On the nature of Fast Blue Optical Transients

Short rise times of Fast Blue Optical Transients (FBOTs) require very light ejected envelopes, $M_{ej} \leq 10^{-1} M_\odot$, much smaller than of a typical supernova. Short peak times also mean that FBOTs should be hydrodynamically, not radioactively powered. The detection by Chandra of X-ray emission in AT2020mrf of $L_X \sim 10^{42} $ erg s$^{-1}$ after 328 days implies total, overall dominant, X-ray energetics at the Gamma Ray Bursts (GRBs) level of $\sim 6 \times 10^{49}$ erg. We further develop a model of Lyutikov & Toonen (2019), whereby FBOTs are the results of a late accretion induced collapse (AIC) of the product of super-Chandrasekhar double white dwarf (WD) merger between ONeMg WD and another WD. Small ejecta mass, and the rarity of FBOTs, result from the competition between mass loss from the merger product to the wind, and ashes added to the core, on time scale of $\sim 10^3-10^4$ years. FBOTs occur only when the envelope mass before AIC is $\leq 10^{-1} M_\odot$. FBOTs proper come from central engine-powered radiation-dominated forward shock as it propagates through ejecta. FBOTs' duration is determined by the diffusion time of photons produced by the NS-driven forward shock within the expanding ejecta. All the photons produced by the central source deep inside the ejecta escape almost simultaneously, producing a short bright event. The high energy emission is generated at the highly relativistic and highly magnetized termination shock, qualitatively similar to Pulsar Wind Nebulae. The X-ray bump observed in AT2020mrf by SRG/eROSITA, predicted by Lyutikov & Toonen (2019), is coming from the break-out of the engine-powered shock from the ejecta into the preceding wind. The model requires total energetics of just few $\times 10^{50}$ ergs, slightly above the observed X-rays. We predict that the system is hydrogen poor.