Next Contents Previous

4. GALAXIES WITH BROAD HI ABSORPTION LINES

4.1. The Group of Nearby Starburst Galaxies

This section describes a group of relatively nearby spiral galaxies showing a variety of activities. The original selection criterion for this sample was the presence of broad HI absorption lines – a quite unusual criterion at first glance, but a very valuable one in the later interpretation of the overall kinematics in these galaxies. Such broad absorption lines are quite peculiar, because they indicate a large range of velocities both blue- and red-shifted in a very tiny region just in front of the radio continuum emission. Since such a broad HI absorption line was detected in NGC 1808 (see below) I wanted to know its origin and the relation to the surrounding medium in the galaxy. The galaxies with broad HI absorption lines can be divided into two different groups, one consisting of relatively nearby luminous spirals (see Tables 1, 2 and 3) and the other of distant ultra-luminous mergers (see Table 4). In the following I will take a closer look at the members of the first group.

Starbursts are thought to be the result of gas accretion toward the central regions of galaxies. There are mainly two mechanisms known which can transport mass into this region: either tidal interactions or bars (see e.g., Combes 1988). I suggest that the broad HI absorption lines in the central region of most luminous spirals (see Table 1) are caused by a fast-rotating ring of cold gas. The accretion of gas in the inner region (probably near the inner Lindblad resonances) could have been induced by the bar potential of these galaxies (see e.g., Combes & Gerin 1985). Since outflow of gas from the central region is another phenomenon observed in this group of galaxies I propose a model in which the neutral gas is fuelling the nuclear starburst where it is partly ionized and then ejected by supernova explosions and stellar winds (see Section 5).

Table 2. Some basic parameters of the selected galaxies

Name α(2000) δ(2000) Type i v21
[ h m s] [° ' "] [°] [km s-1]

NGC 253 00 47 33.1 -25 17 18 .SXS5.. 78 251
NGC 660 01 43 01.4 +13 38 37 .SBS1P. 65 853
NGC 1068 02 42 40.2 -00 00 48 RSAT3.. 40 1137
NGC 1365 03 33 36.6 -36 08 17 .SBS3.. 60 1662
NGC 1808 05 07 42.8 -37 30 51 RSXS1.. 57 1005
NGC 3034 09 55 54.0 +69 40 57 .I.0../ 82 203
NGC 3079 10 01 58.2 +55 40 43 .SBS5./ 85 1125
NGC 3628 11 20 16.3 +13 35 22 .S..3P/ 82 847
NGC 4945 13 05 26.2 -49 28 15 .SBS6*/ 78 560
Circinus 14 13 10.2 -65 20 20 .SAS3*. 60 438
NGC 6221 16 52 46.7 -59 12 59 .SBS5.. 43 1482
NGC 7552 23 16 10.9 -42 35 01 PSBS2.. 28 1585
NGC 7582 23 18 23.3 -42 22 11 PSBS2.. 64 1575

Column (1): Galaxy name. Columns (2) & (3): Galaxy position. Column (4): Galaxy type. Column (5): Approximate galaxy inclination. Column (6): Mean heliocentric radial velocity derived from HI observations. References: All parameters except the inclination are taken from the Third Reference Catalog (RC3) by de Vaucouleurs et al. (1991). The galaxy inclination has either been taken from the HI catalog by Huchtmeier & Richter (1989) or references mentioned in the summary of the individual galaxies.

NGC 253 is a nearby edge-on galaxy and the brightest member of the Sculptor group. Its nuclear region is particularly active, revealing numerous compact radio sources (Antonucci & Ulvestad 1988), as well as bright optical and infrared emission lines. The strong and extended radio emission makes it especially suitable for HI absorption measurements (see e.g., Combes, Gottesman & Weliachew 1977; Dickey, Brinks & Puche 1992; Koribalski, Whiteoak & Houghton 1995). Gardner & Whiteoak (1974) also measured broad H2CO and OH absorption lines. NGC 253 reveals a bar which has been detected in the near-infrared (Scoville et al. 1985; Forbes & DePoy 1992), optical (Pence 1980), and – on a much smaller scale – in the CO emission (Canzian, Mundy & Scoville 1988; see also Mauersberger et al. 1996).

Various authors have also indicated the possibility of large-scale outflow of ionized gas into the halo (see Schulz & Wegner 1992, and references therein). Extended X-ray emission has been observed by Fabbiano & Trinchieri (1984) and Pietsch & Trümper (1993) using the Einstein- and ROSAT satellites, respectively.

