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

6.2. The Importance of Fluids: Radial Motions

The observations strongly suggest that the nucleus of an active galaxy is in communication with its exterior. The galactic environment can both control the rates of energy and fuel delivery to the nucleus and confine and organize the by-products (radiation and energetic particles) of nuclear activity. Some sort of fluid, presumably aboriginal gas or the debris from disrupted stars, is the most efficient agent for delivering kinetic energy, fuel, and angular momentum to the nucleus.

However, a fluid is likely to be ejected as a by-product of nuclear activity, and outward-directed fluid can limit or arrest the delivery process into the nucleus. Consequently, we consider the question and origin of radial motions in both active and nonactive galaxies. Are radial motions dominated by infall or outflow? Can such motions be directly related to the causes or effects of nuclear activity?

OUTFLOW     Oort (1977) summarized the importance of explosive phenomena in a very "normal" galaxy; the Milky Way. Two papers in 1978 (Burton & Liszt 1978, Liszt & Burton 1978) emphasized the need for an expanding, tilted disk to explain gas motions near the Galactic plane. More recently, however, Liszt & Burton (1980) concluded that the data were also compatible with expected radial-streaming motions associated with a bar. Therefore the case for outflow is ambiguous. However, there is much to be understood about motions in the Galaxy. For example, why does ionized gas in the immediate vicinity of the Galactic center have an average radial velocity of more than 50 km s^-1 (e.g. Wollman et al. 1977) with respect to the Sun?

Radial motions in nearby galaxies have been reviewed by van der Kruit & Allen (1978), and Burbidge's (1970) review is also relevant. The data are extensive and cannot be fully discussed here. Although radial motions near the nuclei of galaxies are not uncommon, an interpretation of motions as infall or outflow is impossible without knowledge of the effects of sky projection. The same holds for active galaxies, even where velocities consistently exceed the escape velocity of the system (see e.g. Heckman et al. 1981b). Balick & Heckman (unpublished) find that the ionized gas in several edge-on active galaxies is confined to the plane (e.g. IC 4329A, in which nuclear outflowing gas has been proposed by Pastoriza 1979, but not supported fully by Wilson & Penston 1979). There are some important counterexamples generally associated with unrelaxed systems (e.g. NGC 2992, which is in an interacting system with NGC 2993, exhibits outflow not confined to a disk; see Heckman et al. 1981b). The uncertainty of the gas geometry is a particularly poignant issue in the case of the complicated active galaxy NGC 1068 (Walker 1968, Pelat & Alloin 1980, Alloin et al. 1981) in which very strong and complex radial motions are observed near the nucleus. Yet in one nearby normal galaxy, M31, outflow from the nucleus appears to be incontestable (Rubin & Ford 1971, Morton & Anderdeck 1976), and a strong case has been made for outflow in M81 (Goad 1974, 1976).

Whether outflow is related to nuclear activity is an independent question. In elliptical galaxies, winds driven by supernovae can sweep the ISM outward in a "galactic wind" (Bregman 1978). Global radial motions can and do occur in galaxies with nonaxisymmetric potentials and without an active nucleus (Section 3.4), in interacting systems (Section 2.2), or near Lindblad resonances and streams in spiral arms associated with density waves (e.g. van der Kruit & Allen 1978). Schommer & Sullivan (1976) have, for example, interpreted radial motions in NGC 4736 without need of the explosive activity proposed by van der Kruit (1974a, 1976b).

Radial motions are particularly easy to detect in active galaxies owing to their bright emission lines, However, the caveat of "normal" radial motions has been sometimes ignored in the interpretation of the kinematics of Seyferts, most all of which show evidence for oval distortions. For example, the kinematics of the ionized gas in NGC 4151 (Anderson 1974, Fricke & Reinhardt 1974, Simkin 1975) was first thought to indicate violent outflow (or infall) owing to the incorrect identification of the galaxy's rotation axis. Such motions are now believed to indicate rotation plus some noncircular motions associated with the galaxy's fat bar (Bosma et al. 1977, Heckman et al. 1981b). The disturbed galaxy NGC 3310 has very strong noncircular motions, which might have been misinterpreted as outflow were it not for the detailed H kinematic studies of van der Kruit (1976a), who showed that density-wave motions also explain the data. Global noncircular motions in the luminous disturbed spiral NGC 253 (Ulrich 1978, Beck et al. 1979, Pence 1981) can be qualitatively understood by models of gas flow in a barrel potential (Pence 1981). However, the motions in the core of the galaxy are apparently dominated by outflow (Gottesman et al. 1976, Ulrich 1978), which may be unrelated to the global oval distortion. Further, at least some active galaxies show little evidence for radial gas motions (e.g. Ulrich et al. 1980, Rubin & Ford 1968). Some inactive galaxies (by our criteria) such as M82 (Gottesman & Weliachew 1977, Cottrell 1978, O'Connell & Mangano 1978, Axon & Taylor 1978, Solinger et al. 1977, Watson & Griffiths 1980), NGC 1569, (Hodge 1974, de Vaucouleurs et al. 1974), and NGC 5253 (Sersic et al. 1972, Graham 1981) have strong radial flows.

