Eigenstress Development

Finally, we consider the results for the measurement of eigenstresses developing around the embedded spherical stress sensor, as shown in Fig. 10. In this figure, compressive stresses are shown as negative values and tensile stresses as positive ones, to agree with the sign convention used in the plots of autogenous deformation (i.e., an autogenous shrinkage corresponds to a compressive stress at the sensor surface). Once again, the four different cement finenesses exhibit quite distinct responses. The finest cement (643 m ²/kg) shows a small initial compressive stress, which decays steadily to a zero stress level. It is conjectured that this is due to the very large initial autogenous shrinkage exceeding the tensile strain capacity of the young cement paste, causing local microcracking (and/or debonding) and stress relief at the sensor surface. By comparing the eigenstress and autogenous deformation responses, it is observed that this local microcracking is likely initiated at a strain level of about -560 microstrains. Interestingly, as shown by the dashed colored lines and star data points in Fig. 8, the 387 m²/kg cement shows evidence for microcracking/damage initiating at about -460 microstrains. Thus, for both of these systems, the critical strain capacity appears to be on the order of -500 microstrains or 0.05 %. Conversely, the critical stress levels are seen to be quite different for the two cement pastes, being approximately 0.8 MPa and 4.5 MPa to 5.5 MPa, for the 643 m²/kg and 387 m² /kg cements, respectively.

Figure 10: Measured eigenstress vs. time as a function of cement fineness.

The eigenstress behavior of the coarser cements is even more interesting. First, let us consider the coarsest cement, with a fineness of 212 m ²/kg. Here, from its initial state embedded in the fresh cement paste, the eigenstress sensor first registers a tensile stress (or a release from any initial compressive state) that slowly increases in magnitude, experiences compression relative to its maximum recorded tensile stress at about 200 h, and crosses its initial "zero" stress level after about 400 h, beyond which it continues to develop as a compressive stress. By comparing the stress sensor results in Fig. 10 to the autogenous deformation readings in Fig. 8, it can be clearly seen that the return to the initial "zero" stress level corresponds quite well to the autogenous deformation returning to a value of zero microstrain, as might be expected. In this case, when the deformation level returns to its initial state, the embedded spherical stress sensor finds itself also back to its original value of eigenstress.

Unfortunately, the results for the 254 m² /kg cement are not as clear. Here, the initial expansion observed in the autogenous deformation measurements is not accompanied by an equivalent tensile eigenstress development, with a gradually increasing compressive stress being observed instead. One possible explanation for this observation is that unlike autogenous (due to the continuity of stress in the capillary water phase) or thermal shrinkages/expansions, expansion due to ettringite formation is a local as opposed to a bulk phenomena. Thus, the local environment around the embedded stress sensor will determine whether the growing crystals of ettringite will push cement paste away from the sensor surface (resulting in a tensile stress relative to the initial state) or will compress paste "towards" the sensor surface, resulting in an increase in compressive stresses, as observed for the 254 m²/kg cement. More important from a practical viewpoint, despite the large compressive stress (7 MPa) ultimately developed at the stress sensor surface, there is little indication of microcracking as the strain capacity of the paste is not exceeded either in compression or tension. The autogenous deformation in Fig. 8 reaches a maximum at about 250 microstrains and has only decayed to a value near zero after approximately 700 h of exposure.