Figures 2 and 3 show the HI gas distribution and velocity field of NGC 253, respectively. The emission is distributed rather similar to the optical light and shows a rather regularly rotating spiral galaxy (Koribalski, Whiteoak & Houghton 1995). The HI absorption against the central continuum sources causes the prominent `hole' in the centre of the distribution. The major-axis position-velocity diagram of NGC 253 is rather similar to that of NGC 1808 Fig. 1), with one major difference: here the absorption spectrum is not quite as broad as the emission spectrum and it is asymmetric with respect to the systemic velocity (see Table 1). Whereas the large width of the absorption is most likely caused by a fast-rotating nuclear ring, the asymmetry of the line has been interpreted as due to gas outflow from the nuclear region. A detailed analysis of the HI data is in progress.

Figure 2

Figure 2. (top) HI distribution of the starburst galaxy NGC 253.

Figure 3

Figure 3. (bottom) HI mean velocity field of NGC 253 (for a full description see Koribalski, Whiteoak & Houghton 1995). The data were taken with three different arrays of the ATCA (12 h each).

Table 3. Far-infrared flux, luminosity and derived star-formation rate

D FFIR LFIR SFR
[Mpc] [10-11 Wm-2] [109 L] [Myr-1]

NGC 253 3.4 5.59 (a) 20 5.2
NGC 660 11.2 0.35 (a) 14 3.6
NGC 1068 15.2 0.97 (b) 71 18.4
NGC 1365 20.5 0.51 (a) 68 17.6
NGC 1808 10.9 0.57 (c) 21 5.5
NGC 3034 3.25 5.84 (a) 19 4.9
NGC 3079 15.6 0.25 (b) 19 4.9
NGC 3628 10.2 0.31 (a) 10 2.6
NGC 4945 6.7 3.70 (a) 52 13.5
Circinus 3.6 1.21 (d) 5 1.3
NGC 6221 17.7 0.22 (d) 22 5.6
NGC 7552 20.9 0.36 (d) 50 12.9
NGC 7582 20.7 0.25 (d) 34 8.8

Column (1): Galaxy name. Column (2): Adopted distance, assuming H0 = 75 km s-1 Mpc-1. Column (3): Far-infrared flux as calculated from the IRAS 60 µm fluxes. References: (a) Rice et al. 1988, (b) Young et al. 1989, (c) Danks, Perez & Altner 1990, (d) IRAS Point Source Catalog 1985. Column (4): Far-infrared luminosity (∝ D2). Column (5): Star-formation rate: SFR(M ≥ 0.1 M) = 0.26 × column(4), see Hunter et al. (1986).

NGC 660 is a rather peculiar galaxy for which two kinematical systems exist: a nearly edge-on disk (PA = 46°, i ≈ 80°) and an inclined polar ring or strongly warped outer disk (PA = -7°, i ≲ 60°) (Gottesman & Mahon 1990; Baan, Rhoads & Haschick 1992; Combes et al. 1992). The 6-cm continuum map by Condon et al. (1982) shows quite extended emission along the inner disk with two peaks roughly 4 apart. HI position-velocity diagrams obtained by Gottesman & Mahon along the major axes of both systems show a broad absorption line the varying column density of which resembles the (here unresolved) continuum structure. This and the map of the integrated mean velocity dispersion shown by the authors hints at a fast-rotating nuclear ring at PA = 46°, thus belonging to the disk. Further, there are indications for a stellar bar (Young, Kleinmann & Allen 1988) and very tentatively for nuclear outflow (Gottesman & Mahon 1990). Baan, Rhoads & Haschick (1992) obtained both high-resolution HI and OH absorption measurements with the VLA. They found that a central disk (diameter ∼ 6 or 300 pc) with a large velocity gradient (see Table 1) dominates the absorption signature. The companion UGC 01195, an irregular, distorted galaxy is located 22 or 72 kpc away.

NGC 1068 is a barred galaxy with a very complex nuclear region. Although it is an archetype Seyfert2, at least half of the IR emission arises from star formation (Young, Kleinmann & Allen 1988). High-resolution (∼ 2) VLA HI absorption measurements carried out by Gallimore et al. (1994) reveal several blue- and red-shifted features against the SW radio lobe and a jet as well as broad (FWHM = 128 ± 25 km s-1) double absorption lines against the radio nucleus. The absorption lines cover a total velocity range of ∼ 540 km s-1, from ∼ 760 to 1300 km s-1. The two lines against the core are centred at +66 ± 12 km s-1 and -295 ± 9 km s-1, resulting in a mean offset from vsys of -181 km s-1. Because of this large offset Gallimore et al. reject circular orbits (which are favoured for many of the galaxies presented here), but prefer a model with two distinct, near-nuclear cloud populations, one falling into and the other flowing out of the radio core, both associated with the fuelling of and exhaust from the central engine. In case of a fast-rotating nuclear ring I would expect the radius to coincide with the inner Lindblad resonance (ILR) induced by the bar potential. It is possible that the absorption lines observed in NGC 1068 are caused by outflow and a nuclear ring as suggested for NGC 253. For comparison see also the summary of HI absorption measurements against the Seyfert galaxy NGC 4151 and the peculiar galaxy NGC 4258 in Section 4.3.

