Next: The Concept of Radiation Up: Realistic Models For Expanding Previous: Realistic Models For Expanding
Detailed spectrum syntheses which fully extract the quantitative information contained in the spectral domains of the EUV, the UV, and the optical, are still one of the main objectives in hot-star physics. In particular, the ultraviolet spectra of hot stars, with their hundreds of spectral lines formed far out in the wind at high outflow velocities (e.g.,the resonance lines of the CNO elements; see Fig. 1) as well as in deeper atmospheric layers at velocities comparable to the sound velocity (e.g.,the strongly wind contaminated lines mainly of the iron-group elements; see Fig. 1), can provide important information about stellar parameters, wind properties, and abundances. The aim is to determine these quantities via the application of model atmospheres and to calculate stellar energy distributions and, hence, ionizing fluxes which are of further importance in the formation of emission-line spectra of gaseous nebulae (e.g.,HIIregions and planetary nebulae). However, due to the complexity of the theory of hot-star atmospheres, which requires sophisticated non-LTE calculations and a hydrodynamical treatment that includes the effects of expanding atmospheres, this is a long-standing problem which was only partially - or on the basis of questionable approximations - realized in the past. (A large compilation of references relating to this was recently given by Kudritzki 1997.) Realistic models are now required, by which we mean those which fully reproduce the observed high-resolution spectra of hot stars via consistently calculated synthetic spectra, and which reproduce the energy distribution in the EUV in a detailed and correct way.
Although our theoretical tools are still in an exploratory stage we have already taken first real steps in this direction. So far we have constructed hydrodynamic atmospheric models for seven O stars: the O4I(f) star Puppis (cf. Pauldrach et al. 1994a) and the O3If* star HD 93129A (cf. Taresch et al. 1997) in the Galaxy; the O3If/WN star Mk42 (cf. Pauldrach et al. 1994a), the O3III(f*) star Sk -68deg 137, and the O4If+ star Sk -67deg 166 in the LMC; and the O3III(f*) star NGC 346 No.3 and the O6V star AV 243 in the SMC (for the last four objects cf. Haser et al. 1997).
In the analyses, UV spectra (observed with IUE, ORFEUS, and HST), together with some optical lines, have been investigated, where the quantitative spectroscopic studies, based on our non-LTE hydrodynamic model atmosphere code described in Pauldrach et al. (1994a, 1994b), resulted in the determination of abundances, log g, and Teff. Fig. 2 shows as an example the final fit of the comparison performed for HD 93129A. Although the agreement is quite convincing, and the stellar and wind parameters determined coincide with those obtained from unified atmospheric non-LTE analyses of H and He lines in the visual region (cf. Puls et al. 1996, XXX1997), we still observe deviations for some individual lines (for instance, the subordinate line of OVat 1371Å). Having found from this kind of model calculation that the behaviour of a number of spectral lines is critically dependent on a detailed and consistent description of line blocking and line blanketing, we attribute these deviations to the restrictive approximations which we used for this important part of the method. Hence, the next quantitative approach requires a realistic and consistent description of line blocking and blanketing. In Section 3 we will present a method which fulfills these requirements and, in connection, we will report on problems we have been faced with in calculating the ionizing fluxes of O stars correctly. As a basis for this discussion the general concept and the status of our treatment of hydrodynamic expanding atmospheres is summarized in Section 2.