8 Conclusions

HR 7428 has been included in our program of the H$\alpha$ study of RS CVn binaries because photometric variability attributed to starspots (Hall et al. 1990a) and H$\alpha$ variable emission (Cha et al. 1994) make this binary an interesting active system. Actually HR 7428 is not a typical RS CVn system but an unusual Algol-type system in which the evolved companion is not yet filling its Roche lobe and shows evident signs of magnetic activity.

In this work, in addition to the H$\alpha$ chromospheric study, we have reached three main goals: i) we have determined new accurate orbital parameters; ii) we have given, for the first time, physical parameters of both components; iii) evidence for inter-system material has been inferred from the absorption/emission excesses in the H$\alpha$ profile.

We have found that the H$\alpha$ line behaviour is too complex to be explained simply in terms of chromospheric activity. However, we have isolated an excess emission in the line core that can reflect the chromospheric activity of the cool K2 II-III component. The average activity level corresponding to an H$\alpha$ emission equivalent width EW $_{\rm H\alpha} \simeq 0.3$ Å is in agreement with the emission in the Mg  II h & k and Ca  II H & K lines, and as expected from the long rotation period of 108 $\hbox{$.\!\!^{\rm d}$ }$8.

We have derived complete physical parameters of HR 7428, from new radial velocity data and spectral fit of the UV and optical spectrum. The hotter component has definitely been classified as an A2 V star a little evolved off the main sequence. The spectral model fit lead to a radius of about 40 $R{_\odot }$ for the cooler K2 II-III star. This value is significantly smaller than the previous value of 62 $R{_\odot }$ (Hall 1990), but also a little small to fully justify the observed photometric variations as an ellipticity effect.

The final values for the temperature and luminosity of the two components of the system place the hot component as a little-evolved main-sequence of about 2-3 $M_{\odot}$ and the cooler in the region of the He-burning stars, in agreement with the range of 4-5 $M_{\odot}$ we deduce from the orbital solution. The similar time evolution estimated for the individual components sets the age of the system in the range 100-200 Myr. This also indicates a normal independent evolution of the two stars, yet without any relevant mass exchange or mass loss. As a matter of fact the larger K2 II-III star is still well inside the relative Roche lobe as shown by the scaled diagram in Fig. 13, and no mass flow should be expected.

According to the orbital period and the system separation, the more massive component of HR 7428 should reach its Roche Lobe on the way to the red giant branch, leading to a case-C mass exchange. The evolutionary tracks by Fagotto et al. (1994) predict that the cool star will approach the Roche-lobe limit at an age of about 170 Myr, i.e. about 20 Myr after the present stage. From the H$\alpha$ profile we have inferred the presence of inter-system material that seems to be accumulated mainly on the trailing hemisphere of the K2 II-III star, just near the edge of the lobe. From the width of the excess absorption/emission profile we estimate turbulent velocities in the cloud of up to 150 km$\,$s^-1. Such velocities, very likely, could produce evaporation from the Roche Lobe and mass loss from the cool star.

The evaporated material is expected to accumulate in an envelope or ring around the system. The presence of circumstellar material around HR 7428 has been clearly detected from ISO observations. Rodonò et al. (1998) found significant IR excess at wavelengths greater than 1.6 $\mu$m with a definite inversion of the spectrum slope at 25 $\mu$m.

Since the K2 II-III star is still well inside its Roche lobe, the problem of the origin of the flowing mass arises. The magnetic activity does not appear able to produce extended loops or prominences, as sometime are observed in active systems like AR Lac (Frasca et al. 2000) or SS Boo (Hall et al. 1990b; Hall & Ramsey 1992) and therefore does not seem to be an appropriate mechanism to trigger a significant mass loss.

Evolutionary models including convection-overshooting predict for a 5 $M_{\odot}$star a relevant mass loss at the present evolutionary stage of the HR 7428 cooler component. Models by Maeder & Meynet (1988) and Schaerer et al. (1993) predict mass loss rates of $5\times 10^{-9}$ or 10^-8  $M_{\odot}\,$year^-1 for stars of 4 or 5 solar masses in these evolutionary phases. Such mass loss rates could be responsible for the observed circumstellar effects like the extra-absorption/emission at H$\alpha$ and the infrared excess. A more detailed model of the circumstellar material could be attained with higher resolution spectra and a better phase coverage.

Acknowledgements

We wish to thank the Referee, Dr. D. S. Hall, for carefully reading the manuscript and for valuable comments. We would like to thank G. Carbonaro, A. Distefano, G. Occhipinti, and M. Puleo for valuable help in carrying out our observations. We are grateful to Dr. Peter Hauschildt for the atmosphere models he calculated for us. This work has been supported by the Italian Ministero dell' Università e della Ricerca Scientifica e Tecnologica, the Gruppo Nazionale di Astronomia of the CNR and by the Regione Sicilia which are gratefully acknowledged.