Methods to measure the reaction degree of slag in blended cement
The reaction degree of slag in blended cement is difficult to measure using those methods which are successfully applied for Portland cement. Demoulian et al. described a method based on extraction of the constituents other than unreacted slag with a reagent based on Ethylene Diamine Tetraacetic Acid (EDTA), finding it effective for determining reaction degree of slag in unhydrated blended cements.
EDTA has long been the most used and well-known method to measure the reaction degree of slag in blended cement; however, as stated by Luke and Glasser, special care should be taken when applying this method, e.g. systematic error may occur due to the residue of hydrotalcite in the extraction of unreacted slag.
Lumley et al. reported that EDTA extraction method can be used to determine the degree of reaction of slag under certain corrections, but the accuracy is poorer at degrees of reaction above 70%.
Besides EDTA method, SEM-BSE image analysis could be applied to directly estimate the hydration degree of both plain cement and blended cement paste. This method is based on the fact that cement and slag grains can be distinguished from unhydrated clinkers and paste matrix, though there is some overlap due to the similar gray levels of cement and slag. However, it should be bore in mind that using this method may produces an over-estimated result on the reaction degree of slag due to its fine particles (G. Ye).
To improve the precise of ESEM image analysis determining the reaction degree of slag, V. Kocaba studied SEM-IA-Mg-mapping method to clearly distinguish slag phase from clinkers. The methodology is based on the Mg element of slag does move during the hydration process, thus the presence of Mg can mark the unreacted slag phase. The disadvantage of this method is extremely time consuming, which is the drawback of SEM-BSE image analysis as well.
As a conclusion,there is no perfect method so far to determine the reaction degree of slag in blended cements. The best solution may be using more than one method to validate each other and get reliable result.