On the origin of brown dwarfs and free-floating planetary-mass objects

Briceno et al. report a significantly smaller number of brown dwarfs (BDs) per star in the Taurus-Auriga (TA) pre-main-sequence stellar groups than in the central region of the Orion Nebula cluster (ONC). Also, BDs have binary properties that are not compatible with a star-like formation history. It is shown here that these results can be understood if BDs are produced as ejected embryos with a dispersion of ejection velocities of about 2 km s(-1), and if the number of ejected embryos is about one per four stars born in TA and the ONC. The Briceno et al. observation is thus compatible with a universal BD production mechanism and a universal initial mass function (IMF), but the required number of BDs per star is much too small to account for the one BD per star deduced to be present in the Galactic field. There are two other mechanisms for producing BDs and free-floating planetary-mass objects (FFLOPs), namely the removal of accretion envelopes from low-mass protostars via photo-evaporation through nearby massive stars, and hyperbolic collisions between protostars in dense clusters. The third BD flavour, the collisional BDs, can be neglected in the ONC. It is shown that the observed IMF with a flattening near 0.5 M-. can be reproduced via photo-evaporation of protostars if these are distributed according to a featureless Salpeter mass function above the substellar mass limit, and that the photo-evaporated BDs should have a smaller velocity dispersion than the stars. The number of photo-evaporated BDs per star should increase with cluster mass, peaking in globular clusters that would have contained many stars as massive as 150 M. The required number of embryo-ejected BDs in TA and the ONC can be as low as six ejected BDs per 100 stars if the central ONC contains 0.23 photo-evaporated BDs per star. Alternatively, if the assumption is discarded that embryo ejection must operate equally in all environments, then it can be argued that TA produced about one ejected BD per star, leading to consistency with the Galactic field observations. The dispersion of ejection velocities would be about 3 km s(-1). In the central ONC the number of ejected BDs per star would then be at most 0.37, or less if photo-evaporated BDs contribute. This non-universal scenario would thus imply that the Galactic field BD population may mostly stem from TA-like star formation or modest clusters, the ONC not being able to contribute more than about 0.25 +/- 0.04 BDs per star.