9.1. Molecular gas
Seyfert galaxies are rich in interstellar dust, and as expected from this, also contain large amounts of molecular gas. CO, one of the most abundant molecules in molecular clouds in our Galaxy, and the most readily detectable, has in fact been observed in many Seyferts (Bieging et al 1981, Heckman et al 1989). The observational data seem to show that the Seyfert 2 galaxies have larger molecular gas contents than Seyfert 1s. This is one of the few remaining observational differences which a to conflict with the idea that Seyfert 1 and 2 galaxies are essentially similar systems seen from different orientations. To date no broad-line CO has been detected; the line profiles of the observed CO are most nearly similar to those of the H I = 21 cm lines in the same object. This suggests that the molecular gas is associated more with the disk of the galaxy than with the NLR (Heckman et al 1989). Molecular H2 has only forbidden emission lines in the infrared, and is much more difficult to detect. However, three low-excitation lines of its (1, 0) vibrational band have been measured in NGC 1068, with relatively narrow widths (Oliva and Moorwood 1990). There is no reason to suppose that the abundance ratio CO / H2 is appreciably different from that in our Galaxy.
Very recently high-resolution interferometric measurements have been made of CO emission in two Seyfert galaxies, NGC 3227 and NGC 7469. At 6" resolution the latter shows a concentration of molecular gas with a mass ~ 4 x 109 M and diameter ~ 1.5 kpc, centered on the nucleus (Sanders et al 1988a). The higher resolution (2-3") observations show that in both these galaxies a large fraction of the molecular gas is concentrated within a few hundred parsecs of the nucleus. In NGC 3227 the central structure is partly resolved, and suggests that a significant amount of the CO lies in a roughly toroidal region, centered on the nucleus. Its velocity field includes a rotational component, with apparently other complications as well. In NGC 7469 the molecular gas distribution is more complicated, but seems to have a similar overall pattern. On the other hand in the third Seyfert galaxy of which interferometer measurements have been reported to date, NGC 5033, the CO emission is more diffuse, spread over several kiloparsecs (Meixner et al 1990). Clearly, high resolution maps of the CO emission in more Seyfert galaxies, compared with atomic-line and dust continuum maps, will be very useful for understanding the structure of AGNs and their surroundings.
As stated in section 2.3, all the observational evidence agrees that in AGNs photoionization by hard photons from the central source is the dominant mechanism of energy input to the ionized gas. These high-energy photons are generally considered to be generated by non-thermal processes such as relativistic synchrotron radiation and Compton scattering, or in the inner parts of the accretion disk itself, as described in section 3.3. However, an alternate view is that in Seyfert 2 galaxies and LINERs, the photoionizing radiation may come from stars, rather than from non-thermal sources. The most recent form of this idea is the suggestion by Terlevich and Melnick (1985) that the ionization is produced by high-energy photons from postulated `extremely hot and luminous stars', which they have named `Warmers'. They are supposed to be `bare core Wolf-Rayet stars,' the descendents of massive progenitors (M > 60 M), with temperatures up to 2 x 105 K. Observational evidence that any such stars exist is scanty indeed, and theoretical justification that such high effective temperatures actually occur is very weak. However, if sufficiently many stars of arbitrarily high temperature are postulated, clearly any type of power-law, broken power-law or more complicated form of photoionizing spectrum can clearly be matched by some population of such objects. Of course many starburst galaxies are known to exist, with emission-line spectra like H II regions, photoionized by OB stars. They are the objects marked by circles and asterisks in figures 2, 3 and 4. The AGNs, to be understood on this basis, must have much more luminous and hotter stars, and many more of them.
The Warmer picture of AGNs has been discussed and defended by Terlevich et al (1987). They have attempted to explain the x-ray and radio-frequency properties of AGNs, and some of the emission-line broadening, to supernovae produced from the massive, rapidly evolving stars. However, the rapid optical and especially x-ray variations in many AGNs are extremely difficult to explain by this picture. Likewise jets, polarization of the optical continuum on Seyfert 2 galaxies and the apparent dichotomies of figures 2, 3 and 4 are not, at present, understood on this basis. Neither is the apparent continuity in the physical properties of LINERs, Seyfert 2 and Seyfert 1 nuclei explained by it.