Separating CMB Stokes Q and U polarization signals from Non-Gaussian Emissions

TL;DR

This paper presents a novel foreground minimization technique for CMB polarization maps that leverages the Gaussian nature of CMB and the non-Gaussian nature of foregrounds, improving the separation of CMB signals from foreground contamination.

Contribution

The authors develop a new ILC-based method that utilizes Gaussianity assumptions to effectively separate CMB polarization signals from non-Gaussian foregrounds, validated through simulations.

Findings

Significant foreground removal from polarization maps.

Reduced EE and BB power in cleaned maps compared to WMAP.

Method is robust and independent of foreground modeling uncertainties.

Abstract

In this paper we estimate diffuse foreground minimized Cosmic Microwave Background (CMB) Stokes Q and U polarization maps based upon the fundamental concept of Gaussian nature of CMB and strong non-Gaussian nature of astrophysical polarized foregrounds using WMAP nine year published polarization maps. We excise regions of the sky that define position of the known point sources, regions that are strongly contaminated by either the detector noise or by the diffuse foregrounds or both, and then perform foreground minimizations over the surviving sky regions that constitute approximately $50\%$ of the full sky area. We critically evaluate performance of foreground minimizations in several ways and show that our foreground minimization method removes significant foregrounds from input maps. The cleaned Stokes \{Q, U\} polarization maps have less EE and BB power from relevant sky region compared to WMAP foreground-reduced Stokes \{Q, U\} polarization maps at different multipole ranges. We validate our methodology by performing detailed Monte Carlo simulations. The main driving machinery of our method is an internal-linear-combination (ILC) approach, however, unlike simple variance minimization performed in usual ILC method, the core of the method being dependent on the theoretically well motivated concept of Gaussianity of CMB polarization a direct connection is established between observations and models of inflation. Additionally, the method, like the usual ILC method, is independent on modeling uncertainties of polarized foregrounds. It will be useful to apply our method in future generation low-noise CMB polarization experiments.