MAGNUS III: Mild evolution of the total density slope in massive early-type galaxies since z$\sim$1 from dynamical modeling of MUSE integral-field stellar kinematics

TL;DR

This study delivers a precise measurement of how the total mass density slope $\gamma_{\rm T}$ in massive ETGs evolves from $z\sim 1$ to the present by applying Jeans Anisotropic Modeling (JAM) to a large, homogeneous dataset. The MAGNUS intermediate-redshift sample ($0.24<z<0.75$; $\sim200$ ETGs) is paired with a statistically matched MaNGA local baseline to constrain $\gamma_{\rm T}$ evolution with a consistent methodology across cosmic time. The results show a mild but robust steepening, with $d\gamma_{\rm T}/dz \approx -0.20 \pm 0.03$ when combining MAGNUS with MaNGA, implying dissipative processes contribute to late-time assembly of massive ETGs and challenging some numerical simulations. Across six dynamical models (varying mass profiles and velocity ellipsoid orientations), the finding remains stable, indicating a genuine physical trend rather than model-dependent systematics. These measurements bridge lensing and dynamical studies, providing a cohesive view of ETG mass structure evolution over the last 6–7 Gyr and setting a stringent benchmark for galaxy formation models.

Abstract

We investigate the total mass density slope evolution in massive early-type galaxies (ETGs) over the last 6.5 billion years ($0 < z < 0.75$). We perform a detailed dynamical analysis of approximately 200 ETGs spanning the redshift range $0.24 < z < 0.75$, utilizing spatially resolved stellar kinematics derived from high signal-to-noise ratio (S/N) MUSE-DEEP spectroscopy and surface brightness models from high-resolution HST imaging. We constrain mass distributions using the Jeans Anisotropic Modeling (JAM) technique coupled with Multi-Gaussian Expansion (MGE) method. To rigorously constrain evolutionary trends, we combine this intermediate-redshift dataset with a local ETG sample ($z \sim 0.05$) from the MaNGA survey. We adopt dynamical constraints for the local sample derived using an identical homogeneous methodology, ensuring a strictly consistent comparison. We found that the total density profiles of the intermediate-redshift ETG sample are approximately isothermal and exhibit a median mass-weighted total density slope, $<γ_{\rm T}>=2.19 \pm 0.01$ at $<z>=0.44$, which is shallower than the local baseline of $<γ_{\rm T}> = 2.26 \pm 0.01$ at $<z>=0.04$. This structural shift corresponds to a redshift gradient of $\mathrm{d} γ_{\rm T}/\mathrm{d} z \approx -0.20 \pm 0.03$, detected at $\sim$5-$σ$ significance. We demonstrate that this trend is robust against model assumptions and persists even when restricting the analysis to high-velocity dispersion systems ($σ_e > 150$ km/s). Our findings are consistent with previous lensing-based studies and in tension with cosmological simulations. The observed steepening suggests that dissipative processes, such as gas-rich accretion and mergers, must play a non-negligible role in the late-stage assembly of massive ETGs.

Figures (32)

• Figure 1: Representative MGE surface brightness models for the MAGNUS sample. The first and third panel display the sky-subtracted HST images. The second and fourth panel show the observed surface brightness isophotes (black contours) overlaid with the best-fit MGE model (red contours). Galaxy IDs are indicated above the corresponding HST images. The MGE models for the remaining galaxies in the sample are provided in Appendix \ref{['sec:all_galaxy_model']}.
• Figure 2: Posterior probability distributions for a representative galaxy (ID = 'A2744_4423') derived using three distinct mass models with spherical velocity ellipsoid alignment ($\textsc{JAM}_{\mathrm{sph}}$). The panels correspond to the Power-Law (left), stars+NFW (middle), and stars+gNFW (right) parameterizations. The diagonal subplots display the marginalized 1D distributions, where vertical dashed lines indicate the median and the 16th/84th percentiles. The off-diagonal subplots show the covariances between model parameters. In the top-right corner of each panel display the observed root-mean-square velocity field ($V_{\rm rms}$) compared with the prediction from the respective best-fit model. The observed and best-fit $V_{\rm rms}$ maps for the entire sample are presented in Appendix \ref{['sec:all_galaxy_model']}.
• Figure 3: Example of visual assessment of the dynamical model fit quality. We present three galaxies for each quality class (defined in \ref{['sec:fit_quality']}). Each block of two rows corresponds to a specific quality class, displaying three galaxies arranged side-by-side. For an individual galaxy, the top row shows the HST image (labeled with the Galaxy ID and Quality flag), followed by the observed and modeled root-mean-square velocity ($V_{\rm rms}$) maps. The bottom row displays the MGE fit, where black and red contours indicate the observed and modeled isophotes, respectively, followed by the observed and predicted mean velocity ($V$) fields. In the observed and fitted kinematic maps, the overlaid black contours trace the observed and modeled surface brightness distribution.
• Figure 4: Comparison of the total density slope ($\gamma_{\rm T}$) measured using cylindrically ($\textsc{JAM}_{\mathrm{cyl}}$) versus spherically ($\textsc{JAM}_{\mathrm{sph}}$) aligned velocity ellipsoids for same mass model. The three panels correspond to the power-Law (left), stars + NFW (middle), and stars + gNFW (right) mass models. Galaxies with reliable kinematic fits (Qual $> 0$) are plotted as blue circles, while those with poor or unreliable fits (Qual $\le 0$) are marked with red diamonds. The black dashed line indicates the one-to-one relation. The intrinsic scatter ($\Delta$) between the two orientations, quantified using the LtsFit package, is reported within each panel for both the full sample and the high-quality subset.
• Figure 5: Comparison of the total density slopes ($\gamma_{\rm T}$) derived using same velocity ellipsoid orientation but different mass models. This comparison is restricted to the subsample with reliable dynamical models (Qual $\ge 1$). The top and bottom rows display results for the cylindrically ($\textsc{JAM}_{\mathrm{cyl}}$) and spherically ($\textsc{JAM}_{\mathrm{sph}}$) aligned velocity ellipsoids, respectively. The black dashed line marks the one-to-one relation, and the intrinsic scatter ($\Delta$) between the models is indicated in each panel.