Experimental and numerical investigation of neon-seeded high radiation discharges at the JET tokamak

Generating energy from nuclear fusion is an attractive alternative to energy sources that produce considerable amounts of carbon dioxide or radioactive waste. The currently most advanced concepts for a fusion reactor are based on magnetic confinement. In future fusion devices the power deposited onto the divertor target plates can damage them, if not mitigated. The necessary limitation of the plasma temperature in front of the targets to below about 4 eV can only be achieved by dissipating the fusion power before it reaches the target. In order to increase the power losses by electromagnetic radiation, impurity species are deliberately injected (or "seeded") into the plasma. The seeding of impurities that increase the radiation power from the region of open field lines, the scrape-off layer (SOL), and the divertor is an established technique. However, to mitigate the power of devices like the future demonstration power plant DEMO, the radiation losses from these regions are not sufficient and additional power losses from the region of closed field lines are required. To avoid a degradation of the fusion performance and the plasma stability by impurities in the plasma core, the impurity radiation in the confined region should primarily be located in the edge, close to the separatrix. Currently, efforts are made to investigate ways to increase the radiation power in the edge and the ways it affects the core stability. This thesis contributes to these efforts by the analysis of data from neon-seeded discharges in the currently largest tokamak JET and from numerical simulations of these discharges with the SOLPS-ITER transport code package. In many of the neon-seeded JET discharges, phases of a high energy confinement mode without characteristic periodic edge instabilities (M-mode) were observed. During these M-mode phases the radiation was concentrated in the edge close to the magnetic null (called X-point). The density at the steep edge pressure gradient (pedestal) was degraded and the power flux onto the divertor target plates was reduced (a state called detachment). These three phenomena seemed to occur simultaneously during the M-mode phases and a correlation among them was assumed. Furthermore, the neon seeding led to an increase of the core temperature as well as a reduction of the influx of wall impurities into the core. The comparably low energy confinement time was caused by the experiments' setup, but was not further reduced by the neon seeding. The core and edge radiation powers increased with increased neon seeding, whereas the divertor radiation power saturated, similar to the total radiative power fraction. Due to the high confinement mode in conjunction with high radiation powers and detached targets, this regime appears to be promising with regard to future fusion devices. To investigate the possibility of a correlation between the phenomena observed during the M-mode phases numerical simulations were set up in SOLPS-ITER similar to the experiments. In the simulations the neon seeding rate was ramped up until they became numerically unstable. Like in the experiment higher neon seeding rates led to an increase in core radiation. Although the radiation in the edge was increased as well, a pronounced X-point radiator was only found in cases that were not converged. At highest neon seeding rates the inner target was in detachment, but the outer target was not and remained in the low recycling regime. The pressure losses at the inner target necessary for detachment originated mostly from a region in the divertor close to the target, whereas the pressure losses in the pedestal were much lower. A lack of divertor neutral deuterium density due to the core fuelling assumed in the simulations as well as the neglect of drift effects are discussed to be the main cause for the attached outer target as well as for the low radiation powers in the simulations. Despite these limitations the simulations could, at least partly, reproduce key experimental features. Their results imply that under the given simulation conditions, the correlation between X-point radiator, the drop in pedestal pressure and detachment of the targets is not as strong as the experiments suggest.