Earth system science applications of next-generation SEM-EDS automated mineral mapping

Abstract

Sedimentary rocks contain a unique record of the evolution of the Earth system. Deciphering this record requires a robust understanding of the identity, origin, composition, and post-depositional history of individual constituents. Petrographic analysis informed by Scanning Electron Microscope - Energy Dispersive Spectroscopy (SEM-EDS) mineral mapping can reveal the mineral identity, morphology and petrological context of each imaged grain, making it a valuable tool in the Earth Scientist’s analytical arsenal. Recent technological developments, including quantitative deconvolution of mixed-phase spectra (producing “mixels”), now allow rapid quantitative SEM-EDS-based analysis of a broad range of sedimentary rocks, including the previously troublesome fine-grained lithologies that comprise most of the sedimentary record. Here, we test the reliability and preferred mineral mapping work flow of a modern Field-Emission scanning electron microscope equipped with the Thermofisher Scientific Maps Mineralogy mineral mapping system, focusing on mud/siltstones and calcareous shales. We demonstrate that SEM-EDS mineral mapping that implements 1) a strict error minimization spectral matching approach and 2) spectral deconvolution to produce ‘mixels’ for mixed-phase X-ray volumes can robustly identify individual grains and produce quantitative mineralogical data sets comparable to conventional X-ray diffraction (XRD) analysis (R2 > 0.95). The correlation between SEM-EDS and XRD-derived mineralogy is influenced by mineral abundance, processing modes and mapped area characteristics. Minerals with higher abundance (>10 wt%) show better correlation, likely the result of increased uncertainty for XRD quantification of low-abundance phases. Automated spectral deconvolution to produce ‘mixels’ greatly reduces the proportion of unclassified pixels, especially in the fine-grained fraction, ultimately improving mineral identification and quantification. Mapping of larger areas benefits bulk mineralogy analysis, while customized area size and shape allows high-resolution in situ mineralogical analysis. Finally, we review SEM-EDS-based mineral mapping applications in the Earth Sciences, via case studies illustrating 1) approaches for the quantitative differentiation of various mineral components including detrital (allogenic), syndepositional (authigenic) and burial diagenetic phases, 2) the origin and significance of lamination, 3) the effectiveness and appropriateness of sequential leaching in geochemical studies, and 4) the utility of mineral maps to identify target grains within specific petrological contexts for in situ geochemical or geochronological analysis.

Hydration of MgO: Reaction Kinetics and pH Control on Brucite Crystal Morphology

“…Powdered samples were used to record X-ray diffraction (XRD) patterns using a PANalytical X’Pert PRO diffractometer (Co–Kα radiation; 40 mA, 40 kV) in the range from 5° to 85° 2θ with a step size of 0.017° 2θ and a count time of 15 s per step for each scan, for the identification of the mineralogical composition of the MgO raw materials and its reaction products. For mineral identification and quantification, the PANalytical X’Pert Highscore Plus software (version 3.0.4) and the ICSD database were used, with an analytical error of <3 wt % . Specifically, the relative abundance of MgO versus Mg­(OH) 2 as well as the relevant unit cell parameters ( c - and a -axes) of Mg­(OH) 2 were considered.…”

Microscale Petrographic, Trace Element, and Isotopic Constraints on Glauconite Diagenesis in Altered Sedimentary Sequences: Implications for Glauconite Geochronology

“…Subsequently, regions of interest (10–80 in total) of potentially pristine or altered glauconites in each sample were selected for high‐resolution BSE imaging (10 nm pixel resolution) and mineral mapping (EDS spectra: 1.5 μm step size, 8 ms acquisition time). Quantitative mineral mapping was carried out using FEI Maps Mineralogy software for automated data collection (both BSE imaging and EDS), whereas the FEI Nanomin software was utilized for the classification of the individual EDS spectra to identify minerals and calculate mineral abundances (for details see Abbott et al., 2019; Frank et al., 2020; Han et al., 2022; Rafiei et al., 2020).…”

“…Rb is measured on‐mass while Sr isotopes are measured as mass‐shifted oxides, allowing quantitative on‐line separation of 87 Rb and 87 Sr. When combined with recent advances in sedimentary petrography (e.g., Han et al., 2022; Rafiei & Kennedy, 2019; Rafiei et al., 2020) this approach permits rapid in situ Rb‐Sr dating of carefully screened glauconite grains (Farkas et al., 2018), potentially resolving the limitations of traditional glauconite geochronology and opening up a range of new applications (Scheiblhofer et al., 2022). However, despite the promise of this new approach, glauconite remains susceptible to burial alteration (Bansal et al., 2020; Guimaraes et al., 2000).…”

A free, open-source method for automated mapping of quantitative mineralogy from energy-dispersive X-ray spectroscopy scans of rock thin sections

“…Converting the mineral maps into vector form allows for the calculation of derived parameters such as median grain area for minerals that occur as single grains (e.g., biotite), distance between grains of a mineral, and the types of minerals surrounding a grain or grains in the case of abundant, connected minerals like plagioclase and quartz. This type of data is normally generated by proprietary automated mineralogy systems but could aid in geoscience disciplines beyond ore geology or petroleum geology (Han et al, 2022). An illustrative example is in the analysis of grainscale properties of biotite.…”

“…In practice, automated mineralogy methods use a combination of image analysis and classification methods to identify mineral phases from elemental composition data (or their derivatives), which can be collected with a variety of analytical methods, including energy dispersive X-ray spectroscopy (EDS), wavelength dispersive Xray spectrometry (WDS), micro-X-ray fluorescence (μ-XRF), and laser-ablation inductively-coupled mass spectroscopy (LA-ICP-MS) (Nikonow et al, 2019). Automated mineralogy is being slowly adopted by researchers outside of resource extraction for combined modal analysis of bulk mineralogy, estimates of grain size distribution, and mineral association (Han et al, 2022), which can be useful in a variety of disciplines such as petrology, applied geochemistry, and rock mechanics (Sajid et al, 2016;Elghali et al, 2018;Rafiei et al, 2020).…”