MSA-3D: Uncovering Weak AGNs and Resolved Outflows in Disguise in $z\sim1$ Star-Forming Galaxies

TL;DR

This study uses JWST/NIRSpec MSA-3D slit-stepping to create IFU-like datacubes for 38 star-forming galaxies at $0.5<z<1.7$, enabling spatially resolved maps of $[\mathrm{NII}]/\mathrm{H}\alpha$, $[\mathrm{SII}]/\mathrm{H}\alpha$, $[\mathrm{OIII}]/\mathrm{H}\beta$, and $[\mathrm{NII}]/[\mathrm{SII}]$. Integrated diagnostics place most galaxies in the star-forming regime, but spaxel-level analyses reveal six LLAGN/outflow candidates with localized high excitation and elevated $V_{\mathrm RMS}$, indicating weak AGN activity and/or shocks. Outflow properties are modest ($\dot{M}_{out}\sim1-4\ M_\odot\,\mathrm{yr^{-1}}$, $\dot{KE}_{outflow}\sim(0.7-1.3)\times10^{41}\ \mathrm{erg\,s^{-1}}$) and correspond to $\sim0.1-1\%$ of the inferred $L_{AGN}$, aligning with low-energy tails of known AGN-driven winds and the $\dot{M}_{out}-L_{AGN}$ trend across cosmic time. The analysis highlights both the promise and limitations of detecting faint feedback at $z\sim1$ with spatially resolved spectroscopy, including completeness limits set by luminosity thresholds ($L_{AGN}\gtrsim10^{43}\ \mathrm{erg\,s^{-1}}$) and spatial extent requirements. Overall, the work shows that resolved spectroscopy uncovers hidden AGN/shock signatures and clarifies how excitation, metallicity, and gas kinematics co-evolve in typical $z\sim1$ star-forming galaxies.

Abstract

We present spatially resolved rest-optical spectroscopy of 38 star-forming galaxies at 0.5 < z < 1.7 from the JWST/NIRSpec MSA-3D survey, which uses slit-stepping to build IFU-like datacubes at 0.1'' resolution. We map emission-line morphology, excitation, and kinematics of the warm ionized gas using [N II]/H$α$, [S II]/H$α$, and [O III]/H$β$. Relative to z$\sim$0 galaxies at fixed stellar mass, our sources show systematically lower [N II]/H$α$ and [S II]/H$α$ and elevated [O III]/H$β$, consistent with harder radiation fields and lower metallicities. Radially, [O III]/H$β$ profiles are typically flat or mildly positive, whereas [N II]/H$α$ also remains flat or declines outward, mirroring metallicity trends. On kpc scales, we find a strong positive correlation between [N II]/H$α$ and velocity dispersion ($σ$), linking local excitation to turbulent or shock-driven kinematics. Six galaxies ($\sim$ 16% of the sample) host spatially localized regions with elevated [N II]/H$α$, high EW(H$α$), and V_RMS = $\sqrt{V^2 + σ^2} > 200$ km/s, indicative of weak AGN activity, shocks, or outflows. For these candidates we infer modest warm-ionized outflow rates of 1-4 Msun/yr and kinetic powers $\sim$ 0.1-1% of the AGN bolometric luminosity (from central [O III] or H$α$). These values place our sample at the low-energy tail of known AGN-driven outflows yet in continuity with $\dot{M_{out}}-L_{AGN}$ scaling relations across 0 < z < 6. A completeness assessment shows MSA-3D is sensitive to AGN with $L_{AGN} \geq 10^{43}$ erg/s, underscoring both the promise and current limitations of detecting weak AGN activity in distant galaxies with resolved spectroscopy.

Figures (16)

