Halpha line-profile variations of the two active binary systems UX For and AG Dor

1. Introduction

A large piece of information on the solar chromosphere comes from observations of the Balmer H_alpha line. H_alpha contains information about the photosphere, the chromosphere and about activity tracers such as plages, prominences and macroscopic velocity fields.

The generally accepted explanation of the optical variability of RS CVn stars is that they are covered with large, dark spotted areas. Doyle et al. (1989) found clear evidence for a rotational modulation with spots in chromospheric as well as transition region lines and Rodonò et al. (1987) found evidence of a close spatial correlation between spot and plagelike features while Collier Cameron et al. (1989) have observed transient absorption features in H_alpha due to prominences corotating with the star. In analogy to the Sun we expect that stellar plages are seen in the H_alpha line as emission component while prominences will show up as absorption components (see also Collier Cameron et al. 1989).

2. UX For and AG Dor

UX For (= HD 17084) is a rapidly rotating SB2 system showing strong Ca II H and K emission lines. Bidelman & MacConnell (1973) determined the spectral type of UX For as G5 IV, Houk (1982) as G5-8 V + (G). Lloyd-Evans & Koen (1987) found a photometric period of 0.957 days. The light curve amplitude as well as the mean brightness are variable. From the mean colors Cutispoto (1992) derived the spectral classification G6 V + K1 V or G8 IV + G0 V.

AG Dor (= HD 26354) is a SB1 binary showing only weak Ca II H and K emission lines (Bidelman & MacConnell 1973). Lloyd-Evans & Koen (1987) reported V-light modulations with a period of 2.533 days and an amplitude of 0.05 mag. Houk (1978) determined the spectral type as K1 Vp. The mean colors lead to a classification of K1-2 V + K6 V (Cutispoto 1992).

3. Observations

All observations were made with the ESO CAT telescope during the period November 18 to 24, 1992. We used a 1024 pixel CCD resulting in a spectral resolution of approx. 0.12 A at a resolving power of 50,000. The time resolution of consecutive spectra was between 1200 and 3600 s and we achieved a signal-to-noise ratio of 70:1 for UX For and 150:1 for AG Dor.

4. Spectrum Synthesis

In order to extract activity features from stellar spectra we first eliminate that part of the H_alpha profile originating from the stellar photosphere simply by subtracting an appropriate M-K standard. Further subtraction of a spectrum of the star under investigation itself where the H_alpha emission is weakest, should take out the uniform chromospheric contribution -- and thereby revealing the relative variations from the active chromosphere.

Only a few chromospherically active binary systems show H_alpha as a pure emission line. Commonly this feature is seen in absorption but partly filled in by chromospheric emission when compared with appropriate M-K standards.

We found, that both systems, UX For and AG Dor, are double-lined spectroscopic binaries. In order to eliminate the photospheric H_alpha part of each component of the two binary systems we synthesize the observed spectra with two spectra of non-active standard stars of the same M-K classification. Radial velocity shift, rotational Doppler broadening and relative intensities of the two spectra are adjusted. The combined standard star spectra are then subtracted from our observed spectra presumably representing the chromospheric H_alpha part. The standards used for UX For were HR 2261 (G6 V) for the primary and o2 Eri (K0.5 V) for the secondary; for AG Dor we used HR 857 (K0 V) and TW PsA (K4 V).

The residual spectra are displayed as a two-dimensional image in Fig. 2 and 3. The horizontal grey bars mark holes in the phase coverage. In the case of AG Dor we found no significant variations of the residual \halpha emission lines.

5. Residual \halpha variations

In the case of UX For a second run of the spectrum synthesis technique was applied in order to show the variations in more details. For this purpose we choose the residual spectrum at phase 0.311 where the two H_alpha emission peaks are well separated and appear weakest and presumably represent the least active phase. After the seperated subtraction of the individual \halpha emission profiles only the short-term variations remain and are shown in Fig. 4.

We clearly see some variations from the stronger-emission component (the K star), e.g. between phase 0.023 and approx. 0.15. These variations may be due to additional emission from plages or unresolved flares. There is also evidence in Fig. 4 for additional emission in-between the two H_alpha components (the zero-velocity center) between phase 0.15 through approx. 0.3. We are still in the process of investigating the cause of this emission and whether it could be caused by continuum setting errors but, if real, it could be emission caused by material left over from coronal mass ejections. No phase dependence of the H_alpha variations could be found though, and it is thus unlikely that H_alpha in UX For is rotationally modulated.