Search

EUROFER is a Reduced Activation Ferritic Martensitic (RAFM) steel developed as structural material for future fusion power plants and is considered as first-wall material in recessed areas of DEMO. Recent work has shown a fluence-dependent decrease of the sputter yield for bombardment with deuterium ions in the energy range of 100 to about 500 eV. This decrease was previously attributed to preferential sputtering of the lower mass constituents in EUROFER, such as Fe and Cr, compared to the higher mass alloying elements. This leads to an increase of the surface tungsten concentration. However, it was also observed that, after sputtering, the samples had developed a very rough surface morphology. In this work, the combined influence of surface roughness and W surface enrichment on the sputter yield of EUROFER under bombardment with 200 eV deuterium ions was studied. The influence of surface roughness was determined with the aid of Scanning Electron Microscopy (SEM) and SDTrimSP-3D simulations. W surface enrichment was investigated applying sputter X-Ray Photoelectron Spectroscopy (XPS) depth-profiling and Rutherford backscattering spectrometry (RBS). After bombardment to a fluence of 1024 D m−2 (at 200 eV per deuterium) a reduction of the sputter yield to 29% ± 5% of the initial value was measured by weight-loss measurements. This reduction is in good agreement with published values. Two distinct surface morphologies, consisting of smooth and spiked surfaces, were observed on the EUROFER sample after sputtering. Based on the experimental results, the combined effect of the two factors, surface roughness and W surface enrichment, is estimated to be responsible for a reduction in the sputter yield to 27% ± 4% of the initial value, which is in excellent agreement with the measured value. Our assessment shows that both surface morphology and W surface enrichment contribute significantly to the reduction of the sputter yield of EUROFER under the given experimental conditions, and are sufficient to fully explain the experimentally observed reduction in the sputter yield.

The influence of surface roughness on the sputter yield was investigated as a function of the angle of incidence. In this work, nm-smooth and rough samples with roughnesses on the µm length scale were produced by depositing thin Fe and W films on smooth and rough Si substrates via magnetron sputtering. The surface morphology of the samples was determined by atomic force microscopy. The samples were exposed to a 6 keV D3+ ion beam (2 keV/D) under various angles of incidence ranging from 0° to 75° with respect to the surface normal to fluences of the order of 1022 D/m2. The layer thickness was measured by Rutherford backscattering spectrometry (RBS) before and after erosion. The resulting sputter yields were compared to simulations performed with SDTrimSP (static and dynamic) and SDTrimSP-3D (static), showing good qualitative agreement in all cases, as well as agreement with literature data at normal incidence. This constitutes the first experimental benchmark of SDTrimSP-3D. A discrepancy in the value of the sputter yield for smooth W at normal incidence was observed between the SDTrimSP simulations and the experimental values obtained in this work and found in literature. Analogous experiments were performed to study the sputter yield at normal incidence of 2 keV/D on smooth Au and 6 keV He on smooth W. These sputter yields were also compared to SDTrimSP simulations and literature, showing good agreement in all cases. Keywords: Sputtering, SDTrimSP, Roughness

Time-of-flight Rutherford backscattering spectroscopy (ToF-RBS) with incident heavy ions offers a substantially improved depth resolution compared to conventional RBS with incident He ions. The new Garching ToF-RBS detector is located at a scattering angle of 150° with a free flight path of 1.313 m and a time resolution of 600 ps. The experimentally achieved depth resolution at the surface of W-containing Fe samples is 2–3 nm using incident Si ions. Model systems consisting of Fe layers with 0.7, 1.5 and 4.2 at% W and EUROFER steel (containing 0.34 at% W + Ta) were eroded by 200 eV D ions to a fluence of 1023 D/m2 at 310 K and 900 K. W depth profiles were measured using ToF-RBS, light impurities at the surface were detected using time-of-flight elastic recoil detection analysis (ToF-ERDA). The data sets from the two techniques were analyzed self-consistently. In all cases W enrichment at the surface was observed. The enrichment at 310 K is in good agreement with SDTrimSP simulations if the experimentally achieved depth resolution is taken into account. Annealing of the model systems with 0.7 and 1.5 at% W to 900 K for 10 h in vacuum resulted also in a W-enriched surface layer. Keywords: Rutherford Backscattering Spectroscopy, RBS, ToF-RBS, Tungsten surface enrichment, EUROFER, MultiSIMNRA

