Search

After the 1990/91 discharge period several samples have been cut from the JET vessel walls, the carbon and the beryllium limiters, and the carbon and the beryllium X-point divertor tiles. The surface layers of these samples have been analysed in detail by ion beam techniques, such as PIXE, HIERDA, RBS, SIMS and Sputter-Auger for the absolute amounts as well as for the depth distributions of the different deposits. The major deposits are H, D, Be, C and O, as well as other impurities, such as Ni, Cr, Fe, and C. From the measurements on the carbon samples it was found that the oxygen deposits correlate with the Be deposits, while the measurements on the Be samples show that the C deposits correlate with the D deposits, indicating a codeposition of these elements. In addition H concentrations mostly larger than the D concentrations are measured. These may partly originate also from adsorption during the exposure to air.

After plasma operation in 1991 samples were cut from C and Be tiles of the JET vessel walls, limiters, and X-point tiles along a poloidal cross section in octant 4W, and the surface layers were analysed by the MeV ion beam methods NRA, PIXE, and RBS to determine the D and metal impurity distributions. The amount of trapped deuterium ranges from 5 × 10 15 to 5 × 10 17 D/cm 2 , metal impurities such as Ni, Cr, and Fe are in the range of 10 15 to 3 × 10 16 /cm 2 , Cl in the range of 10 16 to 10 18 /cm 2 , K and Ca in the range of 10 15 to 10 16 /cm 2 and Ti about 10 15 /cm 2 . Except for Cl, all measured concentrations are a factor of up to about ten lower than those observed on JET walls tiles before Be evaporation and Be wall components were used. This observation is in agreement with the reduced impurity concentrations measured in the JET plasma. The D and impurity distributions along the X-point tiles were analysed in some detail. Generally, the deuterium trapping and impurity deposition show minima in areas where erosion dominates and amounts a factor of about ten larger in between.

From a tile of the JET toroidal belt limiter installed during the 1990 operation period, samples were cut from the plasma-facing side along the poloidal direction and the surface layers were analysed by SEM and by the ion beam techniques NRA, PIXE, RBS and HIERD. The poloidal distribution of D is fairly uniform with values of up to 2×1017 cm−2. The contamination of the Be by metallic impurities from the vessel wall is typically up to 2×1016 Ni cm−2, 5×1015 Cr cm−2 and 2×1016 Fe cm−2, giving totally about 0.2 g of metals deposited on the upper limiter. Other impurities amount up to 3×1017 C cm−2, 7×1017 O cm−2, 2×1017 Cl cm−2, 2×1016 K cm−2, 4×1015 Ca cm−2, 3×1015 Ti cm−2 and 1×1015 Zn cm−2. The D deposition is about a factor of 4 smaller than on the C limiter of the previous discharge period, while the deposition of the Inconel components is about one order of magnitude smaller than on the C limiter. O is found to be trapped in the Be at depths up to several micrometers, which may be caused by segregation of pure Be through the oxide to the surface of oxide layers on the limiter. This would improve the O gettering of the Be. The lower D concentration is consistent with the observed degassing of D in the time between discharges.

Depth distributions for deuterium trapped within the Be/C layer on a JET limiter after long-term operation have been measured up to a depth of 7 μm using the D(3He, p)4He reaction in a resonance-like technique. For several points along a JET limiter tile, depth profiles and the total amount of trapped deuterium have been determined.

During 1989, following a brief all carbon phase, JET was operated with an evaporated beryllium film on all the plasma facing surfaces and then, in a third phase, with bulk beryllium used for the belt limiters and one set of antenna protection tiles. Subsequent analysis of wall components and long term samples (LTS) using Beta Backscatter and Ion Beam Analysis has been conducted to determine the condition of the First Wall and its deuterium inventory after beryllium operation. Ex-situ analysis of components where deposition occurs during operation shows mixed carbon and beryllium layers with an approximate ratio 1:1 and some localised higher Z components. The deuterium inventory situation is littled changed by the introduction of beryllium. Similar surface levels of deuterium, ~1 × 1022 atoms m−2, have been seen for both carbon and beryllium belt limiter tiles.

An experimental investigation of metallic impurities has been made in the ASDEX divertor in order to better understand their likely sources and production mechanisms. The impurities were collected on rotatable carbon collector strips which were exposed to the scrape-off layer plasma near the upper outer divertor plate under various discharge conditions. Quantitative ion beam analysis of exposed samples was subsequently carried out by Rutherford backscattering spectroscopy, proton induced X-ray emission and nuclear reaction spectroscopy. Time resolved measurements and spatial profiles of collected impurity fluxes reveal an increase in collected metallic impurities during auxiliary neutral beam heating of plasma discharges causing strong erosion of the copper divertor plate. To a rough approximation, 0.2% of this auxiliary heating power is invested in the production of stainless steel impurities. Measurements made for single-null divertor discharges show a copper flux of 6 × 1017 atomss−1 in the scrape-off layer which is not directly sputtered from the target plates. Collected impurity fluxes in the divertor chamber are found to increase with both the main plasma current and the density. Measurements made after boronization reveal an initial reduction in collected stainless steel impurities although boronization does not have any long term effect on the suppression of such impurities. Some limitations of the collector probe technique are also pointed out, and planned improvements of the existing setup are outlined.

