GA-NIFS: Understanding the ionization nature of EGSY8p7/CEERS-1019. Evidence for a star formation-driven outflow at z = 8.6

TL;DR

This study uses JWST/NIRSpec IFS to obtain spatially resolved spectroscopy of EGSY8p7/CEERS-1019 at $z\approx8.68$, quantifying its ISM properties and testing the ionization source. The data reveal a broad [OIII] component (FWHM $\sim$650 km s$^{-1}$) and a spatially resolved outflow between two main star-forming clumps, while the Hβ broad component remains inconclusive in origin. Photoionization and shock modeling favor a star-formation–driven ionization mechanism over an AGN, and the derived electron density ($n_e\sim3.6\times10^3$ cm$^{-3}$) and gas-phase metallicity ($12+\log(O/H)\approx8.02$) indicate dense, metal-poor conditions typical of early galaxies. The outflow energetics imply that stellar feedback alone can power the observed gas motions, with radiation pressure contributing to momentum transfer, and the low escape fraction suggests gas retention and limited IGM enrichment at this epoch. These results demonstrate that intense, star-formation–driven feedback shapes the ISM structure in early massive galaxies and highlight JWST’s capability to resolve feedback processes at cosmic dawn.

Abstract

Understanding the physical conditions and feedback mechanisms in early massive galaxies is essential to uncover how they formed and evolved during the first billion years of the Universe. In this context, the galaxy EGSY8p7/CEERS-1019 at z=8.6 provides an excellent benchmark, given its stellar mass of $10^{9.3}M_\odot$ and elevated N/O abundance despite its sub-solar metallicity. In this study, we present new JWST/NIRSpec observations offering the first spatially resolved spectroscopy of this galaxy, with higher sensitivity and spectral resolution than previous studies. We identify broad (FWHM=650km/s) H$β$ and [OIII] emission components whose emission is located between the two rest-frame UV clumps of the galaxy and extended over a distance of $\sim1kpc$. The morphology and kinematics of these components indicate that the broad emission arises from outflowing gas rather than from an AGN broad-line region. The kinetic energy injection rate from stellar feedback is an order of magnitude higher than that of the outflow, while the radiation pressure rate is comparable to the outflow momentum rate. These results suggest that stellar feedback alone can drive the outflow, with radiation pressure potentially providing the required momentum transfer. We derive a low mass-loading factor ($η=0.16$) and ionizing photon escape fraction ($f_{esc}=0.021\pm0.014$). Together with the high electron density measured ($n_e=2200cm^{-3}$), these results support the interpretation that most of the gas remains confined within the galaxy. Comparisons of diagnostic emission-line ratios with photoionization and shock models support a star-formation-driven ionization scenario, ruling out any excitation by AGN radiation. Finally, the absence of detectable Wolf-Rayet features suggests that alternative mechanisms must be considered to explain the high N/O ratio in this galaxy.

Figures (13)

• Figure 1: Top: JWST NIRSpec images from collapsed continuum emission (2.2 $\mu$m < $\lambda$ < 2.4 $\mu$m) in the low-resolution PRISM data cube. The central NIRSpec/MOS shutter from program ERS 1345 used by Larson2023 and Marques-Chaves2024 is overlaid as a red solid rectangle, while the JWST/NIRCam F150W contours are shown with blue solid lines. The central point of the extraction aperture used in the following sections is marked with a black cross. Bottom: NIRSpec/MOS data (blue) and NIRSpec/IFS spectrum (black) extracted from a rectangle region as large as the MOS shutter (see text). A best IFS fit-model with two component (tied in velocity and velocity dispersion) for each line is shown in the bottom panel.
• Figure 2: JWST/NIRSpec PRISM and G395H/F290LP spectra of the galaxy (top and bottom panels respectively) extracted from an aperture of R= 0.25 arcsec. In the top panel, the gray shading areas are those excluded in the continuum fitting (see text), the red dashed line shows the power-law continuum model fitted to reproduce the continuum emission, and the solid orange line is the one proposed by Marques-Chaves2024. The brightest nebular lines are marked.
• Figure 3: Spatially resolved fitting of the [OIII]$\lambda$ 5007 Å emission line in the high-resolution data cube (G395H). The narrow and broad emission line flux maps are shown in the top and bottom panels, respectively. The JWST/NIRCam F150W and the emmision map shown contours are shown with blue and white solid lines respectively.
• Figure 4: Emission-line diagnostic diagrams together with theoretical models specifically computed for the nebular properties derived for this source. Photoionization models with a salpeter1955 and a top-heavy IMF kauffmann2024 are shown with blue dots and pink triangles, respectively, for two ionization parameters. Shock models for two different magnetic fields are shown with purple diamonds. The transparency of the colors decreases with age and velocity in the SF and shock models, respectively. Finally, different ionization models with power laws and absorbed AGNs are shown with yellow and grey symbols. Division lines between AGN and SF photoionization from Backhaus2022 and Mazzolari2024 are shown with dashed lines in the panels.
• Figure 5: Electron density $n_{\rm e}$ as a function of redshift. The dotted and dashed curves represent $n_{\rm e}$$\sim$ (1+z)$^{1}$ and $\sim$ (1+z)$^{2}$ respectively. A sample of star-forming galaxies at z = 4–9 observed by the JWST/NIRSpec GLASS, Early Release Observations, and CEERS programs are shown for comparison purposes Isobe2023. Also the means values from Kaasinen2017 and Sanders2016 at lower redshifts. Finally, a sample of GA-NIFS star-forming galaxies is shown Scholtz2025Lamperti2024Marconcini2024delpino2024.