The erosion behaviour of EUROFER steel due to mono-energetic deuterium (D) ion bombardment in the energy range of 100–1000 eV/D was investigated. At low fluences, the sputtering yield is comparable to that of pure iron (Fe). It then decreases with increasing fluence and tends toward a steady state at larger fluences. The largest experimentally investigated fluences are of the order of several 1024 D/m2. The yield reduction is more pronounced for lower D impinging energies. A simple model is presented within which the evolution of the yield can be described by an exponential decay, using empirical values for the fitting parameters. In this model, the yield reduction is caused by preferential sputtering of Fe and the consequent development of a tungsten- (W-) enriched surface layer. SDTrimSP simulations also confirm the appearance of a W-enriched surface layer. However, the experimentally observed fluence dependence of the sputtering yield could not satisfactorily be reproduced in these simulations. The resulting enrichment layer thicknesses below 1 Å stretch the physics model implemented in SDTrimSP beyond its validity range. Experimentally, surface enrichment of W was qualitatively confirmed in Rutherford backscattering and sputter X-ray photoelectron spectroscopy measurements. However, the measured depth profiles are very likely influenced by the observed surface roughening caused by the D irradiation. Electron microscopy revealed the inhomogeneous W distribution on the un-irradiated sample as well as a grain-dependence of erosion and grain-dependent topography development. The inclusion of surface topography in the description of the erosion could be the key to improve the agreement between model predictions and the experiment. Keywords: RAFM steel, Erosion, Plasma-material interactions

We have studied the retention of nitrogen in surfaces with varying roughness and the transport of nitrogen from the divertor to the outer midplane in experiments at the tokamak ASDEX Upgrade. To allow for a reliable identification of nitrogen retained from the plasma exposure, 5.3 · 1021 atoms of the tracer isotope 15N were injected into the private flux region of the plasma. On polished W samples exposed to the plasma in the outer divertor, the N content peaks to both sides of the strike line with an areal density of 1.5 · 1020 15N/m2 and drops to a value of 1.0 · 1020 15N/m2 in the strike line region. In contrast, the N content of samples with a rougher surface peaks at the strike line and reaches areal densities of 3.0 · 1020 15N/m2. The N deposition at the outer limiters was measured via samples exposed on the midplane manipulator. At the limiter position the 15N areal density reaches a value of 0.2 · 1020 15N/m2, only a factor of ten smaller than the areal densities in the divertor. However, a comparison to WallDYN simulations shows that the observed N content is roughly a factor of 4 smaller than the value predicted by these simulations. Possible reasons for this discrepancy are the toroidal asymmetric main wall geometry, which currently cannot be fully included in the simulations, or an enhanced re-erosion of deposited N. Keywords: Tokamak, Impurity migration, Nitrogen, Roughness, WallDYN

Tungsten (W) will be used as material for plasma-facing components (PFCs) in the divertor of ITER and interact with Helium (He) ions either from initial He plasma operation or from Deuterium-Tritium (DT) fusion reactions in the active operation phase. Laboratory experiments reported that in a specific operational window of impact energy, ion fluence, and surface temperature (Ein ≥ 20 eV, ϕ ≥ 1 × 1024 He+mBt=2.5T,Ip=0.8MA,Paux≃8.0MW Tsurf ≥ 1000 K) a modification of W surfaces occurs resulting in the formation of He-induced W nanostructures. Experiments in ASDEX Upgrade H-mode plasmas (Ein=37keV T, ϕ≃0.75×1024He0m−2 MA, Paux ≃ 8.0 MW) in He have been carried out to investigate in detail (a) the potential growth of W nanostructures on pre-damaged W samples incorporating He nanobubbles, and (b) the potential ELM-induced erosion of W nanostructure. Both W surface modifications were generated artificially in the GLADIS facility by He bombardment of W samples at ϕ≃1×1024He0m−2 keV (a) to ϕ ≃ 0.75 × 1024 He0mϕ≃1.6×1024He+m−2 at Tsurf ≃ 1800 K and (b) ϕ ≃ 1 × 1024 He0m−2 at Tsurf ≃ 2300 K prior to exposure in the divertor manipulator of ASDEX Upgrade. Though in part (a) conditions of W nanostructure growth with a total He ion fluence of ϕ ≃ 1.6 × 1024 He+m−2 and peak He ion impact energies above 150 eV were met, no growth could be detected. In part (b) lower density plasmas with more pronounced type I ELMs, carrying energetic He ions in the keV range, were executed with the strike-line positioned on 2 µm thick W nanostructure accumulating a fluence of ϕ ≃ 0.8 × 1024 He+m−2. Post-mortem analysis revealed that co-deposition by predominantly W, and Boron (B), eroded at the main chamber wall and transported into the divertor, took place on all W samples. Erosion of W nanostructure or its formation was hindered by the fact that the outer divertor at the location of the samples was turned under these He plasma conditions into a net deposition zone by W, B and Carbon (C) ions. The surface morphology with large roughness and effective surface area act as a catcher for the impinging impurities. Thus, apart from operation in the existence diagram of W nanostructure with respect to Tsurf, ϕ, and Ein, also the impinging impurity flux contribution needs to be considered in predictions concerning the formation of W nanostructures. Keywords: PSI, ASDEX Upgrade, ITER, W divertor, W nanostructure, Helium

