On the Distribution of Orbital Eccentricities for Very Low-Mass Binaries

We have compiled a sample of 16 orbits for very low-mass stellar (<0.1 Msun) and brown dwarf binaries, enabling the first comprehensive study of the eccentricity distribution for such objects. We find that very low-mass binaries span a broad range of eccentricities (0.03<e<0.83), with a median eccentricity of 0.34. We examine potential observational biases in this sample, and for visual binaries we show through Monte Carlo simulations that appropriate selection criteria result in all eccentricities being equally represented (<5% difference between input and output e distributions). The orbits of this sample of very low-mass binaries show some significant differences from their solar-type counterparts. They lack a correlation between orbital period and eccentricity and display a much higher fraction of near-circular orbits (e<0.1) than solar-type stars, which together suggest a different formation mechanism or dynamical history for these two populations. Very low-mass binaries also do not follow the e^2 distribution of Ambartsumian (1937), which would be expected if their orbits were distributed in phase space according to a function of energy alone (e.g., the Boltzmann distribution). We find that numerical simulations of very low-mass star formation do not completely reproduce the observed properties of our binary sample. The cluster formation model of Bate (2009) agrees very well with the overall e distribution, but lack any high-e (>0.6) binaries at orbital periods comparable to our sample. In contrast, the circumstellar disk fragmentation model of Stamatellos & Whitworth (2009) predicts only high-e binaries and thus is highly inconsistent with our sample. These discrepancies could be explained if multiple formation processes have produced the field population.