Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 359, 1085-1106 (2000)

Previous Section Next Section Title Page Table of Contents

6. Summary and conclusions

An extensive model atom for non-LTE line formation calculations for neutral oxygen in BA-type stars has been developed, based on the most accurate atomic data presently available. The weak OI lines in the visible prove to be highly reliable abundance indicators at all luminosities while the strong near-infrared lines are tracers for the atmospheric structure. They are of only limited use for deriving accurate abundances in supergiants with present atmospheric models.

We have provided a set of non-LTE abundance corrections for the diagnostic OI lines, applicable over a large range of stellar parameters in BA-type stars. Along the main sequence non-LTE effects on OI are negligible for the weak lines, are gradually increasing with luminosity to reach significant values in Ib supergiants and are essential for the analysis of the most luminous objects close to the Eddington limit where non-LTE corrections in excess of 0.5 dex are predicted in some cases. High accuracy in the determination of oxygen abundances can be achieved, within expected uncertainties of [FORMULA]0.10/0.15/0.20 dex for high quality observations of main sequence/Ib/Ia objects including systematic errors as confirmed by the analysis of our three test stars Vega, [FORMULA] Leo and HD 92207. We have derived the following oxygen abundances:


Our results support the findings of a slight oxygen underabundance for Vega in previous studies. [FORMULA] Leo and HD 92207 turn out to show a slightly less than solar oxygen abundance.

The theoretical interpretation of the strong near-infrared lines still poses some problems as they are formed over a large range in the geometrical extent of the stellar atmosphere. Present atmospheric models fail to describe the actual physical conditions realistically, especially in supergiants. A consistent treatment of non-LTE effects and (non-LTE) line blanketing on the atmospheric structure is still lacking.

Nevertheless, we have shown that - apart from weaknesses within our model atom which we cannot completely exclude as for the majority of the highly important collisional processes only approximate data are available - certain simple assumptions are sufficient to remove most of the discrepancies between observation and theory. Significant improvement can be achieved by invoking a depth dependent microturbulence parameter, having a large impact on the saturated lines but leaving the weak lines almost unaffected. For Vega perfect consistency for all oxygen lines has been achieved by this. The technical point of the inclusion of microturbulence for additional line broadening in the statistical equilibrium calculations has also been investigated. In contrast to studies for other elements, no effect on line profiles and equivalent widths is found in our OI non-LTE computations which can be explained from the interplay between occupation numbers and formation depths for varying microturbulence.

The observed asymmetric line profiles with blueward shifted extra absorption in extreme supergiants demand consideration of the hydrodynamic velocity field (and spherical extension) for the spectrum synthesis computations. Some improvement in removing the discrepancies between theory and observation can be achieved by accounting for the macroscopic velocity field where, in the hydrostatic case, the depth dependent microturbulence has to be invoked. This insight cannot be gained from the simple equivalent-width studies performed in the past; detailed line profile analyses have to be carried out. The strong metal lines in the A-type supergiants offer a unique opportunity to study the velocity field at the base of the stellar wind.

A final statement on the line formation of the strong near-infrared OI lines cannot be presented here as from the theoretical standpoint some effort has to be made to describe the physical conditions in a supergiant's atmosphere realistically. But with this study we hope to focus the attention on the importance of non-LTE effects, micro- and macroscopic velocity fields for the line formation.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: July 13, 2000