Title:Constraint on Recurrent Novae as progenitors of Type Ia Supernovae

Abstract:It is known that at low mass accretion rate (M_dot) the material accreted by a white dwarf (WD) is lost in classical nova (CN) explosions. However, some mass accumulation may be possible just below the stable nuclear burning limit. This inspired a corresponding class of Type Ia supernova (SN Ia) models. We point out that in such a regime, the SN Ia progenitor will become a source of recurrent CN explosions. For instance, the envelope mass required for a thermonuclear runaway to start on the surface of a ~1.3M_sun WD is ~10^-6M_sun, thus while accreting ~0.1M_sun needed to reach the Chandrasekhar mass, such a WD will produce ~10^5 recurrent nova explosions. This will result in the CN rate far exceeding the typical rates observed in nearby galaxies. CNe expected from potential SN Ia progenitors will have relatively short decline times and therefore, can be missed in CN surveys of moderate cadence. We investigate their incompleteness and show that for the limiting magnitude of m_R=21, completeness of ~80% can be achieved for detecting novae with decline time t_2<10 days if observations take place every second night. We use these results to interpret the nova statistics from Arp's survey of M31. We find that significant mass accumulation in the unstable nuclear burning regime, at the level of ~10-20%, is only possible for WDs with mass M_WD <1.25M_sun. More massive WDs can only accumulate ~1-4% of the required mass in this regime. Thus, the final stage of the mass accumulation and the SN Ia explosion cannot occur at low M_dot, when the nuclear burning is unstable. We compute the mass function of WDs in the underlying population of cataclysmic variables, required to explain the observed nova statistics in M31. We find that dN/dlnM breaks at M_WD ~1.1-1.15M_sun. For plausible values of the mass accretion rate, the total number of CVs in M31 with M_WD >0.7M_sun is in ~1x10^5-2x10^6 range.