MXDFz4.4: A LyC emitter 250Myr after the epoch of reionization and a first test of Ly-alpha morphology as a tracer of LyC escape at high redshift

Abstract

Assessing the contribution of ionizing sources to cosmic reionization is a central goal of extragalactic astrophysics. Understanding and quantifying ionizing escape remains challenging near the epoch of reionization. We present the highest-redshift Lyman continuum (LyC) emitter detected to date, MXDFz4.4 at z=4.442 in the MUSE eXtremely Deep Field, observed only ~0.25Gyr after the end of reionization. A high confidence Ly-alpha line confirms the redshift. LyC flux is detected at 10.3sigma in the F435W filter with a flux of 4.2+/-0.5nJy, corresponding to a flux measurement at 8.0sigma. After correcting for the intrinsic production of LyC photons and the IGM opacity at z=4.44, we derive high escape fractions, ranging from 50 - 100%. We apply established low-redshift tracers of LyC escape and, for the first time at high redshift, promising Ly-alpha morphological tracers such as the halo fraction. SED fitting indicates the presence of a recent burst of star formation; we explore its impact on the production and escape of ionizing photons. Ly-alpha-based tracers of LyC escape reveal a complex scenario in which the recent burst strong influences LyC production and escape, combined with a more evolved stellar population. This interpretation is supported by UV diagnostics, including the star formation rate surface density and sSFR. Our results provide cautious support for the Ly-alpha halo fraction as a LyC escape tracer at high redshift. Considering the burst-driven enhancement in LyC production and escape, we conclude that stochastic star formation in the early Universe likely plays a significant role in the contribution of galaxies to cosmic reionization.

Figures (14)

• Figure 1: Cutouts of MXDFz4.4 in all filters used for SED fitting, as well as F225W, F275W and F336W to show non-detections. The reduction of the HST data is outlined in Sect. \ref{['sect:MXDF']} and the remaining (JWST) filters are from the JADES program Rieke2023JADES_DR1. All filters have a pixel scale of $0\farcs03$. The LyC flux in F435W, at $8.0\sigma$ significance, is clearly visible.
• Figure 2: MUSE spectrum of MXDFz4.4. Cutouts of the UV lines are shown enlarged below the main spectrum. The Ly$\alpha$ line at $6615.9\,\mathrm{\AA}$ is evident from its asymmetric line shape. No other emission lines are detected in the MUSE spectrum.
• Figure 3: The best fit spectra of MXDFz4.4 from CIGALE using the BC03 and BPASS stellar population models Bruzual2003stellar_popsEldridge2017, plotted in blue. The observations are overplotted in red and model fluxes in orange. The fit was performed without F435W and F606W in order to predict the IGM affected bands and derive $L_{1500}/L_{\mathrm{LyC}}$. The redshift was fixed to the value of $z$ = 4.442 as derived from the Ly$\alpha$ emission line.
• Figure 4: Evolution of $L_{1500}/L_{\mathrm{LyC}}$ over the first $10\,\mathrm{Myr}$ of a starburst, based on the models from pySTARBURST99 models (in color) Hawcroft2025pySTARBURST as well as BPASS Elridge2008BPASSEldridge2017 and BC03 Bruzual2003stellar_pops (in black). Values have been convolved with the F435W and F775W transmission curves and are in $F_\nu$ such that they are comparable to the observations. Metallcities range from Milky Way metallicity ($\sim0.14Z_{\odot}$) or XMP - extremely metal poor ($\sim10^{-5}Z_{\odot}$), with a zero metallicity line included for comparison. VMS indicates models including very massive stars, up to $300\,\mathrm{M_{\odot}}$.
• Figure 5: The transmission of the IGM (defined as $e^{\tau_{\mathrm{IGM}}}$) through F435W to the redshift of MXDFz4.4, based on 10,000 mock sightlines modeled using the TAOIST code as described in S\ref{['sect:IGM']}. We show the results including and excluding the CGM of MXDFz4.4.