JADES: Physical properties of Ly and non-Ly emitters at z4.8-9.6
Abstract
We investigate the physical properties of Lyman-alpha emitters (LAEs) and non-Lyman-alpha emitters (non-LAEs) at z4.8–9.6 via a stacking analysis of 253 JWST/NIRSpec spectra of galaxies observed as part of the JWST Advanced Deep Extragalactic Survey (JADES). We identify a sample of 42 LAEs with the equivalent width of Ly20Å and a sample of 211 non-LAEs, divide each sample further via the median redshift of the LAEs (z6.3), and create composite spectra using the low and medium resolution spectra from NIRSpec. We estimate physical quantities such as dust extinction, UV continuum slope , electron temperatures, ionization parameter, escape fraction of Ly and Lyman Continuum, and the photon production rate for each bin/stack. The existing dust-extinction laws do not appear to be valid at these epochs. The emission line ratio analyses show that active galactic nuclei might dominate all sub-samples, irrespective of Ly emission. LAEs show much higher [O iii]/[O ii] and low [O ii]/H at z6.3 compared to non-LAEs, but these line ratios are not sufficient to distinguish the two populations at z6.3. However, the LAEs samples show large EW([O iii]4959, 5007) (1000Å) compared to the non-LAEs sample at all redshifts. C iv/Ly and C iv/C iii] for LAE population at z6.3 is a factor of 5 larger than that for LAE population at z6.3. The ionizing radiation for LAEs is hard, as revealed from several diagnostics including C iv detection, high [O iii]/[O ii](8), and large values of .
1 Introduction
Building a coherent picture of the Universe’s history is one of astronomy’s longstanding goals. According to the Lambda cold dark model (CDM), the Big Bang happened 13.8 Gyr ago (Planck Collaboration et al., 2020), which was followed by the recombination era at redshift z1100 when photons decoupled from baryons (Peebles & Yu, 1970), and are now observed as cosmic microwave background (CMB; Smoot et al., 1992). The ensuing cosmic dark age ended at z20–30 (Tegmark et al., 1997; Bromm et al., 2009), when the first stars and galaxies formed, resulting in Lyman continuum (LyC) photons capable of ionizing the neutral intergalactic medium (IGM) and thus instigating the process of cosmic reionization which lasted until z 5–6 (Fan et al., 2006b; Bosman et al., 2022).
While the pre-reionization epochs of the Universe can be observationally studied only via CMB, the reionization era can be explored via several direct and indirect probes. These methods encompass the studies of Gunn-Peterson absorption trough in quasar spectra (Gunn & Peterson, 1965), CMB polarization and temperature anisotropy (Kogut et al., 2003), H i 21cm signal from cosmic reionization (Furlanetto et al., 2004), gamma-ray bursts (Totani et al., 2006), and potential sources of high energy photons causing reionization such as quasars, active galactic nuclei (AGN), and early star-forming galaxies including Ly-break galaxies, Ly emitters and dusty-star-forming galaxies. See reviews by Fan et al. (2006a, 2023) and Stark (2016) for details on these observational probes of the reionization era.
The seminal work of Partridge & Peebles (1967) suggested that Ly emission is physically related to early galaxies. However, Ly proved elusive for many years, and when z3 galaxies were discovered and spectroscopically confirmed by (Steidel et al., 1996), it became clear that the equivalent width (EW) of Ly was generally much lower than that predicted from simple stellar population synthesis models and Case B recombination (which predict EW100–200Å), probably through the resonantly-scattered nature of this line. In the past couple of decades, Ly emitting galaxies (LAEs) have been identified well into the epoch of reionization thanks to careful selection techniques and exhaustive follow-up spectroscopic efforts, see e.g., Bunker et al. (2003); Stark et al. (2011); Curtis-Lake et al. (2012); Cassata et al. (2015); Jiang et al. (2017) for z6 LAEs, and Ono et al. (2012); Treu et al. (2013); Pentericci et al. (2014); Jung et al. (2020) for z7 LAEs. These observations allowed us to place constraints on the LAE properties such as Ly equivalent widths (Cassata et al., 2015), Ly fraction (Stark et al., 2010), Ly luminosity density and Ly escape fraction (Konno et al., 2018). However, the information encoded within the rest-frame ultraviolet (UV) and optical spectra of reionization-era LAEs remained largely elusive (though not impossible, see e.g., Stark et al. (2017)) up until the launch of the James Webb Space Telescope (JWST; Gardner et al., 2023).
The near-infrared spectrograph (NIRSpec Jakobsen et al., 2022) on JWST has not only pushed the Ly detection in galaxies out to z=10.6 (GNz11; Bunker et al., 2023a), but has also enabled the UV and optical spectroscopy of reionization era LAEs (see e.g., Jones et al., 2023; Saxena et al., 2023a, b; Jung et al., 2023; Tang et al., 2023; Maseda et al., 2023; Bunker et al., 2023a). These studies have revealed several physical properties of LAEs such as their dust content, ionization state, gas-phase metallicity, UV continuum slope, luminosity and Ly escape fraction. However, these studies are mostly composed of small sample of LAEs, focus on specific set of properties, and none of them compare the properties of LAEs with the galaxies which have weak or no Ly emission (non-LAEs hereafter) within a broad range of redshifts including the reionization era. This renders it difficult to draw broad conclusions about the LAE and non-LAE population in general.
In this paper, we use JWST/NIRSpec spectroscopy of a sample of 253 galaxies in the redshift range of 4.8–9.6 to study the connection between the Ly emission or absence thereof and the rest-frame UV and optical spectral properties. This is the first such study to employ statistically significant samples of LAEs and non-LAEs within different redshifts bins and present a comprehensive analysis of a vast range of properties encoded in the UV and optical spectra of the LAE and non-LAE populations.
The paper is organized as follows: Section 2 describes the observations, sample selection criteria, methodology to identify LAEs and non-LAEs, and the procedure to create the composite spectra on the basis of the strength of LAEs and redshift. In Section 3, we estimate fluxes of emission lines, color-excess, electron temperatures, gas-phase metallicities, ionization parameters, and Ly escape fractions () from the composite spectra. We also estimate average UV continuum slopes , absolute UV magnitudes, and LyC escape fractions () for each bin of galaxies. Section 4 explores correlations between the estimated physical quantities to distinguish the properties of different samples of LAEs and non-LAEs, such as their dust content and geometry, ionization conditions, escape fractions of Ly and LyC. Section 5 summarizes our results.
2 Data
2.1 The JADES survey
We use data observed as part of the JWST Advanced Deep Extragalactic Survey (JADES; Bunker et al., 2020; Rieke, 2020) which is based on guaranteed time observations of 770 hrs (PIDs: 1180, 1181, 1210, 1286, 1287) taken with JADES/NIRSpec micro shutter array (MSA; Ferruit et al., 2022) and JWST/NIRCam. The survey focused on the Great Observatories Origins Deep Survey (GOODS; Giavalisco et al., 2004) deep legacy fields in both hemispheres, GOODS-South and GOODS-North, and is described in detail in Eisenstein et al. (2023). The spectroscopic data were taken with the NIRSpec filter-grating combinations with spectral resolutions varying from 30-2700 with spectral coverage of 0.7-5 m. Please see Bunker et al. (2023b) and (D’Eugenio et al., 2024) for the first and second data release of JADES spectroscopic data. The details of data reduction for the spectroscopic data will be provided in Carniani et al. (in prep). The NIRSpec data used in this work were taken in between 5 October 2022 and 10 January 2024.
