On the typical width of Herschel filaments

Abstract

Herschel studies suggest that nearby (d < 500 pc) molecular filaments have a typical half-power width ~0.1pc, but this finding has been questioned on the ground that the measured widths tend to increase with distance. Here we revisit the dependence of measured filament widths on distance or equivalently spatial resolution, in an effort to determine whether nearby molecular filaments have a characteristic half-power width or whether this is an artifact of the finite resolution of the Herschel data. We perform a convergence test on the B211/213 filament in Taurus, by degrading the resolution of the Herschel data several times and re-estimating the filament width from the resulting column density profiles. We also compare the widths measured for the Taurus filament and other filaments from the Herschel Gould Belt survey to those found for synthetic filaments with various types of simple, idealized column density profiles. We find that the measured filament widths do increase slightly as the spatial resolution worsens and/or the distance to the filaments increases. However, this trend is entirely consistent with what is expected from simple beam convolution for filaments with density profiles that are Plummer-like and have intrinsic half-power diameters ~0.08-0.1 pc and logarithmic slopes 1.5 < p < 2.5 at large radii, as observed in many cases including the Taurus filament. Due to the presence of background noise fluctuations, deconvolution of the measured widths from the telescope beam quickly becomes inaccurate. We conclude that the typical half-power filament width ~0.1 pc measured with Herschel in nearby clouds most likely reflects the presence of a true common scale in the filamentary structure of the cold interstellar medium, at least in the solar neighborhood. We suggest that this common scale may correspond to the magnetized turbulent correlation length in molecular clouds.