RIOJA. Environmental Effects on Stellar Populations and Ionized Gas in a Protocluster at $z=7.88$

Abstract: Protoclusters in the epoch of reionization provide key laboratories for investigating how environment shapes early galaxy formation and evolution, and may also have contributed to cosmic reionization. We analyze 23 member galaxies of A2744-z7p9OD, a protocluster at $z=7.88$, using JWST/NIRCam and NIRSpec to investigate their stellar population properties, rest-frame UV sizes, and ionized-gas properties. We also quantify the internal structure of A2744-z7p9OD using the projected distance to the most massive galaxy ($D_{\rm YD4}$), and to the nearest neighbor ($D_{\rm nei}$), as global and local environmental indicators, respectively. Stellar mass, SFR on a 100 Myr timescale, dust attenuation, and galaxy size show significant correlations ($p<0.05$) with $D_{\rm YD4}$, but not with $D_{\rm nei}$, suggesting that these properties are primarily linked to the global protocluster structure. The member galaxies also show a large galaxy-to-galaxy variation in R23 ($=\log{(([\mathrm{O}\text{\textsc{iii}}]λ\lambda4960,5008\rmÅ+[\mathrm{O}\text{\textsc{ii}}]λ\lambda3727,3730\rmÅ)/\rm{H}β)}$), implying inhomogeneous chemical enrichment in the protocluster environment. O32 ($=\log{([\mathrm{O}\text{\textsc{iii}}]\lambda5008\rmÅ/[\mathrm{O}\text{\textsc{ii}}]λ\lambda3727,3730\rmÅ)}$) correlates with $D_{\rm YD4}$, indicating that the core region is characterized by low-ionization gas. Together with the non-detection of Ly$α$ emission, the possible neutral-gas reservoir traced by ALMA [C{\sc ii}]~$158μ$m emission, and evidence for high-column-density neutral hydrogen in the core, this suggests a neutral-gas-rich protocluster core where the current escape of ionizing photons may be suppressed, even in a overdense environment during the EoR.

• The paper shows that stellar mass, long-term star formation, dust attenuation, and UV size correlate with distance from the protocluster center, but not nearest-neighbor distance, indicating global environmental effects dominate.
• The analysis finds substantial chemical diversity, with ΔR23 ≥ 0.7, and systematically low core O32 values of about 1.0–1.1, suggesting neutral-gas-rich conditions and a multiphase interstellar medium.
• The paper argues that the protocluster core may currently release fewer Lyα and ionizing photons because of high neutral-gas covering fractions, while emphasizing that the proposed mechanism requires larger samples and spatially resolved observations.

Motivation and sample

This paper presents a homogeneous analysis of the stellar populations, rest-frame ultraviolet (UV) sizes, and ionized-gas properties of 23 member galaxies of A2744-z7p9OD, a protocluster at $z=7.88$ behind the lensing cluster Abell 2744, with a galaxy overdensity estimated at $\delta \sim 60$--$70$. The sample comprises 16 spectroscopically confirmed members (from NIRSpec IFS, NIRSpec MOS, and NIRCam/WFSS across several programs, including RIOJA GO-1840 and GTO-4553) and seven photometric member candidates, all within the central 6 cMpc. Gravitational magnification factors of $\mu \sim 1.5$--$2.8$ are derived from the UNCOVER DR4 strong-lensing model. The central question is whether galaxy properties at $z \sim 8$ correlate with the global protocluster structure or with local environments, a distinction the authors operationalize through two projected, source-plane structural parameters: $D_{\rm YD4}$, the distance to YD4 (the most massive member, adopted as a proxy for the protocluster center), and $D_{\rm nei}$, the distance to the nearest neighboring member.

The choice of YD4 as the central reference is justified by several independent arguments: "the Quintet," the compact core association hosting YD4, has a total stellar mass of $\log(M_\ast/M_\odot) \sim 9.8$, roughly four times that of "the Chain" ($\sim 9.2$); the stellar-mass centroid of the full sample lies only $\delta \sim 60$0 pkpc from YD4; and ALMA [C II] $\delta \sim 60$1m mapping indicates that the gas and dust distribution peaks in the Quintet region.

Stellar populations, sizes, and star-formation surface densities

SED fitting was performed with BAGPIPES using BPASS v2.2.1 stellar populations, a Kroupa IMF, Calzetti attenuation, CLOUDY nebular emission, and the continuity non-parametric SFH of Leja et al. (2019), with redshifts fixed to spectroscopic values where available. Stellar masses span $\delta \sim 60$2--$\delta \sim 60$3, with YD4 the most massive galaxy. Recent star-formation histories are diverse: the ratio $\delta \sim 60$4 ranges from $\delta \sim 60$5 to $\delta \sim 60$6. The member-galaxy main sequence is broadly consistent with the $\delta \sim 60$7 relation of Simmonds et al., but several low-mass members lie above it, while YD7E and ZD12E show suppressed recent star formation ($\delta \sim 60$8 and $\delta \sim 60$9, respectively). Dust attenuation spans $70$0--$70$1 mag, with the ALMA dust-continuum-detected core galaxies (YD1, YD4, YD7W) showing the highest values, consistent with their red UV slopes.

