The molecular complex near NGC 3077 discussed here is larger than many complexes in other galaxies; e.g., the complex within NGC 3077 itself has a size (FWHM) of only 320 pc (Becker et al. 1989). The complexes near NGC 3077 are well confined over a narrow range in velocity. Cohen et al. (1988) report a full extent of the 30 Dor complex in the LMC of 2400 pc. Most Galactic molecular clouds have sizes below 60 pc (Solomon et al. 1987). The Orion A & B complexes (see Dame et al. 1987) placed at the distance of NGC 3077 would be detectable with the sensitivity of our observations, however just in one single spectrum. The complex near NGC 3077 thus belongs to one of the largest concentrations of molecular gas in the local universe.
We will now use the ratios of the integrated line intensities to get a handle on the physical parameters. The most stringent limitations on the kinetic temperature and H2 volume density can be derived for complex #1 for which we have observed 3 rotational transitions. In the following we apply a simple large-velocity gradient code (LVG) which adopts constant kinetic temperature and H2 volume density throughout the cloud (see e.g. de Jong et al. 1975for details). We calculate the line temperatures for monotonous velocity gradients of pc-1 and = pc-1, and for CO abundances of [CO/H2]= and [CO/H2]. The higher CO abundance is a representative value for Milky Way clouds (e.g. Blake et al. 1987), the other was chosen to study the effects of lower metallicity.
For complex #1 the observed line ratios can only be matched if the kinetic temperature is about 10 K. The H2 volume density should be between 600 cm-3, in the case of high abundance and low velocity gradient, and about 10000 cm-3, in the case of low abundance and high velocity gradient. The kinetic temperature is similarly low as that for cold dark clouds in the Milky Way with no embedded massive stars, but significantly lower than that of clouds with on-going massive star-formation (e.g. Jijina et al. 1999). Note that this is a large-scale ( kpc) average value for the complex which does not exclude localized small regions which may be warmer and thus may be heated by low-level star-formation.
For complex #2 we observed only two transitions. The observed line ratio can be explained within a wide range of parameters. We can however argue that because the ratio for the outer area of that cloud is similar to that of complex #1 as a whole that also the physical conditions may be similar, i.e. low kinetic temperature and medium to high volume density. In the center of complex #2 the ratio is much higher. Part of this might be due to the different beam areas for the two transitions, 22" for the (10) transition and 11" for the (2) transition. This difference can affect the ratio by a factor of up to 4, if the observed cloud is a point source for both telescope beams.
We argue, however, that this effect only partly can account for the variation within complex #2, because it is extended in both transitions (as indicated by the presence of (2) emission in the averaged spectra surrounding the centre); in addition the line width is too wide in the (2) transition to be due to one single small molecular cloud. It is therefore more likely that indeed the kinetic temperature and/or the H2 volume density rise in complex #2. This could be attributed to suggested on-going low-level star formation in the `Garland' region near complex #2 (Karachentsev et al. 1985b).
One important yet difficult to determine physical parameter is the mass of a molecular complex. Determination of the mass is usually based on either the assumption of virialization and/or application of a conversion factor. In this section we apply both methods and discuss their pros and cons.
To estimate the factor we compare the CO luminosity for a given line width with that of a cloud with the same line width but with known molecular mass (e.g. Cohen et al. 1988). Using Fig. 2 of Cohen et al. we find that complex #1 lies within the range of values spanned by Galactic clouds, whereas complex #2 lies in the range spanned by LMC clouds. This indicates that the tidal arm clouds have CO luminosities for a given line width in between those of Milky Way and LMC clouds. The factor of the Milky Way ( cm-2 (K )-1) is a factor of 6.7 lower than that of the LMC ( cm-2 (K )-1, Cohen et al. 1988). Thus we use an value for the tidal arm clouds which is in between both values, cm-2 (K )-1. This leads to molecular masses for complex #1 of and for complex #2 of , (corrected for the contribution of He).
If we adopt a density profile through the clouds the assumption of virialization leads to masses for complex #1 of and for complex #2 of . Given the uncertainties in both methods the masses for complex #1 agree well, however those for complex #2 are discrepant by a factor of 4. At this point we can only speculate which method gives the better estimate for the true molecular masses.
Fig. 6 shows a comparison of the distribution of the CO gas with that of atomic hydrogen. We use our high-angular resolution (13") 21 cm map as obtained with the VLA (Walter & Heithausen 1999). It is obvious, that there is no direct correlation between the intensities of HI and CO. CO is not concentrated at the peak of the HI column density, but rather found on the outer area of the cm-2 contour. The average HI column density associated with complex #1 and 2 is cm-2, that for complex #3 is cm-2. At position #4 (Table 1), where no CO was found, the associated HI column density is cm-2.
The threshold where to find molecular gas depends on both metallicity and radiation field (e.g. Pak et al. 1998). The shielding HI column density in the tidal arms around NGC 3077 is slightly higher than values in other galaxies. In the Milky Way the H2 threshold appears at cm-2 (Savage et al. 1977). CO observations of M 31 suggest a threshold of about 1021 cm-2 (Lada et al. 1988). Young & Lo (1997) found a threshold of cm-2 for the dwarf elliptical galaxies NGC 185 and NGC 205; they attribute the difference to the Milky Way value and M 31 to the lower radiation field in these dwarf ellipticals. We only can speculate what causes the higher threshold in our complex. The radiation field is probably low because we do not see strong associated star forming regions.
The total HI mass of the tidal arm feature around NGC 3077 was found to be M(HI )= (van der Hulst 1979; Walter & Heithausen 1999), depending on integration boundaries. If we regard only the areas where CO is detected we find an HI mass for complex #1 of about and for complex #2 of about . These atomic masses are comparable to the molecular masses derived by our adopted factor (see Sect. 4.2).
© European Southern Observatory (ESO) 2000
Online publication: October 2, 2000