Link to article: arXiv:1008.3901v1
This paper address a very important question about the built-up and evolution of galaxies at high redshift, by deriving stellar masses from SED fitting for ~400 LBG observed with Hubble-WFC3/IR and Spitzer-IRAC in the GOODS South field. Using Bruzual & Charlot models with a Salpeter IMf, masses are obtained assuming a constant Star Formation History for galaxies.
By fitting the Mass vs Luminosity relation at z~4 - with a slope that makes the M/L ratio increase with luminosity - and finding no clear evidence for evolution with redshift, this work use the UV Luminosity Functions of Bouwens et al. (2007-2010) and Monte-Carlo simulations to derive Mass Functions (MF) at z~4,5,6,7. The results show a significant difference with theoretical expectations from numerical simulations, and especially a much more flatter faint end slope.
The last part of this paper show a comparison of the stellar mass growth computed by 2 different approaches: a quite large disagreement is found between the integration of the SFR density over time [SFRD ~ (1+z)^-6.3] compared with the integration of MF for the 4 bins of redshift [SMD ~ (1+z)^-2.8]. The authors propose that dust extinction overestimates, low duty cycle for SF or shortcomings in the modeling of faint galaxies may explain this noticeable difference.