Galaxy Formation: Cold Dark Matter, Feedback, and the Hubble Sequence

Abstract

TreeSPH simulations of galaxy formation in a standard Λ cold dark matter cosmology, including star formation and the effects of energetic stellar feedback processes and of a metagalactic UV field, have been performed, resulting in a mix of realistic disk, lenticular, and elliptical galaxies at redshift z=0. The disk galaxies are deficient in angular momentum by only about a factor of 2 compared with observed disk galaxies for simulations with fairly strong starbursts in early, protogalactic clouds, leading to ``blow-away'' of the remaining gas in the clouds. In this respect the present scenario is hence doing almost as well as the warm dark matter (WDM) scenarios discussed by Sommer-Larsen & Dolgov. The surface density profiles of the stellar disks are approximately exponential, and those of the bulges range from exponential to r1/4, as observed. The bulge-to-disk ratios of the disk galaxies are consistent with observations, as are their integrated B-V colors, which have been calculated using stellar population synthesis techniques. Furthermore, the observed I-band Tully-Fisher relation can be matched, provided that the stellar mass-to-light ratio of disk galaxies is M/LI~0.8, similar to what was found by Sommer-Larsen & Dolgov from their WDM simulations and in fair agreement with several recent observational determinations of M/LI for disk galaxies. The elliptical and lenticular galaxies have approximately r1/4 stellar surface density profiles, are dominated by nondisklike kinematics, and are flattened as a result of nonisotropic stellar velocity distributions, again consistent with observations. Hot halo gas is predicted to cool out and be accreted onto the Galactic disk at a rate of 0.5-1 Msolar yr-1 at z=0, consistent with upper limits deduced from Far Ultraviolet Spectroscopic Explorer observations of O VI. We have analyzed in detail the formation history of two disk galaxies with circular speeds comparable to that of the Milky Way and find gas accretion rates, and hence bolometric X-ray luminosities of the halos, 6-7 times larger at z~1 than at z=0 for these disk galaxies. More generally, it is found that gas infall rates onto these disks are nearly exponentially declining with time, both for the total disk and for the ``solar cylinder.'' This theoretical result hence supports the exponentially declining gas infall approximation often used in chemical evolution models. The infall timescales deduced are ~5-6 Gyr, comparable to what is adopted in current chemical evolution models to solve the ``G dwarf problem.'' The disk of one of the two galaxies forms ``inside-out,'' the other ``outside-in,'' but in both cases the mean ages of the stars in the outskirts of the disks are >~6-8 Gyr, fairly consistent with the findings of Ferguson & Johnson for the disk of M31. The amount of hot gas in disk galaxy halos is consistent with observational upper limits. The globular cluster M53 and the LMC are ``inserted'' in the halos of the two Milky Way-like disk galaxies, and dispersion measures to these objects are calculated. The results are consistent with upper limits from observed dispersion measures to pulsars in these systems. Finally, the results of the simulations indicate that the observed peak in the cosmic star formation rate at redshift z~2 can be reproduced. Depending on the star formation and feedback scenarios, one predicts either a cosmic star formation rate that decreases monotonically with redshift beyond these redshifts or a second peak at z~6-8, corresponding to the putative Population III and interestingly similar to recent estimates of the redshift at which the universe was reionized. These various scenarios should hence be observationally constrainable with upcoming instruments such as the James Webb Space Telescope and the Atacama Large Millimeter Array.