UC Berkeley

Measurements of the polarization anisotropies in the cosmic microwave background (CMB) provide powerful experimental constraints on the contents and history of the universe. These anisotropies are conventionally split by spatial parity, with even (odd) parity modes called E-modes (B-modes). The degree-scale (multipole moment l ≈ 100) B-mode polarization is expected to receive a contribution from a stochastic gravitational wave background in the early universe. This feature is a generic prediction of theories of cosmic inflation, however its overall amplitude r is a free parameter. Detecting this gravitational wave signal presents a formidable experimental challenge.

This analysis is based on three years of data from the Polarbear experiment using a continuously rotating half-wave plate to modulate the CMB polarization. This dissertation describes the analysis pipeline used to transform the raw detector time ordered data (TOD) into a CMB power spectrum and likelihood on r. This analysis is the tightest constraint on the degree scale B-mode power spectrum yet achieved from a mid-lattitude site or with a polarization modulator. As such, it serves as a pathfinder for future instruments including the Simons Array and Simons Observatory.

This dissertation is structured as follows. Chapter 1 presents an overview of the science case for precise measurements of the CMB B-mode power spectrum. Chapter 2 gives a brief description of the Polarbear instrument and the observation strategy used in this dataset. Chapter 3 describes the data pre-processing and calibration steps used in this analysis. Chapter 4 describes the pipeline used to turn the TOD into maps and subsequently power spectra. Chapter 5 describes the sensitivity degradation seen at large angular scales in this analysis and how it may be mitigated in future experiments. Chapter 6 describes the final calibrations and consistency of this data with existing E-mode measurements. Chapters 7 and 8 describe the detailed internal consistency checks preformed and simulations of a number of known sources of systematic contamination. Chapter 9 shows the recovered B- mode angular power spectrum. Chapter 10 shows the cross correlation with data from the Planck satellite to control contamination due to galactic foregrounds and the likelihood model for the constraint on r. Chapter 11 summarizes these results and describes future experiments.