A SPectroscopic survey of biased halos In the Reionization Era (ASPIRE): Impact of Galaxies on the Circumgalactic Medium Metal Enrichment at z > 6 Using the JWST and VLT

We characterize the multiphase circumgalactic medium (CGM) and galaxy properties at z = 6.0–6.5 in four quasar fields from the James Webb Space Telescope A SPectroscopic survey of biased halos In the Reionization Era (ASPIRE) program. We use the Very Large Telescope/X-shooter spectra of quasar J0305–3150 to identify one new metal absorber at z = 6.2713 with multiple transitions (O i, Mg ii, Fe ii, and C ii). They are combined with the published absorbing systems in Davies et al. at the same redshift range to form a sample of nine metal absorbers at z = 6.03–6.49. We identify eight galaxies within 1000 km s−1 and 350 kpc around the absorbing gas from the ASPIRE spectroscopic data, with their redshifts secured by [O iii] (λ λ4959, 5007) doublets and Hβ emission lines. Our spectral energy distribution fitting indicates that the absorbing galaxies have stellar masses ranging from 107.2 to 108.8 M ⊙ and metallicity between 0.02 and 0.4 solar. Notably, the z = 6.2713 system in the J0305–3150 field resides in a galaxy overdensity region, which contains two (tentatively) merging galaxies within 350 kpc and seven galaxies within 1 Mpc. We measure the relative abundances of α elements to iron ([α/Fe]) and find that the CGM gas in the most overdense region exhibits a lower [α/Fe] ratio. Our modeling of the galaxy’s chemical abundance favors a top-heavy stellar initial mass function and hints that we may be witnessing the contribution of the first generation of Population III stars to the CGM at the end of the reionization epoch.


Introduction
The gaseous halos in the circumgalactic medium (CGM) and the intergalactic medium (IGM) play a critical role in the baryon cycle, subsequently affecting galaxy evolution (for reviews, see Tumlinson et al. 2017;Péroux & Howk 2020 and references therein).Exploring the connection between CGM/IGM with the Universe's earliest stars, quasars, and galaxies offers insights into the metal enrichment, galaxy assembly, and origins of supermassive black holes during the reionization epoch (EoR; see reviews by Fan et al. 2006Fan et al. , 2023;;Inayoshi et al. 2020 and references therein).
The absorbing systems detected in quasar spectroscopy, which arise from intervening gaseous halos toward bright background quasars, provide a sensitive measure of multiphase gas.Different gas phases are characterized by their density, temperature, and ionization parameters.The multiplicity of metal ions in the absorbing system allows for characterization of the different gas Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Signatures of the first generation of metal-free stars (Population III stars) have been reported in galactic metal-poor stars (see Frebel & Norris 2015 for a review), pristine H I- bearing gas such as damped Lyα systems (DLAs) and Lymanlimit systems (LLSs; Cooke et al. 2011;Zou et al. 2020;Welsh et al. 2022;Christensen et al. 2023;Saccardi et al. 2023), and high-redshift galaxies (e.g., Maiolino et al. 2023).Detections of pristine gas around high-redshift galaxies are promising for our understanding of the formation history of the first objects.Previous detections of gaseous halos and their associated galaxy environments have been largely limited by the capabilities of ground-based telescopes, especially at z > 6.
JWST opens a new era for discovering galaxies-particularly dwarf galaxies at high redshift (e.g., Atek et al. 2022;Endsley & Stark 2022;Cameron et al. 2023;Curtis-Lake et al. 2023;Fujimoto et al. 2023;Matthee et al. 2023;Sun et al. 2023).Combining the data from distant galaxies detected by JWST with the absorbing gas in their CGM can enhance our understanding of chemical enrichment in the CGM/IGM and its connection to the formation of the first objects.
The structure of this paper is as follows.In Section 2, we introduce the A SPectroscopic survey of biased halos In the Reionization Era (ASPIRE) program.Section 3 details the absorbers and the detection of surrounding galaxies.The properties of the absorbing gas and its connection with the galaxies are discussed in Section 4. We use the standard ΛCDM model in this work, and the cosmological parameters used are H 0 = 70 km s −1 Mpc −1 , Ω Λ = 0.69, and Ω m = 0.31.

