Quantitative spectroscopy of OB-stars in the optical and the infrared

Massive OB stars are the most luminous stellar objects (10e5 to a
few 10e6 L⊙). Although being rare by number they play a dominant
role in the chemical and dynamical evolution of galaxies through
their input of energy, momentum, and nuclear processed material
into the interstellar medium by means of stellar winds, eruptions,
and (final) explosions. The luminosity of hot massive stars is the
key ingredient to the driving of a dense (10e−6 - 10e−5 M⊙/yr) and
fast (up to 3,000 km/s) outflow lasting a lifetime. This mass loss
imprints unambiguous signatureson the spectral energy distribution
and spectral lines received from these objects. The goal of this
thesis was to investigate and to apply recent, improved methods for
spectral diagnostics in the optical and the infrared by means of
unified model atmospheres, comprising the entire sub- and
supersonic structure from the pseudo-hydrostatic photosphere to the
stellar wind. In particular, we used the nlte-model atmospheres
code fastwind, which is highly computational efficient and updated
to comprise an adequate though approximate treatment of metal line
opacity effects, i.e., metal line-blocking/-blanketing. First we
have tested this code by a comparison with alternative codes (e.g.,
cmfgen by Hillier & Miller 1998 and WM-Basic by Pauldrach et
al. 2001), particularly in terms of temperature stratification,
fluxes, and number of ionizing photons. In almost all cases we
obtained very similar results. Also for the H/He lines which could
be only compared to cmfgen, the coincidence between the codes is
remarkable, except for a subtle discrepancy concerning the He i
singlets, where, in a restricted temperature range, cmfgen predicts
weaker singlet lines. Having tested the improved model atmospheres
code we began our study with a re-analysis of the Galactic O-star
sample presented by Puls et al. 1996 (at that time using pure H/He
models) to investigate the influence of line-blocking/-blanketing.
This re-analysis (by means of profile fitting of photospheric and
wind lines from H and He) resulted in a significant re-definition
of the effective temperature scale due to this line-blanketing
effect. We obtain lower effective temperatures (up to 8,000 K,
depending on spectral type and luminosity class) in combination
with a reduction in either gravity or helium abundance, thus,
making it possible to assign a new Teff - log g and Teff - spectral
type calibration as a function of luminosity class. Further, by
calculating new spectroscopic masses and comparing them with
previous results we find a significant reduction in the so-called
mass discrepancy (Herrero et al. 1992), where the latter describes
the unfortunate situation that spectroscopically derived masses are
lower than those resulting from stellar evolution calculations. For
stars below 50 M⊙ a systematic trend is retained such that the
spectroscopically derived masses are smaller by approx. 10 M⊙
compared to the evolutionary ones. Moreover, the wind momentum
luminosity relation (WLR) changes because of lower luminosities and
almost unmodified wind-momentum rates. Still present, however, is a
separation of the WLR as a function of luminosity class, in
contrast to theoretical simulations which do not predict such a
dependence. From simple arguments and using stellar samples of
different sizes, we find strong indications that for most
supergiants the mass-loss rate is over-estimated by a factor of 2
to 3, whereas the mass-loss-estimates for luminosity class III and
V objects are consistent with our own theoretical expectations and
those by others. The over-estimate is interpreted as an effect of
wind-clumping, and our argumentation is based on the assumption
that the material in the lower wind region is un-clumped, in
accordance with theoretical predictions. As a final step we have
analyzed a large sample of OB stars by means of H and K band
spectroscopy in the infrared (IR) regime, with the primary goal to
investigate to what extent a lone near IR-spectroscopy is able to
recover stellar and wind parameters derived in the optical. Due to
the substantial progress in ground-based IR instrumentation in the
past decade and the extension of model atmosphere codes to the
infrared wavelength regime, IR spectroscopy has become a powerful
diagnostics for the investigation of young and massive stars lying
deeply embedded in the dust-enshrouded environment of molecular
clouds or the Galactic centre, allowing us to take a first step in
the direction of a pure IR analysis. For the stars analyzed we
obtain well-agreeing results between the optical and the infrared,
except for the line cores of Br gamma in early O stars with
significant mass loss, which, again, might indicate the presence of
clumping effects. Having derived the stellar and wind parameters
from the IR, we are now able to constrain the observational
requirements to perform a pure IR-analysis. Most important is a
very high S/N ratio, as the lines to be investigated are extremely
shallow, and a very good resolution, in addition to an adequately
large set of strategic lines. Given this prerequisite, spectral
analyses based on pure IR data could, indeed, be successfully used
as an alternative or support to traditional methods, and will allow
us to proceed towards our ultimate goal of analyzing very young and
highly obscured objects just emanating from their birth places.