4. Physical properties: age, mass, equivalent width, visual extinction, Lithium abundance
Table 5 summarizes the physical properties of the binary components. In Fig. 5 we have placed those binary components in HR diagrams for which we could derive both spatially resolved photometry and spectroscopy. For comparison we show theoretical pre-main-sequence evolutionary tracks from D'Antona & Mazzitelli (1994). These tracks are based on low-temperature opacities from Alexander et al. (1989) or Kurucz (1991) and on the convection model by Canuto & Mazzitelli (1990) or mixing-length theory. The errors in the placement of the individual binary components are indicated. The theoretical tracks allow us to determine the age and mass of the stars. All of our binaries have mass ratios between 0.5 and 1 (cf. Table 5) whereas only 60% of the binaries studied by Hartigan et al. (1994) have mass ratios in that range. Hence our sample appears to be biased somewhat towards equal mass pairs. However, even among the 15 binaries with mass ratios between 0.5 and 1 in the Hartigan et al. (1994) sample only eleven (75%) are coeval.
Table 5. Physical properties of the close binary components. Masses and ages are derived from PMS evolutionary tracks by D'Antona & Mazzitelli (1994) based on opacities by Alexander et al. (1989) and the Canuto & Mazzitelli (1990) description of convection.
Sets of theoretical PMS evolutionary tracks based on different input physics are shown in Fig. 5, bottom. A different convection model (i.e. mixing length theory) affects mainly the effective temperature of the evolutionary tracks (and hence the mass estimates) for stars around 0.5 to 0.7 . In the near future dynamical mass determinations for PMS binaries will allow us to distinguish observationally between these two sets of tracks. Tracks based on opacities from Kurucz (1991) lack molecular opacities and hence provide no adequate description for young late-type stars.
Hartigan et al. (1994) compared their observations also to tracks by Swenson et al. (unpublished). Applying the Swenson tracks to our sample of PMS binaries yield masses above 0.4 and ages between 3 106 yr and 3 107 yr for the binary components. These ages appear to be rather high given the fact that the stars in our sample are still associated with the molecular clouds (e.g. cf. Brandner et al. 1996; Fig. 5 & 6). On the other hand, observations of the spectroscopic binary NTT 155913-2233 suggest that the tracks by Swenson et al. might provide a better description for PMS stars in a mass range between 0.6 and 1.1 than the tracks by D'Antona & Mazzitelli (Prato & Simon, priv. comm.).
Fig. 6 nicely illustrates how age and mass estimates can be wrong due to unresolved binaries. For the objects studied in Chamaeleon I and II we have plotted both the most recent literature values for effective temperature and luminosity for the unresolved binaries (Feigelson et al. 1993; Hughes & Hartigan 1992) and the effective temperatures and luminosities as derived by us for the individual binary components. The literature values given by Feigelson et al. (1993) have been scaled to a distance of 150 pc. Note that the somewhat smaller values for the visual extinction towards the binaries derived by us lead also to somewhat lower estimates for the system luminosity in comparison to Feigelson et al. (1993) and Hughes & Hartigan (1992).
Unresolved PMS binaries could lead to a gross underestimate of the age of the stars involved (cf. Ghez 1994, Simon et al. 1993) and also induce errors in the individual mass estimates and hence in the derivation of the initial mass function.
In Fig. 7 different physical parameters of the primary and secondary are plotted against each other. Fig. 7a (top left) shows the equivalent width of the H emission line of the primary and the secondary. No clear correlation exists. Hence, chromospheric activity and/or accretion rates of primary and secondary appear not to be related. There are a number of binaries in which only the secondary exhibits strong H emission but not the primary. Such stars have only been detected in H objective prism surveys because they are binaries and have a secondary bright in H . This selection effect may explain at least partially why the number of binaries among pre-main-sequence stars appears to be higher than among main-sequence stars (cf. Ghez et al. 1993; Leinert et al. 1993; Reipurth & Zinnecker 1993).
The Lithium I (670.7nm) equivalent width of primary and secondary (Fig. 7b, top right) are not correlated as a test based on rank statistics indicates. Differences in the individual veiling and the non-linear dependence of the Lithium abundance on age and effective temperature might be responsible for that. See below for a more detailed analysis considering curve of growth calculations and Lithium depletion as a function of age and effective temperature. ESO H 281 A is the only star in our sample of 28 binary components which does not show any sign of Lithium absorption. The small value of its A index indicates that ESO H 281 A is a background giant.
The visual extinction towards the primary and secondary (Fig. 7c, bottom left) are in good agreement with each other. This indicates that both components of each binary are embedded equally deep in the dark clouds and thus gives further evidence that the objects studied are indeed physical binaries. Furthermore, the agreement of the extinction values yields that the circumstellar disks around the individual components in each binary system in our sample are aligned.
The age of the primary and secondary (Fig. 7d, bottom right) derived from the theoretical pre-main-sequence evolutionary tracks is generally in good agreement for ages in a range of a few 106 yr to 107 yr. Only for one star (Sz 20) the secondary might be somewhat younger than the primary. However, the deviation amounts to only 1.5 and thus could very well be a purely statistical fluctuation.
Based on the curve of growth calculations by Martín et al. (1994) we were able to derive Lithium abundances for our sample of pre-main-sequence binary components. Fig. 8 shows the Lithium abundances as a function of effective temperature. Overplotted are theoretical isochrones from D'Antona & Mazzitelli (1994). As the curve of growth calculations are based on Kurucz model atmospheres they are only valid down to effective temperatures of about 3700 K. Below that temperature the abundances derived from the model calculations lie probably too low (Martín et al. 1994). Veiling of photospheric lines could additionally lead to an underestimate of the actual Lithium abundance. Hence, the values plotted in Fig. 8 merely represent lower limits in most cases.
For the components of Sz 48 and Sz 59 the Lithium isochrones suggest an age of less than 107 yr. In order to match this with their positions in the HR diagram the distance to these stars (and thus towards the Chamaeleon II cloud) has to be at least 200 pc.
© European Southern Observatory (ESO) 1997
Online publication: June 30, 1998