A detailed discussion of NGC 1068 and NGC 4151 is given in the chapter by Brinks & Mundell in this volume.

NGC 1365 is a southern, strongly barred spiral galaxy with no apparent companions. VLA HI observations by Ondrechen & van der Hulst (1989) show the overall gas extent of the galaxy. The central region contains a Seyfert nucleus as well as circum-nuclear radio continuum and Hα emission, indicating star formation (Saikia et al. 1994; Sandqvist, Jörsäter & Lindblad 1995). Ondrechen & van der Hulst find a strong absorption feature centred at 1570 km s-1 with a width of only 33 km s-1. But their position-velocity diagrams indicate a much broader (Δv ≳ 200 km s-1) feature which is confirmed by Jörsäter & van Moorsel (1995) who find a total velocity width of 313 km s-1 for the HI absorption. Our own ATCA snap-shot observations of NGC 1365 and also NGC 1097, another strongly barred galaxy, clearly reveal HI absorption over a broad velocity range. The broad absorption features are most likely caused by a rapidly rotating ring/disk of neutral gas. For NGC 1365 a similar disk of ionized matter with a radius of 7 has already been inferred by Lindblad (1978). NGC 1097 is in many respects very similar to NGC 1365 (see Ondrechen, van der Hulst & Hummel 1989). The continuum sources of both galaxies are displayed in Fig. 4.

Figure 4

Figure 4. 20-cm radio continuum emission of the starburst galaxies NGC 1365 and NGC 1097. These data were taken with the largest array of the ATCA (snap-shot). The beam has been convolved to 8.

NGC 1808 is a peculiar southern spiral galaxy and rather nearby example of nuclear starburst activity (LFIR ≈ 2 × 1010 L); for a comprehensive overview see Koribalski (1993). Well known are the prominent dust filaments which emerge nearly perpendicular to its central disk (Véron-Cetty & Véron 1985; Laustsen, Madsen & West 1987), similar to those observed in M 82, indicating continuing disk-halo interactions. The nuclear region contains a number of bright `hot spots', which are mostly interpreted as very bright HII regions (Sérsic & Pastoriza 1965; Alloin & Kunth 1979; Koribalski & Dettmar 1993). In the actual nucleus, however, Véron-Cetty & Véron drew attention to the presence of a broad Seyfert-like component of Hα with FWHM about 550 km s-1. Radio observations at 6 cm reveal a number of very small (< 1") compact sources in the central region, most likely to be supernovae and supernova remnants (Saikia et al. 1990).

VLA measurements of its nuclear region reveal that the broad (Δv ≈ 360 km s-1) HI absorption line seen at low angular resolution is a much narrower line which shifts its centre velocity over the face of the continuum (Koribalski, Dickey & Mebold 1993). This is interpreted as a torus of cold, dense gas with a rotation velocity of ∼ 250 km s-1 and radius 500 pc. The gravitational mass required to explain this fast rotation is a few times 109 M. Figures 5 and 6 display the overall HI distribution and velocity field of the galaxy NGC 1808. The central `hole' is caused by HI absorption.

Figure 5

Figure 5. HI distribution of the starburst galaxy NGC 1808.

Figure 6

Figure 6. HI mean velocity field of NGC 1808 (for a full description see Koribalski 1993). These data were taken with the BnC and AnB arrays of the VLA (5 h each).

The radius and velocity of the nuclear ring are comparable to the distribution and kinematics of the optically visible `hot spots'. The possibility of a molecular gas ring has been considered by Dahlem et al. (1990) and is also indicated in new CO(2-1) data by Aalto et al. (1994).

A stellar bar (length ∼ 3 or 6 kpc), which was discovered in the Hα line (Phillips 1993; Koribalski & Dettmar 1993), might be causing the gas to accumulate at the inner and outer Lindblad resonances. The inner Lindblad resonance as determined from the bar pattern speed coincides with the ring radius; the outer resonance lies just outside the pseudo-ring which is formed by the outer spiral arms.

