Planck intermediate results. XLVI. Reduction of large-scale systematic effects in HFI polarization maps and estimation of the reionization optical depth

TL;DR

The paper advances large-scale Planck HFI polarization analysis by implementing SRoll, a polarized destriper that jointly solves for absolute calibration, temperature-to-polarization leakage, and various instrumental systematics via ring-based templates, aided by comprehensive end-to-end HFPS simulations. This yields near-noise-limited maps at 100–217 GHz and enables a robust low-$\ell$ $EE$-based measurement of the reionization optical depth $\tau$, with $\tau = 0.055 \pm 0.009$ from $100\times143$ cross-spectra and cross-instrument consistency with LFI. The study thoroughly tests systematic effects through null tests, transfer-function modeling, and foreground separation (dust and synchrotron), and demonstrates that residuals are small enough to provide a credible $\tau$ estimate with substantial implications for reionization history and cosmology. By combining HFI low-$\ell$ polarization with high-$\ell$ Planck data and lensing, the results constrain $\tau$ while preserving consistency with the base $\Lambda$CDM framework and reducing tensions in other parameters. The work also validates cross-instrument consistency (HFI and LFI) and highlights remaining challenges (e.g., ADC nonlinearity dipole distortion) to be addressed in future releases.

Abstract

This paper describes the identification, modelling, and removal of previously unexplained systematic effects in the polarization data of the Planck High Frequency Instrument (HFI) on large angular scales, including new mapmaking and calibration procedures, new and more complete end-to-end simulations, and a set of robust internal consistency checks on the resulting maps. These maps, at 100, 143, 217, and 353 GHz, are early versions of those that will be released in final form later in 2016. The improvements allow us to determine the cosmic reionization optical depth $τ$ using, for the first time, the low-multipole $EE$ data from HFI, reducing significantly the central value and uncertainty, and hence the upper limit. Two different likelihood procedures are used to constrain $τ$ from two estimators of the CMB $E$- and $B$-mode angular power spectra at 100 and 143 GHz, after debiasing the spectra from a small remaining systematic contamination. These all give fully consistent results. A further consistency test is performed using cross-correlations derived from the Low Frequency Instrument maps of the Planck 2015 data release and the new HFI data. For this purpose, end-to-end analyses of systematic effects from the two instruments are used to demonstrate the near independence of their dominant systematic error residuals. The tightest result comes from the HFI-based $τ$ posterior distribution using the maximum likelihood power spectrum estimator from $EE$ data only, giving a value $0.055\pm 0.009$. In a companion paper these results are discussed in the context of the best-fit Planck $Λ$CDM cosmological model and recent models of reionization.

Figures (27)

• Figure 1: Mean power spectra of the signal-subtracted, time-ordered data from Survey 2 for each polarization-sensitive HFI frequency channel. The spectra are normalized at 0.25Hz. Blue, green, red, and cyan represent 100, 143, 217, and 353GHz, respectively. The vertical dashed line marks the spacecraft spin frequency. The sharp spikes at high frequencies are the so-called 4-K cooler lines. These noise spectra are built before the time transfer function deconvolution.
• Figure 2: Noise cross-power spectra of the 143-GHz bolometers, with the unpolarized spider-web bolometers (SWBs) in red and the polarization-sensitive bolometers (PSBs) in blue. The low-level correlated white noise component of the PSB noise is associated with common glitches below the detection threshold. Auto-spectra are shown in black. The uncorrelated noise is in green.
• Figure 3: Histogram of the noise between 0.018 and 0.062Hz (frequencies at which it is dominated by the uncorrelated $1/f$ noise) for detector 143-1a in blue, together with the best-fit Gaussian distribution in red.
• Figure 4: Auto-power spectra, showing the level of the simulated FSL projected on the maps predicted using the GRASP model. At $\ell<10$, the FSL signal at all frequencies is at least two orders of magnitude smaller than the expected cosmological $EE$ signal.
• Figure 5: Relationship between input and output, for one spare ADC. The plot shows the difference between the measured digitized output signal level and the one with a perfectly linear ADC, as a function of the output level, over a signal range appropriate for the sky signals. Thus, on the one hand, a perfectly linear device would be a horizontal line at zero; in a real device such as shown here, on the other hand, the relationship between input and output is complicated and nonlinear everywhere, especially near the middle of the range around 0ADU.