Although micro-XRF as an analytical technique was developed over 20 years ago, it is with the continuous advancement of computing and hardware technology that it has become more powerful than ever and a routine part of many geoscience characterization workflows. A principal reason for this is the systems capability to analyse large samples at micrometer scales with very minimal sample preparation. This flexibility makes the micro-XRF system ideal for analysing field samples (e.g., hand specimens, drill cores) in the laboratory, and thus easily and timeously enabling relevant decisions to be made about up- or down-scaling information or additional sample analysis in any given workflow. In addition, micro-XRF can analyse a range of sample sizes, from large specimens (over 10´s of centimeters) to those prepared as the commonly used polished thin sections or epoxy briquettes. Furthermore, results of micro-XRF analysis range from major, minor and trace elemental chemistry in semi- and fully-quantified form, to derived mineralogy, thus yielding data-rich information across a range of geoscience fields such as petrology, sedimentology, geochemistry, paleontology, economic geology, amongst others. The visualization of the elemental chemistry and mineralogy on such a large scale is extremely intuitive and relevant in geosciences, as it enables the user to directly link the sample’s visual structure to its chemistry. This presentation will review these capabilities in the world of geoscience and discuss the possibilities for the future.

µXRF is a versatile technique that has been used in various geoscientific fields, particularly for the mapping of larger hand specimen or drill cores. It is easy to use, non-destructive and requires only little sample preparation, enabling the acquisition of 2D element distribution and mineral identification via EDX spectra. However, a limitation is the diffraction of the X-ray beam by the crystal lattice, which can produce peaks that overlap with actual element peaks, thereby affecting chemical quantification and mineral identification. Previous research, such as Nikonow et al. (2016), has demonstrated methods to eliminate these diffraction peaks from µXRF spectra. Additionally, diffraction can be used to identify individual grains within a mono-mineralic domain, such as quartz, without the need for thin section preparation. Since diffraction depends on the angle between the crystal lattice and the X-ray beam, differently oriented grains will produce diffraction peaks at different energies in the spectrum and can be distinguished from each other in the energy-dispersive µXRF spectrum. This technique enables not only the identification of optically similar minerals in the drill core (e.g. magnetite and ilmenite), but also the extraction of grain shapes and measurement of their 2D size, area, and orientation in the cutting plane and allows for quantitative textural analysis, helping to understand e.g. igneous processes (Higgins 1998). The application of this method is demonstrated on drill core sections from magnetitite layers of the Upper Zone of the Bushveld Igneous Complex, South Africa.

Micro X-ray Fluorescence (MXRF) spectrometers enable non-destructive and elemental analysis of a wide variety of solid samples with a lateral resolution of a few tens of micrometers using focusing optics. Confocal MXRF (CMXRF) offers additional depth-dependent measure­ment capabilities, based on a three-dimensional probing volume created by the confocal ar­rangement of a focusing lens in the excitation and detection channel. Therefore, CMXRF pro­vides micro-scalic resolved measurements of complete sample volumes by depth profiles (1D), cross section mappings (2D) and stacked element distribution images (3D). The strengths, challenges and potential of a modified (confocal) MXRF tabletop spectrometer for non-destruc­tive and depth-sensitive element analysis will be illustrated by geoanalytical and geo-related appli­cations:

The first application is the three-dimensional analysis of mineral inclusions. The sophisticated compositional studies and identification of mineral phases by CMXRF provides micro-scalic resolutions with certain limitations due to X-ray absorption, but also preserves the integrity of isolated inclusions for further analysis.

Another example is the study of biomineralization products, due to the biomimetic behavior of deep-sea sponges under extreme conditions resulting in the formation of novel three-dimen­sional composites. Several mineralization products such as atacamite, goethite and lepido­crocite have been studied by three-dimensional reconstruction of the elemental distribution of the formed composites.

The third application is the depth-sensitive analysis of the elemental composition of ce­ment stone corrosion zones simulating the acidic chemical attack on concrete samples. The interest in describing those corrosion processes is motivated by defining the occurring kinetics and deriving information about the persistence, strength and durability of concrete.

