Institution(s): 1. University of Bonn
Recent high resolution observations of dense regions of molecular clouds and massive gaseous clumps with instruments like Herschel and ALMA have revealed intricate and filamentary overdensity structures in them. Such progenitors of massive starburst clusters are in contrast with smooth, centrally-pronounced profiles of the latter. In this work, we intend to constrain massive, substructured stellar distributions that would evolve to cluster-like profiles at very young ages (~Myr), as seen in starburst clusters. Taking the well observed NGC3603 Young Cluster (NYC) as an example, we compute the infall and final merger of filament-like compact (0.1-0.3 pc) subclusters, totalling 10000 M_sun, from a range of spatial scales and modes of sub-clustering, using direct N-body calculations. These calculations infer an allowable span of approx. 2.5 pc from which the subclusters can fall in a gas potential and merge to form a single centrally-dense structure in near dynamical equilibrium, within the young age of NYC (1-2 Myr). However, these merged clusters are too compact and centrally overdense compared to typical young clusters. Our N-body calculations, beginning from such compact initial conditions, show that even stellar wind and supernova mass loss, dynamical heating from retaining black holes, external tidal field and heating due to tight O-star binaries together cannot expand these clusters to their observed sizes, even in 100 Myr. Hence an explosive gas dispersal phase seems essential for forming starburst and other young clusters observed in the Milky Way and in the Local Group which can expand the clusters to their observed sizes and concentrations; including that for NYC with approx. 30% clump star formation efficiency. However, some observed massive but highly extended (>10 pc) , >10 Myr old clusters better fit a slow (several Myr timescale) gas dispersal from parsec-scale initial profiles, which can be the future of embedded systems like W3 Main.