Trade-off between angular resolution and straylight contamination in the PLANCK Low Frequency Instrument

The last generation of CMB anisotropy experiments operating either from space, like the WMAP and PLANCK satellite, from the atmosphere, such as balloons, or from the ground, like interferometers, make use of complex multi-frequency instruments at the focus of meter class telescopes to allow the joint study of CMB and foreground anisotropies, necessary to achieve an accurate component separation. Between ∼70 GHz and ∼300 GHz, where foreground contamination is minimum, it is extremely important to reach the best trade-off between the improvement of the angular resolution, necessary for measuring the high order acoustic peaks of CMB anisotropy, and the minimization of the straylight contamination mainly due to the Galactic emission. This is one of the most critical systematic effects at large and intermediate angular scales (i.e. at multipoles ℓ less than ≈100) and consists in unwanted radiation entering the beam at large angles from the direction of the antenna boresight direction. We consider here the 30 and 100 GHz channels of the PLANCK Low Frequency Instrument (LFI). Assuming the nominal PLANCK scanning strategy, we evaluate the straylight contamination introduced by the most relevant Galactic foreground components for a reference set of optical configurations, accurately simulated as described in Sandri et al. (2004, A&A, 428, 299) (hereafter Paper 1). Given the overall constraints to the LFI optical design, we show that it is possible to improve the angular resolution by 5-7% by keeping the overall peak-to-peak signal of the Galaxy straylight contamination (GSC) below the level of few μK (and about 10 times smaller in terms of rms). A comparison between the level of straylight introduced by the different Galactic components for different beam regions (intermediate and far beam) is presented. We provide approximate relations, both for the intermediate and the far beam, for the rms and the peak-to-peak levels of the GSC as functions of the corresponding contributions to the integrated beam or of the spillover. For some reference cases we compare the results based on Galactic foreground maps derived from radio, IR, and Hα templates with those based on WMAP maps including CMB and extragalactic source fluctuations. The implications for the GSC in the PLANCK LFI polarization data are discussed. Finally, we compare the results obtained at 100 GHz with those at 30 GHz, where the GSC is more critical.