¹JeffA.Berger,²M.E.Schmidt,¹J.L.Campbell,¹E.L.Flannigan,¹R.Gellert,³W.Ming,³R.V.Morris
Icarus (in Press) Link to Article [https://doi.org/10.1016/j.icarus.2020.113708]
¹University of Guelph, Guelph, Canada
²Brock University, St. Catharines, Canada
³NASA Johnson Space Center, Houston, USA
Copyright Elsevier

The Alpha Particle X-ray Spectrometer (APXS), a field instrument onboard four martian rovers, measures largely unprepared, in situ samples on Mars. The APXS has high precision that enables the determination of elemental concentrations in a wide range of geologic materials. However, lack of sample preparation can lead to heterogeneous matrix effects, and understanding the associated uncertainty is essential for interpreting APXS data. Here we use Particle Induced X-ray Emission spectrometry (PIXE) to analyze a suite of geologic samples from Hawai’i as an analogue study to better understand APXS analyses of martian samples. Wavelength-Dispersive X-ray Fluorescence (WDXRF) analyses of fused glass beads establish higher-accuracy standards for the Hawaiian samples. Sulfate-silicate mixtures were made to evaluate sulfur analysis by PIXE. Results show that the PIXE concentrations for most major elements have 2–6% accuracy, which is comparable to the APXS. However, the PIXE concentrations are systematically high in Al and low in Mg, resulting in lower accuracy (13% and 20%, respectively). Olivine-phyric lavas and most of their altered products have the largest discrepancies with Al concentrations up to 25% high and Mg up to 35% low. Sulfur is systematically high (up to 30% in a basalt matrix) compared to gravimetric S concentrations in the sulfate-silicate mixtures. These systematic deviations in Mg, Al, and S are linked to heterogeneous matrix effects, because PIXE and APXS analyses assume all atoms in a sample to be homogeneously mixed on the sub-micrometer scale, which is not the case. Two key implications for APXS results are: (1) Olivine-bearing samples likely have reported concentrations of Mg that is too low and Al that is too high. Thus, olivine-phyric basalts in Gusev crater and the basaltic sand and soil at three landing sites may have Mg and Al concentrations closer to those of the olivine-phyric shergottites and modelled martian crust than previously thought. (2) Sulfate-silicate mixtures may have overestimated S concentrations reported, resulting in greater uncertainty in the stoichiometry of Ca-sulfates, which is used to deduce the geochemical associations of sulfur in samples.

¹Xiaoqing Xu,^1,2Hejiu Hui,³Wei Chen,⁴Shichun Huang,⁵Clive R.Neal,¹XishengXu
Earth and Planetary Science Letters 536, 116138 Link to Article [https://doi.org/10.1016/j.epsl.2020.116138]
¹State Key Laboratory of Mineral Deposits Research & Lunar and Planetary Science Institute, School of Earth Sciences and Engineering, Nanjing University, Nanjing 210023, China
^sCAS Center for Excellence in Comparative Planetology, Hefei 230026, China
³State Key Laboratory of Geological Processes and Mineral Resources, School of Earth Sciences, China University of Geosciences, Wuhan 430074, China
⁴Department of Geoscience, University of Nevada, Las Vegas, NV 89154, United States
⁵Department of Civil and Environmental Engineering and Earth Sciences, University of Notre Dame, Notre Dame, IN 46556, United States
Copyright Elsevier

The lunar magma ocean (LMO) model was proposed after the discovery of anorthosite in Apollo 11 samples. However, the chemical and isotopic compositions of lunar anorthosites are not fully consistent with its LMO origin. We have analyzed major and trace elements in anorthositic clasts from ten lunar feldspathic meteorites, which are related to the solidification of the LMO. The plagioclase rare earth element (REE) abundances and patterns are not correlated with the Mg# of coexisting mafic minerals in anorthosites, implying that mafic minerals and plagioclase may not be in chemical equilibrium, consistent with their textural differences. The REE abundances in plagioclase range approximately fortyfold, which cannot be produced by fractional crystallization of a single magma. Combining plagioclase trace element data from Apollo and meteoritic anorthosites, we propose that plagioclases derived from the LMO floated to the surface to form the primordial crust, which then may have been metasomatized by incompatible-element-rich KREEP (potassium, rare earth element, phosphorus) melts and mantle-derived partial melts. The lunar anorthosites may represent this metasomatized crust rather than solely a derivative from the LMO. Furthermore, silicate melts similar to the metasomatic agents may also have melted the crust to form the Mg-suite rocks. This hypothesis is consistent with overlapping ranges of age and initial ε_Nd between lunar anorthosites and Mg-suite rocks. These events are consistent with an overturn event of the cumulate mantle very early after primordial crust formation to produce the partial melts that metasomatized the crust.