≡ Menu

How Do Brown Dwarfs Form?

Now that we are looking forward to the WISE mission (Wide-Field Infrared Survey Explorer) and its investigations into nearby brown dwarfs, it’s startling to realize that we detected the first of these objects as recently as 1995. Today they’re all the rage, particularly among that small band of us obsessed with missions to nearby interstellar space. A cool, dim brown dwarf could be the closest star to our Sun, an obvious target for a future probe once long-haul propulsion options begin to mature.

Brown dwarfs are too cool to trigger hydrogen fusion, so it takes infrared capabilities like those of WISE or the Spitzer Space Telescope to tell us much about these dim objects. A key question has been whether they form like planets or stars. Spitzer may have found the answer in the form of two ‘proto brown dwarfs’ that have been located in a cloud called Barnard 213, a region of the Taurus-Auriga complex where young objects abound. The finding is significant because we’ve never before found a clear case of a brown dwarf in its earliest formative period.

David Barrado (Centro de Astrobiologia, Madrid) describes the thinking that went into finding these objects:

“We decided to go several steps back in the process when (brown dwarfs) are really hidden. During this step they would have an (opaque) envelope, a cocoon, and they would be easier to identify due to their strong infrared excesses. We have used this property to identify them. This is where Spitzer plays an important role because Spitzer can have a look inside these clouds. Without it this wouldn’t have been possible.”

The result is an infant brown dwarf called SSTB213 J041757, a find that turns out to include not one but two brown dwarfs that are among the faintest and coolest ever observed. Numerous other sites, from the Caltech Submillimeter Observatory in Hawaii to the Calar Alto Observatory in Spain and the Very Large Array in New Mexico took part in the study, which analyzed the dusty envelope around the objects, allowing astronomers to create a spectral density distribution that shows the amount of energy emitted by the objects in each wavelength.


Image: Here we see a long sought-after view of these very young objects, labeled as A and B, which appear as closely-spaced purple-blue and orange-white dots at the very center of this image. The surrounding envelope of cool dust surrounding this nursery can be seen in purple. These twins, which were found in the region of the Taurus-Auriga star-formation complex, are the youngest of their kind ever detected. They are also helping astronomers solve a long-standing riddle: Do brown dwarfs form more like stars or planets? Credit: NASA/JPL-Caltech/D. Barrado (CAB/INTA-CSIC).

Assuming the work stands up to scrutiny, the implication is that brown dwarfs form more like stars than planets — the spectral density distribution matches other young, low-mass stars. The paper, not yet available, is Barrado et al., “A proto brown dwarf candidate in Taurus,” in press at Astronomy & Astrophysics.


Comments on this entry are closed.

  • djlactin November 24, 2009, 13:52

    Welllll…. THESE two formed more like stars than like planets . Sample size is too small to eliminate the possibility that some Brown dwarfs formed within stellar systems and were later ejected by n-body interactions or liberated by gravitational disruption by passing stars. (Is one forming ‘like’ a planet around the other?) Awesome find, though. I await the discovery of Nemesis.

  • Tim Kendall November 24, 2009, 17:25

    Very interesting to find such clear evidence for a brown dwarf pair at what the star formation community would call the Class 0 phase, by analogy to protostars. Low-mass pairs such as these are bound to be common in young (~few Myr) star forming regions. Much evidence suggests that the initial mass function doesn’t change much across the brown dwarf maximum mass (~72MJup) boundary. After age-dependent degrees of ejection rate from the natal cloud (star-forming region, young cluster), the tighter bound pairs, probably a few higher systems and a whole lot of single brown dwarfs are seen dispersing through the interstellar medium with low (few km/sec) velocities. Links to nearby clusters such as the Pleiades can be unravelled via intricate space motion measurements. Some of the local solar neighbourhood very low mass stars and brown dwarf (BD) are these; many of the remainder will turn out to be correspondingly interesting very old, low metallicity objects from the Galactic Halo. The reason to seek young BD pairs is then that they are bright at Spitzer wavelengths while still at low masses, and presumably interruption of accretion by dynamical ejection happens. In this way, BD can be sought to very low masses (certainly pushing a Jupiter mass at 1 Myr on average, I’d say). The real importance of this work is that the object is binary and so evolutionary models are tested at ~few MJup masses – and such objects are formed “as stars” i.e. by the same formation processes. That is, they are common in the Universe. That one is drifting in such a way exactly towards the solar system right now is almost certain, and it is equally probable that quite a few happen to be drifting precisely away.

  • kurt9 November 24, 2009, 17:59

    Is there a minimum size to objects that can form by star formation processes?

