AAS Meeting #194 - Chicago, Illinois, May/June 1999
Session 104. Stellar Atmospheres and Variability
Oral, Thursday, June 3, 1999, 2:00-3:30pm, Marquette

[Previous] | [Session 104] | [Next]

[104.03] Dynamic and Static Conditions in the Atmospheres of Luminous Blue Variables

Th. G\"ang, B. Wolf (LSW), C. Leitherer (STScI)

Luminous Blue Variables (LBVs) have recently been recognized as keys to the evolution of massive stars; understanding the physical mechanisms causing their changing spectral type and huge mass loss rates are thus of great importance.

The typical atmospheric temperature and density characteristics of these objects allow the analysis of these parameters via spectra in the optical wavelength range.

We have obtained two data sets of high resolution spectra to address the following questions:

\vspace*{-2.5mm} \begin{enumerate} \item do LBV winds have inhomogeneities and/or substructures? \vspace*{-2.5mm}

\item which physical mechanisms cause the observed spectral variability? \end{enumerate} \vspace*{-2.5mm}

For the first question we have obtained high resolution spectra (R\approx 60\,000, \lambda\approx3600--7000\,Å) of a sample of 20 LBVs and LBV-candidates in the Milky Way and the Magellanic Clouds. Substructures have been detected in almost all our observed objects suggesting that inhomogeneities might be common features in LBV-winds.

To address the second question we have observed the galactic LBV HD\,160529 over 4 years to study the variability of its optical spectrum. We found that all spectral lines are variable. The lines with P Cygni-type profiles show variations in the emission as well as in the absorption components. Even pure absorption lines, forming deep in the atmosphere (e.g.\ He\,I\,\lambda5876)

show signs of wind-influence. The photospheric line profile variations can be explained by non-radial pulsations which might be triggering temperature fluctuations (order of 2000\,K) that could change the continuum radiation sufficiently to account for the variations seen in the wind lines.

The results from the analyses of this data with static NLTE-codes (Hubeny et al.: TLUSTY, SYNSPEC) and with dynamical and pulsational models (Rivinius et al.: SEIDYNAMIC; Kaufer et al.: Monte Carlo code) will be discussed in this presentation.

If the author provided an email address or URL for general inquiries, it is a s follows:


[Previous] | [Session 104] | [Next]