arXiv2002
We use a Sobolev with Exact Integration model to analyze the winds lines of 25 LMC O stars. The data include FUSE profiles of C III, N III, S IV, P V, S VI, and O VI and IUE or HST data for Si IV, C IV, and N V. Several of the FUSE lines are unsaturated, so meaningful optical depths (equivalently, mass loss rate times ionization fractions), as a function of wind velocity can be determined. Ratios of these quantities give the relative ionization as a function of velocity and demonstrate that, except for O VI in all stars and S VI in the later stars, the wind ionization shifts toward lower stages at higher velocity. Because O VI and S VI do not behave like the other ions, they must be produced by a different mechanism. Using mass-loss rates determined from the Vink et al. relationships, we derive mean ionization fractions. Because these are all less than one, the derived mass loss rates cannot be too small. However, the ion fractions for P V (expected to be dominant in some winds), never exceed 0.20. This implies that either the calculated mass loss rates or the assumed P abundances are too large, or the winds are strongly clumped. We examine correlations between the mean ion fractions and stellar parameters, and find two significant relationships. First, as expected, the mean ionization fraction of lower ions decreases with increasing temperature. Second, the mean ionization fraction of S VI in the latest stars and O VI in all stars increases with terminal velocity, re-affirming Cassinelli and Olson's conjecture that O VI is produced non-radiatively. Finally, we discuss peculiar aspects of three stars, BI 272, BI 208, and Sk-67 166.