We have modelled net and gross erosion of W in low-density l-mode plasmas in the low-field side strike point region of ASDEX Upgrade by ERO and Particle-in-Cell (PIC) simulations. The observed net-erosion peak at the strike point was mainly due to the light impurities present in the plasma while the noticeable net-deposition regions surrounding the erosion maximum could be attributed to the strong E ×B drift and the magnetic field bringing eroded particles from a distance of several meters towards the private flux region. Our results also imply that the role of cross-field diffusion is very small in the studied plasmas. The simulations indicate net/gross erosion ratio of 0.2–0.6, which is in line with the literature data and what was determined spectroscopically. The deviations from the estimates extracted from post-exposure ion-beam-analysis data (∼0.6–0.7) are most likely due to the measured re-deposition patterns showing the outcomes of multiple erosion-deposition cycles. Keywords: ASDEX Upgrade, Tungsten erosion, ERO modelling, PIC simulations, Particle drifts, Cross-field diffusion

A global heat flux model based on a fractional derivative of plasma pressure is proposed for the heat transport in fusion plasmas. The degree of the fractional derivative of the heat flux, α, is defined through the power balance analysis of the steady state. The model was used to obtain the experimental values of α for a large database of the Joint European Torus (JET) carbon-wall as well as ITER like-wall plasmas. The fractional degrees of the electron heat flux are found to be α

In the ITER tokamak, injection of nitrogen is foreseen to decrease the heat loads on the divertor surfaces. However, once dissociated, nitrogen atoms react with hydrogen isotopes to form ammonia isotopologues. The formation of tritiated ammonia may pose some issues with regards to tritium inventory and operation duty cycle. In this paper, we report a study of the effect of three parameters of relevance for the fusion environment on the ammonia production, including the presence of a catalytic surface, sample temperature and noble gas addition. Results of ammonia formation from N2/H2 RF plasma (both with and without tungsten or stainless steel surface) show the importance of the presence of a catalyst in the ammonia formation process. By increasing the temperature of the W samples up to 1270 K, ammonia formation demonstrated a continuous decrease due to two major factors. For high temperatures above 650 K and 830 K, for stainless steel and W, respectively, the reduction results from the thermal decomposition of ammonia. For the lower temperature range, the temperature rise leads to the formation of more stable nitrides that do not tend to react further with hydrogen to form NH2 and NH3. Interestingly, the addition of helium or argon to the N2/H2 plasma show opposite effects on the ammonia production. He effectively decreases the percentage of NH3 by acting as a barrier for the surface processes. On the other hand, argon impacts the plasma processes more, probably by increasing the active nitrogen species in the plasma and as a consequence the percentage of formed ammonia. [ABSTRACT FROM AUTHOR]

Ammonia formation was studied in a series of dedicated nitrogen seeded H-mode discharges at ASDEX Upgrade. The evolution of ammonia formation was investigated with a reference phase before the seeding, and a long, stable flat-top nitrogen-seeded phase. It was monitored with divertor spectroscopy and analysis of the exhaust gas. The amount of the detected ammonia increased continuously over the course of five discharges with the same nitrogen seeding rate. The same trend was observed in the nitrogen density in the core plasma, as measured by charge exchange recombination spectroscopy and other signals, linked to the effects of nitrogen seeding. The results show that the rate of ammonia formation exhibited the same trend as the nitrogen density in the plasma. This density, in turn, was strongly influenced by the nitrogen wall inventory. The spatial distribution of the detected ammonia suggests that a significant contribution to the net ammonia formation is made in plasma-shaded areas, through surface reactions of neutral species. [ABSTRACT FROM AUTHOR]

Gas seeding is often used in tokamaks to reduce the power load onto the divertor target plates. Nitrogen is the preferred seeding species because of its favourable radiative properties as well as its apparent beneficial effect on plasma confinement. However, nitrogen molecules are chemically reactive with hydrogen and its isotopes to form stable ammonia compounds. Since ammonia is a polar molecule, sticking on metal surfaces can be expected, increasing as a consequence the tritium retention which could pose a serious risk for ITER operation and maintenance. It is, therefore, important to understand the adsorption mechanism of ammonia on surfaces, investigate when the surface saturation occurs and whether ammonia adsorbs as a molecule or undergoes a dissociation on the surface. In this contribution, ammonia sticking on different fusion-relevant materials is presented. The results show a pressure-dependent ammonia sticking on tungsten, boron and stainless steel followed by a partial desorption from these surfaces while on gold and beryllium, ammonia molecules weakly adsorb and completely desorb. A detailed explanation of the two interaction mechanisms is addressed. Furthermore, the time dependence of ammonia desorption as well as the chemical state of non-desorbed residuals were investigated with x-ray photoelectron spectroscopy. Tungsten, boron and stainless steel surfaces showed a continuous dissociation process from NH3 to NH2, NH, N and surface nitrides. [ABSTRACT FROM AUTHOR]

