S Stars: Definition

The S stars are the giant stars with the same temperature range as M stars, but their spectra contain absorption bands of zirconium oxide (ZrO) as well as the titanium oxide characteristics of M stars. The overabundance of zirconium, as well as carbon and many heavy elements, is a result of convective mixing which brings products of nuclear reactions in the interiors of stars up to their surfaces. As a result, many S stars show spectral lines of the radioactive element technetium.

M stars with very weak ZrO bands are assigned the intermediate spectral type MS. S stars often show hydrogen emission lined in their spectra and are Mira Variables.

S Stars: Description

Stars of type S were first identified as a distinct spectral type by Merrill (1922). The S stars parallel M stars in temperature, but the absorption bands of the molecule TiO that typify M stars are supplemented and sometimes replaced by absorption bands of ZrO in S stars. In subsequent observations of S stars strong absorption lines of post iron-peak elements were identified in the spectra of S stars as well as molecular bands of YO and LaO. In particular, resonance lines of the unstable element technetium were identified in the spectrum of the S star R And and later in the spectra of other S stars.

Since the half-life of the longest lived isotope of Tc is short (2.6 × 106 yr) compared to the evolutionary timescale of stars, these post iron-peak elements must have been produced in the stars as a consequence of their internal evolution. These observations lay the foundation for the following evolutionary paradigm of late-type stars. During the double shell burning phase in the life of a low-mass star, post iron-peak elements are produced by the capture of thermal neutrons on seed nuclei of iron-peak elements.

These neutron captures occur on a timescale short compared to the half-life of unstable isotopes created during the capture process. Since the review paper of Burbidge et al. (1957), this nucleosynthetic process has been referred to as the s-process. Mixing between the star's interior and surface brings s-processed material and newly created carbon to the surface of the star. This alteration in the chemical composition of the star's atmosphere causes ZrO absorption bands to appear in the spectrum of the star. Thus, an M star is converted into an S star via internal nucleosynthesis and subsequent mixing events. Eventually, the amount of carbon in the star's atmosphere is greater than that of oxygen. At this point, the molecular bands of oxides are replaced by molecular absorption bands of carbon compounds, and the star becomes a carbon star.

Thus, it is widely accepted that late type-stars evolve in spectral type from M to S to C. While the early work on line identifications cited provides a qualitative foundation for the origin of S stars, actual quantitative abundances are more difficult to derive in late-type stars because of the blending of atomic and molecular lines as well as uncertainties in parameters used to calculate model atmospheres. The first study to provide quantitative abundances with which to test the s-process origin of S stars was that of Boesgaard (1970), in which she derived the [Ti/Zr] ratio in a number of late-type stars using the curve-of-growth technique. Analyses using model atmospheres and spectral synthesis have been carried out. The results of these investigations support the general picture that stars evolve from spectral type M to S and finally C via an addition of s-process elements and carbon to their atmospheres. For example, these investigators find Zr enhanced in S stars by up to a factor of 10 over M stars of similar temperatures. However, all these studies confine their attention to a small number of S stars.

The appearance of the spectrum of an S star, unlike that of an M star, depends on several factors. In addition to the temperature, the C/O and [Zr/Ti] ratios influence the strength of TiO and ZrO bands. Several schemes involving two parameters have been devised to classify S. In these schemes a temperature parameter is followed by an abundance parameter that reflects either the C/O ratio or the [Zr/Ti] ratio. Although these empirically determined classification schemes seem to work well, it is not clear how well they correlate to the actual abundance parameters of individual stars.

If one were to consider only those stars that produce carbon and s-process elements by internal nucleosynthesis, the C/O and [Zr/Ti] ratios should be strongly correlated. Theoretical modeling of cool, chemically peculiar stars indicates that the degree to which ZrO bands supersede TiO bands in the spectrum of a star is controlled by the C/O ratio, not the Zr/Ti ratio. Therefore, a classification scheme based on either the C/O ratio or the [Zr/Ti] ratio would be well suited to describing S stars.

However, two discoveries made after the classification schemes of Keenan & Boeshaar and Ake and the theoretical work done by Scalo & Ross indicate that the parameters on which the classification scheme depends for S stars may need to be reconsidered. First, the dissociation energy for ZrO has been found to be lower by about 10% (from 8.47 to 7.85 eV), bringing it closer in value to that of TiO (6.87 eV). This implies that ZrO and TiO will be formed under the same range of conditions in a star's atmosphere. Thus, the ratio of ZrO to TiO band strength will have an increased dependence on the Zr/Ti ratio, and the abundance parameter used in classification schemes must be a reflection of an enhanced abundance of s-process elements. Second, it has been recognized that a considerable fraction of S stars are formed by mass transfer rather than internal nucleosynthesis. In this case, the correlation between Zr/Ti and C/O no longer holds. Because carbon is much more abundant than zirconium in absolute terms, mass transfer onto an unpolluted star will enhance zirconium much more than carbon as a fraction of the total number of atoms in the star's atmosphere. In these stars one expects the band strengths of TiO and ZrO to depend on the [Zr/Ti] ratio to a much greater degree than the C/O ratio. Therefore, it is unclear which abundance parameter in the classification scheme of S stars actually dominates. Clarification of this issue will be aided by the determination of the [Zr/Ti] and C/O ratios in a large sample of S stars.

S Stars: Variability/Peculiarity