S. Ehtesham Hussain’s research while affiliated with Royal Commission for Jubail and Yanbu and other places

Cement pastes with water to cement ratio of 0.60 were prepared using three cements with C3A contents of 2.43, 7.59 and 14 percent. The chloride treatment levels of 0.6 and 1.2 percent by weight of cement, derived from sodium chloride, were used in conjunction with sulfates. Sulfates derived from sodium sulfate, were added in such quantities that for each of the two 0.6 and 1.2 percent chloride-bearing cement pastes the total SO3 content of the cements were raised to 4 and 8 percent on a weight basis. The pastes were allowed to hydrate in sealed containers for 180 days and then subjected to pore solution expression. The expressed pore solutions were analyzed for chloride and hydroxyl ion concentrations. It was found that the alkalinity of the pore solution is significantly increased by the addition of sodium sulfate in the chloride-bearing hydrated cement pastes. This is attributable to the formation of sodium hydroxide as a result of reaction between sodium sulfate and calcium hydroxide liberated during cement hydration. The addition of sulfates also caused a significant increase in the chloride ion concentration in the pore solution, for both chloride levels in all the three cements tested. DTA results show that the sulfate addition reduces the formation of Friedel's salt, which possibly results in an increase in the chloride ion concentration the pore solution. The interactive effect of increase in alkalinity and chloride ion concentration with sulfate addition is not a consistent increase or decrease in the Cl−/OH− ratio of the pore solution. For a given chloride level, whether sulfate addition increases or decreases the Cl−/OH− ratio of the pore solution, and hence the corrosion risk, depends upon the interactive effect of equivalent alkali content and C3A content of the cement.

Cement pastes with a water-cement ratio of 0.6 were prepared using three ordinary portland cements with C3A contents of 2.43, 7.59 and 14%. Three levels of chlorides 0.3, 0.6 and 1.2% by weight of cement, derived from sodium chloride, were added through mix water. The pastes were allowed to cure in sealed containers at 20 and 70°C for 180 days and then subjected to pore solution extraction. The expressed pore solutions were analyzed for chloride and hydroxyl ion concentrations. Results show that increase in temperature from 20 to 70°C increased unbound chlorides and decreased hydroxyl ion concentration of pore solutions for all the three cements. The simultaneous increase in unbound chlorides and decrease in hydroxyl ion concentration drastically increased Cl−/OH− ratio of the pore solution, thereby indicating an increase in corrosion risk. This adverse effect of increase in the Cl−/OH− ratio of the pore solution with increase in temperature is higher in the high 14% C3A cement than in the low C3A cements, and is also higher for the low 0.3% chloride treatment level than the higher chloride inductions. Increase in temperature is also expected to cause an increase in ionic diffusion to steel embedded in concrete as well as in the rate of corrosion reaction. All these factors tend to increase corrosion risk of steel reinforcement in concrete with an increase in temperature.

Accelerated corrosion tests were carried out on reinforced-concrete specimens made with plain and 10% and 20% microsilica-blended cements. Time to corrosion initiation and corrosion rates of reinforcing steel were measured. Pore fluid was also extracted for analysis from chloride-bearing hardened cement pastes. Results show that in terms of corrosion-initiation time, microsilica-blended cement concretes perform threefold better than the plain cement concrete. After the initiation of corrosion, reinforcing steel in microsilica-blended cement concrete shows no significant reduction in corrosion rates. Partial cement replacement by microsilica, sharply increases the Cl- /OH- ratio of the pore solution. In addition to the adverse changes in the chemical environment, microsilica blending also brings about significant improvements in the pore structure of the cementitious matrix. The higher corrosion-initiation time of the microsilica-blended cement concretes is ascribable to observed pore refinement and segmentation, which retards the diffusion of chloride ions to the steel-concrete interface, thereby delaying the onset of corrosion. Data developed in this study show microsilica-blended cement concrete will provide significantly more protection to steel when secondary chlorides permeate into concrete from external sources during service life, compared to situations where primary chlorides are introduced during mixing.

Cement pastes with water-cement ratio of 0.60 were prepared using four cements with C,A contents of 2.04, 7.59, 8.52, and 14 percent. Four levels of chlorides corresponding to 0.3, 0.6, 1.2, and 2.4 percent by weight of cement were added to the mix water. The pastes were allowed to hydrate in sealed containers for 180 days and then subjected to pore solution expression. The expressed pore fluids were analyzed for chloride and hydroxyl ion concentrations. It was found that the free chloride concentration in the pore solution decreases significantly with an increase in the C,A content of the cement. Typically for a 0.6 percent chloride addition, the unbound chlorides decreased from 41 to 12 percent when the C,A content of the cement was increased from 2 to 14 percent. The high C,A content was found to be especially beneficial for binding chlorides in the range of 0.3 to 0.6 percent. With increasing level of chloride addition, although the absolute amount of bound chloride increases, the ratio of bound to total chlorides decreases. For example, in the 14 percent C,A cement, the ratio of bound to unbound chloride is about 14 times higher for the 0.3 percent chloride addition compared to 2.4 percent chloride addition. For a threshold Cl-/OH- ratio of 0.30, the threshold chloride values for the 2.04, 7.59, 8.52, and 14 percent C,A cements were found to be 0.42, 0.62, 0.68, and 1.0 percent by weight of cement. The effect of the C,A content in significantly influencing corrosion is also confirmed by the corrosion initiation times, which were found to be 1.75, 1.93, and 2.45-fold more for the 9, 11, and 14 percent C,A cements compared to 2 percent C,A cement. The pore fluid analysis indicates some chloride binding even in the low 2.04 percent C,A cement when chlorides are added at the time of mixing.