Sample Selection and Data Reduction
The ASPIRE survey encompasses 25 quasars in the range (6.5 < z < 6.8), each with high-quality spectra spanning from optical to radio observations (e.g., Venemans et al. 2019;Yang et al. 2020Yang et al. , 2021)).These quasar fields are observed using JWST's NIRCam/WFSS.Photometric data are collected in three bandpass filters: F115W, F200W, and F356W.The grism R spectra are acquired in the F356W filter, encompassing the [O III] and Hβ emission lines for redshifts approximately between 5.2 and 7.0.The grism spectra offer a resolving power of 1000-1500 at wavelengths of 3-4 μm.Details about this survey can be found in Wang et al. (2023) and Yang et al. (2023).

JWST/NIRCam Data Reduction
We perform data reduction for the JWST/NIRCam and WFSS data using version 1.10.2 of the JWST Calibration Pipeline (CALWEBB).The official CALWEBB pipeline processes data products in three stages.Wang et al. (2023) provides detailed data reduction steps, and we offer a brief overview here.In the NIRCam direct image processing, the 1/f noise pattern is first eliminated.For data calibration, we employ reference files (jwst 1080.pmap)from version 11.16.21 of the standard Calibration Reference Data System.All imaging data in this study align with the Gaia Data Release 3.For the WFSS data, the CALWEBB stage 1 pipeline calibrates detector-level signals and performs ramp fitting for individual NIRCam WFSS exposures.To mitigate the 1/f noise, we subtract the calibrated data spectra columnwise as the spectra disperse rowwise for grism R. It is worth noting that we achieve background subtraction by creating background models based on all ASPIRE observations taken at similar times; we then scale and subtract these models from individual WFSS exposures.
A critical step in the data reduction process for slitless spectra involves the tracing model.We adopt the method outlined in Sun et al. (2022).This model traces point sources observed in the Large Magellanic Cloud field.The sensitivity functions, as determined from both the Early Release Observations calibration and the Cycle 1 calibration, align with an accuracy surpassing 98% across most of the wavelength range.Ultimately, we extract both the 2D and 1D spectra from the catalog of detected imaging.The depth reaches a 5σ line luminosity limit of 9.9 × 10 41 erg s −1 at z = 6.6.

Very Large Telescope/X-shooter Data Reduction
In this work, we analyze the connection between absorbing gas and galaxies using four quasar sight lines (J0305-3150, J0226+0302, J0224-4711, and J0923+0402) for which we have both Very Large Telescope (VLT)/X-shooter and ASPIRE JWST/WFSS data.The latter three sight lines are part of the XQR-30 sample (D'Odorico et al. 2023).The absorber catalog can be found in Davies et al. (2023a).The quasar J0305-3150 is included in the X-shooter Atacama Large Millimeter/submillimeter Array (ALMA) sample of quasars in the epoch of reionization (Schindler et al. 2020;Farina et al. 2022).We reduce the X-shooter near-infrared spectra using the Python Spectroscopic Data Reduction Pipeline, PypeIt (STSCI Development Team 2012), with the reduction details presented in Schindler et al. (2020).

