The usual pattern of cement hydration involves rapid reactions of cement components exposed on the fractured surfaces of the individual ground clinker grains. It is commonly accepted that these fast reactions generate surface hydration products (presumably mostly ettringite or ettringite-like), and that these products serve to temporarily isolate the underlying cement grains from effective contact with the mix solution, inducing the so-called dormant period of restricted hydration. Some hours later the barrier to further hydration is broached, and rapid autocatalytic hydration follows. Other explanations of the dormant period are sometimes advanced. In any event, a few hours after rapid hydration recommences, sufficient hydration product is generated to induce setting, marking a conversion from fresh concrete (a dense suspension) to newly-hardened concrete (a porous viscoelastic solid). Except for air voids, the pores in the newly-hardened concrete are generally full or nearly full of solution, except in the case of low w:c concretes that undergo self-desiccation. Some insight into the spatial arrangement of cement grains in fresh concrete can be derived from Fig. 2.2, kindly supplied by K.O. Kjellsen.27 Figure 2.2 is a backscatter-mode SEM produced from a thin area of freshly-mixed w:c 0.40 fly ash-bearing mortar which was quick-frozen in liquid nitrogen shortly after being mixed. The quick-frozen specimen was sublimed to remove the frozen water, and then impregnated with epoxy resin to fill the space left by the water. The epoxy resin-stabilized preparation was then carefully polished and carbon coated for examination in backscatter SEM.
The smooth gray areas in Fig. 2.2 are sand grains. The spaces between them contain bright white unhydrated cement grains, some fly ash particles of varying gray levels, and black epoxy resin occupying the spaces originally filled with water. To the extent that the pre-existing grain assemblage was not perturbed, the black areas constitute the immediate `ancestors’ of most of the larger pores that will be present in the hardened mortar once it has set. As will be discussed later, the word `most’ is used deliberately. Unknown to Powers and Brownyard,26 a substantial proportion of the larger pores in many concretes arise from space originally within certain cement grains, rather than from the originally water-filled space between them.
Cement is usually ground to particle sizes ranging from ca. 80 um down to ca. 1 or 2 um The particle sizes of the white cement grains seen in Fig. 2.2 are consistent with this expected range. The thickness of the black spaces separating them ± which might be considered in the Powers-Brownyard26 treatment as `capillary pores to be’ ± range downward from ca. 20±30 um.
Within concrete, once setting fixes the cement particles in place, the newly created and highly interconnected pores are also fixed in place. Subsequent hydration and deposition of cement hydration products progressively reduce the sizes and change the connections between the larger pores. These processes turn out to be complex geometrically as well as chemically and, contrary to the traditional assumption, they do not result in straightforward subdivision of the pore spaces.