  • Tim Kendall November 25, 2009, 7:49

    Kurt9 – that is indeed a real question. The likelihood is that too is governed by nuclear processes, in this case the fusion of deuterium. Objects of thirteen Jupiter masses and above are certainly observed in star formation regions (i.e. these are young brown dwarfs), and Dr Barrado is a leader in the difficult observational field of seeking objects below the D-limit. It seems though that 3-5 Jupiter mass objects are quite regularly observed in the youngest, nearest SFRs. Objects of similar mass observed as binary companions to nearby stars in planetary-type orbits are observed but this parameter space is sparsely populated (the so-called “brown dwarf desert”). It seems the same dynamical ejection processes are to blame for perhaps disrupting low mass star/low mass BD pairs in young clusters, or indeed – since it is hard to imagine the preferential ejection of tightly bound pairs – that the wobbly, eccentric orbits followed my many such heavy planets is itself a result of formation-time dynamical interactions. Moreover, wide such pairs, sharing a common proper motion, are beginning to be found in the solar neighbourhood. This shows that relatively weakly bound systems can survive being formed – and indeed persist, and of course there will always be a few results of accidental gravitational “capture” events, although this is uncommon as objects like these don’t often pass close enough to each other once out into the interstellar medium.

    Either way a smooth initial mass function (at formation) across the D-limit is to be expected with no reason to expect a lack of very low mass BD in the neighborhood either.

    With regard to the Nemesis object, one can endlessly speculate on the period of
    say a one Jupiter mass object at aphelion from the Sun in its long, hyperbolic orbit, but I think one can say for certain that there is little real evidence for such a body with, say, a 10 or 20 Myr period, based on re-occurrence of “catastrophic” events on Earth in the geological past or deep near-infrared sky surveys, which would by now have uncovered anything closer than the Centauri system with such a high proper motion, down to say a K magnitude of about 22.

    Fans of the appealing side of the Nemesis concept must not forget Douglas Adams and always remember that space is really big.

  • keith November 25, 2009, 14:05

    Tim Kendall, your statement that the deep near-infrared surveys would by now have uncovered anything closer than the Centauri system seems to be at odds with the WISE people. They believe there is a good chance the nearest “star” will be detected in the forthcoming WISE mission. Do you discount the possibility of old, cool BD’s closer than Centauri?

  • Ron S November 25, 2009, 15:45

    keith, I don’t think Tim said that at all, only that if there is a close BD it is unlikely to be gravitationally bound to our system. That is, it’s just passing through the neighborhood.

  • Tim Kendall November 25, 2009, 16:47

    Keith – not really, I haven’t done the exact calculation, or indeed looked yet at the WISE numbers! I’ve looked for and found L dwarfs at ~10pc distances in 2MASS and SuperCosmos data including the occasional object later than L5, which has to be a brown dwarf given any typical field age. I’ve looked at the whole southern hemisphere and most of the north in this way combined with a reduced proper motion criterion but could certainly have missed an object with a very low tangential velocity. Guessing a bit but I’d say maybe a few very old one Jupiter mass objects could well be found by WISE at only ~ a few pc or even closer than the Centauri system, and if it had K >~ 23-25 say it might be hiding in data from e.g. the UKIDSS Large Area Survey. Things with fast halo kinematics have been looked for, and found, but I’d suspect a lurking Centauri-beating BD would turn out to be more of a thick disk object kinematically.

  • kurt9 November 25, 2009, 20:37

    WISE will certainly tell us more about the availability of sub-stellar objects in or neighborhood. However, I thought of another approach. If there is no roll-off in the independent formation of objects below the D-limit, shouldn’t this result in scattering of stellar light as we see it from the Earth.

    I do recall we had this discussion about a year ago and someone cited a paper that stellar object formation peaks at around 0.07 solar masses (slightly smaller than an M-star) and that the availability of brown dwarfs would be roughly equal to that of M stars.

  • keith November 26, 2009, 9:31

    Ron S, does it make any difference one way or the other if it is gravitationally bound or not. ? Granted a gravitationally bound object is more likely to have tangential motion, but is that the overiding consideration?

    In his reply above, Tim Kendall DOES appear to say that anything significantly more massive than Jupiter and nearer than Centauri would probably already have been discovered. The only possibility for WISE to discover one or more BD’s closer than Centauri is if they happen to have very low tangential velocities. Anyway, at least we should find out one way or the other over the next few months.

    kurt9, yes I’ve seen papers like that, however someone linked to one recently on here about obervations on a young star cluster, and the mass distribution continued to rise at very low masses. These models are poorly constrained by observations and, to me, basically unvalidated at low masses. What you see in some papers may simply reflect the observational bias, ie old BD’s can’t be seen.

  • djlactin November 28, 2009, 11:03

    Tim Kendall: the Nemesis reference was a joke (with a grain of hope – actually discovering a massive distant bound companion of Sol would be …er… earth-shaking).

  • Ron S November 30, 2009, 17:07

    keith, I think I misunderstood your point when I connected your comment to the point made about Nemesis.