Drying shrinkage cracking
As the water in the concrete dries out, the capillary forces set up in the pores increase and cause the concrete to shrink. Such ‘drying shrinkage cracks’ appear later than the previously mentioned cracks, usually weeks, or months after placement. Loss of moisture from cement paste results in shrinkage.
Figure 2.6 Drying shrinkage cracking to a potable water tank and associated lime leach deposition together with localised corrosion risk to reinforcement intersecting cracks. (Courtesy of Chris Weale)
If the shrinkage is restrained cracks may form because tensile stresses will be brought about by differential shrinkage. The risk of drying shrinkage cracks can be reduced by using the maximum possible amount of aggregate since it is only the cement gel that shrinks on curing. Reduced water content and adequate curing assist the prevention of drying shrinkage cracks. Typical drying shrinkage cracks are shown as I in Figure 2.3.
Figure 2.6 shows an example of drying shrinkage cracks to a potable water tank where unsightly lime leach staining has occurred and localised corrosion is a risk where the cracks intersect reinforcement due to reduced pH (i.e. pH < 10 leading to dissolution of the passive film).
Crazing occurs whenever a weak surface layer is formed on the surface and this weak surface layer is unable to withstand quite small stresses which result from the differential shrinkage between the surface and the bulk. This may arise as a result of the separation of aggregate and cement in the very early stages of curing which leaves a thin layer of‘laitance’ on the surface of the concrete or it may result from the very rapid evaporation of water from the surface of the concrete when it is placed under hot conditions. Crazing is shown at J and К in Figure 2.3.
Alkali aggregate reaction cracking
AAR cracking occurs because certain types of siliceous or carbonate rock aggregates can react with the sodium and potassium hydroxides in cement to cause concrete damage, i.e. alkali-silica reaction (ASR) and alkali-carbonate reaction (ACR). Such alkali aggregate reactions are much more common in Australasia with reactive silica aggregates than carbonate rock aggregates which are mainly dolomitic in origin. Reactive siliceous aggregates include opaline and chalcedonic cherts, tridymite, cristobalite, rhyolites, and andesites, and their tuffs, certain zeolites, and certain phvllites. Reactive silicas have a random network of silicon-oxygen tetrahedra, while unreactive silicas such as quartz have orderly tetrahedra. The ASR produces an alkali silicate gel. The gel attracts water from hardened cement paste and causes swelling. Osmotic pressure and expansion result, which in turn produces cracking and spalling of the cement paste, and exudation of gel. The rate of ASR depends on the fineness of siliceous material, alkali content, and water content of the cement paste, aggregate porosity, and cement paste penetrability (permeability). Typical AAR cracks are shown at N in Figure 2.3 and below in Figure 2.7.
Figures 2.8 and 2.9 show examples of ASR cracking to the reinforced concrete substructure elements of a coastal, marine, jetty structure in Western Australia where the effect of restraint to cracking is evident. In Figure 2.8 map cracking (three pronged, similar to crazing) is the
Figure 2.1 Alkali aggregate reaction cracking. (Courtesy of Bungey and Millard, 1996, p.9)
Figure 2.8 ASR map cracking to a transverse beam of a coastal jetty structure
Figure 2.9 ASR cracking to a longitudinal beam of a coastal jetty structure
characteristic pattern for ASR in unrestrained concrete. In other cases, where ligature reinforcement stretch and the element is in compression, longitudinal cracks may form, refer Figure 2.9, where the photograph shows a prestressed beam, whereby the ligatures have yielded and the crack is at the location of maximum stress.
As the above jetty structure was in a coastal marine environment, the ASR-induced cracking permitted the entry of water and chloride ions such that localised reinforcement corrosion presented a serious risk to the structure. With moisture ingress and residual reactive silica within the coarse and fine aggregate ongoing ASR expansion also needed to be managed for this structure.
Restraint has also had an effect with the ASR-induced crack patterns of other structures and elements. Examples include:
- • Longitudinal cracking of prestressed road bridge planks.
- • Mid-face splitting of road bridge headstock elements and map cracking at ends.
- • Vertical cracking of prestressed bridge piles (also leading to pitting corrosion of prestressing steel).
In terms of testing for AAR in new construction, it is advisable to test for an aggregate’s (coarse and fine aggregate) reactivity if this is not already known prior to selection for use in concrete. Various different tests are available for assessing the reactivity of aggregates prior to selection for use in concrete. These include petrographic examination of the aggregate, chemical analysis, and mortar bar, or concrete prism expansion tests based on standard mix designs. For mortar bar and concrete prism tests, expansion limits have been set to indicate the reactivity of the aggregate under test. Different AAR test methods can give contradictory results. There is ongoing research to improve test methods so that they better predict field performance of aggregates. However, interpretation must be based on local knowledge about the alkali reactivity of similar aggregates in the specific test used and in concrete structures.
Petrographic examination of aggregates is undertaken in accordance with ASTM C295-12 (2012) or AS 1141.65 (2008). Where previous use or petrographic tests do not provide sufficient evidence that the aggregates will be acceptable the Concrete Institute of Australia ‘Performance Tests to Assess Concrete Durability’ (2015) recommended practice document proposes that accelerated mortar bar testing to AS 1141.60.1 (2014a) be undertaken to determine the degree to which the aggregate is reactive. Furthermore, that recommended practice advises that where petrographic or mortar bar expansion tests are inconclusive then accelerated concrete prism testing to AS 1141.60.2 (2014b) can be used to provide specific evidence as to whether a proposed concrete mix and cement system will have acceptable resistance to AAR.
Results of accelerated testing can occasionally be misleading, and more than one test method might be needed to resolve the potential AAR risk. For example, in Western Australia a widely-used aggregate that is classified ‘potentially reactive’ by petrography, gives ‘highly reactive’ expansions by mortar bar test, but is classified ‘innocuous’ by concrete prism and has a 40+ years ASR problem free track record in marine and water retaining structures. Thus, wherever possible, the specific tests used should be those upon which knowledge about reactivity of local aggregates is based, and local experience should be utilised in interpreting test results (Concrete Institute of Australia, 2015).
Whichever test methods are used, the art of the testing is in designing a test programme that best meets project needs, rather than simply applying a pass/fail criterion associated with a particular test method. For example, with use of appropriate control mixes, test results can be evaluated by comparison with aggregates or concrete with known site performance (Concrete Institute of Australia, 2015).
Because significant variations in mineralogy can exist within quarries, material is typically characterised by petrography annually, and AAR testing is repeated at a frequency that reflects changes in the source material. For some sources this may be as often as yearly (Concrete Institute of Australia, 2015).