On the detectability of rocky exoplanet surfaces

For the interpretation of upcoming observations of terrestrial — potentially habitable — planets, it is not only crucial to understand how atmospheric species imprint their spectral signatures onto planetary emission spectra, but also whether spectral features of rocks and minerals of various types may be detectable.

We are presenting first results from our new radiative transfer framework, which takes the radiative properties of both the atmosphere and the surface into account. Taking the hot rocky super-Earth LHS 3844b as benchmark — motivated by the recent findings of Kreidberg et al. (2019) — we explore the detectability of various plausible rocky surfaces with JWST. Since the measured Spitzer phase curve of this planet indicates that this planet does not possess a thick atmosphere, the surface should in principle be detectable. Hence, this planet is a first benchmark target for future characterization efforts of exoplanetary geology. However, even atmospheres that are optically thin may provide sufficient extinction within molecular absorption bands to mask the surface.

In order to predict realistic planetary spectra stemming from both atmospheres and surfaces, we a have constructed a large grid of plausible atmosphere/surface models. We have considered O₂- and N₂-dominated atmospheres with various trace species, such as CO₂, SO₂, CO and H₂O, all potential additives from volcanism, and surfaces of basaltic, granitoid and feldspathic crust, as are commonly found in the solar system bodies. Using the JWST noise simulator Pandexo to simulate LHS 3844b secondary eclipse observations we have found the following observables:

(i) The surface albedo in the near-infrared plays a crucial role for the planetary spectrum in that wavelength range. We find that the differences both in the reflected as well as the emitted part of the spectrum are large enough to be detectable with the NIRSpec/G395M instrument. In addition, the surface has a significant effect on the atmospheric temperature profiles and surface temperatures. For instance, due to its high reflectivity, the feldspathic surface leads to surface temperatures up to 60 K cooler compared to the other cases.

(ii) In the mid-infrared, the surface will be detectable with MIRI/LRS. Since both O₂ and N₂ are weak absorbers, there a number of spectral windows in which the planetary emission follows the surface features. For instance, two sweet spots are located at 5–7 micron and 9.5–12 micron. This coincides well with the Si-O bands located at ~ 8–12 microns, resulting in prominent rocky features.

Although focusing on a very hot case study, we believe that our results help gain a broader understanding of the emission spectra of terrestrial planets, as the explored atmospheric compositions and surfaces are typical for a broad range of rocky planets. From the current surface geology we can start learning about the internal structure and past evolution of these planets.