3. The Lithium main features
Fig. 2 shows the distribution of ALi on the HR diagram and Fig. 3 presents ALi versus for three ranges of metallicity. The main features presented in Paper I, i.e., the lithium discontinuity around equal to 3.75, and the dispersion in lithium abundances for subgiants hotter than this value, clearly appear in both figures.
3.1. Lithium discontinuity
The observational lithium discontinuity actually simply reflects the well-known dilution that occurs when the convective envelope starts to deepen after the turnoff and reaches the inner free-lithium layers (Iben 1967a,b). In Fig. 4 we show this behaviour in our models as a function of the effective temperature for different masses. A closer look at the models shows that the beginning of the theoretical lithium dilution is a function of stellar mass, as can be seen in Table 2 where is the effective temperature at the beginning of the lithium dilution for each mass. We also give , which is the effective temperature at which a decrease of the surface lithium abundance by a factor 10 compared to the value on the main sequence is achieved. The dilution is a very fast process, both in terms of age and effective temperature interval. The theoretical beginning of the lithium dilution is in good agreement with the observed abundance drop-off along the subgiant branch, as can be seen in Fig. 5.
Table 2. Relevant characteristics of the models at [Fe/H]=0 shown in Fig. 4. Column 1 gives the stellar masses. Columns 2 and 3 give the effective temperature at the point where the convective envelope just starts deepening and at the onset of the theoretical lithium dilution. The effective temperature at the point where the lithium abundance has decreased by a factor 10 compared to its value at the end of the main sequence is indicated in column 4. Column 5 gives the mass of the convective envelope at the effective temperature of the observed rotational discontinuity
For the considered stellar masses, the predicted dilution factor at the end of the dredge-up (see point a on Fig. 4) ranges between 20 and 60 (these values obtained from our models do not significantly differ from the predictions by Iben 1967a,b). This corresponds to the upper envelope of the observations on the right side of the discontinuity, assuming a cosmic lithium abundance ALi of 3.1.
3.2. Lithium dispersion
In the following, we discuss the lithium dispersion on both sides of the discontinuity as a function of the stellar mass inferred from the evolutionary tracks at the corresponding metallicity. We select four ranges of mass according to the ones defined in Balachandran (1995) for the cluster and field stars with respect to the so-called lithium dip region (Boesgaard & Tripicco 1986). Each subsample corresponds to peculiar observational behaviour of the ALi on the main sequence (see Fig. 5 where both main sequence and subgiants stars are shown). In the four mass ranges a large dispersion of the lithium abundance appears and is independent of the single or binary status.
(i) Stars with masses show different degrees of lithium depletion, which occurs already on the pre-main sequence and on main sequence, as observed in open clusters and field stars (Soderblom et al. 1993; Jones et al. 1999 and references therein). This explains their low lithium content when they reach the subgiant branch (see Fig. 5 a).
(ii) Stars with masses between 1.2 and correspond to the so-called dip region (Boesgaard & Tripicco 1986).
We separate the stars originating from the hot side of the dip region in two mass ranges:
(iii) stars with masses between 1.5 and , and
(iv) above .
We now discuss in more detail the last three cases.
3.2.1. Li-dip stars
Late F- field and open cluster dwarfs with around 6700 K are highly lithium depleted. We find this feature among the dwarfs of our sample which have masses between 1.2 and (Fig. 5 b), in agreement with the observations in the open galactic clusters older than 200 Myr (Balachandran 1995). This lithium depletion persists in our subgiants of the same mass interval, in agreement with the observations in slightly evolved stars of M67 (Pilachowski et al. 1988; Balachandran 1995; Deliyannis et al. 1997). Explanations relying on the nuclear destruction of lithium are thus favoured by these data (see Talon & Charbonnel 1998 and references therein).
3.2.2. Li in stars originating from the hot side of the dip
As can be seen in Fig. 5 c,d (see also Fig. 2), a large lithium dispersion exists among our most massive stars, which show lithium depletion by up to two orders of magnitude before the start of the dilution at Log (point ). Very few observational data are available in the literature for stars with masses higher than . In the Hyades, while on the main sequence these objects show lithium abundances close to the galactic value, except for a few deficient Am-stars (Boesgaard 1987; Burkhart & Coupry 1989). However, our observational result is in agreement with the findings by Balachandran (1990) and Burkhart & Coupry (1991) of a few slightly evolved field stars originating from the hot side of the dip and showing significant lithium depletion. We thus confirm the suggestion by Vauclair (1991, see also Charbonnel & Vauclair 1992) that some extra-lithium depletion occurs inside these stars when they are on the main sequence, even if its signature does not appear at the stellar surface at the age of the Hyades. Among some effects suggested one can quote an unusual mass loss on the main sequence (Boesgaard et al. 1977), and rotational induced mixing (Charbonnel & Vauclair 1992; Charbonnel & Talon 1999).
The behaviour observed in our most massive stars which have not yet reached the beginning of dilution explains the very low lithium content of most of the massive subgiants which cannot be accounted for by dilution alone. The underlying destruction mechanism should also be responsible for the lithium behaviour observed in the Hyades giants (Boesgaard et al. 1977) which have masses higher than and thus correspond to the data we show in Fig. 5 d. This process should however preserve the boron abundance, which is in agreement with the standard dilution predictions in the underabundant Li giants of the Hyades, as shown recently by Duncan et al. (1998). Observations of boron in our sample stars have to be done to confirm the similarity between field and cluster evolved stars.
Finally, we note that Non-LTE effects alone can not explain the very low Li abundance derived for several massive stars, as proposed by Duncan et al. (1998). Indeed, Carlsson et al. (1994) have showed that Non-LTE effects can reach up to dex only for such cool stars. This correction factor is too low to reconcile the derived Li abundances with the standard dilution predictions. Therefore, this confirms that these stars are indeed Li-poor as well as the hotter ones for which only upper limits have been derived (Non-LTE effects are much smaller for such hotter stars).
© European Southern Observatory (ESO) 2000
Online publication: June 5, 2000