MSA-3D: Connecting the Chemical and Kinematic Structures of Galaxies at $z \sim 1$

TL;DR

This study probes how the dynamical state of star-forming galaxies at $z\sim1$ governs their gas-phase metallicity gradients. Using JWST/NIRSpec MSA slit-stepping (the MSA-3D survey), the authors obtain spatially resolved kinematics and metallicity gradients for a representative sample, deriving $v/\sigma$ at $1.5R_e$ and comparing to gradients measured via the PP04 N2 calibration. They find a moderate anti-correlation between metallicity gradients and $v/\sigma$ ($r=-0.43$, $p=0.05$) with a slope of about $0.005$ dex dex$^{-1}$, and a stronger anti-correlation with $R_e/\sigma$ ($r=-0.59$, $p=0.005$), suggesting radial mixing (tied to the mixing timescale) as the primary regulator of chemical stratification. The gradients are uniformly shallow, consistent with efficient turbulent mixing in dynamically settled disks and aligning with FIRE-2 predictions, underscoring the pivotal role of disk dynamics in shaping chemical structure at $z\sim1$.

Abstract

We investigate the connection between ionized gas kinematics and gas-phase metallicity gradients in 21 star-forming galaxies at $0.5 < z < 1.7$ from the MSA-3D survey, using spatially resolved JWST/NIRSpec slit-stepping observations. Galaxy kinematics are characterized by the ratio of rotational velocity to intrinsic velocity dispersion, $v/σ$, measured at $1.5\,R_e$, where $R_e$ is the effective radius. We find that dynamically hotter disks exhibit systematically flatter metallicity gradients, with a moderate anti-correlation between metallicity gradient and $v/σ$ (Pearson $r=-0.43$, $p=0.05$) and a linear fit yields a slope of $\sim 0.005$ dex per dex in $v/σ$, weaker than the dependence on stellar mass. A significantly stronger anti-correlation is observed with $R_e/σ$, interpreted as a proxy for the radial mixing timescale ($r=-0.59$, $p=0.005$), indicating that cumulative radial mixing more directly regulates chemical stratification. The metallicity gradients in our sample are uniformly shallow, indicating that efficient turbulent mixing in kinematically settled disks regulates the chemical structure of typical star-forming galaxies at $z\sim1$.

Figures (24)

• Figure 1: Two-dimensional surface brightness modeling of galaxy ID 8942 using GALFIT. The top row shows the JWST/NIRCam F444W imaging; the bottom row shows the HST/WFC3 F160W imaging. From left to right, the panels display the observed image, the best-fit Sérsic model, and the residual map after model subtraction. The fitting region is $\sim 2\hbox{$^{\prime\prime}$} \times 2\hbox{$^{\prime\prime}$}$ box. The same color scale is used in all panels. The photometric centroids and inclinations derived from the fits are used to extract the kinematic profiles.
• Figure 2: H$\alpha$ velocity fields of the resolved MSA-3D galaxies at $0.5<z<1.7$ shown at their approximate locations in the SFR-M$_*$ plane. The solid line indicates the star-forming main sequence from Whitaker2014 at $z \sim 1$, while the dashed and dotted lines show offsets by factors of ×4 and ×10.
• Figure 3: Example of kinematic modeling for galaxy ID 8942. From left to right: the composite imaging cutout (used for GALFIT modeling), the observed H$\alpha$ velocity field, the best-fit velocity model assuming a rotating disk, and the residual map (data minus model). The observed velocity field exhibits a rotation-dominated morphology with a spider-like pattern. The residual map confirms the quality of the fit, with a reduced $\chi^2 = 1.03$. The gray solid lines indicate the morphological PAs, while the gray dashed lines show the PAs derived from the kinematic models, with a difference of 7.58$^\circ$.
• Figure 4: Comparison of the $v/\sigma$ values obtained by fixing the inclination and by leaving it as a free parameter. The two measurements show good agreement, with most points lying close to the one-to-one relation. Galaxies with $q>0.6$ ($i<53^\circ$) are marked with red circles (10 objects) and show the largest discrepancies between the two fitting approaches, resulting in strong deviations from the one-to-one relation.
• Figure 5: Intrinsic velocity dispersion $\sigma_0$ as a function of redshift. Blue filled circles represent the MSA-3D sample (this work). Galaxy 8512 is shown as a blue point with a red edge in this work, while the purple hexagon indicates the result from Ivana2025. Other samples shown for comparison are KMOS$^{\rm 3D}$Wisnioski2015, and CO-based measurements from CO-PHIBSS Tacconi2013 and Swinbank2011. The gray dashed and solid curves indicate average trends for ionized gas and molecular gas, respectively, from Ubler2019. The gray band shows predictions from a simplified Toomre stability model Wisnioski2025, assuming marginally stable disks with $\log(M_*/M_\odot)=9.5-$10 and constant rotational velocity. The upper and lower bounds correspond to $v_{\rm obs}=110-$202 km/s. Our sample generally lie slightly above the dashed line, which likely reflects systematic differences in measurement methodology.