Extending PLE models into the mid-IR, far-IR & sub-mm

Abstract

Simple pure luminosity evolution (PLE) models, in which galaxies brighten at high redshift due to increased star-formation rates (SFRs), are known to provide a good fit to the colours and number counts of galaxies throughout the optical and near-infrared. We show that optically defined PLE models, where dust reradiates absorbed optical light into infrared spectra composed of local galaxy templates, fit galaxy counts and colours out to 8um and to at least z=2.5. At 24-70um, the model is able to reproduce the observed source counts with reasonable success if 16% of spiral galaxies show an excess in mid-IR flux due to a warmer dust component and a higher SFR, in line with observations of local starburst galaxies. There remains an under-prediction of the number of faint-flux, high-z sources at 24um, so we explore how the evolution may be altered to correct this. At 160um and longer wavelengths, the model fails, with our model of normal galaxies accounting for only a few percent of sources in these bands. However, we show that a PLE model of obscured AGN, which we have previously shown to give a good fit to observations at 850um, also provides a reasonable fit to the Herschel/BLAST number counts and redshift distributions at 250-500um. In the context of a LCDM cosmology, an AGN contribution at 250-870um would remove the need to invoke a top-heavy IMF for high-redshift starburst galaxies, although the excellent fit of the galaxy PLE model at shorter wavelengths would still need to be explained.