Search

Heating with ICRF waves is a well-established method on present-day tokamaks and one of the heating systems foreseen for ITER. However, further work is still needed to test and optimize its performance in fusion devices with metallic high-Z plasma facing components (PFCs) in preparation of ITER and DEMO operation. This is of particular importance for the bulk ion heating capabilities of ICRF waves. Efficient bulk ion heating with the standard ITER ICRF scheme, i.e. the second harmonic heating of tritium with or without ³He minority, was demonstrated in experiments carried out in deuterium-tritium plasmas on JET and TFTR and is confirmed by ICRF modelling. This paper focuses on recent experiments with ³He minority heating for bulk ion heating on the ASDEX Upgrade (AUG) tokamak with ITER-relevant all-tungsten PFCs. An increase of 80% in the central ion temperature Ti from 3 to 5.5 keV was achieved when 3 MW of ICRF power tuned to the central 3He ion cyclotron resonance was added to 4.5 MW of deuterium NBI. The radial gradient of the Ti profile reached locally values up to about 50 keV/m and the normalized logarithmic ion temperature gradients R/LTi of about 20, which are unusually large for AUG plasmas. The large changes in the Ti profiles were accompanied by significant changes in measured plasma toroidal rotation, plasma impurity profiles and MHD activity, which indicate concomitant changes in plasma properties with the application of ICRF waves. When the ³He concentration was increased above the optimum range for bulk ion heating, a weaker peaking of the ion temperature profile was observed, in line with theoretical expectations. [ABSTRACT FROM AUTHOR]

In ASDEX Upgrade H-modes with H980.95, similar effect of the ICRF antenna loading improvement by local gas injection was observed as previously in L-modes. The antenna loading resistance Ra between and during ELMs can increase by more than 25% after a switch-over from a deuterium rate of 7.5-1021 D/s injected from a toroidally remote location to the same amount of deuterium injected close to an antenna. However, in contrast to L-mode, this effect is small in H-mode when the valve downstream w.r.t. parallel plasma flows is used. In L-mode, a non-linearity of Ra at PJCRF<30 kW is observed when using the gas valve integrated in antenna. Application of magnetic perturbations (MPs) in H-mode discharges leads to an increase of Ra>30% with no effect of spectrum and phase of MPs on Ra found so far. In the case ELMs are fully mitigated, the antenna loading is higher and steadier. In the case ELMs are not fully mitigated, the value of Ra between ELMs is increased. Looking at the W source modification for the improved loading, the local gas injection is accompanied by decreased values of tungsten (W) influx Τw from the limiters and its effective sputtering yield Yw, with the exception of the locations directly at the antenna gas valve. Application of MPs leads to increase of Τw and Yw for some of the MP phases. With nitrogen seeding in the divertor, ICRF is routinely used to avoid impurity accumulation and that despite enhanced Τw and Yw at the antenna limiters. [ABSTRACT FROM AUTHOR]

In order to comply with the special challenges (open 4f-shell, configuration mixing) of simulating the spectrum of W23+ an extensive atomic model was implemented using the flexible atomic code (FAC). In detail, the basis functions from 11 configurations were used to model about 12000 levels, which give rise to roughly 60 million transitions including nearly 6 million electric and magnetic dipole transitions. A collisional radiative model has been put together which could handle the size of the input data. The modelled spectra (4-40 nm) show low sensitivity on the electron density, which validates the comparison of EBIT and tokamak spectra. The emissions between 4 and 7 nm are discussed in the context of the observations at fusion plasmas. In this range, the influence of W23+ is limited due to the small contribution to the measurement - however, elements of the presented modelling might explain the second, not understood spectral feature at 6 nm. Further details of the spectra are only briefly discussed as a close comparison to experimental data requires also models for the neighbouring ionisation stages. Additionally, the importance of configuration mixing becomes apparent motivating further investigations on neighbouring ionisation stages with similarly complex models. [ABSTRACT FROM AUTHOR]

W sputtering during ICRF on ASDEX Upgrade (AUG) and temperature rise on JET A2 antenna septa are considered in connection with plasma conditions at the antenna plasma facing components and E| near-fields. Large antenna-plasma clearance, high gas puff and low light impurity content are favorable to reduce W sputtering in AUG. The spatial distribution of spectroscopically measured effective W sputtering yields clearly points to the existence of strong E| fields at the antenna box (“feeder fields”) which dominate over the fields in front of the antenna straps. The picture of E| fields, obtained by HFSS code, corroborates the dominant role of E| at the antenna box on the formation of sheath-driving RF voltages for AUG. Large antenna-plasma clearance and low gas puff are favorable to reduce septum temperature of JET A2 antennas. Assuming a linear relation between the septum temperature and the sheath driving RF voltage calculated by HFSS, the changes of the temperature with dipole phasing (00ππ, 0ππ0 or 0π0π) are well described by the related changes of the RF voltages. Similarly to the AUG antenna, the strongest E| are found at the limiters of the JET A2 antenna for all used dipole phasings and at the septum for the phasings different from 0π0π. A simple general rule can be used to minimize E| at the antenna: image currents can be allowed only at the surfaces which do not intersect magnetic field lines at large angles of incidence. Possible antenna modifications generally rely either on a reduction of the image currents, on their short-circuiting by introducing additional conducting surfaces or on imposing the E| = 0 boundary condition. On the example of AUG antenna, possible options to minimize the sheath driving voltages are presented. [ABSTRACT FROM AUTHOR]