Detection of Absorbers and Associated [O III] Emitters
We identify metal absorbers from the VLT/X-shooter spectra.Due to the sample size, we manually normalize the reduced X-shooter spectra using a spline function and subsequently conduct a visual inspection of the absorbers.We measure the column densities and Doppler parameters by fitting the absorption lines with a Voigt profile utilizing the VPFIT code (Krogager 2018).We add a 10% uncertainty from the continuum fitting when calculating the column densities of different ions.We also incorporate absorbing systems at z > 6.0 along three XQR-30 sight lines (J0226+0302, J0224-4711, and J0923+0402) as detailed in Davies et al. (2023a).Specifically, we measure the upper limits of Fe II column density, N Fe II , by varying the Doppler parameter b to obtain a reasonable fitting curve.
We then search for the [O III] emitters within a velocity window of ±1000 km s −1 and an impact parameter of 350 kpc from the absorbing gas.To determine the expected number of [O III] emitters within this volume, we refer to the [O III] luminosity function at z ∼ 6.2 from Sun et al. (2022).Considering our detection limit (with an [O III] luminosity threshold of 10 42.3 erg s −1 ), the count is 8 × 10 −5 .The detection of one or more galaxies in this context suggests its potential association with the absorbing gas.We use the algorithm described in Wang et al. (2023) to automatically search for the [O III] emitters having a peak signalto-noise ratio >3.0.We then visually checked all the candidates and confirmed the final [O III] targets.We present the detected absorber and associated galaxy candidates in Figures 1 and 2 and Appendix Figures A1, A2, A3, and A4.The full 2D and 1D spectra of all the [O III] emitters can be found in Appendix Figure A5.Characteristics of the detected absorbers and galaxy candidates are presented in Table 1.
J0305-3150-We detect one absorbing system having O I, C II, Mg II, and Fe II at z = 6.2713.The rest-frame equivalent widths (W r ) of the O I (λ1302), Mg II (λ2796), Mg II (λ2803), and Fe II (λ2344) lines are 0.069 ± 0.022, 0.32 ± 0.12, 0.30 ± 0.14, and 0.17 ± 0.10 Å, respectively.The W r ratio of the Mg II doublet l l W W r r 2803 2796 is 0.94, indicating strong saturation in the Mg II absorption.With a Doppler parameter of b = 10−15 km s −1 , the log N O I ranges from 14.26 to 14.38, falling within the slightly saturated region on the curve of growth (COG).The Fe II is slightly saturated, given that log N Fe II > 13.2.There is no significant (W r > 0.1 Å) detection of Si II and C IV lines.The Si IV lines are significantly contaminated by sky lines.The C II (λ1334) is affected by the sky lines, so we measure its column density upper limit in Table 1.We plot the metal lines and fitting curve with a Voigt profile in Figure 1 (right panel).
We detect one possible [O III] emitter pair at impact parameter D ∼ 298 kpc (proper distance) and velocity offset Δv = 708 km s −1 : ASPIRE-J0305M31-O3-5083 (see Figure 1).This emitter displays a merging feature in both the 2D and 1D spectra; two sources are identified using SExtractor.Notably, seven other [O III] emitters reside within D = 1 Mpc.ASPIRE-J0305M31-O3-0623 and ASPIRE-J0305M31-O3-0905 also exhibit a merging feature.Wang et al. (2023) present two galaxy overdensity fields at z = 5.2 and 6.2.This metal-enriched absorbing gas resides in the overdense region at z ∼ 6.2 (a discussion of the z = 5.2 absorbing system is presented in Wu et al. 2023).
