1. Introduction

The least massive star has 75 times the mass of Jupiter. What about objects of intermediate mass? What are their properties and how do they compare with those of stars and planets? How many are there of these objects are there? These questions take us into the realm of the newly discovered “brown dwarfs”. Although theories discussing such objects go back to Kumar 1963, the quest for an observation of an incontrovertible brown dwarf was frustrated by a series of proposed candidates over a 20 year period, which each failed further confirmation. There were several unrelated breakthroughs in 1995, followed rapidly by detection of many further convincing cases. By now the number of truly confirmed brown dwarfs has passed 20, with over 100 very likely detections. There have been several recent conferences and workshops whose proceedings contain valuable reviews on this and related topics. Of particular note is “Brown Dwarfs and Extrasolar Planets” (Rebolo et al 1998a) and “From Giant Planets to Cool Stars” (Griffith & Marley 2000). Other reviews that are useful to consult are Allard et al (1997), Hodgkin and Jameson (1997), Kulkarni (1997), and Oppenheimer et al (2000).

 

1.1 What is a Brown Dwarf?

Before we follow the story of discovery, let us sharpen the definitions of “star”, “brown dwarf” (BD), and “planet”. The defining characteristic of a star is that it will stabilize its luminosity for a period of time by hydrogen burning. A star derives 100% of its luminosity from fusion during the main sequence phase, while the highest mass BD always has gravitational contraction as at least a small part of its luminosity source. The BD is brightest when it is born, and continually dims and cools (at the surface) after that. There can be some hydrogen fusion in the higher mass BDs, and all objects down to about 13 Jupiter masses (jupiters) will at least fuse deuterium (Saumon et al 1996). The lower mass limit of the main sequence lies at about 0.072 times the mass of the Sun (or 75 jupiters) for an object with solar composition. The limit is larger for objects with lower metallicity, reaching about 90 jupiters for zero metallicity (Saumon et al 1994). I refer you for details to the article in this volume that describes the theory of the structure and evolution of these objects, by G. Chabrier.

Amazingly, astronomers are currently somewhat undecided just how to define “planet”. At the low mass end of planets, an example of the difficulty is provided by the recent controversy over Pluto. At the high mass end of planets, we are now aware of extrasolar “giant planets” (Marcy & Butler 1998) ranging up to more than 10 jupiters. At what point in mass should these be more properly called brown dwarfs? The traditional line of thinking holds that brown dwarfs form like stars - though direct collapse of an interstellar cloud into a self-luminous object. As this object forms, the material with higher angular momentum will settle into a disk of gas and dust around it. The dust in the disk can coagulate into planetesimals (kilometers in radius), and these can crash together to eventually form rock/ice cores. When a core reaches 10-15 earth masses, and if the gas disk is still present, it can begin to rapidly attract the gas and build up to a gas giant planet. Because of the nature of this process, one naively expects the planet to be in an almost circular orbit. The layout of our solar system also suggests that a massive enough core can only be produced if icy planetesimals are widely available, which occurs at about the distance of Jupiter (the “ice boundary”).

This traditional picture (based on our own solar system) has been seriously challenged by the discovery of the extrasolar planetary systems. All of them that are not tidally circularized by being too close to the star, have eccentric orbits. They are all inside the ice boundary (though this is largely an observational selection effect). Some are very close to the star (where formation of a giant planet seems nearly impossible). These facts led Black (1997) to claim that most of the extrasolar planets found so far are really BDs (since objects found by Doppler searches only have lower limits on their masses). Such a claim is unsupportable because of the statistics of these lower limits (Marcy & Butler 1998); if they reflected a population of BDs then many others would show up closer to their true masses because of the random inclination of orbits (and would be even easier to detect).  It is possible that neither the size nor shape of the orbits reflect their initial values. This makes it difficult to distinguish giant planets from BDs on an orbital basis.

Stars often form with a companion. This process involves formation in disks (both cirmcumstellar and circumbinary), as does the formation of a lone star. Binary star formation leads to companions at any separation, with eccentric orbits. The difference between them and planets is thought to be the lack of a need to first form a rock/ice core. Unfortunately, there is no current method for determining whether there was such an initial core in extrasolar objects. Thus, formation in a disk does not by itself distinguish star from planet formation, and apparently neither do orbital eccentricity or separation. It is possible that giant planets form both by gas accretion onto a rocky core and by more direct forms of gravitational collapse in gaseous disks (Boss 1997). Even the requirement that a planet be found orbiting a star is now thought overly restrictive; when several giant planets form in a system it is easy for one or more to be ejected by orbital interactions and end up freely floating. For a much more detailed discussion of formation issues, see Protostars and Protoplanets IV (Mannings et al 2000).

Given these difficult issues, there is a rising school of thought that the definition of brown dwarfs should have a basis more similar to the definition of stars (based on interior physics). One intuitive difference between stars and planets is that stars experience nuclear fusion, while planets do not. We can therefore define the lower mass limit for BDs on that basis. Since significant deuterium fusion does not occur below 13 jupiters (Saumon et al 1996), that is the proposed lower mass limit for BDs. It is also thought to be near the lower limit for direct collapse of an interstellar cloud. With this definition, one must only determine its mass to classify an object. We can avoid the observational and theoretical uncertainties associated with a formation-based definition by using the mass-based definition, and that is what I advocate. Nonetheless, there actually is some evidence that most planets and most BDs form by different mechanisms. It is much more probable to find a planet rather than a BD as a companion to a solar-type star (see Section 7.2). It may turn out that there is more than one way to form things, but at least we’ll know what to call them.