Since 2006, 85% of plasma facing components in ASDEX Upgrade are tungsten (W) coated, including all poloidal limiters of ICRF antennas. Sputtering of W in ICRF heated discharges and the increase of W concentration in the plasma are issues. In the discharges with higher heating power and smaller antenna-plasma distance, the parallel RF fields are more localized at the antenna region, but at the same time the absolute values of W influx and sputtering yield at the limiters are higher. To reduce the W sputtering yield, a low plasma temperature at the limiters is preferable, as shown in a set of discharges with a strong gas puff and a variation of antenna-plasma distance. Calculations of near-antenna parallel fields with High Frequency Structure Simulator code shows a significant contribution of the antenna box structure (including the limiters) to the parallel voltages on the long field lines in front of the antenna. This can partially explain the poor localization of RF parallel fields observed in the experiment and the fact of no clear experimental evidence of the reduction of the RF fields when magnetic field is aligned with the Faraday screen inclination angle. [ABSTRACT FROM AUTHOR]

Tungsten is one of very few candidate materials for plasma facing components in future fusion devices. Therefore, investigations have been started at fusion devices and EBITs to provide atomic data for W in fusion plasmas. Usually the influx of impurities is deduced from the intensity of spectral lines from neutrals or ions in a low ionisation state. For this purpose the appropriate ionisation rates and excitation rates have to be known. At the moment, a WI transition (7S-7P) at 400.9 nm is used, but an extension of the method to other lines is under investigation. In the core of present day plasmas ionisation states up to W56+ can be reached and in a reactor states up to around W68+ will be present. In order to extract information on the local W concentrations over the whole plasma radius atomic data (wavelength, excitation, ionisation, recombination) for all the charge states up to the maximum ionisation state are necessary. Similarly, a high sensitivity has to be achieved since the central W concentrations should stay below 10-4. For an unambiguous identification of the transitions EBIT measurements are of great advantage, but due to the lower electron density compared to fusion plasmas, investigations there are indispensable. © 2007 American Institute of Physics [ABSTRACT FROM AUTHOR]

In the ASDEX Upgrade tokamak we use localised electron cyclotron heating and current drive at 140 GHz. Both confinement and stability properties of the plasmas can be favourably altered. Particle transport can be modified with central ECRH. Linear gyrokinetic studies indicate that the transport is determined by collisionality and by the kind of turbulence which is dominant. A similar result is obtained for the electron heat transport. At higher collisionality the turbulent contribution to the diffusivity decreases. With central ECRH accumulation of high Z impurities can be prevented. Trace impurity injection experiments show that this is due to strongly increased impurity diffusion. Stability improvement can be achieved by controlling the sawtooth instability and thus the generation of seed islands for neoclassical tearing modes, by triggering their frequent interruption with ECCD, or by their complete suppression with ECCD on the resonant flux surface. © 2005 American Institute of Physics [ABSTRACT FROM AUTHOR]

In order to comply with the special challenges (open 4f-shell, configuration mixing) of simulating the spectrum of W23+ an extensive atomic model was implemented using the flexible atomic code (FAC). In detail, the basis functions from 11 configurations were used to model about 12000 levels, which give rise to roughly 60 million transitions including nearly 6 million electric and magnetic dipole transitions. A collisional radiative model has been put together which could handle the size of the input data. The modelled spectra (4-40 nm) show low sensitivity on the electron density, which validates the comparison of EBIT and tokamak spectra. The emissions between 4 and 7 nm are discussed in the context of the observations at fusion plasmas. In this range, the influence of W23+ is limited due to the small contribution to the measurement - however, elements of the presented modelling might explain the second, not understood spectral feature at 6 nm. Further details of the spectra are only briefly discussed as a close comparison to experimental data requires also models for the neighbouring ionisation stages. Additionally, the importance of configuration mixing becomes apparent motivating further investigations on neighbouring ionisation stages with similarly complex models. [ABSTRACT FROM AUTHOR]