2016MNRAS.460..902B

We investigate the star-forming properties of 1620 X-ray selected active galactic nuclei (AGN) host galaxies as a function of their specific X-ray luminosity (i.e. X-ray luminosity per unit host stellar mass) - a proxy of the Eddington ratio. Our motivation is to determine whether there is any evidence of a suppression of star formation at high Eddington ratios, which may hint towards 'AGN feedback' effects. Star formation rates (SFRs) are derived from fits to Herschel-measured far-infrared spectral energy distributions, taking into account any contamination from the AGN. Herschel-undetected AGNs are included via stacking analyses to provide average SFRs in bins of redshift and specific X-ray luminosity (spanning 0.01 \lesssim L_X/M_* \lesssim 100  L_{☉}  M_{☉}^{-1}). After normalizing for the effects of mass and redshift arising from the evolving galaxy main sequence, we find that the SFRs of high specific luminosity AGNs are slightly enhanced compared to their lower specific luminosity counterparts. This suggests that the SFR distribution of AGN hosts changes with specific X-ray luminosity, a result reinforced by our finding of a significantly higher fraction of starbursting hosts among high specific luminosity AGNs compared to that of the general star-forming galaxy population (i.e. 8-10 per cent versus 3 per cent). Contrary to our original motivation, our findings suggest that high specific luminosity AGNs are more likely to reside in galaxies with enhanced levels of star formation.