Neutral Hydrogen in Compact Groups

Observations of the neutral hydrogen in compact groups can be very useful for studying the nature of these systems in more detail. Rood and Williams (1987) concluded from single dish observations that many compact groups appear to be poor in neutral hydrogen compared to loose groups of galaxies. Although the mechanism responsible for the HI deficiency is not well understood, because these compact groups have different HI properties compared to loose groups, this could indicate that these groups are genuine compact systems and not projected loose groups. High resolution images of the HI provide a very powerful tool for investigating the dynamical state of compact groups since HI is most sensitive to tidal disturbances. However, such data are available only for very few compact groups (e.g. Williams et al. 1991). There appears to be a trend that in less dense compact groups, although there are signs of tidal interactions, the HI is still associated with the individual galaxies, while in very compact systems the HI is spread out over the whole group (Williams et al. 1991). However, HI images of more compact groups are needed in order to be able to make a firmer statement about this. High resolution HI observations will also help in understanding the reasons for the HI deficiency observed.

Recently, a new sample of compact groups was constructed through a purely automated search using digitised galaxy catalogues (Iovino et al. in prep.). The algorithm used for the extraction incorporates a slightly modified version of the original Hickson criteria, in order to make it more efficient in rejecting chance projections, a problem that gets very serious at fainter magnitudes. The new algorithm has been tested with Monte Carlo simulations and has proved very fast and reliable, while it avoids some of the biases that are present in the Hickson sample (Prandoni et al. 1994; Iovino et al. in prep.). To further study the properties of compact groups and to address some of the questions raised above, we have started a program to observe compact groups from this new catalogue. Here we report on HI observations of a first sub-set of this new sample.

Parkes Observations

To start our observational program, we have observed a first sub-sample of 14 compact groups with the Parkes radio telescope. The aims of these observations were to study the HI content of these groups and to select candidates for observing with the Australia Telescope Compact Array (ATCA). To study the HI content, we do a similar analysis as done by Rood and Williams (1987). In Figure 1 we show the histogram of the HI deficiency, x, for compact groups from the Hickson sample and for the compact groups observed by us. The HI deficiency is here defined as , where is the observed neutral hydrogen mass of a group and is the predicted neutral hydrogen mass for the same group. The predicted HI mass of a group is the sum of the predictions for the galaxies in the group. For spiral galaxies we use the diameter to predict the HI (Roberts & Haynes 1994), while for early type galaxies we assume that they have no HI.

For the Hickson groups we have used the HI data from Rood and Williams (1987), but we have redone their analysis since better optical images are now available for morphological classification (Hickson  et al. 1989), as well as complete redshift information for the galaxies in Hickson groups (Hickson  et al. 1992), although the overall conclusion about the HI content of Hickson groups remains the same. For comparison, we also give the histogram of x for a small sample of field spirals (data taken from Broeils 1992). A Kolmogorov-Smirnov (KS) test (using the upper limits as detections) gives that the distribution of x for the Hickson groups differs from that of the field sample at a level .

 
Figure 1: top: Histogram of the deficiency of neutral hydrogen, x, in Hickson groups. The shaded area indicates the fraction of upper limits to x. middle: as above, but for compact groups observed at Parkes. bottom: as above, but for a control sample of field spirals

Figure 1 shows that the distribution of x has a tail towards negative values, both for the Hickson sample (as expected) and for the 14 groups observed at Parkes. Although a KS test gives that the distribution of x for our sample of compact groups differs from that of the field sample only at the 15% level, also in our new sample there are a few compact groups that are deficient in HI. It can also be seen from Figure 1 that quite a large fraction of compact groups have normal HI content. It is tempting to identify the HI-poor compact groups as the true compact systems. If we do this, this would mean that about a third of the groups studied are true compact systems. But, as we will see below, the situation is somewhat more complicated.

ATCA Observations

>From the 14 compact groups observed with the Parkes telescope, we have observed 7 with the ATCA. In Figure 2 we show the total HI maps of two of these groups. The compact group on the left in Figure 2 is the group that is most deficient in HI as found in the Parkes observations. HI is only detected in one of the galaxies and the morphology of the HI suggest a strong tidal interaction. Interestingly, a deep optical image of this group shows that the galaxies in this group have a common envelope (D. Malin priv. comm.). This suggests that this group is a real compact system and that it is in a quite advanced stage of evolution. Observations with SEST (Oosterloo & van Moorsel in prep.) reveal that CO is not only present in the galaxy we also detect in HI, but also in the southern galaxy in the group. So this southern galaxy appears to be a galaxy that originally was gas rich, but now has lost most of its neutral hydrogen. This is the first example where we can study the HI deficiency in compact groups in much more detail and these observations will help in understanding the mechanisms responsible for the HI deficiency observed in some compact groups. These observations also lend support to the idea the HI deficiency observed in some compact groups could mean that these groups, like the group in Figure 2, are indeed genuine compact systems.

The group on the right in Figure 2 is a compact group that is rich in HI. The interesting feature here is that a significant fraction of the HI is not associated with the galaxies, but instead appears to form a common envelope. This indicates that also this group is very likely a physically compact system. This is also suggested by the morphology of the galaxies, as well as by the fact that one of the galaxies has a relatively strong continuum source in the centre. The observations of this group show that it is not correct to identify only the compact groups that are HI poor with the true compact systems, but that a genuine compact group can also be rich in HI.

In the remaining 5 compact groups observed with ATCA, the morphology of the HI suggests interactions occur in 2 more groups, while in 3 groups the HI does not indicate that the galaxies in these groups are particularly disturbed by interactions.

 
Figure: Total HI map (contours) superposed on an optical image for two of the compact groups observed with ATCA. The group shown on the left is also part of the Hickson sample (HCG 90). The depression in the HI of the largest galaxy in the panel on the right is due to absorption against the central continuum source of this galaxy

© Copyright Astronomical Society of Australia 1997