Delayed ettringite formation - Processes and problems

The scope of the delayed ettringite formation (DEF) problem is reviewed. Based on the experience of the writer and others, DEF is to be expected in both steam cured and non-steam cured concrete. The microstructural details appear identical. The common association with alkali-silica reactions is explored and explained in terms of both prior cracking providing space for ettringite and in terms of the reduction in alkali hydroxide concentration in the pore solution attendant on ASR. The controversy concerning basic DEF mechanism between proponents of 'homogenous paste expansion' and 'crystal pressure' is explored. In the writer's opinion the microstructural evidence is not compatible with homogeneous expansion, but is a distinct consequence of crystal pressure developed by ettringite. The thermodynamic analysis that limits expansive pressure in terms of the degree of supersaturation that can be supported may be applicable to ettringite in air voids, but is inapplicable to ettringite in cracks. Fracture mechanics considerations predict local stress concentrations at the crack tips that are many times those that can be generated in open spaces. An apparent example of such effects in opening up space between mica lamellae in steam cured concrete by ettringite deposition is provided. (C) 1996 Elsevier Science Limited.