Models of high redshift luminosity functions and galactic outflows: The dependence on halo mass function

The form of the halo mass function is a basic ingredient in any semi-analytical galaxy formation model. We study the existing forms of the mass functions in the literature and compare their predictions for semi-analytical galaxy formation models. Two methods are used in the literature to compute the net formation rate of halos, one by simply taking the derivative of the halo mass function and the other using the prescription due to Sasaki [Sasaki, S., 1994. PASJ 46, 427]. For the historically used Press-Schechter (PS) mass function, we compare various model predictions, using these two methods. However, as the Sasaki formalism cannot be easily generalized for other mass functions, we use the derivative while comparing model predictions of different mass functions. We show that the reionization history and UV luminosity function of Lyman break galaxies (LBGs) predicted by the PS mass function differs from those using any other existing mass function, like Sheth-Tormen (ST) mass function. In particular the reionization efficiency of molecular cooled halos has to be substantially reduced when one uses the ST and other mass functions obtained from the simulation instead of the PS mass function. Using chi(2)-minimization, we find that the observed UV luminosity functions of LBGs at 3.0 <= z <= 7.4 are better reproduced by models using the ST mass function compared to models that use the PS mass function. On the other hand, the volume filling factor of the metals expelled from the galaxies through supernovae driven outflows differs very little between models with different mass functions. It depends on the way we treat merging outflows. We also show that the porosity weighted average quantities related to the outflow are not very sensitive to the differences in the halo mass function. (C) 2009 Elsevier B.V. All rights reserved.