Empirical mass-loss rates and clumping properties of O-type stars in the Large Magellanic Cloud

Context. The nature of mass-loss in massive stars is one of the most important and difficult to constrain processes in the evolution of massive stars. The largest observational uncertainties are related to the influence of metallicity and wind structure with optically thick clumps. Aims. We aim to constrain the wind parameters of sample of 18 O-type stars in the LMC, through analysis with stellar atmosphere and wind models including the effects of optically thick clumping. This will allow us to determine the most accurate spectroscopic mass-loss and wind structure properties of massive stars at sub-solar metallicity to date. This will allow us to gain insight into the impact of metallicity on massive stellar winds. Methods. Combining high signal to noise (S/N) ratio observations in the ultraviolet and optical wavelength ranges gives us access to diagnostics of multiple different ongoing physical processes in the stellar wind. We produce synthetic spectra using the stellar atmosphere modelling code FASTWIND, and reproduce the observed spectra using a genetic algorithm based fitting technique to optimise the input parameters. Results. We empirically constrain 15 physical parameters associated with the stellar and wind properties of O-type stars from the dwarf, giant and supergiant luminosity classes. These include temperature, surface gravity, surface abundances, rotation, macroturbulence and wind parameters. Conclusions. We find, on average, mass-loss rates a factor of 4–5 lower than those from theoretical predictions commonly used in stellar-evolution calculations, but in good agreement with more recent theoretical predictions. In the ‘weak-wind’ regime we find massloss rates orders of magnitude below any theoretical predictions. We find a positive correlation of clumping factors with effective temperature with an average fcl = 14 ± 8 for the full sample. It is clear that there is a difference in the porosity of the wind in velocity space, and interclump density, above and below a temperature of roughly 38 kK. Above 38 kK an average 46 ± 24% of the wind velocity span is covered by clumps and the interclump density is 10–30% of the mean wind. Below an effective temperature of roughly 38 kK there must be additional light leakage for supergiants. For dwarf stars at low temperatures there is a statistical preference for very low clump velocity spans, however it is unclear if this can be physically motivated as there are no clearly observable wind signatures in UV diagnostics.