Autogenous Deformation

The measured autogenous deformation vs. time of the four cements is shown in Fig. 8. For the two finest cements, a rapid development of shrinkage is observed while the cement paste is relatively weak, followed by a more gradual development of shrinkage as the elastic modulus of the paste increases. For the coarser cements, after set, a small initial expansion is produced, reaches a maximum, and then is eventually overwhelmed by the developing autogenous shrinkage. These general observations are consistent with comparable measurements of autogenous deformation vs. cement fineness performed by Koenders [26] and by Tazawa and Miyazawa [27]. As with expansive Type K cements [28], the observed expansion is most likely due to ettringite formation [29] (as opposed to readsorption of bleed water which was eliminated as a possibility by comparison of rotated and virgin sealed RH specimens). Even though the cement is very low in C₃A content, substantial ettringite is formed from the reactions between the C₄AF phase and the hemihydrate, as measured by X-ray diffraction techniques for pastes with w/c=0.5 [30], and consistent with the observations of Fukuhara et al. [31] and of Brown [32]. Thus, in all four systems, there is a competition between expansive ettringite formation and autogenous ("drying") shrinkage, with the latter clearly dominating for the finer cements.

Figure 8: Autogenous deformation vs. time as a function of cement fineness. Stars and colored dashed lines indicate the microstrain levels corresponding to the hypothesized local cracking (stress reduction) around the embedded spherical stress sensor.

As with relative humidity, the autogenous deformation results can also be plotted against degree of hydration, as shown in Fig. 9. One interesting feature of these curves is that the initial degree of hydration required to achieve set (as indicated by the intersection of the curves with the microstrain=0 value) is significantly larger for the finer cements, as has been predicted previously based on microstructural modeling [10]. Thus, while the finer cements achieve set in less time due to their much faster hydration rates, they actually require a greater degree of hydration to form a sufficient number of bridges of hydration products between the more numerous initial cement particles, than in the case of coarser cements. After setting, the finest cement (643 m² /kg) shrinks rapidly and the 387 m² /kg cement shrinks gradually, while the two coarser cements actually expand slightly due to the ettringite formation. At higher degrees of hydration ($\alpha >$ > 0.5), the four curves are basically parallel suggesting that the ratio of the incremental capillary stresses to the long term elastic modulus are similar for the four different cement finenesses.

Figure 9: Autogenous deformation vs. model-predicted degree of hydration as a function of cement fineness.