We present a new observational method to evaluate the star formation law as formulated by Schmidt: the power-law expression assumed to relate the rate of star formation in a volume of space to the local total gas volume density. Volume densities in the clouds surrounding an OB association are determined with a simple model which considers atomic hydrogen as a photodissociation product on cloud surfaces. The photodissociating flux incident on the cloud is computed from the far-UV luminosity of the OB association and the geometry. We have applied this "PDR Method" to a sample of star-forming regions in M33 using VLA 21-cm data for the HI and GALEX imagery in the far-UV. It provides an estimate of the total volume density of hydrogen (atomic + molecular) in the gas clouds surrounding the young star cluster. A logarithmic graph of the cluster UV luminosity versus the surrounding gas density is a direct measure of the star formation law. However, this plot is severely affected by observational selection, rendering large areas of the diagram inaccessible to the data. An ordinary least-squares regression fit therefore gives a strongly biased result. Its slope primarily reflects the boundary defined when the 21-cm line becomes optically thick, no longer reliably measuring the HI column density. We use a maximum-likelihood statistical approach which can deal with truncated and skewed data, taking into account the large uncertainties in the derived total gas densities. The exponent we obtain for the Schmidt law in M33 is 1.4 \pm 0.2.
Read full abstract