Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 349, 169-176 (1999)

Previous Section Next Section Title Page Table of Contents

6. Discussion

6.1. Mass-loss rate of the M-giant in SY Mus

We are of the opinion that observations of SY Mus are too few to determine the shape of the velocity law of the red giant wind in a unique way. Additional data during egress at [FORMULA] would better determine the fit parameter k, and therefore the acceleration region. Nevertheless, we suspect that the wind reaches terminal velocity inside the orbit of the hot companion star. We can then derive the ratio [FORMULA]. Adopting [FORMULA], we find that egress data indicate a mass-loss rate of [FORMULA]. A mass loss rate of [FORMULA] has been derived with a modified Zanstra method by Mürset et al. (1991). Mass-loss rates of non-variable single M-giants are in the range [FORMULA] (Dupree 1986). The mass-loss rate of the M-giant in SY Mus is thus high compared to its single counterparts. S-type symbiotic binaries are known to have higher [FORMULA] excess compared to single M-giants (Kenyon et al. 1988). The larger FIR-excess has been interpreted as being due to a higher mass-loss rate.

A mass-loss rate of [FORMULA] has been derived by Pereira et al. (1999), who also determined a velocity law based on the same IUE observations as used by us. The differences to our result are due to their neglect of the asymmetry of the eclipse curve and the effect of the ionization of hydrogen.

6.2. Possible cause for asymmetric eclipses

In SY Mus the continuum flux attenuation is strongly asymmetric with respect to the mid-eclipse of the white dwarf by the red giant. As the recombination time scales for hydrogen at densities of [FORMULA] are too short to produce a significant phase lag of the ionization front, only an asymmetry in the material distribution can lead to the observed asymmetry in the UV-light curve. In their spectropolarimetric data, Harries & Howarth (1996) remarked residual position angle variations, which could be due to an asymmetric red giant wind.

In the following we briefly discuss a scenario which could explain the observed asymmetry.

The hot component in SY Mus has a luminosity of [FORMULA] (Mürset et al. 1991) and [FORMULA] (Schmutz et al. 1994). For these parameters, the wind momentum-luminosity relation for radiation driven winds of hot stars (Kudritzki 1998) predicts a fast wind. This expectation cannot be verified observationally because IUE high resolution spectra of the stellar continuum of SY Mus are severely under-exposed and therefore too noisy to detect P Cygni profiles. Qualitatively the colliding winds in an s-type symbiotic system, where the orbital motion is comparable to the wind velocity of the red giant, can reproduce the observed variation of the column density around the eclipse. This scenario has already been used by Schild & Schmid (1996) to explain the Raman scattering geometry in V1016 Cyg, which they found to be asymmetric with respect to the binary axis. 3D-hydrodynamical simulations by Walder (1995, 1998) show that, during ingress, the line of sight to the hot star passes through a region which shortly before has been evacuated by the fast wind from the hot star. The red giant wind is too slow to refill this region. The large pressure gradient from the red giant atmosphere to the evacuated region, produces a strong acceleration of the material ejected by the red giant. This rarefaction wave thus leads to a steep increase of the column density during ingress. The velocity law derived from ingress data does therefore not directly reflect the wind acceleration due to forces from the red giant.

During egress the line of sight to the hot star passes through the region behind the red giant, where shortly before the red giant deposited its wind material. Densities and velocities in the wake of the red giant wind material are less affected by the wind from the hot star. The velocity law derived from egress data is therefore more representative for an undisturbed red giant wind.

The asymmetry between the variation of the column density during ingress and egress is present even if the wind from the hot star momentum is several times smaller than the red giant wind (see Walder 1998).

ROSAT X-ray data (Mürset et al. 1997) and IUE spectra (Vogel 1993) indicate, that the hot component in the eclipsing symbiotic binary EG And has a mass-loss of about [FORMULA], with a terminal velocity of [FORMULA]. Therefore it would be interesting to search for asymmetries in the UV light curve of EG And. We have inspected all the available IUE low resolution final archive spectra of EG And. Unfortunately, the egress curve is not well observed, and its steepness cannot be determined. The radial velocity data of EG And also do not allow to derive a precise [FORMULA]. The time of conjunction from the radial velocity curve of Munari (1993) is about 30 days ([FORMULA]) after mid-eclipse according to the UV light curve. The two times are nevertheless consistent as Munari (1993) give an error of 34 days for the time of conjunction. Therefore the available data on EG And do at present not imply any asymmetries in the column density.

6.3. Additional attenuation due to line blanketing

In contrast to SY Mus, RW Hya and BF Cyg, which display continuum attenuation larger than can be accounted for by Rayleigh scattering, the attenuation function in EG And does not differ significantly from the Rayleigh scattering function. In the following we discuss two mechanisms which could lead to reduced line blanketing in EG And.

Because of the high system's velocity of [FORMULA] (Munari 1993), EG And is associated with the old population of the Galaxy. EG And thus could have a lower metallicity, resulting in reduced line blanketing. In the available IUE high resolution spectra the continuum is too noisy to detect any absorption lines.

Another possible explanation for reduced line-blanketing can be found in the influence of the electron temperature on atomic level populations. The electron temperature in the red giant wind is dominated by the heating effects due to the companion star. For RW Hya, we found [FORMULA]. According to Mürset et al. (1991) the hot stars in RW Hya and EG And have similar temperatures, but the hot component in EG And is about 25 times less luminous, which implies that the ionization front of hydrogen is farther out in the red giant wind, where densities are smaller. For an estimate, we have calculated the level populations of the material lost by the red giant with an electron temperature, [FORMULA], equal to the red giant temperature [FORMULA]. The other parameters given in Sect. 5 were kept unchanged. This electron temperature halves the iron curtain opacity at most wavelengths in the IUE short wavelength range.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1999

Online publication: August 25, 1999