Turbulent gas motions in galaxy cluster simulations: the role of smoothed particle hydrodynamics viscosity

TLDR

SPH’s artificial viscosity, especially in its original form, is large enough to damp turbulence and suppress random gas motions in the intracluster medium. The study demonstrates that this suppression occurs by comparing results from a time‑variable, particle‑specific viscosity formulation. Test calculations with strong shocks confirm that the variable‑viscosity scheme reproduces shock capture while reducing numerical viscosity in non‑shock regions. In nine high‑resolution cluster simulations, the low‑viscosity SPH produces 5–30 % of the thermal energy in turbulent motions, flattens central entropy, lowers core density, reconciles SPH–AMR differences, and enables efficient electron acceleration to explain diffuse radio emission.

Abstract

Smoothed particle hydrodynamics (SPH) employs an artificial viscosity to properly capture hydrodynamical shock waves. In its original formulation, the resulting numerical viscosity is large enough to suppress structure in the velocity field on scales well above the nominal resolution limit, and to damp the generation of turbulence by fluid instabilities. This could artificially suppress random gas motions in the intracluster medium (ICM), which are driven by infalling structures during the hierarchical structure formation process. We show that this is indeed the case by analysing results obtained with an SPH formulation where an individual, time-variable viscosity is used for each particle (Monaghan 1997). Using test calculations involving strong shocks, we demonstrate that this scheme captures shocks as well as the original formulation of SPH, but, in regions away from shocks, the numerical viscosity is much smaller. In a set of nine high-resolution simulations of cosmological galaxy cluster formation, we find that this low--viscosity formulation of SPH produces substantially higher levels of turbulent gas motions in the ICM, reaching a kinetic energy content in random gas motions (measured within a 1Mpc cube) of up to 5%-30% of the thermal energy content, depending on cluster mass. This has also significant effects on radial gas profile. We find a central flattening of the entropy profile and a reduction of the central gas density in the low--viscosity scheme. Interestingly, this tends to reduce the differences seen in SPH and adaptive mesh refinement simulations of cluster formation. Finally, invoking a model for particle acceleration by MHD waves driven by turbulence, we find efficient electron acceleration to power diffuse radio emission.