Stephen M. Seddio

Petrology and geochemistry of lunar granite 12032,366-19 and implications for lunar granite petrogenesis

Abstract Apollo 12 sample 12032,366-19 is a 21.3 mg granite fragment that is distinct from any other lunar granite or felsite. It is composed of barian K-feldspar, quartz, sodic plagioclase, hedenbergite, fayalite, and ilmenite, with trace amounts of zirconolite, baddeleyite, apatite, and merrillite. The texture of 12032,366-19 is largely a micrographic intergrowth predominantly of K-feldspar and quartz and, to a lesser extent, plagioclase and quartz. Hedenbergite, fayalite, and ilmenite are present in minor but significant quantities-6.0, 3.1, and 1.7 wt%, respectively-and are scattered throughout the feldsparquartz intergrowths. Trace amounts of Zr-bearing phases are found including zirconolite (0.6 wt%) and baddeleyite (0.04 wt%). Incompatible trace-element concentrations are high in 12032,366-19, particularly the high-field-strength elements, e.g., Zr, Sm, and Th (1500, 25, and 61 μg/g, respectively). The chondrite-normalized, rare-earth-element concentrations form a “V-pattern” that is characteristic of other lunar granitic material. By modeling 12032,366-19 as a derivative from a KREEP-like parent melt, the composition and mineral assemblage can be obtained by extended fractional crystallization combined with separation of the low-density minerals plus trapped melt components prior to final solidification. However, this model cannot quantitatively account for the relatively sodic composition of the plagioclase (An34-50) and requires that the starting melt has Na2O of 1.2-1.4 wt%, which is higher than most KREEP compositions. Formation of this assemblage by silicate-liquid immiscibility is neither required nor indicated by petrogenetic modeling.

Silica polymorphs in lunar granite: Implications for granite petrogenesis on the Moon

Abstract Granitic lunar samples largely consist of granophyric intergrowths of silica and K-feldspar. The identification of the silica polymorph present in the granophyre can clarify the petrogenesis of the lunar granites. The presence of tridymite or cristobalite would indicate rapid crystallization at high temperature. Quartz would indicate crystallization at low temperature or perhaps intrusive, slow crystallization, allowing for the orderly transformation from high-temperature silica polymorphs (tridymite or cristobalite). We identify the silica polymorphs present in four granitic lunar samples from the Apollo 12 regolith using laser Raman spectroscopy. Typically, lunar silica occurs with a hackle fracture pattern. We did an initial density calculation on the hackle fracture pattern of quartz and determined that the volume of quartz and fracture space is consistent with a molar volume contraction from tridymite or cristobalite, both of which are less dense than quartz. Moreover, we analyzed the silica in the granitic fragments from Apollo 12 by electron-probe microanalysis and found it contains up to 0.7 wt% TiO2, consistent with initial formation as the high-temperature silica polymorphs, which have more open crystal structures that can more readily accommodate cations other than Si. The silica in Apollo 12 granitic samples crystallized rapidly as tridymite or cristobalite, consistent with extrusive volcanism. The silica then inverted to quartz at a later time, causing it to contract and fracture. A hackle fracture pattern is common in silica occurring in extrusive lunar lithologies (e.g., mare basalt). The extrusive nature of these granitic samples makes them excellent candidates to be similar to the rocks that compose positive relief silicic features such as the Gruithuisen Domes.

Considerations for the Acquisition of Very Large Area EDS Spectral Image Mosaics

Modern SEMs, EDS detectors, and microanalysis software have enabled the acquisition of EDS spectral images over very large areas so that entire samples may be described by composition and phase maps (e.g., [1]). Here, we investigate what count richness (X-ray counts/pixel) is needed to meaningfully extract different map types (counts, quantitative, and phase maps) from spectral image mosaic datasets.

Very Large Area Phase Mapping of a Petrographic Thick Section using Multivariate Statistical Analysis of EDS Spectral Images

There are many applications that require knowing the correct distribution of phases in a sample, such as determining the quality of steel or determining the bulk composition of a sample by modal recombination. Modal recombination is a calculation of a sample’s bulk chemistry based on the assumption that the distribution of phases in a cross section of a sample is a good representation of the distribution of phases in the three dimensional volume of the sample. If the abundance of each phase and the composition of each phase are known, the bulk composition of the sample can be calculated. The greater the area of the section from which the distribution of phases is determined, the more accurate the bulk composition calculated from a modal recombination can be. Large area electron imaging, elemental X-ray mapping, and quantitative elemental X-ray mapping have existed for some time. Historically, most modal recombination calculations have been done relying on the phase abundances determined from the analysis of contrast in backscattered electron image which discriminates phases based on average atomic number. However, many phases that commonly occur in the same sample have similar average atomic numbers (e.g., pyroxene and apatite) and may be impossible to successfully discriminate based on electron image contrast. Here, we demonstrate that modern X-ray microanalytical software, if it includes multivariate statistical analysis algorithms, enables phase mapping based on X-ray microanalysis of very large areas (including petrographic thin or thick sections) in practical amounts of time.