Polarisation spectral synthesis for Type Ia supernova explosion models

We present a Monte Carlo radiative transfer technique for calculating synthetic spectropolarimetry for multi-dimensional supernova explosion models. The approach utilises "virtual-packets" that are generated during the propagation of the Monte Carlo quanta and used to compute synthetic observables for specific observer orientations. Compared to extracting synthetic observables by direct binning of emergent Monte Carlo quanta, this virtual-packet approach leads to a substantial reduction in the Monte Carlo noise. This is vital for calculating synthetic spectropolarimetry (since the degree of polarisation is typically very small) but also useful for calculations of light curves and spectra. We first validate our approach via application of an idealised test code to simple geometries. We then describe its implementation in the Monte Carlo radiative transfer code ARTIS and present test calculations for simple models for Type Ia supernovae. Specifically, we use the well-known one-dimensional W7 model to verify that our scheme can accurately recover zero polarisation from a spherical model, and to demonstrate the reduction in Monte Carlo noise compared to a simple packet-binning approach. To investigate the impact of aspherical ejecta on the polarisation spectra, we then use ARTIS to calculate synthetic observables for prolate and oblate ellipsoidal models with Type Ia supernova compositions.