ASDEX Upgrade has recently finished its transition towards an all-W divertor tokamak, by the exchange of the last remaining graphite tiles to W-coated ones. The plasma start-up was performed without prior boronization. It was found that the large He content in the plasma, resulting from DC glow discharges for conditioning, leads to a confinement reduction. After the change to D glow for inter-shot conditioning, the He content quickly dropped and, in parallel, the usual H-Mode confinement with H factors close to one was achieved. After the initial conditioning phase, oxygen concentrations similar to that in previous campaigns with boronizations could be achieved. Despite the removal of all macroscopic carbon sources, no strong change in C influxes and C content could be observed so far. The W concentrations are similar to the ones measured previously in discharges with old boronization and only partial coverage of the surfaces with W. Concomitantly it is found that although the W erosion flux in the divertor is larger than the W sources in the main chamber in most of the scenarios, it plays only a minor role for the W content in the main plasma. For large antenna distances and strong gas puffing, ICRH power coupling could be optimized to reduce the W influxes. This allowed a similar increase of stored energy as yielded with comparable beam power. However, a strong increase of radiated power and a loss of H-Mode was observed for conditions with high temperature edge plasma close to the antennas. The use of ECRH allowed keeping the central peaking of the W concentration low and even phases of improved H-modes have already been achieved.

Abstract. Particle, electric charge and power fluxes for type-I ELMy H-modes are measured in the divertor of the ASDEX Upgrade tokamak by triple Langmuir probes, shunts, IR-thermography and spectroscopy. The discharges are in the medium to high density range, resulting in predominantly convective ELMs with moderate fractional stored energy losses of 2 % or below. Time resolved data over ELM cycles are obtained by coherent averaging of typically one hundred similar ELMs, spatial profiles from the flush-mounted Langmuir probes are obtained by strike point sweeps. Application of simple physics models is used to compare different diagnostics and to make consistency checks, e.g. the standard sheath model applied to the Langmuir probes yields power fluxes which are compared to the thermographic measurements. In between ELMs, Langmuir probe and thermography power loads appear consistent in the outer divertor, taking into account additional load due to radiation and charge exchange neutrals measured by thermography. The inner divertor is completely detached and no significant power flow by charged particles is measured. During ELMs, quite similar power flux profiles are found in the outer divertor by thermography and probes, albeit larger uncertainties in Langmuir probe evaluation during ELMs have to be taken into account. In the inner divertor, ELM power fluxes from thermography are a factor 10 larger than those derived from probes using the standard sheath model. This deviation is too large to be caused by deficiencies of probe analysis. The total ELM energy deposition from IR is about a factor 2 higher in the inner compared to the outer divertor. Spectroscopic measurements suggest a quite moderate contribution of radiation to the target power load.Shunt measurements reveal a significant positive charge flow into the inner target during ELMs. The net number of elementary charges correlates well with the total core particle loss obtained from highly resolved density profiles.As a consequence, the discrepancy between probe and IR measurements is attributed to the ion power channel via a high mean impact energy of the ions at the inner target. The dominant contributing mechanism is proposed to be the directed loss of ions from the pedestal region into the inner divertor.

In order to assess the contribution of edge localized modes (ELMs) to plasma-wall interaction in future fusion experiments like ITER, a sound experimental data basis for model validation and extrapolation, and, to be prepared for the unfavourable case, the development of tools for ELM mitigation are required. On ASDEX Upgrade, exploiting the fast edge/divertor diagnostics as well as active ELM control tools, especially ELM pacing by cryogenic pellet injection, a large amount of experimental information has been accumulated on the structure and dynamics of natural and induced ELMs, as well as on related wall effects. In this paper, a survey of ELM structure related diagnostics and results is given first. Recent progress in spontaneous ELM analysis is given then with emphasis on magnetic probe analysis, 2d Thomson scattering snapshots of pedestal density and temperature perturbations in comparison with electron cyclotron emision, and combined Langmuir, magnetic and thermographic investigation of filament dynamics in the scrape-off layer wing. Pellet induced ELMs are characterized in comparison to spontaneous ones. The combined experimental evidence is finally discussed in terms of simple considerations based on presently available, still rudimentary models.