Impurity production and transport in ASDEX have been studied using a rotatable graphite collector probe which was exposed to the scrape-off layer plasma in the upper divertor chamber. Metallic deposits on the probe were later analysed using Rutherford backscattering spectroscopy and nuclear reaction analysis. Time resolved information (during tokamak discharges) and spatial profiles of collected impurity fluxes indicate strong erosion of the copper divertor plate during neutral beam heating of the plasma. Both stainless steel and copper deposits increase with neutral beam power. It is shown that, to a rough approximation, 0.2% of this auxiliary heating contributes to the increase in collected stainless steel impurities. Measurements made for single-null divertor discharges show that a copper flux of 6 × 1017 atoms s−1 is present in the scape-off layer which is not directly sputtered from the target plates. Measurements made subsequent to boronization reveal an initial reduction in collected heavy impurities although boronization does not appear to have any significant long-term effect on the suppression of such impurities.

We report on a method to measure the depth profile of deuterium up to a depth of 7 μm in a carbon/beryllium layer using the D( 3 He, p) 4 He nuclear reaction in a resonance-like technique. The deuterium depth profile is determined from a measurement of the proton yield as a function of incident 3 He ion energy for a limiter tile from the Joint European Torus (JET) tokamak.

Surface probes have been exposed at JET with all-carbon walls, when beryllium was evaporated onto graphite limiters, and with beryllium limiters. The aims have been to study the impurity content in the scrape-off, erosion and deposition, and the pumping of deuterium by codeposition with carbon and beryllium. Thermally activated erosion makes it difficult to study time resolved erosion close to the LCFS. In the all-carbon machine, the carbon density in the outboard scrapeoff between the belts was estimated to ~ 20% of the deuterium. With a fresh Be coating on the limiters, carbon in the SOL was initially suppressed, but increased from discharge to discharge. The D/C ratio in codeposition layers was three times larger with Be-evaporation than in all-carbon conditions. Edge profiles determined from impurity deposition became flatter during RF heating and current ramp down.

The amounts of H and D trapped in the surface layers of samples made out of C, Ni, Inconel and Si, installed at the vessel walls of JET and removed after the 1984 and 1985 periods of operation were measured. After the 1984 operation period the D concentration in carbon probes was of the order of 5 × 10 19 /m 2 . while the hydrogen was about a factor of 40 higher. This can be attributed to a final glow discharge cleaning in hydrogen at the end of the operation period. After the 1984 period of operation the amounts of deuterium and hydrogen were nearly equal at about 5 × 10 21 /m 2 . A comparison of the amounts of deuterium and hydrogen with the amount of carbon deposition on the Inconel wall samples shows a ratio of about 0.3 to 0.4 (H + D) per C-atom. The hydrogen concentrations trapped in the vessel wall correspond to more than 100 times the amount of hydrogen isotopes in the plasma during a discharge.

For a great number of ion-target combinations similarities in low energy sputtering have been found. 11Near the threshold energy for sputtering E th, the energy dependence of the sputtering yield can be written as S = Y (E/E th)1/4 (1 – E h/E)7/2, E th and Y depend only on the surface binding energy E B and the ion-target atom mass ratio. Using this relation the energy dependence of the sputtering yield of B4C, SiC, TiC, TaC, WC is compared with the yields from S, Si, Ti, Ta, W, for H, D and He in the energy range of 50 eV to 8 keV. A strong similarity of the energy dependence, i.e. Eth , and Y, can be found between the carbide and the heavier component of the compound. This leads to the assumption, that the target surface is depleted of the light component due to preferential sputtering and the sputtering process is dominated by the heavy component.

In order to evaluate the deuterium and hydrogen distributions in JET after long term operation in deuterium in 1985, as well as after a few weeks with deuterium at the end of the hydrogen-based operations in 1984, limiters have been analysed using nuclear reaction techniques and laser induced mass analysis (LIMA). The results show the presence of very high amounts of deuterium and 1H (up to more than 1022 atoms/m2 on the sides and edges of the limiters). If this amount were the result of the implantation of deuterium into the limiter graphite, it would require ion energies far higher than those expected in the boundary layer. The deuterium concentration shows a broad minimum near the centre of the limiter, where the heat load from the plasma is a maximum. The depth distributions, which are obtained by the 3He(d, p)4He nuclear reaction, indicate that deuterium is present to a considerable depth on the sides of the limiters. Similar amounts have been seen at the inner wall of the vessel — either on small samples mounted on the wall, or on the carbon tiles which covered the wall in 1985 — and on the RF antennae. It appears that co-deposition of carbon and deuterium enables significant trapping of hydrogen isotopes on all exposed surfaces.