Rest-frame UV sizes were measured with GALFIT on the F200W image ($70$2 Å) using single-component Sérsic fits, with all galaxies at least marginally resolved. Effective radii span $70$3 and follow a mass–size relation consistent with field galaxies at similar redshifts. Star-formation rate surface densities, computed as $70$4, span $70$5; a few members fall below the $70$6 relation of Calabrò et al., but the sample shows no significant systematic offset.

Ionized-gas diagnostics

Emission-line fluxes ([O II]$70$73727, 3730, H$70$8, [O III]$70$94960, 5008) were measured from G395H IFS spectra ($\mu \sim 1.5$0) and MOS/PRISM spectra, with dust corrections applied using SED-derived stellar $\mu \sim 1.5$1 scaled by $\mu \sim 1.5$2 (fiducial; the authors verify that $\mu \sim 1.5$3 or 1 changes the conclusions only within uncertainties, shifting O32 by at most $\mu \sim 1.5$4 dex). Two results stand out.

First, R23 varies by $\mu \sim 1.5$5 across members, exceeding measurement uncertainties; under the Nakajima et al. (2022) calibration this corresponds to $\mu \sim 1.5$6, indicating strongly inhomogeneous chemical enrichment among member galaxies already during the epoch of reionization (EoR), mirroring the variation reported in the $\mu \sim 1.5$7 protocluster SPT0311-58. The authors caution, however, that H$\mu \sim 1.5$8 is undetected for several low-O32 galaxies, so R23 is only constrained as a lower limit and metallicity remains only qualitatively probed.

Second, and more consequentially, the member galaxies show systematically low O32 relative to field galaxies at $\mu \sim 1.5$9--$2.8$0 (typical O32 $2.8$1--$2.8$2). The lowest values, O32 $2.8$3 (YD4) and $2.8$4 (YD7W), are comparable to the $2.8$5 MOSDEF median. The core of A2744-z7p9OD therefore hosts gas in an ionization state as low as that of galaxies nearly six billion years later in cosmic time.

Global versus local environment

Correlating galaxy properties with the two structural parameters yields a clear and internally consistent result. Stellar mass ($2.8$6, $2.8$7), $2.8$8 ($2.8$9, $z \sim 8$0), $z \sim 8$1 ($z \sim 8$2, $z \sim 8$3), and $z \sim 8$4 ($z \sim 8$5, $z \sim 8$6) all show significant negative correlations with $z \sim 8$7, while $z \sim 8$8 shows a marginal positive correlation ($z \sim 8$9). None of these quantities correlates significantly with $D_{\rm YD4}$0. For the line ratios, a censored Kendall's $D_{\rm YD4}$1 test gives $D_{\rm YD4}$2 ($D_{\rm YD4}$3) for the O32–$D_{\rm YD4}$4 correlation, again with no significant dependence on $D_{\rm YD4}$5.

The implication the authors draw is that stellar mass assembly, long-timescale star formation, dust enrichment, and size growth in A2744-z7p9OD are tied primarily to the global protocluster structure rather than to local pair-scale interactions — an inside-out growth picture consistent with the framework of Chiang et al. (2017), and consistent with the independent finding of Witten et al. (2025) that the core is more evolved than the outskirts. Notably, this result contradicts the trend reported by Toshikawa et al. (2025) for a protocluster at $D_{\rm YD4}$6, where galaxy properties correlated more strongly with nearest-neighbor distance than with distance from the center. The authors attribute this difference to the compact, core-dominated nature of A2744-z7p9OD and its earlier evolutionary stage, while acknowledging that the comparison is complicated by the $D_{\rm YD4}$7 smaller comoving area covered here, and that the result could change if currently undetected dusty or quiescent members were added to the sample. They also stress that the low-sSFR core galaxies should not be interpreted as fully quenched; they may be in temporary lulls with the potential to reignite.