2.2 Sample Selection
We first determined the redshifts of all targets by inspecting their spectra by eye using the software inzimar developed by the JADES collaboration. The redshifts of the objects were based on either the continuum fit, a single or multiple emission lines from the spectra taken with the PRISM, medium or high-resolution gratings. More information on the redshift determination can be found in the documents provided in the first and second JADES data release. We restricted the analysis to a redshift range of 4.8–9.6, which allows us to access the rest wavelength range of 1215–5007Å. We determined a total of 523 unique NIRSpec targets with robust redshift estimates between the imposed redshift range. We co-added the medium-resolution spectra (G140M, G235M, and G395M) of each target by first resampling them using spectres (Carnall, 2017) on a wavelength grid with a wavelength sampling corresponding to that of G140M and then taking a weighted average. The spectra are not necessarily contiguous throughout the wavelength range, so we further imposed a quality check that the wavelength windows for the following emission lines have finite flux values: Ly1215, C iv 1548,1550, He ii 1640, O iii] 1660,1666, C iii] 1907,1909, [O ii] 3727,3729, [Ne iii] 3869, H4340, [O iii]4363, He ii 4686, H4860, [O iii] 4959,5007. The quality check on wavelength windows resulted in a total of 255 galaxies. As described in Section 2.4, we will use the UV continuum at 1500Å for normalizing the R1000 spectra before creating the composite spectra. So, we excluded two more targets for which the UV continuum at 1500Å was too weak to be detected. The final sample consists of 253 targets. Figure 1 (left-hand panel) shows the redshift distribution for all targets along with the sub-samples of LAEs and non-LAEs, described below. Figure 1 (right-hand panel) shows the tier distribution and the associated PID of JWST observations.
2.3 Identifying LAEs and non-LAEs
Ly emitters are typically defined as the sources with EW(Ly) 20Å (Hayes, 2015), though there are some variations across different works (see e.g., Tang et al., 2023). For estimating EW(Ly), we first estimate the Ly flux from the R1000 spectrum for each target in our sample. However, the continua in the R1000 spectra are too faint, so we use the R100 spectra for estimating the continuum levels at 1215.67 Å by fitting a power law to the spectra in the wavelength range 1300–1500 Å and thereby extrapolating to 1215.67Å. We only use the wavelength range red-ward of the Ly line because the R100 spectra generally show breaks at Ly line in the R100 spectra. We require that the Ly fluxes show signal-to-noise ratios (S/N) 3 and that the continua at 1215.67Å are positive. The EW(Ly) is then estimated from the Ly fluxes from R1000 and continua from R100, which results in 42 galaxies with EW(Ly) 20Å. We note that the EW(Ly) estimated in this work are in broad agreement with those found by Jones et al. (2023) and Saxena et al. (2023a) for the common targets. Figure 1 (left-panel, grey histogram) shows the redshift distribution of these 42 galaxies, which we refer to as the LAE sample hereafter. The rest of the 211 galaxies in the sample are referred to as the non-LAEs in this work (Figure 1, light blue histogram). The median redshift for the LAE sample is z6.28, shown by the dotted vertical line in the figure, and is used to identify sub-samples of LAEs and non-LAEs for creating composite spectra. Hence, we create 6 samples of targets: (a) 42 LAEs with EW(Ly)20Å(b) 21 LAEs lying below the median redshift z6.3 (c) 21 LAEs with z6.3 (d) 211 non-LAEs (e) 118 non-LAEs with z6.3 (f) 93 non-LAEs with z6.3.
2.4 Composite spectra
To create composite spectra, we first de-redshift and re-sample R1000 spectra on a common wavelength grid and normalize them with the continuum at 1500Å obtained from the corresponding R100 spectra. We then create 6 composite spectra corresponding to the 6 sub-samples identified in Section 2.3, by taking the average of all the spectra within a sample weighted by the uncertainties on the fluxes in R1000 spectra and those on continuum determination. Different normalization schemes and weighing methods are used in the literature depending on the goals of the work and data availability (see, e.g., Witstok et al., 2023; Boyett et al., 2024). Our choice of using the continuum 1500Å for normalization instead of using other parameters such as the peak flux density or the luminosity of an emission line, allowed us to include even those targets with faint emission lines by capitalizing over the availability of the R100 data for the continuum detection. Similarly, we chose to use the weighted average instead of median or mean because that allowed us to minimize the impact of noise in the data. Our experiments show that mean results in un-physical features in the stacked spectra while the median smoothens over the potential spectral features.
Figure 2 show six R1000 composite spectra obtained by normalizing with continuum from R100, and corresponds to 6 samples shown by different colors: (a) 42 LAEs with EW(Ly)20Å(black) (b) 21 LAEs lying below the median redshift z6.3 (orange) (c) 21 LAEs with z6.3 (green) (d) 211 non-LAEs (dark blue) (e) 118 non-LAEs with z6.3 (light blue) (f) 93 non-LAEs with z6.3 (magenta). The same color scheme for the samples is used in different figures in the rest of the paper.
We also created six more composite spectra solely using R100 data on a sub-sample of galaxies, for estimating EW([O iii]4959, 5007), EW(H) and production rate of H-ionizing photons per unit UV luminosity, . Since these composite spectra are based on a sub-sample of galaxies, we present the methodology to create the composite spectra and estimates of subsequent quantities in Appendix A.
2.5 Comparison samples from the literature
To compare our analysis with the literature, we selected samples such that all observable redshifts are covered, either via individual observations or composite spectra.
z0: For the local comparison sample, we use the individual galaxies’ data from Yang et al. (2017); Izotov et al. (2020); Flury et al. (2022a). We include five Ly emitters at z0 from Izotov et al. (2020) with EW(Ly)45–190Å. The sample of Yang et al. (2017) consists of the so-called blueberry galaxies at z0.05, which are considered local analogs of high-z Ly emitting galaxies and are characterized by very high ionization ([O iii]/[O ii] 10–60). Since is often associated with the properties of Ly emitters (EW, line profile, ), we include the galaxies from the Low-redshift Lyman Continuum Survey (LzLCS; Flury et al., 2022a) within z0.2–0.4.
z1–2: We use the results from the composite spectra of Tang et al. (2019) obtained by stacking galaxies in the redshift range of 1.3–2.4 and based on the EW([O iii] 5007).
3 Results
3.1 Flux measurements & Reddening correction
We measure the line fluxes for the recombination and collisionally excited emission lines by fitting Gaussian profiles to spectral features after subtracting a linear continuum in the spectral region of interest. We fit single Gaussian profiles to doublets such as C iv, C iii], and [O ii], because as they are unresolved in the composite spectra. However, the rest of the detected emission lines could be fit with the single Gaussian profiles. This is done for all the composite spectra described in Section 2.4.
For reddening correction, we use the extinction curve of Cardelli et al. (1989), which is deemed to be appropriate for high redshift Universe (Reddy et al., 2020), and has been used in related studies (Shapley et al., 2023a; Witstok et al., 2023). We estimate the color excess E(B-V) by combining the parameters of the extinction curve with the observed and intrinsic Balmer line ratio of H/H (=0.473), assuming a Case B recombination and electron temperature and density of 15,000 K and 100 cm-3, respectively. We do not use the H/H line ratio as it is not available for high-z samples. The E(B-V) is then used to estimate the dust-corrected line fluxes for all emission lines. For the subsequent analysis, we do not include the large uncertainties on E(B-V) while propagating errors on dust-corrected fluxes and other different dependent quantities, as they result in unrealistic uncertainties due to the non-Gaussian nature of uncertainty distribution. The random measurement uncertainty of observed flux is propagated to estimate uncertainties on dust-corrected line fluxes and other dependent quantities throughout the work.
LAEs | non-LAEs | |||||
Properties | All | z6.3 | z6.3 | All | z6.3 | z6.3 |
Na | 42 | 21 | 21 | 211 | 118 | 93 |
Derived from R1000 Composite Spectra | ||||||
RO3 | ||||||
R23 | ||||||
O3 | ||||||
O3Hg | ||||||
O2Hb | ||||||
O32 | ||||||
Ne3O2 | ||||||
Ne3Hdb | ||||||
O2Hdb | ||||||
Ly/H | – | – | – | |||
C iv/C iii]c | – | – | – | |||
C iv/Lyc | – | – | – | |||
EW(C iv)c | – | – | – | |||
Te([OIII]) (K) | ||||||
Te([OII]) (K) | ||||||
12+log(O/H)d | ||||||
log(U) | ||||||
Average of properties measured from each spectrume | ||||||
Notes:
a: N indicates the number of galaxies in each bin.
b: For estimating line ratios involving H, we combined the line ratios with respect to H (i.e., [Ne iii]/H and [O ii]/H) and used the theoretical line ratio of H/H= 3.81 obtained using pyneb for Te=15kK and Ne=100cm-3.
c: For the non-LAEs samples, we do not report the line ratios involving C iv as this doublet is not detected with enough S/N in the composite spectra of non-LAEs. We do not put any upper limit on C iv values because the origin of C iv can be both nebular and stellar.
d: Direct method metallicity.
e: Uncertainties for the average properties is , where is the standard deviation of estimated values ( and ) of all sources within a bin and N is the number of sources within the same bin.
