5 Spatial distribution

In Paper I we detected evidence for mass segregation in the 30 Dor cluster. To further investigate this important point we have combined our spectral classification, with that available in the literature, to obtain a total of 235 stars with reliable spectral types within the 30 Dor cluster. This allows us to examine the spatial distribution of stars, grouping them in spectral class bands. The result is shown in Fig. 26 where we have split the stars in four main groups: O dwarfs, B dwarfs, OB giants and OB supergiants (including Wolf-Rayet stars).

  

Figure 26: Spatial distribution of stars classified in this paper. The different distribution of early and late type stars can be seen in this diagram. Coordinates are in arcsec, measured from R136. The dotted circles indicate radial distances from R136, at 50 arcsec steps

It was pointed out by [Underhill (1983)] that determination of evolutionary stages from spectral types is unreliable. Still there are some clear points that come out of the distribution:

1.

Giant and supergiant stars are evenly distributed in the sampled region (our sample is intrinsically stretched in the N-S direction).

2.

O dwarfs are notoriously concentrated towards R136.

3.

There is a large region, south-east of R136, where the number of late type stars is much larger than that of early type ones. These stars can be identified in the center of Fig. 2.

We have also estimated the stellar masses of stars from their spectral types interpolating from the relationship tabulated by [Schmidt-Kaler (1982)]. With these values we analysed the radial distribution of stars according to their masses. As the determination of a star's mass from it's spectral type is not very accurate, we split the dataset in two large groups: high mass and low mass stars, with the transition at $\sim \!23.5\, M_{\odot}$. Figure 27 shows the cumulative distribution of stars from the centre of the cluster. It can be readily seen that the massive stars show a concentrated distribution, while the lower mass stars have a flatter distribution, with few stars in the inner region. Thus, there is evidence for mass segregation in the 30Dor cluster, as it is also suggested from the photometric analysis of Paper I. A Kolmogorov-Smirnov test to both distributions, yields a significance level of 0.007, which indicates that the distribution of high mass stars is significantly different from the less massive ones.

  

Figure 27: Cumulative radial distribution of stars according to masses determined from their spectral types. Both distributions are plotted with dashed and continuous line styles, for the high-mass and low-mass stars respectively. The different distribution for both subsets indicates the presence of mass segregation

In previous works, [Malumuth & Heap (1994)] and [Brandl et al. (1996)] had also found mass segregation in the inner region of 30Dor, close to R136, which agrees with our findings. Malumuth & Heap detected a difference in the slope of the IMF of stars within a 3.3 arcsec radius of R136a and of those outside this area. Brandl et al. detected changes in the IMF slope and a clear trend towards smaller core radii for brighter -more massive- stars. Thus, there is evidence for mass segregation at the core and outer regions of 30 Dor. We believe that these results are also compatible with the lack of change of the IMF slope found by [Hunter et al. (1996)] given the narrow mass range and the restricted radial range of their data.

A more detailed analysis of this important point will be discussed in Paper III where we combine the masses derived from our photometric and spectroscopic observations.