• Figure 1: [Left] Star formation rate (SFR) versus stellar mass ($\rm M_{\star}$) for MSA-3D galaxies (magenta hexagons), derived via multi-wavelength SED fitting koekemoer11stefanon17. For comparison, we show the star-forming main sequence (SFMS) relations at $0.5 < z < 1$ (cyan) and $1 < z < 1.5$ (blue) from whitaker14, along with the $z \sim 0$ SDSS galaxy population as grey 2D historgram. Our sample is consistent with the SFMS between z$\approx 0.5-1.5$. [Right] Emission–line excitation, traced by $\log([\mathrm{N\,II}]\,\lambda6584/\mathrm{H}\alpha)$, (the horizontal axis of the WHAN diagram) $vs.$ gas velocity dispersion ($\sigma$), for MSA–3D galaxies. Both quantities are measured from integrated spectra co–added within $2R_e$ apertures. Only systems with [N ii] and H$\alpha$ detected within the observed spectral window are shown (colored circles). They are color coded by stellar mass. Overplotted for reference are contours for SDSS galaxies at $z \approx 0$ (black) and MOSDEF galaxies at $1.4 < z < 1.7$ (yellow) and $2.0 < z < 2.7$ (green). They show almost no correlation in their integrated measurements.
• Figure 2: [Left] Integrated BPT diagram for MSA-3D galaxies (magenta hexagons), where emission line fluxes are summed within 2$R_e$. 13 out of the 38 galaxies in our sample have wavelength coverage of all four emission lines required for this diagram. For comparison, we overlay contours from SDSS (z $\sim$ 0), and include high-redshift star-forming samples from AURORA shapley25, CEERS sanders23, and SINS/zC-SINF schreiber19. Theoretical and empirical demarcation lines from kewley01schawinski07kauffmann03 are shown in blue, red and dashed black, respectively. MSA-3D galaxies fall below the AGN boundaries and are consistent with ionization predominantly by young stars in relatively low-metallicity environments from their integrated measurements. [Right] WHAN diagram cidfernandes11 showing log([NII]/H$\alpha$) $vs.$ H$\alpha$ Equivalent Width for a larger subset of the sample (31 sources) since G140H/100LP grating captures $\rm H\alpha + [NII]$ for $\sim81\%$ of our sample. Nearly all of our galaxies lie well above the EW(H$\alpha$) = 3Å threshold and within the star-forming region, with two sources (MSAID: 11843 and 13416) crossing the AGN/Seyfert classification boundary.
• Figure 3: Spatially resolved emission-line diagnostics for the MSA-3D sample. (Left) BPT diagram and (Right) WHAN diagram for individual spaxels for all MSA-3D galaxies. Displayed spaxels satisfy $\mathrm{S/N}>3$ for each of H$\alpha$, H$\beta$, [N ii] $\lambda6584$, and [O iii] $\lambda5007$, and any spaxels below this threshold in any line are masked. The flux measurements are obtained from line fitting, following the procedure described in §\ref{['subsec:fitting']}. Spaxels are color-coded by their excitation mechanism based on the kewley01kauffmann03 diagnostics: star formation or HII-region like in blue, AGN and Seyfert-like in red and composite ionization or LINER source in magenta. While the majority of spaxels occupy the star-forming locus, a non-negligible subset of spaxels extend into the composite/LINER and AGN-like regimes. This indicates that AGN, shocks, and mixed ionization sources can be missed in integrated spectra. Sensitive, spatially resolved spectroscopy is required to detect and disentangle these components.
• Figure 4: Example two-dimensional maps for two representative galaxies (one in each row) from the MSA-3D sample, demonstrating the spatially resolved data quality achieved with the NIRSpec slit-stepping strategy. For each galaxy (top: ID 2465; bottom: ID 13416), the panels from left to right show: (1) the HST/WFC3 F160W image; (2) the rest-optical continuum map derived from the IFU datacube by masking strong emission lines and summing along the spectral axis; (3) the H$\alpha$ flux map; and (4) the [NII]$\lambda$6584 flux map. Emission–line fluxes are measured by Gaussian fits to individual spaxels that satisfy our S/N detection thresholds for all required lines, as detailed in §\ref{['subsec:fitting']}. Overplotted contours trace the stellar morphology from the HST image, to enable direct spatial comparison. These examples highlight the morphological diversity in the ionized gas distribution across the MSA-3D sample. We encounter galaxies with well-ordered, disk-like H$\alpha$ emission aligned with the stellar distribution (similar to 2465), while a few sources show clumpy and offset ionized gas structures (13416). These reflect variations in gas excitation, kinematics, or dust geometry across the galaxy.
• Figure 5: Radial profile of log([NII]/H$\alpha$) emission line ratio for a representative galaxy from our sample. We compute the deprojected radial distances for each spaxel and normalize by the effective radius ($R/R_e$). Gray points indicate the line ratio measurements for each individual spaxel, while magenta hexagons denote the mean values of the nebular ratio, computed within annular radial bins of equal width. We adopt this binned radial profile methodology, described in §\ref{['radialbin']}, to characterize radial variations in ionization and abundance-sensitive line ratios.