New PARSEC evolutionary tracks of massive stars at low metallicity: testing canonical stellar evolution in nearby star-forming dwarf galaxies

In this thesis I present new models of massive stars for low metallicity computed with the new code PARSEC: Padova TRieste Stellar Evolution Code. An updated version of this code, known as PARSEC V1.1, has already been released but, in the published version, the range of initial masses does not go beyond 12M . Recently, the PARSEC library of stellar evolutionary tracks has been extended with the inclusion of new models of massive stars, from 14M to 350M . The input physics is the same used in the PARSEC V1.1 version, but for the mass-loss rate which is included by considering the most recent updates in literature. In this thesis I perform a thorough comparison of the new models of massive stars with existing observational data. This is a critical and necessary step because the stellar evolution theory still contains some parameters that need to be tuned to well observed stellar samples. For this reason I focus on the low metallicity environment, Z=0.001, Z=0.004 and Z=0.0005, for which the metal poor dwarf irregular star forming galaxies, Sextans A, WLM, NGC 6822 and SagDIG, provide simple but powerful workbenches: the color-magnitude diagrams (CMDs) of their young stellar populations. From the simulations of these CMDs I draw the following conclusions: While the new models reproduce fairly well the observed CMDs of Sextans A, WLM and NGC 6822, a detailed analysis of the stellar color distributions indicates that the predicted blue loop is not hot enough in models that assume the canonical extent of overshooting from the convective regions. In the framework of a mild extended mixing during central hydrogen burning, the only way to reconcile the discrepancy is to enhance the overshooting at the base of the convective envelope (EO) during the first dredge-UP. Reproducing the features of the observed CMDs with standard values of envelope overshooting would require a metallicity significantly lower than the values measured in these galaxies. I find that the mixing scales required to reproduce the observed loops are large, EO=2HP or EO=4HP . These values are definitely larger than those derived from, e.g., the observed location of the RGB bump in low mass stars. This effect, if confirmed, would imply a strong dependence of the mixing scale below the formal Schwarzschild border, on the stellar mass or luminosity. Other quantities, such as the star formation rate and the initial mass function, are only slightly sensitive to this effect. This result is further validated in the comparison with the observed CMD of SagDIG at lower metal- licity, where I find an overshooting scale EO=2HP to best reproduce the observed loops. I also discuss the dependence of the blue loop extension on the adopted instability criterion and find that, contrary ito what stated in literature, the Schwarzschild criterion, instead of the Ledoux criterion, favours the development of blue loops. Other factors that could affect the CMD comparisons such as differential internal extinction or the presence of binary systems are found to have negligible effects on the results. I thus confirm that, in presence of core overshooting during the H-burning phase, a large envelope overshooting is needed to reproduce the main features of the central He-burning phase of intermediate- and high-mass stars.