In sealed laboratory cement pastes (or in concretes protected against either leaching or drying effects), the exhaustion of sulfate from the pore solution does not yet mark the maximum level of alkali hydroxide concentration. Instead, as hydration proceeds, the limited content of solvent water is progressively depleted, and the concentration of alkali hydroxide in the remaining volume of pore solution progressively increases. The eventual concentration of the hydroxide ions found in the pore solution in sealed pastes and mortars of a given w:c ratio is closely related to the alkali content of the cement used. Some years ago the writer5 collected and published a set of analyses from various literature sources for solutions expressed from w:c 0.50 pastes or mortars. All of these were hydrated at room temperature for the traditional 28-day period, and the chemical compositions of all of the cement used were reported in the original publications. The reported OHÿ ion concentrations were plotted against the alkali contents of the cements used, and a quite good linear relationship was found, especially considering the disparity of the data sources. It was found that the 28-day OHÿ ion concentration (expressed in mol/l) was in each case about 0.7 times the % Na2O equivalent of the cement. Thus in archetypical concretes, particularly high alkali cements can be expected to produce alkali hydroxide concentrations approaching or even exceeding 1 mol/l, i.e. pH values of the order of 14. It was calculated that globally, about 80% of the alkali present in the various cements was found in the pore solutions of sealed specimens at 28 days.5 Presumably some portion of alkali remained fixed in clinker minerals that had not yet hydrated, and some of the alkali hydroxide that had been in solution had been adsorbed by the solid components of the cement paste, primarily by C-S-H. Much more detailed treatments of these phenomena have since been pre- sented by several authors. The development of alkali concentrations in pore solutions of hydrating cements has been modeled by Brouwers and van Eijk6 in terms of calculated rates of release of alkalis from cements and calculated binding coefficients of the ions into hydration products. Rothstein et al.7 have treated the development of pore solutions in terms of saturation indexes, that is, the degree of undersaturation or supersaturation with respect to the solids calculated to be present at each stage based on hydration equilibrium equations derived originally by Taylor.8 Concentrated alkali hydroxide pore solutions have high electrical conductivi- ties. As will be discussed later, electrical methods of assessing permeance in cement paste are influenced by these pore solution conductivity values. Snyder et al.9 proposed a method of calculating the electrical conductivity of pore solutions from the specific potassium and sodium hydroxide concentrations. This is useful, since sufficient pore solution can generally be expressed from sealed samples for chemical analysis, but not necessarily sufficient for the measurement of electrical conductivity. As mentioned previously, many modern concretes contain various added solid components besides the Portland cement used. The effects are generally to reduce the alkali hydroxide concentrations produced, but not always. Most low-calcium fly ashes tend to reduce the pore solution alkali hydroxide concentrations10,11; however, the reverse effect is often found for very high calcium fly ashes12,13 or fly ashes that carry significant contents of available alkali.13 A summary of these effects was provided by Shehata et al.13 The influence of silica fume on alkali hydroxide concentration of pore solutions also appears to be complicated. Many years ago the present writer14 found that incorporation of a silica fume resulted in slightly increased sodium hydroxide concentrations at very early ages, but that after some hours, further reaction reduced alkali hydroxide concentrations to much lower levels. Similar reduced alkali hydroxide levels have been observed by many authors, for example Kawamura et al.15 However the effect may not necessarily be per- manent. Both Duchesne and BeÂrubeÂ16 and Shehata and Thomas17 found that after some weeks of sealed storage the alkali hydroxide concentrations started to increase again, and the secondary increases observed were substantial. This secondary increase in alkali hydroxide concentration may have unexpected consequences with respect to long-term durability. The effect was not found when both silica fume and fly ash were simultaneously incorporated.17 Incorporation of slag reduces the alkali hydroxide content significantly,17 an especially important effect since the proportion of slag usually added is substantial. Incorporation of ground limestone probably has little effect except that due to dilution of the cement. Somewhat surprisingly, the incorporation of certain alkali-metal bearing admixtures such as sodium-neutralized naphthalene sulfonate may add signifi- cantly to the alkali hydroxide content of the pore solution.18 The alkali hydroxide concentrations and the changes in them referred to above reflect sealed storage exposures, i.e. environments where neither water nor dissolved substances are interchanged with the external environment. In real-world exposures, concrete pore solutions can undergo significant changes in water content, in contents of dissolved components, or in both. Concretes may be exposed to leaching, either by virtue of being underwater or through long-continued rainfall. The pores in such concretes will become water solution-saturated if not previously so saturated, and prolonged wet exposures can induce significant loss of the alkali hydroxide leached from the pore solution. A similar leaching effect is often noticed in casual handling of small specimens exposed to laboratory fog-room conditions; an unpleasant `soapy’ feel is induced by contact of the skin with the leached alkali hydroxide. Unexpectedly, it has been found that prolonged partial drying to moderately low RH values, especially when accompanied by some carbonation, `fixes’ much of the pore solution alkali hydroxide in the hydrated cement, significantly reducing the effective concentration found in expressed pore solutions.19,20 Once fixed by such exposure, it is extremely difficult for subsequent re-saturation to redissolve the alkali hydroxide and bring the pore solution concentration back to its original value. This phenomenon may at least partly explain the highly beneficial effects of drying in mitigating the effects of ongoing ASR. Some field concretes are exposed to salt (NaCl), either by deliberate applications of de-icing salts or by contact with sea water or salt spray. Some of the sodium chloride in such salt solutions may penetrate into pore solutions of the outer layers of the concrete, and appear as additional sodium hydroxide in them.21,22 Conversion of the sodium chloride to sodium hydroxide is usually a consequence of the binding of the chloride ions as Friedel’s salt. While the increased hydroxide ion concentration is favorable toward maintaining steel passivation, any significant entry of unbound chloride ions has the opposite effect. Pitting corrosion, as described for example in Ref. 23 may readily result.
Less widely appreciated, but equally important, is the influence of dissolved chloride in augmenting the harmful effects of ASR.22,24,25