Removing beam asymmetry bias in precision CMB temperature and polarisation experiments

Abstract: Asymmetric beams can create significant bias in estimates of the power spectra from CMB experiments. With the temperature power spectrum many orders of magnitude stronger than the B-mode power spectrum any systematic error that couples the two must be carefully controlled and/or removed. Here, we derive unbiased estimators for the CMB temperature and polarisation power spectra taking into account general beams and general scan strategies. A simple consequence of asymmetric beams is that, even with an ideal scan strategy where every sky pixel is seen at every orientation, there will be residual coupling from temperature power to B-mode power if the orientation of the beam asymmetry is not aligned with the orientation of the co-polarisation. We test our correction algorithm on simulations of two temperature-only experiments and demonstrate that it is unbiased. The simulated experiments use realistic scan strategies, noise levels and highly asymmetric beams. We also develop a map-making algorithm that is capable of removing beam asymmetry bias at the map level. We demonstrate its implementation using simulations and show that it is capable of accurately correcting both temperature and polarisation maps for all of the effects of beam asymmetry including the effects of temperature to polarisation leakage.