Next: Effect of freezing
Up: Dielectric amplification mechanism:
Previous: Dielectric amplification mechanism:
As shown by Fig. 4, log (σ) decreases linearly with temperature down to about −5 ºC, drops sharply about 2 orders of magnitude, then decreases linearly again. Some samples were cooled to temperatures below −100 ºC to determine if any other sharp drops in the conductivity existed, but the log (σ) value was linear throughout this region. This is displayed in Fig. 6, which represents a paste of 0.4 w/c ratio, as in Fig. 4, but for a slightly longer hydration time of 18 hours. The large conductivity drop (Δlog σ, defined in Fig. 4) occurs within the same temperature range as the freezing point of the large capillary pores, as determined by DSC [19,20,21,22,23,24,25], so it is reasonable to conclude that the drop is a result of the freezing of bulk and capillary water. When this water freezes, the mobility of the ions that carry the electrical current (mainly Na+, K+, and OH− [1]) is sharply reduced, and thus the conductivity of this phase is sharply reduced as well. As the smaller gel pores freeze at lower temperatures, the conductivity continues to decrease continuously. This large drop at about −8 ºC is probably due to the freezing of larger capillary pores that are percolated or connected at short hydration times. This point is made clearer in Fig. 7 below.
Figure 6: Logarithm of the D.C. conductivity vs. temperature for an 0.4 w/c ratio cement paste hydrated 18 hours, and taken to lower temperatures.
Figure 7: Logarithm of the D.C. conductivity vs. temperature for an 0.4 w/c ratio cement paste hydrated 15 hours, 26 hours, 8 days, and 40 days.
As hydration continues and the capillary water is consumed, more C-S-H is produced, and Δ log(σ) decreases at a given temperature. Figure 7 is a plot of the log(σ) vs. temperature for four pastes of w/c ratio 0.4 with hydration times of 15 hours, 26 hours, 200 hours, and 40 days. Note how Δ log(σ) decreases with increasing time, until, at about 40 days, there is essentially no abrupt large drop in the log(σ) vs. T curve. After 40 days of hydration an 0.4 w/c ratio cement paste has very little capillary porosity remaining, and this porosity is almost certainly discontinuous. When the larger capillary pores are well-percolated and have a relatively large volume fraction, they will dominate the overall value of the conductivity. Switching off their conductivity by freezing causes a steep drop in the conductivity of the paste. As more and more of the capillary pores become discontinuous with the hydration of the paste, less and less of a steep drop in conductivity is seen upon freezing, since the overall conductivity is not dominated anymore by the percolated capillary pores. After the percolation threshold for the capillary pores is reached, there is still a decrease in conductivity upon freezing, but no steep discontinuous drop, as the capillary pores, being disconnected, no longer control the conductivity of the cement paste.
The percolation threshold for the larger capillary pores has been predicted to be 18%, independent of w/c ratio [26]. For a paste of w/c ratio 0.4, the critical degree of hydration needed to achieve this porosity is approximately 0.65, where degree of hydration is the volume fraction of cement that has reacted. This would correspond to a hydration time somewhere between 8 and 40 days, so the above predicted behavior for Δ log(σ) is in agreement with the data shown in Fig. 7.