Astron. Astrophys. 364, 217-224 (2000)

2. Sample

2.1. Accurate masses for M dwarfs

We adopt a 10% mass accuracy cutoff for inclusion in our new M dwarf M/L relations, as a compromise between good statistics and the quality of the individual measurements. To our knowledge 32 M dwarfs fulfill this criterion. They can be divided into four broad categories:

1. Four systems have orbits of a quality that hasn't changed much since Henry & McCarthy (1993), but masses which are now somewhat better determined thanks to the availability of the Hipparcos parallaxes (Henry et al. 1999, and this paper): Gl 65, Gl 661, Gl 702, and Gl 860. One should note however that the Gl 661 masses derived by Martin et al. (1998) represent a 3  correction to the Henry & McCarthy (1993) values. The new masses are much more consistent with the average M/L relation, and we believe that Henry & McCarthy (1993) had underestimated their standard errors for that particular pair. These four systems have long to very long periods, up to 90 years for Gl 702AB. Their inclusion is a testimony to the care and dedication of binary star observers over many decades, but the accuracy of most of these masses is unlikely to be significantly improved over the next few years.

2. The three eclipsing systems have extremely accurate masses, with relative precisions of 0.5% for CM Dra (Metcalfe et al. 1996) and 0.2% for YY Gem and GJ 2069A (Ségransan et al. 2000). Their parallaxes are unfortunately less precisely known, in part due to their slightly larger distances than those of the visual systems. Their luminosities are therefore more uncertain than their masses. Their flux ratios in the near-IR JHK bands have also not yet been determined.

3. The masses of Gl 473 (8%; Torres et al. 1999) and Gl 791.2 (2%; Benedict et al. 2000) result from the effort of the FGS astrometry team on HST (Benedict et al. 1999). We have chosen to temporarily exclude the measurement of the Gl 748 system mass by the same group (5%; Franz et al. 1998): a M/L relation has to be used to estimate individual masses for the two stars in that system. This would introduce an undesirable circular aspect to our discussion. We have on the other hand retained their Gl 473 determination, which formally has the same weakness. The two components of that system are sufficiently similar that this contributes negligible additional uncertainties.

4. Most of the masses in Table 3 result from our own programme: since 1995 we have been monitoring a sample of solar neighbourhood M dwarfs with high precision radial velocity and adaptive optics imaging observations (Delfosse et al. 1999c, for a complete presentation of the project). These observations have resulted in a new orbit for Gl 747, and in improved orbits and masses with 0.5 to 5% accuracy for Gl 234, Gl 644, Gl 831, Gl 866 (Ségransan et al. 2000), Gl 570B and Gl 623 (Ségransan et al. in prep). Some additionnal binaries, including discoveries from that programme (Delfosse et al. 1999c; Beuzit et al. in prep.) are nearing the time when their masses will be known with similar accuracies.

2.2. Photometry

Table 1 lists the basic properties of the selected systems: parallaxes, spectral types and integrated photometry. Table 2 lists the magnitude differences for the V, R, I, J, H and K photometric bands. The near-IR flux ratios were obtained either from litterature infrared speckle observations, or extracted from our adaptics images (Ségransan et al. 2000). For the three eclipsing systems the visible flux ratios were adopted from analyses of the light curves (Leung & Schneider 1978; Delfosse et al. 1999a; Lacy 1977). For the visual binaries, they were preferentially adopted from the FGS work of Henry et al. (1999), with the standard errors quoted in that article. When such measurements were unavailable, we relied instead on spectroscopic magnitude differences, using the relative areas of the ELODIE cross-correlation peaks as a proxy for the V band magnitude difference. The ELODIE cross-correlation has an effective bandpass centered close to the central wavelenghth of the Johnson V filter, but it is significantly broader. A "colour-tranformation" would thus in principle be needed to derive V band magnitude differences. A comparison with the direct measurements of Henry et al. (1999) for the sources in common shows maximum relative errors of 10% from neglecting this transformation: a 0.5 magnitude contrast is in error by at most 0.05 magnitude, and a 2 magnitudes one by at most 0.2 magnitude. We have therefore adopted the larger of 0.05 magnitude and 10% of the magnitude difference as a conservative estimate of the standard error for these spectroscopic magnitude differences.

Table 1. Basic parameters for the M-dwarf systems with accurate masses. The parallaxes mostly originate from the Hipparcos catalog (ESA 1997), the Yale General Catalog of trigonometric Parallaxes (Van Altena et al. 1995), and from Ségransan et al. (2000). In that paper we derived optimal combinations of orbital and astrometric parallaxes, which often have a strong contribution from either an Hipparcos or a Yale catalog astrometric parallax. Some individual entries are additionally taken from Soderhjelm (1999), Forveille et al. (1999), and Benedict et al. (2000). All spectral types are from either Reid et al. (1995) or Hawley et al. (1997) and refer to the integrated light of the system. The photometry is taken from the extensive homogenized compilation of Leggett (1992), except for YY Gem, Gl 702AB and GJ2069A. The photometry of GJ2069A is from Weiss (1991). The optical photometry of YY Gem is from Kron et al. (1957) (RI), Eggen (1968) (UBV), Barnes et al. (1978) (UBVRI), and the infrared photometry from Johnson (1965), Glass (1975), and Veeder (1974). The optical photometry for Gl 702AB is from Bessel (1990) and the infrared photometry from Alonso et al. (1994). All photometry was converted to the Johnson-Cousins-CIT sytem adopted by Leggett (1992) using the colour transformations listed in that paper. Multiple measurements for the same band were averaged with equal weights.

Table 2. Magnitude differences for M-dwarf systems with accurate masses. Reference codes are: TYC for the Tycho catalogue HIP for the Hipparcos catalogue L77 for Lacy (1977), L78 for Leung & Schneider (1978), Hen93 for Henry & McCarthy (1993), C94 for Coppenbarger et al. (1994), B96 for Barbieri et al. (1996), Ben00 for Benedict et al. (2000), Tor99 for Torres et al. (1999) Hen99 for Henry et al. (1999), D99a for Delfosse et al. (1999a), For99 for Forveille et al. (1999), and D00 for the present paper. S stands for spectroscopic magnitude differences, infered from the relative line depths for double-lined spectroscopic binaries.

© European Southern Observatory (ESO) 2000

Online publication: December 15, 2000
helpdesk.link@springer.de