Paper detail

Long term evolution of an interacting binary system

We describe a new code to simulate the stellar evolution of a close interacting binary system. It is then used to calculate the evolution of a classical nova system composed of a 1.25 Msun Main-Sequence (MS) star and a 1.0 Msun white dwarf (WD) companion. The system begins as a well separated non-interacting binary system. Initially, the two stars evolve independently of each other. However, Roche lobe overflow begins as the MS star expands on its way to become a Red Giant. We follow the mass accreted onto the WD and the ensuing nuclear runaways for several thousand flashes. The main finding is that the Roche-Lobe mass transfer rate is modulated by oscillations in the MS star, with a period that is somewhat shorter than the thermal time scale of the star. This periodically modulates the rate of thermonuclear flashes on the WD, between once every 12000 yrs, such that the WD can cool, to once every 300 yrs, such that it cannot. The system is further complicated by the secular drift in the secondary modulation. Such secondary modulation could explain systems like T Pyxidis. Last, we find that the overall process of mass gain by the WD has an efficiency of roughly 9%, thus requiring a donor with an initial mass of larger than about 5 Msun MS for an initial 1 Msun WD, if the WD is to reach the Chandrasekhar mass.