Fragmentation

Since its first determination in 1955 [58], the rather universal validity of the Initial Mass Function (IMF) has been confirmed in many ways. For a recent compilation see [26]. One of the essential questions in general star formation research is how and at what evolutionary stage the shape of the IMF gets established. Turbulent fragmentation theories predict that the turbulence right at the onset of fragmentation processes already produces core mass functions with the same power law like the IMF and that by applying star formation efficiencies these core mass functions directly convert into the later observed IMF [47, 51]. Contrary to that, other groups argue that the the early star-

forming gas clumps fragment down to many cores of more or less the Jeans mass (~O.5M0) and that the IMF will be formed from that stage via competitive accretion from previously unbound gas [17, 19].

In the low-mass regime, dust continuum studies with bolometer arrays installed on single-dish telescopes revealed the core mass distributions in nearby regions like p Ophiuchus and Orion. The power-law distributions dN/dM a: M-a of the pre-stellar dust condensations resemble the Salpeter IMF with typical values of a between 2 and 2.5. While this finding supports the turbulent fragmentation scenario, it only covers a small range of the IMF, and it is necessary to expand such studies to higher-mass regions. Several groups investigated with the same single-dish instruments samples of massive star-forming regions, and the derived cumulative mass distributions of their samples resemble the Salpeter IMF as well [6, 56, 63, 74]. However, since these massive star-forming regions are at about an order of magnitude larger distances, these single-dish studies do not resolve individual protostars but they average over the whole forming protocluster. Hence, these cumulative mass distributions rather resemble protocluster mass functions, and it is far from clear whether they can set any constraints on the formation of the IMF. Therefore, again high spatial resolution is required to resolve individual protostars within the forming protoclusters. So far, only one study resolved and imaged enough sub-sources within a massive star-forming region that a derivation of a core mass function appeared meaningful. The two massive gas condensation in IRAS 19410+3543 were resolved in 1.3mm continuum observations with the PdBI into 24 sub-sources (Fig. 5, [9]), and the resulting core mass function again has a power-law distribution with a —2.5. Although the statistics are still poor, and a few caveats have to be taken into account in this analysis ("Are all dust continuum peaks of protostellar nature"? "Is the assumption of the same temperature for all sub-sources justified?)", it is exciting that all studies so far find core mass distributions resembling the IMF, indicating that turbulent fragmentation processes are really important. However, especially in the high-mass regime, we cannot draw any definitive conclusions yet. Larger source samples observed at high angular resolution are required to base the results on solid grounds. Furthermore, it will be necessary to investigate the temperature structure of the sub-sources in detail, and we need additional spectral line data to determine complementary virial masses and thus establish that the observed sub-sources are really bound and not transient structures.

Another observational curiosity of several high-mass star-forming regions that appear very similar from previous single-dish observations (They are at similar distances, should be at approximately the same evolutionary stage, and have comparable luminosities and other massive star-forming tracers like masers and outflows.) is that they reveal very different sub-structures when observed at high angular resolution. Figure 6 shows a few interferometric example studies of regions, where one would have expected comparable fragmentation results. However, some of the regions show only a single massive

Fig. 5. Single-dish and interferometer mm dust continuum observations toward the young high-mass star-forming region IRAS 19410+2336 [9]. The left panel shows the 1.2 mm map obtained with the IRAM 30 m telescope. The middle and right panel present 3 and 1.3 mm continuum images from the PdBI with increasing spatial resolution. The synthesized beams are shown at the bottom-left of each panel. The resolution difference between the left and right panels are an order of magnitude

Fig. 5. Single-dish and interferometer mm dust continuum observations toward the young high-mass star-forming region IRAS 19410+2336 [9]. The left panel shows the 1.2 mm map obtained with the IRAM 30 m telescope. The middle and right panel present 3 and 1.3 mm continuum images from the PdBI with increasing spatial resolution. The synthesized beams are shown at the bottom-left of each panel. The resolution difference between the left and right panels are an order of magnitude central source, even at the highest angular resolution (Fig. 6 top row), whereas other sources exhibit many sub-sources as expected from a star-forming cluster (Fig. 6 bottom row). While the latter is less of a surprise since we know that massive stars almost always form in clusters, finding only a single peak in other regions is more difficult to explain. While all these studies are sensitivity limited and usually are not capable of tracing any condensation below approximately 1 M0, one would nevertheless expect to find at least some intermediate-mass objects in the vicinity of the central object. One important additional fact is that toward most regions deeply embedded near-infrared clusters have been found, and hence they are no isolated objects. The existence of an embedded near-infrared protocluster in the vicinity of a massive forming star may be interpreted in the direction that the low-mass stars form first and the high-mass objects later (see also [43]). However, currently we do not understand why various apparently similar massive star-forming regions show such a diverse fragmentation behavior. Is it only an observational bias and the selected sources may in fact be not as similar as we believe, or could there be different paths massive star-forming regions can fragment in the first place?

0 0

Post a comment