The 2dF QSO Redshift Survey - IX. A measurement of the luminosity dependence of QSO clustering

In this Paper we present a clustering analysis of QSOs as a function of luminosity over the redshift range z=0.3-2.9. We use a sample of 10566 QSOs taken from the preliminary data release catalogue of the 2dF QSO Redshift Survey (2QZ). We analyse QSO clustering as a function of apparent magnitude. The strong luminosity evolution of QSOs means that this is approximately equivalent to analysing the data as a function of absolute magnitude relative to M* over the redshift range that the 2QZ probes. Over the relatively narrow range in apparent magnitude of the 2QZ we find no significant (>2sigma) variation in the strength of clustering, however, there is marginal evidence for QSOs with brighter apparent magnitudes having a stronger clustering amplitude. QSOs with 18.25<bj<19.80 show a correlation scale length s_0=5.50+-0.79h-1Mpc in an Einstein-de Sitter (EdS) universe and s_0=8.37+-1.17h-1Mpc in a universe with omega_0=0.3 and lambda_0=0.7 (Lambda), while the best fit values for the full magnitude interval (18.25<bj<20.85) over the same spatial scales are s_0=4.29+-0.30h-1Mpc (EdS) and s_0=6.35+-0.45h-1Mpc (Lambda). We can therefore determine that the bias of the brightest sub-sample is a factor 1.22+-0.15 (EdS) or 1.24+-0.15 (Lambda) larger than that of the full data set. An increase in clustering with luminosity, if confirmed, would be in qualitative agreement with models in which the luminosity of a QSO is correlated to the mass of the dark halo in which it resides, implying that the mass of the host plays at least some part in determining a QSO's formation and evolution. These models predict that the clustering in brighter QSO data sets, such as Sloan Digital Sky Survey QSO sample or the bright extension of the 2QZ should show a higher clustering amplitude than the 2QZ.