Raman Scattering in Symbiotic Stars

Symbiotic stars are interacting binary systems consisting of a hot star, typically a white dwarf, and a cool red giant companion. The radiation from the hot star partially ionizes the wind from the cool star, resulting in a characteristic combination of sharp nebular lines and stellar molecular absorption bands in the optical spectrum. Most of the emission lines are readily identifiable with common atoms and ions. However, roughly half of all known symbiotic stars exhibit two strong, broad emission lines at λλ6825 and 7082 (D. A. Allen 1980, MNRAS, 190, 75) which defied identification for a number of years.

H. M. Schmid (1989, A&A, 211, L31) proposed that these long unidentified emission lines result from Raman scattering of the O vi resonance photons at λλ1032, 1038 on the ground state of neutral hydrogen. The O vi photons are produced in the region near the hot star and Raman scattered in the neutral region near the red giant. There was much indirect evidence for the existence of O vi emission in some symbiotic stars (e.g., M. A. Hayes & H. Nussbaumer 1986, A&A, 161, 287; W. A. Fiebelman, F. C. Bruhweiler, & S. Johanssson 1991, ApJ, 373, 649; P. S. Li & S. A. Leah 1997, ApJ, 484, 424). H. M. Schmid & H. Schild (1994, A&A, 281, 145) used spectropolarimetric observations of the λλ6825, 7082 lines to derive orbital parameters for symbiotic systems. However, there remained no simultaneous observations of both the O vi λλ1032, 1038 emission and the emission lines at λλ6825, 7082 in the same symbiotic star, which would provide the ultimate test of the Raman‐scattering hypothesis.

Simultaneous far‐ultraviolet and optical observations of the same objects are rather impractical to obtain. Since the variability of symbiotic binaries occurs on timescales that are of the order of several months, it is not necessary that the optical and far‐UV data be simultaneous. Rather, it is sufficient to demand that the data for each object be taken within a few months of each other. Far‐UV observations of nine symbiotic stars taken with the Hopkins Ultraviolet Telescope (HUT) during the March 1995 Astro‐2 mission and contemporaneous ground‐based optical observations of the same set of symbiotic stars were examined for the presence of both the O vi λλ1032, 1038 and unidentified λλ6825, 7082 lines. Four objects (Z And, V1016 Cyg, AG Dra, RR Tel) exhibit both sets of lines, two objects (HM Sge, CH Cyg) exhibit only the O vi lines, and the remaining three objects (BF Cyg, EG And, AX Per) exhibited neither set of lines. In support of the Raman hypothesis, we find no instance in which the λλ6825, 7082 lines occur without the contemporaneous presence of the far‐UV O vi lines (J. J. Birriel, B. R. Espey, & R. E. Schulte‐Ladbeck 2000, ApJ, 545, 1020).

The Raman scattering efficiency is defined as the number of Raman photons divided by the number of emitted O vi photons. The Raman efficiency in symbiotic systems depends strongly on the ionization geometry of the neutral hydrogen scattering region and the mass‐loss rate of the cool star (H. M. Schmid 1996, MNRAS, 282, 511; T. J. Harries & I. D. Howarth 1997, A&AS, 121, 15; K. W. Lee & H. W. Lee 1997, MNRAS, 287, 211). The derived efficiencies for Z And, V1016 Cyg, AG Dra, and EG And imply ionization geometries X_H⁺∼0.4–4.0, in qualitative agreement with other studies (H. Mürset et al. 1991, A&A, 248, 458; H. M. Schmid et al. 1999, A&A, 348, 950). The relative mass‐loss rates of Z And, V1016 Cyg, AG Dra, and EG And based on their Raman efficiencies are not generally in agreement with those derived from radio measurements (E. R. Seaquist, M. Krogulec, & A. R. Taylor 1993, ApJ, 410, 260). Since both the Raman line intensity and polarization profiles are very sensitive to the mass‐loss rate (T. J. Harries & I. D. Howarth 1997, A&AS, 121, 15), detailed studies of these profiles may provide another useful method to examine the mass‐loss rates of cool red giants in symbiotic systems.

At present, Raman scattering is found to occur only in symbiotic stars, with the exception of the young planetary nebula NGC 7027. D. Péquignot et al. (1997, A&A, 323, 217) observed a Raman scattered He ii line in the optical spectrum of NGC 7027. This unusual planetary nebula appears to have two central stars, one of which is either a main‐sequence star or subdwarf and the other a white dwarf star, and has been discussed as a possible post‐symbiotic system (G. A. Gurzadyan 1997, The Physics and Dynamics of Planetary Nebulae [New York: Springer]). Spectroscopic similarities between bipolar, proto–planetary nebulae and symbiotic stars have lead many researchers to speculate on possible evolutionary links between the two types of objects (H. W. Lee & M. G. Park 1999, ApJ, 515, L89, and references therein). The binarity of symbiotic stars is well‐established while that of bipolar proto–planetary nebulae remains a matter of debate. Nonetheless, it remains an exciting potential that Raman scattering may ultimately reveal evolutionary links between symbiotic stars and bipolar, proto–planetary nebulae.