Predicting Iron Contents from Venus-Temperature Lab Emissivity Spectra: Insights into Igneous Rock Type

Abstract:

Orbital observations through transparent windows in the CO₂ atmosphere of Venus produce viable spectra of the surface below. In Venus missions, spectra can be collected through five windows near 1 μm that coincidentally lie in a wavelength region where visible-near infrared (VNIR) features of both Fe²⁺ and Fe³⁺ occur. These data have the potential to provide understanding of Venus rock type in a manner analogous to the eight-filter imaging on the Pancam instrument of the Mars Exploration rovers. To interpret them, high temperature laboratory data of appropriate well-characterized standards are needed.

In this project, 53 rocks covering a wide range of igneous provenances were characterized using bulk rock x-ray fluorescence for geochemistry and prepared for analysis. Emissivity measurements from 0.7 to 1.5 μm were acquired on powders of various particle sizes and 5 cm disks (slabs 0.5-1 cm thick) of each sample at the Planetary Spectroscopy Laboratory at DLR at temperatures comparable to those to be experienced at the Venus surface (400-480°C).

Igneous rock types are generally classified on the basis of Si, Na, and K using the total alkali versus SiO₂ (TAS) diagram. Although those elements are not active in the VNIR at 1 μm, both Fe²⁺ and Fe³⁺ have features arising from glass, olivine, and pyroxene near that region. To determine the relationship between Fe content and rock type, we contoured a TAS diagram with FeO contents of 486,629 global non-marine volcanic (GeoRoc Database) and oceanic rocks (PetDB Database). Of these, 94.1% lie along the sub-alkaline trend from tephrite/trachybasalt (x̅=11.55/10.39 wt% FeO), picrobasalt/basalt (x̅=9.50/9.86) to basaltic andesite/trachyandesite (x̅=8.69/6.52), andesite/trachyandesite (x̅=6.47/3.68), dacite/trachyte (x̅=4.31/2.29), and rhyolite (x̅=2.19), allowing rock type to be inferred from FeO.

We constructed a partial least squares model to predict FeO contents using Venus-T emissivity spectra. Results indicate that total FeO content of sub-alkaline rocks can be measured with a high degree of accuracy ca. ±0.5 wt% FeO, sufficient to discriminate among relevant rock types. Expanding our spectral library will improve the ability of these and other models, including those based on spectral slopes and band ratios, to distinguish mafic from silicic rocks.