2. The physical ingredients of the stellar models
The evolutionary models are computed with the same physical ingredients as in Meynet et al. (1994). However, the adopted nuclear reaction network is extended, especially in order to include the reactions involved in the production and destruction of (see below).
The differences with respect to the computations of Meynet & Arnould (1993) are twofold:
1) A more extended range of initial masses (from 25 to 120 ) and metallicities (, 0.02 and 0.04) is explored;
2) The present grid of models is computed with the mass loss rates adopted by Meynet et al. (1994), which are twice as large as the values of recommended by de Jager et al. (1988) and Conti (1988) for the pre-WR and WNL phases. This mass loss rate prescription enables to account for the observed variations of WR populations in different environments (Maeder & Meynet 1994).
The metallicity dependence of the mass loss rates during the pre-WR phases is adopted from previous works (e.g. Meynet et al. 1994). More specifically, scales with metallicity Z according to , where is the solar metallicity. This scaling is deduced from stellar wind models (cf. Kudritzki et al. 1987, 1991).
Let us finally add that the models are computed with a moderate core overshooting (, where d is the overshooting distance and the pressure scale height at the boundary of the classical core).
2.1. The thermonuclear production and destruction paths
The CNO mode of H-burning is responsible for the production and destruction of through the reaction chain
The adopted rate is the geometrical mean of the lower and upper limits to that rate proposed by Kious (1990).
Fluorine can also be produced and destroyed during He-burning through the chains (see also Meynet & Arnould 1993)
The synthesis of thus requires the availability of neutrons and protons. They are mainly produced by the reactions and .
The first chain of transformation of into mentioned above is by far the most important in the conditions of relevance in this work, where the -decay lifetime () of is much shorter than its lifetime () or () against () or () reactions. For example, is a few hours only at the center of a 60 model at the beginning of core He-burning, while the corresponding () and () amount to about 1 400 and 18 000 years, respectively.
The NACRE compilation of reaction rates (Angulo et al. 1999) was not available yet at the time of completion of the calculations reported here. This is why most of the necessary nuclear data are taken from Caughlan & Fowler (1988). There are some exceptions to this rule. In particular, the -capture rate is taken from Descouvemont (1987), whose theoretical prediction of an increase of the astrophysical S-factor at low energies is confirmed experimentally (see NACRE). The rate is taken from Brehm et al. (1988). It is a factor of two lower than the one proposed by Koehler and O'Brien (1989), and leads consequently to a lower limit of the calculated yields.
© European Southern Observatory (ESO) 2000
Online publication: March 17, 2000