The hierarchical origins of observed galaxy morphology

Galaxies grow primarily via accretion-driven star formation in discs and merger-driven growth of bulges. These processes are implicit in semi-analytical models of galaxy formation, with bulge growth in particular relating directly to the hierarchical build-up of haloes and their galaxies. In this paper, we consider several implementations of two semi-analytical models. Focusing on implementations in which bulges are formed during mergers only, we examine the fractions of elliptical galaxies and both passive and star-forming disc galaxies as functions of stellar and halo mass, for central and satellite systems. This is compared to an observational cross-matched Sloan Digital Sky Survey+Third Reference Catalog of Bright Galaxies z ˜ 0 sample of galaxies with accurate visual morphological classifications and M* > 1010.5 M⊙. The models qualitatively reproduce the observed increase of elliptical fraction with stellar mass, and with halo mass for central galaxies, supporting the idea that observed ellipticals form during major mergers. However, the overall elliptical fraction produced by the models is much too high compared with the z ˜ 0 data. Since the `passive' - i.e. non-star-forming - fractions are approximately reproduced, and since the fraction which are star-forming disc galaxies is also reproduced, the problem is that the models overproduce ellipticals at the expense of passive S0 and spiral galaxies. Bulge growth implementations (tuned to reproduce simulations) which allow the survival of residual discs in major mergers still destroy too much of the disc. Increasing the lifetime of satellites, or allowing significant disc regrowth around merger remnants, merely increases the fraction of star-forming disc galaxies. Instead, it seems necessary to reduce the mass ratios of merging galaxies, so that most mergers produce modest bulge growth in disc galaxy remnants instead of ellipticals. This could be a natural consequence of tidal stripping of stars from infalling satellite galaxies, a process not considered in our models. However, a high efficiency of quenching during and/or subsequent to minor mergers is still required to keep the passive fraction high.