William G. Hime, Wiss, Janney, Elstner Associates, Inc
Alkali-silica reaction (ASR) was discovered by Stanton* in 1940. Since then, and certainly even now, it has been mischaracterized, misdiagnosed, and probably mistested. At the same time, ways to mitigate its effect have been developed.
Contrary to much of the literature, ASR is not a reaction of sodium or potassium (or other alkali metal ions) with a form of silica. Rather, it is the reaction of the hydroxides of those ions (ammonium ion is the exception) with, almost always, microcrystalline silicon dioxide. Only the hydroxides of these ions are soluble enough to produce the pH levels of 13, or more, that are needed to cause the reaction. The silicate that is produced occupies more space than the silica did, causing “map” cracking in the concrete.
The reaction stops when either the hydroxyl ion is sufficiently depleted (by reaction or carbonation to drop the pH below 13), or when the reactive silica particles have been consumed. Completion of the reaction occurs in hours, weeks, or years, depending upon the thickness of the concrete.
It is obvious that the higher the cement content, the more the alkalies in the concrete. It is less obvious that the higher the alkali metal content, the higher the hydroxyl ion content. But when water is added to portland cement, the alkali metal compounds largely produce alkali metal hydroxides. This is usually not true with mineral admixtures or aggregates. The alkali metals in them do not produce hydroxides.
Test Methods
To some extent, there is an easy method of analysis for reactive silica in aggregate: petrographic microscopy. However, even an excellent petrographer may not be able to predict whether or not some forms of silica will be deleteriously reactive, or if they are of sufficient quantity and reactivity to be of concern. Therefore, several tests have been developed to permit better predictions. Unfortunately, these tests may not correctly predict the duration or extent of the deleterious reaction. Certainly, the first step should be the use of ASTM C 295: Petrographic Examination of Aggregates for Concrete. Such a test, by an experienced petrographer, can provide definitive “yes” or “no” answers in most cases, but a “maybe” in others.
ASTM C 289: Potential Alkali-Silica Reactivity of Aggregates (Chemical Method) involves the chemical determination of the potential reactivity of an aggregate with alkalies in portland cement concrete. The test partially duplicates the chemical reaction of the microcrystalline silica in the aggregate, but also counts some silicates that completed their reaction a long time ago as reactive silica. It fails to detect slowly reactive aggregates, doesn’t measure expansive forces (or the lack of them), and may produce unreliable results with some carbonate aggregates. Many false positives or negatives have led to decreased usage of this test. Its value is in providing results in two days.
ASTM C 227: Potential Alkali Reactivity of Cement-Aggregate Combinations (Mortar-Bar Method) involves measuring the change in length of mortar bars at an elevated temperature. Disadvantages of the test are that it takes at least 14 days and as long as one year to perform, and usually does not provide information on slowly reactive aggregates. The test is advantageous in that it can mimic actual performance of the cement-aggregate combination, can provide specimens for petrographic examination, and is performed at 100°F (38°C), which may double or triple the rate of reaction of concrete compared to that in normal outdoor exposure.
ASTM C 1260: Potential Alkali Reactivity of Aggregates (Mortar-Bar Method) also uses a mortar bar. The test primarily differs from ASTM C 227 in that it greatly accelerates any ASR reaction by immersing two-day old mortar bars in a sodium hydroxide solution at 176°F (80°C) for 14 days. Thus it may provide the same information in 16 days provided by C 227 after six months or a year. It is a current method of choice of many laboratories; those laboratories generally state that the test is conservative. However, a footnote in the ASTM test procedure warns that some reactive aggregates may go undetected. Furthermore, some false positives have been reported.
ASTM C 1293: Concrete Aggregates by Determination of Length Change of Concrete Due to Alkali-Silica Reaction involves the length change of concrete or mortar prisms made with 3/4 in. (19 mm) maximum size aggregate and with cement. Alkali content is increased by the addition of sodium hydroxide. Since the test involves the preparation of concrete, it may use the suspect fine or coarse aggregate with known unreactive counterparts. Because it measures expansion occurring during exposure at a temperature of only 100°F (38°C), it takes about five times longer than C 1260 to provide the data.
ASTM C 33: Standard Specification for Concrete Aggregates includes an Appendix X1 entitled “Methods for Evaluating Potential for Deleterious Expansion Due to Alkali Reactivity of an Aggregate.” It provides a useful discussion of alternative methods.
Finally, optical petrography provides a reliable test to identify ASR in an actual concrete structure but only when performed by an experienced concrete petrographer using the procedures of ASTM C 856: Petrographic Examination of Hardened Concrete.
*Stanton, T. E., “Expansion of Concrete Through Reaction Between Cement and Aggregate,” Proc. ASCE, Vol. 66, No. 10, Dec. 1940, pp. 1781-1811.