2000A&A...362...53S

We present new spectroscopic observations of Mkn 309, a starburst galaxy with one of the largest WR populations known. A highly super solar metallicity of 12+log(O/H)∼9.3-9.4 is derived. Using additional objects from Guseva et al. (2000) we analyse a sample of five metal-rich ([O/H]>0) WR galaxies with the main goal of constraining the basic properties of the massive star populations (IMF slope, M_up) and the star formation history (age, burst duration) of these objects by quantitative comparisons with evolutionary synthesis models. The following main results are obtained: (o) The observations are well explained by extended bursts of star formation with durations Δt∼4-10Myr seen at ages of 7-15Myr or a superposition of several bursts with age differences of ∼4-10Myr including a young (≲5Myr) burst. This naturally explains both the observed WR populations (including WN and WC stars) and the presence of red supergiants. The burst durations, somewhat longer compared to those derived in other WR galaxies using the same models (Schaerer et al., 1999A&A...341..399S), are plausible in view of the physical sizes of the observed regions and the nature and morphology of our objects (nuclear starbursts), and pose no fundamental physical problem. (o) The SEDs in the optical range are very well reproduced for all objects, provided the stellar light suffers from a smaller extinction than that of the gas (derived from the Balmer decrement). This confirms earlier findings from studies combining UV-optical data of other starburst galaxies. (o) All the considered observational constraints are compatible with a Salpeter IMF extending to masses M_up>40M_☉. Adopting a conservative approach we derive a lower limit of M_up>30M_☉ for the Salpeter IMF. From more realistic assumptions on the metallicity and SF history we favour a lower limit M_up>30-40M_☉, which is also in agreement with Hβ equivalent width measurements of metal-rich H II regions in spiral galaxies indicating an upper mass cut-off of at least ∼35-50M_☉. Steep IMF slopes (α>3.3) are very unlikely. The uncertainties of our results are discussed. We compare our findings to other work on massive star populations and the IMF in similar environments. We stress the importance of direct analysis of stellar populations compared to other indirect methods based on properties of ionized gas to constrain the IMF in metal-rich starbursts.