XQR-30 absorbers J0226+0302.We use the absorbing system at z = 6.06111 detected in the QSO sight line J0226+0302 (target name PJ036 +03 in the XQR-30 sample) to search for galaxy counterpart candidates.This absorbing system contains O I, Mg II, and Si II.The W r ratio between the Mg II doublet l l W W r r 2803 2796 is 0.64 ± 0.05, indicating that the Mg II is slightly saturated.The O I, C II, C IV, and Si II lines are not saturated.No Fe II lines are detected.We remeasure the N Fe II and obtain its upper limit in Table 1.From our WFSS data, one [O III] emitter, ASPIRE-J0224M47-O3-4582, is detected at D = 341 kpc and Δv = 155 km s −1 .J0224-4711.We use the absorbing systems at z 6.0 (6.03133, 6.12279, 6.17255, 6.26848, 6.30443, and 6.4820) from Davies et al. (2023a) to search for galaxy counterpart candidates.The O I is detected in the system at z = 6.12279.We refit the metal lines with a Voigt profile and vary the b value to obtain a reasonable profile.The results are presented in Figure A2.With b = 10-15.8km s −1 , log N Fe II varies within 0.04, and log N O I varies between 14.67 ± 0.11 and 14.46 ± 0.06.The Fe lines are likely not saturated, and    contaminated by the sky lines.One galaxy, ASPIRE-J0224M47-O3-6114, is detected at D = 246 kpc and Δv = 41.5 km s −1 .One galaxy, ASPIRE-J0224M47-O3-6315, is detected at D = 347 kpc and Δv = 1047 km s −1 .
For the system at z = 6.03133,only C IV is reported as detected in Davies et al. (2023a).The C II (λ1334) line is also present but likely blended with other lines.We thus give an upper limit on N C II .We remeasured the upper limit for N Si IV and N Fe II .One galaxy, ASPIRE-J0224M47-O3-6114, is detected at D = 246 kpc and Δv = 41.5 km s −1 .
For the system at z = 6.17255,only C IV is detected, and there is no detection for the C II and Fe II lines, so we measure the N C II , N Fe II upper limits for this system.One galaxy, ASPIRE-J0224M47-O3-1674, is detected at D = 341 kpc and Δv = 320.7 km s −1 .
For the system at z = 6.30443,Mg II, C II, and Si II (λ1260) lines are reported as detected in Davies et al. (2023a).
Considering the profile of C IV, we report its N measurement as an upper limit in Table 1.No Fe II lines are detected.The N Mg II is in the slightly saturated region in the COG, and the ratio of For the z = 6.4820 system having only C IV absorption, we do not find any galaxy candidates within 1 Mpc.J0923+0402.This target is in the XQR-30 sample.We select the absorbing systems at z = 6.378 from Davies et al. (2023a) to search for galaxy counterparts.We detect two
For the SED fitting, we incorporate the observed photometry in the F115W, F200W, and F356W bands, as well as the spectroscopically observed fluxes of the [O III] and Hβ lines.For J0305-3150, we also include archived Hubble Space Telescope (HST) photometry from the F105W, F125W, and F160W bands (PropID 15064; PI: Casey).We adopt a "constant" star formation history (SFH) and use the Small Magellanic Cloud UV extinction law.A6).It is worth noting that in BEAGLE, the IMF remains fixed, ensuring that the contributions from different stellar masses and generations are constant.
To quantify the uncertainties in the SED fitting, we randomly selected 100 galaxies from the JAGUAR mock galaxy catalog (Williams et al. 2018) 21 and performed SED modeling on these mock galaxies using the same parameter settings as mentioned above.The [O III] 5007 line flux, redshifts, and stellar mass of these galaxies are consistent with those in our sample.We used their F115W, F200W, and F356W flux densities, assuming a similar photometric uncertainty to our detection depth.Different SFHs, "constant" and "ssp," are adopted in the tests.We compare the best-fit parameters with the intrinsic values in the mock catalog.We find that if a "constant" SFH is adopted, the standard deviation between the fitted and intrinsic stellar properties-stellar mass, age, metallicity, SFR, and M UV -is 0.39 dex, 234 Myr, 0.30 dex, 9.5 M e yr −1 , and 0.17 dex, respectively.These offsets in stellar mass, age, metallicity, and SFR are 0.56 dex, 173 Myr, 0.25 dex, and 11.2 M e yr −1 , respectively, by adopting an "ssp" SFH.We detail the results with added uncertainty on the measurements of the SED fitting for absorbing galaxies in Table 1 and Appendix Figure A6.

