Understanding that all types of AGN are most likely powered by accreting black holes helps to elucidate our picture of these objects. An even clearer picture is possible if we can reduce the variety of classifications by showing that two or more ``types'' which appear distinct are intrinsically the same. Such a unifying scenario has been developed for Seyfert galaxies, and it may have broader implications for AGNs in general because an analogous explanation may apply to other classifications.
There is now much evidence that Seyfert 2 galaxies house Seyfert 1 engines at their cores. Just over a decade ago, Antonucci and Miller [Antonucci and Miller, 1985] identified broad H I emission lines in the polarized light from the Seyfert 2 galaxy NGC 1068. They explained this by asserting that the central engine of NGC 1068 is the same as those observed in Seyfert 1 galaxies, except our line of sight towards this engine is obstructed by a geometrically and optically thick torus. We can't directly observe the engine, but we see its light reflected by regions of the galaxy which are directly exposed to the nucleus and not obstructed by the torus. This explanation contains the essence of what has become the unified model for active galactic nuclei.
The unified model describes all Seyfert galaxies as intrinsically the same creatures, but with different observed properties depending upon our viewing angle. (For an illustration of the structure in the unified model, refer to the figure at the end of this proposal.) An accreting supermassive black hole is posited to be the central engine. The material that fuels it lies in an accertion disk which girdles the black hole. The accreting material produces an intensely luminous continuum. In the volume immediately surrounding the disk are clouds of hot gas that are the sources of the broad lines observed in type 1 Seyferts. This defines the broad line region (BLR). Beyond the BLR and coplanar with the accretion disk is an optically thick torus of material. It is the orientation of this torus which dictates whether the galaxy is observed as a Seyfert 1 or a Seyfert 2. If our line of sight lies close enough to the axis of the torus, we can peer directly through the hole of the donut to see the BLR. On the other hand, if we are too close to the plane of the torus, the direct emissions of the BLR and the accretion disk are attenuated by the torus. Enveloping this whole system is a halo of material. Due to collimation by the geometrically thick torus, conical sections of this halo above and below the opening are directly exposed to the central energy source. The innermost portion of this cone is optically thin and largely photoionized, which provides enough free electrons to scatter a significant fraction of the central engine's light. Denser clouds in this conical region are excited by the high energy photon flux and the collisional excitation of ejecta from the nucleus. These clouds emit the narrow lines. Because the electron scattering and narrow line regions (NLR) lie outside the obscuring torus, the same narrow line features are seen in all Seyfert galaxies, and the emissions of the hidden broad line regions of type 2 Seyferts can be observed in scattered (and hence polarized) light. Intermediate Seyfert types arise when the inclination of the torus is somewhere between the edge-on and the face-on extremes.
Other observations provide further support for the unification scenario. The simple geometry of many Seyfert 2 narrow line regions as conal or bi-conal with their vertices coincident with the galactic nuclei and their elongation parallel to the elongation direction of radio emission fits exactly with the prediction of the unification model. In most Seyfert 1 galaxies, the ratio of the broad line luminosity to the several keV X-ray luminosity is uniform; the ratio observed in the Seyfert 2 galaxy NGC 1068 has the same value [Antonucci, 1993]. Evidence of reflected broad line emissions have been detected in the polarized light of a number of other Seyfert 2 galaxies [Miller and Goodrich, 1990, Tran et al., 1992]. By comparing the observed ionizing radiation with the inferred flux exciting the NLR of type 2 Seyferts, it is clear that much of the direct flux is hidden from our view. Furthermore, the slope of the implied UV spectrum in these objects is indistinguishable from that found in type 1 Seyferts [Kinney et al., 1991]. In addition, X-ray observations of numerous Seyfert 2 galaxies reveal high column densities of intervening matter, which could be evidence of obscuring tori [Mulchaey et al., 1992].
The unified model is very persuasive. Qualitatively it accounts for
many of the variations observed between different objects, but for
years, there have been hints at problems. These include a possible
deficiency of infrared radiation which should be present to account
for the energy absorbed and reprocessed by the tori [Penston et al., 1990];
the morphology of the NLRs for a number of Seyfert galaxies may be
inconsistent with predictions [Pogge, 1989]; evidence that type 2
Seyferts are systematically stronger radio
emitters [Ulvestad and Wilson, 1989, Moran et al., 1992] and type 1's are stronger infrared
sources [Ulvestad, 1986]. Each of these observations may be attributed
to limited samples, selection effects, or intrinsic anisotropies, and
thereby dismissed. However, stronger evidence that there are flaws in
the model are now becoming available. In particular, a study due out
this year by Turner et al. finds that several Seyfert 2 galaxies
exhibit broad Fe K- 
lines whose profiles indicate that they
they are emitted from an accretion disk observed face on,
rather than edge on as predicted by the unified model [Turner et al., 1998].