NGC 3034 (M 82) is a typical and probably the most well-known starburst galaxy and has been extensively studied in nearly all wavelength ranges. Although M 82 is classified as a dwarf irregular galaxy, it is very similar to many of the starburst galaxies discussed here, in particular NGC 1808. High-resolution observations by Kronberg, Bierman & Schwab (1985) show that the continuum emission is concentrated into a large number of compact sources, identified as very young supernova remnants. A cylindrical outflow of dense molecular clouds and ionized gas has been observed, ejected from the disk into the halo at several hundred km s-1 (Nakai et al. 1987; Bland & Tully 1988). HI absorption lines shifting over the continuum source(s) clearly indicate a fast-rotating nuclear ring (Weliachew, Fomalont & Greisen 1984; Yun 1992). Telesco et al. (1991) report the detection of a bar ∼ 1 kpc long; the nuclear ring (radius ∼ 300 pc) lies roughly between the ILR(s) which are found at 40 and 600 pc.

NGC 3079 is a remarkable edge-on spiral galaxy. It is optically disturbed and dusty, has a Seyfert2/LINER spectrum, an active nucleus, and starburst activity (for a summary see Baan & Irwin 1995, and references therein). Two radio lobes extending approximately along the minor axis of the galaxy — very similar to those in Circinus (see below) — suggest the presence of outflow (e.g., Duric & Seaquist 1988; Hummel, van Gorkom & Kontanyi 1983). This is also indicated by the Hα images which reveal a giant loop (Ford et al. 1986) and, on deeper images, several filaments on kpc scales (Armus, Heckmann & Miley 1990). The nuclear activity may be fuelled by an inner molecular disk (Young, Claussen & Scoville 1988) which has a radius of ∼ 400 pc. Broad HI absorption lines detected by Irwin & Seaquist (1991) probably indicate a fast-rotating ring of similar size. High-resolution VLA HI and OH absorption measurements by Baan & Irwin (1995) reveal numerous components, one of which is a rapidly rotating compact disk, extended by about 100 pc or 125 (see also Gallimore et al. 1994). NGC 3079 also contains the most luminous known H2O megamaser.

NGC 3628 is a peculiar edge-on galaxy which belongs to the Leo Triplet, a nearby complex of three interacting spirals. HI observations by Rots (1978) and Haynes, Giovanelli & Roberts (1979) reveal emission in all three galaxies as well as a plume and bridge between NGC 3627 and NGC 3628. CO observations of the group have been published by Young, Tacconi & Scoville (1983). Neutral hydrogen absorption in NGC 3628 was first reported by Dickey (1982); he quotes a line width of 93 km s-1 (although there is a hint of faint absorption over a much larger velocity range) centred at 882 km s-1, close to the systemic velocity of the galaxy. VLA observations against the extended continuum of NGC 3628 by Schmelz, Baan & Haschick (1987) reveal about 10 individual HI absorption features over a velocity range of ∼ 350 km s-1. The velocity gradient of several strong components has been interpreted as a circular, solidly rotating disk. A very detailed study of both the HI emission and absorption in NGC 3628 has been carried out by Wilding, Alexander & Green (1993). Zhang, Wright & Alexander (1993) obtained high-resolution HI and CO measurements of NGC 3627. We obtained ATCA 20-cm radio continuum and HI snap-shot observations, but the angular resolution is rather low.

NGC 4945 is a nearby, southern, edge-on galaxy with an extremely large optical extent. Giant complexes of gas and dust in the disk give it a patchy appearance similar to NGC 253. The presence of a bar was indicated by de Vaucouleurs (1964), but further measurements are necessary to support this finding. The `active' nucleus exhibits both starburst and Seyfert characteristics (Whiteoak 1986). Nakai (1989) indicated the presence of optical filaments on one side of the galaxy emerging from the nuclear region toward the halo. The fan-like morphology of the outflow has a scale-height of 2 kpc and extends ∼ 10 kpc along the plane. A broad HI absorption line has been detected, but not resolved, by Whiteoak & Gardner (1976) and Ables et al. (1987). Similar broad absorption lines have also been found at other wavelengths (see Whiteoak & Wilson 1990, and references therein). Several authors find evidence for a fast-rotating molecular ring: in CO(1-0) by Bergman et al. (1992), in H2 (2.2 µm) by Koornneef (1993), and in CO(2-1) by Dahlem et al. (1993). The nuclear ring is quite thick and has a mean radius of about 100-200 pc and a rotational velocity of 200-250 km s-1. A detailed study of the HI emission and absorption in NGC 4945 has recently been carried out with the ATCA by Ott (1995).