Discussion
In this section, we discuss the connection between the CGM gas properties (the CGM metal budget and ionization state) and the characteristics of surrounding galaxies at the end of the EoR.We focus on two aspects: (1) the metal abundances in the CGM gas and their connection with the SFH of proximate galaxies and (2) the influence of nearby galaxies as local ionizing sources on the absorbing gas.

Relative Abundances between Iron and α-elements in the
CGM at z = 6.0-6.5 Metals residing in the CGM of a galaxy are delivered out of the galaxy by feedback such as stellar winds, supernovae (SNe), and/or AGN feedback (Tumlinson et al. 2017).Stellar feedback is likely to dominate the baryon cycle in dwarf galaxies during early cosmic times (Cen & Bryan 2001).Cen & Bryan (2001) posit that dwarf galaxies formed between z = 7 and 15 contribute significantly to the Universe's metals and energy.Assuming a constant dispersing velocity of v disp ∼ 300 km s −1 for a gas halo in a galaxy formed at z = 11.7 (age ∼ 0.36 Gyr), this gas can be pushed out of the host halo (∼10 10 M e ) and travel as far as 1 comoving Mpc within half of the Hubble time (until z ∼ 7).Such a dispersion velocity might arise from a sustained starburst or galaxy merger.Therefore, the relative abundances in the CGM gas phase can provide insights into nucleosynthesis and the SFH of associated dwarf galaxies.Recent findings from the EIGER (Bordoloi et al. 2023) and XQR-30 (Davies et al. 2023b) surveys indeed indicate rapid metal pollution from galaxies to the IGM at high redshift.
The α-elements are products of Type II SNe.The Fe-peak elements primarily arise from Type Ia SNe, which manifest with a delay of about 1 Gyr compared to the α-elements (Maoz et al. 2012).At z > 6, where the Hubble time is less than 1 Gyr, the gas-phase [α/Fe] abundance becomes an interesting indicator of the gaseous halo's lifetime.It also offers insights into the SFH and the process of galaxy assembly.
For the absorbing gas at z > 6.0 in this work, even though we lack H I column measurements due to H I reionization, we can still determine its relative abundances of α-elements over iron using multiple absorption lines.To compute abundances for any two elements (X and Y) based solely on their column densities, we assume The solar abundances are taken from Asplund et al. (2009).For each species, we refer to the dominant ions for its overall column density N X .For example, we have We also estimate the N C II from N Mg II using the empirical relation in Cooper et al. (2019) when C II is not detected or strongly contaminated: log N C II = 0.811 × (log N Mg II -12) + 13.09.
In the left panel of Figure 3, we plot [C/Fe] and [O/Fe] values of the absorbing gas against the stellar age of their associated [O III] emitter candidates.We find that the [α/Fe] values for the CGM in a less dense region (three galaxies within 1 Mpc) are tentatively about 1 dex greater than those in the most overdense region (nine galaxies within 1 Mpc).This difference might hint at the galaxy assembly and SFH during the cosmic dawn.We suggest two potential explanations for this observation.First, galaxy mergers or clumpy structures might reduce the production of α-elements, particularly from Type II SNe.Alternatively, interactions within galaxy groups might suppress the efficiency of expelling lighter elements such as carbon and oxygen compared to Fe-peak elements into the CGM.Second, the overdensity region somehow triggers the Type I SNe in generated Fe-peak elements earlier than expected.

Accuracy in Relative Abundance Measurements
Note that the accuracy of measurements on the gas-phase relative abundances would be affected by the potential line saturation and metal depletion onto the dust.From Appendix A1-A4, we can tell that for all the systems with detected low ions, there is no >3σ Fe II detection except for two systems (z = 6.12279 and 6.2713).In those systems without Fe II detection, the [O/Fe] or [C/Fe] values in Table 1 will  Dust depletion may also affect the gas-phase relative metal abundance measurements.Different metal species can be locked into the dust grains with varying depletion factors.Iron has approximately 1 dex more depletion onto the dust than carbon and oxygen (Jenkins 2009).The [C/Fe] and [O/Fe] will be slightly enhanced due to possible dust depletion of Fe.Nevertheless, at the metallicities assumed, Fe depletion would be very small and not significantly affect the results.Future ancillary ALMA observations and analyses in these quasar fields will be presented in the EREBUS collaboration (2024, in preparation).