The definition of different line ratios shown in the first column is as follows: RO3 = [O iii]4363/[O iii]5007, R23 =([O ii]3727,3729 + [O iii]4959,5007)/H, O3 = [O iii]5007/H, O3Hg = [O iii]4363/H, O2Hb = [O ii]3727,3729/H, O32 = [O iii]5007/[O ii]3727,3729, Ne3O2 = [Ne iii]3869/[O ii]3727,3729, Ne3Hd = [Ne iii]3869/H, O2Hd = [O ii]3727,3729/H.
3.2 Electron temperature & density
We measure electron temperature Te([O iii]) characterizing the high ionization zone by using the temperature-sensitive [O iii]4363/[O iii]5007 emission lines ratio in the calibration given by Pérez-Montero (2017) which is valid in the range of 7000-25000K. The detection of auroral lines [O ii]7320, 7330 are required to estimate Te of the low-ionization zone. Since this oxygen doublet lies beyond our wavelength range of analysis and hence not available, we simply use the linear relation Te([O iii])-Te([O ii]) from Pilyugin et al. (2009) to estimate Te([O ii]).
The electron density Ne can be determined by the flux ratio of density-sensitive line doublets C iii]1907, 1909 or [O ii]3727, 3729. However, even though both of these line doublets are detected with enough S/N (3) in our composite spectra, the peaks of these lines were not well-resolved to estimate the flux of the individual line of the doublets. This prevented us from estimating the Ne for our composite spectra. In the rest of the analysis, we assume an Ne = 300 cm-3, which has been observed and assumed in other studies of high-redshift studies (see e.g., Sanders et al., 2024).
3.3 Gas-phase metallicity
For estimating gas-phase metallicity via direct method, we use the software pyneb (Luridiana et al., 2015). We measure the O2+/H+ by using the line ratio [O iii] 4959, 5007/H and the high ionization electron temperature Te([O iii]). Similarly, we measure the O+/H+ by using the line ratio [O ii] 3727,3929 and the low ionization electron temperature Te([O ii]). We assume Ne=300cm-3 in both of the above estimates. We estimate the total oxygen abundance as the sum of O+/H+ and O2+/H+, and estimate the gas-phase metallicity, 12+log(O/H).
We also estimate gas-phase metallicities for our composite spectra via the indirect method. In Figure 3, we compare the gas-phase metallicities obtained with the direct Te method with those obtained from strong line metallicity calibrations proposed by Sanders et al. (2024). In particular, we use the emission lines such as O3 (=[O iii]/H), O2 (=[O ii]/H), R23 (=([O ii]+[O ii])/H), O32 (=[O iii]/[O ii]) and Ne3O2 (=[Ne iii]/[O ii]). The calibrations from Sanders et al. (2024) are devised for the galaxies from the reionization era and cosmic noon, though large scatters are found around these empirical relations. Figure 3 shows that the metallicities determined from O2, O32, and Ne3O2 show very high scatter which can not be explained by the intrinsic scatter on these calibrations reported by Sanders et al. (2024). Among these line ratios showing large scatter, the emission lines used in Ne3O2 are close in wavelength, and thus, Ne3O2 is independent of any systematics related to flux calibration. The large scatter in metallicities could be likely due to the poorly-understood ionization conditions in LAEs and non-LAEs (see Section 4.2), which manifest differently in different line ratios, and care should be taken while using these calibrations. The metallicities estimated from O3 and R23 show lesser scatter.
We do not try to compare these results with the calibrations for local galaxies from Curti et al. (2017) or those for the high-z analogs from Pérez-Montero et al. (2021); Bian et al. (2018) and Jones et al. (2015) as our direct method metallicites lie beyond the valid range of those calibrations. In principle, one can also use the strong line calibrations from Nakajima et al. (2022) for high-z galaxies.
3.4 Ionization parameter
The ionization parameter log(U) is expected to correlate with the degree of ionization of the nebula. We estimate log(U) by using the line ratio [O iii]/[O ii] in the linear relation proposed by Díaz et al. (2000) based on the single star photoionization models. We do not use the most recent calibrations from Kewley et al. (2019) as the metallicities derived from the composite spectra of all sub-samples except one (z6.3 non-LAEs) lie beyond the metallicity range of these calibrators (12+log(O/H) = 7.63–8.93).
3.5 Ly & LyC escape fraction
The Ly escape fraction is defined as the ratio of the observed Ly (redenning-corrected) flux to the expected Ly flux derived from the dust-corrected Balmer lines, e.g., H or H (though see Section 4.1). We use the H line flux for this purpose as the stronger H line is beyond the imposed wavelength range of some of our stacks. Hence, we estimate using the following formulae and assuming the theoretical Ly/H=23.3 for a Case B recombination and electron temperature and density of 15,000 K and 100 cm-3, respectively: = fLyα/(23.3 fHβ), where fLyα and fHβ are the dust-corrected line fluxes. The dust-corrected flux line ratio of Ly/H is found to be larger than the theoretical line ratio 23.3 (see Table 1) (see Sections 4.1 for possible causes). We also estimate assuming no dust-extinction for all stacked spectra and discuss them wherever relevant.
We estimate the Lyman Continuum escape fraction by using the relation between the UV continuum slope and given by Chisholm et al. (2022).
3.6 UV continuum slope & absolute magnitude MUV
We estimate the UV continuum slope of all sample targets by using R100 spectra in the spectral windows from Calzetti et al. (1994) in the wavelength range of 1250–2580Å. We estimate the absolute magnitude at 1500Å of all sample targets by using their R100 spectra. To do so, we first smooth the R100 spectra by a median filter of 11 pixels, then estimate the mean flux within a 50Å region centering 1500Å, which is then subsequently converted into the absolute magnitudes. In Table 1, we report the mean of the -slope values for all targets within the six sub-samples as the representative of that sub-sample, and similarly, the mean of MUV within a bin as the representative value of that bin.
4 Discussion
4.1 Interstellar dust
Figure 4 shows the variation of the average UV slope with the color-excess E(B-V) for each sub-sample. Note that the average UV slope is the mean of slopes of all spectra within a sub-sample, while the E(B-V) is directly obtained from the composite spectra. The E(B-V) values for all composite spectra, both LAEs and non-LAEs, lie within the E(B-V) range (0-1.19) found for the star-forming galaxies in the redshift range of z2.7–6.5 by Shapley et al. (2023b). If no redshift separation is applied, LAEs show similar color excess as the non-LAEs. We find that the LAEs at z6.3 show lower dust attenuation than those at z6.3. This could result due to LAEs following peculiar extinction laws (Atek et al., 2014) at high redshifts discussed later. On the other hand, the non-LAEs at z6.3 show higher dust attenuation than those at z6.3. It is possible that the results are affected by systematic uncertainties, though we do not find any direct evidence of such an effect (see Appendix B).
In Figure 4, we also over plot three different versus E(B-V) relations presented in Reddy et al. (2018), and corresponding to different extinction laws for fiducial 0.14 Z⊙ bpass models. These extinction curves include that of Calzetti et al. (2000), Gordon et al. (2003) and Reddy et al. (2015). It is not clear from Reddy et al. (2018) whether their relation corresponding to Calzetti et al. (2000) extinction curve is for nebular or stellar E(B-V), which differs by a factor of 0.44. We consider this factor by including two more straight lines (red and blue) obtained by dividing and multiplying the slope of the relation by 0.44. Note that we have estimated E(B-V) following the extinction law of Cardelli et al. (1989). However, the E(B-V) obtained by H/H will remain the same irrespective of extinction curves used because of similar behavior of extinction curves in the optical range. Figure 4 shows that the relations from Reddy et al. (2018) do not necessarily hold for all sub-samples irrespective of their redshifts or Ly emission. However, the blue dotted straight line seems to roughly satisfy the values reported for composite spectra data from Witstok et al. (2023) at z8. The lack of evident agreement between data and -E(B-V) relations might indicate that the ISM environment of the high-z galaxies does not follow the same attenuation law as the low redshift. Another possibility is that dust is not the prime driver for determining the UV slope at the high redshifts (see e.g., Wilkins et al., 2011). Apart from the dust content, is known or speculated to be sensitive to various parameters such as metallicity, age of the stellar population, initial mass function (Bouwens et al., 2010), nebular and stellar continuum (Raiter et al., 2010), and star-formation history (Schaerer & Pelló, 2005).
