JWST/NIRSpec Reveals a Small Population of Dominant Dust-Obscured Ionizing Sources in Galaxies at 1 < z < 3

TL;DR

This study tests the reliability of rest-frame near-IR line diagnostics, Fe2S3-$\alpha$ and Fe2S3-$\beta$, against traditional rest-frame optical BPT classifications for 55 galaxies at $1<z<3$, highlighting dust as a key source of misclassification in optical tracers. Using JWST/NIRSpec medium-resolution spectra, Pa lines, and high-ionization tracers, the authors identify 89% agreement between near-IR and optical diagnostics, with six galaxies showing disagreements that are best explained by dust geometry or unusual abundances (e.g., elevated N/Fe). The analysis uncovers a small population of dust-obscured ionizing sources, including cases where AGN activity is hidden in the optical but revealed in the near-IR, and vice versa, suggesting complex interplay between AGN and star formation. The work demonstrates the value of rest-frame near-IR diagnostics, combined with multiwavelength data, to more accurately decompose ionizing photon budgets at cosmic noon, and it points to spatially resolved JWST observations as the next step to disentangle AGN and stellar contributions.

Abstract

Rest-frame optical emission line diagnostics are often used to help classify ionizing sources within galaxies. However, rest-frame optical tracers can miss sources with high dust attenuation, leading to misclassification of the dominant ionizing source. Longer wavelength tracers, such as those in the near-infrared, carry the power to diagnose ionizing sources while being more robust than optical tracers to the presence of dust. The diagnostics used in this work employ the ratios of bright near-infrared emission lines [S III] 9530, [Fe II] 12566 and [Fe II] 16443 to Paschen lines in publicly-available JWST/NIRSpec MSA medium-resolution spectroscopy of 55 galaxies at z < 3. We compare the rest-frame near-infrared and rest-frame optical diagnostics and find that ~90% of our sample have consistent classifications across wavelengths (49/55), while the remaining sources can be explained through ionizing radiation obscured by dust and/or elevated N/Fe abundances. We identify three objects classified as star-forming in the rest frame optical and as active galactic nuclei (AGN) in the rest-frame near-infrared, which we interpret as obscured AGN. We also identify three objects which are classified as AGN in the rest-frame optical and star forming in the rest-frame near-infrared. We interpret two of these objects as AGN with obscured star formation and the other with elevated N/Fe. We discuss how future spatially-resolved and/or mid-infrared spectroscopy can test the relative contributions of AGN and stars to the ionizing photon budgets of these disagreeing sources.

Figures (4)

• Figure 1: Rest-frame one-dimensional spectra of 7 galaxies in our sample with an offset for visualization. Each spectrum has its rest-frame optical and near-IR images on the left and its DJA ID and redshift located above and to the right of the images. We mark the locations of several prominent emission lines and note those used in the following figures in bold.
• Figure 2: Stacked redshift distribution of galaxies in our sample. Blue represents galaxies with the [N2]-BPT ([O3] $\lambda5007$, H$\beta$, [N2] $\lambda6583$ and H$\alpha$) and Fe2S3-$\beta$ lines ([S3] $\lambda9530$, Pa$\gamma$, [Fe2] $\lambda12566$ and Pa$\beta$). Green indicates galaxies with the [N2]-BPT and Fe2S3-$\alpha$ lines ([S3] $\lambda9530$, Pa$\gamma$, [Fe2] $\lambda16443$ and Pa$\alpha$). Yellow indicates galaxies with all the lines from all three diagnostics detected. The details of our sample selection are outlined in Section \ref{['sec:data:fluxes']}.
• Figure 3: [N2]-BPT diagram (upper left) and rest-frame near-IR diagnostics Fe2S3-$\beta$, Fe2S3-$\alpha$ (bottom panels). Optical star-forming galaxies and AGNs are green and blue marks, respectively. Sources in [N2]-BPT and Fe2S3-$\alpha$ are diamonds, [N2]-BPT and Fe2S3-$\beta$ are squares, and those in all three are circles. X-ray and mid-IR detected sources have extra thin crosses and hollow crosses, and the marks of AGNs are more opaque and thick than not AGNs. In the [N2]-BPT diagram, the grey dashed line and solid line are the maximum starburst galaxies line from Kewley2001t and the demarcation curve between starburst galaxies and AGNs from Kauffmann2003. In the Fe2S3-$\beta$ and Fe2S3-$\alpha$ diagram, the grey solid line and dotted line are the starburst limits and the maximum AGN models from Calabro2023.
• Figure 4: The relation between Pa$\beta$ and the Balmer decrement for the sources involved in near-IR diagnostics. Green and blue marks with different shapes indicate the same categories as the Figure \ref{['fig: Diagnostic']} . The dotted, dashdot, and dashed grey lines indicate the expected ratios using intrinsic Case B ratios of $\hbox{{\rm H}$\alpha$}/\hbox{{\rm H}$\beta$} = 2.86$ and $\hbox{{\rm Pa}$\beta$}/\hbox{{\rm H}$\alpha$} = 1/17.6$, and Calzetti_2000, Gordon_2003, and Fitzpatrick_1999 attenuation models. The grey shaded region represents line ratios inconsistent with Case B recombination for $n_e = 10^2~\mathrm{cm}^{-3}$ and $T = 10^4$ K Osterbrock2006. There are 5 sources located in the grey shaded region, even for 3$\sigma$ uncertainties, can only be accessed with non case B recombination. Most of our sources are consistent with the expectation within 1$\sigma$ uncertainty.