Abstract:
Almost all astrophysical models depend on the quality of the used data, including atomic
data in particular, which is used, for example, for the calculation of the opacities in a
stellar atmosphere. This atomic data is usually determined using laboratory spectra, such
as those measured at the National Institute of Standards and Technology in the USA.
Hot stars, which exhibit their flux maximum in the ultraviolet (UV) spectral range, also
provide spectra whose lines can be studied in detail. These stellar spectra are taken by
UV spectrographs in space, such as the Space Telescope Imaging Spectrograph, and have
a very high resolution and a very high signal-to-noise ratio, making them comparable to
laboratory spectra. This enables the quality of the available atomic data to be verified
with stellar spectra.
In this work, three hot subdwarf stars, which exhibit high surface abundances of the
elements Ca – Ni, were used as “stellar laboratories” by analysis of high-quality UV spectra
and subsequent evaluation of isolated absorption lines of the aforementioned elements.
After the evolution and properties of hot subdwarfs are presented, the basic theory on
the formation of stellar spectra, with focus on the absorption-line profile, is introduced.
Then, it is shown how the scientific method of spectral analysis is carried out in general,
how it can be implemented numerically and which important role the accuracy of atomic
data, which serves as numerical input, plays for a successful spectral analysis. In this
context, quantum mechanical corrections, so-called “oscillator strengths”, which are necessary
to correctly model individual absorption-line strengths, are introduced.
High-quality spectra of three hot subdwarf stars, namely EC11481–2303, Feige 110, and
PG0909+276 were analyzed, and atmospheric properties and element abundances were determined
by this means. In each of the three stars several thousands individual absorption
lines of Ca – Ni were successfully identified. Afterwards, isolated absorption lines, which
serve as data for examining the line strengths, were to be identified and listed. Due to
stellar rotation, this was not possible for EC11481–2303. In Feige 110 and PG0909+276,
however, a total of 792 isolated absorption lines of Ca – Ni was obtained. For isolated
absorption-line identification, once, a manual procedure is presented, which was then implemented
as an automatic program.
The theoretically modeled, isolated absorption lines were then compared with the observed
ones. Therefore, an automatic fitting procedure, which determines line centers,
widths, and strengths, was established. After some reduction, the remaining 457 absorption
lines, most of which Cr – Ni, were successfully evaluated, whereby a large amount
of important statistics was obtained, which could be used for a more precise abundance
determination. Besides, no systematic deviations of the modeled and the observed line
strengths, and hence the oscillator strengths, were found. For Feige 110, the statistical uncertainty
of the line strengths, among the examined lines, is 45 %, while for PG0909+276
it is 72 %.
At last, the obtained results were used to find strong, isolated absorption lines with good
agreement between observations and models. Using these lines as reference points will
improve the accuracy of abundance determinations for the elements Cr – Ni, especially in
those cases where only a small number of lines is identified in an observation.