Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 322, 576-590 (1997)

Previous Section Next Section Title Page Table of Contents

1. Introduction

Symbiotic stars are characterized by the simultaneous occurrence in an apparently single object of two temperature regimes differing by a factor of 30 or more. The spectrum of a symbiotic star consists of a late type (M-) absorption spectrum, highly excited emission lines and a blue continuum. Generally, symbiotic stars are interpreted as interacting binary systems consisting of a cool luminous visual primary and a hot compact object (white dwarf, subdwarf) as secondary component. Because of mass loss of the giant there is often a common nebulous envelope. Mass transfer from the cool to the hot component is expected. An extensive review on the properties of symbiotic stars can be found in Kenyon (1986).

Many symbiotic stars show outburst events at optical and UV wavelengths. Though several models have been proposed to explain these outbursts, it is the combination of quiescent properties and these outburst properties which make it difficult to derive a consistent picture for some symbiotics.

The symbiotic star AG Draconis (BD [FORMULA] 922) plays an outstanding role inside this group of stars because of its high galactic latitude, its large radial velocity of [FORMULA] =-148 km/s and its relatively early spectral type (K). AG Dra is probably a metal poor symbiotic binary in the galactic halo.

Here, we use all available ROSAT data to document the X-ray light curve of AG Dra over the past 5 years. In addition, we report on the results of the coordinated ROSAT /IUE campaign during the 1994/1995 outbursts. Preliminary results on the IUE and optical observations were given by Viotti et al. (1994a, 1994b). After a description of the relevant previous knowledge of the AG Dra properties in the optical, UV and X-ray range in the remaining part of this paragraph, we present our observational results in paragraph 2-4, discuss the various implications on the AG Dra system parameters in paragraph 5 and end with a summary in paragraph 6.

1.1. AG Dra: the optical picture

Like in most symbiotic stars, the historical light curve of AG Dra is characterized by a sequence of active and quiescent phases (e.g. Robinson 1969, Viotti 1993). The activity is represented by 1-2 mag light maxima (currently called outbursts or eruptions) frequently followed by one or more secondary maxima. It has been noted (Robinson 1969, Iijima et al. 1987) that the major outbursts occur in [FORMULA] 15 yr intervals. This recurrence period has continued with the recent major outbursts of 1981 and 1994. Iijima et al. (1987) suggest that the major outbursts seem to occur at about the same orbital phase, shortly after the photometric maximum, i.e. shortly after the spectroscopic conjunction of the companion (hot component in front of the cool component). Occasionally, AG Dra also undergoes smaller amplitude outbursts, such as those in February 1985 and January 1986.

Between the active phases AG Dra is spending long periods (few years to decades) at minimum light (V [FORMULA] 9.8-10.0, Mattei 1995), with small (0.1 mag) semiregular photometric variations in B and V with pseudo-periods of 300-400 days (Luthardt 1990). However, in the U band regular variations with amplitudes of 1 mag and a period of 554 days have been discovered by Meinunger (1979), and confirmed by later observations (e.g. Kaler 1987, Hric et al. 1994, Skopal 1994). This periodicity is associated with the orbital motion of the system, as confirmed by the radial velocity observations of Kenyon & Garcia (1986), Mikolajewska et al. (1995) and Smith et al. (1996). Because of the color dependence these U band variations certainly reflect the modulation of the Balmer continuum emission. The photometric variability is continuous, suggesting periodic eclipses of an extended hot region by the red giant.

In June 1994 Graslo et al. (1994) announced that AG Dra was starting a new active phase which was marked by a rapid brightening from V=9.9 to V=8.4 on June 14th, and to 8.1 on July 6-10, 1994. After July 1994 the brightness gradually declined reaching the quiescent level (V [FORMULA] 9.8) in November 1994. Like the 1981-82 and 1985-86 episodes, AG Dra underwent a secondary outburst in July 1995, reaching the light maximum of V=8.9 by the end of the month.