A number of carbon tiles of the TFR limiter and antenna protection were analyzed by PIXE, RBS, and SEM to obtain a comprehensive and quantitative picture of the metal coverage in these areas. The results show a metal coverage corresponding to more than a few hundred monolayers fairly homogeneously distributed and/or in the form of droplets (1 to 1000 μm in diameter). Analysis of the chemical composition showed that the layer mainly consisted of Fe, Ni, Cr, and very little Ti. This indicated that the Ti Faraday shield for the ICRH antenna was not responsible for the metal contamination, in agreement with spectroscopic results. The reduction of the metal impurity content of the discharge by a factor of 3 to 10 when carbon limiters are used instead of metal ones seems to be related to the actual chemical composition of the limiter surface. Chemical sputtering of such carbon surfaces is lower than for pure carbon surfaces, whereas their hydrogen content is very high (up to 10 22 D/m 2 ).

Assuming that sputtering of the wall material by energetic charge exchange neutrals is the dominant process for the release of metal impurities in CTR devices, a conversion of the density of atoms in the groundstate, which is the quantity measured by laser-induced fluorescence, into the total flux density of wall material, is possible. For this the knowledge of three quantities is required: (a) the ratio of the groundstate density n 0k to total density n k of one constituent of an alloy; (b) the ratio of n k to the flux density φ k (equal to the transport velocity v k ); and (c) the ratio of φ k to the total flux density φ of the wall material. The quantities n 0k /n k and v k are determined first for the case of iron sputtered by 10 keV Ar-ions. In a second step experimental evidence is presented that the ratio k n k n 0k is independent of the energy (in the range from 80 eV to 8 ke V) and of the mass of sputtering particles (H+, D+, He+). In a third step the iron target is replaced by stainless steel 316 and Inconel 600 targets, respectively; the measurements give as a result that the ratio φ Fe /φ is reduced according to the iron concentration in the alloy.

Hydrogen and deuterium fluxes parallel to the toroidal magnetic field were measured in the plasma boundary of ASDEX using graphite collector probes. Time resolution of the order of 100 ms was obtained by rotating the cylindrical probes behind an aperture during the discharge. The trapped amount of hydrogen was determined by subsequent thermal desorption, in the analysis of deuterium the D( 3 He, p) 4 He nuclear reaction was used. Both methods yield quantitative results. Measurements were done for limiter and divertor discharges in the range of 4 to 20 cm outside the limiter or separatrix. The time distributions show a maximum flux at the beginning and at the end of the discharge. The relatively lower flux during the plateau phase of the discharge is in the range of 10 15 to 2 × 10 17 cm −2 /s , depending on the radial probe position, the maximum values are higher by a factor of 5 to 50. During neutral hydrogen injection an additional maximum can be observed. The radial 1/e- decay length is about 0.9 cm in front and 0.4 cm behind the fixed limiter. The results are compared with independent measurements in ASDEX and other plasma machines.

Wall long term samples (LTS) which were exposed during the 1984 and 1985 operational periods of JET have been investigated by means of surface analysis techniques. The results indicate that deposition of wall material (Fe, Cr, Ni) and its distribution on the vessel wall were strongly affected by wall erosion on bellows protection plates during unstable discharge phases. Erosion sources and deposition sinks are spatially correlated indicating that most of the eroded material is deposited close to its source.

Laser excited fluorescence spectroscopy (LEFS) has been used to measure sputtering yields of iron bombarded with 2 keV hydrogen and 6 keV helium ions as a function of the bombarding fluence and the ratio of oxygen gas pressure to the current density. For a low oxygen partial pressure corresponding to a low ratio of the arrival rate of oxygen molecules to ions on the surface the Fe sputtering yields for a new target are initially low but increase by a factor 10–15 after about 20 monolayers of the surface are sputtered away. This increase in the sputtering yield is also correlated with the appearance of a rough surface structure. At high ratios of the oxygen molecule to ion arrival rates at the surface corresponding to a continuous reoxidation of the surface, the sputtering yields stay low, the surface stays flat during the bombardment, and initially rough surfaces become rounded. Transferring these results to fusion research means that the plasma contamination with metal atoms is becoming a larger problem if the surfaces of the discharge vessel are well cleaned.

Magnetization curves of superconducting wires were determined as a function of the angle α between the field and the axis of the wires. The results show that the magnetization for smaller angles has a behaviour quite different from that for larger angles. Strong transverse magnetization is associated with incomplete penetration of the field. The anistrophy of critical current densities is responsible for the effect. As a result, the loop area shows a deep minimum at an angle α increasing with the wire diameter.

