[Previous] | [Session 11] | [Next]
J. L. Sokoloski (Harvard-Smithsonian Center for Astrophysics), M. Eracleous (Penn State), D. Steeghs (University of Southampton), L. Bildsten (Institute for Theoretical Physics, UCSB)
Stochastic optical variability with minute to second time scales has been observed in a wide variety of accreting white-dwarf binaries, including cataclysmic variables, supersoft X-ray sources and some symbiotic systems. Although the physical origin of this variability is not well-understood, in each case it is thought to be related to accretion onto the white dwarf. For symbiotics, Sokoloski, Bildsten & Ho (2001) found that the amplitude of optical flickering appears to be inversely related to the amount of quasi-steady nuclear burning that is present on the surface of the white dwarf. Quasi-steady nuclear burning produces non-variable (or only slowly variable) extreme-ultraviolet radiation that is reprocessed into the optical by the nebula, thus reducing the amplitude of the rapidly variable accretion light. Symbiotics with large-amplitude optical flickering are therefore those in which the white dwarf burns the accreted fuel in explosive nova events, with little burning between explosions. In order to test these ideas, and also to investigate the physical origin of the fastest optical variability in the symbiotic recurrent novae and other large-amplitude flickering symbiotics, we performed simultaneous high-time-resolution optical spectroscopy and B-band photometry using the 3-meter and 1-meter telescopes at Lick Observatory. We present these observations for the two symbiotic recurrent novae RS Oph and T CrB, and also for the unusual symbiotic CH Cyg. To separate intrinsic source variability from spectral variations due to slit losses and seeing changes, we employ a normalization method that combines fitting of the non-variable red-giant spectral features and use of the broad-band photometry. Timing-analysis techniques and software that have been developed for the study of cataclysmic variables are used to analyze the sets of corrected spectra.
This work was supported in part by NSF grant INT-9902665.