Possibility of Top-heavy IMF Contribution to the CGM at the Cosmic Dawn
In this section, we discuss the stellar population in the CGM host galaxy candidates and its effect on the metal abundances in the CGM.We note that several absorbing systems in our sample exhibit similar features as DLAs with the presence of singly ionized O I and Mg I absorption.Kulkarni et al. (2013) performed a self-consistent model on the Population III star contribution in the IGM at the EoR.They found that the chemical signatures of Population III stars remain in low-mass galaxies (halo mass <10 9 M e ) at z ∼ 6, and the relative abundances of the metal-poor DLAs likely hold the promise of constraining Population III enrichment in the early Universe.Welsh et al. (2022) model the contributions from both Population II and Population III stars in metal-poor DLAs.Saccardi et al. (2023) studied the metal abundances in 37 LLSs and sub-DLAs at z ∼ 3-4.5 and found that among the 14 very metal-poor LLSs or sub-DLAs in their sample, the [C/Fe] and [O/Fe] are similar to those in galactic metal-poor stars and ultrafaint dwarf galaxies.
In the right panel of Figure 3, we compare the relative abundances, [O/Fe], of the four systems with clear O I detection in our sample to those detected in metal-poor DLAs to date and two DLAs detected at z ∼ 5.9 (D'Odorico et al. 2018) and z ∼ 6.4 (Bañados et al. 2019) and low-ion absorbers in Cooper et al. (2019).We note that our observed [O/Fe] is similar to that in metal-poor DLAs and LLSs at lower redshift, possibly suggesting a pristine environment to form the early generation of stars.If these systems are DLAs, their metallicities are plausibly close to the average DLA metallicity at z ∼ 6 as predicted by the empirical relation in Rafelski et al. (2012): 〈Z〉 = (−0.22± 0.05) × z − (0.66 ± 0.15), i.e., −1.95 ± 0.45 at z = 6.
We further consider the effects of different IMFs on the metal abundances of dwarf galaxies at the cosmic dawn and thus the metal abundances in the CGM/IGM.IMFs are generally modeled as a power law of index α.We updated the galactic chemical evolution (GCE) model in Côté et al. (2017) and the algorithm NuPyCEE22 by varying the stellar IMF.Details of the updated algorithm and model assumption are presented in Z. Y. Guo et al. (2024, in preparation).We take the galaxy's stellar mass, SFR, and age from Table 1 as the model inputs.The outputs are the stellar yields of different metals.Specifically, we set three IMFs, top-heavy IMF (α = −1.95),Kroupa IMF (α = −2.3),and bottom-heavy IMF (α = −3.0),with stellar mass ranging between 1 and 40 M e .We plot the modeling results in the left panel of Figure 3.We find that for an isolated dwarf galaxy at z ∼ 6, a top-heavy IMF generates three times higher [C/Fe] than and similar [O/Fe] as the Kroupa IMF does when the stellar age is smaller than 200 Myr.If the stellar feedback occurs simultaneously with the star formation and blows the stellar yields into its CGM/IGM, we may detect a similar [C/Fe] overabundance in its CGM/IGM.Our GCE modeling in dwarf galaxies at z ∼ 6 can account for the observed CGM [O/Fe].The [C/Fe] in the CGM favors a top-heavy IMF in the associated galaxies.However, the top-heavy IMF alone in our GCE model cannot fully account for the detected carbon overabundance in the CGM.The top-heavy IMF can only reach C/Fe values of 0.015, which is smaller than our lowest measured value of 0.06.We note that the [C/Fe] yields Population III stars in the model of Kulkarni et al. (2013) that can potentially explain our measured [C/Fe] values.This may indicate that our detected CGM gas could have been polluted by the Population III stars in nearby galaxies.Schaye (2006) pointed out that the local ionizing source has a significant effect on the ionization state of the IGM gas.Finlator et al. (2016) examined the effects of different UV background (UVB) sources on the strengths of ions (C IV, C II, and Si IV) in the CGM/IGM.They found that the UVB from Haardt & Madau (2012) only overpredicts the C II/C IV abundance in highly ionized C IV systems, suggesting a local amplification of UV radiation from galaxies.Their simulations also suggest that the strong C IV-galaxy correlation extends to at least 300 pkpc.
We thus perform a simple test on the ionizing radiation from our detected absorber-associated galaxy candidates and its effect on the CGM gas ionization state, in particular, the systems with only C IV systems detected (log N C IV > 13.0).First, we estimate the ionizing radiation bubble size generated by the [O III] emitters.We calculate the galaxy radiation radius using its Strömgren radius (R s ) from the analytical model described in Bolton & Haehnelt (2007).The Strömgren radius is defined by ) , where t r = 1/(n H α) and α is the recombination rate (2.6 × 10 −13 cm −3 s −1 ).Then the timescale of R s expands to 300 pkpc around 3 Myr, which is reasonable within our galaxy age.
We then conduct photoionization modeling to obtain the physical properties of the absorbing gas using CLOUDY (Ferland et al. 2017).In the models, we include the metagalactic UVB from Khaire & Srianand (2019), the cosmic microwave background at z = z abs , and cosmic rays as the extragalactic UVB.The UV flux from the detected [O III] emitters is estimated from its UV magnitude M UV from our SED fitting.For the neutral systems (N C II /N C IV > 3), we find that the observed N C II /N C IV can be reproduced without adding a local UV flux.For the highly ionized C IV gas (N C II /N C IV < 1), an additional ionizing source is required to reproduce the C IV/C II ratio.We show the fitting result of the z = 6.17255 system toward J0224M4711 in Figure 4.
Direct detections of galaxies around highly ionized C IV systems (N C II /N C IV < 3) at z > 6 are rare; see Díaz et al. (2011Díaz et al. ( , 2015) ) and Meyer et al. (2019) for connections between C IV and Lyα emission.The galaxy around the highly ionized C IV system (log N C IV ∼ 13.4) at 5.9784 in Díaz et al. (2011) has an M UV of −20.66 ± 0.05 and D = 311.4pkpc.The galaxy detected around the z = 5.7242 C IV system (log N C II ∼ 14.52) in Díaz et al. (2015) has M UV = −20.65 ± 0.52 and D = 212.3pkpc.Meyer et al. (2019) statistically calculated the correlation between Lyα emission around C IV systems at 5.7 < z < 6.2 and concluded that C IV absorbers with log N C IV > 13.2 are associated with galaxies having M UV < −16.The system at z = 6.4821, that has only C IV absorption detected and no galaxy within 1 pMpc with the lowest N C IV , may associated with galaxies fainter than M UV = −16.0.In summary, our detection of highly ionized C IV gas [O III]-emitting galaxies and their effects on the local UVB are consistent with cosmological simulations and previous limited detections.

