ILLUSTRating red nugget assembly through observations and simulations

TL;DR

This work investigates how compact massive galaxies (CMGs) and massive relics assemble and evolve across cosmic time by applying uniform selection to both observations and the TNG50 simulation at z=0, 0.3, 0.7, and 2. It analyzes mass-size, velocity dispersion, and stellar metallicity relations, and classifies simulated CMGs by the fraction of stellar mass growth since z=2 to mirror observational DoR categories. Key findings show that simulated CMGs follow the observed mass-size trend and increase in number density with redshift, while velocity dispersions are comparatively uniform across accretion histories and simulated CMGs are generally more metal-rich than the quiescent population, with the metallicity offset shrinking toward higher redshift. Observational relics, however, tend to exhibit higher velocity dispersions and stronger metallicity enhancements than their simulated counterparts, indicating partial alignment but notable discrepancies that highlight the need for improved forward-modeling and larger samples for robust cross-epoch comparisons.

Abstract

The properties of massive and compact early-type galaxies provide important constraints on early galaxy formation. Among these, massive relic galaxies, characterized by old stellar populations and minimal late-time accretion, are considered preserved compact galaxies from the high-$z$ Universe. We investigate compact and massive galaxies (CMGs) using the TNG50 cosmological simulation, applying uniform selection criteria matching observational surveys at $z=0$, $z=0.3$, and $z=0.7$, enabling direct comparisons with observed compact galaxies. CMGs are classified according to their stellar mass assembly histories to examine how compactness relates to dynamical properties and chemical enrichment across cosmic time. Our results show that simulated CMGs follow the observed mass-size relation, with the number of objects increasing at higher redshifts, in line with observational trends. Dynamically, while observations suggest relic galaxies are outliers in the stellar mass-velocity dispersion plane, simulated compacts show relatively uniform velocity dispersions across different accretion histories. Observed relics are more metal-rich than other compact galaxies with extended star formation, deviating from the local mass-metallicity relation. In contrast, simulated CMGs are overall more metal-rich than the quiescent population, regardless of accretion history. The deviation from the mass-metallicity relation decreases with redshift. These results suggest that the extreme characteristics of CMGs in TNG50, particularly in metallicity and dynamics, are less pronounced than in observed relics. Nonetheless, these results offer a theoretical framework to assess the properties of such extreme objects from different epochs, highlighting both alignment with and deviations between the models.

Figures (9)

• Figure 1: Stellar mass--size relation across redshift bins. All TNG50 subhalos are shown as light gray circles in the background, while selected compact subhalos are represented by dark gray circles. Observational CMGs from the different samples are indicated with white diamond markers. The solid black lines follow the mass--size relation from vanderWel2014, while the dashed gray lines highlight the selection region for the simulated sample ($R_e<2\,\rm{kpc}, log\,\Sigma_{1.5}>10.3\,dex$)
• Figure 2: Example of the selected sample of compact TNG50 simulated (left column) and observed (right column) galaxies across redshift bins: $z=0$ (top row), $z=0.3$ (middle row), and $z=0.7$ (bottom row). Observed galaxies include NGC 1277 (top-right, HST ACS/WFC instrument), 2024spiniello (middle-right), and 2018Scodeggio (bottom-right).
• Figure 3: Number density of CMGs as a function of redshift, in units of co-moving Mpc$^{-3}$. The black line and markers represent this work TNG50 CMGs. Colored lines and markers correspond to observational estimates from VIPERS 2023Lisiecki in red, 2020AScognamiglio in orange, 2014Damjanov in green, vanderWel2014 in purple, 2013Barro in blue, and Ferre-Mateu2017 in dark green for the local relics, and in magenta for 2023GrebolTomas compacts. Simulated CMGs show increasing number density with redshift, in line with models predicting more compact objects in the early Universe. A similar trend is seen in the observational data.
• Figure 4: Stellar mass assembly for the simulated and observed samples at $z=0$. The upper panel shows the stellar mass fraction over time for each compact subhalo selected from TNG50, color-coded by the stellar mass growth fraction since $z=2$. The lower panel shows the stellar mass assembly over time for the MaNGA compact galaxies of 2023GrebolTomas, color-coded by their groups: blue represents groups B and C, which show extended and/or late SFHs. Red color represents galaxies in their group A, which show early and fast formation (and thus negligible accretion after $z=2$), being dark red those extreme cases whereby the galaxy is fully assembled early on (and thus has the minimum amount of accreted material, $<10$%). In both panels, the black dashed vertical line indicates $z=2$ and the dashed gray horizontal line indicates the stellar mass fraction of 90%.
• Figure 5: Stellar mass--stellar velocity dispersion relations for simulated and observed galaxies in the different redshift bins. The dashed black line represents the median $\sigma_\star$ for all TNG50 quiescent subhalos, while the solid black line corresponds to the 2016AZahid relation across all redshifts. All TNG50 subhalos are shown as gray dots, being the compact sample color-coded by their mass growth since $z=2$ (except for the last bin, which is considered as the starting point and thus are left un-colored). Observational data are represented by diamonds color-coded as: dark red for the extreme relics (e.g., extreme cases of group A for MaNGA, DoR$>$0.7 for INSPIRE); red for relics (rest of A of MaNGA and INSPIRE galaxies with 0.5$<$DoR$<$0.7); and blue for the non-relic group (classes B and C of MaNGA, DoR$<$0.5 for INSPIRE).