| The microstructure of cement-based materials determines most of their macroscale properties, such as strength, permeability, drying shrinkage, and creep. As new types of cementitious materials are developed, it becomes increasingly important to develop predictive models that connect the microstructure of concrete to its macroscale behavior. Calcium silicate hydrate (C--S--H), the primary hydration product of cement, is responsible for most of the cohesive properties and durability of hardened concrete. This thesis demonstrates that by modeling the complex and variable nanostructure of C--S--H, it becomes possible to predict macroscale properties of cement and concrete.C--S--H can be viewed from the perspective of both colloidal science and granular mechanics. This thesis examines first the effect of humidity and drying rate on C--S--H morphology using environmental scanning electron microscopy (ESEM). Micrographs indicate that the fibrillar structure commonly seen in an electron microscope is suppressed by drying cement paste very slowly. This characteristic, combined with experimental results from the literature, suggests that C--S--H is a discrete or particulate material. Using this information, a numerical model of C--S--H is introduced as an assemblage of discrete particles. Created using a self-designed autocatalytic growth algorithm, the model includes nanoscale material properties such as elastic modulus, particle friction, and cohesive surface charge. As nanoindentation simulations show, macroscale properties of hardened cement paste, notably its elastic indentation modulus and hardness, are quantitatively predicted by this model. |
