Multifrequency observations of the eclipsing symbiotic triple system CH Cyg during the 1992-94 active phase

We present UBV photoelectric photometry, plus optical and ultraviolet low- and high-resolution spectroscopy and MERLIN, VLA and JCMT radio/mm-wave observations of the symbiotic star CH Cyg during the period 1989.7 - 1995.5 which covers the recent (1992) outburst. On 1990 July 22 the spectrum displayed only absorption features characterized by broad hydrogen profiles, indicating that CH Cyg underwent a dramatic transient change, losing its circumstellar shell. In spring 1992 the star's brightness increased by about 2.5 mag in the U band, and then gradually increased from U similar to 8.5 at that time to U similar to 6.7 in the summer of 1994. In 1992 September-November and 1994 October-December, the two deep minima caused by an eclipse of the active component by the cool component in the symbiotic pair of the triple CH Cyg system were observed. A decrease of the star's brightness to U similar to 11 in the summer of 1995 indicates that this active phase is over. During the outburst the energy distribution of the hot component exhibits two maxima: the first in the far-UV (1500 - 1900 Angstrom) region, and the second one around the U band. Evolution of the spectrum and variation in the hydrogen line profiles result from an irregular, high-velocity and in part bipolar outflow from the central star. Radio observations indicate a radio flux increase and suggest an expansion of the outer radius of the wind material around the active component at an average velocity of 310 +/- 50 km s(-1) and a mass loss rate of 1.9 +/- 0.3 x 10(-6) M. yr(-1). The luminosity of the active component during the outburst is a few tens of L(.), well below the Eddington limit. It is suggested that its source is the liberated gravitational potential energy of the material accreted on to the low-mass hot component. The mass accretion rate required to power the outburst is of the same order as the mass loss rate inferred from radio observations, with particularly good agreement if the accreting object is a main-sequence dwarf. Our observational results contradict the prediction from an oblique rotator model that CH Cyg will now be in an inactive phase for several decades.