July 28, 2026 Dataset Open Access

Data publication from the Master's thesis "Mapping Metal Phase Distribution in Historical Muldenhütten Slag Using X-ray Computed Tomography" by Mahsa Abdolahzade.

Abdolahzade, Mahsa; Renno, Axel D.; Debastiani, Rafaela; Ebert, Doreen; Guy, Bradley; Moeckel, Robert

Data publication from the Master's thesis "Mapping Metal Phase Distribution in Historical Muldenhütten Slag Using X-ray Computed Tomography" by Mahsa Abdolahzade.

The data is described in the file "Data publication description relative to the Master thesis Mahsa Abdolahzade"

Abstract: 
Historical metallurgical slags are heterogeneous multiphase materials in which metal-rich constituents may occur as highly localised internal domains. This thesis developed and evaluated an X-ray computed tomography (XCT) workflow for the non-destructive three-dimensional characterisation of selected historical Muldenhütten slags. The methodological focus was the distinction and quantification of three broad attenuation-based classes: internal pores, the dominant slag matrix and high-attenuation features.

Multiscale XCT was applied to samples MS15 and MS18 from slag group SL7 and MS35 from slag group SL3. Fragment scans were used to visualise internal heterogeneity and guide region-of-interest selection, followed by higher-resolution scanning of 11 extracted cylindrical cores. Scan-specific grey-value thresholds, spatial masks and connected-component analysis were used to quantify class volumes and evaluate feature connectivity. Complementary micro-X-ray fluorescence mapping and targeted SEM–EDS and mineral liberation analysis (MLA) constrained the chemical and mineralogical interpretation of selected XCT-visible regions, while existing X-ray diffraction data provided sample-level context.

The slag matrix dominated all analysed cores, accounting for 86.38–98.02 vol.%. Internal pore fractions ranged from 0.62 to 7.57 vol.%, while high-attenuation material generally represented 0.59–2.80 vol.%. MS15-01-C1 was a clear exception, containing 12.90 vol.% high-attenuation material, of which 88.2% occurred in one connected region. Correlative µXRF and MLA results identified this feature as a localised Pb-rich metallic domain containing native Pb, Sb-bearing alloys and subordinate sulphide-bearing phases. The remaining investigated volumes were predominantly silicate dominated and contained smaller, more dispersed high-attenuation features.

The results demonstrate that conventional grey-value segmentation is suitable for quantifying broad structural and attenuation classes where sufficient contrast exists, but not for complete mineral-specific classification of complex slag. Overlapping attenuation responses, scan-dependent grey values, spatial-resolution limits, artefacts and partial-volume effects restricted phase separation. The developed workflow is therefore most effective as a non-destructive localisation and volumetric characterisation method integrated with targeted chemical and mineralogical analyses. It provides a basis for representative sample selection and future metal-deportment, recovery and environmental studies, while the results remain specific to the analysed fragments and core volumes.