EIGER II. first spectroscopic characterisation of the young stars and ionised gas associated with strong H$β$ and [OIII] line-emission in galaxies at z=5-7 with JWST

TL;DR

This study leverages the first deep JWST/NIRCam wide-field slitless spectroscopy observations to characterize 117 [OIII]-emitting galaxies at z = 5.33–6.93 around the quasar J0100+2802. Through dual detection approaches, 133 [OIII] components are identified and grouped into 117 systems, with strong Hβ and [OIII] emission yielding rest-frame EWs of order 10^2–10^3 Å and young, low-dust stellar populations. SED fitting with Nebular emission reveals typical ages ~100 Myr, M⋆ ~ 10^6.8–10^10.1 M⊙, low metallicities (12+log(O/H) ~ 7.3–7.9), and a high ionization state (ξ_ion ≳ 10^25.3 Hz erg^−1), driving strong [OIII] outputs. The [OIII] luminosity function at z ≈ 6 shows little evolution relative to z ≈ 3, and the L_[OIII]–L_UV relation indicates higher [OIII] luminosities at fixed UV brightness than at lower redshift, implying an ISM and metallicity-driven enhancement of line emission during reionization. Overall, the work demonstrates that JWST/NIRCam WFSS is a highly efficient, photometry-independent probe for mapping the high-z galaxy density and chemical/ionization state during the epoch of reionization.

Abstract

We present emission-line measurements and physical interpretations for a sample of 117 [OIII] emitting galaxies at $z=5.33-6.93$, using the first deep JWST/NIRCam wide field slitless spectroscopic observations. Our 9.7-hour integration is centered upon the $z=6.3$ quasar J0100+2802 -- the first of six fields targeted by the EIGER survey -- and covers $λ=3-4$ microns. We detect 133 [OIII] doublets, but merge pairs within $\approx$10 kpc and 600 km s$^{-1}$, motivated by their small scale clustering excess. We detect H$β$ in 68 and H$γ$ emission in two galaxies. The galaxies are characterised by a UV luminosity M$_{\rm UV}\sim-19.6$ ($-17.7$ to $-22.3$), stellar mass ~$10^8$ $(10^{6.8-10.1})$ M$_{\odot}$, H$β$ and [OIII] EWs $\approx$ 850 Angstrom (up to 3000 Angstrom), young ages (~100 Myr), a highly excited interstellar medium ([OIII]/H$β\approx6$) and low dust attenuations. These high EWs are very rare in the local Universe, but we show they are ubiquitous at $z\sim6$ based on the measured number densities. The stacked spectrum reveals H$γ$ and [OIII]$_{4364}$ which shows that the galaxies are typically dust and metal poor (E(B-V)=0.1, 12+log(O/H)=7.4) with a high electron temperature ($2\times10^4$ K) and a production efficiency of ionising photons ($ξ_{\rm ion}=10^{25.3}$ Hz erg$^{-1}$). We further show the existence of a strong mass-metallicity relation. The young highly ionising stellar populations, moderately low metallicities, low dust attenuations and high ionisation state in z~6 galaxies conspire to maximise the [OIII] output from galaxies, yielding an [OIII] luminosity density at z~6 that is significantly higher than at z~2, despite the order of magnitude decline in cosmic star formation. Thus, [OIII] emission-line surveys with JWST prove a highly efficient method to trace the galaxy density in the epoch of reionization.

Figures (28)

• Figure 1: Demonstration of the JWST/NIRCam imaging and grism data and the continuum-filtering efficiency in a small $16\times16$ arcsec$^2$ region that constitutes 0.3 % of the data in the J0100+2802 field. The left panel shows a false-color composite of the F115W/F200W/F356W imaging and highlights the locations of two [OIII] emitting systems identified in our data. These are particularly red due to the strong line-emission that falls in the F356W filter. The middle panel shows the dispersed grism image on the same sub-region, while the right panel shows the result of our continuum-filtering methodology which reveals various emission-lines detected in the data.
• Figure 2: Example emission-line 1D spectra of three representative [OIII] emitters in our sample. Vertical dotted lines highlight the locations of H$\beta$ and [OIII]$_{4960,5008}$. Shaded regions show the noise level. The integrated S/N of H$\beta$ ([OIII]$_{4960}$) in each panel are 13.8 (38.3), 4.5 (10.8) and 3.3 (7.4), respectively.
• Figure 3: Example false-color F115W/F200W/F356W stamps of regions where we detect multiple [OIII] emitting systems within 2$"$, highlighting the diversity of these groups. The images are oriented with the position angle 236 degrees. Each horizontal dashed line marks an [OIII] emitting component that is resolved in the grism data.
• Figure 4: The cumulative distribution of the number of object pairs as a function of their angular separation, normalised to the maximum number of pairs. The red symbols show the distribution of separations for all [OIII] emitters, the purple symbols show the separations for pairs of [OIII] emitters that have velocity differences less than $1000$ km s$^{-1}$ and the black symbols for the full source catalog. Error bars represent poisson noise. The grey dashed line shows the expectation for a random distribution. The [OIII] emitters show a significant excess in the number of pairs below a scale of 2$"$, which corresponds to $\approx10$ kpc. All [OIII] pairs within 2$"$ also happen to be closely separated in redshift ($\Delta v = 600$ km s$^{-1}$ at max).
• Figure 5: Overview of the information that we can measure for two example [OIII] emitters identified in the JWST data. For each object, we fit the spectral energy distribution using a composite stellar and nebular emission and dust attenuation model (see § \ref{['sec:sed']}) to the three photometric data-points and the H$\beta$ and [OIII] emission-line fluxes. Open squares show the modeled flux in the F115W, F200W and F356W filters, respectively, while black squares show the measured photometry and its uncertainties. The main parameters that we derive from the SED models are the UV luminosity, the stellar mass and the emission-line EWs. The false-color stamps reveal the diverse morphologies of the [OIII] emitters whereas the generally red colors highlight the regions within the systems with strong line-emission. We display the 2D continuum-filtered spectra in both modules A and B (when available). The modules have opposite dispersion directions, where module B mirrors the image in the spatial direction. This is clearly illustrated in the spectrum of the object on top. Red and orange lines highlight the locations of H$\beta$ and [OIII], respectively. Both objects in this Figure are identified as groups with three line-emitting components.