FORMATION AND EVOLUTION OF MASSIVE STARS

Pre-main-sequence evolutionary tracks of massive stars have been calcu lated under the canonical theory and the mass accretion paradigm. The canonical tracks for 15, 20, 25, 30, 45, and 60 M. stars have been cal culated with both Cox-Stewart (1970) and the Rogers and Iglesias (1992 a, b) opacities. Stellar models less massive than about 30 M. develop convective cores only during the onset of central hydrogen burning. Be cause of the increasing importance of radiation pressure, stars more m assive than 30 M. develop convective cores before the onset of hydroge n burning. A generalized Naur-Osterbrock ( 1953) criterion has been us ed to understand the development of the convective core. It is estimat ed that stars more massive than approximate to 95 M. evolving canonica lly to the main sequence will have a convective core throughout the wh ole pre-main-sequence phase. The canonical 30 M. model has been evolve d through the main-sequence phase using the mass-loss algorithm of de Jager, Nieuwenhuijzen, and van der Hucht( 1988). The pre-main-sequence evolution has also been calculated under the accretion paradigm: begi nning with a collapsed core, mass was accreted until a final stellar m ass was reached. The initial hydrodynamic evolution of the protostar w as not followed. The calculation was begun with a 1 M. core, and a con stant accretion rate of 10(-5) M. yr(-1) was used. The initial model i s fully convective and burning deuterium in the core. The accretion tr ack intersected the canonical main sequence at M = 8.5 M.. For masses from 8.5 to approximate to 15 M., the accretion and canonical zero-age main-sequence models were essentially the same. Beyond approximate to 17 M., the accretion models became increasingly more luminous and coo ler with smaller convective cores. The main-sequence evolution of the 30 M. star, begun with an accretion model, occurs at a slightly lower luminosity and a lower effective temperature compared with canonical e volution, The accretion sequence was calculated up to the final mass o f 45 M.. The locus where massive stars first become optically visible, the upper stellar birth line, was determined from these models. It is pointed out that the upper mass limit of stable stars must be determi ned by the formation environment of the star rather than the onset of nuclearly energized pulsation instability (Stothers 1992), since the i ncreasingly large chemical inhomogeneity that develops as a massive st ar accretes matter evolving along the upper stellar birth line will st abilize nuclearly driven pulsations.