Alkali Aggregate Reaction in Concrete

Most of the structures severely cracked by AAR are exposed to the whether or underground in contact with damp soil. This is because for significant expansion to occur sufficient presence of moisture is essential. Apart from the moisture, high content of alkali in the concrete is also essential. No case as been found where the alkali content, in terms of equivalent sodium oxide (Na₂O_equi), is below 3-4 kg per cubic meter of concrete.

It is also found that, when there are sufficient moisture and alkali, maximum expansion of concrete due to AAR occurs when the content of reactive minerals in aggregate is within a sensitive region, some refer to this as “pessimum” content. Content of reactive minerals below or greater than the pessimum value, AAR expansion reduces.

From the above, it can be seen that, for a damage AAR expansion to occur, it is necessary to have

• sufficient moisture supply,

• high content of alkali in concrete, and

• pessimum amount of reactive minerals in aggregate.

Sources of alkalis in concrete

a. Cement

All ingredients of concrete may contribute to the total alkali content of the concrete, the major source of alkali is from cement. The chemical composition of cement is usually expressed in terms of oxides. In relation to AAR, alkali content in cement is determined from Na₂O and K₂O. Alkali content is described as total mass of “equivalent sodium oxide”, Na₂O_equ, which is determined from the following expression.

Na₂O_equ = Na₂O + 0.658 K₂O

b. Pozzolans

A pozzolan is a siliceous or siliceous and aluminous material which react with lime released from cement hydration forming a compound possessing cementitious properties. Pozzolanic materials are used as a cement replacement or as part of cenmentitious material to modify or improve properties of concrete, sometimes for economical consideration.

Common pozzolanic material used in concrete include PFA (palverized fuel ash, or fly ash), silica fume, GGBS (ground granulated blastfurnace slag). Other pozzolans include volcanic ash (the original pozzolan), opaline shale and chert etc.

Pozzolan consumes alkali when react with lime. When considering pozzolan contribution of alkali to concrete, a reduction to the alkali content of the pozzolan should be allowed for.

c. Aggregate

Aggregate containing feldspars, some micas, glassy rock and glass may release alkali in concrete. Sea dredged sand, if not properly washed, may contain sodium chloride which can contribute significant alkali to concrete.

d. Admixtures

Admixture in the context of AAR in concrete means chemical agents added to concrete at the mixing stage. These include accelerators, water reducers (plasticizers), retarders, superplasticizers, air entraining, etc. Some of the chemicals contain sodium and potassium compounds which may contribute to the alkali content of concrete.

e. Water

Water may contain certain amount of alkali.

f. Alkalis from outside the concrete

In area of cold whether, de-icing salt containing sodium compounds may increase alkali content on the surface layer of concrete. Soils containing alkali may also increase alkali content on the surface of concrete.

Alkali content and AAR

Research show that when the total alkali content, in terms of equivalent sodium oxide, is less than 3 kg/m³, damage expansion due to AAR is unlikely to happen, provide that known highly alkali-reactive minerals, such as opal and glass, are not present in the concrete.

Because of the above research findings, and also because no reliable universal testing method have been established for the determination of reactivity of an aggregate, limiting alkali content in concrete has become the most widely used approach for the control of AAR.

Reactive aggregate

General types of AAR

There are mainly three types of AAR found in concrete. These are

• alkali-silica reaction

• alkali-silicate reaction and

• alkali-carbonate reaction

Alkali-silica reaction is a reaction between alkali hydroxides and free silica in aggregate form a alkali-silica gel.

Alkali-silicate reaction is the same as alkali-silica reaction except that in this case the reactive constituent is not free silica present in the combined form of phyllosilicates.

Alkali-carbonate reaction occurs in concrete when alkalis react with certain dolomitic lime stones containing clay. Reaction causes cracks allowing water to enter which causes the clay to swell and disrupt the aggregate.

Majority of the structures affected by AAR is found due to alkali-silica reaction. Alkali-silicate and alkali-carbonate reaction is relatively rear.

Tests for aggregate reactivity

Many test methods have been proposed by researchers for identifying potential reactivity of aggregate all over the world. These may be classified into three types.

• Petrographic examinations

• Chemical tests

• Expansion tests

Petrographic examination is the process of identifying the types of minerals in aggregate or concrete section by observation using microscope or other aids. This method can identify types of minerals in the aggregate and give suggestions as for whether the aggregate is potentially reactive or not. Because the uncertainties involved in the test, the method is generally used as a screening test as a part of an investigation.

Many proposed test methods using chemical analyses to identify potential reactivity of aggregate. ASTM C389-87 chemical test (ASTM 1987a), for example, evaluates aggregate reactivity by measuring the amount of dissolved silica and the reduction of alkalinity in the reaction alkali solution.

In an expansion tests, mortar bars or concrete prisms are made using the aggregate to be investigated. There specimens are then put in to a specified condition and the expansion of the specimens are measured. Since at normal climate conditions the reaction will take a few years or even longer to complete, measures to accelerate the reaction sometimes adopted for such tests. One of such measures is to place the specimens at 38?C or 40?C and 100% relative humidity. In such a condition the AAR and its expansion complete within a few months time. In another such test using mortar bar as specimens, the mortar bars are immersed in a NaOH solution at 80?C. This method is also referred to as Accelerated Mortar Bar Test.

It should be noted that the methods that have been proposed so far have their limitations. Some succeeds in identifying reactivity for certain aggregates whereas fails for others. Therefore it is difficult to ascertain an aggregate is absolutely non-reactive using the currently available testing methods.