Astron. Astrophys. 364, 443-449 (2000)

3. The Seyfert 2 galaxies

While the Catalog lists 228 Seyfert 1s and QSOs, only 16 Seyfert 2s were identified as a primary optical counterpart of one of the observed X-ray sources. Moreover, our classification criteria for Seyfert 2s (Seyfert spectrum without detectable broad lines or line components and without detectable Fe II emission) could have resulted in a contamination of the Seyfert 2 sample by Seyfert 1.5-1.9 and NLS1 objects with weak BLR components and weak Fe II emission (since broad line components and weak Fe II blends are more strongly affected by a low than narrow spectral features covering fewer pixels). Moreover, while (as pointed out in Sect. 1) the Seyfert 1 sample is practically unaffected by the definitions used, a Seyfert 1 classification based entirely on the [O III ]/H flux ratio may result in a different Seyfert 2 sample. In order to estimate these potential effects we re-inspected the spectra of all objects classified as Seyfert 2s in our Catalog and derived [O III ]/H flux ratios and approximate limits for the presence of undetected BLR components and Fe II blends. In this process we found one case (RX J2218.6+0802) which definitely should have been classified differently. New measurements of the corresponding spectrum resulted in FWHM(H) = 1520 120 kms^-1, FWHM([O III ]) = 878 200 kms^-1 (both uncorrected for the instrumental profile) and [O III ]/H = 2.0. Obviously, this object should have been classified as a NLS1. Five additional objects do not show detectable BLR or Fe II features but definitely [O III ]/H 3.0. In a classification based on the [O III ]/H ratio alone, these objects would also have to be classified as NLS1s. On the other hand, at least 7 of the objects listed in our Catalog as Seyfert 2s have [O III ]/H 3.0, Balmer BLR contributions 10% and Fe II 0.2. Hence, we conclude that at least about 50% of the optical counterparts classified as Seyfert 2s in our Catalog are bona fide type 2 objects, independently of the exact classification criteria and a possible contamination with type 1 objects due to the inadequate of some of our spectra.

In view of the uncertainties and ambiguities of the Seyfert 2 sample, in the following all statistical results are listed separately for the full original sample of the Catalog (16 objects) followed [in brackets] by the corresponding value for the minimal sample of 7 reliable objects. Because of the small size of the samples, the discussion will be restricted to statistical means.

Compared to the Seyfert 1 and QSO sample the most obvious difference is the significantly lower mean luminosity (and correspondingly lower mean redshift) of the Seyfert 2s (reflecting also the scarcity of narrow line QSOs). The visual luminosities of our Seyfert 2s covered the range 18.1 22.8 [18.1 22.1] with a median value of -20.5, as compared to -22.9 for the Seyfert 1 and QSO sample. Hence the observed Seyfert 2s have on average only about 1/10 of the visual luminosity of the observed Seyfert 1s and QSOs. The Seyfert 2 redshifts are all below 0.5 with an average of 0.18 [.22]. On the other hand, the distribution of the Seyfert 2s (median value is not significantly different from that of the Seyfert 1s and QSOs ().

For our flux limited sample the (visual and X-ray) Seyfert 2 to 1 luminosity ratio of 1/10 means that the volume in which we observe the Seyfert 2s is only about 3% of that of the Seyfert 1s and QSOs. Since the relative fraction of the Seyfert 2s in our Catalog ([3%] to 7%) is rather close to this number the Seyfert 2s detected in the RASS, although on average much fainter, are found to have about the same (or slightly larger) space density than the observed Seyfert 1s. Equal space densities of Seyfert 1s and Seyfert 2s have also been estimated for the general AGN population (Simkin et al. 1980; Véron & Véron-Cetty 1986). In this respect our X-ray selected AGN obviously show the same behavior as the general AGN population.

© European Southern Observatory (ESO) 2000

Online publication: January 29, 2001
helpdesk.link@springer.de