3.4. Relation between Dust and Gas Absorption

According to the standard paradigm of the unified model, X-ray absorption should be observed in optically type II objects. This is indeed what has been found in many local AGN. However, recent observations challenge this simple view and suggest a more complex scenario.

A direct measurement of the dust-to-gas ratio is possible in objects with intermediate optical classification, i.e., Seyfert 1.8 or 1.9. These show clear signs of absorption in the optical/near-IR but still have broad components in some of the brightest emission lines. In these cases, an estimate of the dust absorption can be obtained from the ratio of the emission line fluxes (typically, the hydrogen lines) and can be compared with the absorbing column density measured in the X-rays.

For a Galactic dust composition and dust-to-gas ratio, the relation between optical extinction and X-ray absorption is A_V ~ 4.5 × 10^-21 N_H. Therefore, one would expect objects with optical broad lines to have X-ray column densities not higher than 10²² cm^-2. Maccacaro, Perola, & Elvis (1982) and Maiolino et al. (2001) analyzed a sample of bright, intermediate Seyferts and found that X-ray absorption is systematically higher than expected from optical extinction by a factor of ~ 10. The physical explanation of this result may be a lower than Galactic dust-to-gas ratio, or a different composition of dust grains. The observed SEDs of these objects can be significantly different than standard type I and type II templates, with lower UV, larger IR, and heavily absorbed X-rays (Ward et al. 1982).

Moving to more extreme cases, BAL quasars are objects with no or little dust extinction in the optical/UV, but with broad, blueshifted, and often saturated absorption lines. Many of them also only have extremely faint X-ray emission (note, however, that there are exceptions, such as the BAL quasars found in deep X-ray surveys; Barger et al. 2002).

There is now convincing evidence that these objects are intrinsically normal quasars, covered by a high column density of dust-free gas that is responsible for heavy absorption in the X-rays and absorption lines in the optical/UV. A strong correlation between these two absorption features has been found by Brandt et al. (2000), and recent X-ray observations are starting to directly measure the X-ray column densities of these objects. A clear example is the BeppoSAX observation of the X-ray weak BAL quasar MKN 231, which revealed powerful hard X-ray emission above 10 keV obscured by a column density N_H > 10²⁴ cm^-2 (Braito et al. 2004, see Fig. 12).

Figure 12. XMM-Newton and BeppoSAX spectrum of MKN 231. Figure obtained using data and model of Braito et al. (2004). Model shown here is a best fit to the 0.5 - 10 keV emission. The large excess at E > 10 keV is due to the intrinsic emission of the AGN, which is absorbed by a column density NH ~ 2 × 1024 cm-2.

Other examples of objects with strongly differing optical and X-ray absorption are found in quasar surveys where selection criteria other than optical/UV color are used. An interesting example is the sample of "red" quasars discovered with the 2MASS near-IR survey (Cutri et al. 2001). The selection criterion adopted in this case is J - K > 2, which is efficient at low redshifts where the minimum in the quasar emission, due to the sublimation temperature of dust, is observed in the J band. Chandra observations of a sample of these "red", yet broad-line objects (Wilkes et al. 2002), revealed that they are extremely faint in the X-rays, probably due to absorption by a column density of order 10²³ cm^-2.

Another case of X-ray weakness in optically type I quasars is found in the sample of quasars from the Hamburg survey (Hagen et al. 2001) which also have ROSAT observations (Risaliti et al. 2001). Most of these objects, which are slightly redder in the optical than standard blue quasars (the selection criterion was based both on blue color and on low resolution spectroscopy), are undetected by ROSAT, contrary to what was expected assuming a "normal" _OX. Subsequent Chandra observations of a subsample of these objects revealed that they are underluminous in the X-rays by a factor of ~ 3 to ~ 100 with respect to PG quasars (Risaliti et al. 2003a). It is not clear whether the observed X-ray weakness is an intrinsic property of these objects (as most of the Chandra spectra seem to suggest) or whether it is due to absorption. In any case, they are part of a population of quasars with SEDs different than that of standard type I or type II AGN.

In addition to the cases described above of optically type I AGN with the X-ray properties of type II AGN, several examples are known of the opposite case, i.e., objects with type II optical properties and no hint of absorption in the X-rays. Panessa et al. (2003) described a sample of local sources optically classified as Seyfert 2 galaxies with no measured X-ray absorbing column density in excess of the Galactic value.