The physics of galaxy formation

Abstract

We investigate improved, more physical methods for modelling key processes in galaxy formation that take place in the interstellar medium, and study their effects on the observed properties and evolution of galaxies. The topics we investigate are: (i) improvements to the relation between the star formation rates (SFRs) and the cold gas contents of galaxies; (ii) how to predict the atomic and molecular gas masses in galaxies with different properties, (iii) how to predict the emission of widely used molecular tracers, such as carbon monoxide (CO); and (iv) the gas outflows from galaxies driven by supernovae and their dependence on local and global properties of galaxies. We perform these studies using the semi-analytic model of galaxy formation GALFORM, and fully embed our calculations in a cosmological scenario, the Lambda cold dark matter paradigm. This is done with the dual aims of understanding how the physical processes above affect galaxy formation and evolution in a statistical fashion, and to improve and extend the predictive power of galaxy formation models. We find that by calculating the SFR from the molecular gas content and relating the molecular-to-atomic mass ratio in the interstellar medium to the hydrostatic pressure in the midplane of the disk, we can explain the observed atomic gas mass function and clustering of galaxies selected by their atomic hydrogen mass, the SFRs of local and high-redshift galaxies, the evolution of the molecular hydrogen gas fraction and the global atomic hydrogen abundance of the universe, and the local scaling relations between gas contents and other galaxy properties. We also find that by coupling GALFORM with a radiative transfer and interstellar chemistry code describing photon dominated regions, our new model can explain the observed emission of CO from different types of galaxy. Finally, based on a physical description of the dynamical evolution of bubbles created by supernovae in the interstellar medium, we find that the outflow rate driven by supernovae depends strongly on the surface density of gas plus stars and the gas fraction. We critically revise the phenomenological prescriptions widely used to describe supernova feedback in the literature and propose new physically motivated prescriptions.