Precision Foreground Removal in Cosmic Microwave Background Polarization Maps

The latest Cosmic Microwave Background (CMB) experiments, such as the Advanced ACTPol experiment (AdvACT) and Simons Observatory (SO), will advance the state of the art in mapping the CMB polarization and temperature by offering unique combinations of sensitivity, resolution, spectral channels, and sky coverage. These upgraded combinations will enable precise measurements of the conditions of the early universe, and will also constrain the large-scale structure of the Universe, thereby allowing us to measure the sum of the neutrino masses and dark energy [1]. Accurate subtraction of polarized foreground emission is contingent upon precise knowledge of an experiment’s detectors’ responses as a function of frequency over the chosen frequency bands. Previous photometric calibration methods, including those used by Planck and other leading experiments, are not precise enough for the next generation of ground-, balloon-, and space-based CMB experiments due to the power-law frequency dependence of foreground components. Attaining the precision necessary to achieve the science goals of modern CMB experiments requires a dedicated effort to push the state of the art in the calibration of the detectors’ spectral responses. A Fourier transform spectrometer (FTS) was used to perform detector calibration in the first-generation ACTPol experiment [1] and achieved 2-GHz accuracy on the 150 GHz band centers. A new FTS with larger etendue is required to calibrate the low-frequency (LF) channels of AdvACT and to enable fast calibration of the full focal plane. New coupling optics are also needed to couple the full etendue of the FTS onto our detectors with a matched f-number while controlling loading. My research began with characterizing the new LF FTS for AdvACT and automating the bandpass analysis to handle large data sets with ease. I then extended this work to the much larger Simons Observatory (SO). I designed a new set of coupling optics to couple an FTS to the SO Large Aperture Telescope Receiver (LATR), where the cryogenic detectors are housed, to eventually perform the SO precision bandpass calibration. I modeled our new LF FTS using optical software Code V (shown in Figure 1), to understand exactly what the FTS beams were doing and how they needed to couple to each telescope’s detector focal plane. Each set of coupling optics should have an f-number that ensures the beam from the optics will fill the detector feedhorns. This requirement, along with the consideration of excess thermal loading on the detectors from lenses in the coupling optics, put limitations on lens thicknesses and design layout. I designed refractive coupling optics with an f-number of 1.52 for AdvACT, as well as the reflective coupling optics for calibration of the larger SO detector focal planes displayed in Figure 1. Reflective coupling optics will reduce thermal loading on the detectors and eliminate systematics introduced by refractive lenses. At Cornell I took measurements with two FTS’s, using an FTS at Cornell that is identical to our FTS at Michigan. By measuring each FTS’s spectral response separately, as well as the spectral response of both FTS’s in series (i.e. the output of one FTS is sent into the input of a second FTS), one can get a measure of the absolute transfer function of each FTS, as illustrated in Figure 2. We were still able to obtain a 2D interferogram and a 2D FFT (Figure 3), which shows each FTS’s single spectral response (horizontal and vertical), as well as the response of both FTS’s in series (along the diagonal), shown in Figure 4. There is still much to be done to continue this work and achieve the precision necessary to achieve SO and future CMB experiments’ science goals. Continued research into high-precision photometric calibration technology will benefit all CMB, millimeter wave, and submillimeter wave experiments.