What turns galaxies off? The different morphologies of star-forming and quiescent galaxies since z~2 from CANDELS
(30 votes over all institutions)
We use HST/WFC3 imaging from the CANDELS multicycle treasury survey, in conjunction with the SDSS, to explore the evolution of galactic structure for galaxies with stellar masses M*>3e10M_sun from z=2.2 to the present epoch, a time span of 10 Gyr. We explore the rest-frame optical colors, stellar mass, star formation (SF) activity and the structural parameters of galaxies. We confirm the dramatic evolution from z=2.2 to the present day in the number density of non-star-forming galaxies above 3e10M_sun reported by other authors. The vast majority of these quiescent systems have concentrated light profiles, as parameterized by the Sersic index, and the population of concentrated galaxies grows similarly rapidly. We examine the joint distribution of SF activity, Sersic index, stellar mass, mass divided by radius (a proxy for velocity dispersion), and stellar surface density. Quiescence correlates poorly with stellar mass at all z<2.2 (given the <0.2dex scatter between halo mass and stellar mass at z~0 inferred by More et al. 2009; MNRAS, 392, 801, this argues against halo mass being the only factor determining quiescence). Quiescence correlates better with Sersic index, `velocity dispersion' and stellar surface density, where Sersic index correlates the best (increasingly so at lower redshift). Yet, there is significant scatter between quiescence and galaxy structure: while the vast majority of quiescent galaxies have prominent bulges, many of them have significant disks, and a number of bulge-dominated galaxies have significant SF. Noting the rarity of quiescent galaxies without prominent bulges, we argue that a prominent bulge (and, perhaps by association, a supermassive black hole) is a necessary but not sufficient condition for quenching SF on galactic scales over the last 10Gyr; such a result is qualitatively consistent with the expectations of the AGN feedback paradigm.
