3D mineralogy based on microCT, XRF elemental mapping and XRD
Rail Kadyrov
Unlike classical methods for determining the mineral composition of rocks, 3D-mineralogy, obtained based on X-ray computed tomography, allows to study the spatial arrangement of mineral phases in associations and the geometric features of the structure of the rock without violating its integrity. Based on CT results, it becomes possible to segment phases significantly different in X-ray attenuation coefficients corresponding to different minerals. However, since the X-ray radiation generated by the x-ray tube is not monochromatic, the absorption coefficients of various minerals will overlap, which excludes the unitary application of this approach.
The problem of 3D mineralogy for tube sources can be solved by using additional methods for determining the mineralogy on the surface of the sample, one of which is QEMSCAN - integrated automated mineralogy and petrography method that is based on the use of scanning electron microscopy (SEM) with energy-dispersive X-ray spectroscopy (EMF) detectors and software that controls automatic data collection and processing. However, the cost of such a solution can be very high.
We suggest an alternative approach, which based on prevalent laboratory methods as XRF elemental mapping of the surface of the sample and the subsequent use of XRD to identify the mineral phases and conversion the resulting elemental maps into mineralogical. Next, mineralogical maps of the surface are combined with the spatially linked voxels with certain grayscale values, based on which the voxels are classified for the entire volume. Figure 1 shows three-dimensional mineralogy for a fragment of the Chelyabinsk meteorite, made according to the described technique. As can be seen from the figure, despite the use of additional methods, minerals with similar composition are still difficult to segment, since their grayscales overlap. The way out of this situation can be tomographic imaging of the sample at 2 different energies, which should improve the classification capabilities.
This work was funded by the subsidy allocated to Kazan Federal University for the state assignment in the sphere of scientific activities (Project No. 0671-2020-0048).
