Astron. Astrophys. 358, 242-256 (2000)

3. Results

3.1. Large scale morphology

The 1.1 mm map, presented in Fig. 1, gives a good overview of the region. The northern-most part of the map has a much higher noise level, but we can see that the emission still continues to the north outside the area mapped by us. There is no dust emission in the western part of the map, except for a faint "dust-cloudlet" 240 " west of NGC 6334 I, seen in both the 1.3 mm and the 1.1 mm maps. Some additional faint emission blobs in the same area are likely to be spurious, i.e. artifacts from the DBMEM reduction. The map shows that the dust emission is concentrated to two main emission centers, I and I(N), and that the dust emission is sharply bounded to the west by a long narrow linear filament. Additional ridges connect this filament to I and I(N). These ridges and lumps give the sub-mm images the appearance of a cob web. The crossing points are regions of intense sub-mm emission, which appear to contain embedded high to intermediate mass proto-stars or clusters of protostars. These regions, especially in the south, are connected by lumpy ridges of dust emission. In addition to the two strong sub-mm sources NGC 6334 I and I(N), we identify eight additional sources. These are labelled with the prefix SM in Fig. 1.

3.2. The linear filament

The narrow straight filament (width 15 "-20 ") that bounds the dust emission to the west is especially striking. This filament has the appearance of an ionization front, e.g. similar to the Orion bright bar, yet there is nothing visible in optical plates of the region, nor in the near-IR surveys. The high resolution "True colour" JHK image by Tapia et al. (1996) shows that the extinction falls off rapidly to the west, approximately where we see the filament, but the near-IR images do not show a sharp boundary between the low and high opacity region. The filament is narrow and lumpy, and appears to continue to the south, outside the area covered by our maps. The 800 µm-map (Fig. 2), which has higher dynamic range in the north, shows that it still continues straight into the dust cloud even though here the dust emission also curves to the north. The filament bends off to the east, and then makes a sharp turn back approximately at the position of SM 8 (see Fig. 1 and Table 1). The northermost part of the filament is again perfectly aligned. This filament appears to coincide with the boundary of high visual extinction, as mapped by Straw & Hyland (1989) from their near-IR survey of the NGC 6334 complex. Their high extinction region also turns towards north east as well as westward at a declination of -35^o 40´, and continues to the south-west as far as NGC 6334 V. The total length of the filament, as seen in our 800 µm-map is 7´ (3.5 pc), but it is possible that it may follow the high extinction boundary all the way to NGC 6334 V. Nevertheless, it covers a large distance on the sky and is remarkably straight. There is no sign of bending, or curvature, apart from the apparent lumpy structure of the filament.

3.3. Cores, ridges and holes

NGC 6334 I is by far the strongest source at all wavelengths, and barely resolved in our maps. It is surrounded by relatively little extended emission, although there is a ridge of emission extending towards the more evolved UC HII region E, and continuing past it to connect with the linear filament. There is also another dust bridge that connects I to the extended cloud core surrounding I(N). In the sub-mm I appears as a single source, but high resolution molecular line and maser observations shows that it harbours at least two young stars. Our maps of the I(N) region resolve at least two different sites of star formation, both of which may harbour multiple sources.

The I(N) region continues to the north through another bridge of emission, which is fainter than the one to the south. This bridge merges in the north into a more extended region bounded to the west by the linear ridge. The I(N) region connects through another dust bridge to the west, which terminates more directly on the linear filament. Further north the emission gets fainter, and the dust peaks appear less point-like and fainter. These may be regions, which are both colder and less massive, and which have not yet formed any stars.

There seems to be several empty regions void of gas and dust, bounded by I, I(N) and the linear filament. These regions void of dust stand out even clearer on the 450 and 350 µm-maps. Overall, there is very little dust emission east and northeast of NGC 6334 I, until we get up to I(N) which shows a spur of emission extending almost straight east. The region south of I is also devoid of emission. The location of the more evolved HII region E, powered by a cluster of B-type stars (Tapia et al. 1996), on a ridge of emission with areas void of gas both to the north-east and the south-west would suggest that the area cloud have been cleared by an energetic wind from E, which now is almost void of dust. However, the morphology of the voids do not lend much support to such an idea. This looks more like a pre-existing structure, which was present even before E was formed.

3.4. Compact sources

We can identify more than 10 cores in the 1.3 mm - 800 µm maps, which appear to be massive compact cloud cores. Table 1 lists average source sizes and positions for 10 of these cores ordered from south to north, which have been identified on at least two different maps. The positions are all relative to our adopted position of NGC 6334 I (see Sect. 2.2). The relative positions should be accurate to 2 " except for the northern part of the cloud, which was not covered by our 450 and 350 µm-maps. The S/N and positional accuracy of these sources are much poorer than for the rest of the cloud. Table 2 gives flux densities derived from Gaussian fits for the sources that we could fit with a Gaussian. For the fainter sources the flux estimates are rather uncertain, because the fits at long wavelengths tend to give rather large source sizes. In order to ensure that we sample the same region, we have crudely corrected these fluxes so that they refer to the same source size.

Table 2. Integrated fluxes of compact sources and the I(N) cloud core.

© European Southern Observatory (ESO) 2000

Online publication: June 26, 2000
helpdesk.link@springer.de