On the universal outcome of star formation: is there a link between stars and brown dwarfs?

Given the current consensus that stars form from pre-stellar cloud cores that fragment into small-N groups which decay within a few x 10(4) yr, and taking the observed properties of approximately 1-Myr-old stars in the Taurus-Auriga (TA) star-forming region as empirical constraints, we suggest a model that describes the multiplicity properties of the disintegrated groups. This model concisely describes the outcome of star formation in terms of dynamically unevolved binary properties. Two variants of the model are tested against data on very young stars in TA and the Orion Nebula cluster (ONC), as well as the older Pleiades and the Galactic field populations. The 'standard model' (SM) assumes that cloud-core fragmentation only produces stellar systems, while the 'standard model with brown dwarfs' (SMwBDs) assumes that cloud-core fragmentation proceeds down to substellar-mass cores. Brown dwarfs (BDs) enter the SM by being a separate, dynamically unimportant population. The models produce a very high initial binary proportion among stars (SM), and stars and BDs (SMwBDs), and both reproduce the measured initial mass function (IMF) in TA, the ONC and the Pleiades as well as the Galactic field. Concentrating on the SMwBDs, it is shown that the Briceno et al. result that TA appears to have produced significantly fewer BDs per star than the ONC is reproduced almost exactly without calling for a different IMF. The reason is that star-BD and BD-BD binaries are disrupted in the dense ONC. The model, however, fails to reproduce the observed star-star binary period distribution in TA, because it contains too many star-BD pairs. Also, the SMwBDs leads to too many wide star-BD and BD-BD systems. This is a problem if most stars form in clusters because Galactic field very low-mass star and BD binaries have a low binary fraction and do not contain wide systems. The SM, on the other hand, finds excellent agreement with the observed mass ratio and period distribution among TA and Galactic field stellar binaries, as well as the observed stellar period distribution in the ONC and the Pleiades. The conclusion of this work is therefore that the SM describes the initial, dynamically unevolved stellar population very well indeed for a large range of star-forming conditions, suggesting (1) a remarkable invariance of the star formation products, and (2) that BDs (and some very low-mass stars) need to be added as a separate population with its own kinematical and binary properties. This separate population may vary with star-forming conditions.