A Measurement of the Cosmic Microwave Background B-Mode Polarization Power Spectrum at Sub-Degree Scales with POLARBEAR

TL;DR

Polarbear measures the CMB B-mode power spectrum $C_\ ell^{BB}$ in the sub-degree regime ($500<\ell<2100$) with a high-resolution, low-noise polarization map over ~25 deg$^2$ at 150 GHz. The analysis employs a MASTER-based pseudospectrum pipeline, comprehensive calibration (including CMB self-calibration via $C_\ ell^{EB}$ and Tau A/Cen A polarization calibrations), foreground assessments, and extensive null tests and cross-checks with a second pipeline. The main result is a lensing amplitude $A_{BB}=1.12\pm0.61\,({\rm stat})^{+0.04}_{-0.12}\,({\rm sys})\pm0.07\,({\rm multi})$, with the $C_ ell^{BB}$ band powers consistent with the $\Lambda$CDM expectation and a 97.1% confidence to reject zero B-modes from lensing alone. By combining with non-Gaussian lensing measurements from the same data, the study achieves a total lensing-detection significance of $4.7\sigma$, illustrating robust control of systematics and establishing CMB lensing B-modes as a precise probe of structure formation and inflationary physics.

Abstract

We report a measurement of the B-mode polarization power spectrum in the cosmic microwave background (CMB) using the POLARBEAR experiment in Chile. The faint B-mode polarization signature carries information about the Universe's entire history of gravitational structure formation, and the cosmic inflation that may have occurred in the very early Universe. Our measurement covers the angular multipole range 500 < l < 2100 and is based on observations of an effective sky area of 25 square degrees with 3.5 arcmin resolution at 150 GHz. On these angular scales, gravitational lensing of the CMB by intervening structure in the Universe is expected to be the dominant source of B-mode polarization. Including both systematic and statistical uncertainties, the hypothesis of no B-mode polarization power from gravitational lensing is rejected at 97.1% confidence. The band powers are consistent with the standard cosmological model. Fitting a single lensing amplitude parameter A_BB to the measured band powers, A_BB = 1.12 +/- 0.61 (stat) +0.04/-0.12 (sys) +/- 0.07 (multi), where A_BB = 1 is the fiducial WMAP-9 LCDM value. In this expression, "stat" refers to the statistical uncertainty, "sys" to the systematic uncertainty associated with possible biases from the instrument and astrophysical foregrounds, and "multi" to the calibration uncertainties that have a multiplicative effect on the measured amplitude A_BB.

Figures (12)

• Figure 1: The three Polarbear patches overlaid on a full-sky 857 GHz intensity map 2013arXiv1303.5062P. Patches were chosen for low dust emission, overlap with other observations, and to allow nearly continuous CMB observations from the James Ax Observatory in Chile.
• Figure 2: Top: Beam profiles measured from Jupiter (red) and from fitting point sources with a Gaussian-smoothed Jupiter beam (black). Bottom: Beam uncertainties given as 1-$\sigma$ uncertainties normalized by the beam profile.
• Figure 3: Full season co-added polarization intensity map of Tau A observed by Polarbear. The orientations of bars in map pixels represent polarization angles at each map pixel.
• Figure 4: First season Polarbear power spectra used for calibration and cross-checks of the calibration. Black dots show the measured band powers, with horizontal bars representing the bin widths, and vertical bars representing the uncertainty due to noise, sample variance, and beam uncertainty, the diagonal of the band power covariance matrix. The red curve is the wmap-9 $\Lambda$CDM theory, and the red crosses are the expected binned band powers. The data are consistent with $\Lambda$CDM, as described in Section \ref{['sec:self_calibration']}.
• Figure 5: (a) The $C_\ell^{EB}$ band powers measured using the Tau A-derived instrument polarization angle calibration, combining data from all three patches. A theoretical $C_{\ell '}^{EB}$ spectrum, expected if the instrument polarization angle calibration is incorrect, is fit to the data (best fit curve shown in red, with binned band powers as red crosses). The measured $C_\ell^{EB}$ band powers fit the model that the true instrument polarization angle is -1.08$^\circ$ from the Tau A-derived instrument polarization angle, with a PTE of 55%. (b) $C_\ell^{EB}$ spectrum, after self-calibrating using the $C_\ell^{EB}$ data above.