The nature of the calcium-silicate-hydrate phase in hydrated Portland cement has been the subject of considerable debate for decades. Various nanostructural models have been proposed including those constructed from colloidal-based particulate systems and those formulated on the basis of layered calcium-silicate-hydrates. These are examined in detail in the literature review section of the thesis. Relatively recent composition-based models have been proposed by Taylor and Richardson-Groves. These models contain structural elements comprised of 1.4 nm tobermorite and jennite. Details are also provided in the literature review. There is however a paucity of data on the engineering properties of pure calcium-silicate-hydrate phases and virtually none on the mechanical performance of 1.4 nm tobermorite and jennite. The global objective of this thesis was to examine the compatibility of the composition-based models with the engineering behaviour of the pure tobermorite and jennite phases. Pure phases of a variety of layered calcium-silicate-hydrates were synthesized and novel techniques developed to determine their engineering characteristics in a variety of test environments. The silicate phases investigated included high temperature silicates e.g. gyrolite as these layered structures are known to be cross-linked. Investigation of the role of ‘structural’ water in layered silicates was also a part of these studies. The thesis is based on a series of twelve refereed journal papers by the candidate (eight are published or accepted and four have been submitted for publication). The research is reported in four parts with each part comprised of three papers. Each part provides insight into the nanostructure of C-S-H in hydrated cement. The arguments developed evolve from an assessment of various factors including aging and the state of water in the layered silicates. The first part of the thesis focuses on the development and application of dynamic mechanical thermo-analysis methods that are sensitive to phase changes and are useful for assessing the compatibility of engineering behaviour with model composition based on 1.4 nm tobermorite and jennite. The second part represents a study of volume stability and mechanical property-porosity relationships for the pure silicate phases that are germane to these studies. The third part focuses on prolonged aging and role of structural water in cement paste hydrated for 45 years. The fourth and final part attempts to address the role of layer structure e.g. cross-linking of silicate sheets on engineering behaviour. The non-uniqueness of modulus of elasticity with respect to equilibrium moisture content is demonstrated. Structurally related irreversible effects that are dependent on drying history are rationalized. A summary chapter is provided wherein the evidence for a composition-based model with tobermorite and jennite structural units is rationalized in terms of the experimental evidence provided in this study and suggestions for future research are discussed.

The nanostructure, physical properties and mechanical performance of C-S-H, 1.4 nm tobermorite, jennite, and ettringite were studied. C-S-H of variable stoichiometries was examined as a model system in comparison with that produced in the hydration of Portland cement. The current Master’s thesis is comprised of four research papers designed to improve the current understanding of the nanostructure and engineering properties of C-S-H systems and modified C-S-H systems. Many of the controversial issues in cement science were identified and were addressed in a comprehensive research study, which examined the key features of the C-S-H systems at the nano-structure level. In Chapter 4, each paper presented new evidence for a number of mechanical aspects of C-S-H materials. Numerous advanced analytical tools were used in order to verify the observations made in each section. The major achievements of the current work are mentioned briefly as follows: 1. It was determined that microindentation is a useful method for determining the creep behavior of C-S-H of various stoichiometries, 1.4 tobermorite, jennite, and ettringite. 2. Microindentation parameters i.e. creep modulus, indentation modulus and indentation hardness are porosity dependent. 3. Microindentation creep measurements on C-S-H (C/S = 0.80 and 1.20) demonstrated that creep modulus, indentation modulus, and indentation hardness are all dependent on mass-loss from the 11%RH condition. 4. Evidence was presented that the nanostructural role of interlayer water in C-S-H has a significant influence on the creep process.

