An Adaptive Optics Survey for Close Protostellar Binaries

In order to test the hypothesis that Class I protostellar binary stars are a product of ejections during the dynamical decay of non-hierarchical 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 5,000 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.

💡 Research Summary

The paper investigates whether close (≤200 AU) binary companions among Class I protostars are linked to the presence of widely separated companions, testing the hypothesis that many close binaries arise from ejections during the dynamical decay of non‑hierarchical multiple systems. Using adaptive‑optics (AO) imaging from the Subaru natural‑guide‑star system and the Keck laser‑guide‑star system, the authors observed 47 embedded young stellar objects (YSOs) and combined these data with previously published AO observations, yielding a total sample of 71 targets. High‑resolution near‑infrared images (≈0.05″ resolution) allowed detection of companions down to ≈50 AU at distances ≤500 pc, minimizing incompleteness.

The analysis divided the environment into two scales: a “near‑wide” regime (companions within 5 000 AU) and a “far‑wide” regime (companions within 25 000 AU). Statistical tests showed no significant correlation between the presence of a near‑wide companion and a close companion (χ² p≈0.34). In contrast, the far‑wide regime displayed a strong association: protostars with a close companion were three times more likely to have another YSO within 25 000 AU, and Fisher’s exact test gave p < 0.01. Remarkably, every close binary in the sample (27 systems) had at least one additional YSO within the 25 000 AU projected separation, with only one exception in the full dataset.

To control for observational bias, the authors restricted the sample to objects within 500 pc and to close companions wider than 50 AU, where AO completeness is high. The same correlation persisted, confirming that the result is not an artifact of resolution or sensitivity limits.

Regional differences were also examined. The Orion star‑forming complex exhibited a pronounced excess of both close binaries and wide YSOs compared with other regions (e.g., Taurus, Perseus). This suggests that high‑density environments promote the formation of non‑hierarchical multiples, which subsequently undergo dynamical decay and ejection.

The authors interpret these findings as strong evidence that many close Class I binaries are the remnants of dynamical interactions in unstable multiple systems. In this picture, an initially chaotic group of three or more protostars experiences close encounters; one member is ejected, while the remaining two settle into a tight orbit (≤200 AU). The ejected component may remain bound on a very wide orbit (up to tens of thousands of AU) for some time, but many become unbound during the Class I phase, explaining the observed scarcity of very wide companions at later evolutionary stages.

Overall, the study supports a hybrid formation scenario: while core fragmentation and disk instability undoubtedly produce some binaries, a substantial fraction of close Class I binaries likely originate from dynamical ejection processes. The work highlights the importance of considering both small‑scale (≤200 AU) and large‑scale (≤25 000 AU) companion statistics, and it underscores the role of the natal environment—particularly high‑density clusters like Orion—in shaping the binary population of young stars.