Spatially Resolved AGN Ionization and Star Formation at Cosmic Noon with JWST/JEMS

Abstract

At Cosmic Noon ($z\approx 2-3$), both star formation and Active Galactic Nuclei (AGN) activity peaked, each playing a significant role in ionizing interstellar gas on galaxy-wide scales. The spatial distribution of this ionized gas provides a direct probe of how AGN and stellar ionization shape the gaseous reservoirs of their host galaxies. Using JWST/NIRCam imaging from the JWST Extragalactic Medium-band Survey (JEMS) we spatially map two complementary tracers of ionized gas, [O III]$+\mathrm{H}β$ and Pa$β$, in $\sim200$ galaxies at $2.5 < z < 2.9$. We apply multiwavelength AGN diagnostics to divide the sample into AGN hosts (33 galaxies), Pa$β$-detected systems (32 galaxies), and control objects (175 galaxies). We measure the [O III]$+\mathrm{H}β$ and Pa$β$ spatial extents in each population and relate them to AGN and host properties derived from Spectral Energy Distribution (SED) modeling. Both tracers exhibit systematically larger maximum radial extents in AGN hosts than in control galaxies (by $\sim0.3$ dex), with [O III]$+\mathrm{H}β$ emission modestly more extended than Pa$β$ by $\sim0.1$ dex. With this statistically robust AGN sample, we measure the [O III]$+\mathrm{H}β$ radial extent-AGN luminosity relation at $z\sim3$ and derive a slope of $\sim0.2$, consistent with the shallow end of values reported at low redshift. The larger ionized gas extents among AGN hosts relative to the control sample, combined with the strong correlation between [O III]$+\mathrm{H}β$ extent and AGN luminosity suggest that AGN activity may dominate gas ionization in galaxies with mixed AGN and star-forming activity at Cosmic Noon, although stellar processes can still contribute significantly on kiloparsec scales.

Figures (5)

• Figure 1: Color-redshift plots showing $m(\mathrm{F182M}) - m(\mathrm{F210M})$ (top) and $m(\mathrm{F460M}) - m(\mathrm{F480M})$ (bottom) colors for objects in JEMS as a function of redshift. Points are colored by their y-axis values, with red corresponding to more positive values and blue to more negative values. This demonstrates that [O$\;$]$+\mathrm{H}\beta$ and Pa$\beta$ can be targeted simultaneously in the range $2.52 < z < 2.88$.
• Figure 2: Example EAZY (left panels) and CIGALE (right panels) SEDs for an AGN and control sample galaxy at $z=2.7$ with NIRCam F182M, F210M, F430M, F460M, and F480M filter curves overlaid. At this redshift, [O$\;$]$+\mathrm{H}\beta$ falls into F182M, while Pa$\beta$ is targeted by F460M. The adjacent F210M and F430M filters capture the continuum emission without contamination from other lines.
• Figure 3: JWST/NIRCam observations of [JADES]199443, an AGN at $\mathrm{z}=2.65$. Top Panel: Shows the [O$\;$]$+\mathrm{H}\beta$ emission. First column: RGB image with [O$\;$]$+\mathrm{H}\beta$ emission represented by the green color. Second column: F182M image showing both emission-line $+$ continuum emission. Third column: F210M continuum image. Fourth column: Continuum subtracted [O$\;$]$+\mathrm{H}\beta$ map. Bottom Panel: Shows the Pa$\beta$ emission. First column: RGB image with Pa$\beta$ emission represented by the red color. Second column: F460M image showing both emission-line $+$ continuum emission. Third column: F430M continuum image. Fourth column: Continuum subtracted Pa$\beta$ map.
• Figure 4: Example of size calculations for [JADES]199443 ($z=2.65$), an AGN with three discrete [O$\;$]$+\mathrm{H}\beta$ features at a surface brightness limit of $8.7\times10^{-17}$ erg s$^{-1}$ cm$^{-2}$. The [O$\;$]$+\mathrm{H}\beta$ emission-line map is shown in the first column and each of the three features are shown in columns 2-4 in order of decreasing area. The average radial size of each features is shown by a black circle and the maximum radial size is shown by a red square. For features that are offset from the galaxy's center, we plot a light blue 'x' at the center and a black 'x' at the feature's centroid. For these offset features, we also plot the maximum extent from the galaxy's center to the furthest edge pixel of the feature represented by a light blue star.
• Figure 5: AGNfitter’s best-fit SED in the observed frame for [JADES]199443 using the AGN and galaxy components described in Section \ref{['subsec:SED_modeling']}. The cold-dust emission from star formation is shown in green, and the stellar emission is shown in yellow. The accretion-disk component is modeled in blue, and the hot-dust torus emission is shown in purple. The width of each curve reflects the uncertainty in that component. The red line shows the total model fit, with residuals displayed in the bottom panel as red points. The total model is computed by summing the four emission components at each wavelength. Observed photometry is plotted as black points, including medium- and wide-band measurements from HST/ACS, JWST/NIRCam, and JWST/MIRI (downward arrows indicate MIRI upper limits at $\mathrm{SNR}<3$). This object does not have Chandra or VLA observations. For this AGN, the AGNfitter solution favors strong stellar emission and a weak accretion-disk component in the rest-frame UV–optical, while at mid-IR wavelengths the model finds that a prominent torus component best reproduces the MIRI data. We note that an independent SED fit with CIGALE yields consistent results for this object, including a similarly strong torus component at mid-IR wavelengths.