Braginskii viscosity on an unstructured, moving mesh accelerated with super-time-stepping

Abstract

We present a method for efficiently modelling Braginskii viscosity on an unstructured, moving mesh. Braginskii viscosity, i.e., anisotropic transport of momentum with respect to the direction of the magnetic field, is thought to be of prime importance for studies of the weakly collisional plasma that comprises the intracluster medium (ICM) of galaxy clusters. Here anisotropic transport of heat and momentum has been shown to have profound consequences for the stability properties of the ICM. Our new method for modelling Braginskii viscosity has been implemented in the moving mesh code Arepo. We present a number of examples that serve to test the implementation and illustrate the modified dynamics found when including Braginskii viscosity in simulations. These include (but are not limited to) damping of fast magneto-sonic waves, interruption of linearly polarized Alfv\'en waves by the firehose instability and the inhibition of the Kelvin-Helmholtz instability by Braginskii viscosity. An explicit update of Braginskii viscosity is associated with a severe time step constraint that scales with $(\Delta x)^2$ where $\Delta x$ is the grid size. In our implementation, this restrictive time step constraint is alleviated by employing 2nd order accurate Runge-Kutta-Legendre super-time-stepping. We envision including Braginskii viscosity in future large-scale simulations of Kelvin-Helmholtz unstable cold fronts in cluster mergers and AGN-generated bubbles in central cluster regions.