Nanotechnology refers to the use of the materials or particles ranging from a few nanometers (nm) to 100 nanometers (nm) in a wide range of applications. Use of nanomaterials in cement composite to enhance the mechanical properties, fracture toughness and other functionalities has been studied for decades. In this regard, one of the carbon-based nanomaterials, Graphene Oxide (GO), has received attentions from researchers for its superior mechanical properties (e.g. tensile strength, yield strength, and Young's modulus). Although GO is not lucrative in increasing electrical conductivity (EC) of cement paste compared to that of graphene- another derivative of GO, reduced graphene oxide (rGO), might be a solution to increase EC. Another derivative of GO is the solution to the problem. In this research, the compressive strength and flexural strength of GO-cement composite (GOCC) and rGO-cement composite (rGOCC) have been investigated with 0.01% and 0.05% GO and rGO content. GOCC-0.05% showed 27% increase in compressive strength compared to the control cement paste after 28 days (d) of hydration. GOCC-0.01% showed only 3.4% increase in compressive strength compared to the control. rGOCC-0.05% showed 21% increase in compressive strength and 15.5% increase in Modulus of Rupture (MOR) compared to the control cement paste after 28 d of hydration. On the other hand, rGOCC-0.01% showed 7% increase in compressive strength and 0.35% increase in MOR after 28 d. GOCC-0.05% showed increasing trends in compressive strength after 28 d indicating continuation of hydration. Similarly, rGOCC-0.05% also showed increasing trends in compressive and flexural strength after 28 d, possibly due to the reason described earlier. Microstructural investigation on GOCC-0.05% and GOCC-0.01% by X-ray Diffraction (XRD) illustrated that the crystallite sizes of tobermorite-Å and jennite, which are mineralogical counterpart of disordered Calcium-Silicate-Hydrate (C-S-H), increases from 3 d to 28 d, representing the crystallite growth due to continued hydration. However, the crystallite size of GOCC-0.05% was smaller than that of GOCC-0.01% at both 3 d and 28 d, indicating finer nucleated grains. According to Hall-Petch equation, mechanical strength increases with decreasing particle size. Finer particles or grains can increase the strength in cement composites in several other ways: (1) GO acted as heterogeneous nucleation sites because of reactive functional groups. Activation energy was decreased by these "defects" in the cement paste, and consequently, numerous nuclei of C-S-H. with high surface area were formed, (2) because of finer grains, cracks are forced to move along a tortuous path, which makes the structure difficult to fail, and strength increased consequently (3) Finer grains of GOCC-0.05% created compacted hydration products decreasing porosity which can indirectly increase the strength. The above reasons, separately or in conjunction, might increase the strength of GOCC-0.05% and proved that GO is responsible for increasing heterogeneous nucleation sites during cement hydration. Early age hydration (EAH) characteristics were investigated for rGOCC specimens with 0.1% and 0.5% rGO content. Scanning Electron Microscope (SEM), Energy Dispersive X-ray analysis (EDX), and X-ray Diffraction (XRD) were employed to study the EAH characteristics. SEM/EDX, and XRD analysis were performed after 15 min, 1 h, 3 h and 24 h of hydration. (EAH) study on rGOCC-0.1% showed that at 15 min hydration, numerous precipitates of, possibly, C-S-H formed along the grain boundary (GB) of unhydrated cement grains. This served as visual confirmation of Thomas and Scherer's Boundary Nucleation and Growth (BNG) model that hydration of cement grains was initiated by the short burst of nucleation of C-S-H embryos along GB. EDX on rGOCC-0.1% and rGOCC-0.5% showed that Ca/Si ratio in C-S-H was ~2.0. This finding indicated that C-S-H structure in this study was concurrent with that of impure jennite. XRD analysis also evidently showed that jennite was present, possibly possessing a short range ordered (SRO) structure, referring to local crystalline structure in a very short area. After consulting Chen's work, it would be appropriate to say that C-S-H found in this study resembled more as C-S-H (II), which is disordered jennite. It was also observed that as expected with cement with nanomaterials, with continuing hydration, pore spaces were filled with hydration products such as C-S-H, ettringite, CH, sulfoaluminates etc,. Lastly, Electrical resistivity (ER) testing on 9 sets of rGOCC specimens was conducted. The specimen includes 0.5%, 1%, 5% rGO content, and the control conditioned in both oven dry (OD) and saturated surface dry (SSD). ER increased with the increase of rGO content from 0.5% and 1% compared to that of the control. However, the ER of rGOCC-5% was significantly decreased, showing 93% reduction compared to the control, which can be interpreted as a threshold value for sensing applications to be explored. As expected, large reduction of ER value occurred on the specimens with the SSD condition. This reduction can be attributed to the ionic conduction though the pore solution of the composites. As the rGO content increased, so did the potential nucleation sites for hydration (as can be seen in SEM images), which might block the number of contact points among the rGO, resulting in low conduction and high resistivity. However, as rGO content increased to 5%, the contact areas/points increased to a degree that could trump the nucleation seeding sites, resulting in decreased ER. The ER measured with the rGOCC specimens was comparable to that of cement composites incorporating carbon fibers (CF), and steel fibers, but higher content of rGO are required to have a similar ER range of those fiber cement composites. This might be due to smaller sizes of rGO sheets and lower aspect ratio compared to other nanofibers causing drastic reduction of electron tunneling mechanism compared to other fibers.