CMBPol Mission Concept Study: Prospects for polarized foreground removal

TL;DR

This study assesses the potential for a future CMBPol satellite to detect primordial gravitational waves by analyzing polarized foregrounds, emphasizing the importance of frequency range and component separation techniques.

Contribution

It provides an analysis of polarized foreground contamination and explores optimization strategies for frequency channels to improve primordial signal detection.

Findings

Polarized foregrounds are lower than primordial signals at ~100 GHz in small sky patches.

Over 75% of the sky, foregrounds exceed primordial signals by a factor of about eight.

Detection of r=0.01 signals with high significance is promising with multi-scale observations.

Abstract

In this report we discuss the impact of polarized foregrounds on a future CMBPol satellite mission. We review our current knowledge of Galactic polarized emission at microwave frequencies, including synchrotron and thermal dust emission. We use existing data and our understanding of the physical behavior of the sources of foreground emission to generate sky templates, and start to assess how well primordial gravitational wave signals can be separated from foreground contaminants for a CMBPol mission. At the estimated foreground minimum of ~100 GHz, the polarized foregrounds are expected to be lower than a primordial polarization signal with tensor-to-scalar ratio r=0.01, in a small patch (~1%) of the sky known to have low Galactic emission. Over 75% of the sky we expect the foreground amplitude to exceed the primordial signal by about a factor of eight at the foreground minimum and on scales of two degrees. Only on the largest scales does the polarized foreground amplitude exceed the primordial signal by a larger factor of about 20. The prospects for detecting an r=0.01 signal including degree-scale measurements appear promising, with 5 sigma_r ~0.003 forecast from multiple methods. A mission that observes a range of scales offers better prospects from the foregrounds perspective than one targeting only the lowest few multipoles. We begin to explore how optimizing the composition of frequency channels in the focal plane can maximize our ability to perform component separation, with a range of typically 40 < nu < 300 GHz preferred for ten channels. Foreground cleaning methods are already in place to tackle a CMBPol mission data set, and further investigation of the optimization and detectability of the primordial signal will be useful for mission design.