A unified theory of cataclysmic variable evolution from feedback-dominated numerical simulations

The envelopes accreted by white dwarf stars from their hydrogen-rich companions1 experience thermonuclear-powered runaways2,3 observed as classical nova eruptions4,5 peaking at 105–106 solar luminosities6–9. Virtually all nova progenitors—‘nova-like variables’—exhibit high mass transfer rates to their white dwarfs before and after an eruption10. Surprisingly, 10–1,000 times lower mass transfer rate11 binaries, exhibiting accretion-powered ‘dwarf nova’ outbursts12, exist at identical orbital periods. Nova shells surrounding dwarf novae13–16 demonstrate that at least some novae metamorphize into dwarf novae17,18, though the mechanisms and timescales governing mass transfer rate variations are poorly understood. Here, we report simulations of the multi-Gyr evolution of novae modelling every eruption’s thermonuclear runaway, mass and angular momentum losses, feedback due to irradiation and variable mass transfer rate, and orbital size and period changes. These feedback-dominated simulations reproduce the observed range of mass transfer rates at a given orbital period, with large and cyclic kyr–Myr timescale changes. They also demonstrate Myr-long deep hibernation (complete stoppage of mass transfer), but only in short-period binaries; that initially different binaries converge to become nearly identical systems; low-mass-transfer-rate dwarf novae occasionally generate novae; and that the masses of white dwarfs decrease monotonically, but only slightly while their red dwarf companions are consumed. Cataclysmic variables—a binary pairing of a white dwarf and a hydrogen-rich donor star—experience mass transfer and other complex interactions. This numerical simulation by Hillman et al. models in particular the feedback between the stellar pair, and succeeds in reproducing many of the observed characteristics of cataclysmic variables.