2009A&A...508..107G

We present an extensive search to identify the counterparts of all the microwave foreground sources listed in the WMAP 3-year catalogue using literature and archival data. Our work has led to the identification of 309 WMAP sources, 98% of which are blazars, radio quasars, or radio galaxies. Only 7 WMAP detections were identified with other types of cosmic sources (3 starburst galaxies and 4 planetary/LBN nebulae). At present, 15 objects (<5%) still remain without identification because of the unavailability of optical spectroscopic data or a clear radio counterpart. Our results allow us to define a flux-limited sample of 203 high Galactic latitude microwave sources (f_41GHz≥1Jy, |b_II|>15°) that is virtually completely identified (99%). The microwave band is ideally suited to blazar statistical studies since this part of the electromagnetic spectrum is least affected by the superposition of spectral components of different origins, and therefore by selection effects. Using this data-set, we derived number counts, luminosity functions, and cosmological evolution of blazars and radio galaxies at microwave frequencies. Our results are in good agreement with those found at radio (cm) frequencies. The 5GHz bivariate blazar luminosity functions are similar to those derived from the DXRBS survey, which shows that this sample is representative of the blazar population at 41GHz. Microwave selected broad-lined quasars are about six times more abundant than BL Lacs, a ratio that is similar to, or larger than, that seen at radio and gamma-ray frequencies, once spectral selection effects are taken into account. This strongly suggests that the mechanism responsible for the generation of gamma-rays is, to first order, the same in all blazar types, leaving little room for models (such as external Compton radiation) that predict very different gamma-ray emission in broad-lined and lineless blazars. Our results confirm, and strengthen on a more solid statistical base, that blazars and radio galaxies are the largest contaminants of CMB anisotropy maps. We predict that these sources are also bright gamma-ray sources, most of which will be detected by the AGILE and GLAST satellites.