The origin of the low-ionization core

The paper systematically tests whether the low O32 values in the core can be attributed to any single galaxy property, and finds that none suffices:

• Stellar mass: the members lie below the $D_{\rm YD4}$8–O32 relation fitted to JWST-observed galaxies at $D_{\rm YD4}$9; three of five SPT0311-58 core galaxies do as well, so mass alone does not explain the offset.
• sSFR: YD4 and YD7W have $D_{\rm nei}$0 and $D_{\rm nei}$1, at or above the epoch-typical value of $D_{\rm nei}$2, ruling out suppressed ionizing-photon production as the cause.
• $D_{\rm nei}$3: galaxies with moderate O32 follow the Reddy et al. (2023) relation, but YD4 and YD7W fall below it, implying additional physics; the authors note the comparison relation is calibrated at $D_{\rm nei}$4 and its redshift evolution is uncertain.
• Electron density: [O II] doublet ratios give $D_{\rm nei}$5 (YD1), $D_{\rm nei}$6 (YD4), and $D_{\rm nei}$7 (YD7W), spanning a wide range around the epoch-typical $D_{\rm nei}$8; density alone cannot uniformly explain the low O32.
• Metallicity: plausible but unconstrained, given the R23 lower limits and possible elevated metallicities in YD1/YD4 reported by Venturi et al. (2024).

The remaining explanation favored by the authors is structural: extended [C II] $D_{\rm nei}$9m emission around the Quintet implies a substantial neutral-gas reservoir. If compact H II regions are embedded in optically thick neutral gas with high H I covering fraction, ionization-bounded conditions would boost [O II] relative to [O III] through contributions from H II–H I transition layers and diffuse ionized gas, lowering the integrated O32. The authors are explicit that [C II] and [O III] are not expected to be co-spatial and that the current data do not directly demonstrate this mechanism; it remains a consistent interpretation rather than a demonstrated cause. The broader conclusion is that protocluster cores at $\log(M_\ast/M_\odot) \sim 9.8$0--$\log(M_\ast/M_\odot) \sim 9.8$1 host a distinct low-ionization galaxy population whose ISM conditions are not captured by existing field-galaxy scaling relations.

Implications for cosmic reionization

Combining the non-detection of Ly$\log(M_\ast/M_\odot) \sim 9.8$2 from the core (Ly$\log(M_\ast/M_\odot) \sim 9.8$3 is reported only from ZD4, outside the core), the low O32 values, the [C II]-traced neutral reservoir, and evidence for damped Ly$\log(M_\ast/M_\odot) \sim 9.8$4 absorption from high-column-density H I in core members, the authors argue that the current escape of Ly$\log(M_\ast/M_\odot) \sim 9.8$5 and ionizing photons from the core is likely suppressed. They are careful to state that this does not contradict the inside-out reionization picture nor rule out a past contribution by A2744-z7p9OD, since the ionization state of the surrounding IGM reflects cumulative ionizing output; the claim is specifically that the gas-rich core is not an efficient present-day source of escaping ionizing photons.

Limitations and open questions

Several caveats bear directly on the strength of the conclusions. The environmental analysis relies on 2D projected distances because line-of-sight positions cannot be uniquely reconstructed from redshifts contaminated by peculiar velocities. The sample of 23 members, while the largest for any $\log(M_\ast/M_\odot) \sim 9.8$6 protocluster, is small for correlation statistics, and the member census is likely incomplete for dusty or faint sources — an incompleteness that the authors note could alter the $\log(M_\ast/M_\odot) \sim 9.8$7 versus $\log(M_\ast/M_\odot) \sim 9.8$8 comparison. The nebular dust correction depends on the uncertain stellar-to-nebular attenuation ratio $\log(M_\ast/M_\odot) \sim 9.8$9, adopted from a $\sim 9.2$0 empirical relation; although the fiducial results are robust to the choice of $\sim 9.2$1, individual line ratios carry this systematic. Metallicity constraints are qualitative because H$\sim 9.2$2 is undetected in the key low-O32 galaxies, and the [C II]-based neutral-gas interpretation lacks direct spatial correspondence with the optical line-emitting gas. Whether low-ionization cores are a generic property of EoR protoclusters, and what physical mechanism produces them, remain open questions that this single system cannot answer; the authors point to larger protocluster samples from wide-field surveys with deep JWST spectroscopic follow-up as the required test.

Conclusion

This work provides the most detailed environmental analysis to date of a spectroscopically characterized protocluster in the EoR. Its principal findings are that (i) stellar mass, long-timescale SFR, dust attenuation, and UV size correlate significantly with distance from the protocluster center but not with nearest-neighbor distance, indicating global-structure-dominated environmental effects; (ii) gas-phase enrichment is strongly inhomogeneous across members ($\sim 9.2$3); (iii) the core hosts systematically low-ionization gas (O32–$\sim 9.2$4: $\sim 9.2$5, $\sim 9.2$6) that no single galaxy property explains, pointing instead to a neutral-gas-rich multiphase ISM; and (iv) the current escape of ionizing photons from the core is plausibly suppressed despite the system's overdensity. A2744-z7p9OD thus emerges as an evolved, dusty, gas-rich, low-ionization protocluster core at $\sim 9.2$7, demonstrating that environmental differentiation of galaxy populations was already established within 650 Myr of the Big Bang.