Figure 5 shows UV slope versus Ly/H derived for three LAEs sub-samples. We show Ly/H before and after dust correction via light and dark legends, respectively. We find that dust-corrected Ly/H is sometimes much larger than the theoretical value for a Case B recombination, which results in 1. In particular, we find that increases with E(B-V) for the LAE sub-samples when we compare Figures 4 and 5. The result is in agreement with previous studies and has been attributed to several physical processes or their combination at play. Firstly and as discussed earlier in the context of versus E(B-V) relation (Figure 4), LAEs at high redshifts may follow a peculiar dust extinction law, which is shaped by the ISM geometry of an inhomogeneous multi-phase ISM (Atek et al., 2014). In this scenario proposed by Neufeld (1991), dusty gas clouds reside within neutral small clouds in an intercloud medium with negligible absorption and scattering coefficients. Under suitable conditions, Ly photons, being prone to resonant scattering, will bounce off from one cloud to the other without suffering any dust attenuation, while the non-resonant photons from the Balmer series (such as H or H) pass through the neutral medium unaltered and suffer dust attenuation, thus resulting in a higher . Such an environment will also result in an enhanced EW(Ly), i.e., line-to-continuum photons, with an increased E(B-V), which can be explored further by studying dust attenuation in individual LAEs. The enhanced Ly/H has also been attributed to other arguments related to ISM geometry such as clumpy ISM resulting in a non-typical effective dust attenuation (Scarlata et al., 2009; Natta & Panagia, 1984), and accounting for more realistic ISM configurations (other than spherical geometry) favoring Ly transmission including ionized cones and sources’ inclination angles (Atek et al., 2014).
Another possibility for large values of Ly/H could be the origin of Ly emission. While Ly escape fraction is generally estimated assuming the theoretical value of Ly/H flux ratio in the case B recombination, this ratio can be higher in the case of collisional excitation (Dij, 2019) causing the Ly emission in the outer region of a galaxy (Ouchi et al., 2020) via cold accretion (see e.g., Dijkstra & Loeb, 2009; Goerdt et al., 2010) and possibly even outflows (Daddi et al., 2021). If the ISM temperature is higher than 2104K, the collisional excitation and de-excitation can dominate the recombination process (Atek et al., 2014). All sub-samples of LAEs, including those with 1, show the electron temperatures 2104K (see Table 1).
is expected to be connected to , because both of these quantities are affected by the geometry and the nature of gas in and around the young star-forming region. A positive correlation is found between and in both observational (Flury et al., 2022b; Begley et al., 2024) and theoretical studies (Dijkstra et al., 2016; Kimm et al., 2019; Maji et al., 2022). In Figure 5, we also show corresponding to values from the prescriptions of Chisholm et al. (2022). We find that of the three LAE samples agree with each other considering uncertainties, though it appears that LAEs at z6.3 have higher than those at z6.3. The -derived for the sub-sample containing all LAEs and those at z6.3 are less than 10%, and contradict the suggestion from Schaerer et al. (2022), which would categorize both of these LAE samples as strong LyC leakers (i.e., 0.1) owing to their high C iv/C iii] values (0.75, see Figure 6). We do not further comment upon estimated here because Choustikov et al. (2023) show that is a good indicator of only for metallicities 0.1 solar metallicity. Kumari et al. (2024) also find that estimated from is negligible compared to that estimated from far-infrared emission lines for a local analog of reionization era galaxy, Pox 186. Similarly, C iv/C iii] for Pox 186 was too low to categorize this dwarf as a LyC leaker (Kumari et al., 2024). Physical measurables (other than ) are also found to correlate with , though they are also sensitive to other physical quantities. We could not estimate from a recent 6-parameters-based recipe from Choustikov et al. (2023) as three of these parameters (O32, R23, and E(B-V)) are determined from the composite spectra and the other two parameters (, MUV) are the average values from galaxies in different samples, thus leading to a heterogeneous determination of parameters and hence not suitable. Moreover, the sixth parameter H luminosity is not possible to determine for each of the sub-samples as H is not necessarily detected for all galaxies. Such a comparison will be presented in Bunker et al. (in prep).
4.2 Ionization source
The composite spectra of all LAE samples (Figure 2) show clear detection of C iv thus manifesting that the LAEs are capable of producing very high energy photons (with energies in excess of 47.9 eV). In Figure 6, we find that there is a large variation in both C iv/C iii] and C iv/Ly line ratios. The LAEs at z6.3 have C iv/C iii] and C iv/Ly about 5 times larger than the LAEs at z6.3. Similarly, EW(C iv) of LAEs at z6.3 is 4 times larger than that at z6.3 (see Table 1), and are consistent with the narrow-lined AGNs (Alexandroff et al., 2013). C iv/C iii] versus C iii]/He ii has been explored for distinguishing the sources of ionization as star-forming, composite, and AGN (Feltre et al., 2016; Hirschmann et al., 2023; Scholtz et al., 2023). He ii1640 is detected in only 2 of the composite spectra, i.e., the one consisting of all LAEs and the other one with LAEs at z6.3. Their log(C iii]/He ii) values are too low (-0.35), thus classifying the two samples as AGN-dominated (Hirschmann et al., 2023).
Unlike UV lines, several optical lines are detected in all composite spectra, enabling us to explore the emission line ratio diagnostics diagrams involving the optical lines. However, several works have demonstrated that most of the existing optical emission line diagnostic diagrams fail to distinguish the source of ionization in high-redshift objects. The current work also confirms this (see Appendix C). In Figure 7, we explore [O iii]4363/H versus [O iii]/[O ii] diagnostic diagram proposed recently by Mazzolari et al. (2024), to distinguish the AGN-only versus SF or AGN population. This figure shows that AGN may dominate the nebular emission for a significant fraction of galaxies, and hence dominate the auroral emission in the composite spectra.
From the above analysis, it is not obvious what is the dominant feature that may preferentially select AGN. However, it seems to suggest that at high redshifts, the presence of an AGN might aid in creating ionized bubbles that help Ly to escape and become visible.
4.3 Ly versus non-Ly emitters
LAEs are known to show large values of [O iii]/[O ii] (e.g., Nakajima et al., 2016; Erb et al., 2016). Figure 8 shows the variation of log(O32) (=log([O iii]/[O ii])) with respect to temperature-sensitive line ratio log(RO3) (= log([O iii]4363/[O iii]5007)) for all sub-samples. Since RO3 and O32 are sensitive to Te([O iii]) and log(U), respectively, we also show their values on corresponding opposite axes, along with the individual galaxies’ data at different redshifts from literature including Yang et al. (2017, z), Izotov et al. (2020, z0) and Katz et al. (2023, z7.5). The RO3 values are relatively small for all samples, which results in larger electron temperatures compared to those found typically at lower redshifts (7000-14000K). These large electron temperatures manifest in low gas-phase metallicities of the composite spectra. The variation of gas-phase metallicity is relatively small (0.3 dex) compared to the variation in [O iii]/[O ii]( a factor of 10) across the six sub-samples. We find the same behavior in Figure 9, which shows the variation of O32 versus the indirect strong-line metallicity diagnostic R23, thus confirming the results of Nakajima et al. (2016) that low metallicity is not solely responsible for extreme values of [O iii]/[O ii] within LAEs. On the other hand, log(U) shows a relatively large variation (0.7 dex), thus indicating that a high ionization parameter could lead to high [O iii]/[O ii]. The ionization parameter depends on several factors, such as the hardness of the ionizing spectrum (Kumari et al., 2021), ionization photon production rate, gas density, and the volume filling factor of ionizing gas (Nakajima et al., 2016; Tang et al., 2019).