Circinus is a nearby spiral galaxy only 4° below the Galactic plane. It was found while inspecting a Schmidt plate and shortly afterwards observed in HI with the Parkes 64-m telescope. Freeman et al. (1977) measured a half-width of at least 32' × 15' for the HI extent of Circinus, much larger than the optical diameter (∼ 10) of the galaxy. K. Jones et al. (1996, in prep.) and Koribalski & Whiteoak (1996) only recently mapped Circinus with several configurations of the ATCA; a few preliminary results are presented here. Fig. 7 shows part of the overall HI distribution, clearly outlining the two spiral arms. The elongated central disk (length ∼ 5) indicates a bar, and the central `hole' is caused by HI absorption against the bright central region of Circinus (see Fig. 8). The nuclear activity is caused by star formation, a Seyfert2 nucleus and giant radio lobes (perhaps associated with outflow), very similar to those in NGC 3079. Harnett et al. (1990) observed Circinus with Parkes at 1665 and 1667 MHz and found strong OH absorption lines over a velocity range of nearly 200 km s-1. High-resolution ATCA measurements reveal HI absorption over about the same velocity range. The data show a shift of the line over the face of the continuum, indicating a fast-rotating nuclear ring (Koribalski & Whiteoak 1996). This is supported by Hα images taken with AAT Taurus Fabry-Perot interferometer (Fig. 9) and with the ESO New Technology Telescope (Marconi et al. 1994) showing a broken ring or spiral arms at ∼ 200 pc from the nucleus. Marconi et al. also found an OIII emission cone toward the NW indicating highly ionized gas flowing out from the nuclear region.

Figure 7

Figure 7. HI distribution of the Circinus galaxy (courtesy: K. Jones, B. Koribalski, M. Elmouttie, R. Haynes). Contour levels are 0.042, 0.063, 0.084, 0.105, 0.126, 0.169, 0.211 Jy beam-1 km s-1. This is a very preliminary map was taken with the 750-m array of the ATCA (12 h). The angular resolution is ∼ 40. Please note that the HI emission is extended over a much larger field than displayed.

Figure 8

Figure 8. HI absorption spectrum toward the nuclear region of the Circinus galaxy. These data were taken with the largest array of the ATCA (Koribalski & Whiteoak 1996).

Figure 9

Figure 9. Hα emission from the nuclear region of the Circinus galaxy. This is a very preliminary image taken with the TAURUS-2 Fabry-Perot instrument at the Anglo-Australian Telescope (courtesy: B. Koribalski, K. Taylor, J. Whiteoak). The angular resolution is ∼ 1.5".

NGC 6221 is a barred spiral galaxy which is interacting with its brightest neighbour, NGC 6215, and possibly also with two newly discovered low-surface brightness galaxies (see Fig. 10). The galaxy group lies only ∼ 10° below the Galactic plane which explains the rather crowded field in the optical. Recent measurements with the ATCA 375-m array revealed an HI bridge between the two major galaxies and several extensions of the gas envelope of NGC 6221 which can be attributed to tidal forces between these galaxies (B. Koribalski et al., in prep.). Detailed studies of the Hα line emission in NGC 6221 by Pence & Blackman (1984) revealed very large non-circular motions of the ionized gas, possibly as a result of tidal interactions and streaming motions along the bar. The HI distribution of NGC 6221 (Fig. 11) is quite similar to the optical emission; the central depression is caused by HI absorption which is observed over several hundred km s-1. A nuclear ring is the most likely explanation regarding the similarity of this galaxy to others discussed here, but further data analysis is needed to confirm this idea. The HI velocity field looks much more regular than the Hα field which was reconstructed from several long-slit spectra. The nuclear activity is characterised by star formation and a weak Seyfert2 nucleus.

Figure 10

Figure 10. HI distribution (contours) of NGC 6221 and neighbouring galaxies overlaid onto the optical emission (greyscale) from the Digitized Sky Survey (DSS). The contour levels are 0.15, 0.3, 0.6, 1.2, 2.25, 4.5, and 9 Jybeam-1 km s-1. Please note that no primary beam correction has been applied here. A tidal HI bridge is visible between NGC 6221 and its brightest neighbour, NGC 6215 (vsys = 1555 km s-1). The two low-surface brightness galaxies (BK_1: vsys ∼ 1645 km s-1 and BK_2: vsys ∼ 1510 km s-1) are newly discovered members of this interacting group of galaxies. These data have been obtained with the 375-m array of the ATCA (12 h). The angular resolution is about 1.5"

Not much is known about NGC 6215, which is classified as a non-barred spiral galaxy of type SAS5 (RC3). Its HI emission extends from ∼ 1450 to 1650 km s-1 (Fig. 10).

Figure 11

Figure 11. Left: HI distribution of NGC 6221. The contour levels are 0.15, 0.3, 0.6, 0.9, 1.2, and 1.5 Jy beam-1 km s-1. Right: Mean HI velocity field of NGC 6221. The contour levels range from 1315 to 1615 km s-1, step 15 km s-1. These data have been obtained with the 1.5-km array of the ATCA (12 h). The resolution is 30" × 25" × 6.6 km s-1.

