Alkali-silica reaction in concrete
Alkali-silica reaction (ASR) can cause serious expansion and cracking in concrete, resulting in major structural problems and sometimes necessitating demolition.
ASR is caused by a reaction between the hydroxyl ions in the alkaline cement pore solution in the concrete and reactive forms of silica in the aggregate (eg: chert, quartzite, opal, strained quartz crystals).
A gel is produced, which increases in volume by taking up water and so exerts an expansive pressure, resulting in failure of the concrete. In unrestrained concrete (that is, without any reinforcement), ASR causes characteristic 'map cracking' or 'Isle of Man cracking'.
Gel may be present in cracks and within aggregate particles. The best technique for the identification of ASR is the examination of concrete in thin section, using a petrographic microscope. Alternatively,polished sections of concrete can be examined by scanning electron microscopy (SEM).
Figure 1 Concrete thin-section, viewed with a petrographic microscope, showing a chert aggregate particle (at the right of the image) from which alkali-silica gel has extruded into adjacent cracks.
Figure 2 Polished section of concrete, viewed with a scanning electron microscope, showing a chert aggregate particle with extensive internal cracks due to alkali-silica reaction. The cracks extend from the aggregate into the nearby concrete (arrowed).
Figure 3 Detail of the chert particle in the previous image and adjacent cement paste, showing alkali-silica gel extruded into cracks within the concrete. Ettringite is also present within some cracks.
The conditions required for ASR to occur are:
- A sufficiently high alkali content of the cement (or alkali fromother sources)
- A reactive aggregate, such as chert
- Water - ASR will not occur if there is no available water in the concrete, since alkali-silica gel formation requires water
The use of pozzolans in the concrete mix as a pertial cement replacement can reduce the likelihood of ASR occurring as they reducethe alkalinity of the pore fluid.
Keeping other factors constant, but varying the proportion of reactive aggregate in test mixes, shows a maximum expansion at a particular aggregate content - higher or lower levels of reactive aggregate will give a lower expansion. This is the so-called 'pessimum effect'.
The process of alkali-silica reaction is believed to be in many respects similar to the pozzolanic reaction, such as occurs normally in concrete containing pulverised fuel ash (PFA), for example. However, there is an important difference. In the pozzolanic reaction small pozzolanic particles are reacting in a Ca-rich environment, while ASR occurs in mature concrete and involves larger particles of aggregate.
The pozzolanic reaction mechanism is believed to be a process in which silicate anions are detached from the reactive aggregate by hydroxyl ions in the pore fluid. Sodium and potassium ions are the ions most readily-available to balance the silicate anions and an alkali-silicate gel is formed. This can take up (imbibe) water and is mobile. The alkali-silicate gel is unstable in the presence of calcium and calcium silicate hydrate (C-S-H) is formed.
In the pozzolanic reaction where a pozzolan is used as a partial cement replacement, the particles are small. As there is much calcium available in young concrete, the alkali-silicate gel forms in a thin layer around the pozzolanic particle and quickly converts to C-S-H. No expansion results.
In the case of alkali-silica reaction, the reaction usually occurs much later, possibly years after the concrete was placed. Large aggregate particles (large, that is, compared with cement-sized pozzolan) generate a significant volume of gel which then takes up water and expands within the hardened, mature concrete.
Because the concrete is mature, calcium availability is limited as most of the calcium is bound up in stable solid phases. The rate of supply of calcium is therefore insufficient to convert the gel quickly to C-S-H.
Expansion of the gel as water is taken up, may result in damage to the surrounding concrete. Over time, the gel slowly does take up calcium; eventually the composition of the alkali-silica gel may become very similar to that of the calcium silicate hydrate in the cement paste (see Figure 4). By then, though, the damage to the concrete may have already been done.
Figure 4 Same image as in Figure 3, with X-ray spectra superimposed showing how alkali-silica gel composition changes with time to become more like the surrounding calcium silicate hydrates. In a) the gel spectrum shows large peaks due to silicon and potassium (the alkali) and only a very weak peak due to calcium. In b) the calcium peak had become much stronger and the potassium peak much weaker. In c) the potassium peak has disappeared entirely and the gel has approximately the same composition as the normal calcium silicate hydrate comprising the bulk of the cement paste. Clearly, the gel is older with increasing distance from the aggregate particle in which it originated - the 'oldest' gel has had more time in which to take up calcium from the surrounding paste and has become calcium silicate hydrate.
A standard work on the subject of alkali-silica reaction, published by the Concrete Society in the UK, is:
ALKALI-SILICA REACTION - Minimizing the risk of damage to concrete: Report TR30