2018ApJ...854..108G

Planets reflect and linearly polarize the radiation that they receive from their host stars. The emergent polarization is sensitive to aspects of the planet's atmosphere such as the gas composition and the occurrence of condensates and their optical properties. Extracting this information will represent a major step in the characterization of exoplanets. The numerical simulations presented here show that the polarization of a spatially unresolved exoplanet may be detected by cross-correlating high-dispersion linear polarization and intensity (brightness) spectra of the planet-star system. The Doppler shift of the planet-reflected starlight facilitates the separation of this signal from the polarization introduced by the interstellar medium and the terrestrial atmosphere. The selection of the orbital phases and wavelengths at which to study the planet is critical. An optimal choice, however, will partly depend on information about the atmosphere that is a priori unknown. We elaborate on the cases of close-in giant exoplanets with non-uniform cloud coverage, an outcome of recent surveys of brightness phase curves from space, and for which the hemispheres east and west of the substellar point will produce different polarizations. With integration times of the order of hours at a 10 m telescope, the technique might distinguish among some proposed asymmetric cloud scenarios with fractional polarizations of 10 parts per million for one such planet orbiting a V-mag = 5.5 host star. Future 30-40 m telescopes equipped with high-dispersion spectro-polarimeters will be able to investigate the linear polarization of smaller planets orbiting fainter stars and look for molecular features in their polarization spectra.