Theoretical requirements for calculating heavy impurity transport in rotating plasmas

Toroidal plasma flow shear is known to have a profound stabilizing effect on drift-wave turbulence and radial transport. In practice, gyrokinetic theory and simulation operate almost exclusively in the weak-rotation limit, retaining only the ExB flow, Coriolis drift and toroidal rotation shear. Proper treatment of sonic rotation, however, requires inclusion of centrifugal effects, which are quadratic in the Mach number. In 1998, Sugama derived a comprehensive formulation of gyrokinetic theory including sonic rotation and associated centrifugal terms valid for general electromagnetic perturbations. We show that implementation of this complete formulation is critical for the study of heavy impurity transport. In particular, turbulent fluxes of tungsten at finite Mach number are heavily modified by the new terms, even though deuterium ions and electrons are mostly unaffected. To this end, we discuss implications for core tungsten accumulation in a reactor, and remark that for realistic tungsten modeling both turbulent and neoclassical transport must be considered. These claims are based on neoclassical NEO simulations together with nonlinear CGYRO simulations, and suggest that tungsten transport calculations with existing reduced transport models may be unreliable. In addition, we discuss a new approach for the implementation of ExB flow shear. This is different than the previous rotation terms and cannot be treated simply or directly in a flux-tube. In the past, ExB shear has been simulated using either non-periodic boundary conditions or with a discontinuous wavenumber shift method. We report on the development of a new algorithm that is continuous and can treat the shear with spectral accuracy. This new method sheds light on recent gyrokinetic code disagreements.