Astron. Astrophys. 351, L5-L9 (1999)

4. Discussion

4.1. LHS 102B: an L dwarf companion to an M dwarf

One of the two L dwarfs is within 20 arcsec of a previously known high proper motion star, LHS 102 (M3.5V). It shares its proper motion of towards PA=, and the two objects are thus physical companions. The trigonometric parallax of LHS 102 (van Altena et al. 1995) provides a distance for the system of  pc, and LHS 102B is thus a rare case of an L dwarf of known distance and luminosity. Just a few months ago only two other L dwarfs had known distances: GD 165B (Becklin & Zuckerman 1988) through its association with GD 165A, and Roque 25 which Martín et al. (1998b) established to lie (at 94% C.L.) in the Pleiades, whose distance pc is known through main-sequence fitting (Pinsonneault et al. 1998, Soderblom et al. 1998) and whose radius is pc (Narayanan & Gould 1999). Kirkpatrick et al. (1999) have since presented preliminary parallaxes for another three L dwarfs. They are also shown in Fig. 3, though Martín et al. (1999) suggest that they might perhaps have problems, as the stars would be very young for field objects (Gyr) and would have preserved lithium contrary to model expectations and observations in the Pleiades (see Martin et al. 1998a). It is difficult to assess their reliability from the limited information in Kirkpatrick et al. (1999), but possible sources of trouble include a relatively short timespan, and strong differential colour refraction from the extreme colours difference between the L dwarfs and their reference frames (as the USNO uses a very broad filter). Alternatively those sources could be binaries, though it seems unlikely that all three are.

Fig. 3. a  :I-K HR diagram for our objects, along with M and L dwarfs with known distances (from Leggett 1992, Tinney et al. 1993 and Kirkpatrick et al. 1999). Models are overlaid for both dust-free and NG-DUSTY atmospheres and for ages of 5 Gyr (appropriate for field objects) and 120 Myr (appropriate for the Pleiades brown dwarf Roque 25). b  Colour-colour diagram of M and L dwarfs (Fig. 6 of Delfosse et al. 1999) with the two new L dwarfs (square symbols).

Fig. 3 shows M dwarfs of known distance and the six L dwarfs in an vs I-K HR diagram, together with two sets of theoretical tracks, NextGen and NG-DUSTY. Dust condenses in the atmospheres of very cool dwarfs, with two main consequences: depletion from the atmosphere of the refractory elements; such as Ti and V; decreases line opacities; and dust continuum opacity changes the atmospheric structure through a greenhouse effect. The NextGen models (Hauschildt et al. 1999) ignore dust condensation altogether, while the NG-DUSTY models (Leggett et al. 1998, Allard & Hauschildt 1999) account for its effect on both the chemical equilibrium and the continuous opacity. As can be seen in Fig. 3, the NextGen models provide an excellent fit to near-IR colours and luminosities of M dwarfs, but fail to reproduce the J-K reddening of the late M and L dwarf sequence. The NG-DUSTY models in contrast provide an impressive fit of the near IR colours and luminosities of L dwarf, especially when one considers the still preliminary nature of these complex models. Clearly dust condensation plays a dominant role in the atmospheric physics at these temperatures.

Comparison with the NG-DUSTY models gives an effective temperature of  K for LHS 102B, consistent with that derived from the optical spectrum (Basri et al. 1999). The best fit is obtained for the 5 Gyr isochrone and a mass of (just at the stellar/substellar mass limit for models using NG-DUSTY atmospheres (Allard & Hauschildt 1999)), but the data are also consistent with a 1 Gyr age and a (substellar) mass just above . Since LHS 102B has depleted its lithium (Basri et al. 1999), its mass must be larger than and its age therefore cannot be less than 1 Gyr. The optical spectrum indicates shows weak emission, and this low level chromospheric activity may indicate that LHS 102B is not very much older than that minimum age. It can be either a star or a brown dwarf and we cannot presently say on which side of the border it stands.

4.2. EROS-MP J0032-4405: a field brown dwarf

The second object, EROS-MP J0032-4405, has I-J and J-K. From comparison with NG-DUSTY atmospheric models, we obtain an effective temperature of  K. This is only marginally consistent with the effective temperature of 2200+-100 K, which corresponds to the L0 spectral type derived by Martin et al. 1999, with 1- error bars extending to M9.5-L0.5 classes. The 670.8 nm lithium line absorption in the optical spectrum (see Fig. 4 and Martín et al. 1999) indicates that this fully convective very cool dwarf has not depleted its lithium. Since lithium is destroyed by proton capture at lower temperature than needed for hydrogen fusion (Rebolo et al. 1992), EROS-MP J0032-4405 has to be a brown dwarf, less massive than . Since models show that brown dwarfs cool down to effective temperatures of  1800 K at an age of 1 Gyr (and less massive ones cool faster), it must also be younger than 1 Gyr.

Fig. 4. Spectrum of EROS J0032-4405 and LHS 102B (shifted by 10 units), as in Martin et al. 1999. The flux has been normalized to the counts in the region 738-742 nm. See Martin et al. 1999for details.

© European Southern Observatory (ESO) 1999

Online publication: November 2, 1999
helpdesk.link@springer.de