Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 345, 137-148 (1999)

Previous Section Next Section Title Page Table of Contents

5. Variations of AW UMa light-curves

The LCs variations are quite common in W UMa systems. Differences between the LCs obtained at different epochs were also detected in AW UMa.

Mauder (1972) pointed out that the LC variations in W UMa systems can occur in the course of a single night. To check this possibility, we have estimated the standard deviations of counts for AW UMa and for the comparison star by fitting them with trigonometric polynomials. This computation was done in each passband separately for all nights in which our observations were taken. As the brightness of AW UMa and the comparison star were nearly the same, we can expect the same scatter of individual observations. The scatter of U and B observations of AW UMa was markedly larger than that of the comparison star only in May, 1995. Except for this period, AW UMa was in a non-active stage during our observations.

To find a phase dependence of the scatter we have divided the orbital period into 50 intervals and determined the standard deviation of the individual observations in each interval and passband taking into account all the observations. As shown in Fig. 4, the largest scatter occurred around the secondary minimum (explained by the long-term LC variations or possible mid-eclipse brightening), and the smallest scatter around the maximum II.

[FIGURE] Fig. 4. Phase dependence of the scatter of observations in UBV passbands. The phases were calculated using the ephemerides (2) and (3)

5.1. Mid-eclipse brightening?

A very interesting feature of the LC is the variability in the secondary minimum, firstly reported in AW UMa by Kudzej (1987) as mid-eclipse brightening. As an explanation, he proposed the refraction of the radiation of the secondary in the atmosphere of the primary. He argued that the effect is measurable only in the presence of a strong magnetic field. IUE spectroscopy of AW UMa has shown that its magnetic activity is negligible (Rucinski et al., 1984).

The mid-eclipse brightening does not occur at every secondary minimum. Moreover, its position and relative height is variable (Fig. 5). The average amplitude of the brightening, best visible in the V light (due to the lowest scatter of observations), is about 0.02 mag. We have found a mid-eclipse brightening also in published LCs taken in February 1963 (Paczynski, 1964) and in March-April 1985 (Bakos et al., 1991). The mid-eclipse brightening is also visible in Fig. 1.

[FIGURE] Fig. 5. "Mid-eclipse brightening" (indicated by arrows). The dates of observations (upwardly) Mar 1, 1995; Feb 25, 1996; Mar 18, 1996; Dec 10, 1996; Jan 16, 1997 and Feb 6, 1997

5.2. Long-term changes of the LC and photometric elements

Intrinsic variations of the AW UMa LC were observed and discussed by Hrivnak (1982), Derman et al. (1990) and Bakos et al. (1991). Hrivnak (1982) observed and separately analyzed two types of LC: faint and bright. He found that the faint LC corresponds to a larger fill-out factor of the system and to a lower temperature of the secondary component. To explain the photometric variability of the system, the author proposed two alternative models: (i) variations in temperature and size of the secondary component and intermittent outflow of matter from L2 (ii) variable hot spot on the back of the secondary component. The most pronounced changes of the LC and colour indices, as large as 0.15 mag, were observed in 1989-90 by Derman et al. (1990). They explained them by fast mass transfer and variable spots.

Bakos et al. (1991) observed four types of LCs: shallow, deep, intermediate and extraordinary. They differ mainly in the depth of the primary minimum. Transitions from shallow to deep LC lasted a few days. The authors explained LC variations by a non-uniformly distributed matter around the primary component, without attempt at determining the photometric elements. We have analyzed their LCs by the W&D code to find the photometric elements most affected by the LC changes. We have fixed the geometrical elements determined in Sect. 4 (q = 0.08, i = 78.3o). While we have got [FORMULA] = 6970[FORMULA]30 K and a fill-out factor f = 0.95[FORMULA]0.01 for the shallow LC, we were not able to obtain a physical solution for the deep LC (the differential corrections did not converge). It is impossible to explain the large dip of the brightness during primary minimum in the deep LC solely by changes of the photometric parameters of the system. The best explanation provides the stream of matter escaping from L2, when the system fills up the outer critical surface.

In this paper, we refer to the epochs of large variations of the LC as active stages.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1999

Online publication: April 12, 1999