We set out with the hope of unlocking the mysteries of AM CVn, a helium binary with the same optical colours and temperature as observed for pulsating helium white dwarfs. The Whole Earth Telescope, which in twelve days doubled the available photometry on the object, provided a new tool to eliminate gaps in data coverage and identify periodic modulations in the system. As is the case with many past WET runs, the data reveals unexpected complexities and problems and will be a great source for future studies of the object.
AM CVn's temporal spectrum is simple, with precise integer relations between frequencies of orbital resonances, some of which with sidebands with a spacing of 20.8 µHz, which we believe are the result of superhump modulations. The stable part of the temporal spectrum can be explained by 4 independent frequencies, 2 of which we identify as the disk's response to the and 3 terms of the tidal potential, where is the strongest and gives a two-fold azimuthal structure as two spiral arms, is a far weaker tidal response in the outer part of the disk, forcing it into a triangular shape. The mode is assumed to be generated somehow, and then amplified by wave interaction as proposed in Lubow's (1991) model. This leads to a weak prograde precessing wave in the disk.
The best model which describes the data remains an ultrashort period binary undergoing mass transfer. We have improved on the original model of FFW by including a precessing elliptical accretion disk with a stationary spiral pattern, created in response to the system's small mass ratio and the close orbit. The mass transfer rate is large enough to keep the disk in a permanent superoutburst state. The temporal spectrum observed gives us means to investigate structural variations within the disk.
In addition to the stable frequencies detected and analysed, the object also displays quasi periodic modulations, which we interpret as signs of mass accretion from the disk onto the primary object. We could not identify with certainty any modulations from the central accreting star, but propose the modulation with frequency 988.8 µHz as the most likely candidate for a g -mode pulsation on the accretor, supporting the classification of the central object as a DO white dwarf. The detection of the superhump period and the orbital period, gives a mass of 1.0 for the primary object. Some parameters deduced for the system are given in Table 5.
Table 5. Some parameters for AM CVn
On the observational side, we must continue monitoring AM CVn in both white light and multicolour photometry. Of particular importance is an understanding of AM CVn's photometric behaviour in the far UV, leading to determination of the temperature and hence the origin of the various modulations. To investigate the possibility of a g -mode pulsation at 988.8 µHz, a future WET campaign should be organised at a time when the amplitude of this modulation is at its highest, as was the case in 1987 and in 1995. One could then search for harmonics, multiplet splitting, and amplitude as function of wavelength for this modulation frequency.
In addition we must intensify the investigation of the other AM CVn stars. If our model is correct, they should also possess precessing elliptical disks - at least in their high state, but a superhump period may be more difficult to detect if the objects are observed at lesser inclination, which certainly is the case for V803 Cen, where we do not observe emission inside the absorption lines, as we do in AM CVn (Solheim 1993a). The most promising candidate for identification of a superhump period is EC 15330-14, which appears to have an even simpler temporal spectrum than AM CVn, and shows strong modulation at the fundamental frequency (O'Donoghue et al. 1994).
On the theoretical side, we require detailed models of helium disks atmospheres appropriate for AM CVn systems (Bard 1995). We must also produce a better model for permanent superoutbursts in helium disks (Osaki 1995; Tsugawa and Osaki 1995), and consider a detailed study of a hole in the center of the AM CVn's disk, as discussed by Provencal et al. (1995) and indicated in our analysis. In addition, we know that, whatever the surface temperature, the central white dwarfs' internal temperature profile must differ from a normal single white dwarf. We challenge theoreticians to investigate the possibility that some of the observed periods (as 1011.4 s) arise from g -mode pulsations in the mass gaining white dwarf.
We cannot continue to ignore the secondary object. Evolutionary calculations of ultrashort period binary systems under- going mass transfer, shows that the secondary is burning helium in its core until the approach of a minimum orbital period of 10.6 minutes (Savonije et al. 1986). If the binary period is now increasing, it must have passed the minimum period. The secondary is extremely underluminous, but is blasted with radiation (Nymark 1997). Since it is still transferring helium, it must have a helium atmosphere on top of a carbon-oxygen enriched core.
Finally, we reflect upon the future of AM CVn and related systems. Although we know of only 6 such objects, there appears to be a relationship between behaviour and photometric or orbital period, and hence mass transfer rate. AM CVn's general characteristics should remain unchanged until the orbital period increases above s (Eq. (3)) when disk instabilities begin to occur. AM CVn will reach this threshold in about 200 000 years, if we accept the rate of period change determined by Provencal et al. (1995) as the time scale for orbital evolution.
As it evolves, AM CVn will then look similar to V803 Cen and CR Boo, exhibiting large changes in mean magnitude. The semi-degenerate secondary will not inhibit mass transfer, and in 50-100 Myr most of its mass should be transferred. The final product may look like a single helium white dwarf with a thick, rapidly rotating atmosphere - possibly with a brown dwarf or a planet companion. The descendants of AM CVn stars should be DO or DB stars. If the mass of AM CVn is , and this is typical for the primary of AM CVns, then the descendants may be distinguished as DO or DB white dwarfs with higher than average masses and somewhat different pulsation pattern.
Another, less likely scenario for its evolution which needs to be explored, is that the secondary is likely to start transferring carbon-rich material to the disk when the helium envelope is lost (Savonije et al. 1986). The increased opacity of carbon compared with helium may lead to an expanding He-C mixed atmosphere - and the system may look like a R CrB object (Solheim 1996).
© European Southern Observatory (ESO) 1998
Online publication: March 30, 1998