Astron. Astrophys. 347, 532-549 (1999)

4. Conclusions

1. A new photometric method was designed that allows the determination of masses for even the most massive stars from UBV photometry alone without introducing biases towards smaller masses. As a check of the method we obtained a distance modulus of (m-M)₀ = for 30 Doradus in excellent agreement with the best estimate for distance to the LMC, (m-M)₀ = 18.50.

2. For we find the reddening distribution to be independent of stellar mass. The data are not complete to allow an accurate determination of this distribution below . We find a different reddening distribution for the stars near the center of the cluster (larger reddening close to the center) confirming the conclusions from Paper I. If not considered, this differential reddening distribution could introduce serious systematic effects in the IMF.

3. We have uncovered a systematic effect due to variable reddening which is responsible for the flattening of the IMF close to the limit of the photometry and we have designed and applied a method to correct for this effect which we call the magnitude-limit correction.

4. We use our data to constrain the ages of stars with , and find that the star-formation history is dominated by three bursts of increasing strength occurring 5My, 2.5My, and 1.5My ago. The number of high mass stars produced in each burst are approximately in proportion to the intensity of the burst, supporting Elmegreen's (1997) stochastic sampling interpretation of the IMF (with the exception of the somewhat flatter core and the 6 pc shell discussed below). The three bursts appear to be spatially disjoint with the youngest stars concentrated towards the center while the 2.5My stars appear to delineate a spherically symmetric structure of 6 pc radius slightly off-center. The observations are consistent with a picture where star formation propagated inwards in the cluster. The shell at 6 pc indicates a very complex structure for the ISM before the starburst.

5. For " we find the IMF to be well represented by a power law with for consistent with the Salpeter slope. For and we find . Combining our data with published HST observations we find and in this region, somewhat flatter than Salpeter. We reanalyze the Massey & Hunter (1998a) data and conclude, at variance with those authors, that the large number of O3 stars they observed in the core is not consistent with a Salpeter slope, and it is most likely due to a flattening of the IMF. The concentration of the most massive stars towards the center, irrespective of their ages, indicates that dynamical effects probably play an important role in shaping the spatial variations of the IMF.

6. The spatial density profile between 0.4 pc and 6 pc is well represented by a single power law with slope -2.85, steeper than isothermal. The density profile also shows an excess of stars within the 6 pc shell. Moreover, this shell appears to have a larger than normal proportion of high mass stars, providing evidence that Elmegreen's (1997) simple stochastic picture of star-formation is not the only mode at work in this complex region. The steeper overall profile of the cluster, together with the features in the density profile in the outer radial bins, indicates that the cluster is not relaxed, and that it might still have memory of its collapse.

7. We find that, if current evolutionary models are correct, there is evidence that the upper mass cut-off of the IMF is larger than 120. It appear that the most massive stars in 30 Doradus have , or that we are seeing the effects of rotation which for a given mass makes stars evolve at higher luminosities. Alternatively, the most massive stars in the cluster may all be binaries.
   

© European Southern Observatory (ESO) 1999

Online publication: June 30, 1999
helpdesk.link@springer.de