Optical emission spectroscopy can be applied to determine in situ tungsten particle fluxes from erosion processes at plasma-facing materials. Inverse photon efficiencies convert photon fluxes of WI and WIT line transitions into W and W+ particle fluxes, respectively, dependening on the local plasma conditions. Experiments in TEXTOR were carried out to determine effective conversion factors for different WI and WIT transitions with the aid of WF6 injection into deuterium scrape-off layer plasmas in the electron temperature T-e range between T-e = 20 eV and T-e = 82 eV. The inverse photon efficiencies or so-called effective s/XB-values have been determined for WI lines at lambda = 400.9 nm, 429.5 nm, 488.7 nm, 498.3 nm, and 522.5 nm as well as for WIT at lambda = 434.6 nm and compared with theoretical calculations from the ADAS data base. Moreover, a multi-machine scaling for the s/XB-value in the range of T-e between 2... 100 eV has been determined for the most prominent WI line at A = 400.9 nm to s/XB(T-e) = 53.63 56.07 x e((0.045 x Te[ev])) considering experimental data from TEXTOR, ASDEX Upgrade, PSI and PISCES. Comparison with ADAS calculations for the same transition reveal a good qualitative agreement with the dependence on T-e, but an underestimation of ADAS calculations of less than 25% over the full covered range of experimentally accessible T-e in the multi-machine scaling. A good agreement within the experimental uncertainties is found between TEXTOR and ADAS s/XB-values for WI at lambda = 429.5 nm and lambda = 488.7 nm whereas an underestimation of up to a factor two of ADAS values for WI at lambda = 522.5 nm and lambda = 498.3 nm was measured. Potentially, reasons for the discrepancy are an overestimation of applied ionisation rate coefficients in ADAS for neutral W and a stronger electron dependence n(e) for these transitions.

The ERO code was modified for modeling of plasma–surface interactions and impurities transport in the PSI-2 installation. Results of experiments on tungsten target irradiation with argon plasma were taken as a benchmark for the new version of the code. Spectroscopy data modeled with the code are in good agreement with experimental ones. Main factors contributing to observed discrepancies are discussed.

Laser induced ablation spectroscopy (LIAS) is a diagnostic to provide temporally and spatially resolved in situ measurements of tritium retention and material migration in order to characterize the status of the first wall in future fusion devices. In LIAS, a ns-laser pulse ablates the first nanometres of the first wall plasma-facing components into the plasma edge. The resulting line radiation by plasma excitation is observed by spectroscopy. In the case of the full ionizing plasma and with knowledge of appropriate photon efficiencies for the corresponding line emission the amount of ablated material can be measured in situ. We present the photon efficiency for the deuterium Balmer alpha-line resulting from ablation in TEXTOR by performing LIAS on amorphous hydrocarbon (a-C:D) layers deposited on tungsten substrate of thicknesses between 0.1 and 1.1 mu m. An experimental inverse photon efficiency of [D/XB](D alpha (EXP))(a-C:D -> D)(LIAS) = 75.9 +/- 23.4 was determined. This value is a factor 5 larger than predicted values from the ADAS database for atomic injection of deuterium under TEXTOR plasma edge conditions and about twice as high, assuming normal wall recycling and release of molecular deuterium and break-up of D-2 via the molecular ion which is usually observed at the high temperature tokamak edge (T-e > 30 eV).

Linear plasma devices such as PSI-2 [1] are widely used for studying of plasma-surface interaction (PSI) because of their flexibility, con tinuous plasma operation and divertor relevant plasma conditions. However, due to complex physics of plasma surface interaction, the numerical modeling is still an essential tool to in terpret the experimental results. The 3D Monte-Carlo code ERO [2] is a recognized too l f r the predictive modelling of the life time of ITER wall components and long-term tritium retention. It is also used for the interpretation of PSI experiments involving erosion , l cal impurity transport, spectroscopy and surface content evolution. It has been applied earl ier to linear plasma devices [3], however for PSI-2 simulations further modifications of the code have been made to consider the specific shape of plasma density and temperature profiles, t arge geometry and position, and locations of available spectroscopy diagnostics. Tungsten (W) is one of the most important materials for ITER wall construction. Therefore, understanding of its interaction with plasma is ess ential for the project. However, some of data concerning these interactions need verification or are not known at all. For example, the role of metastable states in tungsten light emission, which is important for interpretation of spectroscopy results, is still not clear; different databases give different values for the ionization rate coefficients etc. Dedicated experiments were performed at PSI-2 facil ity and analyzed using the modified ERO code. In these experiments a tungsten target was ex pos d to argon plasma, which has been characterized by scanning Langmuir probe measuremen ts. WI (emission from neutral tungsten atoms) line intensity profiles along the axial axis of PSI-2 were obtained using a spectrometer, and 2D picture of WI irradiation was observed by a camera with a filter. ERO modeling shows good qualitative agreement with observed 2D picture. However, the decay rate for WI emission profiles shows some disc repancy between experimental and simulation results. Different factors responsible f or this are discussed. The main contribution is associated with an uncertainty of atomic data for t ungsten ionization and light emission taken from ADAS [4] and ATOM [5] databases. It is interes ting to note, that the simulations show that the most of sputtered particles are deposited on the walls of the chamber and only about of 6% are coming back on the target surface.

