed for Si and C. The temperature of the maximum increases with ion energy from about 350 K for 20 eV to 570 K above 100 eV for Si, which is always about 250 K lower than for graphite. The erosion yields of Si are in general 10 times smaller than those of graphite. Chemical erosion species were observed mass-spectroscopically: silane was found for Si, but not for SiC, where only hydrocarbons were observed.

Total erosion yields of graphite and carbon materials under hydrogen and deuterium bombardment measured with the weight-loss method are presented for ion energies between 15 eV and 8 keV in the temperature range 300 to above 1000 K. The temperature of the maximum of the chemical erosion increases from below 600 to above 850 K with ion energies from 15 to 300 eV. Chemical erosion yields obtained by weight-loss measurements exceed yields measured mass-spectrometrically always by a factor of about two. Collector experiments show that a fraction of the eroded particle sticks to walls and, therefore, reduces the yield measured by mass spectrometry. A synergistic effect of neutrals in the ion beam on the chemical erosion yield can be excluded.

The co-deposition of deuterium with silicon doped carbon for silicon concentrations between 0–100 at.% in the temperature range from room temperature to 1000 K has been investigated. The eroded material from various different targets was caught on collectors together with the reflected D to build up the co-deposited layers, which were analysed with MeV ion beam techniques. The amount of trapped D per re-deposited target atom depends weakly on the Si concentration. The maximum of about 0.7 D/(Si + C) was found at Si/C ≈ 1. For pure C and pure Si the D concentration is about 0.45 and 0.5 D atoms per re-deposited target atom at room temperature, respectively. For increasing deposition temperature the D concentration does not decrease significantly until about 600 K. At about 1000 K the D concentration for pure carbon layers is still about 30% of the concentration at room temperature. Also, co-deposited layers of stainless steel and of titanium–carbon mixtures were investigated.

The total hydrogen isotope inventories in the tungsten coated divertor tiles of ASDEX-Upgrade were measured by thermal desorption spectroscopy (TDS) and - in the near surface region - by nuclear reaction analysis (NRA). The highest amounts of deuterium (≤ 5 x 10 22 D/m 2 ) are measured in the inner divertor region outside the separatrix due to the formation of a C:H layer on the tungsten by codeposition of deuterium with carbon. The inventories in the outer divertor - where erosion dominates - are generally smaller by an order of magnitude.

The thermal stability of the 8–10 at.% Si doped 3D CFC material NS31 was investigated using scanning electron microscopy (SEM) and ion beam analysis. Total erosion yields and CD4 production yields were determined by bombarding with monoenergetic deuterium ions. During heating to 1800 K SiC crystallites of a few μm size are created on the surface. A Si depletion at the surface and an enrichment layer (20–50 at.% Si) beneath the surface of about 5–10 μm thickness were observed. They are explained by the distribution of the crystallites of μm size and the evaporation of Si from the surface. After heating a NS31 mock-up tile to temperatures above 2300 K in a power load test, the Si concentration in a surface layer of more than 5 μm thickness drops below 3 at.% Si. The chemical erosion is reduced for the initial and heated samples by a factor of 2–3 compared to pure graphite. NS31 shows sputtering yields between those of graphite and silicon carbide depending on the Si concentration of the individual surface.

Tungsten coated graphite tiles were mounted in the divertor of the ASDEX Upgrade tokamak and exposed to approximately 800 plasma discharges with varying average heat fluxes of up to 6 MW/m 2 . After the experimental campaign, cracks were detected on almost all tiles in accordance with results of electron beam tests. The W-coated tiles in the inner divertor were found to be covered by deposited layers of low-Z material with thicknesses in the μm range and only a few percent tungsten remaining on the surface. In contrast the outer divertor is dominated by erosion. Here, the tungsten concentration at the surface was found to be close to the initial level. The global W-migration pattern was determined by analyzing the deposition of tungsten on a complete poloidal set of plasma facing wall components. Both divertor regions display broad deposition peaks with deposited layers at least one order of magnitude thicker than the main chamber values.