Round robin on Rietveld quantitative phase analysis of Portland cements

Interlaboratory studies on the precision and accuracy of Rietveld quantitative phase analysis (R-QPA) of mixtures of crystalline phases have already been carried out. However, cement-related materials are samples of variable complexity, ranging from three phases in white Portland clinkers to more than eight phases in grey cements, that need to be specifically investigated. Here, the results are reported from a round robin on the R-QPA of cement-related materials from laboratories with experience in this type of analysis. The aim of the work was to evaluate the levels of precision and accuracy associated with Portland clinkers and cements. Two sets of samples have been investigated, artificial mixtures and commercial samples. Artificial mixtures were prepared by mixing (weighing) synthesized single-crystalline phases in the appropriate proportions: (i) white clinker (Ca₃SiO₅, Ca₂SiO₄ and Ca₃Al₂O₆) and (ii) grey cement (Ca₃SiO₅, Ca₂SiO₄, Ca₃Al₂O₆, Ca₄Al₂Fe₂O₁₀, CaCO₃ and CaSO₄·2H₂O). These two samples were used to assess the accuracy and uncertainty of the procedure, as an expected mineralogical phase fraction, the `true mineral­ogical percentage', is available under the assumption of negligible non-diffracting contents. In order to assess the validity and limitations of the Rietveld-based approach for cement materials, three commercial samples were measured: (i) white Portland clinker, (ii) grey Portland clinker and (iii) a type-I grey Portland cement. The samples studied have been chosen in order to cover most of the different typologies of binders. Reproducibilities and general uncertainty values, with a level of confidence of 95%, are reported and discussed.