NGC 7552 is a prominent starburst galaxy; no active nucleus has been reported. Because of its nearly face-on inclination the strong bar is easy to identify. Although it has no close neighbours, the group of galaxies associated with NGC 7582 (see below) at a distance of ∼ 30' plays an important role. Forbes et al. (1994) found a 1-kpc starburst ring in NGC 7552 consisting of supernova remnants and optical `hot spots'. We observe HI absorption against the ring and find a velocity shift of ∼ 120 km s-1 betweenits eastern and western parts.

NGC 7582 is a barred spiral which is interacting with several bright neighbours, namely NGC 7590, NGC 7599 (see Fig. 12), and most likely also NGC 7552 (see above). Recent ATCA measurements reveal several HI bridges between the group members (B. Koribalski et al., in prep.). NGC 7582 has a Seyfert2 nucleus, but it is clearly dominated by starburst activity as shown by strong radio emission, which probably originates from many supernova remnants, and by the presence of numerous HII regions (Morris et al. 1985). TAURUS Fabry-Perot observations of the circum-nuclear Hα emission suggest a fast-rotating disk of HII regions (diameter ∼ 1 kpc) in the plane of the galaxy. Morris et al. also suggest that blue-shifted OIII line emission originates from high excitation gas expanding outwards from the nucleus in a cone. The HI distribution of NGC 7582 (see Fig. 13) is rather asymmetric (because of tidal interaction) and extends over a velocity range of at least 400 km s-1. The absorption lines cover nearly the same range, but are ∼ 50 km s-1 blue-shifted with respect to the emission. We find the HI absorption line clearly shifting over the face of the continuum emission indicating a fast-rotating ring. Circumnuclear Hα emission, suggesting a similar feature (see above), hints at a mixture of gas phases in the nuclear region.

Figure 12

Figure 12. HI distribution (contours) of NGC 7582 and two nearby galaxies, NGC 7590 (vsys = 1596 km s-1) and NGC 7599 (vsys = 1654 km s-1) overlaid onto the optical emission (greyscale) from the Digitized Sky Survey (DSS). The contour levels are 0.15, 0.3, 0.6, 1.2, 2.25, and 4.5 Jy beam-1 km s-1. Please note that no primary beam correction has been applied here. Several tidal tails are visible near NGC 7582, pointing toward the neighbours in the east and toward NGC 7552 which lies at a projected distance of ∼ 30'. These data have been obtained with the 375-m array of the ATCA (12 h). The angular resolution is about 1.7' × 1.3'.

Figure 13

Figure 13. HI distribution and mean velocity field of NGC 7582. These data were taken with the 1.5-km array of the ATCA. The angular resolution is 38" × 23".

Recent high-resolution 8.64-GHz observations with the ATCA by Norris & Forbes (1996) suggest a double nucleus.

4.2. Strongly Interacting Galaxies

Table 4 lists a few strongly interacting or merging galaxies which also show broad HI absorption lines. These galaxies happen to be some of the nearest and more famous examples of very luminous FIR-galaxies with far-infrared luminosities of more than 1012L (see also Heckman et al. 1990). They are the sites of exceedingly powerful bursts of star formation. The RC3-Type is added to show how peculiar these systems are. Several of these galaxies have also been detected in hydroxyl absorption (see e.g., Schmelz et al. 1986).

I could easily add a few more galaxies to this group (see e.g., Dickey 1982, 1986; Mirabel 1982; Heckman et al. 1983). Mirabel & Sanders (1988) find that 18 out of their 92 luminous far-infrared galaxies show HI absorption with line widths of 100-600 km s-1, although some cases are very weak (e.g., U 9618, Mrk 331), and many of the detections were known before (see the references in Table 4). Interferometric studies need to be carried out to separate the HI absorption from the emission to determine more accurate line widths and centre velocities. For example, the study fails to pick up the broad absorption in the nearby polar-ring galaxy NGC 660, and there may well be others.

Table 4. Strongly interacting pairs of spirals

Name D RC3-Type abs. line width References
[Mpc] [km/s]

NGC 520 30 .P..... 100 1,5,7
Mkn 231 160 .SAT5$P 150 1
NGC 2623 74 .P..... 525 2,5,7
NGC 3256 37 .P..... 300 10
NGC 3690 / IC 694 42 .IB.9P. 300 1,5,8
IC 4553 (=Arp 220) 72 .S?.... 743 2,(4),5,6,7
IC 883 92 .I..9*P 460 3,7
NGC 6240 100 .I.0.*P 700 1,3,(9)

References: (1) Dickey 1982; (2) Mirabel 1982; (3) Heckman et al. 1983; (4) Norris et al. 1985; (5) Dickey 1986; (6) Baan et al. 1987; (7) Mirabel & Sanders 1988; (8) Baan & Haschick 1990; (9) van der Werf et al. 1993; (10) English 1994.

