Constraints on reionisation from the z=7.5 QSO ULASJ1342+0928

TL;DR

This paper analyzes the spectrum of a high-redshift quasar to constrain the epoch of reionisation, using a Bayesian approach that accounts for uncertainties in quasar emission and reionisation morphology, finding weak evidence for residual neutral hydrogen.

Contribution

It introduces a novel Bayesian framework focusing on the Lyα damping wing redward of Lyα to constrain the neutral hydrogen fraction during reionisation, differing from previous methods.

Findings

Weak evidence for damping wing absorption at z=7.5

Neutral hydrogen fraction estimated at ~0.21 with uncertainties

Results support a moderately extended epoch of reionisation

Abstract

The recent detection of ULASJ1342+0928, a bright QSO at $z=7.54$, provides a powerful probe of the ionisation state of the intervening intergalactic medium, potentially allowing us to set strong constraints on the epoch of reionisation (EoR). Here we quantify the presence of Ly$\alpha$ damping wing absorption from the EoR in the spectrum of ULASJ1342+0928. Our Bayesian framework simultaneously accounts for uncertainties on: (i) the intrinsic QSO emission (obtained from reconstructing the Ly$\alpha$ profile from a covariance matrix of emission lines) and (ii) the distribution of HII regions during reionisation (obtained from three different 1.6$^3$ Gpc$^3$ simulations spanning the range of plausible EoR morphologies). Our analysis is complementary to that in the discovery paper (Ba\~nados et al.) and the accompanying method paper (Davies et al.) as it focuses solely on the damping wing imprint redward of Ly$\alpha$ ($1218 < \lambda < 1230$\AA), and uses a different methodology for (i) and (ii). We recover weak evidence for damping wing absorption. Our intermediate EoR model yields a volume-weighted neutral hydrogen fraction at $z=7.5$ of $\bar{x}_{\rm HI} = 0.21\substack{+0.17 \\ -0.19}$ (68 per cent). The constraints depend weakly on the EoR morphology. Our limits are lower than those presented previously, though they are consistent at ~1-1.5$\sigma$. We attribute this difference to: (i) a lower amplitude intrinsic Ly$\alpha$ profile obtained from our reconstruction pipeline, driven by correlations with other high-ionisation lines in the spectrum which are relatively weak; and (ii) only considering transmission redward of Ly$\alpha$ when computing the likelihood, which reduces the available constraining power but makes the results less model-dependent. Our results are consistent with previous estimates of the EoR history, and support the picture of a moderately extended EoR.