Annu. Rev. Astron. Astrophys. 1982. 20: 431-468 Copyright © 1982 by Annual Reviews. All rights reserved

4.2. Neutral and Molecular Gas

Gas, like dust, is a common constituent of all spiral galaxies, active and inactive. Unlike dust, gas can be studied through emission lines, making it possible to determine column densities, excitation temperatures, and kinematics.

We first consider early-type galaxies whose gaseous content is normally (but not uniformly) below ~ 10^8.5 M, i.e. under present detection limits, Nonetheless some ellipticals do, in fact, show substantial amounts of HI. Sanders (1980) has shown that the distance-independent ratio of HI mass to blue optical luminosity (M_HI / L_B in elliptical galaxies has a bimodal distribution: 70% of ellipticals are gas-poor (M_HI / L_B < 0.003 in solar units) and the remainder are gas-rich galaxies (M_HI/L_B ~ 0.03). As noted by Sanders (1981a) and Raimond et al. (1981), gas-rich elliptical galaxies are generally active. O'Connell & Dressel (1978) find a significant link between the presence of [O II] 3727 emission (generally a sign of nuclear activity according to Heckman 1980a) and detectable H I in ellipticals. The analogous link between HI and a compact nuclear radio source is less well established, however.

Spatial-kinematic maps of NGC 4278 and NGC 1052 are now available (Raimond et al. 1981, Knapp et al. 1978a, b, Reif et al. 1978, Bottinelli & Gougenheim 1980, and references within). In the case of the relatively simple system NGC 4278, the HI appears to lie in a spinning disk whose rotation axis does not lie along the minor axis of the galaxy. The observations require a radial component of motion, which may be related to oval orbits in a bar-like gravitational potential, or "sloshing" motions as a recently captured cloud settles into a stable configuration.

Although gas-rich ellipticals are generally active, it is not clear whether active ellipticals, as a class, are gas-rich. Strong radio emission, characteristic of active ellipticals, makes the detection of HI emission difficult in practice. By the same token, such ellipticals are ideally suited for searches of HI in absorption. However, a surprisingly small fraction of radio galaxies, whether ellipticals or spirals, exhibit HI absorption (see reviews by Roberts & Steigerwald 1977, Burke 1978; see also Heckman et al. 1978, Bieging & Bierman 1981, and Burns et al. 1981b). Even so, several radio galaxies exhibit HI absorption: NGC 1275 (Crane et al. 1981, van Gorkom & Ekers 1981 and references therein), Cen A (see below), 3C 293 (Baan & Haschick 1981), 3C 305 (Heckman et al. 1981a), and NGC 315 and NGC 1052 (van Gorkom, private communication) are some recently detected examples. However, HI absorption is as yet undetected even in some bright radio ellipticals with dust lanes crossing near their nuclei (e.g. Cyg A). Presumably the paucity of HI absorption detections reflects the geometry of the HI distribution relative to the radio source (searches for HI absorption in systems in which the radio source is mostly within the galaxy would be interesting in this regard), or it may be that hydrogen is not atomic in most ellipticals (see the recent discussion by Burns et al. 1981b).

Cen A is of particular interest since it shows both HI emission and absorption (Gardner & Whiteoak 1976). Aside from velocities affected by absorption, the HI emission line has a shape typical of late-type systems which, for such galaxies, could suggest that HI extends through the inner regions of the galaxy. An H I mass of ~ 10⁹ M and a total virial mass 10^11.5 M are suggested. Assuming that no H I has been ejected or converted into nonatomic form, the original spiral may have been a small late-type galaxy.

Gardner & Whiteoak (1976) observed a series of HI and molecular absorption lines against the central parts of Cen A. All of these lines are seen in absorption between 541 and 551 km s^-1, close to the HI systemic emission-line velocity of 535 km s^-1. Additional weaker HI absorption lines extend redward some 200 km s^-1 in velocity. High spatial resolution observations at the VLA (van der Hulst & Haschick, private communication) show that the clouds producing these highly redshifted absorption lines are seen only against the radio nucleus, even though there are several other nearby radio sources lying behind the dust lane. Similar redshifted HI absorption lines are seen in the radio galaxies NGC 315, NGC 1052 (van Gorkom, private communication), NGC 1275 (Crane et al. 1981), and in several other kinds of galaxies (e.g. Haynes & Giovanelli 1980, Thuan & Wadiak 1982), but the spatial distribution of the redshifted absorbing material is not known. Gordon & Gottesman (1981) suggest that infalling HI may be common, at least in very-late-type galaxies.

