Astron. Astrophys. 322, 256-265 (1997)

1. Introduction

The analysis of white dwarf and subdwarf B (sdB) star atmospheres is traditionally the domain of LTE techniques. Usually, deviations from LTE are kept small by the high densities of the atmospheres. NLTE calculations, and comparisons with LTE models in the white dwarf and subdwarf O star regions were performed by Kudritzki (1976), Wesemael et al. (1980), and Wesemael (1981). Kudritzki (1976) tried to define a LTE domain in which the stellar photospheres should be well represented by LTE models. According to his Fig. 3 LTE is valid for white dwarfs with K (but see below). Today LTE calculations are used for the analysis of even the hottest white dwarfs.

In this paper a detailed discussion of the influence of NLTE on white dwarf atmospheres is given. It is about time to revisit this question for several reasons. First of all, the spectroscopic material available today for many white dwarfs has now reached a quality which was Utopian some twenty years ago. For instance, Kudritzki's criterion for the importance of NLTE was a deviation of 15% in the equivalent width of H . This was certainly a realistic error estimate for typical photographic spectra, but today we can do much better! Moreover, Kudritzki (1976) restricted the discussion to atmospheres with and 1.0, but some effects are present only for the generally very low helium abundances in the nearly pure hydrogen atmospheres of DA white dwarfs (see Sect. 4.1).

The performance of NLTE model atmosphere calculations have increased dramatically, too. The calculations of Kudritzki (1976), Wesemael et al. (1980), and Wesemael (1981) are all based on the pioneering complete linearization (CL) method developed by Auer & Mihalas (1969). This method limits the calculations to only a few NLTE levels ( 15) and line transitions due to its numerical properties. New methods have been developed in the meantime, which are much more powerful tools for the calculation of NLTE atmospheres. Important are the so-called accelerated lambda iteration (ALI) method (Cannon 1973; Hamann 1985; Werner & Husfeld 1985) and the hybrid CL/ALI approach (Hubeny & Lanz 1995). Both methods now allow the consistent treatment of more than a hundred atomic levels and many hundreds of lines. Even the inclusion of highly complex atoms like iron with millions of lines is now possible through a statistical approach (Anderson 1989; Dreizler & Werner 1993; Hubeny & Lanz 1995). NLTE atmospheres have now reached a level of sophistication which is comparable to that of present day LTE atmospheres. However, the computational effort is still much larger than for LTE calculations. Lanz & Hubeny (1995) compared some selected metal line blanketed NLTE atmospheres of white dwarfs with simple hydrogen LTE models without any line blanketing. Koester & Herrero (1988) showed that line blanketing effects dominate over NLTE effects for white dwarfs at least at K.

During the last years hot white dwarfs have become a field of intense research. LTE model atmospheres are widely used for the analysis of hot white dwarfs (spectral types DA, DAO, and DO) by, e.g., Wesemael et al. (1985), Kidder (1991), Bergeron et al. (1994), and Finley et al. (1996) as well as for sdB stars by Heber (1986) and Saffer et al. (1994). NLTE analyses have been performed for hydrogen-rich white dwarf central stars of old planetary nebulae by Napiwotzki (1993a, 1993b, 1995a), and for hot helium-rich DO white dwarfs by Werner et al. (1994, 1995) and Dreizler & Werner (1996).

The aim of this article is a comparison of hydrogen and helium composed NLTE and LTE models to detect the influence of NLTE effects on important spectroscopic features. The effects of deviations from LTE on the EUV flux of hot DA white dwarfs were already discussed in Napiwotzki et al. (1993). In Sect. 2 our model calculations and the construction of consistent NLTE and LTE models are described. After some general considerations in Sect. 3 the hydrogen-rich white dwarfs are discussed in Sect. 4. Special emphasis is laid on the LTE results of Bergeron et al. (1994) that Balmer lines of DA stars are strongly influenced by small, undetectable traces of helium in the atmosphere. We will discuss this effect and demonstrate that it is caused by the assumption of LTE and disappears in NLTE calculations. The investigation is extended to the Balmer and helium lines of DAO stars. In Sect. 5 we will show that deviations from LTE are important for (nearly) all helium-rich DO stars. The effects of NLTE and metal line blanketing on the spectral lines of the subdwarf B stars is discussed in Sect. 6. Finally, the use of NLTE line formation calculations on LTE atmospheres is checked in Sect. 7. A short summary of our results is given in Sect. 8.

© European Southern Observatory (ESO) 1997

Online publication: June 30, 1998
helpdesk.link@springer.de