Neoclassical level poloidal rotation measurements based on the inboard-outboard asymmetry of toroidal rotation in the TCV tokamak

Direct and indirect poloidal rotation measurements with improved accuracy were performed and compared in the TCV tokamak. The indirect measurement argues that, provided the plasma flow is divergence free on a flux surface, poloidal rotation can be inferred from the toroidal rotation at the high and low field sides of a flux surface. The key advantage of the method is an intrinsic amplification factor: instead of measuring poloidal rotation directly (typically ~few km/s i.e. of the order of the measurement accuracy), a difference in toroidal rotation is measured that is 4 to 10 times larger. Here, the main uncertainties arise from the flux surface mapping that are, however, largely compensated by this amplification factor. In TCV, the $C^{6+}$ toroidal rotation was measured across the whole plasma diameter by charge exchange (CX) spectroscopy for a series of low collisionality ($0.1<\nu^*<1.5$) OH and ECH L-mode plasmas, including positive and negative plasma current and toroidal magnetic field. Interestingly, the inferred poloidal rotation and the neoclassical theory predictions from the NEOART code, a variant of NCLASS, agree within $\pm 1$ km/s. In particular, a reversal of poloidal rotation is observed with the toroidal magnetic field direction, as predicted by theory.