Astron. Astrophys. 318, 819-834 (1997)

3. Stellar parameters

For the analysis of the stellar wind and its variability, the stellar parameters are required. The distances were adopted from the table of Bieging et al. (1989) (  Sco , ), from Federman & Lambert (1992) (HD 169454, ), and from Barlow & Cohen (1977) (HD 190603, ).

The extinction values were taken from Leitherer & Wolf (1984) (  Sco and HD 169454) and from Lennon et al. (1992) (HD 190603). With these extinction values we derived good continuum fits for all three stars. The glactic extinction law given by Cardelli et al. (1989) was used.

To derive the effective temperature and gravity , a - curve for the measured equivalent width of   is interpolated. The calculated equivalent widths were taken from the BALMER 6 grid based on hydrostatic plane-parallel line-blanketed atmospheres in LTE by Kurucz (1979).

The abundances required for the equivalent widths of the Si lines listed in Table 2 were calculated using the LTE-code by Baschek et al. (1966). We used several atmospheres located on the - curve for the   equivalent width and a set of microturbulences . For HD 169454 the SiII lines were not measurable due to the high noise in this wavelength range. The measured equivalent widths are given in Table 2.

Table 2. Equivalent widths of the   and the Si lines

This method gives curves for every Si line in the - plane with different slopes for different ionization stages. These curves converge to a common intersection only for one microturbulence , as can be seen in Fig. 2. The position of the intersection provides . The accuracy of the method in temperature is about 1000 K. Interpolating the above mentioned - curve gives with an accuracy of about 0.3 dex, which is sufficient for our purpose. The values for should be regarded as upper limits, because the model atmospheres do not converge for lower values at this temperature due to the high radiative acceleration. The atomic data were taken from Kaufer et al. (1994), who describe the method more in detail. The Si abundances obtained are in the range of . They correspond with solar values within the uncertainties of the analysis.

Fig. 2. The silicon ionization equilibrium for  Sco. The plot shows the abundances necessary to yield the equivalent widths given in Table 2. The abundance required to produce a given equivalent width is decreasing with temperature for higher ionization stages. The -value for a given is constrained by the equivalent width of  .

Radius and the luminosity were derived by fitting a model continuum to low-resolution IUE spectra and Johnson and Strömgren photometry. For the model continuum a final atmosphere was calculated using the observed values for and (cf. Table 3). For HD 190603 only Johnson photometry was available. The mean Strömgren photometry was taken from Sterken (1977). The values were converted to absolute fluxes using the conversion factors published by Szeifert et al. (1993). The Johnson UBV data was presented by Fernie (1983). The conversion factors were taken from Bessell (1979).

Table 3. Stellar parameters of the observed stars

A mean flux-calibrated IUE spectrum was used for the fit. For  Sco this was created using averages of four SWP (6065, 8829, 8968, 10342) and three LWR (7611, 7665, 7720) spectra. For HD 190603 we had one SWP (14587) and one LWR (7299) spectrum. For HD 169454 five SWP (2985, 6580, 6581, 6582, 23524) and two LWP (3879, 3880) spectra were used. In our fit we concentrate on the near-UV wavelength range and the optical photometry.

The systemic velocities were approximated by the average velocities of the emission lines of the FeIII multiplets 115 and 117. The standard deviation of these velocities is about 5 , presumably due to intrinsic variations. These lines are expected to be pumped by two HeI transitions in the far UV (cf. Wolf & Stahl 1985) and therefore should form in the lowest part of the photosphere. In the following we call these lines "photospheric emission lines".

For  Sco and HD 190603 we adopted the terminal velocities published by Prinja et al. (1990) and corrected them for the systemic velocity . For HD 169454 no high-resolution IUE spectrum in the short wavelength range is available. So we derived the terminal velocity from , in which the fastest components can be detected at in the rest frame of the star.

For our analysis we neglected the effects of atmospheric extension, non-LTE and the fact that the atmosphere is not in a hydrostatic state. Our set of parameters has been derived in a unified way, so the systematic errors are of the same order for all three stars. For the analysis below, we expect the absolute values of the stellar parameters to be less important compared with the advantage of having an homogeneous derived set of parameters for all stars. Nevertheless, our values agree well with the values given in the literature. A summary of all adopted and derived stellar parameters is given in Table 3.

© European Southern Observatory (ESO) 1997

Online publication: July 3, 1998
helpdesk.link@springer.de