The optical spectrum of AG Dra was largely investigated especially in recent years, during both the active phases and quiescence. The spectrum is typical of a symbiotic star, with a probably stable cool component which dominates the yellow-red region, and a largely variable "nebular" component with a strong blue-ultraviolet continuum and a rich emission line spectrum (e.g. Boyarchuk 1966). According to most authors the cool component is a K3 giant, which together with its large radial velocity (-148 km s-1) and high galactic latitude ([FORMULA] = [FORMULA]), would place AG Dra in the halo population at a distance of about 1.2 kpc, 0.8 kpc above the galactic plane. While Huang et al. (1994) suggested that the cool component of AG Dra may be of spectral type K0Ib, which would place it at a distance of about 10 kpc, Mikolajewska et al. (1995) from a comparison with the near-infrared colors of M3 and M13 giants concluded that the cool component is a bright giant with [FORMULA] [FORMULA] -3.5, placing AG Dra at about 2.5 kpc. Most recently, Smith et al. (1996) performed a detailed abundance analysis and found a low metallicity ([Fe/H]=-1.3), a temperature of [FORMULA] =4300 [FORMULA] 100 K and log g = 1.6 [FORMULA] 0.3 consistent with the classification of AG Dra's cool component as an early K giant. In the following we shall assume a distance of 2.5 kpc for the AG Dra system.

1.2. AG Dra: the UV picture

The hot dwarf companion is a source of intense ultraviolet radiation which produces a rich high-temperature emission line spectrum and a strong UV continuum. AG Dra is in fact a very bright UV target which has been intensively studied with IUE (e.g. Viotti et al. 1983, Lutz et al. 1987, Kafatos et al. 1993, Mürset et al. 1991, Mikolajewska et al. 1995).

Ultraviolet high-resolution spectra with the IUE satellite revealed high-ionization permitted emission lines such as the resonance doublets of NV, CIV and SiIV. The strongest emission line is HeII [FORMULA] 1640 which is composed of narrow and broad (FWHM = 6 Å, or equivalently [FORMULA] 1000 km/s) components. The NV line has a P Cygni absorption component displaced 120 km/s from the emission peak (Viotti et al. 1983).

The origin of this feature is unclear. It can either arise in the red-giant wind ionized by the hot dwarf radiation, or in some low-velocity regions of the hot component's wind.

The UV continuum and line flux is largely variable with the star's activity. Viotti et al. (1984) studied the IUE spectra of AG Dra during the major 1980-1983 active phase, and found that the outburst was most energetic in the ultraviolet with an overall rise of about a factor 10 in the continuum, much larger than in the visual, and of a factor 2-5 in the emission line flux. A large UV variation was also a characteristics of the minor 1985 and 1986 outbursts (e.g. Mikolajewska et al. 1995).

1.3. AG Dra: the X-ray picture

First X-ray observations of AG Dra during the quiescent phase were done with the HEAO-2 satellite (Einstein Observatory) before the 1981-1985 series of eruptions (0.27 IPC counts/sec). The spectrum was found to be very soft (Anderson et al. 1981). The data are consistent with a blackbody source of kT=0.016 keV (Kenyon 1988) in addition to the bremsstrahlung source (kT=0.1 keV) suggested by Anderson et al. (1981). The X-ray temperature of [FORMULA]  200 000 K is in fair agreement with the source temperature of [FORMULA] 100 000-150 000 K inferred from IUE observations, although the blackbody radius deduced from IUE data is nearly a factor of 10 larger than the X-ray value (Kenyon 1988). Unfortunately, the major 1980 outburst was "lost" by the HEAO-2 satellite because of a failure of the high voltage power.

A comparison of the X-ray flux with the observed HeII [FORMULA] 4686 flux indicates that the luminosity in absorbed He [FORMULA] photons is larger by a factor of two (Kenyon 1988). It has been concluded that the X-rays are degraded by the surrounding nebula thus causing an underestimate of the actual X-ray luminosity.

EXOSAT was pointed on AG Dra four times during the 1985-86 minor active phase, which was characterized by two light maxima in February 1985 and January 1986. These observations revealed a large X-ray fading with respect to quiescence (Piro 1986), the source being at least 5-6 times weaker in the EXOSAT thin Lexan filter in March 1985, and not detected in February 1986 (Viotti et al. 1995). Simultaneous IUE observations have on the contrary shown an increase of the continuum and emission line flux, especially of the high temperature NV 1240 A and HeII 1640 A lines at the time of the light maxima. According to Friedjung (1988) this behaviour might be due to a temperature drop of a non-black body component, or to a continuous absorption of the X-rays shortwards of the N [FORMULA] ionization limit. No eclipse of the X-ray source was found at phase 0.5 (beginning of November 1985) of the orbital motion of the AG Dra system, implying a limit in the orbit inclination. The weakness of the countrate in the Boron (filter #6) EXOSAT filter during quiescence implies a low (2-3 [FORMULA] 105 K) temperature of the source (Piro et al. 1985).

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1997

Online publication: June 5, 1998