Summary
1. We detected a new absorber at z = 6.2713 from quasar sight line J0305-3150.Along with absorbers from Davies et al. (2023a), we have a sample of nine spanning z = 6.03-6.49.Using ASPIRE JWST/WFSS data, we found 8 (11) [O III] emitters within D = 350 (1 Mpc) of these absorbers, with stellar masses log M * /M e from 7.2 to 8.8 and metallicities Z/Z e from 0.02 to 0.4 solar metallicity.2. We find that the absorbing gas with a higher [α/Fe] is possibly associated with a less overdense region than the most overdense region.3. We examine various IMFs in galaxies at z ∼ 6 with ages less than 500 Myr.Our findings indicate that the topheavy IMF produces a [C/Fe] that is three times higher than the Kroupa IMF (Kroupa 2001) during the initial 200 Myr.The yields from the Population III IMF suggest its potential contribution to the [C/Fe] overabundance in our observed CGM. 4. We perform photoionization modeling for all the systems and find that a local ionizing source is plausible for the highly ionized C IV systems, which is consistent with the cosmological simulation results at z ∼ 6. 5. We note that fainter galaxies below our detection limit may reside closer to our absorbing gas in the CGM; this will not affect our discussion of the CGM metal abundance and its ionization source.

Figure 1 .
Figure 1.The absorbers detected at z = 6.2713 toward quasar J0305-3150 and associated [O III] emitter candidate.The purple, orange, and red curves represent the fitting of absorption lines in this system with a Voigt profile and Doppler parameters of 5, 10, and 15 km s −1 , respectively.The letters U and L represent the upper and lower limits on the column density measurement of the lines.The galaxy candidates detected within 350 kpc and 1000 km s −1 of the absorbing gas are presented in the zoom-in red, green, and blue imaging boxes.

Figure 2 .
Figure2.We include the absorbing systems at z > 6.0 inDavies et al. (2023a) along the quasar sight lines J0226+0302, J0923+0402, and J0224-4711.The galaxies detected within 350 kpc and 1000 km s −1 of the absorbing gas are presented in the zoom-in red, green, and blue imaging boxes.
[O III] +Hβ emitters (ASPIRE-J0923P04-O3-2112 and 1423) at D = 108 and 217 kpc, respectively.This absorbing system has O I, Mg II, C II, and Si II absorption.The O I (λ1302) equivalent width is 0.074 ± 0.004 Å, and the N O I is in the saturated region in the COG.The W r ratio between the Mg II doublet 77 ± 0.05, indicating a partial saturation.No Fe II lines are detected, and the C IV lines are significantly contaminated by the sky lines.
be even higher if the O I, C II, and C IV lines are saturated/partially saturated.The two systems having both O I and Fe II detected show similar N O II .The Fe II absorption detected at z = 6.2713 is more saturated than that at z = 6.12279.The [O/Fe] values are 0.56 ± 0.22 and 0.36 ± 0.17 with fixed b values of 10 and 15.8 km s −1 (the b value in Davies et al. 2023a) for the z = 6.12279 system.The [O/Fe] values are −0.32 ± 0.34 and −0.23 ± 0.32 with fixed b values of 5 and 15 km s −1 for the z = 6.2713 system.In summary, with careful consideration of line saturation, we confirm that our newly detected system in the J0305−3150 sight line has the lowest [O/Fe] value in this work.

