In this article, the three types of Alkali–Aggregate Reactions are described: (1) alkali–carbonate reaction (ACR), (2) alkali–silicate reaction, and (3) alkali–silica reaction (ASR).
Water in the pores of hardened concrete contains quantities of dissolved ions largely deriving from the cement matrix. In mature concrete, the vast majority of dissolved cations are those of the alkali metals potassium and sodium. These alkali ions derive from cement. During the early periods of cement hydration, sulphate ions are removed from solution through their incorporation into hydration products such as ettringite and monosulphate, which means that the anions balancing the cations in the solution are soon exclusively hydroxide ions. As a consequence, the pH value of concrete pore solutions can be as high as 13.9. Under these highly alkaline conditions, durability problems can arise as a result of reactions between aggregate minerals and hydroxide ions. Products of the reactions are capable of absorbing water, leading to expansion and cracking of concrete. These alkali–aggregate reactions can be divided into three types:
(1) alkali–carbonate reaction (ACR), (2) alkali–silicate reaction, and (3) alkali–silica reaction (ASR).
ASR involves the breaking of bonds in the framework of certain silica bearing minerals to produce an expansive gel. Under high pH conditions, siloxane bonds at the surface of silica minerals are attacked by hydroxide ions in the following manner:
≡Si-O-Si≡ + OH– + R+ → ≡Si-OH + R-O-Si≡
where R denotes either sodium or potassium. The reaction continues in the following manner:
≡Si-OH + OH– + R+ → ≡Si-O-R + H2O
The reaction reduces the original silica network to an open, gel like network that is more accessible to water molecules. The gel undergoes hydration:
≡Si-O-R + H2O → ≡Si-O–-(H2O)n + R+
As water is absorbed into the gel, it swells considerably. It should be noted that the attack of siloxane bonds continues to an extent that the outer layer of the gel will eventually begin to break down completely, releasing silicate groups into solution:
≡Si-O-Si-OH + OH– → ≡Si-OH + O–-Si-OH
The silicate groups removed from the gel are likely to rapidly be involved in reactions between calcium ions to form CSH gel. Thus, the reactions occurring during ASR are also those that occur during the pozzolanic reactions of materials such as ﬂy ash. An alternative mechanism has been proposed in terms of the accumulation of osmotic cell pressure. Under this mechanism, water and both alkali and hydroxide ions can move from the cement matrix into a reacting aggregate particle, but the movement of silicate ions out of the particle into the matrix is prevented by a layer of calcium alkali silica gel formed by a reaction of calcium ions with ASR gel at the cement paste–aggregate interface. Thus, the reaction product layer acts as an osmotic membrane. As water enters the ASR gel formed within the membrane, hydraulic pressure increases to a point where fracture of the aggregate and its surrounding matrix occurs, leading to expansion.
The ACR is often referred to as ‘dedolomitization’ because it can involve the decomposition of dolomite to form brucite (Mg(OH)2) and calcite (CaCO3) in the following manner:
CaMg(CO3)2 + 2ROH → Mg(OH)2 + CaCO3
where R is either sodium or potassium. However, magnesite (MgCO3) will also undergo a similar process:
MgCO3 + 2ROH → Mg(OH)2 +R2CO3
In this case, the alkali carbonate products will react with the cement hydration product portlandite (Ca(OH)2) to give calcite:
R2CO3 + Ca(OH)2 → CaCO3 + 2ROH
There is a net reduction in the volume of the products of these reactions relative to the reactants, which has led to speculation that expansion is the result of expansive clay particles within the carbonate mineral matrix–exposed and unconstrained by dedolomitization – absorbing water.
Other proposed mechanisms include the suggestion that the source of expansion is ASR between microscopic quartz particles in the carbonate matrix, or the expansion of assemblages of colloidal particles present in pores within the carbonate minerals. However, research carried out using carbonate rocks of high purity has led to the general conclusion that the precipitation of brucite in confined spaces is, indeed, the cause of expansion.
The alkali–silicate reaction involves rocks that can contain quantities of minerals with a layered phyllosilicate structure. It has been observed that, under high pH conditions, these minerals exfoliate, permitting water to occupy the space between the layers. However, gel is often also present, indicating that expansion could be the result of ASR, possibly involving strained microcrystalline quartz.