As in all fields of science, materials science and civil engineering have benefited greatly from the evolution of computer hardware and software in the past few decades. Thus, it is no surprise that a new paradigm, computational materials science [1,2], has been applied to the complex phenomena of cement hydration and microstructure development. Cement hydration is inherently complex [3]: a reaction system consisting of multi-size, multi-phase particles that evolves from a viscous suspension to a rigid load-bearing solid. However, this microstructure development needs to be better understood, because of its large influence on the transport and mechanical properties of these materials. Such an understanding could provide a major contribution to an integrated knowledge system for predicting concrete performance, such as the one currently being developed at the National Institute of Standards and Technology (NIST) under the auspices of the Partnership for High-Performance Concrete Technology program [4].
While the hydrated microstructure is easily observed in two dimensions using scanning electron microscopy (SEM), direct observation of the three-dimensional microstructure is extremely difficult. Since the percolation and other stereological properties of the three-dimensional structure greatly influence properties [5,6], computer simulation of the three-dimensional microstructural development can provide critical knowledge on the relationships between microstructure and properties and lead to the design of improved materials. At NIST, a ten year research program has resulted in the development of a comprehensive computer model for the microstructural development of portland cement based systems [7]. The model is readily available to researchers for their use and further enhancement [8]. This paper briefly describes the computational tools employed in the model and presents a summary of experimental and computer model results obtained to date.