2020MNRAS.494.3675C

In typical environments of star-forming clouds, converging supersonic turbulence generates shock-compressed regions, and can create strongly magnetized sheet-like layers. Numerical magnetohydrodynamic simulations show that within these post-shock layers, dense filaments and embedded self-gravitating cores form via gathering material along the magnetic field lines. As a result of the preferred-direction mass collection, a velocity gradient perpendicular to the filament major axis is a common feature seen in simulations. We show that this prediction is in good agreement with recent observations from the CARMA Large Area Star Formation Survey (CLASSy), from which we identified several filaments with prominent velocity gradients perpendicular to their major axes. Highlighting a filament from the north-west part of Serpens South, we provide both qualitative and quantitative comparisons between simulation results and observational data. In particular, we show that the dimensionless ratio C_v ≡ Δv_h²/(GM/L), where Δv_h is half of the observed perpendicular velocity difference across a filament, and M/L is the filament's mass per unit length, can distinguish between filaments formed purely due to turbulent compression and those formed due to gravity-induced accretion. We conclude that the perpendicular velocity gradient observed in the Serpens South north-west filament can be caused by gravity-induced anisotropic accretion of material from a flattened layer. Using synthetic observations of our simulated filaments, we also propose that a density-selection effect may explain observed subfilaments (one filament breaking into two components in velocity space) as reported in recent observations.