Evolution of Massive Stars into Luminous Blue Variables and Wolf-Rayet Stars for a Range of Metallicities: Theory versus Observation
Evolutionary tracks for 30-90 Msun stars, computed along the lines described in our recent series of papers, suggest that luminous blue variables (LBVs) are the blue remnants of a prior phase of heavy mass loss. They are now intermittently experiencing ionization-induced dynamical instability within their hydrogen-poor envelopes, while their massive helium cores are in an advanced stage of central helium burning. In the H-R diagram, our new evolutionary models successfully explain the high- and low- temperature limits, the low-luminosity limit, and the relative luminosity function of the observed LBVs at quiescence which have log (L/Lsun) < 6.3. Metal-poor LBVs are predicted (correctly, in the only case known) to be significantly cooler than LBVs with normal metallicities. In the (mass, luminosity) plane, the models and observed stars agree closely, and show no discernible metals dependence. Predicted values of the surface hydrogen abundance and of the expelled mass during the LBV phase match reasonably well the available observational data, which, however, are still very crude. The observed cycles of mass loss are predicted correctly. Pre-LBV and post-LBV stars can be identified with the hydrogen-poor and hydrogen-free WN stars, respectively. Post-red-supergiant lifetimes, derived from published star counts and kinematical ages of nebulae around the WN stars and LBVs, are used to infer an approximate mean value of the rate of stellar wind mass loss from luminous red supergiants. This parameter is crucial for understanding stellar evolution at log (L/Lsun) < 5.8. Stars more luminous may never become red, and may spend all, or nearly all, of their post-main-sequence lives as blue supergiants. Our theory of LBVs makes a number of other predictions that have not yet been tested. A critical comparison with the very different theory of Langer et al. is presented.