We have used non-destructive 3D imaging via X-ray Microscopy (XRM) to obtain data of various cements and concretes to understand the spatial distribution of internal components (chemical compounds, pores, cracks, pathways for corrosive agents). We have then applied subsequent analyses using Mineralogic 3D automated quantitative mineralogy software, combined with a machine learning based reconstruction (DeepRecon Pro), to specifically identify, quantify, and understand the ‘mineralogical’ phases in 3D.