Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 336, 637-647 (1998)

Previous Section Next Section Title Page Table of Contents

4. Radial velocity curves and stellar masses

Initially we had taken high resolution data only in the red spectral range [FORMULA] Å in order to determine a radial velocity curve for the cool giant (Table 1). It turned out that our data disagreed with those of Garcia (1986) when interpreted as orbital motion of the cool giant. In fact Garcia's measurements are compatible with an opposite radial velocity curve, as expected for a companion in a double lined spectroscopic binary. Garcia's (1986) measurements were centered at 5200 Å, where the hot component could contribute significantly to the spectrum. Subsequently, we succeeded on two occasions to measure the radial velocity of the hot component from absorptions in the blue spectral range (Table 2). In the following we discuss the radial velocity data for the hot and cool component separately.

4.1. Radial velocity curve of the M star

To derive the orbital parameters of the cool giant we use only observations in the red and near IR ([FORMULA] Å). For these wavelengths we can safely assume that the absorption lines of the red giant are not significantly disturbed by the A-type spectrum of the hot star. The radial velocity of the red star in BX Mon was determined by cross-correlating observations of the symbiotic star with the radial velocity standard HD 108 903. The radial velocity curve analysis was done in the same way as described in Schmutz et al. (1994, Paper I). For the velocity standard we adopt a radial velocity of [FORMULA] from the Astronomical Almanac (1994). Because of the weak continuum in some of the [FORMULA] settings, not all of these spectra are suited for radial velocity determinations.

In order to find the orbital parameters: time of periastron passage, [FORMULA], systemic velocity, [FORMULA], radial velocity semi-amplitude, [FORMULA], eccentricity, e and position angle, [FORMULA] we have performed a least squares fit. For this we used a fixed period of [FORMULA] days as derived in the previous section. The best fit parameters are listed in Table 4 and the corresponding radial velocity curve is shown in Fig. 3. The radial velocity solution predicts the red star in front of the hot component at


This is consistent with the reduced flux in the IUE spectra taken on Jan. 6, 1979 and Dec. 1, 1990. The red giant eclipses the hot component at [FORMULA], periastron passage takes place at phase [FORMULA], the hot component is in front of the red star at [FORMULA] and the apastron is at [FORMULA].


Table 4. Orbital parameters of the M star in BX Mon, assuming [FORMULA] days. [FORMULA] gives the Julian date at which the cool star is in front of the hot component, [FORMULA] gives the time of periastron passage.

[FIGURE] Fig. 3. Radial velocity of the M5 III star in BX Mon based on observations centered redwards of 6000 Å

4.2. Radial velocities for the hot component

Since the spectrum of the hot component stretches well into the visible spectral range it is, in principle, amenable to radial velocity observations with high spectral resolution. We have compared the BX Mon spectra taken below [FORMULA] with that of M-giants and A-supergiants. In two spectra from March 1996 we detected absorption lines, that we attribute to the atmosphere of the hot component. In the spectrum centered at [FORMULA] taken at phase [FORMULA] and the spectrum centered at [FORMULA] taken at phase [FORMULA] we see that at wavelengths at which A-stars have strong absorption lines, the spectrum of an M-standard does not fit well the spectrum of BX Mon. In these two spectra we also have to assume an additional continuum contribution in order to scale the BX Mon spectrum to that of the comparison M-star. In these two observations we are thus seeing a superposition of the M-star spectrum and an A-star spectrum caused by the hot component. These observations allow us to determine radial velocities for the red and the hot component in BX Mon. Cross-correlation with a M3III and an A8Ia star leads to the radial velocity values listed in Table 2.

The two values in Table 2 for the hot component, were used to determine the radial velocity semi-amplitude of the hot component [FORMULA]. We employed the parameters [FORMULA], e, and [FORMULA] as determined for the red giant. We set [FORMULA]. The only parameter we have to fit is thus [FORMULA]. As shown in Fig. 4, we find


Further measurements at phases when the system is bright in the optical are needed to improve the accuracy of our result. Unfortunately, Garcia's (1986) measurements do not improve the accuracy of the [FORMULA]-determination. As both components may contribute to the spectrum, his cross-correlations may suffer from a blending effect. This would explain his values which lie between our radial velocity curves In fact, our measurements of the red star velocity from settings at wavelengths [FORMULA] also differ from the solution derived in Sect. 4.1, but we cannot derive additional radial velocities of the hot component from these settings.

[FIGURE] Fig. 4. Radial velocity curve of the hot and cool components. The symbols represent our two radial velocity measurements of the hot component (squares), the radial velocity measurements of the M-star derived from spectra centered blueward of [FORMULA] (vertical lines) and Garcia's (1986) radial velocity values derived from spectra centered at [FORMULA] (diamonds).

4.3. Mass function and stellar masses

We determine first the binary mass function based on the well defined radial velocity curve of the red giant. The mass function is defined as


where [FORMULA] and [FORMULA] stand for the masses of the hot star and the red giant, i for the orbital inclination, P for the period, [FORMULA] for the radial velocity semi-amplitude of the red star, and G for the gravitational constant. The values from Table 4 yield for the BX Mon system.


To derive stellar masses we need in addition the mass ratio and the orbital inclination. From the radial velocity amplitude we obtain a mass ratio of


For the inclination we can determine a lower limit from the observed eclipse in the UV. The radius as derived in Sect. 6.2together with the binary separation at eclipse calculated with the upper limit [FORMULA] leads to an inclination [FORMULA] which corresponds to a lower limit of [FORMULA] for the inclination.

We can thus determine both stellar masses according to



The above mass [FORMULA] is typical for a white dwarf. Hot components with similar mass values have indeed been found in other symbiotic systems (Mikolajewska & Kenyon 1992, Schmutz et al. 1994, Schild et al. 1995) The mass-ratio and the total system mass [FORMULA] are slightly higher than for most other symbiotics (Mikolajewska 1997, Schmid 1998).

All newly derived orbital and stellar parameters of the BX Mon system are summarized in Table 5 and the orbits of the two components are illustrated in Fig. 5. The separation between the two components varies between 2.0 and 5.9 AU. At periastron the inner Lagrangian point [FORMULA] is at 1.1 AU, that is 1.6 times the radius of the cool component. Thus even at periastron, [FORMULA] is well detached from the red giant.


Table 5. Summary of newly derived parameters for the BX Mon system.

[FIGURE] Fig. 5. Orbit of the red giant (+) and hot component (*) in the BX Mon system in steps of [FORMULA]. In this representation, the stars move anti-clockwise. The dotted circle represents the red giant boundary at [FORMULA]. The square marks the center of gravity. Axes are in units of the semi-major axis a.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1998

Online publication: July 20, 1998