We use photometric data from the Spitzer mission to explore the mid- and far-infrared properties of 10 red QSOs ( J - K > 2 , R - K > 5 mag ) selected by combining the 2MASS in the near-infrared with the SDSS at optical wavelengths .
Optical and/or near-infrared spectra are available for 8/10 sources .
Modeling the Spectral Energy Distribution ( SED ) from UV to far-infrared shows that moderate dust reddening ( A _ { V } = 1.3 - 3.2 ) can explain the red optical and near-IR colours of the sources in the sample .
There is also evidence that red QSOs have 60 / 12 \mu m luminosity ratio higher than PG QSOs ( 97 per cent significance ) .
This can be interpreted as a higher level of star-formation in these systems ( measured by the 60 \mu m luminosity ) for a given AGN power ( approximated by the 12 \mu m luminosity ) .
This is consistent with a picture where red QSOs represent an early phase of AGN evolution , when the supermassive black hole is enshrouded in dust and gas clouds , which will eventually be blown out ( possibly by AGN driven outflows ) and the system will appear as typical optically luminous QSO .
There is also tentative evidence significant at the 96 % level that red 2MASS QSOs are more often associated with radio emission than optically selected SDSS QSOs .
This may indicate outflows , also consistent with the young AGN interpretation .
We also estimate the space density of red QSOs relative to optically selected SDSS QSOs , taking into account the effect of dust extinction and the intrinsic luminosity of the sources .
We estimate that the fraction of red QSOs in the overall population increases from 3 % at M _ { K } = -27.5 mag to 12 % at M _ { K } = -29.5 mag .
This suggests that reddened QSOs become more important at the bright end of the Luminosity Function .
If red QSOs are transition objects on the way to becoming typical optically luminous QSOs , the low fractions above suggest that these systems spent less than 12 % of their lifetime at the “ reddened ” stage .