What is happening in these galaxies? Are they scaled-up versions of the group of nearby luminous galaxies discussed in the previous section? I think the answer is yes. But in these galaxies the cold gas which feeds the starburst is driven inwards by strong tidal interaction or merging between galaxies (see e.g., Rieke et al. 1985), whereas in the nearby galaxies, bars and/or weak galaxy interactions may be responsible.

Are those broad absorption lines also caused by fast-rotating nuclear rings? Interferometer maps of the integrated CO emission have shown that cold gas is strongly concentrated toward the `active' nuclei (starburst nuclei or AGN) of those systems (Sargent & Scoville 1991; Scoville et al. 1991). The kinematics, which are more clearly revealed by high-resolution HI absorption measurements, indicate that at least part of the gas rotates around the nuclei (see e.g., Baan & Haschick 1990). A very nice example is the peculiar galaxy IC 4553, where the HI and OH absorption measurements clearly indicate a fast-rotating edge-on disk close to the nucleus; the total velocity shift across the source is 145 km s-1 over a distance of 1.9" or ∼ 700 pc (Baan et al. 1987, and references therein). But there are also inward and outward motions of the cold gas. The central activity of IC 4553, which is also the best studied OH megamaser galaxy, is caused by a Seyfert2 or starburst nucleus (see e.g., Heckman et al. 1983).

The gas distribution and kinematics in interacting systems could provide information about the merger evolution. As the interaction progresses, a large fraction of the interstellar medium (ISM) of each galaxy is expected to sink into a common centre of mass. I suggest that the cold gas is accumulated in a ring near the central source from where it is spiralling inwards, eaten up by the hungry starburst, and then partly ejected into the halo.

4.3. Other Galaxies with HI Absorption Lines

Seyfert Galaxies. NGC 4151 is a weakly-barred spiral galaxy (type PSXT2). It is also one of the best-studied examples of a nearby active galactic nucleus (AGN), and one of the original nuclei noted by Seyfert (1943). The Seyfert type is possibly intermediate between 1 and 2. Pedlar et al. (1992) investigated the large-scale HI structure of the galaxy which covers a velocity range of about 920-1080 km s-1 in emission and absorption; the galaxy is viewed nearly face-on (i ∼ 20°). They find that the most striking feature of the Hsc i emission structure is a `fat' bar, or oval, which occupies the central 5 kpc. The distortions of the velocity field are most likely caused by the non-circular gas flows in the bar. No evidence for tidal interactions with the companions at a distance of ≳ 20' has been found. Pedlar et al. suggest that the massive, gas-rich bar of NGC 4151 could have caused the inflow of gas to the centre (similar to the model I discuss for the nearby starburst galaxies in Section 4.1), and that the neutral hydrogen gas present within the few hundred parsec of the nucleus (as shown by the HI absorption measurements) constitutes a fuelling reservoir for the AGN.

Using the VLA B-array HI absorption toward the nuclear region of NGC 4151 was detected by Dickey (1986), with a centre velocity of 997 km s-1 and width of 52 ± 15 km s-1. To resolve the nuclear region Mundell et al. (1995) obtained Merlin HI absorption measurements with an angular resolution of 0.15". They found absorption (FWHM ∼ 90 km s-1, centred at 993 ± 6 km s-1) against the nucleus only. Several other, slightly weaker continuum sources which are part of a jet pointing toward us show no absorption. The authors therefore conclude that the neutral hydrogen disk in which the nucleus is embedded can be no thicker than 50 pc.

The HI survey of Seyfert galaxies by Heckman, Balick & Sullivan (1978) shows how the low angular resolution of even the biggest single-dish telescopes makes it nearly impossible to detect HI absorption against the nuclear continuum sources of galaxies. Out of 58 Seyfert and Seyfert-like galaxies, 25 were detected in HI emission and only three in HI absorption.

The peculiar galaxy NGC 4258: A black hole candidate. NGC 4258 is a nearby (D = 6.4 Mpc) active spiral galaxy with high luminosity H2O maser emission. VLBI observations by Greenhill et al. (1995) and Miyoshi et al. (1995) find the masers residing in a very fast-rotating (vsys ± 900 km s-1), nearly edge-on toroid/disk of radius 0.2 pc. This provides the most compelling evidence so far for a black hole in the centre of a galaxy. — The masers are probably formed in relatively dense gas clouds that are forming new stars.

