The galaxy-IGM connection in THESAN: the physics connecting the IGM Lyman-$α$ opacity and galaxy density in the reionization epoch

TL;DR

The paper investigates how Lyα forest opacity along quasar sightlines relates to the surrounding galaxy density during the tail of cosmic reionization. It uses THESAN radiation-hydrodynamical simulations to map the τ_los–n_gal relation and its redshift evolution, and to test how different ionizing source populations and dark matter models affect it. The main finding is a characteristic proximity scale around 15 Mpc/h where opacity is most sensitive to galaxies, with transparent sightlines tracing earlier, outside-in reionization and opaque sightlines tracing local overdense, inside-out regions; neutral islands are not strictly required to produce large optical depths. The results show robustness across physics variations but highlight volume limitations and the need for more sightlines to constrain reionization timing and sources. The work thus links the early galaxy population to the IGM during reionization and informs future observational campaigns.

Abstract

The relation between the Lyman-$α$ effective optical depth of quasar sightlines ($τ_\mathrm{los}$) and the distribution of galaxies around them is an emerging probe of the connection between the first collapsed structures and the IGM properties at the tail end of cosmic reionization. We employ the THESAN simulations to demonstrate that $τ_\mathrm{los}$ is most sensitive to galaxies at a redshift-dependent distance, reflecting the growth of ionized regions around sources of photons and in agreement with studies of the galaxy--Lyman-$α$ cross correlation. This is $d \sim 15 \, h^{-1} \, \mathrm{Mpc}$ at the tail end of reionization. The flagship THESAN run struggles to reproduce the most opaque sightlines as well as those with large galaxy densities, likely as a consequence of its limited volume. We identify a promising region of parameter space to probe with future observations in order to distinguish both the timing and sources of reionization. We present an investigation of the IGM physical conditions around opaque and transparent spectra, revealing that they probe regions that reionized inside-out and outside-in, respectively, and demonstrate that residual neutral islands at the end of reionization are not required to produce optical depths of $τ_\mathrm{los} > 4$, although they facilitate the task. Finally, we investigate the sensitivity of the aforementioned results to the nature of ionizing sources and dark matter.

Figures (13)

• Figure 1: Top: Galaxy overdensity as a function of distance from each of the 600 lines of sight extracted at $z=5.7$ from the thesan-1 simulation, color-coded by their Ly$\alpha$ optical depth ($\tau_\mathrm{los}$). The black dashed line indicate the average overdensity in the simulation (i.e. 1 by definition) and is used to guide the eye. The inset shows the normalized Ly$\alpha$ flux in the most transparent (blue line) and most opaque (orange line) lines of sight. Middle and bottom: Distribution of galaxies and gas properties around two simulated sightlines. The sightline position is indicated by a red cross, and its direction (perpendicular to the plane of the figure) is the projection axis. Circles indicated the projected galaxy position and are color-coded by the galaxy H$\alpha$ luminosity. The projected density of all galaxies (reconstructed using a Gaussian kernel density estimator) is shown by the contours. The background maps show the average gas density (left), temperature (centre) and @series H I H I H I fraction (right) along the projection axis. We also plot circles with radius increasing by steps of 5 Mpc, up to 30 Mpc. The middle row shows the most transparent sightline in the simulation ($\tau_\mathrm{los} = 1.6$), while the bottom one refers to the most opaque line of sight ($\tau_\mathrm{eff} = 6.0$).
• Figure 2: Pearson correlation coefficient between the number of galaxies in a cylindrical annulus around the LOS and the total Lyman-$\alpha$ optical depth in the LOS, as function of the annulus radius. The Figure shows that the LOS Ly$\alpha$ optical depth is maximally sensitive to galaxies at distance $\approx 15 \, h^{-1}\,{\rm Mpc}$ from it.
• Figure 3: Two-dimensional distribution of large-to-small-scale-overdensity ratio ($\delta_\mathrm{gal}^\mathrm{far} / \delta_\mathrm{gal}^\mathrm{close}$) and optical depth ($\tau_\mathrm{los}$) for each of the investigated lines of sight. The color reflects the number density of sightlines estimated using a Gaussian kernel density estimator, while crosses show individual sightlines in regions where the estimated density is below 5% of the maximum.
• Figure 4: As Fig. \ref{['fig:ngt_corrcoeff']} but only showing the curve including all sightlines and for different redshifts (i.e. different values of the volume-averaged hydrogen neutral fraction $x_\mathrm{HI}$).
• Figure 5: Distribution of galaxy overdensity $\Sigma_\mathrm{gal} (d) / \langle \Sigma_\mathrm{gal} \rangle$ at different radii $d$ in thesan-1, compared to the observed values by Ishimoto+2022 and Christenson+2023 (slightly offset in the horizontal direction for visual clarity). The galaxy overdensity is computed as in observations, after selecting galaxies to have the same number density as observed (see text for details on this procedure) and using the same bins in radial distance. The red left-side violins are computed using all sightlines in the simulation, while the yellow and purple right-side violins are computed using only, respectively, the 100 most transparent and most opaque lines of sight.