2002MNRAS.336..467Y

In many symbiotic stars there appear broad wings around Hα, of which the formation mechanisms proposed thus far include a fast outflow, motion of the inner accretion disc, electron scattering and Raman scattering of Lyβ. We adopt a Monte Carlo technique to simulate the Raman scattering of ultraviolet photons that are converted into optical photons around Hα, forming broad wings, and compute its polarization. Noting that many symbiotic stars exhibit a bipolar nebular morphology and polarization flip in the red-wing part of the Raman-scattered O vi features, we assume that the neutral scattering region is composed of two components. The first component is a static cylindrical shell with finite thickness; and the second component is a finite slab that is moving away with velocity v_p= 100 km.s^–1 along the symmetry axis of the first component. The cylindrical shell component yields polarization in the direction parallel to the cylinder axis. The strongest polarization is obtained in the limit where the height of the cylinder approaches zero and the scattering region effectively becomes a circular ring. As the height of the cylinder increases, the resultant polarization decreases and becomes negligible in the limit of the infinite cylinder. The polarization near the line centre is weaker than in the far-wing regions because of the large Rayleigh scattering numbers arising from the large scattering cross-sections near the line centre. The receding polar scattering component produces strong polarization in the direction perpendicular to the cylinder axis. In the presence of a Lyβ emission-line component with an equivalent width ∼ 0.5 Å, the polarized flux exhibits a local maximum at λ= 6578 Å that corresponds to a recession velocity 6.4v_prelative to Hα. When both of the scattering components co-exist, the polarization is characterized by weak parallel polarization near the line centre and strong perpendicular polarization in the red part. We discuss the observational implications of our computation.