QACD-quack

Consider a gabbroic rock that is observed to be variably textured and has subsequently been subsampled for serial thin sections. Petrographic analysis of the thin sections reveals several distinct textural domains. Full thin section element maps have been collected for each section with the goal of examining the distribution of An (calculated as molar Ca / (Ca + Na + K)) and Mg# (Mg / (Mg + Fe)) in plagioclase and pyroxene respectively in order to narrow down the potential origin of the textural domains. The supplied software used to collect the XSIs has been used to extract background corrected maps for individual elements and classify phases, but there is no means of carrying out the calculations needed for the creation of ratio maps like Mg# and An. Further, the software has the ability to quantify the bulk element maps for the analysed area, but lacks the ability to apply different quantitative corrections to each of the extracted phase maps. In order to accommodate the large number of full-thin section quantitative element maps needed for such an investigation, we require both a method for the rapid collection of element maps and software which could facilitate the efficient post-processing of element maps collected by EDS-SEM. In particular, the ‘varitextured’ nature of such samples require software which would facilitate post-processing methods to produce a fully quantitative assessment of compositional distribution for element maps without the need of multiple software platforms and allow for the flexibility to easily adapt the software’s functionality, which none of the existing platforms provide.

​

This contribution presents a method for the processing of large area X-ray element maps obtained by EDS to produce a quantitative assessment of compositional distribution (QACD) of crystal populations within geologic materials. Such maps can be used to accurately identify all phases and calculate mineral modes for a sample. The QACD method has been incorporated into a python-based easy-to-use graphical user interface (GUI) called Quack (see section 3 for an overview), which facilitates every step of the QACD process from processing for ‘bad’ pixels and noise in maps through manual/automatic phase identification, calculation of modal phase abundances, quantification of elements and molar ratios without the need for standard analyses, and the production and export of both element/ratio maps and histograms. Most importantly, the Quack software allows for users to quickly quantify elemental concentrations and molar ratios for individual phases instead of having to quantify the entire map with a single broad standardisation or several times over for mineral-specific standardisations. The python code for Quack allows for users to adapt the software to their needs and more readily create new functionality when it is needed, thus aiming to curb the perpetual development of more disparate software methods.

​

By optimising the conditions at which the EDS X-ray element maps are acquired, we are able to obtain fully quantitative element maps for most major elements. Although fully quantified maps of absolute element distributions typically require some form of standardisation prior to analysis, the X-ray element maps acquired by our method can be processed without the analysis  of standards to create maps of absolute element ratios (e.g. An in plagioclase, Mg# in pyroxene) accurate to within a 2σ error of their accepted values. Instead, corrections for individual elements and common element ratios are derived from a growing database of standard analyses and models calculated from first principles. This method is applied to maps of standard materials (e.g. Albite, Plagioclase, Diopside, and Enstatite), a zoned plagioclase within a dolerite, and a mafic tuff. By applying the QACD method of processing to these maps, we are able to separate out phases of interest and assess the nature and distribution of chemical populations across a given rock. The resulting element ratio maps and chemical population histograms provide a means of determining the relative compositions and volumes of melts which contributed to the crystallisation history of the rock.