On the TEXTOR tokamak various experiments aimed at investigation of tungsten erosion and transport are performed. In one experiment a spherical W/C twin limiter positioned close to the last-closed flux surface in the near scrape-off layer was exposed to a number of comparable plasma discharges with stepwise variations of edge plasma parameters. Spatial distribution of tungsten and carbon light emission was recorded with two dimensional CCD cameras and spectrometer systems with high spectral and spatial resolution. Penetration depths, tungsten sputtering fluxes and erosion yields were measured. Comparison between experimental data and the results of modelling with the 3D Monte-Carlo code ERO is performed. The main objective of this study was to test the adequacy of the existing atomic data for neutral tungsten. The modelled penetration depths of the light emission of tungsten are a factor of 2–3 smaller than in experiment, which may indicate the overestimation of ionization rates.

Resonant Magnetic Perturbations (RMPs) are applied with the Dynamic Ergodic Divertor (DED) at TEXTOR to control the plasma edge transport and the plasma surface interaction. This leads to the formation of a three-dimensional (3D) topology of the scrape-off layer (SOL). To quantify the erosion/deposition balance and the material migration in this 3D boundary, spherical test limiters were exposed to plasmas with and without RMP fields applied. Methane doped with 13C as tracer element was injected through a gas inlet in the test limiter. The local gas source was monitored by spatially resolving spectroscopy and the resulting deposition patterns on the limiters were analysed with colourimetry and nuclear reaction analysis. These measurements were compared to simulations of the magnetic field topology simulations. The data provide evidence of a particle migration dominated by an ExB drift within stochastic zones of the 3D plasma boundary.

Tracer experiments have been carried out by injection C-13 marked methane through test limiters exposed to the scrape-off-layer in TEXTOR. The influence of impact energy and flux on depositing C-13 species has been studied. One experiment has been performed with biased test limiter (-300 V) in order to increase energy of positively charged ions and the other one with 10 times reduced (CH4)-C-13 injection rate compared to previously used injection rate. Biasing of the test limiter increases the resulting C-13 deposition by a factor of similar to 6 - post-mortem analysis yields a C-13 deposition efficiency of similar to 1.7% compared to similar to 0.3% without biasing. Reducing the injection rate increases C-13 deposition efficiency to similar to 0.7%, which is more than two times larger compared to experiments with previously used injection rate. ERO modelling shows that enhanced re-erosion of redeposits is still necessary to reproduce measured C-13 deposition efficiencies. (C) 2013 Elsevier B. V. All rights reserved.

To quantify tungsten (W) sputtering measurements by spectroscopy, two experiments have been carried out in the tokamak TEXTOR. The erosion of a W limiter in the plasma edge was studied spectrocopically under different plasma conditions. Conversion of the photon fluxes of neutral W lines was performed [1] with the aid of inverse photon efficiencies, measured in situ in a second experiment with WF6 injection to realise a controllable W source. Penetration depths, particle velocities and line ratios of W I were determined and compared between the two sources of W. The velocities of injected and sputtered W differ by a factor of 3 as the analysis of the penetration depth shows, while the line ratios for different W I lines are comparable. A comparison of the e-folding length with the model code GKU was performed which reveals a deviation of at least a factor of 6. Inverse photon efficiencies at Te=40eV were determined to be about 44 for W I (400.88nm) and about 63 for W I (429.46nm).

Laser-based methods are investigated for the development of an in situ diagnostic for spatially and temporally resolved characterization of the first wall in fusion devices. Here we report on the first systematic laser-induced ablation spectroscopy (LIAS) measurements carried out on various surface layers in the TEXTOR tokamak. These materials include a-C: D, mixed W/C/Al/D-2, Oerlikon Balzers 'Balinit' diamond-like carbon layers and EK98 fine-grain graphite. In LIAS, the bulk or deposited material is evaporated during the plasma discharge by intense laser radiation. The light emitted by particles entering the edge of the ionizing tokamak plasma is then observed by optical spectroscopy. In the measurements taken, it was found that the studied layers can be identified by their characteristic line emission. A good correlation between the observed line intensity and layer thickness is found. The observed plumes show target material dependence. To analyze layers formed during tokamak operation, further investigation of the ablation process and reference materials for cross calibration is required.

In future fusion devices like ITER deposition of impurities will likely occur in areas, remote from plasma and on the sensitive components of optical diagnostics, like mirrors and windows. Deposition in remote areas may lead to the tritium retention and therefore represent a safety issue. Deposition on optical components will severely deteriorate their optical properties potentially leading to the shutdown of the respective diagnostic systems. An active control over deposition is therefore highly desirable. The paper contains an overview of experiments on active control over carbon deposition undertaken in TEXTOR tokamak. A prototype of diagnostic duct was exposed in the scrape-off layer plasmas of TEXTOR where several techniques were applied to mitigate carbon deposition on diagnostic mirrors located inside this prototype. The complete suppression of carbon deposition on the surface of diagnostic mirror was achieved by feeding the deuterium gas inside the diagnostic duct.

