Cross-correlating the patchy screening and kinetic Sunyaev-Zel'dovich effects as a new probe of reionization

TL;DR

This work introduces a new CMB-based probe of reionization by cross-correlating a reconstructed patchy screening map $\tau$ with the squared kSZ field $K$, i.e., the cross-spectrum $C_L^{K\tau}$. Using AMBER EoR simulations and Agora post-EoR maps, the authors show that $C_L^{K\tau}$ is non-zero and positive, sensitive to the first half of reionization and to the ionized-gas bispectrum, and they forecast detectability for upcoming surveys. Forecasts indicate $C_L^{K\tau}$ could reach $\sim 1.8\sigma$ for CMB-S4 (without foregrounds) and $\sim 12$–$14\sigma$ for CMB-HD, with foregrounds and lensing biases mitigated via polarization-based $\tau$ reconstruction and bias-hardening techniques. Overall, $C_L^{K\tau}$ provides a high-signal, non-Gaussian probe of reionization that complements traditional auto- and cross-correlations, potentially constraining the midpoint, duration, and skewness of reionization with next-generation CMB data.

Abstract

The kinetic Sunyaev-Zel'dovich effect (kSZ) and patchy screening effect are two complementary cosmic microwave background (CMB) probes of the reionization era. The kSZ effect is a relatively strong signal, but is difficult to disentangle from other sources of temperature anisotropy, whereas patchy screening is weaker but can be reconstructed using the cleaner polarization channel. Here, we explore the potential of using upcoming CMB surveys to correlate a reconstructed map of patchy screening with (the square of) the kSZ map, and what a detection of this cross-correlation would mean for reionization science. To do this, we use simulations and theory to quantify the contributions to this signal from different redshifts. We then use the expected survey properties for CMB-S4 and CMB-HD to make detection forecasts. We find that, for or our fiducial reionization scenario, CMB-S4 will obtain a hint of this signal at up to 1.8$σ$, and CMB-HD will detect it at up to 14$σ$. We explore the physical interpretation of the signal and find that it is uniquely sensitive to the first half of reionization and to the bispectrum of the ionized gas distribution.

Figures (9)

• Figure 1: A demonstration of the map-level effects of patchy screening and the kSZ effect. The left panel shows the primary, unmodified CMB. Panel A (center) demonstrates the screening that happens to the primary CMB when a bubble of electrons is simulated in the foreground, located inside the dashed green circle. Note that the CMB fluctuations, both temperature (colors) and polarization (black vectors), are "washed out" by the Thomson scattering. Panel B (right) demonstrates the shift in the CMB temperature fluctuations that happens when the bubble of electrons is given a positive LOS velocity with respect to the Hubble flow. This leads to a colder (redder) CMB spot in temperature; the polarization is unaffected at this order in scattering.
• Figure 2: A 3-panel display of $2^{\circ} \times 2^{\circ}$ cutouts of the fiducial AMBER maps used in this work. Left: The reionization kSZ map showing the changes to CMB temperature that the kSZ effect makes. Center: The $\tau$ map showing the projected electron density fluctuations centered about $\bar{\tau} = 0.06$. We have added the low-z value of $\bar{\tau}_{z<5} = 0.029$ to the AMBER map to include the optical depth from redshifts below those simulated by AMBER. Right: In purple, the $\hat{K}(\hat{\mathbf{n}})$ map reconstructed using the kSZ map on the left and overlaid with colored $\tau$ contours from the center map to display the correlations between the two.
• Figure 3: Comparison of the $\tau$, reconstructed $\hat{K}$, and $\hat{K}\times\tau$ signal power spectra from the EoR (AMBER) and low-redshifts (Agora) alongside their effective noise curves at $0 < L < 3000$. The orange signal curves are power spectra from three different AMBER models shown by the different shades in the legend. The black noise curves represent quadratic reconstructions of these fields, assuming white noise levels and beam for CMB-S4 temperature and polarization maps. The "S4+FG" noise curve represents the post-ILC noise forecast for CMB-S4, which includes CMB foreground power. The blue dashed curves are the power spectra of the three Agora simulations (discussed in Section \ref{['sec:agora']}), demonstrating that they contribute roughly the same amount to the total as the EoR.
• Figure 4: The dimensionless cross-correlation coefficient demonstrating the strength of the correlation between the fiducial $\hat{K}$ and $\tau$AMBER fields up to $L = 6000$. This correlation peaks from $500 < L < 1000$ at close to 50%. This is not a perfect correlation because the $\tau$ and $\hat{K}$ fields trace the electron density fields with different weights.
• Figure 5: A demonstration of all the components used to forecast the signal-to-noise ratios of $C_L^{KK}$, $C_L^{\tau\tau}$, and $C_L^{K\tau}$. A.) A plot of the lensed primary CMB temperature power spectrum, the fiducial kSZ power spectrum, and the four temperature noise curves for the different surveys used in this work. B.) The effective kSZ filters, $W_l^\mathrm{kSZ} \equiv \sqrt{C_l^{TT,\mathrm{kSZ}}} / C_l^{TT\mathrm{,t}}$ used to reconstruct $\hat{K}$ in this work, demonstrating the sensitivities of the different survey configurations to the kSZ signal. C.) The filters on the CMB E-mode polarization, $W_l^{\tau_{E}} \equiv C_l^{EE} / C_l^{EE\mathrm{,t}}$ for S4 and HD $\tau$ reconstruction calculations. D.) The filters on the CMB B-mode polarization, $W_l^{\tau_{B}} \equiv 1 / C_l^{BB\mathrm{,t}}$ for S4 and HD $\tau$ reconstruction calculations. Together these curves demonstrate the higher sensitivity that CMB-HD would have compared to CMB-S4. For kSZ science specifically, CMB-HD has much better sensitivity at much smaller scales than CMB-S4, even when foreground power is considered.