Motions exceeding ~ 10³ km s^-1 are far in excess of the escape velocity of a galaxy of any reasonable mass. It is difficult to envisage such motions as the result of free-fall or rotation. Examples of active nuclei near which such motions are seen include DA 240 (Burbidge et al. 1975, 1978), AO 0235+164 (Smith et al. 1977), and NGC 1275 (Section 4.3). NGC 1275 is apparently an example of a high-velocity encounter of a spiral and elliptical (i.e. infall) which, if seen at much larger distances, could go unrecognized. Explanations for the other objects are needed.

Despite our warnings concerning the existence and interpretation of radial motions, there are some fairly clear examples of nuclear ejecta in active galaxies. The anomalous H / radio arms of NGC 4258 appear to result from collimated outflow from the nucleus (van der Kruit et al. 1972, van der Kruit 1974b, van Albada & Shane 1975, Icke 1976, 1979, van Albada 1980, Sanders 1981b). Heckman et al. (1981a) similarly interpret the bright radio / H lobes in 3C 305. Graham & Price (1981) and De Young (1981) propose that nebulae along the NE jet in Cen A are a result of outflow. And, of course, the geometry of radio lobes, jets, etc. strongly supports the idea of collimated outflow in all extended radio galaxies (e.g. Miley 1980). Direct evidence of outflow is seen in expanding nuclear radio sources (cf. Kellermann & Pauliny-Toth 1981). Heckman et al. (1981b), Capriotti et al. (1979, 1980, 1981) and others have argued that the line shapes of Seyfert galaxies are best explained as nuclear outflow, although this interpretation is model-dependent (e.g. Whittle 1981). Blueshifted H I absorption lines observed in three Seyferts (Heckman et al. 1978, 1981b) and a blueshifted radio recombination line in the Seyfert OQ 208 = Mrk 668 (e.g. Bell & Seaquist 1980) further suggest that outflow occurs in at least some Seyferts. The broad blueshifted optical absorption lines in front of a few bright, distant QSOs (Turnshek et al. 1980, Turnshek 1981, Weymann et al. 1981) and their narrower counterparts in the Seyfert galaxies Mrk 231 (Boksenberg et al. 1977) and NGC 4151 (Anderson & Kraft 1969, Cromwell & Weymann 1970, Anderson 1974, Penston et al. 1979) also provide direct evidence for outflow in active systems.

INFALL     The evidence for infall is sketchy and infrequent. Of course, the caveats concerning projection effects and the naive interpretation of radial motions being related to galactic activity apply to infall as well as outflow. There is strong indirect evidence for infall in massive cluster-centered ellipticals, in which accretion flows are alleged to exist (Sections 2.1, 4.2, and 4.3). Direct evidence is sometimes found in redshifted H I lines, most notably in Cen A where such lines are unique to the nuclear radio source (Section 4.2). Redshifted optical absorption lines exist, but are very rare in QSOs (see references in the previous paragraph), and the link between such features and galactic activity is speculative.

Of course, there may be selection effects that hide infalling gas from detection. For example, a free-falling fluid will have characteristic speeds approaching the escape velocity, and will heat up until the sound speed and escape velocities are comparable. Such a medium will be characterized by temperatures near 10⁶ K (e.g. Silk & Norman 1979, Bregman 1980). A fluid of this type is too hot to detect optically, yet too cool or underluminous to be observed as an X-ray halo. Potentially observable cooling instabilities in the fluid can lead to infall under some conditions (Bailey 1980, Loose & Fricke 1981). Of course, infall will be common if not ubiquitous in interacting systems, because gas in one galaxy is likely to have small angular momentum with respect to the nucleus of the second galaxy for wide ranges of impact conditions. VLA H I studies of radio nuclei in interacting systems will be important in this regard.

It is also possible that both infall and outflow are germane. Three appropriate mechanisms have been suggested.

1. Sanders (1981a) has argued that infall and outflow can occur in alternate cycles. Outflow dominates until the reservoir of fluid is exhausted, then infall takes over until the mechanism that sustains outflow is reactivated. This is the "hot gust" model. Bailey (1980) has a similar model in which intermittent thermal instabilities allow gas to fall periodically onto the nucleus.

2. The outflow may be collimated; this is nearly always the ease in radio galaxies (Miley 1980) and is common in radio Seyferts (Wilson et al. 1980, Ulvestad et al. 1981, Wilson & Ulvestad 1981, and references therein). As suggested by the radio observations, the degree of collimation can be so high that infall to the nucleus in most directions can occur without resistance.

The infalling material may arrive near the nucleus as numerous small dense clouds with a high ratio of mass to surface area. Such clouds, perhaps produced initially in thermal instabilities in an accretion region (Ford & Butcher 1979) or behind shocks (Marscher & Weaver 1979), might move freely through a low-density outbound wind. Taken to an an extreme, the infalling clouds might actually be stars that are disrupted on close approach to the nucleus (Hills 1978, Shields & Wheeler 1978, McMillan et al. 1981, Norman & Silk 1981, Roos 1981b). This last mechanism might be very important in BL Lac objects, which prefer luminous gas-free galaxies as their hosts.