Small-scale density and velocity structure of a molecular cloud edge

This paper presents a detailed analysis of a random piece of molecular gas, chosen in a weakly CO-emitting part of a cloud edge in the Perseus-Auriga complex. The data set consists of high angular resolution observations in the (CO)-C-12 J = 1-0, J = 2-1, J = 3-2, and J = 4-3 and CS J = 2-1 and J = 3-2 lines, combined with CO (J = 3-2) and (CO)-C-13 (J = 2-1) observations obtained previously. The observational results can be summarized as follows: (i) At many locations the CO line profiles exhibit weak, broad line wings, superimposed on a narrow intense line core. (ii) The CO line emission is highly structured down to the resolution of the observations (similar to 0.014 pc), but the spatial distribution of the line core emission is different from that of the wing emission. The former is concentrated mostly in two distinct structures of size similar to 0.06 pc, while the latter is concentrated in a long filamentary structure (l similar to 0.2. pc) that crosses the mapped held and is barely resolved in its transverse dimension. (iii) The J = 2-1 to J=1-0 CO Line ratio is approximately constant across the entire field and has the same value, R = 0.62+/-0.08, in the GO-bright areas as in the almost 10 times weaker areas. (iv) A weak CS (2-1) line has been detected at the core velocity, and only an upper limit has been obtained for the CS (3-2) line, and (v) the low column density gas that emits in the line wings has been detected clearly in CO (3-2) and detected tentatively in the CO (4-3) line. The present work strengthens the results of our previous study, that the edges of molecular clouds are weak CO emitters when observed at the parsec scale, say, because the CO emission there is beam-diluted emission of dense (nH(2) > 10(4) cm(-3)), cold (T-k < 15 K) structures. The moderate CO beam-averaged optical depth and the smoothness of the CO profiles set an upper limit of similar to 35 AU for the size of the cells within which each CO photon interacts with the gas. This result, though, does not necessarily imply that the GO-free regions are structured similarly and, indeed, a moderate-opacity region of GO-free gas and dust is required to surround the GO-emitting structures in order to provide shielding from the interstellar radiation held. Our work also reveals, for the first time, that the line wing emission also originates in small-scale structure. The CO (4-3) wing measurement is critically important in that it constrains the wing gas temperature to be in the range from T-k similar to 25 K, for dense gas at n(H2) similar to 10(3) cm(-3), up to T-k similar to 250 K for n(H2) similar to 200 cm(-3). Its filamentary morphology is consistent with the idea that it is dilute and warm gas confined to the specific structures in which the energy of turbulence is being dissipated, in an intermittent way. The large body of observational results presently available on the CO emission properties of non-star-forming molecular clouds, from the smallest to the largest scales, is not consistent with a picture of randomly moving dense clumps immersed in a lower density medium. The AU scale dense CO cells have to be distributed on a fractal set with correlated velocities. They are likely to be dynamically connected to the turbulent velocity field of the gas that fills the volume (atomic and molecular hydrogen) and probably trace the active regions of large vorticity either because they would tend to be trapped in these structures or because those may be regions of enhanced CO formation rate.