Figure 3 .
Figure 3. Left: relative metal abundances between α (C and O) elements with iron (Fe) in the absorbing gas vs. the age of associated galaxy candidates.The solid black, dashed blue, and long dashed purple curves are the Kroupa, bottom-heavy, and top-heavy IMF in the GCE models, respectively.Right: the comparison of the [α/Fe] ratio with that in metal-poor DLAs in Welsh et al. (2022) at z < 5.7 and those detected at z > 5.7.

4. 4 .
Surrounding Galaxies as Local Ionizing Sources of the CGM escape fraction f esc .The f esc of the galaxy at z 6 has a large variety (0.01%-20%;Ma et al. 2015).If an [O III]-emitting galaxy has M 1450 = −20 (luminosity L 1450 ∼ 10 45 erg s −1 ), and if we assume its f esc = 0.1 and adopt the gas density n H = 10 −3 cm −3 from our CLOUDY modeling, its Strömgren radius R s is ∼1 Mpc.The H II expansion rate can be calculated following the relation

Figure 4 .
Figure 4. Photoionization modeling of the absorbing gas in the absorber system at z = 6.17255 toward J0224-4711 using CLOUDY.The left and right panels are the models with and without the UV flux from the detected nearby galaxies, respectively.We can tell that without an external UV flux, the metagalactic UVB only cannot reproduce the observed [C IV/C II] ratio.

Figure A2 .
Figure A2.The fitting of the metal lines with a Voigt profile in the absorbing system at z = 6.12279 and 6.03133 toward quasar J0224-4711.The red, purple, and orange curves are the Voigt fitting profile with different Doppler parameters.The red curve is the b value used in Davies et al. (2023a).The letters U and L represent the upper and lower limits on the column density measurement of the lines.

Figure A3 .
Figure A3.The fitting of the metal lines with a Voigt profile in the absorbing system z = 6.17255 (left) and 6.26848 (right) toward quasar J0224-4711.The O I λ1302 and Si II λ1304 lines are strongly blended with the C IV doublet at z = 5.10911.The Si IV doublet is strongly blended with the sky lines; therefore, we do not plot it in the right panel.The red, purple, and orange curves are the Voigt fitting profile with different Doppler parameters.The red curve is the b value used in Davies et al. (2023a).The letters U and L represent the upper and lower limits on the column density measurement of the lines.

Figure A4 .
Figure A4.The fitting of the metal lines with a Voigt profile in the absorbing system z = 6.30443 toward quasar J0224-4711.Weak Mg II, C II, and Mg II (λ1260) lines are detected.The letters U and L represent the upper and lower limits on the column density measurement of the lines.

Figure A6 .
Figure A6.SED fitting of all detected absorbing gas associated with galaxy candidates.The uncertainty is derived from the BEAGLE fitting results, taking into account the uncertainty when compared with the JAGUAR mock catalog (Williams et al. 2018).

Table 1
Measurements of the Properties of Absorbers and Associated Galaxy Candidates in This Study Davies et al. (2023a)e, we remeasure the upper limits of ion column density using the same b value as inDavies et al. (2023a); these values are highlighted in bold.In the bottom table, the stellar mass, SFR, metallicity, ionization parameter, and age are results from the SED fitting using BEAGLE.
a Quasars are included in the XQR-30 sample.b Remeasured column density with the same b value as in Davies et al. (2023a).c Impact parameter within 1 Mpc.