An upper limit to polarized submillimetre emission in Arp 220

TL;DR

This paper establishes an upper bound on polarized submillimetre emission from the ULIRG Arp 220 by observing at 850 μm with the SCUBA polarimeter on the JCMT. Through careful data reduction, systematic checks, and a Bayesian analysis, the authors derive a 99% confidence upper limit of 1.54% for the polarization fraction within a 15 arcsec beam, indicating no detectable polarized signal. The result constrains the magnetic-field geometry in Arp 220 and suggests polarized extragalactic submillimetre sources will not be a major contaminant for CMB polarization measurements, while acknowledging the need for more measurements with upcoming instruments like SCUBA-2 to generalize across ULIRGs.

Abstract

We report the results of pointed observations of the prototypical ultra-luminous infrared galaxy (ULIRG) Arp 220 at 850 microns using the polarimeter on the SCUBA instrument on the James Clerk Maxwell Telescope. We find a Bayesian 99 per cent confidence upper limit on the polarized emission for Arp 220 of 1.54 per cent, averaged over the 15 arcsec beam-size. Arp 220 can serve as a proxy for other, more distant such galaxies. This upper limit constrains the magnetic field geometry in Arp 220 and also provides evidence that polarized ULIRGs will not be a major contaminant for next-generation cosmic microwave background polarization measurements.

Figures (2)

• Figure 1: Timestreams for the extinction corrected, sky-subtracted on-source bolometer from the two runs. The top two panels show the 450 and 850$\,\mu$m signal (in Volts). The third panel shows the ratio of the 2 signals, and should be constant under ideal conditions. The bottom plot shows the measured 225 GHz CSO $\tau$ (black circles) and the polynomial fit (dotted line), together with the elevation angle of the observation (solid line). Vertical dashed lines indicate where pointing checks were performed. Left: The 2000 data are shown, with a solid black vertical line at sample 1664, denoting the transition from observations taken with H7 as the on-source bolometer to using bolometer G15. Right: The 2001 data highlights the problems associated with observing the target near the zenith. In particular, the data between samples ${\sim}\,750$ and ${\sim}\,1250$ demonstrate that the source probably drifted away from the bolometer centre around transit. These data were not included in the analysis. The opacity was much lower in the 2001 observations, and one can actually see the instrumental polarization in the 450$\,\mu$m data.
• Figure 2: Theoretical CMB angular power spectra with potential contributions from polarized dusty galaxies and with noise estimates from Planck. The solid lines show the expected CMB power spectrum from (top to bottom) temperature fluctuations, $E$-mode polarization, $B$-mode polarization with lensing, and $B$-mode without lensing. A standard cosmological model constrained by WMAP three year data has been assumed, along with a tensor-to-scalar ratio of 0.1. Also shown are the pixel polarized noise estimates from Planck at 353 GHz (upper dashed line) and 143 GHz (lower dashed line). The dotted line shows an estimate of the contribution of polarized dusty galaxies at 350 GHz assuming all the sources are 1.5% polarized with random orientation angles, the sources are not clustered, and that sources above 100 mJy have been removed. The two dot dash curves show two estimates from de Zotti et al. (1999) of the contribution of dusty polarized galaxies at 143 GHz, assuming 2% galaxy polarization.