ATLASGAL - properties of a complete sample of Galactic clumps

The APEX Telescope Large Area Survey of the Galaxy (ATLASGAL) is an unbiased 870 mu m submillimetre survey of the inner Galactic plane (vertical bar l vertical bar < 60 degrees with vertical bar b vertical bar < 1 degrees.5). It is the largest and most sensitive ground-based submillimetre wavelength Galactic survey to date and has provided a large and systematic inventory of all massive, dense clumps in the Galaxy (>= 1000 M-circle dot at a heliocentric distance of 20 kpc) and includes representative samples of all of the earliest embedded stages of high-mass star formation. Here, we present the first detailed census of the properties (velocities, distances, luminosities and masses) and spatial distribution of a complete sample of similar to 8000 dense clumps located in the Galactic disc (5 degrees < vertical bar l vertical bar < 60 degrees). We derive highly reliable velocities and distances to similar to 97 per cent of the sample and use midand far-infrared survey data to develop an evolutionary classification scheme that we apply to the whole sample. Comparing the evolutionary subsamples reveals trends for increasing dust temperatures, luminosities and linewidths as a function of evolution indicating that the feedback from the embedded protoclusters is having a significant impact on the structure and dynamics of their natal clumps. We find that the vast majority of the detected clumps are capable of forming a massive star and 88 per cent are already associated with star formation at some level. We find the clump mass to be independent of evolution suggesting that the clumps form with the majority of their mass in situ. We estimate the statistical lifetime of the quiescent stage to be similar to 5 x 10(4) yr for clump masses similar to 1000 M-circle dot decreasing to similar to 1 x 10(4) yr for clump masses > 10000 M-circle dot. We find a strong correlation between the fraction of clumps associated with massive stars and peak column density. The fraction is initially small at low column densities, but reaching 100 per cent for column densities above 10(23) cm(-2); there are no clumps with column densities above this value that are not already associated with massive star formation. All of the evidence is consistent with a dynamic view of star formation wherein the clumps form rapidly and are initially very unstable so that star formation quickly ensues.