Foreground Component Separation for the SPIDER Balloon-Borne CMB Polarimeter

Abstract

Detecting primordial B-modes in the polarized cosmic microwave background anisotropies would be an indirect measurement of the gravitational wave background generated in the very early Universe, as can be accommodated within Inflationary theories. This measurement is complicated by the presence of bright Galactic foregrounds that obscure this cosmological signal. The first half of this dissertation focuses on hardware development for foreground measurements. The first is the design and development of a 30,40\,GHz diplexed camera for measuring synchrotron radiation; detectors of this type have been successfully integrated in BICEP Array.The second section discusses the development of \Spider II, a CMB balloon-polarimeter. It successfully flew in December 2022, and observed 10\% of the sky with three frequency bands: 95, 150\, and 280\,GHz. The data acquired from \Spider II are anticipated to make the deepest maps of polarized Galactic dust emission to date, which will be invaluable for refining our understanding of dust emission for the next generation of experiments. The second part of this dissertation focuses on the development of analysis techniques. First, it discusses the development and validation of a harmonic-based component separation method, SMICA. This analysis has provided an upper limit on the tensor-to-scalar ratio $r_{0.05} < 0.24$ at 95\% confidence using data from \Spider's first flight (95 and 150\,GHz) and the four \Planck HFI polarization maps. Additionally, this dissertation discusses the potential for ICA as a method to remove radio-frequency interference, a common problem across many microwave telescopes with a transition-edge sensor bolometer array architecture. Lastly, this dissertation demonstrates the use of Gaussian Mixture Models to identify the presence of spatially distinct dust populations using intensity data from \Planck HFI and \IRAS 100$\mu m$.