A short mean free path at $z=6$ favors late and rapid reionization by faint galaxies

TL;DR

This paper uses simulations to show that a short mean free path at redshift 6 indicates a late, rapid reionization driven mainly by faint galaxies, requiring a high photon budget and possibly increased IGM clumpiness.

Contribution

It introduces a model for sub-resolution ionizing sinks and links the short mean free path at z=6 to late, rapid reionization by faint galaxies, refining understanding of reionization timing.

Findings

Reionization likely completed between z=7.1 and z=5.1.

A short mean free path at z=6 suggests faint galaxies drove reionization.

Achieving the observed mean free path may require increased IGM clumpiness.

Abstract

Recent measurements of the ionizing photon mean free path ($\lambda_{912}^{\rm mfp}$) at $5 < z < 6$ suggest that the IGM was rapidly evolving at those times. We use radiative transfer simulations to explore the implications for reionization, with a focus on the short value of $\lambda_{912}^{\rm mfp} = 3.57^{+3.09}_{-2.14}$ cMpc/$h$ at $z=6$. We introduce a model for sub-resolution ionizing photon sinks based on radiative hydrodynamics simulations of small-scale IGM clumping. We argue that the rapid evolution in $\lambda_{912}^{\rm mfp}$ at $z=5-6$, together with constraints on the metagalactic ionizing background, favors a late reionization process in which the neutral fraction evolved rapidly in the latter half. We also argue that the short $\lambda_{912}^{\rm mfp}(z=6)$ points to faint galaxies as the primary drivers of reionizaton. Our preferred model, with $\lambda_{912}^{\rm mfp}(z=6) = 6.5$ Mpc/$h$, has a midpoint of $z= 7.1$ and ends at $z= 5.1$. It requires 3 ionizing photons per H atom to complete reionization and a LyC photon production efficiency of $\log(f^{\rm eff}_{\rm esc} \xi_{\rm ion}/[\mathrm{erg}^{-1} \mathrm{Hz}]) = 24.8$ at $z>6$. Recovering $\lambda_{912}^{\rm mfp}(z=6)$ as low as the measured central value may require an increase in IGM clumpiness beyond predictions from simulations, with a commensurate increase in the photon budget.