Characterization of Fast Ion losses in various ELMy regimes on the TCV tokamak

The large gradient region at the plasma edge of H-mode plasmas can suffer from cyclic instabilities, the so-called Edge Localized Modes (ELMs), that reduce the overall confinement for H-mode operation and provoke large heat loads towards the wall during the ELM crash. These can damage the Plasma Facing Components (PFCs) critically. The ELMs also lead to the undesired localised expulsion of fast ions (FI) further damaging the PFCs. Thus, the interplay between the FI, the ELMs and the present MHD activity has been extensively investigated on the TCV tokamak. Various H-mode scenarios displaying many ELM frequencies and amplitudes have been explored. First, in the presence of large type-I ELMs, MHD modes precursing the ELMs identified as ballooning modes [1], are observed in the magnetic pick-up coils and the Soft X-Ray (SXR). For the first time coherent diffusive transport of confined FIs preceding the ELM crash correlated with these modes has been observed. The novel TCV Fast Ion Loss Detector (FILD) measures a non-negligible amount of FI losses before the ELM crash, consistent with the decrease in the neutron level possibly indicating a significant FI transport. The characteristic frequency of these FI losses is consistent with the one observed with the SXR and the magnetic pick-up coils.

Additionally, FI non-neoclassical transport in the inter-ELM phase has also been observed in smaller ELM regimes with higher ELM frequency and lower plasma density. These modes observed on the FILD before and during the ELM crash are also in the magnetic pick-up coils. The inter-ELM modes lead to significant FI losses in between ELMs indicating that the FI confinement is adversely affected between ELMs and not only by the ELM crash. These losses are similar in amplitude to the ones produced by the ELMs potentially increasing the overall power deposition on the PFCs. Experimental analyses of the FI population using the FILD and the fast neutron detector are presented. A numerical analysis of the neoclassical transport of these scenarios is performed using ASCOT5 and TRANSP codes to compare with the experimental observations qualitatively. The MISHKA code is used to identify the inter-ELM modes as AEs.

[1] Z. Yan et al., Phys. Rev. Lett. 107, 055004 (2011).