The Origin of OB Cluster

The majority of gas in molecular clouds is not involved in forming stars, es- pecially the most massive OB stars. OB stars are observed to form in very special regions of molecular clouds —massive molecular clumps, typically of ~1 parsec scale size. The large mass and high density are needed for the collapsing clumps to feed the forming massive stars and the associated stellar cluster at a high enough rate. How the dense massive molecular clumps with internal conditions conducive to massive star formation, come into existence? How the massive clumps fragment into clusters, have been open questions. We have derived extensive imaging programs, to resolve the structures and the kinematics in several L >106 L⊙ OB cluster forming molecular clouds, from 50 pc scale to 0.05 pc scale. Based on the IRAM 30m and the PMO 14m single dish telescope observations on the molecular clouds G10.6-0.4, G33.92+0.11, G20.08-0.14, W33 and the Galactic mini starburst region W49A, we found that the most active high-mass star-forming regions are those >10^4 M⊙ massive molecular clumps located at the convergences/confluences of >5pc scale, 10^3-10^4 M⊙ molecular filaments. We resolved these confluences of >5 pc molecular filaments in molecular clouds G10.6-0.4 and G33.92+0.11, by higher angular resolution NRAO VLA/EVLA(or JVLA) and SMA observations. Respective of these two targets are centrally em- bedded with an edge-on and an face-on candidates of geometrically flattened OB cluster forming molecular clumps. Series of previous observations on G10.6-0.4 have established the ground work, suggesting the radial inward motion of the molecular gas. We have improved the angular resolutions of the observations at all scales. Our analysis of the interferometric and single dish data further show: ‧ 5-10 pc Scale: The molecular clouds show a Hub-Filament geometric con- figuration. The >5 pc scale molecular filaments (several times 10^3 M⊙ each) have a radiative distribution and are gravitationally accelerated toward their convergences. In G10.6-0.4, the >5pc molecular filaments merge into an extremely massive (>20,000M⊙ in a 2 pc area) molecular clump, which are 10-100 times more massive than the massive molecular clumps embedded in the molecular filaments. The brightest sources of infrared emission or radio continuum emission are concentrated in those centralized molecular clumps at confluences. The observations of the 1.2 mm dust continuum emission show the individual filaments are embedded with parsec scale, regularly spaced massive molecular clumps (200-500 M⊙ each). Higher angular resolution observations of selected massive molecular clumps embedded in filaments further unveiled groups of embedded regularly spaced (~0.1pc) massive molecular cores (10-50 M⊙ each). Some of these cores emanates (proto)stellar outflows with terminal velocities up to 110kms^−1. The fragmentation of the massive molecular filaments appear to be hierarchical/fractal. ‧ 1-2 pc Scale: The dense filaments occupy 3-27% of the spatial volume in the central 2 pc scale region in G10.6-0.4. The localized gas concentration in filaments leads to a 1.9-5.8 times shorter local free-fall timescale than the global contraction timescale, which is conducive to fragmentation and the formation of the OB cluster. We detected multiple sources of molecular outflows and water maser. The detected 6.6×10^46 erg total energy of high velocity molecular outflows can replenish the dissipated turbulence energy in ~10^4 years. In the face-on sample G33.92+0.11, we explicitly resolved several molecular arms winding from the ~1pc radii to the central 0.1pc massive molecular cores. At the locations where the exterior filaments enter the 0.3pc radius of the flattened rotating molecular clump, we detected enhanced emission of shock tracers such as SiO, OCS. ‧ 0.05-0.5 pc Scale: At this scale, the spectral line observations on the edge-on sample G10.6-0.4 show that the dominant systematic motion of the molecular gas is already relaxed to the co-planar rotational motion that marginally balances the gravity. The radial (infall) velocity (1-2 kms−1) of gas is about 30% of the gas speed. We found a radially rapid decrease of specific angular momentum of 5.2 pc km s−1 per parsec within the turnover radius of rT ∼0.3pc. The densest gas is geometrically flattened. The Toomre''s Q parameter decreases with the radius, becoming smaller than 1.5-2.0 at a 0.015-0.03 pc (~1") radius, that is conducive to the existence of the grand–designed spiral arms; the Toomre''s Q parameter becomes less than 1.0 at >0.03-0.06 pc radii, that will allow the localized fragmentation. We performed the deepest SMA observations of the linearly polarized dust emission on G10.6-0.4, and did not detect organized magnetic field within the central 0.5 pc radius. Inside the 0.3 pc turnover radius where the exterior molecular arms enter the geometrically round face–on candidate G33.92+0.11, we explicitly resolved the predicted spiraling grand-designed molecular arms embedded with the localized molecular cores. At the face-on sample G33.92+0.11, the spectra minimally biased by the global infall/rotational motions allow us to constrain the turbulence velocity to be FWHM10 times larger molecular mass than the virial mass