Herschel observations of nearby molecular clouds suggest that interstellar filaments and prestellar cores represent two fundamental steps in the star formation process. The observations support a picture of low-mass star formation according to which ~ 0.1 pc-wide filaments form first in the cold interstellar medium, probably as a result of large-scale compression of interstellar matter by supersonic turbulent flows, and then prestellar cores arise from gravitational fragmentation of the densest filaments. Whether this scenario also applies to regions of high-mass star formation is an open question, in part because Herschel data cannot resolve the inner width of filaments in the nearest regions of massive star formation. We used the bolometer camera ArTeMiS on the APEX telescope to map the central part of the NGC6334 complex at a factor of > 3 higher resolution than Herschel at 350 microns. Combining ArTeMiS data with Herschel data allowed us to study the structure of the main filament of the complex with a resolution of 8" or < 0.07 pc at d ~ 1.7 kpc. Our study confirms that this filament is a very dense, massive linear structure with a line mass ranging from ~ 500 Msun/pc to ~ 2000 Msun/pc over nearly 10 pc. It also demonstrates that its inner width remains as narrow as W ~ 0.15 +- 0.05 pc all along the filament length, within a factor of < 2 of the characteristic 0.1 pc value found with Herschel for lower-mass filaments in the Gould Belt. While it is not completely clear whether the NGC 6334 filament will form massive stars or not in the future, it is two to three orders of magnitude denser than the majority of filaments observed in Gould Belt clouds, and yet has a very similar inner width. This points to a common physical mechanism for setting the filament width and suggests that some important structural properties of nearby clouds also hold in high-mass star forming regions.

ArTeMiS is a wide-field submillimeter camera operating at three wavelengths simultaneously (200, 350 and 450 microns). A preliminary version of the instrument equipped with the 350 microns focal plane, has been successfully installed and tested on APEX telescope in Chile during the 2013 and 2014 austral winters. This instrument is developed by CEA (Saclay and Grenoble, France), IAS (France) and University of Manchester (UK) in collaboration with ESO. We introduce the mechanical and optical design, as well as the cryogenics and electronics of the ArTeMiS camera. ArTeMiS detectors are similar to the ones developed for the Herschel PACS photometer but they are adapted to the high optical load encountered at APEX site. Ultimately, ArTeMiS will contain 4 sub-arrays at 200 microns and 2x8 sub-arrays at 350 and 450 microns. We show preliminary lab measurements like the responsivity of the instrument to hot and cold loads illumination and NEP calculation. Details on the on-sky commissioning runs made in 2013 and 2014 at APEX are shown. We used planets (Mars, Saturn, Uranus) to determine the flat-field and to get the flux calibration. A pointing model was established in the first days of the runs. The average relative pointing accuracy is 3 arcsec. The beam at 350 microns has been estimated to be 8.5 arcsec, which is in good agreement with the beam of the 12 m APEX dish. Several observing modes have been tested, like On-The-Fly for beam-maps or large maps, spirals or raster of spirals for compact sources. With this preliminary version of ArTeMiS, we concluded that the mapping speed is already more than 5 times better than the previous 350 microns instrument at APEX. The median NEFD at 350 microns is 600 mJy.s1/2, with best values at 300 mJy.s1/2. The complete instrument with 5760 pixels and optimized settings will be installed during the first half of 2015.