Sub-luminous `1991bg-Like' Thermonuclear Supernovae Account for Most Diffuse Antimatter in the Milky Way

Observations by the INTEGRAL satellite reveal that the Galaxy glows with the radiation from the annihilation of $(5.0_{-1.5}^{+1.0}) \times 10^{43}$ electron-positron pairs every second. Constrained to be injected into the interstellar medium (ISM) at only mildly relativistic energies, it is highly plausible most positrons originate from the $\beta^+$ decay of radionuclides synthesised in stars or supernovae. However, none of the initially most likely candidates -- massive stars, core-collapse (CC) supernovae (SNe) or ordinary thermonuclear supernovae (SNe Ia) -- have Galactic distributions that match the spatial distribution of positron injection across the Milky Way. Here we show that a class of transient positron source occurring in stars of age >5 Gyr can explain the global distribution of positron annihilation in the Galaxy. Such sources, occurring at a present Galactic rate $\sim$ 0.002 year$^{-1}$ and typically synthesising $\sim$ 0.03 solar masses of the $\beta^+$-unstable radionuclide $^{44}$Ti, can simultaneously explain the absolute positron luminosity of the Galaxy and the abundance of $^{44}$Ca in mainstream solar system material. Our binary evolution models show that mergers of helium-white dwarf (WD) and carbon-oxygen (CO) WD binary systems satisfy these age and rate requirements and plausibly match the $^{44}$Ti yield requirements. The $^{56}$Ni yield, delay time, and rates of these mergers strongly suggests they are the sub-luminous, thermonuclear SNe known as SN1991bg-like (SNe 91bg). These supernovae are, therefore, likely the main source of Galactic positrons. ONeMg WDs from the same WD population plausibly birth (via accretion induced collapse) the millisecond pulsars plausibly responsible for the 'Galactic Centre Excess'.