Pore fluid analysis and 6-month expansion tests were carried out on plain, 10 and 20% microsilica, and 60 and 70% blast furnace slag cements. The objective was to study the correlation between expansions and hydroxide concentrations in the pore solution, alkali-silica reaction (ASR) behaviour of microsilica- and slag-blended cements, optimisation of microsilica addition, and alkali-removing capacity of blast furnace slag. Incorporation of 10–20% microsilica and 60–70% slag reduced expansions from nine times the permissible expansion to safe values ranging from one-tenth to one-half the allowable expansion. The data developed in this study confirm a broad correlation between the hydroxide ion concentration and ASR-generated expansions. However, there are obvious deviations which are strongly indicative of other concurrently occurring mechanisms in addition to the alkali removal action. Blast furnace slag is shown to be an active remover of alkalis and is especially effective in medium-alkali cements, where for equal alkali contents, the performance of 60% slag cement is comparable with that of 10% microsilica cement. However, the effectiveness of the slag decreases as alkali content of the cement increases. Addition of 20% microsilica sweeps virtually all the hydroxide ions from the pore solution. Ten per cent microsilica significantly reduces OH− concentration, although in 1·2% and 1·5% alkali cements the hydroxide ion concentrations remain above the postulated threshold of 250 mM/l. However, the expansions for 10% microsilica cement with OH− concentration of 368 mM/l is well below the allowable limit. This suggests that, for blended cements, the 250 mM/l value is excessively conservative and a value close to 400 mM/l would be more realistic. This is possibly a result of the fact that the incorporation of reactive silica in the form of microsilica or slag, in addition to chemical factors, brings about a marked refinement and segmentation of the pore structure, thereby effectively retarding the transportation of alkalis to reaction sites. From both the standpoint of expansion and the 400 mM/l threshold value, it seems that for most cases of medium- to high-alkali cements, 10% cement replacement by microsilica is adequate for ASR control.

Plain and microsilica blended cement pastes with water-cement ratio of 0.6 were prepared using a 14% C3A cement. Two levels of chloride from NaCl corresponding to 0.6% and 1.2% by weight of cement were added through mix water. The pastes were allowed to hydrate in sealed containers for 180 days and then subjected to pore solution expression. The expressed pore fluids were analyzed for chloride and hydroxyl ion concentrations. The results show that the OH− ion concentration in the pore solutions of both chloride-free and chloride-bearing pastes drop steeply with increasing cement replacement by microsilica. For 10% microsilica cement pastes the pH for both 0.6% and 1.2% chloride addition was found to be around 13.30. However, the pH drops to a level below that of saturated Ca(OH)2 solution when cement replacement by microsilica is increased from 10% to 20%. This is ascribable to the consumption of Ca(OH)2 by microsilica as shown by the DTA/TGA results. 10% and 20% microsilica blending more than doubles the free chloride ion concentration in the pore solutions of the chloride-bearing pastes. 10% microsilica replacement raises the Cl−/OH− ratio 4 to 5 fold, whereas for 20% microsilica replacement, the Cl−/OH− ratio is increased to 77 and 39 folds over the corresponding values for the plain cement pastes for 0.6% and 1.2% chloride additions respectively. Accelerated corrosion monitoring tests carried out on steel bars embedded in plain and microsilica blended cement concretes exposed to 5% NaCl solution show a 3 fold superior performance of microsilica blended cement concretes in terms of corrosion initiation time. This corrosion behaviour is contrary to the prediction from the increased aggressivity of pore solution composition in terms of highly elevated Cl−/OH− ratios. This is attributable to the densification of cement matrix by the pozzolanic reaction between microsilica and calcium hydroxide. No discernable advantage in terms of corrosion initiation time is evident by increasing microsilica blending from 10% to 20%.

Pore solution study has been carried out on 2.43 and 14% C3A hardened cement pastes. Data have been analyzed in conjunction with the data developed in two pore solution studies made by Page and Vennesland and Diamond using 7.37 and 9.1% C3A mature cement pastes. The results show that C3A and alkali contents of a cement have significant effect on its chloride-binding capacity. For similar alkali content, the levels of free chlorides in the pore solutions of 2.43 and 9.1% C3A cement pastes are respectively 4.7 and 2.8 times more than in a 14% C3A cement. The alkali content of a cement appears to have an inhibiting effect on its chloride-binding capacity. However, this effect is overshadowed by a conjoint strong elevation of the OH− ion concentration in the pore solution due to cement alkalies, causing a net lowering of the Cl−/OH− ratio which roughly ascertains corrosion risk. Threshold chloride values have been evaluated for different C3A cements. The threshold chloride content for a typical Type I portland cement with C3A upto 8% and Na2O equivalent upto 0.60%, may be taken as 0.4% chlorides by weight of cement. However, for a similar alkali cement with a high C3A content of about 14%, the chloride threshold value is 2.5 times higher and may be taken as 1.0% by weight of cement.