The composite spectrum corresponding to all LAEs has [O iii]/[O ii] about a factor of 2 larger than that containing all non-LAEs. At z6.3, [O iii]/[O ii] is 8 times stronger in LAEs than in non-LAEs. At z6.3, both LAEs and non-LAEs have comparable [O iii]/[O ii], which is likely because Ly emission is purely dictated by the intervening IGM at such high redshifts. This means that [O iii]/[O ii] can not be used to identify the LAEs and non-LAEs at high redshifts (z6.3).
Like [O iii]/[O ii], [Ne iii]/[O ii] is also a proxy for ionization parameter. In Figure 10, we show the variation of [O iii]/[O ii] with respect to [Ne iii]/[O ii]for all composite spectra along with the literature data. The relative behavior of [Ne iii]/[O ii] for different sub-samples is similar to that of [O iii]/[O ii]. However, the dynamic range of [Ne iii]/[O ii] across different samples is much smaller than that of [O iii]/[O ii]. For example, even at z6.3, [Ne iii]/[O ii] is only a factor 3 larger for LAEs than non-LAEs, compared to a variation of a factor 8 for [O iii]/[O ii]. In the same figure 10, we also overplot the [O iii]/[O ii] versus [Ne iii]/[O ii] relations for star-forming galaxies and Seyferts derived from their local (z0) Sloan Digital Sky Survey (SDSS) samples by Witstok et al. (2021), along with literature data at different redshifts (0z9). The measurements of high-redshift galaxies do not necessarily follow the empirical relations based on the measurements of local galaxies. This is likely because there are only a few data points at the higher end of [O iii]/[O ii] and [Ne iii]/[O ii] values in the sample considered in Witstok et al. (2021), and they do not extend as high as the ratios considered in this work. These empirical relations should be used in high-z studies with caution.
As for [O iii]/[O ii], the composite spectra of all LAEs sample and all non-LAEs sample show distinct line ratios [O ii]/H as shown in Figure 11. At z6.3, [O ii]/H for LAEs is 7 times weaker than for non-LAEs. The local LAEs (Yang et al., 2017; Izotov et al., 2020) show similarly stronger [O iii]/[O ii] and weaker [O ii]/H as shown by z6.3 composite LAE spectrum. We propose that in the absence Ly line coverage, a galaxy at z4.8–6.3 can be classified as an LAE if it has [O iii]/[O ii]8 and [O ii]/H2 (based on the composite spectrum of all LAEs sample), though detailed modeling of ionization sources and comparison with larger dataset would be needed to test this proposition. At z6.3, [O iii]/[O ii] and [O ii]/H remain comparable for both LAEs and non-LAEs; however, the degeneracy can be broken by inspecting EW([O iii]4959, 5007), which is typically high (1000Å) for LAEs than for non-LAEs (Table 2).
The for LAEs samples are systematically higher than the non-LAEs (Table 2), indicating that the LAEs are galaxies with intense radiation fields. The log() for LAE samples lies in the range of 25.63-25.75 (Hz erg-1), and are in agreement with those derived by Stark et al. (2015, 2017) for three LAEs. The composite spectra corresponding to all non-LAE samples and that of z6.3 non-LAEs show log() closer to the canonical values used in the reionization calculations (e.g., Robertson et al., 2013; Bouwens et al., 2015), but lower than that of z6.3 non-LAEs.
5 Summary
In this work, we have studied the UV and optical properties of a sample of 253 sources at z4.8–9.6 observed with JWST/NIRSpec with the aim of understanding the differences between the LAE and non-LAE populations at different redshifts. The main results of this study are summarized as follows:
-
1.
We identified 42 LAEs with EW(Ly)20 and 211 non-LAEs with weak or no Ly emission. We further subdivided each sub-sample of LAEs and non-LAEs on the basis of the median redshift of LAEs, i.e., z6.28.
-
2.
We created six composite spectra by taking the variance-weighted mean of the R1000 spectra normalized by the UV continua estimated from R100 spectra. We used these composite spectra to estimate UV and optical emission line ratios and other dependent quantities such as electron temperatures, direct and indirect metallicities, and ionization parameters.
-
3.
We created another set of six composite spectra by stacking the R100 spectra normalized by the continua at [O iii]+H complex. We used these composite spectra to determine the EW(H), EW([O iii]4959, 5007) and .
-
4.
We estimated average UV continuum slope and UV magnitude MUV by simply taking the mean of the corresponding values for each target within a sub-sample. The versus E(B-V) for LAE and non-LAEs do not follow the local relations, indicating that the local dust-attenuation laws are not valid for high-z environments.
-
5.
Ly/H for two LAE samples (EW(Ly)20Å, and z6.3) show values higher than the theoretical prediction for Case B recombination, which might be related to a clumpy dust geometry or the collisional excitation dominating the recombination processes. This results in 1 for these LAE samples.
-
6.
C iv doublet remains undetected in the non-LAE composite spectra while they are well-detected in the LAE samples. C iv/Ly and C iv/C iii] for LAE population at z6.3 is a factor of 5 larger than that for LAE population at z6.3. C iv/C iii] line ratios and UV slope for a given sub-sample of LAEs give contradictory information about the escape fraction of LyC.
-
7.
Both UV and optical line ratios suggest that AGN might dominate all sub-samples of galaxies under study irrespective of Ly emission and redshift. However, it is not clear how these AGNs are selected within each sub-sample. The hint of AGN in each sub-sample also suggests that they might be responsible for creating ionized bubbles and hence Ly escape.
-
8.
LAEs have much stronger [O iii]/[O ii]and much lower [O ii]/H compared to the non-LAEs at z6.3. However, both of these line ratios are comparable for LAEs and non-LAEs at z6.3, where a high value of EW([O iii]4959, 5007) (1000Å) is useful in breaking the degeneracy and distinguishing LAEs from non-LAEs.
-
9.
LAEs have harder ionizing radiation field as revealed by C iv detection, high [O iii]/[O ii] and large compared to the non-LAEs.
In summary, the present analysis allowed us to identify criteria to distinguish between the LAE and non-LAE populations in general, as well as at different redshifts. It also raised questions about the applicability of the existing dust attenuation laws at high redshifts, the use of typical emission line ratio diagnostics in identifying the ionization mechanisms within distant galaxies, and the likely role of AGN in creating ionized bubbles in the reionization era. Further detailed studies of LAEs and non-LAEs involving deeper spectroscopy ensuring the detection of continuum and He ii features, combined with modeling and simulations, are required to address these questions and better understand the populations of LAEs and non-LAEs across different redshifts and particularly their role in reionizing the Universe.
Appendix A R100 Composite Spectra & Properties
For creating the R100 composite spectra, we de-redshift and resample all R100 spectra on a common wavelength grid, normalize them by the continuum level at the [O iii]+H complex, and take the mean weighted by the variance in the spectra. We follow the above procedure for all six sub-samples described in Section 2.2. However, the continua were not detected in some of the galaxies. So, the number of galaxies within R100 composite spectra is slightly different than that in the R1000 composite spectra but still statistically significant within each bin (compare N in Tables 1 and 2).
We use the R100 composite spectra obtained by normalizing the spectra by the continua at [O iii]+H to estimate the EW([O iii]+H). We integrate the flux under the H line and that under [O iii] doublet to estimate the corresponding EWs. The total fluxes from these continua-normalized composite spectra correspond to the EWs.
We estimate for each composite R100 spectra using EW([O iii]4959,5007) in the following relation from Chevallard et al. (2018): .
Table 2 summarizes the properties of the R100 composite spectra estimated in this section.
LAEs | non-LAEs | |||||
Properties | All | z6.3 | z6.3 | All | z6.3 | z6.3 |
N | 38 | 18 | 20 | 179 | 104 | 75 |
EW([O iii]4959, 5007) (Å) | ||||||
EW(H) (Å) | 2 | 1 | ||||
log( /Hz erg-1) | 25.68 | 25.75 | 25.63 | 25.31 | 25.29 | 25.45 |
Appendix B Uncertainties analysis
D’Eugenio et al. (2024) report a relative flux uncertainty of 15% between the prism and gratings data. To investigate its effect on E(B-V) values of the composite spectra, we also create composite spectra weighted by the variance from the uncertainties accounting for the uncertainties in the R1000 spectra, dispersion in R100 spectra in the 1500Å wavelength window and a 15% relative uncertainty between R1000 and R100 spectra. We then fit single Gaussian to the H and H emission lines by taking the above-estimated uncertainties into account and, thus, we estimate E(B-V) from these line fluxes. We compare these E(B-V) with those estimated from composite spectra where the systematic flux uncertainty of 15% is not taken into account, and we give equal weight to every data point while fitting the emission lines. The E(B-V) obtained from the two methods agree with each other within uncertainties.