Tungsten is foreseen as the plasma-facing component material for baffles, the dome and strike-point area in the ITER divertor. Quantification of the W source, which is connected with the components lifetime and W plasma concentration, is one of the outstanding issues in the qualification process. A dedicated experiment in TEXTOR with the exposure of a W/C twin limiter to the near scrape-off layer plasma has been carried out in order to address the W sputtering and local material mixing in the electron temperature range between T-e = 30 and 85 eV, achieved with deuterium fueling in four steps. The T-e range is comparable to the baffle region and the strike-point area during non-detached transient phases of the ITER divertor plasma. Quantification of the W sputtering yield and the impinging impurity fluxes was performed with the aid of optical spectroscopy, in particular by observation of WI and WII lines. As no inverse photon efficiencies in the plasma parameter range of the twin limiter experiment exist, we performed in a second experiment for the first time a calibration of WI and WII photon efficiencies with local injection of WF6 through a gas inlet into the TEXTOR edge plasma. The in situ determined effective inverse photon efficiency of about 85 for the most prominent WI line at 400.9 nm, which is in good agreement with GKU modelling for the covered T-e range and 650 for the WII line at 434.8 nm, has been applied to the corresponding photon fluxes in the twin limiter experiment. The W sputtering yield decreases from 5.2 to 0.5%, thus by about one order of magnitude, with a reduction of T-e from 85 eV down to 30 eV and a simultaneous increase of the impinging deuterium ion flux by 50% occurs. A lower limit for the prompt redeposition has been estimated at 50% by analyzing the WI to WII flux ratio. Local measurement of OII (441.6 nm) and CII (426.7 nm) provided impurity flux ratios of 0.6% for O and 5.2% for C related to the deuterium recycling, respectively ion flux. Both flux ratios remain constant for all phases of the discharge with plasma edge cooling. W erosion is predominantly caused by sputtering of higher ionization stages of O and C impinging on the W limiter half and not by the fuel species itself. Plasma cooling below the physical sputtering threshold could not be achieved without impurity seeding.

Mirrors will be used in all optical and laser-based diagnostic systems of ITER. In the severe environment, the optical characteristics of mirrors will be degraded, hampering the entire performance of the respective diagnostics. A minute impurity deposition of 20 nm of carbon on the mirror is sufficient to decrease the mirror reflectivity by tens of percent outlining the necessity of the mirror cleaning in ITER. The results of R&D on plasma cleaning of molybdenum diagnostic mirrors are reported. The mirrors contaminated with amorphous carbon films in the laboratory conditions and in the tokamaks were cleaned in steady-state hydrogenic plasmas. The maximum cleaning efficiency of 4.2 nm/min was reached for the laboratory and soft tokamak hydrocarbon films, whereas for the hard tokamak films the carbidization of mirrors drastically decreased the cleaning efficiency down to 0.016 nm/min. This implies the necessity of sputtering cleaning of contaminated mirrors as the only reliable tool to remove the deposits by plasma cleaning. An overview of R&D program on mirror cleaning is provided along with plans for further studies and the recommendations for ITER mirror-based diagnostics.

Behaviour and characteristics of W plasma-facing components under impinging high heat fluxes are investigated in view of the material choices for the divertor in future devices such as ITER and DEMO. Experiments have been carried out in the plasma edge of the TEXTOR tokamak to study melt-layer motion, macroscopic tungsten erosion from the melt layer as well as the changes in material properties such as grain size and abundance of voids or bubbles. The parallel heat flux at the radial position of the plasma-facing components (PFCs) in the plasma ranges around q(parallel to) similar to 45 MW m(-2) allowing samples to be exposed at an impact angle of 35 degrees to 20-30 MW m(-2). Melt-layer motion perpendicular to the magnetic field is observed following a Lorentz force originating from thermoelectric emission of the hot sample. Up to 3 g of molten W are redistributed forming mountain-like structures at the edge of the sample. The typical melt-layer thickness is 1-1.5 mm. Those hills are, due to the changes in the local geometry, particularly susceptible to even higher heat fluxes of up to the full q(parallel to). Locally the temperature can reach up to 6000 K, high levels of evaporation are causing significant erosion in the form of continuous fine-spray (similar to 1 x 10(24) atoms m(-2) s(-1)). Strong evaporation cooling is observed hindering the further heating of the samples. In addition, the formation of ligaments and splashes occurs several times during the melt phase ejecting droplets in the order of several 10 mu m up to 100 mu m probably caused by an instability evolving in the melt. In terms of material degradation several aspects are considered: formation of leading edges by redistributed melt, bubble formation and recrystallization. Bubbles are occurring in sizes between 1 and 200 mu m while recrystallization increases the grain size up to 1.5 mm. The power-handling capabilities are thus severely degraded. Melting of tungsten (W) in future devices is highly unfavourable and needs to be avoided especially in light of uncontrolled transients and possible unshaped PFCs