A method is presented to examine consecutive in-situ X-ray diffraction (XRD) diffractograms using exploratory factor analysis. Systematic changes in the diffractograms are described numerically by score values that could be used to correlate diffraction data with other non-stationary sample properties. Phase and structure evolution in a reacting material can be studied by in-situ XRD. The consecutively collected data can be considered a time series of datasets. Time series are non-stationary data. Such non-stationary data are often hard to examine fully by conventional evaluation methods including applications of the Rietveld method. Here a method is presented to avoid shortcomings of conventional evaluation methods. The new method helps to identify and describe significant systematic changes in in-situ XRD datasets by numerical values. These systematic changes can represent structural changes as well as changes in phase composition. The method can be used to describe the development of the complex processes of compositional and structural changes. The method is demonstrated using the example of a hydrating Portland cement mortar. This hydration process involves at least 11 phases including non-crystalline phases. In the presented example factor analysis of in-situ XRD data results in three variables (factors) describing the observed changes numerically.

A revised structure model of ettringite is presented, in order to provide quantitative X-ray diffraction (QXRD) of this mineral in cement pastes. The model is derived from two different existing structure models, both of which are suitable for restricted use but are inferior to the refined ettringite structure presented. In the first published ettringite structure proposed by Moore and Taylor [Acta Crystallogr. B 26, 386–393 (1970)], none of the 128 positions for H are given in the unit cell, which results in reduced scattering power for use of this model for quantification purposes. For the precise quantification of ettringite in samples together with anhydrous phases, the scattering factors of all atoms including the H positions are indispensable. The revised structure model is based on the data of Moore and Taylor, supplemented by the H positions determined by Berliner (Material Science of Concrete Special Volume, The Sydney Diamond Symposium, American Ceramic, Society, 1998, pp. 127–141) on the basis of a neutron diffraction structural investigation of deuterated ettringite at 20 K. Berliner’s (Material Science of Concrete Special Volume, The Sydney Diamond Symposium, American Ceramic Society, 1998, pp. 127–141) thermal parameter should not, however, be used, since a normal application is at room temperature. In order further to improve the structure model of ettringite, Rietveld refinement with the rigid body approach for OH and H2O molecules and SO4 tetrahedra was employed. The refined and improved ettringite structure model was tested for quantitative phase analysis by the determination of actual ettringite contents in mixtures with an internal standard. Synthesized and orientation-free prepared ettringite powders were investigated by X-ray powder diffraction analysis and quantified in four different blends with zircon. The quantification results with the new structure model demonstrate the superior quality of the revised ettringite structure.

The two-dimensional XRD (GADDS) was used to characterize the microstructure of an applied self-leveling compound (SLC). A calcium aluminate cement based SLC was prepared on two different substrates (water absorbent and nonwater absorbent) to determine the vertical distribution of the crystalline phases. The application of the GADDS enables the detection of the phase composition of the hydrating mortar in horizontal slices. Thus the analysis could be carried out in position-sensitive mode at three different areas: near the bottom, in the center, and at the top of the mortar. For investigation of SLCs from the very early hydration stage up to 10 h of hydration, a custom-made in situ sample holder for the measurements was designed and constructed. The combination of the GADDS and the custom-made in situ sample holder provides the possibility to characterize additionally the time-dependent phase composition within the SLC. The nonabsorbent substrate has no effect on the hydration of the binder phases but the absorbent substrate influences the formation of ettringite. In the top layer of the SLC the ettringite content is reduced during the first hours of hydration. The absorbing forces of the substrate lead to migration of the mix water to the substrate. This lack of water results in the reduced formation of ettringite.