Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 351, 1075-1086 (1999)

Previous Section Next Section Title Page Table of Contents

7. Ionization models

In principle, O is the only element for which we can observe in the optical range all the ionization states expected to be present inside HII regions: O+ and O++ . These nebulae are not expected to contain a significant number of photons with enough energy to ionize O++, a fact confirmed by the absence from their spectra of HeII lines (like HeII  [FORMULA]), since He ++ and O3+ require similar energies to be produced (54.4 and 54.9 eV, respectively).

To derive the total abundances for the other elements, it is necessary to estimate the contribution of the unobserved ions. These estimates can be made using ionization models, but the results of these models depend on poorly known parameters like the characteristics of the ionizing radiation field and the nebular structure. This dependence can be minimized by using model grids (where each model is characterized by the number of ionizing stars, the stellar effective temperatures, [FORMULA], and the density and metallicity of the gas) to study the relative variations of the ionization fractions of the element of interest and a reference element with two observed ions whose ionization potentials are similar to those of the first element. This can be a good approximation, but it must be taken into account that the ionization equilibrium for the different ions is affected by factors which are not always well known, like the detailed dependence on energy of the ionization and recombination coefficients and the effects of charge-exchange reactions.

Three series of ionization models have been considered: those from Rubin (1985), Stasinska (1990) and Gruenwald & Viegas (1992). None of these authors considers the effects of dust on his calculations, but dust absorption cross-sections are expected to behave in a hydrogenic fashion in such a way that dust will not significantly alter the ionization balance between the different ions (Mathis 1986). The ionization models of the different authors can differ in the model atmospheres used, the values adopted for the atomic data, the kind of atomic processes considered, etc. Therefore, when the results of these models differ from each other or from the observational data, it is difficult to infer the reasons and no attempt will be made to do so in this paper. However, there is one characteristic of the work of Gruenwald & Viegas (1992) that can be significant: their results are presented for several lines of sight across each model, a better approach to resolved sources than the models just giving the volume-averaged ionic fractions.

The models selected from Gruenwald & Viegas (1992) are those with solar metallicities [FORMULA] or [FORMULA]; these models are ionized by a single star with [FORMULA] or [FORMULA] and have densities [FORMULA], 100, or 1000 cm-3. The models selected from Stasinska (1990) have metallicities [FORMULA] or [FORMULA], are ionized by 1, 102 or 104 stars with [FORMULA] and have densities [FORMULA] or 1000 cm-3. Two series of models have been selected from Rubin (1985), both with [FORMULA] Lyman-continuum photons per second. The first series is characterized by solar abundances, [FORMULA] and [FORMULA], 1000, or [FORMULA]. The second series is characterized by `nebular' abundances, [FORMULA] or [FORMULA] and mean densities [FORMULA], 1000, or [FORMULA], but including clumps denser by a factor of 10.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1999

Online publication: November 16, 1999