4 The type of variability in symbiotic stars

As a guideline for observers not familiar with the symbiotic stars, a few simplified notes may be of interest concerning the types of variability one may expect from the latter and the best way to observe them. We will limit the discussion to the photometric bands of the $UBV(RI)_{\rm C}$ system.

The variability ascribed to the cool giant is best observed at longer wavelengths (i.e. the I band. The R band is affected by the usually very strong H$\alpha$ emission), where the contamination from the white dwarf companion and the circumstellar material become less important. Basically, two types of variability of the cool giant may be observed:
intrinsic, like the pulsations of a Mira variable (about $\sim$ 20% of the known symbiotics harbor a Mira). The amplitude of variability generally decreases toward longer wavelengths. Popular examples are R Aqr (pulsation period of 386 days, minima as faint as V = 12, maxima as bright as V =5 mag) or UV Aur (pulsation period of 395 days, minima as faint as V=11, maxima as bright as V=7.5 mag);
ellipsoidal, when the cool giant fills its Roche lobe. Due to the orbital motion the area of the Roche lobe projected onto the sky varies continuously, with two maxima (when the binary system is seen at quadrature) and two minima (when the cool giant passes at superior or inferior conjunctions) per orbital cycle. Because the reason for variability is a geometrical one, the amplitude of variability is not strongly dependent upon wavelength. Popular examples of symbiotic stars showing ellipsoidal distortion of their lightcurve are T CrB (orbital period 227 days, amplitude $\bigtriangleup m= 0.3$ mag) and BD-21.3873 (orbital period 285 days, amplitude $\bigtriangleup m= 0.2$ mag).

The variability ascribed to the hot white dwarf companion to the cool giant is best observed at shorter wavelengths. There are several type of manifestations, among which:
outbursts, with amplitudes $\bigtriangleup B = 2 - 5$ mag and duration from half a year to many decades. The amplitude, duration and lightcurve shape are usually unpredictable. The same system might show completely different type of outbursts one after the other. For example, QW Sge had an outburst extending from July 1962 to March 1972 characterized by a rapid rise and a linear and smooth decline, followed by another one from May 1982 to September 1989 showing a complex lightcurve with more than one maximum and deep minima in between;
reflection effect, when the hard radiation field of the hot and luminous white dwarf (radiating mainly in the X-ray and far ultraviolet domains) illuminates and heats up the facing side of the cool giant (which reprocesses to the optical domain the energy received by the white dwarf). The heated side of the cool giant is therefore a bit brighter and bluer than the opposite one (which is not illuminated by the white dwarf radiation field). During an orbital period the heated side comes and goes from view, causing a sinusoidal lightcurve. The effect is strongly wavelength dependent, being maximum in the U band and undetectable in R and I bands. The amplitude may be fairly large, as in LT Del where $\bigtriangleup U=1.6$, $\bigtriangleup B=0.5$ and $\bigtriangleup V= 0.2$ mag. It should be observable in the majority of symbiotic stars (more and more easily as the white dwarf gets hotter and the orbital inclination increases) and it is a powerful way to measure orbital periods;
eclipses of the white dwarf by the cool giant. In quiescence the eclipses generally escape detection by optical photometry because the white dwarf is radiating mostly at shorter wavelengths (X-rays and far ultraviolet). During the outbursts the white dwarf emission shifts to longer wavelengths and becomes conspicuous in the optical, thus allowing the eclipses to be detected if the orbital inclination is sufficiently high. Classical examples of symbiotic stars for which the eclipses passed undetected in quiescence and instead became outstanding features of the outburst lightcurve are FG Ser and V1413 Aql. Because the eclipsing body is cool and the eclipsed one is hot, the visibility of eclipses increases toward shorter wavelengths (for example for FG Ser in outburst it was $\bigtriangleup V=1.4$, $\bigtriangleup B=1.9$ and $\bigtriangleup U=2.3$ mag);
re-processing by the circumstellar nebula of the energy radiated by the white dwarf. Sometimes there is so much circumstellar gas ionized by the radiation field of the white dwarf that its brightness completely overwhelms that of the binary system, as it is for the popular cases of V1016 Cyg and V852 Cen (the Southern Crab). Both these symbiotic binaries harbor a Mira variable, whose variability however does not at all affect the optical photometry because of the immensely brighter circumstellar ionized gas. When the white dwarf becomes progressively cooler and dimmer, the amount of ionizing photons that it releases goes down, and the ionized fraction of the circumstellar nebula decreases and consequently its brightness (the scenario is valid for radiation bounded nebulae). This is the case for HM Sge that over the last 25 years has gradually become fainter by $\bigtriangleup V$ = 0.031 and $\bigtriangleup B = 0.086$ mag yr^-1.