Abstract: Scanning electron microscopy coupled with microanalysis (SEM-EDX) is an important analytical tool for the morphological and chemical characterization of different types of materials. In many applications, SEM-EDX elemental maps are usually used and processed as images, thus flattening and reducing the spectroscopic information contained in EDX hyperspectral data cubes. The exploitation of the full hyperspectral dataset could be indeed very useful for the study of complex matrices like soil. In order to maximize the information attainable by SEM-EDX data cubes analysis, the software package “Datamuncher Gamma” was implemented and applied to study soil aggregates. By using this approach, different phases (silicates, aluminosilicates, Ca-carbonates, Ca-phosphates, organic matter, iron oxides) inside soil aggregates were successfully identified and segmented. The advantages of this method over the common ROI imaging approach are presented. Finally, this method was used to compare different aggregates in a Cr-polluted soil and understand their possible pedological history. The present method can be used for the analysis of every type of SEM-EDX data cubes, allowing its application to different types of samples and fields of study.