The impact of Spitzer infrared data on stellar mass estimates - and a revised galaxy stellar mass function at 0 < z < 5

Aims. We estimate stellar masses of galaxies in the high redshift universe with the intention of determining the influence of newly available Spitzer/IRAC infrared data on the analysis. Based on the results, we probe the mass assembly history of the universe. Methods. We use the GOODS-MUSIC catalog, which provides multiband photometry from the U-filter to the 8 mu m Spitzer band for almost 15 000 galaxies with either spectroscopic (for approximate to 7% of the sample) or photometric redshifts, and apply a standard model fitting technique to estimate stellar masses. We than repeat our calculations with fixed photometric redshifts excluding Spitzer photometry and directly compare the outcomes to look for systematic deviations. Finally we use our results to compute stellar mass functions and mass densities up to redshift z=5. Results. We find that stellar masses tend to be overestimated on average if further constraining Spitzer data are not included into the analysis. Whilst this trend is small up to intermediate redshifts z less than or similar to 2.5 and falls within the typical error in mass, the deviation increases strongly for higher redshifts and reaches a maximum of a factor of three at redshift z approximate to 3.5. Thus, up to intermediate redshifts, results for stellar mass density are in good agreement with values taken from literature calculated without additional Spitzer photometry. At higher redshifts, however, we find a systematic trend towards lower mass densities if Spitzer/IRAC data are included.