Since the advent of high-energy astronomy, active binaries have been studied in more and more detail thanks to their very high activity levels, several orders of magnitude higher than e.g. in the Sun. In the Sun the modest rotation level induces a correspondingly modest level of magnetic activity and sunspot cover, even at solar maximum, a modest fraction of the solar surface. In active binaries on the other hand the tidally locked rotation is likely to induce significant transformations in the stellar structure, and also manifests itself through much higher levels of magnetic activity. Doppler imaging techniques have also produced evidence for very large starspot groups (see e.g. Rice & Strassmeier 2001 and previous papers in the series). A good understanding of the structure and evolutionary state of active binaries can thus help to understand coronal phenomena at their extreme levels.
One new area of investigation which has opened up since the advent of enhanced resolution X-ray spectroscopy is the study of the metal abundance in coronal plasmas. Abundance values significantly lower than solar have been derived (see e.g. Jordan et al. 1998 for a review). In the Sun it is observed that the coronal abundances are not the same as the photospheric ones, with elements being selectively enhanced (Feldman 1998) on the basis of their first ionisation potential (FIP). The possibility that some elements are actually depleted is also not ruled out (Jordan et al. 1998). Detailed individual coronal abundances of other stars are now becoming available, and thus a new question arises: the relationship between photospheric and coronal abundances.
To study this issue it is, however, obviously necessary to have a detailed knowledge of the photospheric abundances of the stars under investigation. The iron photospheric abundances of most of the active stars which are the subject of detailed investigation in e.g. X-rays, are not at all well known, and the situation is worse for the other metals. There are several reasons for this, including (for the active binaries) the difficulty of performing detailed spectroscopic studies of binary stars with a significant rotation velocity, which tends to broaden and blend the spectral lines, making the analysis difficult. Note that, in the case of cool active stars, photometric abundance indices seem to produce biased results, giving lower abundances than determined spectroscopically (Favata et al. 1997) and thus cannot be used for a detailed comparison of photospheric and coronal abundances.
Randich et al. (1994) have conducted a study on a significant sample of objects (67 components in 54 systems), using effective temperatures derived from the B-V index (which is, however, sensitive to both metal abundances and interstellar reddening) and surface gravities estimated from the spectral type and the luminosity class or when radii and masses were known these were directly computed. Iron and lithium abundances were derived by comparison with a grid of synthetic spectra covering a domain of 25 Å around the 6708 Å Li I doublet. The main result from the above study is that a large fraction of its sample is composed of metal deficient stars (around 60% of the stars display a metallicity equal or lower than -0.4 dex). At the same time a large fraction of their stars are Li rich.
Ottmann et al. (1998) studied a much more limited sample of (in general hotter)
objects, but with a much more detailed and purely spectroscopic
analysis, determining Fe, Mg and Si abundances through an equivalent
widths based analysis. Temperatures were derived fitting the wings of
the H
and H
Balmer lines, and gravities were estimated
in a similar way, by comparing the wings of two strong lines to
synthetic profiles. Their study presents results for 5 systems in
common with Randich et al. (1994), deriving, in 4 cases out of the 5, [Fe/H]
abundances larger by 0.1 to 0.3 dex. They also compared the
metallicities of the photosphere and corona (the latest extracted from
the literature) in II Peg and 
And, showing that the X-ray
derived coronal abundance is lower than the photospheric value by one
order of magnitude.
We have thus started a detailed study of the photospheric abundances in a sample of 28 RS CVn stars, aiming at providing consistently-derived abundances of several elements in their photosphere. They will be useful to address the issues of RS CVn metallicity distribution, abundance pattern(s) and mixing processes. They will also serve as a starting point for the comparison with the coronal abundances once our sample objects will have been observed by either Chandra or XMM. The sample selection as well as the observations are presented in Sect. 2.
RS CVn stars are characterised by a high level of activity, which may modify their photospheric properties. Therefore, the methods commonly used for the study of non active stars should be carefully checked before being applied to RS CVn. The aim of this paper is to compare different spectral analysis techniques, to investigate their reliability in the study of single line RS CVn stars. In order to do so, two stars of the sample, the slow rotators, single lines spectroscopic binaries (SB1) IS Vir and V851 Cen, are analysed using three different methods (Sect. 3). In addition, the consistency of the B-V, V-R and V-I indices as temperature indicators is studied on a larger set of 8 single lined sample stars (Sect. 3.6). The results of the three methods are compared in Sect. 4. The parameters and chemical compositions of IS Vir and V851 Cen are discussed in Sect. 5. The analysis of the remaining SB1 sample stars as well as the study of double lines spectroscopic binaries (SB2) will be the subject of future papers.
Copyright ESO 2003