UC Berkeley

Portland cement (PC) based concrete is the second most consumed commodity after water, and is the foundation of modern construction industry. The present status of the cement and concrete industry is unsustainable: the production of Portland cement (PC) accounts for ~8% of global anthropogenic CO2 emissions and the deterioration of existing reinforced concrete infrastructure is a constant, environmentally and economically costly problem. Optimizing the existing technology requires a breakthrough in the understanding of the cement hydration mechanisms, and in the structure-property correlation of the hydration products, especially at nano and atomistic scale. However despite decades of study, mechanisms underlying the cement hydration and all-stage performance remain unclear to a large extent, mainly due to the lack of experimental data at the fundamental level.

The aim of this dissertation is to provide new evidence to clarify unsolved questions along the service life timeline of PC concrete, using novel experimental methods that are based on synchrotron radiation. Three specific topics at distinct stage of concrete service life are investigated: 1) at early age the hydration mechanism of tricalcium aluminate (C3A) in the presence of gypsum. C3A is the most reactive clinker phase of PC whose rapid hydration greatly controls the workability of fresh concrete mixes, yet more evidence is needed to understand how C3A hydration is retarded by gypsum; 2) at mature age, the composition-structure-property correlation of calcium (alumino)silicate hydrates (C-(A-)S-H). C-(A-)S-H is the dominant binding phase of PC concrete, and therefore determines the mechanical property and volumetric stability of modern concrete infrastructure. It is hierarchically porous down to single-nanometer, making it rather difficult to experimentally probe the intrinsic mechanical properties of pore-free C-(A-)S-H; 3) at late age, the microstructure and chemistry of hydrated tricalcium silicate (C3S). C3S is the major clinker phase in PC and is the major source for the creation of C-S-H. Its chemistry and structure have been extensively studied for the first few months of hydration, but the long-term (beyond 10 years) characterization is rarely done.

The research reported in the present dissertation uses synchrotron-radiation-based characterizing techniques to provide multi-scale evidence to answer the above mentioned questions. In-situ Wide and Small Angle X-ray Scattering, X-ray spectromicroscopy coupled with the single-nanometer resolved X-ray Ptychography imaging tool, and nano-scale 3D Tomography are used, to investigate the dissolution-precipitation front of partially hydrated C3A particles. The crystal-chemistry and 3D-morphology of the hydration products are quantified. Ettringite is found to be the only stable phase during the induction period. Regardless of the water to cement ratio, ettringite needles grow to a maximum length of 1-2 μm. Ettringite crystallization is not homogeneous on the C3A surface, and the crystallization rate depends on the amount of relative surface site. Surface of pores/cavities inside the anhydrous C3A particles contributes significantly to the hydration reaction, as soon as the dissolution connects them with the solution.

The knowledge of both the properties of the (pore-free) material at fundamental scale, and the multi-scale pore structure, are critical to the theoretical prediction of the macro-scale properties of concrete. Here, high pressure X-ray diffraction is utilized to study the anisotropic nanomechanical property of C-(A-)S-H, by tracking the response of the nanocrystal lattice to hydrostatic pressure up to ~10 GPa. For the first time, the nanomechanical properties of pore-free C-(A-)S-H are investigated as a function of chemistry and atomic configuration. The experimental results show that, contrary to the predictions of most molecular simulations, the stiffness of C-S-H increases along with its bulk Ca to Si molar ratio (Ca/Si). Under hydrostatic pressure, the c-axis (perpendicular to the layer structure) deforms much more readily, compared with the a- and b-axis (i.e. the in-plane directions of the layer structure). The densification of the interlayer spacing significantly decreases the incompressibility along the c-axis, and thus increases the overall bulk modulus. When Al is incorporated, the Al-induced crosslinking of adjacent dreierketten chains increases the stiffness of C-(A-)S-H. Molecular simulations in this dissertation unveil the atomistic mechanism for the C-S-H deformation, that the in-plane deformation of the silicate tetrahedra chain is through rigid rotation instead of the shortening of the stiff Si-O bond. Therefore neither the vacancies nor the Al incorporation in the silicate chain, alters the in-plane stiffness of C-(A-)S-H layer structure.

Lastly, Scanning Transmission X-ray Microscopy is integrated with conventional methods (lab XRD, SEM, TEM and NMR) to study the composition and structure of a late age (50 years) hydrated C3S paste. After 50 years of curing, the mean chain length (MCL) of the C-S-H is 4.18, and the XRD reveals no resolved diffraction peaks for C-S-H. This indicates that the crystallization of C-S-H under ambient condition is negligible during concrete service life. Although the average Ca/Si of the C-S-H is ~1.7, much higher than that of tobermorite (0.67-0.8) but closer to that of jennite (1.5), the coordination environment of Ca in C-S-H is similar to that of tobermorite rather than jennite.

The results in this dissertation enhance the understanding of cement chemistry, as well as provide direct validation to existing cement hydration modelling and molecular level C-(A-)S-H modelling. The work also develops protocols for systematically studying modern construction material using innovative synchrotron-radiation-based methods.