The peculiar motions of gas in the dust lane (Mollenhoff 1981) cannot account for the high-z HI features in Cen A. Thus it is tempting to associate these HI absorption features with clouds falling into the active nucleus, as the observations of van der Hulst & Hasehick strongly suggest. However, several caveats must be stated. First, Mollenhoff (1981) and H. Harris (private communication) present evidence that the systemic velocities of the disk emission-line gas and the system of globular clusters are significantly higher than that of the HI emission. Second. several cases of HI absorption blueshifted by 100 km s^-1 or more are also known among Seyfert (Heckman et al. 1978, 1981b) and normal (Gottesman et al. 1976) galaxies. In 3C 293 (Baan & Haschick 1981), HI absorption occurs to the blue and red side of the systemic velocity. Detailed maps of HI absorption, together with the collection of a sample of cases that is large enough for statistical investigation, is required in order to understand the significance of these HI absorption features.

The gas-rich disks of spiral galaxies can be a handy reservoir of fuel for galactic activity. HI surveys have been conducted by Heckman et al. (1978 - hereafter HBS), Biermann et al. (1979 - hereafter BCF), and Bieging & Biermann (1981 - hereafter BB), and observations of individual galaxies abound in the literature. The common properties of HI in active spirals, and some notable exceptions, include the following.

1. The HI masses and M_HI / L_B ratios of active galaxies are generally normal for galaxies of their morphological classes. Yet Mrk 348 remains one of the most extreme examples of large M_HI / L_B and HI-to-Holmberg radius ratios (HBS, Morris & Wannier 1980, Heckman et al. 1981c). Because of its normal colors (Huchra 1980), Mrk 348 has not recently formed, but star formation may have been delayed or suppressed (Hawarden et al. 1979b). Heckman et al. (1981c) propose that much of the HI in Mrk 348 was stripped from the presently gas-poor spiral NGC 266 about 3 × 10⁹ yr ago; however, the gas has not yet relaxed. Other active galaxies with grossly anomalous surplusses of HI are PKS 1718+649 (Fosbury et al. 1977), Mrk 298 = IC 1182 (Bothun et al. 1981b), the "quasar" 0351+026 (Bothun et al. 1981a), and the type I Seyfert galaxy NGC 931 (G. Bothun, private communication). Another important oddball is NGC 1068, which is surprisingly depleted of HI for its luminosity (HBS). In this case, L_B may be enhanced by the disk of young stars lying near the nucleus (see Section 3.5).

2. HI profiles of several active galaxies show evidence of missing horns and high-velocity wings, consistent with tidal disruption of their disks (see Section 2.2).

3. HBS noted a propensity for the galaxies with the most luminous nuclear activity to exhibit the most unusual HI properties. It is possible to interpret this in terms of either the causes or results of galactic activity. BB find no correlation between the HI properties of active galaxies and the characteristics of their nuclear spectra or their IR luminosities at 10 µm. However, they did notice a relationship between HI mass, nuclear X-ray luminosity, and optical stellar luminosity, which suggests that the galaxies with the largest bulge luminosities (i.e. masses) are the most active (Section 3.3).

Recent HI studies of other individual active disk and peculiar galaxies include NGC 4151 (Bosma et al. 1977), IC 5063 (Danziger et al. 1981), and NGC 4258 (van Albada & Shane 1975, van Albada 1980). These galaxies have notable characteristics (e.g. 0351 + 026 has a HI line width exceeding 1000 km s^-1). NGC 2685, a peculiar SO with a mildly active nucleus, appears to have disks of HI perpendicular to its major and minor stellar axes (Shane 1980).

Finally we mention observations of molecules in active galaxies. Three of the five initial detections of CO (Rickard et al. 1977) were made in highly peculiar galaxies, and galactic activity was first suspected to be related to the production of CO. More recent surveys suggest that active galaxies are probably similar to normal galaxies in their CO properties. NGC 1068 is an outstanding exception; in addition to its abundant CO, H₂ is also plentiful (Thompson et al. 1978), perhaps collisionally excited (Carlson & Foltz 1979). The H₂ is concentrated in a zone comparable in extent to the ring of young stars and HII regions about one kpc from the nucleus (Section 3.5). However, the relevance of this region to the question of the evolution of activity is dubious.