Micro-X-ray fluorescence spectroscopy (XRF) represents a well-established and complementary analytical technique to electron beam energy dispersive spectroscopy (EDS) for the detailed characterization of elemental composition in samples. The integration of the X-ray source (namely XTrace) facilitates the application of XRF technology within a scanning electron microscope (SEM). Micro-XRF excitation analysis is a specialized small-area/volume technique, particularly suitable for beam-sensitive samples due to the absence of charging effects. The technique offers significant advantages, including enhanced sensitivity for trace element detection, the capability to excite higher energy X-ray lines (spanning a full spectral range to 40 keV), and the acquisition of information from greater sample depths even in centimeter level.

The deployment of advanced X-ray polycapillary optics enables the focal spot size of the X-rays to be reduced to 10 microns, all within an X-ray source compatible with SEM ports. X-ray energy detection is performed using the existing EDS detector integrated into the SEM system. Consequently, the SEM system attains dual-source capability, encompassing both electron and X-ray sources (as illustrated in Fig. 1), thereby expanding the possibilities for material characterization. This dual-source capability is termed "Full Range EDS," leveraging the novel analytical potential arising from the combined dual excitation of micro-XRF and electron beam sources alongside an EDS detector. This dual-beam system, allowing samples to interact with either the SEM's electron beam, the XTrace’s X-ray photons, or both simultaneously. Full Range EDS confers numerous advantages over traditional EDS, providing researchers with deeper insights into the elemental and compositional intricacies of their samples.

Brandenburg’s surface geology predominantly consists of Quaternary sediments, with sequences averaging 50 to 80 meters (locally up to 500 meters) in thickness. Research up to 2008 on heavy mineral composition facilitated the lithostratigraphic classification of fluvial deposits, revealing frequency and compositional variations. Stratigraphic classification in Brandenburg relies primarily on pollen analysis of interglacial, predominantly limnic deposits, and small-scale gravel counts of (glacio-)fluvial and till sediments, leaving sandy components unrepresented methodologically.

To establish a comprehensive provenance analysis, the method development presented here includes both the heavy and light mineral fractions. The geochemical composition of the samples is determined semi-quantitatively using spectral analysis. In this project, 24 sand samples from the drill core Kb Borgisdorf 1/06 were examined to reconstruct the distribution patterns of Saale Late Glacial to Weichsel Early Glacial sediments in Brandenburg. The focus of method development is on sand deposits that cannot be classified by pollen and clast analysis. All samples were prepared for both polarization microscopy and Mineral Liberation Analysis (MLA), maintaining a grain size range of <200 μm. This saves time and provides a comprehensive dataset that is better comparable with conventional analyses. The data produced by the MLA are compiled into large databases and statistically analyzed, utilizing mineralogy and grain parameters such as size, length, width, and roundness. By comparing with comprehensive geochemical and mineralogical data, the method was validated. Initial results show that additional preparation yields comparable results and that samples without density separation are statistically reliable for heavy mineral analysis.

Elemental overview of a thin section (2.5 x 2 cm) typically requires large-area mapping over numerous fields, which can take several hours with conventional approaches. To shorten the measurement time without compromising data quality, we utilize the annular EDS FlatQUAD detector, capable of collecting up to 2.4 million counts per second. This speed reduces the required time-per-pixel, dwell time per frame, and overall measurement time, enabling the mapping of major elemental distributions across an entire thin section in under a minute.

In this example of a garnet-spinel peridotite from South Africa, we compare measurements of the entire thin section which took several hours to cover the 8 x 14 fields, resulting in high statistical accuracy compared with ultra-high-speed mappings revealing elemental distribution in less than one minute measurement time for the entire thin section. This example showcases the efficiency and capability of advanced EDS technology in geological studies. Extending the measurement time will result in much better statistics, and the software can detect mineral phases automatically; however, the major elements distribution is clearly visible in a short analytical run time for the entire thin section, including offline extraction of spectra from each pixel in the map for further quantification.