Spectroscopic evolution of massive stars on the main sequence

Context. The evolution of massive stars depends on several parameters, and the relation between different morphological types is not fully constrained. Aims. We aim to provide an observational view of evolutionary models in the Hertzsprung–Russell diagram, on the main sequence. This view should help compare observations and model predictions. Methods. We first computed evolutionary models with the code STAREVOL for initial masses between 15 and 100 M⊙. We subsequently calculated atmosphere models at specific points along the evolutionary tracks, using the code CMFGEN. Synthetic spectra obtained in this way were classified as if they were observational data: we assigned them a spectral type and a luminosity class. We tested our spectral classification by comparison to observed spectra of various stars with different spectral types. We also compared our results with empirical data of a large number of OB stars. Results. We obtain spectroscopic sequences along evolutionary tracks. In our computations, the earliest O stars (O2-3.5) appear only above ~50 M⊙. For later spectral types, a similar mass limit exists, but is lower. A luminosity class V does not correspond to the entire main sequence. This only holds for the 15 M⊙ track. As mass increases, a larger portion of the main sequence is spent in luminosity class III. Above 50 M⊙, supergiants appear before the end of core-hydrogen burning. Dwarf stars (luminosity class V) do not occur on the zero-age main sequence above 80 M⊙. Consequently, the distribution of luminosity class V in the HR diagram is not a diagnostic of the length of the main sequence (above 15 M⊙) and cannot be used to constrain the size of the convective core. The distribution of dwarfs and giants in the HR diagram that results from our calculations agrees well with the location of stars analyzed by means of quantitative spectroscopy. For supergiants, there is a slight discrepancy in the sense that luminosity class I is observed slightly earlier (i.e., at higher Teff) than our predictions. This is mainly due to wind densities that affect the luminosity class diagnostic lines. We predict an upper mass limit for dwarf stars (~60 M⊙) that is found consistent with the rarity of O2V stars in the Galaxy. Stars with WNh spectral type are not predicted by our models. Stronger winds are required to produce the characteristic emission lines of these objects