Assessing molecular line diagnostics of triggered star formation using synthetic observations

We investigate observational signatures of triggered star formation in bright rimmed clouds (BRCs) by using molecular line transfer calculations based on radiation-hydrodynamic radiatively driven implosion models. We find that for BRCs the separation in velocity between the line profile peak of an optically thick and an optically thin line is determined by both the observer viewing angle and the density of the shell driving into the cloud. In agreement with observations, we find that most BRC line profiles are symmetric and that asymmetries can be either red or blue, in contrast to the blue dominance expected for a collapsing cloud. Asymmetries in the line profiles arise when an optically thick line is dominated by the shell and an optically thin line is dominated by the cloud interior to the shell. The asymmetries are red or blue depending on whether the shell is moving towards or away from the observer, respectively. Using the known motions of the molecular gas in our models we rule out the 'envelope expansion with core collapse' mechanism as the cause of the lack of blue-asymmetry in our simulated observations. We show that the absence of a strong photon-dominated region (PDR) around a BRC may not rule out the presence of triggered star formation: if the BRC line profile has a strong blue component then the shell is expected to be driving towards the observer, suggesting that the cloud is being viewed from behind and the PDR is obstructed. This could explain why BRCs such as SFO 80, 81 and 86 have a blue secondary peak and only a weak PDR inferred at 8 mu m. Finally we also test the use of (CO)-C-12, (CO)-C-13 and (CO)-O-18 as diagnostics of cloud mass, temperature and column density. We find that the inferred conditions are in reasonable agreement with those from the models. Calculating the cloud mass assuming spherical symmetry is shown to introduce an error of an order of magnitude whereas integrating the column density over a given region is found to introduce an error of up to a factor of 2.