We have investigated four models to explain the photospheric features of HS 0209+0832: a homogeneously mixed hydrogen-helium atmosphere, a stratified atmosphere in diffusion equilibrium, a helium abundance varying over the stellar surface, and a DA+DB double star. A stratified model with hydrogen and helium in diffusion equilibrium and a double star with extreme differences of chemical compositions can be clearly ruled out. Abundance variations over the stellar surface like single spots of pure helium surrounded by hydrogen are also not successful. The simplest model gives the best results: All spectral features can be reproduced very well with a homogeneous atmosphere.
Since the time scales for downward diffusion of helium are much shorter than the cooling times there must be a process which can compensate for the gravitational settling of helium. Such mechanisms can be ongoing accretion, meridional circulation, or surface inhomogeneties possibly caused by magnetic fields. As shown in Sect. 3 the latter two possibilities are rather unlikely so that we consider ongoing accretion as the most probable reason for the unusual photospheric composition of HS 0209+0832. This scenario has the advantage that it can explain both the helium content and the presence of heavier elements. Our analysis shows also that the abundances for He II, C IV, and Ti IV are higher than the abundances for He I, C III, and Ti III, respectively. This is probably either a non-LTE effect or an indication that significant amounts of these elements have already been diffused into deeper and hotter layers (see Sect. 3.2). However, diffusion equilibrium has not been reached due to ongoing accretion.
Accretion can occur from the interstellar/circumstellar environment or from mass exchange with a companion star. Since there are no indications for a companion and since only two velocity systems (ISM and photospheric) could be detected in the UV spectrum, the interstellar accretion scenario seems to be most probable - although we cannot fully exclude that some of the unidentified lines originate in an interstellar/circumstellar environment with a different velocity. It is possible that the interstellar lines detected in our HST spectrum originate in the cloud surrounding the white dwarf.
The variation of the helium abundance as observed by Heber et al. (1997) can be explained, if HS 0209+0832 currently crosses a region of the interstellar medium with an inhomogeneous density. The time scale for separation of hydrogen and helium is about nine months (Koester 1989) so that a lower accretion rate results immediately in a reduced helium abundance. This can explain the different measurements of helium in observations separated by about eight months. Since the variability measurement rests on only one observation and could not be repeated it cannot, however, be fully excluded that the helium abundance is actually constant. The only hint for a variable metal abundance is the different strength of C IV 1548 Å in the IUE and HST spectra, but the reality of this measurement is doubtful due to the low S/N ratio of the IUE spectrum.
The accretion rate necessary to explain our observations can be estimated, since the thickness of the visible hydrogen envelope is about (Sect. 5). In order to account for 1 % of helium (by numbers), of matter with solar helium abundance must be accreted on a time scale of about nine months. This value is similar to the accretion rate during an encounter with an interstellar cloud assumed by Dupuis et al. (1993) to explain the presence of metals in cool DZ white dwarfs.
How does HS 0209+0832 fit into the scenario for the evolution of the spectral sequence? The prototype DAB star GD 323 is not compatible with a homogeneously mixed hydrogen-helium atmosphere (Liebert et al. 1984, Koester et al. 1994) whereas the hotter DAOs can be explained better with homogeneous than with stratified atmospheres (Bergeron et al. 1994). Mass loss may play an important role for delaying diffusion equilibrium between hydrogen and helium in DAOs (Unglaub & Bues 1998), but at 35 000 K this can be excluded for HS 0209+0832.
Considering the isolated place of this star inside the DB gap it seems probable that it is not related to other DABs/DAOs and that the important physical mechanisms are different from that of other DAB or DAO white dwarfs. If this is true HS 0209+0832 is not representative for a former DO star which is currently hiding as a DA (with traces of the former atmospherical helium content) before He II convective mixing at K will transform it into a helium-rich DB again. Since the helium is probably accreted only recently it is not possible to tell whether the star has a thick hydrogen envelope and therefore has been a DA during the whole cooling sequence or whether it has been a DO with a hydrogen mass just thick enough that the hydrogen layer is not transparent for the underlying helium.
For a test of our favored scenario it would be very useful to observe HS 0209+0832 on a time scale of a few months in order to monitor possible variations of the helium abundance due to an inhomogeneous interstellar medium or due to slow rotation. Additional UV observations would help to find out whether the metal content remains constant or is also changing with time.
© European Southern Observatory (ESO) 2000
Online publication: October 2, 2000