Plasma-wall interactions have been studied in the full-W ASDEX Upgrade during its dedicated helium campaign. Relatively clean plasmas with a He content of  >80% could be obtained by applying ion cyclotron wall conditioning (ICWC) discharges upon changeover from D to He. However, co-deposited layers with significant amounts of He and D were measured on W samples exposed to ICWC plasmas at the low-field side (outer) midplane. This is a sign of local migration and accumulation of materials and residual fuel in regions shadowed from direct plasma exposure albeit globally D was removed from the vessel. When exposing W samples to ELMy H-mode helium plasmas in the outer strike-point region, no net erosion was observed but the surfaces had been covered with co-deposited layers mainly consisting of W, B, C, and D and being the thickest on rough and modified surfaces. This is different from the typical erosion-deposition patterns in D plasmas, where usually sharp net-erosion peaks surrounded by prominent net-deposition maxima for W are observed close to the strike point. Moreover, no clear signs of W nanostructure growth or destruction could be seen. The growth of deposited layers may impact the operation of future fusion reactors and is attributed to strong sources in the main chamber that under suitable conditions may switch the balance from net erosion into net deposition, even close to the strike points. In addition, the absence of noticeable chemical erosion in helium plasmas may have affected the thickness of the deposited layers. Retention of He, for its part, remained small and uniform throughout the strike-point region although our results indicate that samples with smooth surfaces can contain an order of magnitude less He than their rough counterparts. [ABSTRACT FROM AUTHOR]

The impact of lithium (Li) on plasma performance was investigated at the ASDEX Upgrade tokamak, which features a full tungsten wall. Li pellets containing 1.6  ×  1020 Li atoms were launched with a speed of 600 m s−1 to achieve deep penetration into the plasma and minimize the impact on the first wall. Homogeneous transient Li concentrations in the plasma of up to 15% were established. The Li sustainment time in the plasma decreased with an increasing heating power from 150 to 40 ms. Due to the pellet rate being restricted to 2 Hz, no Li pile-up could take place. No significant positive impact on plasma properties, as reported from other tokamak devices, could be found; the Li pellets rather caused a small reduction in plasma energy, mainly due to enhanced radiation. Due to pellet injection, a short-lived Li layer was formed on the plasma-facing components, which lasted a few discharges and led to moderately beneficial effects during plasma start-up. Most pellets were found to trigger type-I ELMs, either by their direct local perturbation or indirectly by the altered edge conditions; however, reliability was less than 100%. [ABSTRACT FROM AUTHOR]

This work presents ASDEX Upgrade experiments, where the nitrogen deposition and re-erosion on divertor manipulator samples and the effect of its transport through the plasma were studied. These results are compared to WallDYN-DIVIMP simulations based on SOLPS plasma backgrounds and employing an improved WallDYN model, which includes the effusion of nitrogen from saturated surfaces. On one hand, this allows the WallDYN code and the new saturation model with a comprehensive data set to be benchmarked, on the other hand the simulations help in the interpretation of the experimental results. Both, experimental results and simulations, show that the N content in the region of the outer strike line reaches its steady-state value within one discharge. The simulations also reproduce the experimentally observed nitrogen content in samples exposed to N2-seeded discharges. With respect to the boron deposition, the nitrogen deposition in a non-seeded discharge and the re-erosion of nitrogen discrepancies to the WallDYN-DIVIMP simulations are observed. Based on SDTrimSP simulations, these are attributed to the missing depth resolution of the WallDYN surface model. A detailed comparison of spectroscopic measurements to WallDYN simulations, based on a novel synthetic spectroscopy diagnostic for WallDYN, shows that the nitrogen fluxes in the plasma are well described by the simulations. From a comparison of several WallDYN-DIVIMP simulations employing customized onion-skin model plasma backgrounds the physical processes controlling the nitrogen concentration in the core plasma and the applicability of onion-skin model plasma backgrounds are discussed. From these simulations the private flux zone with the gas valve, the outer baffle and the high field side main wall are identified as the main sources for the nitrogen content of the core plasma. [ABSTRACT FROM AUTHOR]