TCV – Development of new plasma shapes

“Snowflake” divertor

The novel “snowflake‘ divertor topology, recently proposed as a possible route towards reducing the heat loads on a reactor’s divertor plates, has been successfully created and studied for the first time in TCV. This configuration is characterized by a second-order null point in which both the poloidal magnetic field and its gradient vanish, resulting in six separatrix branches − visually resembling a snowflake − and four divertor legs. Stable ELMing H-modes have also been obtained and investigated in this topology in TCV.

The primary merit of the snowflake topology is an increase in the flux expansion in the X-point region by a factor of typically 2–5, accompanied by a comparable increase in the connection length, potentially reducing the heat load on appropriately situated divertor plates. The plasma perturbations in the divertor region are also expected to be more decoupled from those in the scrape-off layer (SOL) and radial transport generally slowed down as a result.

The exact snowflake configuration is topologically unstable, in that divertor coil current variations cause it to drift into a so-called snowflake-plus (SF+), with a first-order null and an unconnected secondary X-point, or a snowflake-minus (SF−), possessing two X-points that share a separatrix. These configurations, of course, chart a continuum, which can be conveniently parametrized by the distance between X-points normalized to the plasma minor radius, which we label σ. We have obtained all of these configurations in TCV, indeed exploring the continuum dynamically in a single discharge, by employing current combinations in the 16 independently powered external shaping coils.

Video images clearly confirmed the separatrix geometries, and inversion of tomographic data from quasi-bolometric AXUV detector arrays placed the maximum radiated power in the X-point region. The magnetic properties, particularly an expected increase in magnetic shear at the edge, were verified through equilibrium reconstruction.

More recently, we have applied Electron Cyclotron Heating to SF+ plasmas with the express purpose of investigating the possibility of sustaining a stable H-mode. A value σ = 0.5 was chosen as a compromise between flux expansion and stable controllability. A combination of 1MW top-injected third harmonic (X3) heating and 0.5MW X2 heating localized at the edge (the plasma core being beyond the density cutoff layer) was employed to induce the L–H transition. A standard single-null (SN) and a SF+ plasma, with closely matched shapes, were generated in the same discharge for ease of comparison; in both cases an ELMy H-mode was sustained. The measured properties were verified to be independent of the time history, i.e. the order in which the two topologies were obtained. A modest (15%) enhancement in the stored energy occurs in the SF+ as opposed to the SN, which could be consistent with the effect of the residual shape difference. An important by-product of this study was a thorough mapping of the L–H power threshold over a broad density range, which also, remarkably, revealed no systematic difference between the two cases.

One of the primary goals of this study was to investigate the ELM phenomenology, which proved to differ markedly between the two cases. While ELMs can be classified as type I in both configurations, owing to their frequency exhibiting an increasing dependence on power, the frequency is lower by a factor of 2–3 for the SF+, whereas the fractional energy loss per ELM is only 20–30% larger than in the SN plasma. These two observations combine to yield a favorable average energy loss scaling for the snowflake configuration. MHD stability calculations predict that SF plasmas are slightly, but systematically, more stable to intermediate-toroidal-number kink-ballooning modes, mainly as a result of higher edge magnetic shear.

Studies of the edge and SOL properties of the snowflake configuration are ongoing, to extend previous ELM studies using infrared imaging, AXUV tomography and Langmuir probes.

AK, May 15th, 2012 from S. Coda for the TCV team; Progress and scientific results in the TCV tokamak; Nucl. Fusion 51 (2011) 094017