White dwarfs are separated in two distinct spectroscopic sequences, a hydrogen- rich and a helium-rich one. The hydrogen white dwarfs can be subdivided into two classes, the DA stars with no spectroscopically detectable helium and the DAO stars, in the spectra of which helium lines are visible. For hot DA's helium abundances have sometimes been derived under the assumption that this element provided most of the EUV opacity. However, Vennes (1992) and Koester (1989) have shown that the EUV flux deficiencies typical for DA's with can be explained either with the presence of heavy elements or by stratified H/He atmospheres. The presence of helium in the DAO's, which are with hotter than most of the DA's, is still unexplained. Vennes et al. (1988) have shown that the observed helium abundances, which are typically , cannot be explained by radiative levitation. The predicted abundances are too low by at least a factor of ten. Fontaine & Wesemael (1987) suggested that most white dwarfs enter the cooling sequence as helium-rich objects with some traces of hydrogen. In the course of time gravitational settling causes hydrogen to float up. This led to the hypothesis of thin hydrogen layers (Vennes & Fontaine, 1992), according to which the DAO's have stratified atmospheres characterized by an equilibrium between gravitational settling and ordinary diffusion. This scenario requires ultrathin hydrogen layers with masses smaller than . However, Napiwotzki & Schönberner (1993) and Bergeron et al. (1994) concluded from an analysis of helium line profiles, that the atmospheres of most DAO's are more likely chemically homogeneous rather than stratified. Mass loss has been invoked as a possible explanation. The question "thick or thin hydrogen layers in white dwarfs" is rediscussed in Fontaine & Wesemael (1997) and Shipman (1997).
A similar problem occurs in hot subdwarfs with and (for a review see Heber, 1992 and Saffer & Liebert, 1995). They may be subdivided into the intermediate helium-deficient sdOB stars ( would be a typical value) which are preferably near the hot end of the range and the helium-poor sdB's. Michaud et al. (1989) have shown that at least in the case of sdOB's the observed abundances of helium cannot be levitated by radiative forces.
In a previous paper (Unglaub & Bues, 1996, Paper I) we have shown that the effect of the chemical composition on the temperature structure may be of importance for the predicted surface composition. Therefore in addition to hydrogen and helium the elements C, N and O are taken into account. The abundances of all these elements are predicted simultaneously. As this is an improvement of the work Vennes et al. (1988), we decided to do new diffusion calculations for white dwarfs and subdwarfs with and . For the elements hydrogen and helium the calculation of the radiative forces described in Paper I has been improved and ionization effects are taken into account (see Sect. 2). The diffusion calculations are time dependent with a numerical method similar to the one of in Unglaub & Bues (1997; Paper II). and remain constant during the calculations. As in Paper II a plane-parallel stratification is assumed. In Sect. 4 the influence of mass loss on the surface composition will be investigated.
© European Southern Observatory (ESO) 1998
Online publication: September 8, 1998