Michele Cappellari

Abstract:

We present a series of 10 × 2 cosmological zoom simulations of the formation of massive galaxies with and without a model for active galactic nucleus (AGN) feedback. Differences in stellar population and kinematic properties are evaluated by constructing mock integral field unit maps. The impact of the AGN is weak at high redshift when all systems are mostly fast rotating and disc-like. After z ∼ 1 the AGN simulations result in lower mass, older, less metal rich, and slower rotating systems with less discy isophotes – in general agreement with observations. 2D kinematic maps of in situ and accreted stars show that these differences result from reduced in-situ star formation due to AGN feedback. A full analysis of stellar orbits indicates that galaxies simulated with AGN are typically more triaxial and have higher fractions of x-tubes and box orbits and lower fractions of z-tubes. This trend can also be explained by reduced late in-situ star formation. We introduce a global parameter, ξ3, to characterize the anticorrelation between the third-order kinematic moment h3 and the line-of-sight velocity (Vavg/σ), and compare to ATLAS3D observations. The kinematic correlation parameter ξ3 might be a useful diagnostic for large integral field surveys as it is a kinematic indicator for intrinsic shape and orbital content.

Abstract:

We investigate the nature of star-forming galaxies with reduced specific star formation rate (sSFR) and high stellar masses, those ‘green valley’ objects that seemingly cause a reported bending, or flattening, of the star-forming main sequence. The fact that such objects host large bulges recently led some to suggest that the internal formation of bulges was a late event that induced the sSFRs of massive galaxies to drop in a slow downfall, and thus the main sequence to bend. We have studied in detail a sample of 10 galaxies at 0.45 < z < 1 with secure SFR from Herschel, deep Keck optical spectroscopy, and HST imaging from CANDELS allowing us to perform multiwavelength bulge to disc decomposition, and to derive star formation histories for the separated bulge and disc components. We find that the bulges hosted in these systems below main sequence are virtually all maximally old, with ages approaching the age of the Universe at the time of observation, while discs are young (〈 T50〉 ∼ 1.5 Gyr). We conclude that, at least based on our sample, the bending of the main sequence is, for a major part, due to rejuvenation, and we disfavour mechanisms that postulate the internal formation of bulges at late times. The very old stellar ages of our bulges suggest a number density of early-type galaxies at z = 1–3 higher than actually observed. If confirmed, this might represent one of the first direct validations of hierarchical assembly of bulges at high redshifts.