2011MNRAS.415.2932R

We find that the recently reported departure from the Eddington luminosity limit for the highest quasar black hole masses at a given redshift is an artefact due to biases in black hole mass measurements. This sub-Eddington boundary (with non-unity slope) in the quasar mass–luminosity plane was initially reported by Steinhardt & Elvis using the full width at half-maximum (FWHM) based black hole mass catalogue of Shen et al. However, the significance of the boundary is reduced when the FWHM-based mass scaling relationship is recalibrated following Wang et al. and using the most updated reverberation mapping estimates of black hole masses. Furthermore, this boundary is not seen using mass estimates based on the line dispersion of the same quasars' Mg II emission lines. Thus, the initial report of a sub-Eddington boundary with non-unity slope was due to biases in estimating masses using the FWHM of a fit of one or two Gaussians to quasar Mg II emission lines. We provide evidence that using the line dispersion of the Mg II line produces less biased black hole mass estimates.