Simulating TeV gamma-ray morphologies of shell-type supernova remnants

Supernova remnant (SNR) shocks provide favourable sites of cosmic ray (CR) proton acceleration if the local magnetic field direction is quasi-parallel to the shock normal. Using the moving-mesh magneto-hydrodynamical (MHD) code AREPO we present a suite of SNR simulations with CR acceleration in the Sedov-Taylor phase that combine different magnetic field topologies, density distributions with gradients and large-scale fluctuations, and -- for our core-collapse SNRs -- a multi-phase interstellar medium with dense clumps with a contrast of $10^4$. Assuming the hadronic gamma-ray emission model for the TeV gamma-ray emission, we find that large-amplitude density fluctuations of $\delta\rho/\rho_0\gtrsim75$ per cent are required to strongly modulate the gamma-ray emissivity in a straw man's model in which the acceleration efficiency is independent of magnetic obliquity. However, this causes strong corrugations of the shock surface that are ruled out by gamma-ray observations. By contrast, magnetic obliquity-dependent acceleration can easily explain the observed variance in gamma-ray morphologies ranging from SN1006 (with a homogeneous magnetic field) to Vela Junior and RX J1713 (with a turbulent field) in a single model that derives from plasma particle-in-cell simulations. Our best-fit model for SN1006 has a large-scale density gradient of $\nabla{n}\simeq 0.0034~\mathrm{cm}^{-3}~\mathrm{pc}^{-1}$ pointing from south-west to north-east and a magnetic inclination with the plane of the sky of $\lesssim10^\circ$. Our best-fit model for Vela Junior and RX J1713 adopts a combination of turbulent magnetic field and dense clumps to explain their TeV gamma-ray morphologies and moderate shock corrugations.