Polarization of the WMAP Point Sources

TL;DR

This paper reports the detection of polarized sources in WMAP data using a new filtering technique, revealing 22 sources with significant polarization and estimating their polarization degrees, with implications for future Planck observations.

Contribution

Introduction of the 'filtered fusion technique' for detecting polarized sources in WMAP data and providing statistical polarization estimates.

Findings

Detected 22 polarized sources with high significance in WMAP data.

Estimated polarization degrees of bright sources are a few percent.

Identified a polarized flux limit of 300 mJy for Planck/LFI at 30 GHz.

Abstract

The detection of polarized sources in the WMAP 5-year data is a very difficult task. The maps are dominated by instrumental noise and only a handful of sources show up as clear peaks in the Q and U maps. Optimal linear filters applied at the position of known bright sources detect with a high level of significance a polarized flux P from many more sources, but estimates of P are liable to biases. Using a new technique, named the "filtered fusion technique", we have detected in polarization, with a significance level greater than 99.99% in at least one WMAP channel, 22 objects, 5 of which, however, do not have a plausible low radio frequency counterpart and are therefore doubtful. Estimated polarized fluxes P < 400 mJy at 23 GHz were found to be severely affected by the Eddington bias. The corresponding polarized flux limit for Planck/LFI at 30 GHz, obtained via realistic simulations, is 300 mJy. We have also obtained statistical estimates of, or upper limits to the mean polarization degrees of bright WMAP sources at 23, 33, 41, and 61 GHz, finding that they are of a few percent.