The Q/U Imaging Experiment: Polarization Measurements of Radio Sources at 43 and 95 GHz

TL;DR

This paper reports polarization measurements of extragalactic radio sources at 43 and 95 GHz, revealing diverse polarization behaviors and challenging simple frequency-scaling models, based on QUIET survey data.

Contribution

First polarization measurements of radio sources at these frequencies from QUIET, providing new insights into their polarization properties and variability.

Findings

Detected polarization in 6-7 sources per band at >3σ significance.

Polarization fractions up to 10-20% observed, with variability effects noted.

Simple models with fixed polarization fractions across frequencies are inadequate.

Abstract

We present polarization measurements of extragalactic radio sources observed during the Cosmic Microwave Background polarization survey of the Q/U Imaging Experiment (QUIET), operating at 43 GHz (Q-band) and 95 GHz (W-band). We examine sources selected at 20 GHz from the public, $>$40 mJy catalog of the Australia Telescope (AT20G) survey. There are $\sim$480 such sources within QUIET's four low-foreground survey patches, including the nearby radio galaxies Centaurus A and Pictor A. The median error on our polarized flux density measurements is 30--40 mJy per Stokes parameter. At S/N $> 3$ significance, we detect linear polarization for seven sources in Q-band and six in W-band; only $1.3 \pm 1.1$ detections per frequency band are expected by chance. For sources without a detection of polarized emission, we find that half of the sources have polarization amplitudes below 90 mJy (Q-band) and 106 mJy (W-band), at 95% confidence. Finally, we compare our polarization measurements to intensity and polarization measurements of the same sources from the literature. For the four sources with WMAP and Planck intensity measurements $>1$ Jy, the polarization fraction are above 1% in both QUIET bands. At high significance, we compute polarization fractions as much as 10--20% for some sources, but the effects of source variability may cut that level in half for contemporaneous comparisons. Our results indicate that simple models---ones that scale a fixed polarization fraction with frequency---are inadequate to model the behavior of these sources and their contributions to polarization maps.