1996A&A...307..470N

A series of IUE observation from 1992 to 1995 has definitely established PU Vul as an eclipsing binary. The outburst of this symbiotic nova began in 1977. An extended fading in 1980 gave rise to various interpretations, the eclipse scenario being one of them, dust formation being another. From AFOEV and AAVSO observations we find a period of 4900±100days, or 13.42±0.27years. An eclipsing object of such a long period signifies that we see the binary system at an orbital inclination close to 90deg. ESO observations in the near infrared give an orbital velocity of 4.7km/s and a mass function of m_f≃0.05. Assuming a white dwarf mass between 0.4M_☉ and 0.5M_☉ gives for the red giant 0.7≤M/M_☉≤1.1. From the length of the eclipse the radius of the red giant is determined as R_giant≥82R_☉. We discuss IUE, HST and ground based observations of PU Vulpeculae before and during its second observed eclipse of the hot component by the cool giant which lasted from 1993 to 1995, mid-eclipse was in April 1994. Line profiles, particularly those taken by HST, allow a neat distinction between narrow nebular lines and broader wind lines which prove the existence of a fast wind from the hot star in the binary system of v≃1000km/s. That wind has relatively high densities (N_e>10¹²cm^–3) and is optically thick to radiation at λ<228A. Nebular lines have half widths corresponding to v≃70km/s. During the 1994 eclipse the more highly ionized lines were strongly eclipsed, whereas the lowly ionized nebular lines were hardly affected. This proves that the lowly ionized nebular lines are emitted in a very extended region, and not only close to the cool giant. From 1990 to 1994 relative C/N/O abundances of the nebular and wind emission regions have not changed beyond observational uncertainties.