2007ApJ...668..492A

I have undertaken a comprehensive statistical investigation of the ultracool dwarf companion distribution (spectral type M6 and later). Utilizing a Bayesian algorithm, I tested models of the companion distribution against data from an extensive set of space and ground-based imaging observations of nearby ultracool dwarfs. My main conclusions are fivefold: (1) confirm that the concentration of high mass ratio ultracool binary systems is a fundamental feature of the companion distribution, not an observational or selection bias; (2) determine that the wide (≳20 AU) binary frequency can be no more than 1%-2%; (3) show that the decreasing binary frequency with later spectral types is a real trend; (4) demonstrate that a large population of currently undetected low mass ratio systems are not consistent with the current data; and (5) find that the population of spectroscopic binaries must be at least 30% that of currently known ultracool binaries. The best-fit value for the overall M6 and later binary frequency is ∼20%-22%, of which only ∼6% consists of currently undetected companions with separations less than 1 AU. If this is correct, then the upper limit of the ultracool binary population discovered to date is ∼75%. I find that the numerical simulation results of the ejection formation method are inconsistent with the outcome of this analysis. However, dynamics do seem to play an important role, as simulations of small-N clusters and triple system decays produce results similar to those of this work. The observational efforts required to improve these constraints are shown to be primarily large spectroscopic binary surveys and improved high-resolution imaging techniques.