The first systematic work on binary clusters in the Magellanic Clouds started approximately a decade ago. The first catalogue of binary star clusters in the LMC was presented by Bhatia & Hatzidimitriou (1988) and Bhatia et al. (1991) who surveyed the cluster system (consisting of 1200 objects known at that time) and listed 69 binary cluster candidates. Their selection criterion was a maximum separation between the components of a proposed pair of approximately 18 pc (assuming a distance modulus of 18.4 mag). Following Page (1975), out of these 69 double objects only 31 can be explained as optical pairs, i.e., clusters that appear as close pairs on the plane of the sky due to projection effects. Ages were available only for some of the clusters and suggested that the pairs are young (between 107 to a few 108 yr), consistent with expected time scales for merger or disruption of binary clusters (Bhatia 1990).
In the following years, more studies on binary cluster candidates were performed but concentrated mainly on one or a few individual objects in order to establish their binarity (Kontizas et al. 1989; Lee 1992; Bhatia 1992; Kontizas et al. 1993; Vallenari et al. 1994; Hilker et al. 1995; Grebel 1997; Vallenari et al. 1998; Leon et al. 1999; Dieball & Grebel 1998,2000; Dieball et al. 2000). Few theoretical studies concerning formation, gravitational interaction and dynamical evolution of binary clusters are available (Sugimoto & Makino 1989; Bhatia 1990; Fujimoto & Kumai 1997; de Oliveira et al. 1998; Theis 1998).
However, since the investigation of Bhatia & Hatzidimitriou (1988) and Bhatia et al. (1991) many more clusters have been discovered in the LMC. Thus, it is time to perform a new study on the nowadays better known LMC cluster system, aiming at the question of how many close cluster pairs exist and how many of these might be explained as chance superpositions.
Recently, Pietrzynski & Udalski (2000b) provided a new but spatially limited catalogue of multiple cluster candidates in the LMC. They based their studies on the OGLE (see Udalski et al. 1992) dataset which covers 5.8 square degrees of the inner part of the LMC and contains 745 star clusters (Pietrzynski et al. 1999). Out of these, a total of 100 multiple cluster candidates with a maximum separation of 18 pc, assuming a distance modulus of 18.24 mag, were selected. The cluster groups consisted of 73 pairs, 18 triple systems, 5 systems containing four components, 1 with five and 3 systems with six clusters. Assuming that all 745 clusters are distributed uniformly in the 5.8 square degree region and adopting the same statistical approach as Bhatia & Hatzidimitriou (1988), 51 chance pairs can be expected. A more detailed investigation of the cluster distribution led to nearly the same result of 53 random pairs. The number of all detected candidates is 153 and thus significantly larger than expected from chance superposition. Ages for the components were taken from Pietrzynski & Udalski (2000a). 102 components are coeval, 53 have very different ages, and most objects are younger than 300 Myr with a peak at 100 Myr. This suggests that most of the multiple clusters have a common origin and are quite young objects.
A catalogue of multiple cluster candidates in the SMC was published by
Pietrzynski & Udalski (1999), containing 23 binary and 4
triple cluster candidates. A comparison of both the LMC and SMC binary
cluster lists reveals that the distribution of the components'
separation, the fraction of cluster groups (
12%) and their
ages are very similar. The similar ages of the binary cluster
candidates in both the LMC and SMC might be connected with the last
close encounter between these two galaxies. De Oliveira et al.
(2000b) presented an isophotal atlas of 75 binary and multiple
clusters (comprising 176 objects) from the Bica &
Dutra (2000) catalogue of SMC clusters. Bica & Dutra (2000)
included also new discoveries from the OGLE catalogue of SMC clusters
(Pietrzynski et al. 1998).
Investigating the isophotes of the binary and multiple cluster
candidates, de Oliveira et al. (2000b) found isophotal distortions,
connecting bridges, or common isophotal envelopes for 25% of the suggested
multiple clusters. The authors interpreted this as signs of
interaction between the components of a supposed binary or multiple
cluster, in agreement with the findings from previous N-body
simulations (de Oliveira et al. 2000a).
Ages for 91 out of the 176 clusters that are part of pairs or
groups were investigated based on the OGLE BVI maps. 40 clusters are
in common with Pietrzynski & Udalski (1999), and
de Oliveira et al. (2000b) found good agreement with the
study of the OGLE group. Most clusters are young, and the age
distribution shows a relevant peak around 200 Myr that can be
attributed to the last close encounter between SMC and LMC.
The components of groups with more than two members are younger than
100 Myr, which might be an indication that multiple clusters coalesce
into binary or single clusters within this timescale. 55% of the binary and
multiple cluster candidates were found to be coeval. From this the authors
concluded that tidal capture is a rare phenomenon.
In this paper, we present a statistical study of close pairs and
multiple clusters in the LMC. We decided to base our analysis on the
new, extended catalogue of stellar clusters, associations, and
emission nebulae in the LMC provided by Bica et al. (1999,
hereafter BSDO). The authors surveyed the ESO/SERC R and J Sky
Survey Schmidt films, checked the entries of previous catalogues and
searched for new objects. The resolution of the measurements was
(Bica & Schmitt 1995). The resulting new catalogue
unifies previous surveys and contains 6659 entries, out of which
3246 are new discoveries that are not mentioned in previous catalogues and
lists. Thus, the BSDO catalogue can be considered as the
so far most complete catalogue of LMC stellar clusters and associations.
We restricted our study to bound stellar systems, which
means that we selected only objects which are categorized as "C''-type
(cluster-type), and left out associations and emission nebulae, which are not
of interest in the context of the present study. This reduces the number of
objects found in the BSDO catalogue from a total of 6659
to 4089.
Based on this catalogue, we performed a statistical study of cluster pairs and groups and provide a complete list of all multiple cluster candidates in the LMC.
In the following sections, we address a number of questions: How many cluster pairs can be found with a projected separation of less than 20 pc between the components of a pair (Sect. 4)? Following Bhatia & Hatzidimitriou (1988) and Bhatia et al. (1991), we consider this to be a good selection criterion. Several cluster pairs may form a larger cluster group, e.g., if a component of a cluster pair is less than 20 pc distant from any component of another pair. In this way, three clusters may form a triple cluster, but they also might constitute three cluster pairs if each cluster is seen within 20 pc from each other cluster. How many "multiple'' clusters, consisting out of more than two single objects, are present, and how many single clusters are involved in these pairs and groups (Sect. 4)? How many of these pairs and multiple systems can be expected statistically, and of how many individual components do they consist (Sect. 5)? Are there any correlations between the properties of the cluster systems such as ages, radii and separations between the components (Sect. 6)? What is the fraction of coeval pairs or groups compared with the number of multiple clusters whose internal age differences exceed the protocluster survival time (Sect. 6.4)? Does the percentage of coeval systems agree with the number of statistically expected groups (Sect. 6.4)? And finally, do our results favour or give hints at a specific cluster formation scenario (Sect. 7)? For instance, can cluster pairs be explained with statistically expected cluster encounters in the LMC, which lead to tidal capture and thus to bound pairs of different ages (see Sect. 3)? Or are the multiple cluster candidates predominantly found to be coeval, favouring the formation scenario of Fujimoto & Kumai (1997) or of Theis (1998)?
Copyright ESO 2002