Context. Theory predicts, and observations confirm, that the column density ratio of a molecule containing D to its counterpart containing H can be used as an evolutionary tracer in the low-mass star formation process. Aims. Since it remains unclear if the high-mass star formation process is a scaled-up version of the low-mass one, we investigated whether the relation between deuteration and evolution can be applied to the high-mass regime. Methods. With the IRAM-30 m telescope, we observed rotational transitions of N 2 D + and N 2 H + and derived the deuterated fraction in 27 cores within massive star-forming regions understood to represent different evolutionary stages of the massive-star formation process.Results. The abundance of N 2 D + is higher at the pre-stellar/cluster stage, then drops during the formation of the protostellar object(s) as in the low-mass regime, remaining relatively constant during the ultra-compact HII region phase. The objects with the highest fractional abundance of N 2 D + are starless cores with properties very similar to typical pre-stellar cores of lower mass. The abundance of N 2 D + is lower in objects with higher gas temperatures as in the low-mass case but does not seem to depend on gas turbulence. Conclusions. Our results indicate that the N 2 D + -to-N 2 H + column density ratio can be used as an evolutionary indicator in both lowand high-mass star formation, and that the physical conditions influencing the abundance of deuterated species likely evolve similarly during the processes that lead to the formation of both low-and high-mass stars.

The unprecedented sensitivity of Herschel coupled with the high resolution of the HIFI spectrometer permits studies of the intensity-velocity relationship I(v) in molecular outflows, over a higher excitation range than possible up to now. In the course of the CHESS Key Program, we have observed toward the bright bowshock region L1157-B1 the CO rotational transitions between J=5-4 and J=16-15 with HIFI, and the J=1-0, 2-1 and 3-2 with the IRAM-30m and the CSO telescopes. We find that all the line profiles I CO (v) are well fit by a linear combination of three exponential laws ∝ exp(−|v/v 0 |) with v 0 = 12.5, 4.4 and 2.5 km s −1 . The first component dominates the CO emission at J ≥ 13, as well as the high-excitation lines of SiO and H 2 O. The second component dominates for 3 ≤ J up ≤ 10 and the third one for J up ≤ 2. We show that these exponentials are the signature of quasi-isothermal shocked gas components : the impact of the jet against the L1157-B1 bowshock (T k ≃ 210 K), the walls of the outflow cavity associated with B1 (T k ≃ 64 K) and the older cavity L1157-B2

Observations of molecular clouds reveal a complex structure, with gas and dust often arranged in filamentary rather than spherical geometries. The associations of pre-and proto-stellar cores with the filaments suggest a direct link with the process of star formation. Any study of the properties of such filaments requires a representative samples from different enviroments and so an unbiased detection method. We developed such an approach using the Hessian matrix of a surface-brightness distribution to identify filaments and determine their physical and morphological properties. After testing the method on simulated, but realistic filaments, we apply the algorithms to column-density maps computed from Herschel observations of the Galactic Plane obtained by the Hi-GAL project. We identified ∼ 500 filaments, in the longitude range of l=216.5 o to l=225.5 o , with lengths from ∼1 pc up to ∼30 pc and widths between 0.1 pc and 2.5 pc. Average column densities are between 10 20 cm −2 and 10 22 cm −2 . Filaments include the majority of dense material with N H 2 >6×10 21 cm −2 . We find that the pre-and proto-stellar compact sources already identified in the same region are mostly associated with filaments. However, surface densities in excess of the expected critical values for high-mass star formation are only found on the filaments, indicating that these structures are necessary to channel material into the clumps. Furthermore, we analyze the gravitational stability of filaments and discuss their relationship with star formation.

In order to study the fragmentation of massive dense cores, which constitute the cluster cradles, we observed with the PdBI in the most extended configuration the continuum at 1.3 mm and the CO (2-1) emission of four massive cores. We detect dust condensations down to ∼ 0.3 M ⊙ and separate millimeter sources down to 0.4 ′′ or 1000 AU, comparable to the sensitivities and separations reached in optical/infrared studies of clusters. The CO (2-1) high angular resolution images reveal high-velocity knots usually aligned with previously known outflow directions. This, in combination with additional cores from the literature observed at similar mass sensitivity and spatial resolution, allowed us to build a sample of 18 protoclusters with luminosities spanning 3 orders of magnitude. Among the 18 regions, ∼ 30% show no signs of fragmentation, while 50% split up into 4 millimeter sources. We compiled a list of properties for the 18 massive dense cores, such as bolometric luminosity, total mass, and mean density, and found no correlation of any of these parameters with the fragmentation level. In order to investigate the combined effects of magnetic field, radiative feedback and turbulence in the fragmentation process, we compared our observations to radiation magneto-hydrodynamic simulations, and obtained that the low-fragmented regions are well reproduced in the magnetized core case, while the highly-fragmented regions are consistent with cores where turbulence dominates over the magnetic field. Overall, our study suggests that the fragmentation in massive dense cores could be determined by the initial magnetic field/turbulence balance in each particular core.