Experimental and theoretical study of delayed ettringite damage in concrete

An experimental study is conducted to investigate the effect of a range of parameters on concrete expansion and premature deterioration associated with delayed ettringite formation (DEF). These parameters were identified from a comprehensive literature review; they include concrete curing conditions and cement chemistry. Specifically, the study investigated the effect of potassium and magnesium contents as well as the effect of water and steam curing conditions on concrete damage associated with DEF. An accelerated test method for DEF was evaluated and modifications are suggested.; Concrete specimens were prepared with varying amounts of potassium and magnesium and subjected to different curing conditions. To accelerate any potential expansion, a beat cycle was employed. The change in length of the specimens was monitored for the duration of the study. Specimens were analyzed using the scanning electron microscope, x-ray dispersive analysis, and x-ray diffraction to determine the mechanism responsible for deterioration. Computed tomography x-ray was also employed as a non-destructive method for examining the internal crack patterns associated with expansion.; The expansion data is fitted using the Kohnogorov-Avrami-Johenson-Mehl (KAJM) equation. This theory makes it possible to identify individual expansion processes and to identify the contribution of different expansion mechanisms.; All concretes showed expansion following the beat cycle. Subjecting concrete to the heat cycle has a primary role in accelerating expansion. The age at which concrete is subjected to the beat cycle as well as the sample dimensions have a profound effect on the resulting expansion. The use of the concrete prisms as suggested in the modified test method offers a better alternative to cores suggested by Duggan.; Increasing the potassium content has a deleterious effect on concrete expansion and premature deterioration and is also associated with dramatic reduction in compressive strength. Expansion results clearly indicate that steam curing is not a necessary factor for damaging expansion to occur. Increasing the MgO content up to 4.38% and the use steam curing have a minor effect on concrete expansion.; Extensive petrographic investigation of the expanded concrete revealed the presence of signs suggestive of DEF with no signs of alkali silica reaction or brucite. KAJM equation parameters were collected from fits to the experimental data and used to predict concrete expansion.