Appendix C Ambiguous ionization sources
In Figure 13, we use two emission-line diagnostic diagrams: [O ii]/H versus [Ne iii]/H (left-hand panel) and [Ne iii]/[O ii] versus[O iii]/H (right-hand panel). The former was proposed by Pérez-Montero et al. (2007) and provided an empirical SF-AGN separator (black curve on the left-hand panel in Figure 13). The source of ionization for the LAE and non-LAE samples can not be distinguished as their line ratios lie within the uncertainty range of 0.15 dex along the SF-AGN separator. However, they occupy the same parameter space as the local star-forming galaxies presented in Izotov et al. (2020) and Yang et al. (2017). Similarly, the SF-AGN separator curve devised by Backhaus et al. (2022) classifies the LAE and non-LAE samples as AGNs, which occupy the same parameter space as those occupied by star-forming galaxies at high-redshifts (Tang et al., 2019; Shapley et al., 2023a) and their local analogs (Izotov et al., 2020; Yang et al., 2017). Our result agrees with that of Cleri et al. (2023), who also report that the [O iii]/H-[Ne iii]/[O ii] is not a good diagnostic of the ionizing source.
References
- Dij (2019) 2019, Saas-Fee Advanced Course, Vol. 46, Lyman-alpha as an Astrophysical and Cosmological Tool, doi: 10.1007/978-3-662-59623-4
- Alexandroff et al. (2013) Alexandroff, R., Strauss, M. A., Greene, J. E., et al. 2013, MNRAS, 435, 3306, doi: 10.1093/mnras/stt1500
- Asplund et al. (2009) Asplund, M., Grevesse, N., Sauval, A. J., & Scott, P. 2009, ARA&A, 47, 481, doi: 10.1146/annurev.astro.46.060407.145222
- Astropy Collaboration et al. (2013) Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al. 2013, A&A, 558, A33, doi: 10.1051/0004-6361/201322068
- Astropy Collaboration et al. (2018) Astropy Collaboration, Price-Whelan, A. M., Sipőcz, B. M., et al. 2018, AJ, 156, 123, doi: 10.3847/1538-3881/aabc4f
- Atek et al. (2014) Atek, H., Kunth, D., Schaerer, D., et al. 2014, A&A, 561, A89, doi: 10.1051/0004-6361/201321519
- Backhaus et al. (2022) Backhaus, B. E., Trump, J. R., Cleri, N. J., et al. 2022, ApJ, 926, 161, doi: 10.3847/1538-4357/ac3919
- Begley et al. (2024) Begley, R., Cullen, F., McLure, R. J., et al. 2024, MNRAS, 527, 4040, doi: 10.1093/mnras/stad3417
- Bian et al. (2018) Bian, F., Kewley, L. J., & Dopita, M. A. 2018, ApJ, 859, 175, doi: 10.3847/1538-4357/aabd74
- Bosman et al. (2022) Bosman, S. E. I., Davies, F. B., Becker, G. D., et al. 2022, MNRAS, 514, 55, doi: 10.1093/mnras/stac1046
- Bouwens et al. (2015) Bouwens, R. J., Illingworth, G. D., Oesch, P. A., et al. 2015, ApJ, 811, 140, doi: 10.1088/0004-637X/811/2/140
- Bouwens et al. (2010) —. 2010, ApJ, 708, L69, doi: 10.1088/2041-8205/708/2/L69
- Boyett et al. (2024) Boyett, K., Bunker, A. J., Curtis-Lake, E., et al. 2024, arXiv e-prints, arXiv:2401.16934, doi: 10.48550/arXiv.2401.16934
- Bromm et al. (2009) Bromm, V., Yoshida, N., Hernquist, L., & McKee, C. F. 2009, Nature, 459, 49, doi: 10.1038/nature07990
- Bunker et al. (2020) Bunker, A. J., NIRSPEC Instrument Science Team, & JAESs Collaboration. 2020, in Uncovering Early Galaxy Evolution in the ALMA and JWST Era, ed. E. da Cunha, J. Hodge, J. Afonso, L. Pentericci, & D. Sobral, Vol. 352, 342–346, doi: 10.1017/S1743921319009463
- Bunker et al. (2003) Bunker, A. J., Stanway, E. R., Ellis, R. S., McMahon, R. G., & McCarthy, P. J. 2003, MNRAS, 342, L47, doi: 10.1046/j.1365-8711.2003.06664.x
- Bunker et al. (2023a) Bunker, A. J., Saxena, A., Cameron, A. J., et al. 2023a, A&A, 677, A88, doi: 10.1051/0004-6361/202346159
- Bunker et al. (2023b) Bunker, A. J., Cameron, A. J., Curtis-Lake, E., et al. 2023b, arXiv e-prints, arXiv:2306.02467, doi: 10.48550/arXiv.2306.02467
- Calzetti et al. (2000) Calzetti, D., Armus, L., Bohlin, R. C., et al. 2000, ApJ, 533, 682, doi: 10.1086/308692
- Calzetti et al. (1994) Calzetti, D., Kinney, A. L., & Storchi-Bergmann, T. 1994, ApJ, 429, 582, doi: 10.1086/174346
- Cardelli et al. (1989) Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245, doi: 10.1086/167900
- Carnall (2017) Carnall, A. C. 2017, arXiv e-prints, arXiv:1705.05165, doi: 10.48550/arXiv.1705.05165
- Cassata et al. (2015) Cassata, P., Tasca, L. A. M., Le Fèvre, O., et al. 2015, A&A, 573, A24, doi: 10.1051/0004-6361/201423824
- Chevallard et al. (2018) Chevallard, J., Charlot, S., Senchyna, P., et al. 2018, MNRAS, 479, 3264, doi: 10.1093/mnras/sty1461
- Chisholm et al. (2022) Chisholm, J., Saldana-Lopez, A., Flury, S., et al. 2022, MNRAS, 517, 5104, doi: 10.1093/mnras/stac2874
- Choustikov et al. (2023) Choustikov, N., Katz, H., Saxena, A., et al. 2023, arXiv e-prints, arXiv:2304.08526, doi: 10.48550/arXiv.2304.08526
- Cleri et al. (2023) Cleri, N. J., Olivier, G. M., Hutchison, T. A., et al. 2023, ApJ, 953, 10, doi: 10.3847/1538-4357/acde55
- Curti et al. (2017) Curti, M., Cresci, G., Mannucci, F., et al. 2017, MNRAS, 465, 1384, doi: 10.1093/mnras/stw2766
- Curtis-Lake et al. (2012) Curtis-Lake, E., McLure, R. J., Pearce, H. J., et al. 2012, MNRAS, 422, 1425, doi: 10.1111/j.1365-2966.2012.20720.x
- Daddi et al. (2021) Daddi, E., Valentino, F., Rich, R. M., et al. 2021, A&A, 649, A78, doi: 10.1051/0004-6361/202038700
- D’Eugenio et al. (2024) D’Eugenio, F., Cameron, A. J., Scholtz, J., et al. 2024, arXiv e-prints, arXiv:2404.06531, doi: 10.48550/arXiv.2404.06531
- Díaz et al. (2000) Díaz, A. I., Castellanos, M., Terlevich, E., & Luisa García-Vargas, M. 2000, MNRAS, 318, 462, doi: 10.1046/j.1365-8711.2000.03737.x
- Dijkstra et al. (2016) Dijkstra, M., Gronke, M., & Venkatesan, A. 2016, ApJ, 828, 71, doi: 10.3847/0004-637X/828/2/71
