Astron. Astrophys. 351, 607-618 (1999)

3. Discussion

The analysis of our data shows that the "positive superhump" in TT Ari light curves has remained "active" for more than a year. A comparison between Figs. 4 and 5 shows that the "positive superhump" is more stable than the "negative" one; while the data in 1997-1998 are almost perfectly fitted, the fits of 1995-1996 data deviate in some parts of the curves. Moreover, the amplitude of the wave is larger than that of the signal. On the basis of the fact that with a "positive superhump" period of TT Ari follows the "orbital period-fractional period excess" relation for the CVs showing "positive superhumps", Skillman et al. (1998) supposed that the new wave in TT Ari could be a "positive superhump" of the same type. There are, however, some uncertainties related to this interpretation. The "positive superhumps" observed in SU UMa type CVs during superoutburst are thought to arise from a slowly precessing under the gravity of the orbiting secondary eccentric accretion disc. The theoretical and numerical analyses have shown that eccentric accretion discs are formed under the action of 3:1 resonance, where the accretion disc matter orbits at a period 3 times shorter than the orbital one (Whitehurst 1988; Hirose & Osaki 1990; Lubow 1991). This resonance lies within the accretion disc only for systems with mass ratio and therefore only these systems are expected to display "positive superhumps". The mass ratio of TT Ari components is most probably (Ritter 1990) and the 3:1 resonance lies outside the disc. The "positive superhump" phenomenon has been observed in other systems with also (Patterson et al. 1993; Skillman et al. 1995) and this suggests that another mechanism of eccentric instability could operate. The origin of the "negative superhumps" in TT Ari light curves is unclear also. Lubow (1992) has shown that fluid discs are tilt unstable at the 3:1 resonance. Patterson et al. (1993) supposed that "negative superhumps" in CVs arose from precession of a tilted disc, but this model has the same problem - in TT Ari the 3:1 resonance most likely lies outside the disc. Murray & Armigate (1998) have investigated the tidal instabilities in CV accretion discs by 3D numerical simulations and found that the tilt instability could not grow fast enough to generate a significant tilt. Thus, the questions concerning the origin of the "positive" and "negative superhumps" remain to be resolved by future investigations.

Our photometry before and after the "3-hour" period change provides us with an opportunity to search for changes in the observational properties of the star. In Table 3 are listed the mean values of the quantities characterising the flickering, as discussed in Sect. 2.4, during the "negative" and "positive" regimes. It is seen that except for the time scale of the flickering, all above quantities have decreased. As we have seen, the flickering in TT Ari can be modelled by a general "shot noise" process with shots a little different from exponents. The standard deviation of a light curve generated by some "shot noise" process roughly depends on the overlapping parameter and the squared amplitude of the shots. The overlapping parameter is a multiplication of the shots rate and duration. A decrease of any of these parameters will produce a decrease of standard deviation of the light curve. From Table 3 it is seen that the characteristic length of the shots most probably remains a constant during both regimes. Thus, the decrease of the standard deviation in the TT Ari light curves could be produced by a decrease of either shots rate or shots amplitude. Unfortunately, since the strong shots overlapping these two parameters cannot be directly measured.

Table 3. Flickering properties: mean values

The power slope decreases also. It depends on the particular shape of the shots. If the shots shape is fixed, the PS can be flattened by introduction of some statistical distribution of the shot durations (Terebizh 1989). Bruch (1992) has shown that the most promising physical mechanisms that can generate the flickering in CVs are the turbulence in the accretion discs and/or an unstable accretion onto the white dwarf surface. Both mechanisms are expected to produce shots with varying durations and following some statistical distribution (usually described by a peak function). Thus, the PS flattening could be a result of a broadening of the distribution function of the shots durations.

Rozhen spectroscopic observations hint at possible changes of the astrophysical parameters of the structures responsible for the hydrogen emissions after 1997. Emission lines from CV accretion discs are believed to be composed of two components - one double-peaked component arising either in a hot corona above and below the accretion disc, or in optically thin parts of the disc itself, and a superimposed single-peaked emission component arising near the hot spot. In high-inclination systems these three peaks are clearly seen. In low-inclination systems, as TT Ari, emission lines are single-peaked as we observed in 1996. The spectra of TT Ari taken in 1998-1999, however, suggest more than one component. At that time the star was in a minimum of 6 yr cycle (Fig. 1) suggesting a lower mass transfer rate. Then it is possible for the relative contribution of the two emission sources to change and the hot spot to contribute less. In this case multiple-peaked emission lines could be seen. The small number of observations, however, makes these conclusions somewhat speculative. It is also not clear if the lower mass transfer rate could be connected with the "3-hour" period and the flickering activity change. Therefore regular, long photometric and spectroscopic observations of TT Ari are needed to clarify the reason for the transition from "negative" to "positive superhump" regime.

© European Southern Observatory (ESO) 1999

Online publication: November 3, 1999
helpdesk@link.springer.de