JPL and NASA News

Bill Wheaton, IPAC

Methane Dwarfs: A New Spectral Type (T?)

Until recently GL 229B was the most extreme cool substellar dwarf known. It has strong methane (CH₄ ) features in its spectrum.

Both methane and water strongly absorb light in the near IR, between say 1.5 and 2.5 microns. Without those molecules, you tend to see down into the "star"'s atmosphere to a certain level, and then, when the atmosphere becomes hot enough to be opaque, it looks more or less like a smooth "Planck function" -- the universal black body curve -- for the temperature at the photosphere, which is the surface you see. (Actually there are always other lines and features, due to atoms, which complicate things a lot.) The smooth Planckian black body curve of emission versus wavelength is called "the continuum", to distinguish it from any lines, emission or absorption. That is, the continuum is the level of the spectrum between the lines. For the Sun this temperature T is about 5700 K, whereas for extreme dwarf M stars (dM stars, as they are often indicated) T is down about 2000 K. For the new class of L dwarfs, defined last year, T is in the range of roughly 1500-2000 K.

Probably about half of the L dwarfs are "brown dwarfs" -- too small (mass M <7.5 % the mass of the Sun, approximately) to make it to H burning and become real stars. Unfortunately you can't tell whether they are really brown dwarfs just from the colors, because near the dividing line the outside of the star looks pretty much the same, and you don't know their age. The brown dwarfs continue to cool off slowly with time, whereas real stars stop cooling as the nuclear power source takes over, and then they just hang there, with T nearly steady, for many tens of billions of years. Near the hot end of the L dwarf range (L0, say), most of the objects are real stars, and near the bottom (L9), most or all are brown dwarfs. Near the middle however, an L5 object could be either a younger brown dwarf, just passing through on its way to L9 and below, but which has not yet had a chance to cool off; o,r it could be an extremely old star, gently simmering away, where it will remain apparently unchanged till "the end of time" or close enough for human purposes.

But, if T is still lower, molecules, especially water (H₂O) and CH₄, can start to form in a star's atmosphere. If it's too hot, these molecules simply can't hold together against the violent shaking high T implies. At a T below about 1300 K, methane starts to be important, affecting the IR color of the object. CH₄ absorbs and scatters light so strongly in the near IR, along with water (ie, steam in this context), that there are big bands where you cannot see down to the deeper levels, because the upper atmosphere is not transparent. High up, where you do see, the atmosphere is yet much cooler, so the emission due to the Planckian is for a lower T, and therefore much less intense. Thus observationally, the result is that there is a big hole in the continuum at those wavelengths.

Stars having this strong CH₄ absorption at about 2.2 microns look "blue" when you compare their J band (1.2 micron) and Ks band (2.2 micron) brightness. But they usually don't appear on the R (red) band visual image from the Palomar Sky Survey at all -- they are too red and cool to show up, even in R.

With 2MASS (the Two Micron All Sky Survey), Davey Kirkpatrick and coworkers had been looking for redder and redder dwarfs, and not finding them, although many extremely cool L objects have been turning up. It is clear now that the reason is that when the CH₄ absorption is strong, below 1300 K or so, they become bluer in the J-Ks color comparison. It's important astrophysically, because for stars as cool as GL 229B (900-1000 K), we can be sure they will never make it to H burning.

Ultimately, in order to make sure about these candidates, you have to take an object with these funny colors, check that it is not on the Palomar R plate, and then go to a large telescope with an infrared spectrometer - Kirkpatrick and his colleagues have been using the 10 m Keck mostly - and get a spectrum. Then you can see the methane and water absorption bands eating away at the continuum in the Ks band, and be sure nothing else strange is happening that would affect the IR colors ("J-H", say, comparing the J and H bands, and "H-Ks", from comparison of H and Ks). Only then can you obtain an accurate temperature and be certain. This is obviously a laborious process, limited by the amount of large telescope time you can get, in particular. Nevertheless, to speed up the process it is quite helpful to know in detail how the colors change with T. Then you can take only the best candidates to the telescope and spectrometer.

For the past few years GL 229B has been the only "methane" dwarf known, cooler than even than L9, the lowest T the new spectral class L can accommodate. After the summer meeting of the American Astronomical Association in Chicago, we have now 6 new ones, a total of 7 in all. Unlike GL 229B (the binary companion of the far-brighter GL 229A), the six new methane dwarfs are all "field" objects, isolated and floating alone in space. Two were discovered by the SLOAN Digital Sky Survey, a new and very sensitive 5-band optical survey between the near UV (0.35 micron) and the near IR (0.9 micron). The other four were discovered by 2MASS. At least one is cooler than GL 229B. Insofar as they can be estimated, their distances appear to be on the order of 10 pc.

It is clear that the venerable OBAFGKM spectral sequence, extended only a year ago to include the L dwarfs, will have to be extended yet again, as it is defined only above about 1500 K. While no official letter has yet been decided, there are only about three letters remaining that are not already in use one way or another. Eight are in the present sequence OBAFGKML, with three more for the carbon stars (C,S,N), making 11 for stellar spectral sequences alone. Another dozen or so are effectively used for other purposes, such as D for white dwarfs (DA, DB, DC), E for elliptical galaxies (E0-E9), and so forth, which would be bound to cause confusion. Kirkpatrick and his co-workers believe T is the best available choice; if that is accepted, we will than have OBAFGKMLT.

Next Month:

A year or so ago, based on the experience with 1997 XF11, we predicted that many more asteroid impact false alarms would be forthcoming, based mostly on the realization that the chances of an object passing rather close (e.g. 65,000 km from Earth, say) is of course much larger than (e.g., about 100 times) the chance of our actually being hit. In the good old days we never noticed such quiet passages. Now that we can actually take action to affect the result if need be, it is inevitable that lots of close calls are going to be detected, evaluated, and worried about. In almost all cases these will eventually be dismissed as more accurate measurements become available showing that no impact is possible. Sure enough, two new potential impactors have turned up in the past month or so, the latest only in the past few days as I write. It is too early to evaluate the very newest threat (except to say the danger appears practically negligible, probability P <10^{-7 of impact), but the earlier one, 1999 AN10, is interesting -- though not at all alarming as yet.}

We will leave the details of the story about 1999 AN10 until next month. Also, by the time this reaches you, the Far Ultraviolet Spectroscopic Explorer, FUSE, should be in orbit; its Delta II launch vehicle was just stacked last week. FUSE is one of those missions that has been patiently progressing towards fruition, year after year, for nearly two decades. Next month it should be safely in orbit, and unless something more urgent intervenes we will take a look at it in the August issue.

Bill Wheaton
Waw@ipac.caltech.edu