Alkali-Aggregate Reactivity Of The Major Rocks In Sudan

Abstract

Reaction in concrete between alkali solution derived from the cement and reactive silica contained in the aggregate is now known to be capable of causing considerable damage to concrete structures, which may sometimes even lead to their failure. Thus, in many countries, the alkali-aggregate reaction has been thoroughly investigated. It is, therefore, important to consider at the design stage the possible damages arising from alkali-aggregate reaction, and to minimize the risk of its occurrence by choice of suitable materials. The reported investigation is an attempt to examine the likely reactivity of selected Sudanese aggregates obtained from different parts of the country. The aggregate samples used in the testing programme include granitic, volcanic and sedimentary rock materials and natural aggregates. These rock materials which were collected from operating quarries and outcrops are of unknown service record and probably represent the major rock types in the Sudan which have been or might be used for concrete making. Since the significance of the reaction depends on the alkalinity of the cement three Portland cements of different alkali content were used in the investigation. Testing the potential reactivity of these resources which this study seeks to provide, will add significantly to the factual background against which planning policies can be decided. Petrographic properties of the samples were tested by ASTM C 295 method of aggregates for concrete to identify potentially alkali-silica reactive components in the aggregates. Reaction experiments (chemical method: ASTM C 289, mortar-bar methods: NBRI and ASTM C 227) were performed to further identify the magnitude of the reaction and the expansive potential of the aggregates. Furthermore, supplementary information such as visual inspection and thin section examination of the test specimens were developed to confirm that the observed expansion (if any) is due to alkali-silica reaction. Moreover, a correlation of the results using the expansion tests with results of the petrographic examination is made and hence conclusions and recommendations are drawn. In addition, the various test methods used and modification of some of them are reviewed in the light of the results obtained. Also, the major rock units in the Sudan are considered with regard to their alkali reactive potential. Expansion tests demonstrate that some aggregates have been found to be potentially reactive and their expansion was shown to be a direct result of alkali-silica reaction. However, other aggregates were found to be not harmful despite the fact that they formed some amounts of gel but this is not accompanied by excessive expansion. In the ASTM C 289 chemical method all the aggregates were shown to be innocuous. However, the petrographic examination revealed that known potentially alkali-silica reactive constituents are present. Furthermore, when the ASTM mortar-bar test was run at an elevated temperature, in all cases, differences were observed in both the ultimate level and the early rate of expansion, compared with the results of the test when performed according to the relevant standard method. It has been shown that the aggregates which would be classified, according to the ASTM chemical and expansion tests as not harmful have been found to be slowly expansive in the NBRI accelerated test. Therefore, in this study the ASTM chemical and expansion tests might have underestimated the potential reactivity and expansiveness of these aggregates. It is concluded that alkali-silica reaction has been found in some rock aggregates and the rocks which showed expansive reactivity or those found to be marginal should be considered harmful and may produce excessive expansion if used in mortar or concrete. It is recommended that the rocks that were found to be innocuous should be used in mortar or concrete with a Portland cement of lower alkali level, such as 0.6% or less, because they also contain known potentially alkali-silica reactive components.