- Dijkstra & Loeb (2009) Dijkstra, M., & Loeb, A. 2009, MNRAS, 400, 1109, doi: 10.1111/j.1365-2966.2009.15533.x
- Eisenstein et al. (2023) Eisenstein, D. J., Willott, C., Alberts, S., et al. 2023, arXiv e-prints, arXiv:2306.02465, doi: 10.48550/arXiv.2306.02465
- Erb et al. (2016) Erb, D. K., Pettini, M., Steidel, C. C., et al. 2016, ApJ, 830, 52, doi: 10.3847/0004-637X/830/1/52
- Fan et al. (2023) Fan, X., Bañados, E., & Simcoe, R. A. 2023, ARA&A, 61, 373, doi: 10.1146/annurev-astro-052920-102455
- Fan et al. (2006a) Fan, X., Carilli, C. L., & Keating, B. 2006a, ARA&A, 44, 415, doi: 10.1146/annurev.astro.44.051905.092514
- Fan et al. (2006b) Fan, X., Strauss, M. A., Becker, R. H., et al. 2006b, AJ, 132, 117, doi: 10.1086/504836
- Feltre et al. (2016) Feltre, A., Charlot, S., & Gutkin, J. 2016, MNRAS, 456, 3354, doi: 10.1093/mnras/stv2794
- Ferruit et al. (2022) Ferruit, P., Jakobsen, P., Giardino, G., et al. 2022, A&A, 661, A81, doi: 10.1051/0004-6361/202142673
- Flury et al. (2022a) Flury, S. R., Jaskot, A. E., Ferguson, H. C., et al. 2022a, ApJS, 260, 1, doi: 10.3847/1538-4365/ac5331
- Flury et al. (2022b) —. 2022b, ApJ, 930, 126, doi: 10.3847/1538-4357/ac61e4
- Furlanetto et al. (2004) Furlanetto, S. R., Zaldarriaga, M., & Hernquist, L. 2004, ApJ, 613, 16, doi: 10.1086/423028
- Gardner et al. (2023) Gardner, J. P., Mather, J. C., Abbott, R., et al. 2023, PASP, 135, 068001, doi: 10.1088/1538-3873/acd1b5
- Giavalisco et al. (2004) Giavalisco, M., Ferguson, H. C., Koekemoer, A. M., et al. 2004, ApJ, 600, L93, doi: 10.1086/379232
- Goerdt et al. (2010) Goerdt, T., Dekel, A., Sternberg, A., et al. 2010, MNRAS, 407, 613, doi: 10.1111/j.1365-2966.2010.16941.x
- Gordon et al. (2003) Gordon, K. D., Clayton, G. C., Misselt, K. A., Landolt, A. U., & Wolff, M. J. 2003, ApJ, 594, 279, doi: 10.1086/376774
- Gunn & Peterson (1965) Gunn, J. E., & Peterson, B. A. 1965, ApJ, 142, 1633, doi: 10.1086/148444
- Harris et al. (2020) Harris, C. R., Millman, K. J., van der Walt, S. J., et al. 2020, Nature, 585, 357, doi: 10.1038/s41586-020-2649-2
- Hayes (2015) Hayes, M. 2015, PASA, 32, e027, doi: 10.1017/pasa.2015.25
- Hirschmann et al. (2023) Hirschmann, M., Charlot, S., Feltre, A., et al. 2023, MNRAS, 526, 3610, doi: 10.1093/mnras/stad2955
- Hunter (2007) Hunter, J. D. 2007, Computing in Science and Engineering, 9, 90, doi: 10.1109/MCSE.2007.55
- Izotov et al. (2020) Izotov, Y. I., Schaerer, D., Worseck, G., et al. 2020, MNRAS, 491, 468, doi: 10.1093/mnras/stz3041
- Jakobsen et al. (2022) Jakobsen, P., Ferruit, P., Alves de Oliveira, C., et al. 2022, A&A, 661, A80, doi: 10.1051/0004-6361/202142663
- Jiang et al. (2017) Jiang, L., Shen, Y., Bian, F., et al. 2017, ApJ, 846, 134, doi: 10.3847/1538-4357/aa8561
- Jones et al. (2023) Jones, G. C., Bunker, A. J., Saxena, A., et al. 2023, arXiv e-prints, arXiv:2306.02471, doi: 10.48550/arXiv.2306.02471
- Jones et al. (2015) Jones, T., Martin, C., & Cooper, M. C. 2015, ApJ, 813, 126, doi: 10.1088/0004-637X/813/2/126
- Jung et al. (2020) Jung, I., Finkelstein, S. L., Dickinson, M., et al. 2020, ApJ, 904, 144, doi: 10.3847/1538-4357/abbd44
- Jung et al. (2023) Jung, I., Finkelstein, S. L., Arrabal Haro, P., et al. 2023, arXiv e-prints, arXiv:2304.05385, doi: 10.48550/arXiv.2304.05385
- Katz et al. (2020) Katz, H., Ďurovčíková, D., Kimm, T., et al. 2020, MNRAS, 498, 164, doi: 10.1093/mnras/staa2355
- Katz et al. (2023) Katz, H., Saxena, A., Cameron, A. J., et al. 2023, MNRAS, 518, 592, doi: 10.1093/mnras/stac2657
- Kewley et al. (2019) Kewley, L. J., Nicholls, D. C., Sutherland, R., et al. 2019, ApJ, 880, 16, doi: 10.3847/1538-4357/ab16ed
- Kimm et al. (2019) Kimm, T., Blaizot, J., Garel, T., et al. 2019, MNRAS, 486, 2215, doi: 10.1093/mnras/stz989
- Kogut et al. (2003) Kogut, A., Spergel, D. N., Barnes, C., et al. 2003, ApJS, 148, 161, doi: 10.1086/377219
- Konno et al. (2018) Konno, A., Ouchi, M., Shibuya, T., et al. 2018, PASJ, 70, S16, doi: 10.1093/pasj/psx131
- Kumari et al. (2021) Kumari, N., Amorín, R., Pérez-Montero, E., Vílchez, J., & Maiolino, R. 2021, MNRAS, 508, 1084, doi: 10.1093/mnras/stab2495
- Kumari et al. (2024) Kumari, N., Smit, R., Leitherer, C., et al. 2024, MNRAS, doi: 10.1093/mnras/stae252
- Luridiana et al. (2015) Luridiana, V., Morisset, C., & Shaw, R. A. 2015, A&A, 573, A42, doi: 10.1051/0004-6361/201323152
- Maji et al. (2022) Maji, M., Verhamme, A., Rosdahl, J., et al. 2022, A&A, 663, A66, doi: 10.1051/0004-6361/202142740
- Maseda et al. (2023) Maseda, M. V., Lewis, Z., Matthee, J., et al. 2023, ApJ, 956, 11, doi: 10.3847/1538-4357/acf12b
- Mazzolari et al. (2024) Mazzolari, G., Übler, H., Maiolino, R., et al. 2024, arXiv e-prints, arXiv:2404.10811, doi: 10.48550/arXiv.2404.10811
- Nakajima et al. (2016) Nakajima, K., Ellis, R. S., Iwata, I., et al. 2016, ApJ, 831, L9, doi: 10.3847/2041-8205/831/1/L9
- Nakajima et al. (2022) Nakajima, K., Ouchi, M., Xu, Y., et al. 2022, ApJS, 262, 3, doi: 10.3847/1538-4365/ac7710
- Natta & Panagia (1984) Natta, A., & Panagia, N. 1984, ApJ, 287, 228, doi: 10.1086/162681
- Neufeld (1991) Neufeld, D. A. 1991, ApJ, 370, L85, doi: 10.1086/185983
- Newville et al. (2014) Newville, M., Stensitzki, T., Allen, D. B., & Ingargiola, A. 2014, LMFIT: Non-Linear Least-Square Minimization and Curve-Fitting for Python, 0.8.0, Zenodo, doi: 10.5281/zenodo.11813
- Oke (1990) Oke, J. B. 1990, AJ, 99, 1621, doi: 10.1086/115444
- Ono et al. (2012) Ono, Y., Ouchi, M., Mobasher, B., et al. 2012, ApJ, 744, 83, doi: 10.1088/0004-637X/744/2/83
- Ouchi et al. (2020) Ouchi, M., Ono, Y., & Shibuya, T. 2020, ARA&A, 58, 617, doi: 10.1146/annurev-astro-032620-021859
