ALMA Observations of Fragmentation, Substructure, and Protostars in High-mass Starless Clump Candidates

Abstract

The initial physical conditions of high-mass stars and protoclusters remain poorly characterized. To this end, we present the first targeted ALMA Band 6 1.3 mm continuum and spectral line survey toward high-mass starless clump candidates, selecting a sample of 12 of the most massive candidates (4 x 10(2) M-circle dot less than or similar to Mcl less than or similar to 4 x 10(3) M-circle dot) within d(circle dot) < 5 kpc. The joint 12 + 7 m array maps have a high spatial resolution of.3000 au (0.015 pc,.syn 0.8) and have high point-source mass-completeness down to M >> 0.3 M. at 6srms (or 1srms column density sensitivity of N = 1.1 x 10(22) cm(-2)). We discover previously undetected signposts of low-luminosity star formation from CO J = 2 -> 1 and SiO J = 5 -> 4 bipolar outflows and other signatures toward 11 out of 12 clumps, showing that current MIR/FIR Galactic plane surveys are incomplete to low-and intermediate-mass protostars (L-bol less than or similar to 50 L-circle dot), and emphasizing the necessity of high-resolution follow-up. We compare a subset of the observed cores with a suite of radiative transfer models of starless cores. We find a high-mass starless core candidate with a model-derived mass consistent with 29(15)(52)M(circle dot) when integrated over size scales of R < 2 x 10(4) au. Unresolved cores are poorly fit by radiative transfer models of externally heated Plummer density profiles, supporting the interpretation that they are protostellar even without detection of outflows. A high degree of fragmentation with rich substructure is observed toward 10 out of 12 clumps. We extract sources from the maps using a dendrogram to study the characteristic fragmentation length scale. Nearest neighbor separations, when corrected for projection with Monte Carlo random sampling, are consistent with being equal to the clump average thermal Jeans length (lambda(j,th); i.e., separations equal to 0.4-1.6 x lambda(j,th)). In the context of previous observations that, on larger scales, see separations consistent with the turbulent Jeans length or the cylindrical thermal Jeans scale (approximate to 3-4 x lambda(j, th)), our findings support a hierarchical fragmentation process, where the highest-density regions are not strongly supported against thermal gravitational fragmentation by turbulence or magnetic fields.