Probing for Evidence of Spatial Variability in the Spectral Energy Distribution of CMB Foreground Components

Abstract

Recent characterization of the Cosmic Microwave Background (CMB) has successfully constrained cosmological models. However, it remains to be seen whether measurements of the CMB polarization anisotropy uncover cosmological B-modes, a key prediction of inflationary theories. A major limiting factor in the search for B-modes has been contamination by foreground components, especially Galactic dust. In this paper, we perform a joint analysis of first-flight SPIDER and High Frequency Instrument Planck data in an attempt to separate CMB and foreground dust signals. In particular, we seek to determine whether our data can constrain the dust spectral index spatial variation. As part of our analysis pipeline, we develop a novel map space noise masking routine that should have application in future studies. Following data processing, we perform pixel by pixel fits across 6 frequency bands (2 SPIDER bands and 4 Planck bands) in each of the I, Q, and U Stokes parameters. We repeat the fits at various characteristic angular scales. Our results indicate that a CMB plus dust component model is well defined only at small angular scales. At large scales, significant subpixel variation of model parameters induces unphysical x2 scaling. We find strong suggestions of spatial variation in the Stokes I dust spectral index at degree angular scales. While we also observe a dispersion in the polarization spectral indices, our data are inconclusive as to its statistical significance. We note that our conclusions are limited by bias in our pixel by pixel errorbar estimates; eliminating this bias is a priority for future analyses.