1999A&A...341..399S

We report the detection of WC stars in five Wolf-Rayet (W-R) galaxies: He 2-10, NGC 3049, NGC 3125, NGC 5253 and Tol 89. The faint broad C IV λ5808 line requires sufficiently high S/N (>40) to be detected explaining the non-detection of this WC feature in previous observations. From the measurement of W-R emission lines (N III λ4640 + C III λ4650, He II λ4686, and C IV λ5808), we conclude that all W-R regions contain a mixed population of WNL, and early WC stars. The exception is the high-metallicity region NGC 3049 where late WC stars prevail. A spatial offset between the multiple peaks of the nebular emission and the stellar light in He 2-10 and Tol 89 is observed. These nebular emission structures are likely due to the existence of bubbles and loops, owing to the injection of mechanical energy in the ISM through the W-R winds and/or supernovae. Due to age differences and likely smaller energy deposition the structures around the W-R regions are possibly smaller than the ones predominantly energized by SNe. The spatial distribution of W-R stars closely follows the stellar continuum with no significant distinction between WN and WC stars. From the luminosity of the W-R signatures we have estimated the absolute number of W-R stars of the different subtypes. The WC/WN number ratios have typical values between 0.2-0.4, and show no clear trend with metallicity. For low-metallicity objects (Z∼1/5Z_☉), these values are larger than the observed WC/WN ratios in Local Group objects, but are compatible with expectations for star forming events with short duration if stellar evolution models with high mass loss are used. We derive ages for the starburst regions in the range of 3 to 6Myr and confirm that the burst duration must not exceed ∼2-4Myr to account for the high population of W-R stars observed in starburst regions, even if emission line stars similar to those observed in R136 and NGC 3603 are common in starbursts. Within the uncertainties the majority of the observed quantities is reasonably reproduced by models with a Salpeter IMF. Although some W-R lines in few regions are stronger than predicted by the models no clear case requiring a significantly flatter IMF is found. IMF slopes much steeper than Salpeter may, however, not be compatible with our data.