P. Wienhold

Hydrogen inventories in nuclear fusion devices

Hydrogen retention in tokamaks is due to implantation into plasma-facing materials and trapping in deposited layers. In the limiter tokamak TEXTOR-94 hydrogen-rich deposited layers with thicknesses up to 1 mm are observed on recessed parts of the limiters, areas perpendicular to the magnetic field in the scrape-off layer (SOL), neutralizer plates of the pumped limiter and inside the pumping ducts. In the divertor tokamak JET the main deposition is observed in the divertor, additional deposits are observed in the main chamber on the sides of the guard limiters. Codeposition of carbon ions with hydrogen is the major mechanism of layer growth at areas with direct plasma contact. At remote areas without direct plasma contact, sticking of neutral hydrocarbon radicals seems to play an important role for hydrogen trapping.

Molybdenum test limiter experiments in TEXTOR

Limiter experiments with a Mo test limiter have been carried out in TEXTOR under various conditions with ohmic and NBI heating. Maximum power loads reached about 20 MW/m2 resulting in surface temperatures up to the melting point. A maximum fraction of 8% of the total convective energy in the plasma was deposited onto the Mo limiter. Molybdenum impurities are mainly produced by physical sputtering due to the impact of C and O impurities. Under ohmic heating conditions the Mo impurity radiation increased with increasing plasma density, though the relative source strength of the Mo release decreased. Approaching a critical density of about 3*1013 cm-3 Mo accumulated in the plasma centre leading to strong central radiation with hollow temperature profiles, which resulted in a minor disruption. Under NBI heated plasma conditions (co-NBI, 1.3 MW) Mo radiation decreased with increasing plasma density as did the relative Mo source at the limiter; at high plasma densities the influence of the Mo limiter on plasma impurities and plasma performance became negligible. No instability was observed under these conditions. The different behaviour between ohmic and NBI heating conditions is explained by the different relative importance of neoclassical and anomalous transport under ohmic and NBI heating, respectively. The observed instability in high density ohmic plasmas is caused by Mo accumulation in the plasma centre and is attributed to neoclassical transport processes

Plasma edge physics with siliconization in TEXTOR

Abstract By using silicon as a material for wall coatings significant improvements in the tokamak performance have been obtained. These include the lowest oxygen level ever achieved in TEXTOR, an improved density limit and an enhanced energy confinement also at high central electron densities. The present paper gives a survey on the plasma properties under siliconized wall conditions with special emphasis on plasma edge physics. The impurity fluxes released from the limiter have been determined as a function various parameters. An outstanding property of silicon as an impurity in the plasma is its strong influence on plasma edge parameters caused by line radiation. A high radiation level of up to 70% of the input power (1.5 MW) has been achieved, limited by the mechanisms which are responsible for the release of silicon. To describe and understand these processes at the plasma edge Od- and 1d-models are used to simulate the highly nonlinear system. The properties are compared to plasmas in a boronized machine. Silicon and neon are compared with respect to radiation characteristics and transport to the centre. Remarkable differences in the penetration depths and central densities are reported.

Borontrimethyl B(CH3)3 — A less hazardous substance for boronization

Abstract Boronization, i.e. plasma induced deposition of amorphous boron containing carbon films a-C/B : H on all inner surfaces of fusion devices, has proven to be a powerful conditioning method to achieve very pure fusion plasmas. The use of the highly toxic and explosive gas diborane (B2 H6) in conventional boronization requires considerable precautions for safe handling. We report on a new technique for boronization, using less hazardous organic boron compounds, namely borontrimethyl B(CH3)3 and borontriethyl B(C2H5)3. First results indicate that films can be produced from borontrimethyl showing a performance in fusion devices similar to that of layers produced in the conventional way. The new technique thus offers an easier and more convenient way for the boronization of surfaces.

Short and long range transport of materials eroded from wall components in fusion devices

Abstract Carbon sources and the sinks have been quantified in TEXTOR and are discussed in terms of short and long range transport. The major source (22 g/h) is the graphite belt limiter, but part (10 g/h) of the carbon is directly re-deposited after short range transport. Long range transport causes flake formation on obstacles and neutralisers, but little and deuterium rich (D/C≈0.7) deposition in remote areas. The rest is leaving via the pumps in gaseous form. This behaviour is different from that in JET where large amounts of deuterium rich deposits were found in the louvers. Tungsten is favoured for the ITER divertors because of its low sputtering yield for hydrogen, but melting and erosion by carbon may be an additional concern. The short range transport of tungsten has been investigated in a well defined experiment and quantitatively re-constructed by means of the ERO-TEXTOR code. Code validation is necessary in order to increase the confidence and the applicability to JET and ITER.

