Abstract |
Herschel, the ESA/NASA submillimeter satellite, imaged the continuum dust emission
of the interstellar medium with unprecedented detail. Studies of Herschel images revealed a complex
structure for interstellar clouds. Topological-analysis tools applied on Herschel maps of molecular
clouds traced many elongated density enhancements (filaments), along which most of the density
peaks (cores) were found to lie. Since molecular cloud cores are the birthplaces of protostars, this
picture was taken to signify a paradigm shift in our understanding of star formation, with clouds
breaking up into filaments first, and then cores forming within those filaments. Filaments have thus
been viewed as the first stage of star formation. In this thesis, we have undertaken a study of such
filamentary structures not only in integrated intensity (as in the Herschel data) but also in line-of-sight
velocity, using 13CO line emission data in the nearby Taurus molecular cloud. Our study constitutes
the largest area (~100 deg2) analyzed for filamentary structures in position-position-velocity
space. We have used the topological analysis tool employed in many Herschel studies, the DisPerSe
software. However, in addition, we have post-processed our results to include: (a) a more strict definition of filaments (at least 3:1 aspect ratio and structures with cross-section density profiles peaked
on the spine of the filament); (b) studies of different velocity components. Our results are surprising:
although DisPerSe identifies, as in Herschel maps, an intricate “spider web” of hundreds of
filamentary structures, only 5 of them comply with our filament definition criteria. In addition, unlike
Herschel analyses, which find a characteristic width for their filaments of ~ 0:1 pc, we find a
much broader distribution in profile widths in our structures, with a wide peak at much higher values
(0.4-0.5 pc). The filaments that do survive have a complicated velocity structure, with different
parts of the filaments being grouped at different velocities. Even if these filaments are cylindrical
today, the velocity dispersion along the line of sight implies that their shape would not be retained
even for one free fall time. We have compared our results with those obtained from simulated maps
consisting only of randomly placed “cores” plus noise, using the same methodology (DisPerSe + postprocessing).
Apparent filaments can also be traced in such maps, and they have similar profile shapes
with the Herschel and 13CO filaments, while most cores do get included in the filaments. These results
indicate that care has to be applied when interpreting the nature of filaments in Herschel maps.
|