Self-consistent modeling of metastable helium exoplanet transits

Absorption of stellar X-ray and extreme ultraviolet (EUV) radiation in the upper atmosphere of close-in exoplanets can give rise to hydrodynamic outflows, which may lead to the gradual shedding of their primordial light element envelopes. Excess absorption by neutral helium atoms in the metastable 2 S-3 state [He I(2 S-3)], at similar to 10 830 angstrom, has recently emerged as a viable diagnostic of atmospheric escape. Here we present a public add-on module to the 1D photoionization hydrodynamic code ATES, designed to calculate the He I(2 S-3) transmission probability for a broad range of planetary parameters. By relaxing the isothermal outflow assumption, the code enables a self-consistent assessment of the He I(2 S-3) absorption depth along with the atmospheric mass-loss rate and the outflow temperature profile, which strongly affects the recombination rate of He II into He I(2 S-3). We investigate how the transit signal can be expected to depend upon known system parameters, including host spectral type, orbital distance, and planet gravity. At variance with previous studies, which identified K-type stars as favorable hosts, we conclude that late M dwarfs with Neptune-sized planets orbiting at similar to 0.05-0.1 AU can be expected to yield the strongest transit signal, well in excess of 30% for near-cosmological He-to-H abundances. More generally, we show that the physics that regulates the population and depletion of the metastable state, combined with geometrical effects, can yield somewhat counterintuitive results, such as a nonmonotonic dependence of the transit depth on orbital distance. These are compounded by a strong degeneracy between the stellar EUV flux intensity and the atmospheric He-to-H abundance, both of which are highly uncertain. Compared with spectroscopy data, now available for over 40 systems, our modeling suggests either that a large fraction of the targets have helium-depleted envelopes or that the input stellar EUV spectra are systematically overestimated. The updated code and transmission probability module are available publicly as an online repository.