Constraints on Electromagnetic Parity Violation in the Cosmic Microwave Background using BICEP3

Abstract

Conservation of parity in the electromagnetic (EM) force is a generally accepted assumption for the Standard Model of physics. This need not be true, however, and a number of frameworks have been invented which predict observables in the universe for cases of EM parity violations. For instance, a coupling of the EM Lagrangian with a Chern-Simons term manifests as a frequency-independent rotation of linearly polarized light as it travels through space and is called Cosmic Birefringence. The rotation of light in the presence of Cosmic Birefringence increases over the distance of travel and is thus expected to have the largest rotation from far away objects. As such, the Cosmic Microwave Background (CMB), which is the oldest and farthest source of light in the universe, is a natural candidate in the search for birefringent signatures. The CMB is naturally polarized and can be separated into parity-even and parity-odd (i.e., E-mode and B-mode) polarization states. Under parity violation, this results in a leakage between polarization states in a predictable way. However, this leakage signal as seen in experimental data is degenerate with an incorrect calibration of the polarization response angle of a CMB experiment. Thus, in order to be sensitive to celestial sources of polarization rotation, a CMB experiment must have a precise measurement of the polarization response.

In this thesis, we explore our potential to establish constraints on Cosmic Birefringence using the BICEP3 telescope which measures the linearly polarized light from the CMB at 95 GHz. We break the degeneracy between instrumental systematics and parity-violating physics by characterizing BICEP3's polarization response with respect to gravity using a Rotating Polarized Source (RPS). The RPS is a quasi-thermal and highly polarized source and its co-polar axis is registered with respect to gravity to within 0.09°. The RPS is capable of rotating 360° and we observe the source with BICEP3 at many different rotations of the RPS to derive the orientation of the co-polar axes of BICEP3's ~2000 linearly polarized detectors. We carry out a thorough assessment of the errors in the measured polarization angles, examining a variety of potential sources of systematic uncertainty. Most of the errors are well understood and controlled, with a minimum total statistical + systematic uncertainty of 0.086°. However, empirical benchtop characterization of the RPS's performance suggests there may be a large uncertainty that so far has been difficult to quantify accurately.

We construct an analysis to constrain Cosmic Birefringence by applying angle calibrations to BICEP3's full-year 2017 and 2018 CMB data. The uncertainties on this constraint due to instrument noise and CMB foregrounds are calculated through an extensive suite of end-to-end CMB simulations. In order to maintain a high degree of confidence in the analysis, we blind ourselves to real CMB data and only unblind to certain aspects of the real data when pre-established criteria are met. We are able to progress the analysis to the point where real data is used to complete a number of cross-checks that lend high confidence to calibration techniques. However, as a result of the uncertainty arising from benchtop measurements, we are unable to unblind fully to the birefringence angle. Instead, we establish a mock birefringence constraint for demonstration purposes by substituting a single realization from our suite of CMB simulations in lieu of real CMB data. The calibration techniques we have described in this thesis are shown to have the potential to measure birefringence angles with a precision of 0.13°, which would be the most powerful constraint supported by real telescope data. This potential will be realized in future work with continued advancements in the instrumental calibration, which were dominated entirely by the instrumental error from empirical benchtop measurements. Behind the instrumental systematic uncertainty, our next most limiting uncertainty comes from the statistical noise in our CMB maps. With the inclusion of more CMB data, the precision of future analyses will quickly be limited by extra-galactic gravitational lensing. Therefore, follow-up birefringence analyses will also need to consider CMB delensing techniques before significant improvement on birefringence constraints can be achieved.