Properties of X-ray selected active galactic nuclei .4. The local optical luminosity function of broad-line active galactic nuclei

We have selected a local (z less than or equal to 0.3) subsample of 226 broad line active galactic nuclei (BLAGNs) from the Einstein Observatory Extended Medium Sensitivity Survey. This sample represents the largest unbiased and complete sample of local BLAGNs ever assembled and has allowed us to derive their space density in regions of the m(B) - z plane where, with the usual optical selection criteria, it is very difficult to obtain complete samples of BLAGNs. Using total integrated magnitudes (i.e., nucleus + host galaxy), we have computed the local optical luminosity function of this X-ray selected sample and compared it with those derived from local optical samples. Thanks to the large number of objects at our disposal we can set more stringent constraints on the space density of BLAGNs than has previously been possible. The luminosity function derived from our sample is in good agreement with the composite luminosity function which can be derived from optically selected samples only by using different selection criteria in different ranges of absolute magnitude. In particular, at low luminosity (M(B) greater than or equal to -22) we confirm the flattening of the local optical luminosity function originally suggested by Meurs & Wilson (1984) while in the magnitude range from M(B) similar to-23 to -25 we find a very good agreement with the optical spatial density derived using data from the Bright Quasars Survey. By convolving our luminosity function with the distribution of the ratio of nuclear to total flux of a sample of similar to 40 Seyfert 1 and 1.5 galaxies from the literature, we have also derived an estimate for the nuclear luminosity function of BLAGNs. This nuclear luminosity function is in rather good agreement with the nuclear luminosity functions previously derived, using a much smaller number of objects, from optical samples of low-luminosity BLAGNs. A reasonably good agreement is also found between our luminosity function and the extrapolation to low redshift (z = 0.15, the average redshift of our sample) of the quasar luminosity function derived from more than 1000 optically selected quasars. The integration of our nuclear luminosity function over the M(B) - z plane shows that good agreement is obtained with the observed number counts of low luminosity (M(B) greater than or equal to -23) BLAGNs at faint magnitudes, if the M(B) greater than or equal to -23 population evolves similarly to the QSO population.