Tracing the mass during low-mass star formation. IV. Observations and modeling of the submillimeter continuum emission from class I protostars

We present results from the observations and modeling of 17 Class I cores with the Submillimetre Common Users Bolometer Array (SCUBA) on the James Clerk Maxwell Telescope ( JCMT). Our sample consists of cores with 64 K < T-bol < 270 K, 0 : 2 < L-obs/L. < 12, and 50 < L-obs/L-smm < 1000. By modeling the transfer of radiation through the envelope for nine cores, we find, for a power-law distribution n(r) = n(f) (r/r(f))(-p), the average and standard deviation [p] = 1.6 +/- 0.4 and a median of p = 1.8. However, the inclusion of a disk or other pointlike component can cause the derived p to be shallower by as much as 0.5. We discuss uncertainties due to the interstellar radiation field (ISRF), disks, dust opacity, and outer radii in our modeling results. We find no evidence for a truncated outer radius or radially variant dust properties in most sources. Uncertainty in the strength of the ISRF and possible existence of a disk contribute the greatest uncertainty in p. In addition, we test the Shu collapse model for our sources and discuss the application of simpler analyses that derive a density power-law distribution directly from the slope of the intensity radial profile. The total mass of the envelope in our sample has a range of 0.04 < M-env/M. < 5.0, but these masses disagree with the virial masses derived from molecular line observations, indicating that observations of molecular lines do not trace the mass in some Class I cores. We also discuss several sources individually. In particular, IRAS 03256+3055, with its unique morphology, is an ideal object for testing theories of fragmentation in the formation of low-mass protostars. Also, we note the possibility, through some simple calculations, that IRAS 04385+2550 is a young, forming substellar object. Finally, we discuss the nature of these sources in light of various evolutionary indicators and find that T-bol and L-obs/L-smm are often inconsistent in distinguishing Class 0 from Class I cores. We note that, in this sample, the L-obs/L-smm criterion redefines many of these Class I sources (by T-bol) as Class 0 sources.