Transmission Spectra as Diagnostics of Extrasolar Giant Planet Atmospheres

Atmospheres of transiting extrasolar giant planets (EGPs) such as HD 209458b must impose features on the spectra of their parent stars during transits; these features contain information about the physical conditions and chemical composition of the atmospheres. The most convenient observational index showing these features is the "spectrum ratio" ℜ(λ), defined as the wavelength-dependent ratio of spectra taken in and out of transit. The principal source of structure in ℜ is the variation with wavelength of the height at which the EGP atmosphere first becomes opaque to tangential rays—one may think of the planet as having different radii, and hence different transit depths, at each wavelength. The characteristic depth of absorption lines in ℜ scales with the atmospheric scale height and with the logarithm of the opacity ratio between continuum and strong lines. For close-in EGPs, line depths of 10^-3 relative to the stellar continuum can occur. The atmospheres of EGPs probably consist mostly of molecular species, including H₂, CO, H₂O, and CH₄, while the illuminating flux is characteristic of a Sun-like star. Thus, the most useful diagnostics are likely to be the near-infrared bands of these molecules, and the visible/near-IR resonance lines of the alkali metals. I describe a model that estimates ℜ(λ) for EGPs with prescribed radius, mass, temperature structure, chemical composition, and cloud properties. This model assumes hydrostatic and chemical equilibrium in an atmosphere with chemistry involving only H, C, N, and O. Other elements (He, Na, K, Si) are included as nonreacting minor constituents. Opacity sources include Rayleigh scattering, the strongest lines of Na and K, collision-induced absorption by H₂, scattering by cloud particles, and molecular lines of CO, H₂O, and CH₄. The model simulates Doppler shifts from height-dependent winds and from planetary rotation, and deals in a schematic way with photoionization of Na and K by the stellar UV flux. Using this model, I investigated the diagnostic potential of various spectral features for planets similar to HD 209458b. Clouds are the most important determinants of the depth of features in ℜ; they decrease the strength of all features as they reach higher in the atmosphere. The relative strengths of molecular lines provide diagnostics for the heavy-element abundance, temperature, and the vertical temperature structure, although diagnostics for different physical properties tend to be somewhat degenerate. Planetary rotation with likely periods leaves a clear signature on the line profiles, as do winds with speeds comparable to that of rotation. Successful use of these diagnostics will require spectral observations with signal-to-noise ratio (S/N) of 10³ or better and resolving power R = λ/δλ ranging from 10³ to 10⁶, depending on the application. Because of these stringent demands, it will be important to evolve analysis methods that combine information from many lines into a few definitive diagnostic indices.