Effect of type of aggregate on delayed ettringite formation

Abstract

Portland cement mortars were made from natural quartz and crushed limestone sands, and combinations of thetwo. Three types of concrete were made from granite and quartz sand, coarse and fine flint, and coarse and finelimestone. Mortar and concrete prisms were cast and cured at 95–100°C and subsequently stored at roomtemperature under water. The quartz sand mortars exhibited large expansions within the first year but thelimestone mortars remained stable even after 6 years. The ultimate expansion of those mortars with combinationsof limestone and quartz aggregate had intermediate values. The granite aggregate concretes expanded readily butthe limestone and the flint concretes started to expand only after very long induction periods. X ray diffraction,scanning electron microscopy with back-scattered electron imaging and X ray microanalysis were used to examinethe composition and microstructure of the hydration products. Differences in the features of ettringite bands thatdeveloped in the various heat-cured mortars and concretes were observed. It was shown that the rough surface ofthe limestone sand and aggregate particles was able to interrupt the development of an ettringite band in thepaste-aggregate transition zone. Microstructural observations give support to the growth of ettringite bands asthe mechanism responsible for the observed expansion.

Five year monitoring of curing solutions of heat-cured mortars affected by delayed ettringite formation

“…Some studies were carried out on specimens treated at 1008C to accelerate the phenomena even though such temperatures are never reached in precast industries or inside massive pieces of concrete (Lawrence, 1995;Yang et al, 1996Yang et al, , 1999aYang et al, , 1999b. The temperature threshold for the development of DEF remains unclear.…”

Surely we know all about cement – don't we?

“…The prisms were heat cured at 95uC for 6 h and subsequently stored in water for 6 years, before being examined by Yang at the University of Sheffield, using scanning electron microscopy in the back scattered electron image (BEI) mode. 15 The heat cured concrete samples had all expanded, but to different extents, whereas the control sample, made from the granite aggregate and stored at room temperature continuously throughout the six years did not show any significant dimensional change (Fig. 7).…”

“…This expansion is a serious corrosion problem, because it can be enough to disrupt the concrete or mortar; the phenomenon is known as delayed ettringite formation or DEF. Typical expansion curves 15 are shown in Fig. 7 for three concretes that were made by Lawrence at the British Cement Association using the same cement and storage conditions, but three different aggregates.…”

“…8-10, modified from those published by Yang, Lawrence and Sharp. 15 It can be seen from Fig. 8 that the concrete made using the granite aggregate that had expanded the most, a network of microcracks had formed which passed through the cement paste matrix, dividing it into many sections with dimensions of hundreds of micrometres.…”

“…The limestone concrete was only just starting to expand after storage for six years, so the aggregate is not merely an inert filler but controls the rate of expansion. In view of these observations, Yang, Lawrence and Sharp 15,17 concluded that: (i) few bands extend completely around the aggregate grains (ii) the width of the bands is not related to the aggregate grain size (iii) the bands do not exist in isolation, but form part of a crack network involving paste cracks as well. To meet these criticisms, the uniform paste expansion hypothesis has been modified by Famy and Taylor 18 and Taylor, Famy and Scrivener.…”