Fast Rotators at Cosmic Noon: Stellar Kinematics for 15 Quiescent Galaxies from JWST-SUSPENSE

TL;DR

This study uses ultra-deep JWST-SUSPENSE NIRSpec/MSA spectroscopy, combined with high-resolution imaging, to recover intrinsic stellar kinematics for 15 massive quiescent galaxies at $z\sim1.2-2.3$ via a forward-modelling framework. They constrain rotation for 10 galaxies, finding $V_{r_e}$ in $114-336$ km s$^{-1}$ and $(V/\sigma)_{r_e}$ up to $1.57$, implying that quenching did not erase disc structures and that younger quiescent populations are more rotation-supported. Dynamical masses imply substantial dark matter within $1\,r_e$ and potential compatibility with a bottom-heavy IMF, supporting a evolutionary scenario where distant quenched galaxies grow their outskirts and gradually lose rotation through predominantly minor mergers to become the massive slow rotators seen today. The work also highlights the importance of aperture corrections in dynamical mass estimates from NIRSpec/MSA spectra and outlines future steps (e.g., JAM modelling, IFU data) to refine constraints on IMF and dark matter in high-redshift galaxies.

Abstract

We present spatially-resolved stellar kinematics of 15 massive ($M_*=10^{10.5-11.5}M_{\odot}$) quiescent galaxies at $z\sim1.2-2.3$ from the JWST-SUSPENSE program. This is the largest sample of spatially-resolved kinematic measurements of quiescent galaxies at cosmic noon to date. Our measurements are derived from ultra-deep NIRSpec/MSA stellar absorption line spectra, using a forward modelling approach that accounts for optics, source morphology, positioning, and data reduction effects. 10 out of 15 galaxies are orientated such that we can measure rotational support. Remarkably, all 10 galaxies show significant rotation ($V_{r_e}=117-345$km/s, $σ_0 = 180-387$km/s), and are classified as "fast rotators" from their spin parameter. The remaining galaxies are too misaligned with respect to the slit to constrain their rotational velocities. The widespread rotational support in our sample indicates that the process responsible for quenching star formation in early massive galaxies did not destroy rotating disc structures. When combined with other quiescent galaxy samples at $z\sim0.5-2.5$, we find a trend between rotational support and age, with younger quiescent galaxies being more rotationally supported. This age trend has also been found at $z\sim0$, and likely explains why our high-redshift galaxies show more rotational support compared to massive ETGs at $z\sim0$, which are, on average, older. Our kinematic modelling also enables us to calculate dynamical masses. These dynamical masses greatly exceed the stellar masses for our sample (median $M_{\text{dyn}}/M_*=2.7$); they even allow for the bottom-heavy IMF found in the cores of low-$z$ massive ellipticals. Altogether, our results support a scenario in which distant quiescent galaxies evolve into nearby massive ETGs, gradually building up their outskirts and simultaneously losing rotation, due to a series of (mostly minor) mergers.

Figures (15)

• Figure 1: LOS velocity profiles for a model galaxy that is perfectly aligned and centred (left, black circles), a model galaxy that is perfectly centred but misaligned (middle, dark grey crosses), and a model galaxy that is offset and misaligned (right, light grey squares) with respect to the central micro-shutter. The left panels show the models for a galaxy with an intrinsic velocity field with a maximum rotational velocity of 300 km s$^{-1}$. The right panels show the LOS velocity profiles for a galaxy without intrinsic rotational velocity, representing the effects from optics, the PSF, and data reduction steps. These effects can lead to observed LOS velocities of up to $\sim50$ km s$^{-1}$. The modelled galaxies have $r_e = 0.13$ (1.1 kpc at $z\sim1.5$), a Sérsic index of 1.5, and an axis ratio $q$ of $0.3$. The grey bands indicate the areas of the velocity profile that are affected by bar shadows from the MSA.
• Figure 2: 2D absorption line profiles of Ca ii K (left) and H (middle), and the best-fit model (right) for two example galaxies. The red points (lines) show the best-fit velocity (velocity dispersion) profile from ppxf that was used as input for our forward modelling in Section \ref{['sec:forward_model']}. We note that the ppxf fits were obtained from the entire wavelength range of the spectra, and the two absorption lines illustrated in this figure serve as an example.
• Figure 3: Observed kinematics and best-fit models for the ten distant quiescent galaxies for which we can constrain rotational velocities. In the left panels, we show the observed velocities (dispersions) as the black (orange) data points, and the best-fit models are represented by the corresponding lines. In the top right panels we show NIRCam RGB imaging CCasey2023 where available, or HST F814W imaging otherwise NScoville2007. In the bottom right panels we show the inferred 2D line of sight velocity fields. We overlay the MSA microshutter positions in both panels on the right.
• Figure 3: Continued.
• Figure 4: $V_{r_e}/\sigma_0$ as a function of stellar mass for the 15 distant quiescent galaxies in our sample (squares). The grey squares indicate galaxies for which we can only obtain lower limits to $V_{r_e}/\sigma_0$. We also show quiescent galaxies at $z\sim0.8$ from LEGA-C JvanHoudt2021 and three lensed quiescent galaxies at $z\sim2$ANewman2018b. The dashed line indicates a ratio of 1, corresponding to the definition of rotational support. The symbols are coloured by $V_{r_e}$.