Massive binary stars as a probe of massive star formation

Massive stars are among the largest and most influential objects we know of on a sub-galactic scale. Binary systems, composed of at least one of these stars, may be responsible for several types of phenomena, including type Ib/c supernovae, short and long gamma ray bursts, high-velocity runaway O and B-type stars, and the density of the parent star clusters. Our understanding of these stars has met with limited success, especially in the area of their formation. Current formation theories rely on the accumulated statistics of massive binary systems that are limited because of their sample size or the inhomogeneous environments from which the statistics are collected. The purpose of this work is to provide a higher-level analysis of close massive binary characteristics using the radial velocity information of 113 massive stars (B3 and earlier) and binary orbital properties for the 19 known close massive binaries in the Cygnus OB2 Association. This work provides an analysis using the largest amount of massive star and binary information ever compiled for an O-star rich cluster like Cygnus OB2, and compliments other O-star binary studies such as NGC 6231, NGC 2244, and NGC 6611. I first report the discovery of 73 new O or B-type stars and 13 new massive binaries by this survey. This work involved the use of 75 successful nights of spectroscopic observation at the Wyoming Infrared Observatory in addition to observations obtained using the Hydra multi-object spectrograph at WIYN, the HIRES echelle spectrograph at KECK, and the Hamilton spectrograph at LICK. I use these data to estimate the spectrophotometric distance to the cluster and to measure the mean systemic velocity and the one-sided velocity dispersion of the cluster. Finally, I compare these data to a series of Monte Carlo models, the results of which indicate that the binary fraction of the cluster is 57 +/- 5% and that the indices for the power law distributions, describing the log of the periods, mass-ratios, and eccentricities, are --0.2 +/- 0.3, 0.3 +/- 0.3, and --0.8 +/- 0.3 respectively (or not consistent with a simple power law distribution). The observed distributions indicate a preference for short period systems with nearly circular orbits and companions that are not likely drawn from a standard initial mass function, as would be expected from random pairing. An interesting and unexpected result is that the period distribution is inconsistent with a standard power-law slope stemming mainly from an excess of periods between 3 and 5 days and an absence of periods between 7 and 14 days. One possible explanation of this phenomenon is that the binary systems with periods from 7--14 days are migrating to periods of 3--5 days. In addition, the binary distribution here is not consistent with previous suggestions in the literature that 45% of OB binaries are members of twin systems (mass ratio near 1).