- Partridge & Peebles (1967) Partridge, R. B., & Peebles, P. J. E. 1967, ApJ, 147, 868, doi: 10.1086/149079
- Peebles & Yu (1970) Peebles, P. J. E., & Yu, J. T. 1970, ApJ, 162, 815, doi: 10.1086/150713
- Pentericci et al. (2014) Pentericci, L., Vanzella, E., Fontana, A., et al. 2014, ApJ, 793, 113, doi: 10.1088/0004-637X/793/2/113
- Pérez-Montero (2017) Pérez-Montero, E. 2017, PASP, 129, 043001, doi: 10.1088/1538-3873/aa5abb
- Pérez-Montero et al. (2021) Pérez-Montero, E., Amorín, R., Sánchez Almeida, J., et al. 2021, MNRAS, 504, 1237, doi: 10.1093/mnras/stab862
- Pérez-Montero et al. (2007) Pérez-Montero, E., Hägele, G. F., Contini, T., & Díaz, Á. I. 2007, MNRAS, 381, 125, doi: 10.1111/j.1365-2966.2007.12213.x
- Pilyugin et al. (2009) Pilyugin, L. S., Mattsson, L., Vílchez, J. M., & Cedrés, B. 2009, MNRAS, 398, 485, doi: 10.1111/j.1365-2966.2009.15182.x
- Planck Collaboration et al. (2020) Planck Collaboration, Aghanim, N., Akrami, Y., et al. 2020, A&A, 641, A6, doi: 10.1051/0004-6361/201833910
- Raiter et al. (2010) Raiter, A., Schaerer, D., & Fosbury, R. A. E. 2010, A&A, 523, A64, doi: 10.1051/0004-6361/201015236
- Reddy et al. (2015) Reddy, N. A., Kriek, M., Shapley, A. E., et al. 2015, ApJ, 806, 259, doi: 10.1088/0004-637X/806/2/259
- Reddy et al. (2018) Reddy, N. A., Oesch, P. A., Bouwens, R. J., et al. 2018, ApJ, 853, 56, doi: 10.3847/1538-4357/aaa3e7
- Reddy et al. (2020) Reddy, N. A., Shapley, A. E., Kriek, M., et al. 2020, ApJ, 902, 123, doi: 10.3847/1538-4357/abb674
- Rieke (2020) Rieke, M. 2020, in Uncovering Early Galaxy Evolution in the ALMA and JWST Era, ed. E. da Cunha, J. Hodge, J. Afonso, L. Pentericci, & D. Sobral, Vol. 352, 337–341, doi: 10.1017/S1743921319008950
- Robertson et al. (2013) Robertson, B. E., Furlanetto, S. R., Schneider, E., et al. 2013, ApJ, 768, 71, doi: 10.1088/0004-637X/768/1/71
- Rohatgi (2022) Rohatgi, A. 2022, Webplotdigitizer: Version 4.6. https://automeris.io/WebPlotDigitizer
- Sanders et al. (2023) Sanders, R. L., Shapley, A. E., Topping, M. W., Reddy, N. A., & Brammer, G. B. 2023, ApJ, 955, 54, doi: 10.3847/1538-4357/acedad
- Sanders et al. (2024) —. 2024, ApJ, 962, 24, doi: 10.3847/1538-4357/ad15fc
- Saxena et al. (2023a) Saxena, A., Bunker, A. J., Jones, G. C., et al. 2023a, arXiv e-prints, arXiv:2306.04536, doi: 10.48550/arXiv.2306.04536
- Saxena et al. (2023b) Saxena, A., Robertson, B. E., Bunker, A. J., et al. 2023b, A&A, 678, A68, doi: 10.1051/0004-6361/202346245
- Scarlata et al. (2009) Scarlata, C., Colbert, J., Teplitz, H. I., et al. 2009, ApJ, 704, L98, doi: 10.1088/0004-637X/704/2/L98
- Schaerer & Pelló (2005) Schaerer, D., & Pelló, R. 2005, MNRAS, 362, 1054, doi: 10.1111/j.1365-2966.2005.09373.x
- Schaerer et al. (2022) Schaerer, D., Izotov, Y. I., Worseck, G., et al. 2022, A&A, 658, L11, doi: 10.1051/0004-6361/202243149
- Schmidt et al. (2021) Schmidt, K. B., Kerutt, J., Wisotzki, L., et al. 2021, A&A, 654, A80, doi: 10.1051/0004-6361/202140876
- Scholtz et al. (2023) Scholtz, J., Maiolino, R., D’Eugenio, F., et al. 2023, arXiv e-prints, arXiv:2311.18731, doi: 10.48550/arXiv.2311.18731
- Shapley et al. (2023a) Shapley, A. E., Reddy, N. A., Sanders, R. L., Topping, M. W., & Brammer, G. B. 2023a, ApJ, 950, L1, doi: 10.3847/2041-8213/acd939
- Shapley et al. (2023b) Shapley, A. E., Sanders, R. L., Reddy, N. A., Topping, M. W., & Brammer, G. B. 2023b, ApJ, 954, 157, doi: 10.3847/1538-4357/acea5a
- Smoot et al. (1992) Smoot, G. F., Bennett, C. L., Kogut, A., et al. 1992, ApJ, 396, L1, doi: 10.1086/186504
- Stark (2016) Stark, D. P. 2016, ARA&A, 54, 761, doi: 10.1146/annurev-astro-081915-023417
- Stark et al. (2010) Stark, D. P., Ellis, R. S., Chiu, K., Ouchi, M., & Bunker, A. 2010, MNRAS, 408, 1628, doi: 10.1111/j.1365-2966.2010.17227.x
- Stark et al. (2011) Stark, D. P., Ellis, R. S., & Ouchi, M. 2011, ApJ, 728, L2, doi: 10.1088/2041-8205/728/1/L2
- Stark et al. (2015) Stark, D. P., Walth, G., Charlot, S., et al. 2015, MNRAS, 454, 1393, doi: 10.1093/mnras/stv1907
- Stark et al. (2017) Stark, D. P., Ellis, R. S., Charlot, S., et al. 2017, MNRAS, 464, 469, doi: 10.1093/mnras/stw2233
- Steidel et al. (1996) Steidel, C. C., Giavalisco, M., Pettini, M., Dickinson, M., & Adelberger, K. L. 1996, ApJ, 462, L17, doi: 10.1086/310029
- Tang et al. (2019) Tang, M., Stark, D. P., Chevallard, J., & Charlot, S. 2019, MNRAS, 489, 2572, doi: 10.1093/mnras/stz2236
- Tang et al. (2023) Tang, M., Stark, D. P., Chen, Z., et al. 2023, MNRAS, 526, 1657, doi: 10.1093/mnras/stad2763
- Taylor (2005) Taylor, M. B. 2005, in Astronomical Society of the Pacific Conference Series, Vol. 347, Astronomical Data Analysis Software and Systems XIV, ed. P. Shopbell, M. Britton, & R. Ebert, 29
- Tegmark et al. (1997) Tegmark, M., Silk, J., Rees, M. J., et al. 1997, ApJ, 474, 1, doi: 10.1086/303434
- Totani et al. (2006) Totani, T., Kawai, N., Kosugi, G., et al. 2006, PASJ, 58, 485, doi: 10.1093/pasj/58.3.485
- Treu et al. (2013) Treu, T., Schmidt, K. B., Trenti, M., Bradley, L. D., & Stiavelli, M. 2013, ApJ, 775, L29, doi: 10.1088/2041-8205/775/1/L29
- Wilkins et al. (2011) Wilkins, S. M., Bunker, A. J., Stanway, E., Lorenzoni, S., & Caruana, J. 2011, MNRAS, 417, 717, doi: 10.1111/j.1365-2966.2011.19315.x
- Witstok et al. (2021) Witstok, J., Smit, R., Maiolino, R., et al. 2021, MNRAS, 508, 1686, doi: 10.1093/mnras/stab2591
- Witstok et al. (2023) Witstok, J., Shivaei, I., Smit, R., et al. 2023, Nature, 621, 267, doi: 10.1038/s41586-023-06413-w
- Yang et al. (2017) Yang, H., Malhotra, S., Rhoads, J. E., & Wang, J. 2017, ApJ, 847, 38, doi: 10.3847/1538-4357/aa8809