Disentangling the metallicity and star formation history of H ii galaxies through tailor-made models

Abstract

We present a self-consistent study of the stellar populations and the ionized gas in a sample of 10 H ii galaxies with, at least, four measured electron temperatures and a precise determination of ionic abundances following the 'direct method'. We fitted the spectral energy distribution of the galaxies using the program starlight and starburst99 libraries in order to quantify the contribution of the underlying stellar population to the equivalent width of Hβ [EW(Hβ)], which amounts to about 10 per cent for most of the objects. We then studied the Wolf-Rayet (WR) stellar populations detected in seven of the galaxies. The presence of these populations and the EW(Hβ) values, once corrected for the continuum contribution from underlying stars and ultravoilet dust absorption, indicate that the ionizing stellar populations were created following a continuous star formation episode of 10 Myr duration, hence WR stars may be present in all of objects even if they are not detected in some of them.The derived stellar features, the number of ionizing photons and the relative intensities of the most prominent emission lines were used as input parameters to compute tailored models with the photoionization code cloudy. Our models are able to adequately reproduce the thermal and ionization structure of these galaxies as deduced from their collisionally excited emission lines. This indicates that ionic abundances can be derived following the 'direct method' if the thermal structure of the ionized gas is well traced, hence no abundance discrepancy factors are implied for this kind of objects. Only the electron temperature of S+ is overestimated by the models, with the corresponding underestimate of its abundance, pointing to the possible presence of outer shells of diffuse gas in these objects that have not been taken into account in our models. This kind of geometrical effects can affect the determination of the equivalent effective temperature of the ionizing cluster using calibrators which depend on low-excitation emission lines.