TiC coating were produced by chemical vapour deposition (CVD) and plasma spray techniques on AISI 316, 310, 304, Inconel 600, and Mo. The 10–30 μm thick coatings formed by CVD at 1323 K show a stoichiometric composition, high density, and preferential orientation (200). Coatings made by plasma spraying also have a stoichiometric composition but traces of titanium oxides and nitrides have been observed. The testing of the coatings involved thermal shock and sputtering at different temperatures. Thermal shock tests were performed by electron beam heating in a 10 −5 Torr vacuum system with a heat load of 1 kW/cm 2 for 1–4 s. In no case disintegration, blistering, or local spall of the coating was observed, even where the substrate was melted during the heat pulse. However, owing to differential thermal expansion, cracks were observed in the case of CVD coatings. Cracks were not evident in plasma spray coatings. Sputtering at medium temperature is in agreement with that previously measured at room temperature.

Deuterium fluxes parallel to the toroidal magnetic field in the plasma boundary of the midplane of ASDEX were collected with graphite collector probes positioned at a distance of about 5 to 10 cm from the separatrix, and then quantitatively analysed by the nuclear reaction D( 3 He, p) 4 He. The time resolution of about 0.1 s of the measuring setup allows the fluxes to be collected during the ohmic heating plateau phase (about 1 s) of the ASDEX discharge (typically 2 to 3 s) and during the neutral injection phase (about 0.2 s). The measurements yield fluxes of 10 16 to 10 17 D/cm 2 s in the plateau phase, and fluxes one order of magnitude larger during the neutral injection phase. In the ohmic heating phase the e-folding length of the curve described by the measured points is observed to be about 1.5 cm. During neutral injection in discharges of the H-type the e-folding length was found to be about 8 cm. A series of such measurements is presented and discussed, with particular regard to the plasma-wall behavior during L-type and H-type discharges.

The surface layers of tiles from the carbon limiters in octant 1, ocant 4 and the RF-antenna limiters were analysed after the end of the first three experimental sessions of JET, 1983, 1984 and 1985. On all tiles metallic contaminants were found from the vessel walls and the RF antennae. In general on all areas deposition and erosion was observed. At the part of the tiles in direct contact with the plasma the erosion dominates, whereas deposition dominates on the outer parts of the limiter radially further away from the plasma. In the small gaps between adjacent limiter tiles, deposited layers of carbon having a thickness of several μm were found, which were partly flaked and contained metallic impurities. Electrical are tracks were observed on the limiter tiles and in the area between limiter tiles.

Cu and Ag were sputtered at normal incidence by H + , D + , He + , Ne + , and Ar + ions at energies between 3 and 6 keV. The target temperature was varied between room temperature and 80–90% of the melting temperature. The sputtering yield was measured by the weight loss method. Within the accuracy of the measured data, no temperature dependent change of the sputtering yield was found, but the surface structure did show a significant change. Only in the case of recrystallisation could a change of the sputtering yield of about 20% be detected.

Eight long-term graphite probes mounted on the metal limiter of the FT tokamak were exposed to about 60 ohmically heated deuterium discharges mostly characterized by off-normal events such as disruptions and runaway fluxes. A poloidally and radially asymmetric pattern of both the amount and the shape of the metal coverage and the deuterium content is found from the PIXE and NRA surface analyses.

The surfaces of two Be limiters and of a test sample from the stainless steel torus wall used in UNITOR have been analyzed by RBS, PIXE, AES, and SEM/EIXE. The main foreign atoms found on the limiters are O (2 × 10 17 to 2 × 10 18 /cm 2 ), Fe, Ni Cr, and Sn (2 × 10 15 to 2 × 10 16 /cm 2 ). There are no large differences of the foreign atom coverage on the different surface areas of the limiters. Generally the front surface of the limiters shows a slightly lower coverage than the sides. On limiter 2, the Sn coverage is about a factor of 4 larger than on limiter 1, indicating that the Sn source was close to this limiter. The stainless steel test sample from the torus wall shows islands of BeO, which can be attributed to splashes or droplets eroded from the limiters during disruptions. Major erosion of the limiters occurred only during the first 100 shots, when many discharges were terminated by disruptions. Redeposited wall material is found on the wall sample which partially covers the BeO deposits. The surface modifications of limiters and wall samples were discussed by taking into account the particle fluxes from and to the walls and limiters.

This paper describes the application of superconductivity to controlled fusion research at Max-Planck-Institute of Plasma Pnysics in Garching, i.e. mainly in the quadrupole (Wendelstein 6) and stellarator (Wendelstein T) experiments, which are respectively under construction and at the advanced planning stage.