2019A&A...628A..21W

Context. Stellar feedback from high-mass stars shapes the interstellar medium, and thereby impacts gas that will form future generations of stars. However, due to our inability to track the time evolution of individual molecular clouds, quantifying the exact role of stellar feedback on their star formation history is an observationally challenging task.
Aims. In the present study, we take advantage of the unique properties of the G316.75-00.00 massive-star forming ridge to determine how stellar feedback from O-stars impacts the dynamical stability of massive filaments. The G316.75 ridge is 13.6pc long and contains 18900M_☉ of H₂ gas, half of which is infrared dark and half of which infrared bright. The infrared bright part has already formed four O-type stars over the past 2 Myr, while the infrared dark part is still quiescent. Therefore, by assuming the star forming properties of the infrared dark part represent the earlier evolutionary stage of the infrared bright part, we can quantify how feedback impacts these properties by contrasting the two.
Methods. We used publicly available Herschel/HiGAL and molecular line data to measure the ratio of kinetic to gravitational energy per-unit-length, α_vir^line, across the entire ridge. By using both dense (i.e. N₂H⁺ and NH₃) and more diffuse (i.e. ¹³CO) gas tracers, we were able to compute α_vir^line for a range of gas volume densities (∼1x10²-1x10⁵cm^–3). Results. This study shows that despite the presence of four embedded O-stars, the ridge remains gravitationally bound (i.e. α_vir^line≤2) nearly everywhere, except for some small gas pockets near the high-mass stars. In fact, α_vir^line is almost indistinguishable for both parts of the ridge. These results are at odds with most hydrodynamical simulations in which O-star-forming clouds are completely dispersed by stellar feedback within a few cloud free-fall times. However, from simple theoretical calculations, we show that such feedback inefficiency is expected in the case of high-gas-density filamentary clouds.Conclusions. We conclude that the discrepancy between numerical simulations and the observations presented here originates from different cloud morphologies and average densities at the time when the first O-stars form. In the case of G316.75, we speculate that the ridge could arise from the aftermath of a cloud-cloud collision, and that such filamentary configuration promotes the inefficiency of stellar feedback. This does very little to the dense gas already present, but potentially prevents further gas accretion onto the ridge. These results have important implications regarding, for instance, how stellar feedback is implemented in cosmological and galaxy scale simulations.