The new spectroscopic class L

Neill Reid
inr@dowland.caltech.edu
June 19, 1998

Originally posted to sci.astro.

Maybe the following will cast a little more light on the L-dwarfs issue:

First, it's important to remember that the spectral type sequence is a morphological description based on the spectroscopic appearance of star-like objects. That depends on two properties - chemical composition and surface temperature. For members of the Galactic disk population, which all have chemical properties very similar to those of the Sun, surface temperature is the dominant factor. One defines a spectral type by looking for patterns in the way that certain features grow and recede in strength, and identifying certain objects as `standards'. The latter provide a reference grid, so one can take observations of a star of interest and estimate its properties by comparing the spectrum against the standard set.

M-dwarf stars, the coolest objects in the standard sequence, have surface temperatures between about 4000 and 2500 degrees Kelvin (the lower limit is uncertain by at least 200 K). Their spectra are dominated by molecular bands, notably titanium oxide and vanadium oxide.

The L-dwarfs which were the subject of the recent press release have cooler surface temperatures, probably as low as 1600 to 1800 K, and at those lower temperatures, different molecules are present - metal hydrides, like iron hydride, chromium hydride and, to a lesser extent, calcium hydride. Theoretical models predict that titanium and vanadium actually precipitate out of the atmosphere as dust at such low temperatures.

How does that fit in with brown dwarfs? I think it's important to hold two physical concepts separate while considering these objects:

This is a slightly tricky distinction which didn't quite make it into the press reports.

Isolated brown dwarfs and stars form (as far as we know) in the same way - by gravitational contraction within a molecular cloud. The only difference is that stars have sufficient mass that the potential energy lost in collapse raises the central temperature to levels where hydrogen fusion starts. Brown dwarfs never achieve that level, so they evolve rapidly (in astronomical terms). Initially, their surface temperatures are over 3000 K, so brown dwarfs look like M-dwarfs, but they cool rapidly and enter the L-dwarf regime after only a few million years. They also become significantly fainter as their surface temperature falls. After a time which depends on their mass (anywhere from a few hundred thousand to a few billion years), the brown dwarfs will cool to below 1500 K - and methane will form in the atmosphere and dominate the spectrum (at least at infrared wavelength). At that point, the brown dwarfs cease to be L-dwarfs and will enter a new spectral class. We can't define that class at present since we only have a single example - the brown dwarf Gl 229B.

Stars, on the other hand, will settle into a relatively fixed position in temperature and luminosity as hydrogen fusion stabilises. It turns out that theoretical models predict that the lowest-mass stars (objects with masses between about 0.095 M(sun) and 0.075 M(sun)) have equilibrium temperatures below 2500K - so they should look like L-dwarfs.

So both the M-dwarf and L-dwarf classifications encompass stars and brown dwarfs, although only a very small fraction of M-dwarfs are brown dwarfs. Note that this doesn't mean that spectral typing breaks down - spectral types give you an estimate of temperature. It just happens that for stars of a given mass, most of their lifetime is spent (almost) at a fixed temperature. It's that monotonic mass/luminosity/temperature relation which breaks down for brown dwarfs.

Amongst the 25 or so L-dwarfs currently known (20 of which are from our analysis of the 2MASS infrared sky survey), at least 9 are brown dwarfs, since they have measureable quantities of lithium in their atmosphere: lithium is destroyed if the central temperature exceeds 2.5 million K, which is the case for objects with masses of more than 0.06 solar masses. Some of the other L-dwarfs (probably 8-10 of the remaining 17) are also likely to be brown dwarfs, but with masses between 0.06 and 0.075 M(sun); the remainder are very low-mass stars.

Tying in these results to 'super-planets' is difficult. My opinion is that there is a significant difference between most of the objects which are being found as companions and those found as isolated objects. The companion object searches should be very efficient at finding brown dwarfs as massive as our field objects (which are between 0.075 and probably 0.04 M(sun)), but only one or two have been found - as compared with about a dozen planetary mass (0.02 M(sun), or 20 Jovian masses, or less). So my feeling is that the latter are a different sort of beast.

One final important point: while there are still a lot of uncertainties in our analysis, I think it's clear that these objects don't provide sufficient total mass to contribute significantly to any missing mass problem. Our current estimate is that there's only about 10 percent as much mass in brown dwarfs as there is in stars in the Galactic disk.

Davy Kirkpatrick is currently working on the final details of the classification scheme. We (Kirkpatrick, Reid, Liebert, Cutri, Beichmann, Monet et al) will be writing up our results for the Astronomical Journal.

Much of this work is based on data from the 2MASS Survey.