Massive quiescent galaxies at z ∼ 3: A comparison of selection, stellar population, and structural properties with simulation predictions

We study stellar population and structural properties of massive log (M-*/M-?) > 11 galaxies at z approximate to 2.7 in the Magneticum and IllustrisTNG hydrodynamical simulations and GAEA semi-analytic model. We find stellar mass functions broadly consistent with observations, with no scarcity of massive, quiescent galaxies at z ti2.7, but with a higher quiescent galaxy fraction at high masses in IllustrisTNG. Average ages of simulated quiescent galaxies are between approximate to 0.8 and 1 . 0 Gyr , older by a factor approximate to 2 than observed in spectroscopically confirmed quiescent galaxies at similar redshift. Besides being potentially indicative of limitations of simulations in reproducing observed star formation histories, this discrepancy may also reflect limitations in the estimation of observed ages. We investigate the purity of simulated UVJ rest-frame colour-selected massive quiescent samples with photometric uncertainties typical of deep surveys (e.g. COSMOS). We find evidence for significant contamination (up to 60 per cent ) by dusty star-forming galaxies in the UVJ region that is typically populated by older quiescent sources. Furthermore, the completeness of UVJ-selected quiescent samples at this redshift may be reduced by approximate to 30 per cent due to a high fraction of young quiescent galaxies not entering the UVJ quiescent region. Massive, quiescent galaxies in simulations have on average lower angular momenta and higher projected axis ratios and concentrations than star-forming counterparts. Average sizes of simulated quiescent galaxies are broadly consistent with observations within the uncertainties. The average size ratio of quiescent and star-forming galaxies in the probed mass range is formally consistent with observations, although this result is partly affected by poor statistics.

Automated summary

We study the properties of massive galaxies at z~=2.7 using hydrodynamical simulations and a semi-analytic model. The simulations show that there is no shortage of massive, quiescent galaxies at high redshift, but the fraction of quiescent galaxies is higher in some models. The average ages of simulated quiescent galaxies are older than observed, indicating some limitations in the simulations or in estimating observed ages. We also find that there is contamination in the UVJ selection method for quiescent galaxies, and that the completeness of UVJ-selected samples may be reduced due to young quiescent galaxies not entering the quiescent region. The simulated quiescent galaxies have different properties compared to star-forming galaxies, with lower angular momenta and higher concentrations.