3. The globules and the HII region
IC 2944 and the adjacent IC 2948 comprise a bright and very extensive HII region (Fig. 1). Within its boundaries there are numerous OB stars located, the brightest being HD 101205, a sixth magnitude O7 III star (Walborn 1973). Thackeray & Wesselink (1965) studied these stars and suggested a distance of 2 kpc. However, Ardeberg & Maurice (1977, 1980, 1981) suggested that the concentration of OB stars is a line-of-sight effect, in fact consisting of 5 different sub-groups at different distances. This interpretation has been supported by Perry & Landolt (1986). More recently, Walborn (1987) has argued that at least the O-type stars in the region constitute a significant physical cluster, which is responsible for the IC 2944 HII region. The present paper cannot resolve these differing views, but we note that the HII region is likely to be associated with the most luminous and the earliest of the O-stars, HD 101205, and that the globules certainly are associated with the HII region. As discussed by Ardeberg & Maurice (1980), the nebular kinematic distance is about 1800 pc, a distance we adopt in the following for the globules.
The large-scale plate in Fig. 1 demonstrates that the globules are very isolated in the HII region, situated away from the receding walls of neutral material, where the HII region is ionization-bounded. This is a situation different from the Rosette Nebula, where the small globules are closely associated with the dense shell surrounding the HII region.
Fig. 2a shows the region of Thackeray's globules on our CCD images obtained through an H filter. The major ones are identified in Fig. 2b, and the positions of the principal ones are listed in Table 1. The angular scale is related to a physical length assuming a distance of 1.8 kpc. Within the very localized region of the globules, about 50 globules, or fragments or splinters of globules are visible. At an assumed distance of 1.8 kpc, one arcsecond corresponds to about 0.01 pc. The projected separation of the eastern- and western-most globules is 7.5 arcminutes, corresponding to 3.9 pc.
There are three basic features to be noted in the deep H CCD image. First, the largest globule is clearly surrounded by bright rims. Second, the edges of all globules are generally extremely sharp and highly structured. Third, well-defined fragments of globules are visible in all sizes down to the resolution limit.
A large number of theoretical studies of isolated dense globules immersed in HII regions around O stars have been performed, among the pioneering studies are Kahn (1969), Dyson (1973) and Tenorio-Tagle (1977). They found that at the interface between neutral and ionized material an ionization-shock front is set up, from which ionized gas streams away supersonically. Because of recombination in the downflow region the neutral globule is largely shielded from the incident uv radiation, and the front moves only slowly into the globule, which can then survive for a considerable time.
If a globule is situated towards the edge of an HII region, it takes a cometary shape, with a rather smooth front facing the ionizing source(s), and a tail streaming in the opposite direction (e.g. Reipurth 1983). But such a morphology is not seen for Thackeray's globules.
What is the origin of these globules? In principle, there are two possibilities. Firstly, the globules could be cloud cores originally located inside a larger more tenuous cloud, which has by now been evaporated, thus exposing its skeleton of dense structures at their original locations. The dimensions of the largest globule, 0.3 x 0.5 pc, is typical of dense cloud cores (Myers et al. 1991). The remarkably sharp edges displayed by most of the globules is unlikely to have existed prior to exposure, and would in this picture be a result of subsequent erosion by ionization-shock fronts driven into the cores. The large number of very small globules down to the resolution limit (2 pxl = 0.007 pc) may not have pre-existed inside the larger cloud, but could be products of a fragmentation process or an instability in the ionization-shock front.
Another, and as we shall argue, more likely possibility is that the globules are remnants of an elephant trunk, similar to the well known "Pillars" in M16 (e.g. Hester et al. 1996), which formed as the HII region interacted with the walls of the neutral material out of which the O-star cluster was born. This scenario finds support in our CO data.
© European Southern Observatory (ESO) 1997
Online publication: April 6, 1998