Nickel-rich outflows from accretion discs formed by the accretion-induced collapse of white dwarfs

A white dwarf (WD) approaching the Chandrasekhar mass may in several circumstances undergo accretion-induced collapse (AIC) to a neutron star (NS) before a thermonuclear explosion ensues. It has generally been assumed that such an AIC does not produce a detectable supernova (SN). If, however, the progenitor WD is rapidly rotating (as may be expected due to its prior accretion), a centrifugally supported disc forms around the NS upon collapse. We calculate the subsequent evolution of this accretion disc and its nuclear composition using time-dependent height-integrated simulations with initial conditions taken from the AIC calculations of Dessart and collaborators. Soon after its formation, the disc is cooled by neutrinos and its composition is driven neutron rich (electron fraction Y_e ∼ 0.1) by electron captures. However, as the disc viscously spreads, it is irradiated by neutrinos from the central proto-NS, which dramatically alters its neutron-to-proton ratio. We find that electron neutrino captures increase Y_e to ∼0.5 by the time that weak interactions in the disc freeze out. Because the disc becomes radiatively inefficient and begins forming α-particles soon after freeze out, powerful winds blow away most of the disc's remaining mass. These Y_e ∼ 0.5 outflows synthesize up to a few times 10 ^-2 M_⊙ in ⁵⁶Ni. As a result, AIC may be accompanied by a radioactively powered SN-like transient that peaks on a time-scale of ∼1 d. Since few intermediate mass elements are likely synthesized, these nickel-rich explosions should be spectroscopically distinct from other SNe. The time-scale, velocity and composition of the AIC transient can be modified if the disc wind sweeps up a ∼0.1 M_⊙ remnant disc created by a WD-WD merger; such an 'enshrouded' AIC may account for sub-luminous, sub-Chandrasekhar Type I SNe. Optical transient surveys such as the Panoramic Survey Telescope and Rapid Response System and the Palomar Transient Factory should detect a few AIC transients per year if their true rate is ∼10^-2 of the type Ia rate, and Large Synoptic Survey Telescope should detect several hundred per year. High cadence observations (≲1 d) are optimal for the detection and followup of AIC.