Massive binary stars and the kinematics of Young Massive Clusters
Located in the Large Magellanic Cloud, R136 is a rare example of a nearby young and dense massive star cluster in which individual stars can be resolved. Often suggested as a globular cluster in formation, its study is of great interest and promises to provide insights into the early dynamical evolution of massive star clusters. This is crucial to understand more extreme and distant starburst clusters, which contribute to a significant fraction of all current star formation in the Local Universe, in particular in interacting galaxies. The majority of this thesis is based on multi-epoch spectroscopic observations in and around R136 obtained as part of the VLT-FLAMES Tarantula Survey (VFTS), an ambitious programme which targeted nearly 1 000 massive stars in the intricate 30 Doradus star-forming region. The motivations and observing strategy of this survey, designed to address key questions about the evolution of massive stars and clusters, are first introduced. The data reduction procedures applied to VFTS data are described, with an emphasis on the tasks accomplished in the context of this thesis. The VFTS data are first used to perform a detailed kinematic study of R136, determine its dynamical state, and evaluate the importance of gas expulsion in the early evolution of massive star clusters. Orbital motions of binary stars are found to dominate the line- of-sight velocity dispersion of the cluster, illustrating the risk of interpreting velocity dispersion measurements for unresolved extragalactic young massive clusters. However, once the detected binaries are rejected and the contribution of undetected binaries is accounted for through Monte Carlo simulations, the true velocity dispersion of the cluster is found to be low and consistent with it being in virial equilibrium. This suggests that gas expulsion has not had a dramatic effect on the early dynamical evolution of R136. Using the velocity measurements of R136 as a test case, a maximum likelihood method that fits the velocity dispersion of a cluster from a single epoch of radial velocity data is then tested. The method must be applied with care given the high binary fraction of massive stars and the large uncertainties in their binary orbital parameter distributions, but for typical velocity dispersions of young massive clusters (& 4 kms−1), it is shown that the velocity dispersion can be measured with an accuracy of 40% or better. This offers an efficient way of constraining the dynamics of these systems. The radial velocity measurements of apparently single stars in R136 are also used to investigate the internal rotation of the cluster, a potentially important but largely unexplored characteristic of young clusters. Evidence is found, at the 95% confidence level, for rotation of the cluster as a whole. A simple maximum likelihood method is presented to fit rotation curves to the data, from which a typical rotational velocity of 3 kms−1 is found. When compared to the low velocity dispersion of R136, this suggests that star clusters may form with as much as 20% of their kinetic energy in rotation. Finally, a smaller-scale survey of massive stars in the Wing of the Small Magellanic Cloud is introduced. As an example of the particularly interesting massive binaries that can be revealed by the synergy between large optical spectroscopic surveys of young clusters and observations at other wavelengths, the discovery of a new Be/X-ray pulsar binary and associated supernova remnant is reported. With a long spin period of over 1 000 seconds and a young age of 104 years constrained by its association with the supernova remnant, the pulsar in this system is quickly emerging as a unique object that challenges our understanding of the spin evolution of accreting neutron stars.