2. Shell Stars
According to Merrill (1953), Otto Struve first introduced the term shell star to describe a star with an extended atmosphere. Now the term is generally interpreted to mean a late O, B or A star of luminosity class III--V with the following properties. Usually emission is seen in one or more of the Balmer lines; the star is thus a Be star. As a group Be stars rotate rapidly and are surrounded by circumstellar material which is not expected to be distributed with spherical symmetry. Recently the circumstellar matter around Cas (Quirrenbach et al. 1993, Stee et al. 1995) and several other Be stars has been resolved, directly confirming this expectation.
In addition the hydrogen lines also have narrow, deep absorption cores; in one star absorption lines up to H42 have been seen. Frequently these lines have zero central depth, implying that the matter in which these lines arise, referred to as a shell, covers the star in the line of sight. The spectrum at optical wavelengths also contains a large number of narrow, deep absorption lines of singly ionized metals such as Fe, Ti, Ca, Sc, Cr, etc. For these lines the lower level of the particular transition is either the ground state of the ion or is metastable. The gas producing these lines has a lower degree of ionization than the photosphere, with the result that the overall spectrum resembles an early A supergiant, except that lines whose lower level is neither the ground state nor metastable are much weaker in the spectrum of the shell star than they are in the spectrum of the supergiant. In the ultraviolet the spectrum is also greatly affected by the presence of large amounts of circumstellar matter. This is well illustrated by the comparison of the absolute energy distribution of the Be star 59 Cyg before and during a shell phase (Beeckmans 1976).
There are a small number of shell stars whose spectra have not changed over long periods of time (Gulliver 1981, and references therein). In general, however, variability predominates with the shell features appearing and later disappearing completely. Well known examples are Cas and 59 Cyg (Underhill and Doazan 1982 and references therein).
It has frequently been proposed that Be stars are interacting binaries (Harmanec 1987 and references therein). According to one version of this suggestion the Be star is the mass gaining star in an interacting binary system in which the companion fills its Roche lobe. However, despite extensive study of some systems thought to be prime candidates, no such cool giant companion has yet been detected (Hubert 1994). Based upon a consideration of numerous cases, Hubert (1994) concluded that ".... it is clear that the circumstellar envelope of the majority of Be stars in binary systems is produced by ejection from the star itself". Some Be stars are known to be members of binary systems, specifically the Be/X-ray binaries (van den Heuvel and Rappaport 1987). In these systems, however, the Be star is the mass losing star. Cas has long been suspected as being a member of a binary system of this type due to its hard X-ray emission (Haberl 1995 and references therein). The hypothetical companion might therefore be either a neutron star or a white dwarf. Neither possibility, however, seems to be a likely candidate for the source of the circumstellar material responsible for the variability which occurred in 1933-1942.
© European Southern Observatory (ESO) 1997