Reconstructing the Assembly of Massive Galaxies. III: Quiescent Galaxies Loose Angular Momentum as They Evolve in a Mass-dependent Fashion

We study the evolution of stellar kinematics of a sample of 952 massive quiescent galaxies with $M_>10^{10.5}M_\odot$ at $0.6<z<1$. Utilizing spatially integrated spectroscopy from the LEGA-C survey, we focus on the relationship between the observed integrated stellar velocity dispersion ($σ^\prime_{star}$) and the morphological axial ratio ($q$), and its variation with the stellar age and mass of quiescent galaxies. For the youngest quiescent galaxies, regardless of stellar mass, $σ^\prime_{star}$ decreases with increasing $q$, a trend that is consistent with a system having significant rotation and hence suggests that massive galaxies still retain significant amount of angular momentum in the aftermath of quenching. As they continue to evolve, the variation of the $σ^\prime_{star}$-$q$ relationship depends on stellar mass. For quiescent galaxies with $M_<10^{11.3}M_\odot$, $σ^\prime_{star}$ decreases with $q$ in all stellar-age bins, suggesting that the quiescent populations of this mass regime retain significant rotation even long time after quenching. In contrast, for more massive quiescent galaxies with $M_*>10^{11.3}M_\odot$, the relationship between $σ^\prime_{star}$ and $q$ becomes significantly flattened with increasing stellar age. This indicates that, as the very massive galaxy populations continue to evolve after quenching, angular momentum gradually reduces, which eventually transforms them into velocity-dispersion supported systems. We suggest that incoherent, continuous merging and accretion events onto the galaxies are the main drivers of the observed mass-dependent, posting-quenching dynamical evolution, because more massive galaxies are more likely to undergo such interactions. We are witnessing the early formation epoch of fast and slow rotators at $z \sim 0.8$, when the Universe was only half of its age nowadays.

💡 Research Summary

This paper investigates how the stellar kinematics of massive quiescent galaxies evolve after quenching, focusing on the loss of angular momentum as a function of galaxy mass. Using the Large Early Galaxy Astrophysics Census (LEGA‑C) DR3, the authors select a clean sample of 952 quiescent galaxies at 0.6 < z < 1 with stellar masses M* > 10¹⁰·⁵ M⊙, reliable morphological classifications (FLAG_MORPH = 0), and no AGN contamination (FLAG_SPEC = 0). Morphological axial ratios (q = b/a) are measured from HST/ACS I₈₁₄ imaging via single‑Sérsic fits, while integrated stellar velocity dispersions (σ′₍star₎) are obtained from the pPXF fitting of the spatially integrated LEGA‑C spectra. σ′₍star₎ is defined as the quadrature sum of the intrinsic random motion (σ) and the line‑of‑sight contribution from ordered rotation (σ_rotation). In a rotating system σ_rotation decreases as the galaxy is viewed more face‑on (higher q), causing σ′₍star₎ to decline with increasing q; a dispersion‑dominated system shows little or no σ′₍star₎–q trend. Thus the slope of the σ′₍star₎–q relation serves as a proxy for the V/σ ratio.

Stellar ages are derived through full‑spectral energy‑distribution (SED) fitting with the Bayesian code Prospector, employing the Flexible Stellar Population Synthesis (FSPS) framework, MIST isochrones, and MILES stellar libraries. The authors adopt a non‑parametric star‑formation history (SFH) with nine look‑back time bins and a Dirichlet prior, which has been shown to recover unbiased ages for high‑z quiescent galaxies. Each galaxy’s mass‑weighted age is used to rank the sample; absolute ages may shift systematically, but relative ordering is robust.

The sample is split at M* = 10¹¹·³ M⊙, the empirical mass that separates fast and slow rotators in the local universe. Within each mass bin, galaxies are further divided into three equal‑size age subsamples (young, intermediate, old) using the 33rd and 67th percentiles of the age distribution. For each of the six subsamples, the authors compute a locally weighted scatterplot smoothing (LOWESS) curve of σ′₍star₎ versus q, thereby quantifying the median σ′₍star₎–q trend.

Key results:

1. Low‑mass quiescent galaxies (M < 10¹¹·³ M⊙):* All three age groups exhibit a clear negative σ′₍star₎–q slope. The youngest galaxies show the steepest decline, but even the oldest retain a measurable trend, indicating that rotation remains a significant component of their dynamical support long after quenching.

2. High‑mass quiescent galaxies (M > 10¹¹·³ M⊙):* Young galaxies also display a negative σ′₍star₎–q slope, consistent with substantial rotation. However, as stellar age increases the slope flattens dramatically; the oldest high‑mass galaxies show virtually no dependence of σ′₍star₎ on q. This flattening signals a progressive loss of ordered rotation, culminating in a dispersion‑dominated (slow‑rotator) dynamical state.

The authors interpret these mass‑dependent trends as evidence that post‑quenching dynamical evolution is driven primarily by continued merging and accretion. Massive galaxies reside in denser environments and experience more frequent dry mergers, which efficiently randomize stellar orbits and diminish angular momentum. Lower‑mass galaxies, experiencing fewer and less disruptive interactions, preserve their rotational structure.

Systematic uncertainties are addressed: measurement errors in σ′₍star₎ (instrumental resolution, template mismatch), projection effects in q, and assumptions in the SED fitting (IMF, metallicity, dust law) could affect absolute σ′₍star₎ values, but the relative change of the σ′₍star₎–q slope with age remains robust. The study also compares its findings with earlier work at z ∼ 2 (Belli et al. 2017), noting that the larger, uniformly selected LEGA‑C sample provides stronger statistical leverage.

In conclusion, the paper provides direct observational evidence that the dichotomy between fast and slow rotators is already emerging at z ≈ 0.8. While low‑mass quiescent galaxies retain rotation for several gigayears after quenching, high‑mass systems progressively lose angular momentum, likely through continuous dry mergers, and evolve into the slow‑rotator population observed locally. The authors suggest that future integral‑field spectroscopy with JWST, ELT, and upcoming surveys will be essential to map V/σ directly and to test the merger‑driven angular momentum loss scenario quantitatively.