Collaborateurs: Uli Heber (Bamberg); Heinz Edelmann (Bamberg).
Project description: Subluminous B stars dominate the populations of
faint blue stars (V $<$\magpt{16}{5}) and are found both in the old disk
(field sdBs) and in the halo populations (globular cluster members). Recently,
HUT observations suggested that the ``UV upturn phenomenon" observed in
elliptical galaxies and galaxy bulges is caused sdB stars (Brown et al., 1997,
ApJ 482, 685). Spectroscopic analyses have revealed that sdB stars are
extreme (EHB) horizontal branch stars (see Heber 1986, A\&A 155, 33). Their
origin, however, is still unclear and strongly debated. A large fraction of
sdB stars in the field are found in binary systems. Both interacting and
(non-interacting) single star evolutionary scenarios have been proposed for
the origin of the sdB stars (see Bailyn et al.\ 1992, AJ 103, 1564). A HST
Snapshot survey is executed in cycle 7 (7409, PI: Heber) to test predicted
distributions of sdB binary separations by R band imaging with the PC for 40
stars. Recently, 9 sdB stars have been found to be pulsating (Koen et al.,
1997, MNRAS 285, 645), defining a new instability strip in the HRD.
NGC\,6752 was the only globular cluster known to contain sdB stars (Moehler et
al., 1997, A\&A 319, 109) until two sdBs were found in M\,15 (Moehler et al.,
1997, A\&A 317, L83) and a large population of EHB stars in the central parts
of several metal-poor globular clusters were found by HST (Ferraro et al.,
1997, ApJ 484, L145). In contrast, the discovery of four sdB stars in the
very metal-rich open cluster NGC~6791 %was a surprise since previous estimates
of the sdB space density predicted %less than one sdB even for a rich open
cluster like NGC~6791. This fact led Liebert et al.\ (1994, AJ 107, 1408) to
postulate that sdB stars are formed preferentially in a metal-rich stellar
population.
First estimates of the space density were derived from small stellar samples
by Heber (1986, 12 stars at the South Galactic Pole, SGP) %and Downes (1986,
ApJS 61, 569; galactic plane) and yielded 4$\times$10$^{-6}$\,pc$^{-3}$ %and
2$\times$10$^{-6}$\,pc$^{-3}$, respectively. The scale height from the SGP
sample was derived by Heber (1986) to be about 200\,pc. Subsequent analyses
(e.g. Moehler et al., 1990, A\&A, 239, 265) which were based mainly on small
subsamples of the Palomar Green (PG) survey gave results consistent with the
first estimates. For a review see Heber (1992, LNP 401, 233). Vileneuve et
al.\ (1995, ApJ 446, 646) have presented the first analysis of a large sample
of PG sdB stars. Temperatures and distances of 209 sdBs were determined from
Str\"omgren photometry by assuming plausible gravities and masses of their
programme stars. Their scale height (450$\pm$150\,pc) is twice as large as
previous work indicated and the space density
(3$\pm$1$\times$10$^{-7}$\,pc$^{-3}$) is a factor of 5--10 lower than previous
estimates. Such a low space density is difficult to understand in view of the
sdBs in the open cluster NGC\,6791. Moreover, the birthrate of sdB stars
becomes so small that only 0.1\% of the white dwarfs have evolved through the
sdB channel. Vileneuve et al.\ (1995), however, show convincingly that all
previous work was hampered by small number statistics and biased. However,
they also found a problem with their PG sample, in that it appeared to be
incomplete at bright magnitudes due to saturation effects of the photographic
plates at about \magpt{13}{}. This had already been suspected by Moehler et
al.\ (1990). Missing bright stars, however, may lead to an overestimate of
the scale height.
We propose to improve the scale height/space density determination from a
SGP sample. Starting from Heber's (1986) sample (flux limited,
V$_{lim}=$\magpt{14}{}) we will increase the sample size by a factor of 10 by
using two ongoing objective prism surveys, the Beers et al.\ (1985, AJ, 90,
2089) survey (150\,\AA\ band around the H and K lines) and the Hamburg ESO
(HE) survey (Wisotzki 1994, IAU Symp.\ 161, 723). The HK survey covers the
magnitude range from \magpt{11}{}\ to \magpt{15}{5}, while the HE survey
covers B$=$\magpt{13}{} to \magpt{17}{}. Both surveys secured several plates
of the SGP region and are cross-checked for candidate selection, which is
important for the completeness estimate. We have isolated all sdB candidates
in selected SGP fields down to 16th mag and found $\approx$ 100 sdB
candidates. %(including the 12 sdBs of Heber, 1986). A second (smaller) sky
area at lower latitude (at b$=$-40\dg) is chosen for comparison and to allow a
more efficient observing strategy at the 1.5m telescope. During the first two
observing run (96 + 97) spectra of 80 stars were obtained. The proposed
observations will complete the sample. Since statistical completeness is the
most important issue the proposed observations are essential to the project.
{\bf Norbert: Kannst du bitte etwas \"uber Vollst\"andigkeitsabsch\"atzungen
sagen.}
Although being smaller in size, the proposed investigation has two advantages
over Vileneuve et al.'s study: i) we can derive individual gravities and,
hence, distances instead of relying on mean gravities or a particular
distribution of gravities and ii) the incompleteness of the PG survey at the
bright end is avoided since the Beers et al.\ survey find also the brighter
stars.
\noindent {\bf B.} Medium resolution spectra of 60 stars in the blue
spectral range shall be obtained with the B\&C spectrograph at the 1.5m
telescope with S/N=30--70. Our previous runs have indicated that the
misclassification (due to hot DA white dwarfs and sdO stars) of the objective
prism spectra is small, i.e. below 20$\%$.
The spectra shall be analysed for the atmospheric parameters $T_{\rm eff}$,
$\log g$, and helium content by using model atmospheres of Heber (1986) and
the fitting procedure of Saffer et al.\ (1994, ApJ, 432, 321). The procedure
determines the atmospheric parameters simultaneously by fitting the Balmer and
He\,{\sc i} lines. It is important to observe several Balmer lines
(H$_{\beta}$ -- H$_9$) because higher Balmer series members are most sensitive
to $T_{\rm eff}$ and $\log g$. New broadening tables for high $n$ Balmer
lines (principal quantum number $n<23$) calculated with the unified theory of
Vidal et al.\ recently became available (Lemke, 1997, A\&AS, 122, ) and shall
be used for the analysis of high $n$ Balmer lines. Adopting the canonical
mass for the EHB (0.5\,$M_{\odot}$) allows individual spectroscopical
distances to be determined. The distribution of distances shall be analysed
for scale height and space density either from the differential $N(z)$
distribution (binned histogram) following Bixler et al.\ (1991) or Vileneuve
et al.\ (1995), or alternatively from the cumulative $N(z
Last modified 15.1.1999 by
nchristlieb@hs.uni-hamburg.de