We present a survey of 28 molecular outflows driven by low-mass protostars, all of which are sufficiently isolated spatially and/or kinematically to fully separate into individual outflows. Using a combination of new and archival data from several single-dish telescopes, 17 outflows are mapped in 12 CO (2-1) and 17 are mapped in 12 CO (3-2), with 6 mapped in both transitions. For each outflow, we calculate and tabulate the mass (M flow ), momentum (P flow ), kinetic energy (E flow ), mechanical luminosity (L flow ), and force (F flow ) assuming optically thin emission in LTE at an excitation temperature, T ex , of 50 K. We show that all of the calculated properties are underestimated when calculated under these assumptions. Taken together, the effects of opacity, outflow emission at low velocities confused with ambient cloud emission, and emission below the sensitivities of the observations increase outflow masses and dynamical properties by an order of magnitude, on average, and factors of 50-90 in the most extreme cases. Different (and non-uniform) excitation temperatures, inclination effects, and dissociation of molecular gas will all work to further increase outflow properties. Molecular outflows are thus almost certainly more massive and energetic than commonly reported. Additionally, outflow properties are lower, on average, by almost an order of magnitude when calculated from the 12 CO (3-2) maps compared to the 12 CO (2-1) maps, even after accounting for different opacities, map sensitivities, and possible excitation temperature variations. It has recently been argued in the literature that the 12 CO (3-2) line is subthermally excited in outflows, and our results support this finding.
Data ReductionLow-order polynomial baselines were subtracted from all of the raw data using the default software package for each telescope: Continuum and Line Analysis Singledish Software (CLASS 11 ) for APEX, CSO, SRAO, and SMT, NEWSTAR 12 for ASTE, and Starlink 13 for the JCMT. These packages were then used to combine together all observations of a particular source and write out FITS datacubes on grids of Nyquist sampled spatial pixels. Further analysis was performed using custom IDL procedures.
RESULTS