Jeffrey W. Bullard
National Institute of Standards and Technology, Gaithersburg, Maryland, USA
Microstructure models of cement paste hydration, such as CEMHYD3D, HYMOSTRUC, or μIC, generally do a good job at generating 3-D microstructures that can be used to compute mechanical and transport properties. However, although the correlation between microstructures and properties is modeled well, the kinetics of the microstructure development itself are handled either empirically, by calibrating property development with experiments, or with the help of phenomenological equations of the Avrami category. In most other materials science fields, microstructure evolution phenomena are modeled in a more fundamental way, by evaluating the local driving forces for microstructural changes, allowing the microstructure to evolve over a small time step according to those driving forces, and then re-evaluating the driving forces for the next time step. With these models, only localized equations for the basic laws of irreversible thermodynamics are required, and the microstructure evolution proceeds by motion of the various interfaces under the influence of the local driving forces.
Anyone trying to apply these same principles to model the hydration of Ca3SiO5 paste is confronted with some frustrating problems, many of which are related to the nature of the C-S-H product. The compositional variability of C-S-H is not itself the problem, for microstructure models are capable of modeling solid solutions. A more serious challenge is the seeming ability of C-S-H to adapt its growth morphology and its apparent physical properties to the local growth conditions. Some experiments  report a standard heterogeneous nucleation and growth behavior of C-S-H, although a strong dependence of the anisotropy of the nanoscale growth rate is inferred as a function of pH. Other experiments , on nominally the same material and using very similar growth conditions, report an almost instantaneous production of large quantities of a porous fibrillar product, the rate of growth of which drops to nearly zero for several hours afterward. These independent experiments and others like them seem to provide little clue about what can cause such striking differences in growth rates and morphology of C-S-H.
To make advances in fundamental microstructure modeling of cement hydration, therefore, it is important to identify the basic chemical and physical principles, i.e. driving force and material response, that govern the growth of C-S-H over length scales from 10 nm to 10 μm. A series of questions about C-S-H growth will be posed. Very few answers will be offered, but it is hoped that the questions will stimulate discussion and provide ideas for future research.
 S. Garrault and A. Nonat, Langmuir 17 (2001) 8131-8138
 M.C.G. Juenger, P.J.M. Monteiro, E.M. Gartner, and G.P. Denbeaux, Cem. Concr. Res. 35 (2005) 19-25.