This thesis shows the use of advanced characterisation instrumentation to provide novel and detailed insights about the structural features of hydrated Portland cements and other cementitious systems where Portland cement has been partially substituted for an alternative, in this case blast furnace slag. Three methods were applied in this thesis to characterise the formation and chemistry of hydrates across the first hours and days after the reaction has been triggered, the properties of the pore volume, and the uptake of water into the specimen.
The thesis begins with the identification of a number of gaps in the literature and areas in to which research might extend, in Chapter \ref{chap:Literature}. For example, it was apparent in the literature that very few data are available regarding the morphology and distribution of hydrates within the material across the first hours of the hydration reaction, and that the uptake of moisture in similar materials has only been observed a single time in a semi in-situ manner.
In Chapter \ref{chap:Materials}, the blast furnace slag and Portland cement materials used were chemically and physically characterised. This involves qualitative X-ray diffraction, thermogravimetry, and particle size and shape analysis. In hydrated pastes based on blended cements, chemical analysis was conducted and the consumption of clinker phases and the formation of hydrates across the first year of the hydration reaction shown.
In Chapter \ref{chap:XRD}, in-situ X-ray diffraction was applied to the hydrating material to reveal the chemistry of the early period of the cement hydration reaction, which is very challenging to otherwise observe. The application of the partial or no known crystal structure method to observe the early age consumption of the blast furnace slag was demonstrated, which in combination with the Rietveld method provides the phase assemblage across the first days of the reaction. The consumption of anhydrous materials and the formation of hydrates was correlated to the thermal output and the formation of hydrates showed good agreement with the observed heat flow.
In Chapter \ref{chap:Tomography}, computed microtomography was applied in-situ to temporally changing samples of both pure Portland cement and blended cementitious systems. This was carried out to directly observe the morphology of the hydration reaction, and how the hydration reaction forms and defines the pore volume within the material. The formation of precipitates and deposition of hydrates was shown across the first 12 hours of the reaction. Similarly to the results presented in Chapter \ref{chap:XRD}, the data produced were correlated with the thermal output, revealing the morphological processes occurring within the microstructure and the times at which these are triggered within the reaction. In completing this, automated methods of reconstruction, segmentation, and data rendering were applied, which are shown in the supplementary information. This removes the human from the process beyond the application of original parameters, which subsequently remain constant throughout. This appears to be the first application of true in-situ microtomography to the field of cement hydration.
Microtomography was then applied for a second time in Chapter \ref{chap:Tomography2} to study the pore volume in closer detail, characterising a distribution of pore sizes by direct imaging and avoiding the assumptions made by the Washburn equation. Mercury porosimetry requires assumptions to be made regarding the pore geometry which while perhaps correct for some materials are incorrect for cements. This invalidates the pore size distribution. Again the entire process was automated, and is shown in supplementary information.
In the final chapter moisture uptake was directly observed and quantified in the specimen by neutron radiography, which reveals the accessible pore volume and some of the potential flaws in the capillary suction test. This links the formation of hydrates to the geometric properties of the pore structure, and finally to the dynamic uptake process which occurs within the structure itself. Data were captured in-situ across 24 hours by neutron imaging, which provides good greyscale contrast between permeating moisture and the specimen and allows for a straightforward segmentation. Data were calibrated to a known moisture content, which allows for the production of spatially resolved profile within a specimen, and the extraction of the total porosity assuming full saturation in non-zero array elements. It was also possible to generate a semi-quantitative distribution of moisture contents within the specimen and observe how this changes as the degree of hydration within the specimen increases.
