The IRAM key-project: small-scale structure of pre-star-forming regions. I. Observational results

This paper presents the observational results of the first IRAM key-project and a straightforward interpretation of the most salient features of the data. The project is devoted to the analysis of the processes which drive the dissipation of the non-thermal support of molecular clouds, a mandatory step toward the formation of almost thermally supported cores. The selected fields therefore all contain a starless dense core of small internal velocity dispersion. The maps include the core (of size ~ 0.1 pc) or a fraction of it, and extend over large areas of their environment (several arc minutes, or several tenths of pc at the distance of the clouds, d<150pc). Maps have been completed in five transitions, ¹²COJ=1-0e\ and (J=2-1), ¹³COJ=1-0e\ and (J=2-1) and C¹⁸OJ=1-0e, at high angular resolution (22'' and 11'' at low and high frequency respectively, with a sampling of 7.5'') and a velocity resolution of 0.05km s^{-1}. The spatial resolution of the high frequency maps is ~ 1700 AU. The data set, because of its size, the good signal-to-noise ratio of the spectra and the multiplicity of the lines observed, provides several new results, as follows: (1) there is little unresolved structure left in the maps of line integrated emission, but unresolved structure is still present in the channel maps of all the fields and all the lines. The velocity gradients involved reach values as large as 10km s^{-1} pc(-1) , implying large accelerations never observed before at small scale in non star-forming clouds, (2) the texture and velocity dispersion of the gas bright in ¹²CO and barely detected in ¹³CO are significatively different from those of the gas bright in the three isotopes. The gas bright in ¹²CO only exhibits filamentary structure with, in some cases, unresolved transverse dimensions, and aspect ratios larger than ~ 5. Its velocity dispersion is much larger than that of the latter. Unexpectedly, it is in the more opaque transitions and in the gas component of larger velocity dispersion that the smallest scale structure has been observed, (3) the dense cores are not isolated structures but are connected, in space and velocity, to another kind of filamentary structures, bright in ¹³CO and C¹⁸O, (4) the brightness temperature ratio of the two lowest CO rotational transitions is remarkably uniform: R(2-1/1-0)=0.65+/-0.15 for 80% of the data points in the three fields, from the brightest to the weakest detected lines, across the whole profiles and for both ¹²CO and ¹³CO isotopes, (5) the ¹³CO lines reach intensities as large as those of the ¹²CO lines, though the line profiles are in general neither flat-topped nor self-reversed. >From these well defined spectral properties, we infer that the lines have to form in very small cells, weakly coupled radiatively to one another, optically thick in the ¹²CO lines and that the line shapes are governed mostly by the spatial and velocity dilution of the emitting cells in the beam. Under the simple assumption that the cells are statistically independent, we estimate that they are smaller than ~ 200 AU with H_2 densities n_{H_2} ~ a few 10(3) cm^{-3} in the gas component barely detected in ¹³CO, and are up to two orders of magnitude denser in the component bright in the three isotopes. We also notice an anticorrelation between the intensities of the ¹³CO lines and their linewidths which we interpret as a signature of a gradual loss of the non-thermal support which increases the phase-space radiative coupling of the cells.