The close environment of Seyfert galaxies and its implication for unification models

In this Letter, we present a statistical analysis of the circumgalactic env ironment of nearby Seyfert galaxies based on a computer-aided search of com panion galaxies on the Digitized Sky Survey. We defined a sample of 72 near by Seyfert I galaxies (redshift 0.007 less than or equal to z less than or equal to 0.034) and a sample of 60 Seyfert 2 galaxies (0.007 I z I 0.020), which include only high galactic latitude objects. In addition, we built tw o control samples of nonactive galaxies that match the number of sample mem bers, the redshift, morphological type, and diameter distribution of the Se yfert 1 and Seyfert 2 galaxy samples separately. We stress how our sample s election introduces important methodological improvements that avoid severa l sources of strong bias. An intrinsic difference between the environment o f Seyfert 1 galaxies and that of Seyfert 2 galaxies, suggested by previous work, is confirmed as statistically significant. For Seyfert 2 galaxies, we find a significant excess of large companions (D-C less than or similar to 10 kpc) within a search radius less than or similar to 100 kpc of projecte d linear distance as well as within a search radius equal to 3 times the di ameter D-s of each Seyfert galaxy. For Seyfert 1 galaxies, there is no clea r evidence of any excess of companion galaxies either within 100 kpc or wit hin 3D(s). For all samples, the number of companions that are actually coun ted within a search radius of 3D(s) is a factor of approximate to 2 above t he expectation values derived from the number density of galaxies over the 1 deg(2) fields that are centered on the sample galaxies, suggesting a mark edly non-Poissonian distribution for galaxies on scales less than or simila r to 100 kpc. This difference in environment is not compatible with the sim plest formulation of the unification model for Seyfert galaxies: both type 1 and type 2 should be intrinsically alike, the only difference being the r esult of the orientation of an obscuring torus. We propose an alternative f ormulation.