In this work we investigate whether the infrared and radio continua
of hot stars are compatible with a smooth wind. We have gathered from the
literature all
infrared, millimetre and radio continuum
observations for a sample of 18 OB stars. These
observations are supplemented with UBV data.
We have developed a spherically symmetric stellar wind code that
self-consistently solves the equations of radiative transfer and statistical
equilibrium for a mixture of H and He.
A Hubeny model photosphere provides the input of radiation
at the base of the wind.
The density and velocity are given
by the force multiplier formalism, with empirical force multiplier
parameters. We call this the smooth wind model.
We compare the continuum observations with the results of our stellar wind
model. The significance of discrepancies between theory and observations is
judged on the basis of the following criteria:
there should be more than one discrepancy (as a single discrepancy
could in principle be due to an observational mismeasurement) and
the discrepancies should be large to the scatter on the observations.
For four stars from our sample (HD 66811, HD 38771, HD 36486 and HD 30614)
the far infrared (larger than 20 µm) fluxes are significantly underestimated
by the theoretical model. This points to an additional emission mechanism, not
present in the smooth wind model. For the other stars from our sample,
there is insufficient observational evidence to confirm -- or exclude -- the
presence of such additional emission. Hence, the fact that we find evidence
for additional emission in 4 out of 18 stars, should not be interpreted in
a statistical sense.
In the light of other observational evidence, we consider wind structure to
be the most plausible source of additional continuum radiation. A structured
wind will have a stronger infrared and radio continuum, due to the
density-squared dependence of the free-free and bound-free emission.
Other explanations, such as a more gradual acceleration of the stellar wind
(i.e. a larger value of ß) are shown to be inadequate.
It is useful to distinguish between two kinds of structure.
The influence of stochastic structure has already been
studied, i.a. by Lamers & Waters.
In this work we study larger, localised structure in the form
of complete or partial shells. By fitting the models to the observations we
estimate the position and the strength of the shells.
Due to the integrated nature of the infrared
continuum radiation, one cannot differentiate between a shell and a
co-rotating interaction region on the basis of these observations alone.
The presence of wind structure could considerably lower radio mass loss rates,
if this structure persists into the radio formation region. We give a first
estimate of the effect of structure on radio mass loss rates and outline how
we will derive a more precise estimate.
This PhD thesis was presented at the Vrije Universiteit Brussel,
28 November 1997, under the direction of Prof. W. van Rensbergen
and Dr. R. Blomme