Investigation of carbon transport in the scrape-off layer of TEXTOR-94

Abstract 13 CH 4 was injected into 18 consecutive ohmic discharges (108 s) through a hole in a graphite block equipped with an aluminium plate. CH and CII radical emission adjacent the hole was observed by spectroscopy, while the incremental grow of the deposit was measured in situ by colorimetry. SIMS depth profiling yielded the fractions of 13 C and of the 12 C out of the background. Ratios up to 13 C / C ≈0.42 were found although the deposition efficiency for 13 C was only about 0.2%. H and D were co-deposited up to ratios of 0.4 and 0.2, respectively, as measured by NRA and ERDA. The results can be simulated quantitatively by means of the ERO-TEXTOR code if high re-erosion is assumed for the hydrogen rich radicals CH x + . Within the interaction depth the 13 C/C ratio seems to become quasi-stationary. 13 C transported to obstacles creates a ratio of 13 C / C ≈0.27 after 108 s.

Erosion and redeposition of wall material in controlled fusion devices

Processes of erosion and redeposition and their impact on plasma facing materials in devices for magnetically confined fusion are discussed. Volatile molecules formed in the erosion process are partly pumped out but the majority of species released from the wall components returns to the surface causing the modification of its morphology. Prompt redeposition and redeposition after global transport reduce the gross erosion at any surface. Detailed analysis shows that erosion and deposition can coexist on otherwise erosion-dominated surfaces due to local inhomogenities. The erosion yield of redeposited material of sub-monolayer thickness significantly differs from that observed for thick targets. On deposition dominated areas one observes the formation of thick co-deposits containing a mixture of trapped fuel atoms (hydrogen isotopes) and species removed originally from the wall. This leads to a large and long-term fuel accumulation (tritium inventory) in a device. Reduced mechanical integrity of such layers stimulates their flaking and peeling-off if a critical thickness has been reached. This, in turn, results in the formation of hydrogen-rich dust particles.

Removal of redeposited layers and hydrogen release by oxygen ventilation of TEXTOR-94

Abstract Ventilation of the TEXTOR torus with oxygen at pressures between 0.007 and 0.3 mbar and at wall temperatures between 500 and 700 K has been investigated to remove redeposited carbon material and to release the incorporated hydrogen. It has been observed that a significant part of the injected oxygen adsorbs on the walls due to formation of stable oxygen compounds. Part of the oxygen reacts with CO and CO 2 . The CO reaction rate is independent of the filling pressure whereas the CO 2 production increases with increasing pressure. Plasma operation after the oxygen baking has been achieved after GDC in D 2 and He (15–30 min). The oxygen impurity content was initially about a factor of 2–3 higher than before but also showed a decreasing behaviour shot by shot.

Limiter Lock Systems at TEXTOR: Flexible Tools for Plasma-Wall Investigation

Abstract Limiter lock systems on the top and the bottom of the TEXTOR vessel are essential elements for experimental investigations of plasma-wall interaction in a tokamak. The lock systems are designed as user facilities that allow the insertion of wall elements (limiter) and tools for diagnostic (electrical probes, gas injection) without breaking the TEXTOR vacuum. The specially designed holder on top of the central carrier and a powerful vacuum pump system permit the exchange of components within ~1 h. Up to ten electrical signals, four thermocouples, and a gas supply can be connected at the holder interface. Between discharges, the inserted component can be positioned radially and turned with respect to the toroidal magnetic field. Additionally, the central carrier is electrically isolated to apply bias voltages and currents up to 1 kV and 1 kA, respectively. An important feature of the lock system is the good access for optical spectroscopic observation of the inserted components in the vicinity of the edge plasma. The whole spectrum from ultraviolet to infrared is covered by spectrometers and filters combined with cameras. Toroidally and poloidally resolved measurements are obtained from the view on top of the probes while the tangential poloidal view delivers radially and toroidally resolved information. A programmable logic controller (Simatic S5) that is operated inside the TEXTOR bunker and from remote locations outside the concrete wall drives all possible features of the lock system.

Experiments with tungsten limiters in TEXTOR-94

Abstract The release of tungsten and light impurities from tungsten limiters exposed into the plasma edge of TEXTOR-94 has been measured by spectroscopic methods. Absolute effective tungsten sputtering yields are compared with model calculations on physical sputtering. The agreement is reasonable: however the observed strong decrease of tungsten release with increasing density cannot be fully explained. Erosion areas are clearly separated from carbon deposition zones. Surface analysis found neither carbon nor deuterium on the shiny metallic areas: A very sharp transition from “clean” metallic areas to carbon deposition zones within about 2–4 mm is found, instead. The carbon deposit is about 200–300 nm thick and contains deuterium with a D/C ratio of 0.05–0.1.