Abstract

We present a study of the density structure of the protostellar collapse candidate B335 using continuum observations from the IRAM Plateau de Bure Interferometer made at wavelengths of 1.2 and 3.0 mm. We analyze these data, which probe spatial scales from 5000 to 500 AU, directly in the visibility domain by comparison with synthetic observations constructed from models that assume different physical conditions. This approach allows for much more stringent constraints to be derived from the data than from analysis of images. A single radial power law in density provides a good description of the data, with a best-fit power-law density index p = 1.65 ± 0.05. Through simulations, we quantify the sensitivity of this result to various model uncertainties, including assumptions of temperature distribution, outer boundary, dust opacity spectral index, and an unresolved central component. The largest uncertainty comes from the unknown presence of a centralized point source. The maximal point source, with 1.2 mm flux of F = 12 ± 7 mJy, reduces the power-law density index to p = 1.47 ± 0.07. The remaining sources of systematic uncertainty, of which the most important is the radial dependence of the temperature distribution, likely contribute a total uncertainty at the level of ?p ? 0.2. Taking into account the uncertainties, we find strong evidence that the power-law index of the density distribution within 5000 AU is significantly less than the value at larger radii, close to 2.0, from previous studies of dust emission and extinction. Images made from the data show clear departures from spherical symmetry, with the globule being slightly extended perpendicular to the outflow axis. The inclusion of a crude model of the outflow as a hollow bipolar cone of constant opening angle improves the fit and leaves the resulting density power-law index unchanged. These results conform well to the generic paradigm of isolated, low-mass star formation, which predicts a power-law density index close to p = 1.5 for an inner region of gravitational free fall onto the protostar. However, the standard inside-out collapse model does not fit the data as successfully as a simple p = 1.5 power law, because of the relative shallowness of the predicted density profile just within the infall radius.