Selection of chemical species for Europa’s surface using Galileo/NIMS
Résumé
Europa’s surface shows evidence of active resurfacing which can be explained by either endogenic or exogenic
processes. Apart from water ice, several compounds have been proposed to account for the potential complex
chemistry that could take place if a connection with the subsurface salty ocean would occur. Previous
spectroscopic studies that investigated the surface composition were limited by the number of compounds
to consider due to the unavailability of laboratory measurements. We now have access to optical constants
and laboratory spectra of synthetic chemical compounds such as hydrated sulfates and chlorinated salts under
Europa’s surface conditions.
In this study, we test for the first time the relevance of 15 potential endmembers on a Galileo/NIMS
observation of a dark lineament of the Trailing Anti-Jovian hemisphere using a realistic radiative transfer
modeling and a robust Bayesian inference framework. We consider an intimate mixture of 3, 4 and 5
endmembers and perform a fitting procedure on all possible combinations among the list of 15 compounds in
the 1.0–2.4 μm spectral range. At the end, 5000 combinations have been tested and analyzed. The parameters
we are fitting for are the surface roughness, the volumetric abundances and the grain size. Given the fact
that NIMS noise level and absolute calibration are uncertain, we have distinguished two scenarios: assuming
a pessimistic signal-to-noise ratio (SNR) of 5 and an optimistic SNR of 50. For each scenario, we use several
criteria to highlight the contribution of each endmember and thus identify the most relevant compounds.
We first show that no combinations using only 3 endmembers are able to accurately reproduce the
observation. Then, when considering 4 and 5 endmembers together and assuming the SNR=50, there is 21
and 153 combinations that evenly reproduce the observation. From this result, we discuss the relevance of
each endmember using criteria such as the Root-Mean-Square (RMS) deviation of the best-fits, the spectral
contribution at each wavelength, the occurrences distribution of the selected best-fits and the numerical
abundances retrieved by the fitting procedure. By grouping criteria together, we show that only sulfuric acid
octahydrate and water ice appear as essential compounds. Then, hydrated sulfates are in general preferred
over others compounds. This result is only valid for a lineament of the Trailing Anti-Jovian hemisphere and
can be possibly extended to other lineaments in the same hemisphere. However, we show that chlorinated
salts but also ammonium sulfate mascagnite ((NH4)2SO4), sodium chloride (NaCl) and even magnetite (Fe3O4)
are not excluded, their contributions from the selected criteria are just less pronounced than hydrated sulfates.
Nevertheless, it is clear that it is not possible to correctly distinguish between these endmembers, and even
less between endmembers with similar chemical composition, at these wavelengths. We therefore propose for
future investigation to either consider a mixture that would include only one representative of each chemical
compound, rather than using all possible endmembers and overfitting the data.