Evolution of chemical abundances in massive stars. I. OB stars, Hubble-Sandage variables and Wolf-Rayet stars. Changes at stellar surfaces and galactic enrichment by stellar winds.

Evolutionary models with mass loss and a detailed study of changes in the abundances of ¹H, ³He, ⁴He, ¹²C, ¹³C, ¹⁴N, ¹⁵N, ¹⁶O, ¹⁷O, ¹⁸O, ²⁰Ne, ²²Ne, ²⁵Mg, and ²⁶Mg are made for initial stellar masses of 120, 85, and 60 M_sun. The evolution has been followed through stages corresponding to OB stars, blue super- giants, Hubble-Sandage variables and various Wolf-Rayet stages (WNL, WNE, WC, WO).

The evolutionary status of the Hubble-Sandage variables is examined. In the case of η Car, the C/N and 0/N ratios observed by Davidson et al. (1982) agree remarkably well with the theoretical abundance ratios for a massive post-main sequence supergiant having experienced strong mass loss. The origin of the variability of the Hubble-Sandage variables is discussed in relation with the instability limit (de Jager's limit) for deep turbulent external convective zones in supergiants. This limit may also be responsible for the fact that the brightest red supergiants are 2 mag fainter in M_bol than the brightest OB stars.

A theoretical mass-luminosity relation is found for the blue, non-degenerate stellar cores left after the loss of the envelopes by stellar winds and which are likely to correspond to WR stars.

Due to the removal of the outer layers by mass loss, matter produced by the CNO tri-cycle is revealed at stellar surfaces in OB stars, supergiants and WN stars. ³He, ¹⁵N, ¹⁸O disappear. The C/N ratio suddenly changes from about 4.1 to 0.03 (in mass), and 0/N from 9.1 to less than 0.1. ¹³C keeps a factor of 3.3 lower than ¹²C. The abundance of ¹⁷O strongly rises and becomes nearly equal to that of ¹⁶O. A very large discontinuity (whose physical origin is explained) marks the appearance of the various products of He-burning at the surface, what is likely to correspond to the beginning of the WC stage. We note then a very steep disappearance of ¹³C and ¹⁴N, a very temporary peak of ¹⁸O, and above all a vertiginous rise by more than 2 orders of magnitude of ¹²C, ¹⁶O, and ²²Ne. The abundances of ²⁵Mg and ²⁶Mg also rise strongly, particularly in the most massive WC stars, where s-elements are therefore to be expected.

Comparisons are made between the theoretical C/He, N/He, and C/N ratios and those observed by Smith and Willis (1982) and by Nugis (1982) for WNL, WNE, and WC stars. The general agreement strongly supports the advanced evolutionary stage of WR stars as left-over cores resulting from the peeling of massive stars by stellar winds. If confirmed, the presence of ¹⁴N in WC stars may bring some indication on mixing. Observations of other interesting elements are desirable. The contributions of the winds of WR star to the galactic enrichment in various elements and isotopes is estimated. The main part of the ²²Ne in the Galaxy probably results from the winds of WC stars, which also strongly contribute to the galactic enrichment in ⁴He and ¹²C and moderately to those in ²⁶Mg, ²⁵Mg, and ¹⁶O. WN stars are likely to be major contributors to the ¹⁷O-enrichment with a modest yield in ¹⁴N.