Astron. Astrophys. 361, 877-887 (2000)

6. Discussion and concluding remarks

In this work we have obtained sub-arcsecond angular resolutions (0".13 and 0".26 FWHM) comparable to those of the adaptive optics owing to the MCS deconvolution algorithm. This technique has allowed us to push the resolution of the Sk-66^o41 cluster further into 15 components, while at the same time yielding accurate photometry of the components. It has also enabled us to present a first study of HNT, the other tight star cluster lying towards the core of the LH 13 association. With its 70 components, HNT is richer than Sk-66^o41, but its stars are fainter and less massive, the brightest components being A-F types.

Fig. 8. Rectified spectrum of Wo 647. We classify this star as an F7 - F8 foreground star.

It is interesting to note that the ROSAT -HRI X-ray observations of Mac Low et al. (1998) revealed a point-like source in N 11C. Although the number of HRI counts is quite low and Mac Low et al. (1998) caution that they cannot confirm the point nature of this source, its position is in very good agreement with the optical position of the Sk-66^o41 cluster. Interestingly the same HRI observations revealed no X-ray emission associated with the famous tight cluster HD 32228 at the core of LH 9 (south of N 11B) which contains at least 16 early-type stars with the brightest components being of spectral type O9 Ib and O8.5 II(f) (Walborn et al. 1999, see also Parker et al. 1992). These results suggest that the X-ray emission seen in N 11C is most probably due to the interaction of the stellar winds of the components of the Sk-66^o41 cluster with the relatively dense ambient interstellar medium, whereas the lack of X-ray emission from HD 32228 is due to the lack of a sufficiently dense interstellar medium in the LH 9 region. Further constraints on the nature of this X-ray source will have to await the XMM observations of the N 11 complex.

The Sk-66^o41 cluster harbors a very hot star of spectral type O3 and therefore provides the main exciting source of the N 11C H II region contrarily to the finding of Paper I. The ionized gas streaming from N 11C has a radial velocity of 288 km s^-1 (Rosado et al. 1996). We have measured mean nebular line radial velocities of 296.0 and 293.6 km s^-1 in the spectra around the HNT cluster and Wo599. Given our spectral resolution, these are in very good agreement with Rosado et al.'s (1996) results.

An inevitable question is whether the two compact clusters, one high mass the other low mass, belong to the same star formation region. This question is crucial for better understanding star formation in the LMC OB association LH13. The radial velocity derived for the HNT cluster is 172 15 km s^-1 (rms), based on the five measurements listed in Table 3. Putting aside star #91, which has a somewhat discrepant velocity, yields 177 11 (rms) km s^-1. On the other hand, the global radial velocity of the Sk-66^o41 cluster is 326 13 km s^-1 (Table 4). Although it cannot be ruled out that the multiplicity and the internal motion of the stars within the Sk-66^o41 cluster could alter the measured radial velocities, the velocity difference between the two clusters is probably due to their non-association. This is in line with the drastic age difference between the two clusters (see Sect. 5).

Apart from Sk-66^o41, the brightest stars towards LH 13 are Wo597, Wo599, Wo600, Wo622, and Wo647. There are two late types among them, Wo600 and Wo647, which are both Galactic F types (Sect. 4.3 and Paper I). The remaining three bright stars are O types most probably associated with LH 13 and the H II region N 11C, although we notice that the measured radial velocities of these stars are slightly less positive than the radial velocity of the N 11C nebular lines (288 km s^-1, Rosado et al. 1996) and that of the Sk-66^o41 cluster. We conclude that these three stars belong to the LH 13 association as deduced from the C-M diagram. Moreover, the probability that three early type stars not belonging to LH 13 lie by chance towards this OB association should be very low. These stars have probably formed along with Sk-66^o41 during the same burst.

Models studying formation of massive stars predict that these stars should never form in isolation, and that those found in isolation have been ejected from dense stellar clusters (Bonnell et al. 1998). Wo599 has a projected distance of 15" (3.8 pc) from Sk-66^o41. Let us consider that Sk-66^o41 represents the core of the massive stars resulting from the same starburst and assume that Wo599 is escaping from its birthplace. Escape velocities can be larger than 200 km s^-1 (Leonard & Duncan 1990; Kroupa 1995), and if we arbitrarily take 50 km s^-1, Wo599 needs a travel time of 75 000 years to reach its present position. Wo597, would need a comparable timespan for its journey. The most distant candidate, Wo622, has a projected distance of 50" (12.5 pc), and requires a longer travel time of 250 000 years. These travel-time estimates are of course lower bounds, since we deal with the image on the sky of a three-dimensional configuration in space. Inversely, one can calculate the minimum velocities at which the stars could have reached their current locations. Assuming a lifetime of 3 Myr, one gets a lower velocity of 4 km s^-1 for Wo622. One might wonder whether the observed star density of the Sk-66^o41 cluster is sufficiently high for the dynamical ejection mechanism to work. However, Leonard & Duncan (1988) have shown that binary-binary collisions required to produce high velocity escapees occur in low density clusters, even though simple estimates suggest that such interactions are unlikely. Furthermore, the ejection of the OB stars that we observe around Sk-66^o41 must have happened during an earlier evolutionary stage when the cluster was most probably more compact than today (Portegies Zwart et al. 1999).

In summary, our high resolution images reveal two tight clusters with significantly different ages within the core of the LH 13 association. The physical connection between these two clusters is presently not clear. The younger one, Sk-66^o41 (age Myr) is most likely the core of a compact starburst event. The surrounding OB-stars might have been ejected from the Sk-66^o41 cluster, probably as a result of dynamical interaction. The older cluster, HNT (age 100 25 Myr), contains no stars earlier than spectral types A-F and its kinematical properties as well as its color-magnitude diagram suggest that this cluster has no direct connection to the Sk-66^o41 starburst and could rather be a line of sight object.

© European Southern Observatory (ESO) 2000

Online publication: January 29, 2001
helpdesk.link@springer.de