Abstract
We have observed 150 regions of massive star formation, selected originally by the presence of an H2O maser, in the J = 5 → 4, 3 → 2, and 2 → 1 transitions of CS, and 49 regions in the same transitions of C34S. Over 90% of the 150 regions were detected in the J = 2 → 1 and 3 → 2 transitions of CS, and 75% were detected in the J = 5 → 4 transition. We have combined the data with the J = 7 → 6 data from our original (1992) survey to determine the density by analyzing the excitation of the rotational levels. Using large velocity gradient models, we have determined densities and column densities for 71 of these regions. The gas densities are very high (⟨log n⟩ = 5.9), but much less than the critical density of the J = 7 → 6 line. Small maps of 25 of the sources in the J = 5 → 4 line yield a mean diameter of 1.0 pc. Several estimates of the mass of dense gas were made for the sources for which we had sufficient information. The mean virial mass is 3800 M☉. The mean ratio of bolometric luminosity to virial mass (L/M) is 190, about 50 times higher than estimates made using CO emission, suggesting that star formation is much more efficient in the dense gas probed in this study. The depletion time for the dense gas is ~1.3 × 107 yr, comparable to the timescale for gas dispersal around open clusters and OB associations. We find no statistically significant line width-size or density-size relationships in our data. Instead, both line width and density are greater for a given size than would be predicted by the usual relationships. We find that the line width increases with density, the opposite of what would be predicted by the usual arguments. We estimate that the luminosity of our Galaxy (excluding the inner 400 pc) in the CS J = 5 → 4 transition is 15-23 L☉, considerably less than the luminosity in this line within the central 100 pc of NGC 253 and M82. In addition, the ratio of far-infrared luminosity to CS luminosity is higher in M82 than in any cloud in our sample.
Export citation and abstract BibTeX RIS