A Survey for Spectroscopic Binaries among Very Low Mass Stars

We report on the results of a survey for radial velocity variability in a heterogeneous sample of very low-mass stars and brown dwarfs. One distinguishing characteristic of the survey is its timespan, which allows an overlap between spectroscopic binaries and those which can be found by high angular-resolution imaging. We are able to place a new constraint on the total binary fraction in these objects, which suggests that they are more likely the result of extending the same processes at work at higher masses into this mass range, rather than a distinct mode of formation. Our basic result is that there are $6 \pm 2$ out of 53, or $11^{+0.07}_{-0.04}$% spectroscopic binaries in the separation range 0-6 AU, nearly as many as resolved binaries. This leads to an estimate of an upper limit of $26 \pm 10$% for the binary fraction of VLM objects (it is an upper limit because of the possible overlap between the spectroscopic and resolved populations). A reasonable estimate for the very low-mass binary fraction is $20 - 25$%. We consider several possible separation and frequency distributions, including the same one as found for GK stars, a compressed version of that, a version of the compressed distribution truncated at 15 AU, and a theoretical distribution which considers the evaporation of small-N clusters. We conclude that the latter two bracket the observations, which may mean that these systems form with intrinsically smaller separations due to their smaller mass, and then are truncated due to their smaller binding energy. We do not find support for the ``ejection hypothesis'' as their dominant mode of formation, particularly in view of the similarity in the total binary fraction compared with slightly more massive stars, and the difficulty this mechanism has in producing numerous binary systems.