Signatures of reionization on Lyman alpha emitters

TL;DR

This study uses semi-analytic models of Lyman alpha emitters to investigate reionization history, suggesting the universe was already highly ionized by z=6.56 based on observed luminosity functions and line profiles.

Contribution

It compares early and late reionization scenarios using LAE data, constraining the timing of reionization and the properties of LAEs with detailed modeling.

Findings

The early reionization model fits observed data without redshift evolution of star formation efficiency.

The late reionization model requires an unphysically large drop in star formation rate.

Line profile skewness correlates inversely with neutral hydrogen fraction.

Abstract

We use a semi-analytic model of Lyman alpha emitters (LAEs) to constrain the reionization history. By considering two physically motivated scenarios in which reionization ends either early (ERM, z_i ~ 7) or late (LRM, z_i ~ 6), we fix the global value of the IGM neutral fraction (e.g. chi_{HI}=3 times 10^{-4}, 0.15 at z=6.56 for the ERM and LRM, respectively) leaving only the star formation efficiency and the effective escape fraction of Lya photons as free parameters. The ERM fits the observed LAE luminosity function (LF) at z=5.7 and 6.56 requiring no redshift evolution or mass dependence of the star formation efficiency, and LAE star formation rates (SFR) of 3-103 solar masses/year, contributing approximately 8% of the cosmic SFR density at z=5.7. The LRM requires a physically uncomfortable drop of approximately 4.5 times in the SFR of the emitters from z=6.5 to 5.7. Thus, the data seem to imply that the Universe was already highly ionized at z=6.56. The mass-dependent Lya transmissivity is between 0.36-0.51 (ERM) and less than 0.26 (LRM) at z=6.56. The LF data at z=4.5 imply an extra Lya line damping factor of approximately 0.25 possibly due to dust; the presence of a (clumpy) dust component with E(B-V) ~ 0.28 is also required to reproduce the observed large Lya equivalent widths at the same redshift. Additional useful information can be extracted from the line profile (weighted) skewness, found to be S_W=10-17 Angstrom for the two reionization models, which shows an interesting L_alpha-chi_{HI} anti-correlation, holding under the model assumptions. The shortcomings of the model and strategies to overcome them are discussed.