Domain size as a parameter for studying the potential alkali-silica reactivity of chert-bearing aggregates

Abstract

The presence of chert or flint in aggregates used for concrete manufacture has long been a source of concern in the context of alkali-silica reactivity. The global in-service performance of chert and flint-bearing aggregates, however, varies from innocuous to deleterious. This could be due to variations in the degree of crystallinity of the silica, which influences potential reactivity. Chert occurs in significant Irish sources of aggregate used for concrete, yet no cases of deleterious expansion have been reported in the Republic of Ireland to date. Many of these sources fail to conclusively demonstrate innocuous behaviour in the laboratory expansion tests despite satisfactory in-service behaviour. A previous X-ray diffraction study, employing quartz crystallinity index, had indicated a significant difference between Irish Carboniferous cherts and English Cretaceous flints. This paper reports a further study of crystallinity, correlated with mortar bar expansion tests, but employing both quartz crystallinity index and domain size in the characterisation of crystallinity. The influence of chert content was also studied. Twenty-three Irish aggregate sources were sampled, and petrographic analysis revealed that 17 of these contained chert. Three flint-bearing aggregates, from sources in England that had demonstrated alkali-silica reactivity, were also sampled. Chert and flint were extracted from the aggregate samples by a petrographer experienced in the field of concrete technology. Twenty-six distinct samples were successfully extracted from 13 of the Irish and the three English sources. X-ray diffraction testing confirmed the previous findings (using the quartz crystallinity indices) that the Irish cherts are,more crystalline than the English flints. Domain size determination, however, showed that the difference in crystallinity of the national sets was less than that anticipated. The influence of chert content is advanced as a hypothesis to explain the apparent reactivity of Irish cherts in screening tests. The potential value of combining domain size and chert content determination in helping to classify aggregate reactivity is advanced for cases where in-service behaviour is in significant conflict with findings from standard laboratory tests.