Astron. Astrophys. 321, 485-491 (1997)

1. Introduction

In a recent paper (Unglaub & Bues, 1996; Paper I) we have investigated the influence of gravitational settling and selective radiative forces in PG 1159 stars (for a review about these objects see Werner et al., 1996). The results have shown that under certain circumstances diffusion processes may lead to large abundances of heavy elements in the outer regions with surface mass fractions of about . However, radiative levitation and other diffusive processes alone cannot explain the observed abundances in these stars, especially the observed and predicted carbon abundances are in clear disagreement.

For PG 1159-035 and PG 1520+525, both with and the spectroscopic mass determination by theoretical evolutionary tracks yields (Werner et al., 1991). For PG 1159-035, a pulsator, this result has been confirmed by pulsation analysis (Kawaler & Bradley, 1994). According to the corresponding evolutionary track (helium burners, mass loss type B) of Wood & Faulkner (1986) these objects have passed through a region in the HRD where the effective temperature is also 140000 K, but the surface gravity is only . For these model parameters, we predicted in Paper I a thin metal-rich region floating ontop of a helium-rich mantle. However, stars with surface gravities significantly lower than cross this region within less than 30000 y. In the present paper a decision will be obtained, if diffusion can transform an originally helium-rich atmosphere with about solar number ratios of the elements C, N, O and Ne into a metal-rich one within this time-scale. If this were possible, diffusion processes could be an explanation of the surface chemistry alternative to the "born-again AGB star" scenario proposed by Werner et al. (1991) and Kawaler & Bradley (1994).

The object H1504+65 with , has a spectoscopically determined mass of from model atmospheres with carbon and oxygen only (Werner, 1991). According to the evolutionary tracks this star has passed through a region in the HRD with a maximal effective temperature of 350000 K at . For these model parameters we predicted in Paper I an atmosphere which consists of heavy elements with no detectable amount of helium, whereas for , carbon and oxygen should be trace elements only. Nevertheless diffusion processes could be a possible explanation for the observational results, if the heavy elements were enriched rapidly when the star crossed the ultrahot region, and sank slowly until it has cooled down to .

In a direct comparison of theoretical predictions with observational results for objects with , and , the time-independent calculations predicted surface abundances of carbon and oxygen, which are clearly smaller than the observed ones. These discrepancies may be a consequence of too restrictive assumptions in the diffusion calculations. For example, the effects of possible hydrodynamical instabilities and mass loss are ignored. In the present paper we investigate another possible reason for the failure of the time-independent calculations to explain the observations: Do the heavy elements not sink, because the diffusion time-scales are long in comparison to the time scales of stellar evolution? This question will be clarified by calculations which start with a chemical composition typical for PG 1159 stars.

Recently Dehner & Kawaler (1995) have computed evolutionary sequences of white dwarf models which include time-dependent gravitational settling. They published results which refer to surface mass fractions between and stellar masses. In deep regions it may indeed be a good approximation to neglect the effect of radiative forces on the composition profile, because the elements helium, carbon and oxygen are almost completely ionized. Our paper, however, deals with the outer regions where the effect of radiative forces has to be taken into account as shown in Paper I and detailed opacity calculations are necessary to give proper attention to the influence of the composition on the radiation flux.

© European Southern Observatory (ESO) 1997

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