Next: Fluid permeability Up: Comparison with experimental Previous: Comparison with experimental

Electrical conductivity

There has been very little experimental work in which both the electrical properties of mortars and the underlying cement paste have been measured. It is hoped that the model presented in this paper will encourage such measurements. However, there are two papers in which measurements of the correct parameters were attempted [10,45].

The first [45] contains chloride diffusivity measurements made for cement pastes and mortars with several different sand volume fractions. The ratio of the overall diffusivity to the cement paste diffusivity is, via the Nernst-Einstein relation, the same as the ratio σ / σ_p [2]. It was found that this ratio was on the order of 1-2, for both 0.4 and 0.5 w/c mortars, at about 50% sand content. Using Fig. 7, this result implies, assuming that the sand particle size distribution was similar to that used in our random mortar model, that σ_s / σ_p was roughly between 10 and 20, in agreement with the experimental results discussed earlier on flat aggregate interfaces [11,12]. However, another paper that attempted to measure the conductivity of mortar as a function of sand volume fraction found completely different results [10].

Ping and Ming-shu [10] found a roughly linear dependence of mortar conductivity on sand volume fraction, with the slope always negative. They used an approximate formula for the electrical conductivity of the mortar, as a function of sand content, σ_s / σ_p , and h, to try to extract information about the interfacial zone conductivity. As this formula does not even agree with the exact result for the dilute limit shown in eq. (3) and Fig. 8, any information obtained by it on the interfacial zone conductivity is suspect [10]. The experimental data also appears to have problems. For quartz aggregate mortars, which almost certainly have high porosity interfacial zones, Ping and Ming-shu find mortar conductivity values that are up to four times lower than the cement paste values at sand fractions of 50%. This value seems too low, especially when compared to the 3/2 power law [27] that predicts a conductivity, (0.5)^3/2 = 0.35, only three times lower than the paste value when the interfacial zone has the same conductivity as the bulk cement paste. In these quartz sand mortars, we expect that the interfacial zone conductivity is indeed greater than the bulk cement paste value, so that the overall conductivity should be higher than the 3/2 power law result, not lower.

A possible problem with the experimental data in Ref. [10] is that a fixed frequency measurement of the conductivity was used. It has been shown that variable frequency A.C. measurements (impedance spectroscopy) must be used to properly define the D.C. response of ionic conduction systems like cement paste and mortars [5,46]. Fixed frequency measurements do not necessarily correctly eliminate the effect of the electrode-sample impedance, so that, in general, too high values of resistance, and therefore erroneously low values of conductivity, will be measured [46].