E-XQR-30: Evidence for an Increase in the Ionization State of Metal Absorbers from z~6 to z~2

TL;DR

This study uses a large, deep absorber catalog (E-XQR-30) to quantify how the ionization state of metal-enriched circumgalactic gas changes from the tail end of reionization ($z\sim6$) to cosmic noon ($z\sim2$). By classifying absorbers as low-, high-, or mixed-ionization and analyzing column-density ratios such as $N_{CII}/N_{CIV}$ and $N_{SiII}/N_{SiIV}$, it shows a strong shift toward higher ionization with time, including a $\sim$20× decline in $N_{CII}/N_{CIV}$ for mixed systems. Cloudy photoionization models demonstrate that this evolution can be driven by a combination of rising metallicity in CIV-bearing gas and decreasing $N_{HI}$ as the UV background strengthens and the universe becomes more ionized, consistent with the end of reionization around $z\sim5.4$. The results offer robust constraints on the timing and nature of reionization, the chemical enrichment of the CGM, and the evolving interplay between density, metallicity, and the ionizing radiation field in the early universe.

Abstract

We investigate the evolution of the ionization state of metal-enriched gas in and around galaxies near the epoch of reionization using a sample of 488 metal absorption systems at 4.3<z<6.3 from the E-XQR-30 survey. We classify the absorption systems based on whether they display only low-ionization absorption (CII, SiII, MgII), only high-ionization absorption (CIV, SiIV), or both. The percentage of low-ionization-only systems decreases from 24% at $z\sim$6 to 2% at $z\sim$4.3, whilst the fraction of high-ionization-only systems increases from 52% to 82%. For mixed absorbers (with both low and high ionization absorption), we use the column density ratios log(N_CII/N_CIV) and log(N_SiII/N_SiIV) to quantify the average ionization as a function of redshift. The log(N_SiII/N_SiIV) ratio does not change significantly over 5$\lesssim z \lesssim$6.3. We combine the E-XQR-30 log(N_CII/N_CIV) measurements with literature measurements at $z\sim$2-4 and find that the log(N_CII/N_CIV) ratio declines by a factor of $\sim$20 between $z\sim$6 and $z\sim$2. To explore possible drivers of this evolution, we run photoionization models of gas slabs illuminated by a uniform UV background at fixed density, metallicity and HI column density. We find that the increase in the ionization state of metal absorbers towards lower redshifts can likely be explained by some combination of 1) an increase in the metallicity of CIV-absorbing gas and 2) a decrease in the typical HI column densities of the absorbing gas, driven by the declining cosmic mean density and a rapid rise in the strength of the UV background during the final stages of reionization.

Figures (7)

• Figure 1: The column densities of C ii and Mg ii for 29 absorption systems at 5.3 $< z <$ 6.4 from the E-XQR-30 metal absorber catalogue. Measurements for systems with saturated absorption or evidence of partial covering are plotted as open markers. The orange line represents the linear fit to the E-XQR-30 data and the green line indicates the fit from Cooper_2019. The bottom panel shows the residual scatter of the C ii column densities from the linear regression model.
• Figure 2: Evolution of log(NCII/NCIV) (left) and log(NSiII/NSiIV) (right) as a function of redshift with upper and lower limit values denoted by downward and upward arrows respectively. Our E-XQR-30 log(NCII/NCIV) and log(NSiII/NSiIV) measurements are shown in black, and measurements where Mg ii is used as a proxy for C ii are shown in orange. The Cooper_2019 measurements (C19) are shown in grey and the Boksenberg_2015 (B15) measurements are shown in blue. The dashed horizontal lines mark the location for equal quantities of singly and triply ionized metals.
• Figure 3: The fraction of absorbers detected in carbon/magnesium (left) and silicon (right) with low-ionization, high-ionization or mixed absorption, split into redshift bins 4.34 $<z<$ 5.17 (carbon only), 5.17 $<z<$ 5.7 and 5.7 $<z<$ 6.5. The color-coding of the samples is the same as in Figure \ref{['fig:carbon_silicon_main_plot']}. Horizontal error bars show the span of redshifts occupied by absorbers and vertical error bars show the 1$\sigma$ error on the fraction measurements. Some points have been offset by up to 0.02 on either axis for clarity.
• Figure 4: Measurements of the log(NCII/NCIV) ratio for mixed absorption systems with column densities above the 50% completeness limit of the E-XQR-30 catalog (log(NCII) $>$ 13.0, log(NCIV) $>$ 13.2 (background)). Green solid markers show medians and associated error (1.253$\sigma$/$\sqrt{N}$) in bins of redshift. The linear fit to these medians reveals a clear decline in log(NCII/NCIV) towards lower redshift.
• Figure 5: Top: log(NCII/NCIV) correlates strongly with log(NCII) (left) but shows little dependence on log(NCIV) (right). Bottom: accounting for the correlations with log(NCII (left) or log(NCIV (right) reduces the scatter in log(NCII/NCIV) at fixed redshift but does not significantly impact the observed redshift evolution. Plot symbols are the same as in Figure \ref{['fig:trendline_fig']}.