The chemical composition of white dwarfs is rather peculiar. Normally, only the lightest element shows up in optical spectra. The reason for this almost mono-elemental composition has been known for a long time. Schatzman (1949, 1958) has shown that the huge gravitational acceleration leads to downward diffusion of heavier elements, which is fast compared to the evolutionary time scale.
Whether the atmosphere of a white dwarf is hydrogen-rich (spectral type DA) or helium-rich (non-DA: DB, DO) at the hot end of the cooling sequence is probably decided by the pre-white dwarf evolution, particularly by the exact phase, when the star leaves the AGB. According to evolutionary calculations of hydrogen is left over when nuclear burning finally ceases in the hydrogen shell. However, a late thermal pulse in a post-AGB star may lead to an almost complete loss of hydrogen and therefore to a non-DA atmosphere (Iben & McDonald 1995, Herwig et al. 1999).
It is clear that not all white dwarfs remain in their spectral class during the whole cooling time. Particularly, between 28 000 K and 45 000 K not a single non-DA star is known (Liebert et al. 1986). Up to now, no comprehensive explanation has been found for this DB gap. The most popular explanation is that small amounts of hydrogen are hidden in the atmosphere of a DO white dwarf, which float up until the star has cooled down to about 45 000 K. A hydrogen layer of on top of the helium envelope is sufficient to let the white dwarf look like a DA. At K the convection zone of He II begins to reach into the photosphere so that hydrogen and helium are mixed completely. The white dwarf develops again a helium-rich atmosphere (Fontaine & Wesemael 1987).
In order to provide further constraints for an understanding of the atmospheric composition of white dwarfs, the very few objects with hybrid atmospheres consisting of both lightest elements are of great importance. Among these, HS 0209+0832 is particularly interesting because it is besides GD 50 (Vennes et al. 1996) the only white dwarf between 28 000 and 45 000 K where helium (%) could be detected in the otherwise hydrogen rich atmosphere (Jordan et al. 1993). At the temperature of HS 0209+0832 ( K), helium is expected to diffuse out of the hydrogen atmosphere on a time scale of only a few months (see e.g. Vennes et al. 1988, Koester 1989). Physical processes must exist preventing helium from settling down or helium must be supplied continuously, e.g. by accretion.
HS 0209+0832 became even more interesting, when Heber et al. (1997) found that the strength of the He I line at 4471 Å may not be constant. In one spectrum helium was reduced by a factor of 2-3 relative to the other observations. Eight months later the previous helium abundance was measured again. A variable helium line at 4471 Å has also been reported for the DAB G 104-27 with an effective temperature just below the DB gap. While Holberg et al. (1990) detected helium, neither Kidder et al. (1992) nor Finley et al. (1997) confirmed its presence.
Besides helium, C IV could also be observed in the atmosphere of HS 0209+0832 with a spectrum of low signal-to-noise taken with the IUE satellite (Jordan et al. 1993). As in the case of helium, carbon cannot be supported by radiative acceleration in the atmosphere. According to the calculations of Chayer et al. (1995) only can be explained by radiative levitation - compared to as determined by Jordan et al. (1993).
The aim of our HST observation of HS 0209+0832 was to look for He II 1640 Å, to confirm the detection of carbon, to search for additional metals, and to precisely determine the effective temperature from the energy distribution and the shape of the hydrogen lines. Unfortunately, the observation was delayed for more than a year due to the problems with the NICMOS detector and the resulting preference for infrared observations.
In this paper we present an analysis of the HST spectra. Sect. 2 contains a description of the observations. These are analyzed in Sect. 3 and in Sect. 4 with different assumptions for the atmospheric structure. In Sect. 5, we discuss possible explanations for HS 0209+0832 and conclude with a general discussion in Sect. 6.
© European Southern Observatory (ESO) 2000
Online publication: October 2, 2000