HI observations of NGC 4258 (van Albada 1978, 1980) reveal a rather flat rotation curve (vrot ∼ 200 km s-1) in the outer parts of the galaxy, but a steep rise in the nuclear region. The latter prompts van Albada (1978) to speculate ``that the rotational velocities near the nucleus may be significantly higher than measured, implying a larger central mass concentration.'' Streaming motions in the HI velocity field are attributed to the weak bar in the disk of NGC 4258 (type SXS4, see Table 2).

NGC 4258 is also well known for its anomalous arms, visible only in Hα and radio continuum, which are usually interpreted in terms of collimated nuclear outflow or jets (Cecil, Wilson & Tully 1992; Dettmar & Koribalski 1990, and references therein).

As NGC 4258, the galaxies NGC 3079, NGC 1068, NGC 4945, and Circinus each host an H20 megamaser (apparent isotropic luminosity > 10 L). A position-velocity diagram of the brightest H2O masers in the central parsec of NGC 1068 (Gallimore et al. 1996) shows a similar, but much less extreme phenomenon than NGC 4258. Other megamaser galaxies are currently investigated. Weaker H20 masers are also known in NGC 253, M 82 and a few other galaxies (see Greenhill et al. 1990, and references therein).

Elliptical Galaxies. The nearest and most prominent radio elliptical galaxy is NGC 5128 (Cen A). HI, OH, and H2CO absorption features have been detected near the systemic velocity of the galaxy (550 km s-1). The HI absorption line extends from about 500 to 700 km s-1, whereas the HI emission is much broader, extending rather symmetrically over ∼ 500 km s-1 (Gardner & Whiteoak 1976, and references therein). High-resolution VLA observations by van der Hulst, Golisch & Haschick (1983) show three rather narrow absorption features, one at the systemic velocity of CenA and the other two at +26 and +46 km s-1. The red-shifted components possibly arise from clouds falling into the nucleus where they fuel the central engine; this has also been observed in several other radio elliptical galaxies (see van Gorkom et al. 1990).

Several narrow absorption features, both at the systemic velocity and at red-shifted velocities, have been observed against the nucleus of the early-type galaxy NGC 1052 (van Gorkom et al. 1986).

Other radio elliptical galaxies detected in HI absorption are NGC 315, NGC 1275, NGC 3894, NGC 5363, 3C 236, and 4C 31.06, bringing the number of detections up to eight (van Gorkom et al. 1989, and references therein).

Quasars. The first detection of HI in a quasar spectrum was made by Brown & Roberts (1973), who detected a narrow absorption line against 3C 286 at a red-shift of z = 0.69. HI absorption has also been detected against the QSO PKS 1157+014 (Wolfe, Briggs & Jauncey 1981) and the BL Lac object AO 0235+164 (Wolfe, Davis & Briggs 1982). For a summary see e.g., Giovanelli & Haynes (1988). Another very interesting study is that of HI absorption lines against strong radio continuum sources, many of them quasars, in the vicinity of spiral galaxies; see e.g., Corbelli & Schneider (1990) and Dickey, Brinks & Puche (1992).

Others. VLBI observations of HI absorption (3420-3590 km s-1) against the active nucleus (0.016" or 5 pc) of the highly luminous spiral galaxy NGC 5793 (vsys = 3521 km s-1; D = 70 Mpc) reveal several components (Gardner et al. 1992). The authors claim that these are probably caused by various clouds moving outwards, and not by a single rotating cloud complex surrounding the nucleus. But a higher sensitivity study and larger bandwidth are definitely needed to solve the issue.

Several strong absorption lines in the radio galaxy 3C 293 (z = 0.045) have been detected by Haschick & Baan (1985). They can resolve the broad (Δv = 480 km s-1) HI absorption spectrum into more than 10 components, which can be attributed to two different structures: —a) the strongest features are part of a rapidly rotating disk surrounding the nucleus of the galaxy and —b) several high-velocity features which are either red- or blue-shifted are interpreted as clouds falling into and being expelled from the nucleus of 3C 293. The observed velocity gradient across the whole 22 continuum structure is 179 km s-1.

Recent VLA observations of Cygnus A, the nearest powerful FRII galaxy (z = 0.0565), by Conway & Blanco (1995) also revealed a broad (Δv = 270 km s-1) absorption line against its 15-pc nucleus. It most likely consists of two components arising from a rotating nuclear torus.

Another radio galaxy, NGC 4261 (3C270), shows a rather narrow (Δv = 65 km s-1) HI absorption feature against its nucleus, near the systemic velocity at ∼ 2200 km s-1 (Jaffe & McNamara 1994).

Next Contents Previous