2.3.8. Mineral Recipe

The Mineral recipe is the central part of the Mineralogic Mining plugin. It regulates the EDS and mapping type parameters, holds the list of minerals, and allows for morphochemical classification criteria (Figure S6). The Mineralogic analysis can be performed in five di fferent ways:


classification might be generated. The method can be nearly as fast as a spot centroid analysis, depending on the pre-set amount of counts in the spectrum.


For all EDS-based methods, the EDS spectrum deconvolution can be specified in the same way as for regular EDS analyses (Figure S6a). Thus, elements can be excluded from quantification (e.g., carbon used for coating is set to deconvolution-only), matrix quantification methods can be chosen (ZAF vs. Phi-Rho-Z), spectra output data can be normalized, and elements can be chosen to be always or never included, if wished for, thus potential sources of peak overlap can be avoided. EDS dwell time, and detector throughput rate can be chosen by the user. Every single generated spectrum during the analysis is fully quantified and the weight percentages of each element in each pixel on the false coloured mineral map is available after the analysis. User can choose to add standards-based quantification of the spectra. These latter two points are unique for the Mineralogic software. The matrix quantification for each spectrum also allows the user to switch between di fferent acceleration voltages in between samples, without the need to specify a new mineral list. Most AQM systems require the user to make separate mineral lists for each acceleration voltage.

The mineral list for the sample analysis is created by the user based on the element wt% of the mappable phases and can be used to produce a variety of visualizable informative image and data outputs. For each mineral the range of tolerated concentrations for each element, as well as for ratios between two elements, can be used to define the mineral phase (Figure S6b). Minerals are placed in a list, which is checked against the chemistry of each analysed point, following a first-match principle and based on the order of minerals in the mineral list, which is defined by the operator. The same principle to build mineral lists can also be used to define element concentration maps or element ratio maps, where individual phases are identified and coloured by the concentration of one or a few elements. Mineral lists can be exported and imported between projects and adjusted to fit the exact mineral specifics for the sample area of interest, allowing for slight di fferences in mineral chemistry resulting from di fferent whole rock compositions or metamorphic temperatures.

In an addition tab in the recipe, minerals can be clustered into groups (e.g., albite, labradorite and anorthite as plagioclase). Clustered minerals can be integrated as a group when particle data is exported after the analysis.

Additionally, morphological criteria can be used in the classification of minerals. For example, textural properties can be defined based on the chemically mapped grains, where the grain size or shape properties are used to synergize the chemical and textural characteristics. For example, zircons of a size large enough to be dated with a laser connect to a mass spectrometer can be classified from the main zircon population separately. All di fferent ways of describing the sample (e.g., element map, mineral map, morphochemical map) can be calculated o ffline from the generated data. There is no need to reanalyze the sample after the mineral list was changed.

Many parameters a ffect the speed and quality of an AQM analysis, and a Mineralogic analysis is no di fferent. Each analysis must be optimized prior to an AQM run. Di fferent types of analyses (see above) will yield a di fference in analysis speed and data quality. Acceleration voltage and aperture size both control how much signal reaches the EDX detectors, where a high acceleration voltage and a large aperture size result in a greater signal. These parameters are typically tailored to the nature of the question (i.e., the data required), according to which the resolution and speed can be adapted. Detector throughput rate can be adjusted to optimize the measurement accuracy and precision. The throughput amount (i.e., counts input) impacts the elemental energy peaks full width half maximum (FWHM) and therefore the accuracy and precision. The dwell time (time the beam stays at one spot before moving on) has a major impact on the analysis speed. The lower a dwell time the faster the analysis is, however if the dwell time is too low it will result in insufficient spectrum counts for a good quality analysis. Image capture time, determining the quality of the BSE image, and frame magnification, determining the amount of stage movements, also affect the analysis time, but these effects are minor and have no real influence during a mapping analysis. The last parameter to affect the analysis speed, is the step size (pixel size) in the mineral map, where a smaller step size will increase the analysis time, but at the same time will provide a more detailed in-depth analysis.
