As mentioned in the introduction, in the present chapter the writer has deliberately confined himself to findings based on experimental evidence, rather than on modeling. The bases of at least some of the current models seem to him overly unrealistic in terms of their representations of the features found in concrete. However, he sees as a future trend a progressive development of more realistic models, based on keener appreciations of actual microstructures in concrete, and of how the microstructural features actually control the transport processes involved in permeating water, ions, etc. through them. Anticipated further increases in computer power and speed would presumably be helpful in facilitating such developments. Another future trend that may be realized is the ability to visualize the actual, three-dimensional structure of cement paste ± including pores ± in concretes, by computed tomographic methods at a resolution fine enough to be definitive. As was indicated earlier, computed tomography renderings can currently be pro- duced at voxel sizes close to 1 um and at dynamic ranges adequate to distinguish cement grains and at least the coarser pores. Potential improvements in resolution (and dynamic range) for computed tomography could provide three dimensional information with much the same level of detail as is provided by current backscatter SEM. Meanwhile, improvements in SEM instrumentation should certainly provide the facility for more detailed two-dimensional study of pores and other microstructural features. In particular, it is hoped that some combination of computed tomography, higher resolution SEM capability, and realistic modeling will clarify details that govern permeation in actual concretes.
