Development of realistic simulations for the polarization of the cosmic microwave background

TL;DR

This paper develops a detailed simulation pipeline for CMB polarization experiments using HWPs, analyzing how non-idealities impact measurements of cosmic birefringence and tensor-to-scalar ratio, and proposing mitigation strategies.

Contribution

It introduces a new TOD simulation framework for LiteBIRD-like experiments and derives analytical models to assess HWP non-idealities effects.

Findings

HWP non-idealities can bias the cosmic birefringence angle by a few degrees.

HWP effects can lead to underestimation of the tensor-to-scalar ratio, r.

Gain calibration can partially mitigate HWP-induced biases.

Abstract

Polarization of the cosmic microwave background (CMB) can help probe cosmic inflation (via primordial $B$ modes) and test parity-violating physics (via cosmic birefringence), but realizing the potential of these opportunities requires precise control and mitigation of systematic effects. To this end, some experiments (including LiteBIRD) will use rotating half-wave plates (HWPs) as polarization modulators. Ideally, this choice should remove the $1/f$ noise component in the observed polarization and reduce intensity-to-polarization leakage, but any real HWP is characterized by non-idealities that, if not properly treated in the analysis, can lead to new systematics. In this thesis, after briefly introducing the science case, we discuss the macro steps that make up any CMB experiment, introduce the HWP, and present a new time-ordered data (TOD) simulation pipeline tailored to a LiteBIRD-like experiment that returns TOD and binned maps for realistic beams and HWPs. We show that the simulation framework can be used to study how the HWP non-idealities affect the measured cosmic birefringence angle, resulting in a few degrees bias for a realistic choice of HWP. We also derive analytical formulae to model the observed temperature and polarization maps and test them against the simulations. Finally, we present a simple, semi-analytical end-to-end model to propagate the non-idealities through the macro-steps that make up any CMB experiment (observation of multi-frequency maps, foreground cleaning, and power spectra estimation) and compute the HWP-induced bias on the estimated tensor-to-scalar ratio, $r$, finding that the HWP leads to underestimating $r$. We also show how gain calibration of the CMB temperature can be used to partially mitigate the non-idealities' effect and present a set of recommendations for the HWP design that can help maximize the benefits of gain calibration. [abridged]