Accretion of a giant planet onto a white dwarf

Abstract

The detection of a dust disc around G29-38 and transits from debris orbiting WD1145+017 confirmed that the photospheric trace metals found in many white dwarfs arise from the accretion of tidally disrupted planetesimals. The composition of these planetesimals is similar to that of rocky bodies in the inner solar system. Gravitationally scattering planetesimals towards the white dwarf requires the presence of more massive bodies, yet no planet has so far been detected at a white dwarf. Here we report optical spectroscopy of a $\simeq27\,750$K hot white dwarf that is accreting from a circumstellar gaseous disc composed of hydrogen, oxygen, and sulphur at a rate of $\simeq3.3\times10^9\,\mathrm{g\,s^{-1}}$. The composition of this disc is unlike all other known planetary debris around white dwarfs, but resembles predictions for the makeup of deeper atmospheric layers of icy giant planets, with H$_2$O and H$_2$S being major constituents. A giant planet orbiting a hot white dwarf with a semi-major axis of $\simeq15$ solar radii will undergo significant evaporation with expected mass loss rates comparable to the accretion rate onto the white dwarf. The orbit of the planet is most likely the result of gravitational interactions, indicating the presence of additional planets in the system. We infer an occurrence rate of spectroscopically detectable giant planets in close orbits around white dwarfs of $\simeq10^{-4}$.