Constraints on Reionization from Planck Low-$\ell$ EE Polarization and the Lyman-Series Forests

Abstract

CMB measurements tightly constrain the Thomson scattering optical depth ($τ_{\rm es}$), but they still permit a broad range of degenerate reionization histories. Incorporating Lyman-series forest constraints on the end of reionization can alleviate the degeneracy and narrow the range of allowed histories. We explore this effect using forest measurements combined with full Planck likelihoods. We apply upper limits on the neutral fraction at $4.9\lesssim z\lesssim6.9$ derived from the dark pixel fraction, as well as a lower limit at $z=5.8$ from damping-wing-like features on the edges of long dark gaps in stacked forest spectra. We find that neither the choice of parametric vs. non-parametric reionization model, nor the choice among different forest data combinations, produces a significant shift in $τ_{\rm es}$ or the other cosmological parameters. Instead, forest constraints primarily redistribute $τ_{\rm es}$ in redshift. Combining dark pixel, damping-wing-like, and CMB measurements, we find $z_{50}=7.7^{+0.7}_{-0.8}$ and a 95\% upper limit of $Δz_{90}<11.4$ for the midpoint and duration of reionization, respectively. The asymmetry of the reionization history ($A_z$) is not well constrained by these data. Stronger constraints can be obtained using Ly$α$ forest measurements of the mean IGM photoionization rate, at the cost of requiring more detailed modeling. Using a new semi-analytic reionization model, we obtain (model-dependent) constraints of $z_{50} = 7.4^{+0.2}_{-0.3}$, $Δz_{90}=6.0^{+0.8}_{-1.0}$, and $A_z = 2.4^{+0.3}_{-0.7}$. We also find evidence that the ionizing emissivity of reionization sources declines after $z=7-8$, perhaps signaling a decline in the escape fraction and/or the onset of a feedback process. Our results motivate the development of forest likelihood products and fast forward models for its transmission statistics.