2009AJ....138.1193C

In order to test the hypothesis that Class I protostellar binary stars are a product of ejections during the dynamical decay of nonhierarchical multiple systems, we combined the results of new adaptive optics (AO) observations of Class I protostars with our previously published AO data to investigate whether Class I protostars with a widely separated companion (r > 200 AU) are more likely to also have a close companion (r < 200 AU). In total, we observed 47 embedded young stellar objects (YSOs) with either the Subaru natural guide star AO system or the Keck laser guide star AO system. We found that targets with a widely separated companion within 5000 AU are not more likely to have a close companion. However, targets with another YSO within a projected separation of 25,000 AU are much more likely to have a close companion. Most importantly, every target with a close companion has another YSO within a projected separation of 25,000 AU. We came to the same conclusions after considering a restricted sample of targets within 500 pc and close companions wider than 50 AU to minimize incompleteness effects. The Orion star-forming region was found to have an excess of both close binaries and YSOs within 25,000 AU compared to other star-forming regions. We interpret these observations as strong evidence that many close Class I binary stars form via ejections and that many of the ejected stars become unbound during the Class I phase.