Elsevier

Cement and Concrete Research

Volume 99, September 2017, Pages 172-182
Cement and Concrete Research

Phase assemblage of composite cements

https://doi.org/10.1016/j.cemconres.2017.05.009Get rights and content

Abstract

The phase assemblage of binary Portland cements with 45 wt% replacement by calcareous fly ash and slag has been investigated by experiment, mass balance and thermodynamic simulations. The input for these calculations was based on clinker reaction measured by XRD-Rietveld refinement, on SCM reaction measured by SEM-EDS full element mapping and on C-S-H composition by SEM-EDS point analyses. A sensitivity analysis estimates the uncertainty of determination of phase volume as at best ± 2 cm3/100 cm3, and thus of rather semi-quantitative character. The differences between experiment and calculations regarding AFt/AFm, C-S-H and portlandite were assessed. Gel/space ratios computed using the volumes of all hydrates showed a clear correlation to compressive strength, which was unaffected by the mix composition. This suggests that the type of hydrates formed has little or no influence on the compressive strength and that the key factor is the space filling.

Introduction

The quantitative determination of the phase assemblage in hydrated composite cements is a challenging task due to the presence of several amorphous phases. These mainly concern one or more glassy phases from the anhydrous supplementary cementitious material (SCM) and the principal hydration product, C-S-H. Some of the AFm hydrates may also form in a poorly crystalline form, which is important in composite pastes as their content can increase significantly if the incorporated SCMs are rich in reactive alumina.

The currently available techniques for characterization of phase assemblage of hydrated cements often struggle to distinguish and accurately quantify multiple amorphous phases. Thermogravimetric analysis (TGA) enables identification of hydrate phases in cement paste but due to proximity of the temperatures of thermal decomposition of many of the hydrates, quantitative results can only be obtained for portlandite. X-ray powder diffraction (XRD) with Rietveld refinement is a powerful technique to quantify crystalline phases in cement paste but fails to distinguish between multiple amorphous components without prior calibration. Despite advanced approaches such as PONKCS (Partial Or Not Known Crystal Structure) [1], [2] an accurate quantification of individual amorphous phases remains very challenging [3]. Electron microscopy with image analysis can be used to analyse the main hydration products [4] and to extract anhydrous SCM particles [5]. However, the quantification of hydrates is rendered impossible by fine intermixing of C-S-H and other hydrates and porosity at a sub-micron scale below the resolution limit of the scanning electron microscope.

To overcome the limitations of experimental techniques, analytical approaches such as thermodynamic simulations [6], [7], [8] and mass-balance [9], [10] have been proposed to calculate the phase assemblage from the quantities of clinker and SCMs reacted. Thermodynamic simulation models an equilibrium phase assemblage whereas mass-balance distributes the reacted elements among phases assumed to precipitate in the analysed system. Studies of composite cements containing siliceous fly ashes [11], [12] showed that XRD-Rietveld and thermodynamic simulations for the evolution of amount of portlandite and ettringite over time, agree within ± 5–10 wt.-%. Further, in the above studies, De Weerdt et al. [11] showed good correlation of coarse porosity measured by electron microscopy with compressive strength, while Zajac and Ben Haha [12] used the modelled phase volumes to calculate total porosity and relate it to strength. It is generally accepted that compressive strength is a function of the amount of hydrates formed. This dependence has been studied by many researchers and notably Powers, who provided an empirical relationship of compressive strength vs. gel/space ratio of Portland cements [13], [14]. More recently, Lam and Poon have calculated these values for fly ash blended cements but observed significant scatter [15]. This was likely because the gel/space ratio was calculated from the degree of reaction of clinker and fly ash, assuming that the volume of hydrates is around 2.06 of the volume of binder reacted. However, the type and amount of hydration products formed in the presence of SCMs may differ significantly from those in Portland cements and may vary according to the type and amount of SCM. It thus seems that the assessment of strength vs. gel/space of composite cements should be based on more precise a determination of the amount of “gel”. A promising result has been presented by Schöler et al. [16], who compared compressive strength of fly ash and slag cements with the total volume of hydrates calculated by thermodynamic modelling.

The aim of this study was to quantify and compare the phase assemblages of binary cements with high replacement (45 wt%) by calcareous fly ash and slag using thermodynamic modelling and mass balance and to look at their correlation to compressive strength. The input for the calculations was the amount of clinker reacted measured by XRD-Rietveld and the amount of slag and fly ash reacted measured by scanning electron microscopy using the technique described in [5]. The calculated phase assemblages were compared with experimental phase assemblages by XRD-Rietveld refinement. A sensitivity analysis was carried out to determine how an uncertainty linked to the modelling input affects the total volume of hydrates predicted. Capillary porosity obtained from the thermodynamic and mass-balance simulations and that measured by mercury intrusion porosimetry were compared to the compressive strength of mortars. Further, Power's gel/space ratio was calculated from the total volume of hydrates modelled and compared to the compressive strength.

Section snippets

Materials

One Portland cement (PC), one slag (S) and two calcareous fly ashes (FA1 and FA2) characterised in a previous study [5], [17] were the primary materials in this study. Their chemical composition was measured by X-ray fluorescence and their phase composition was analysed by X-ray diffraction with quantitative Rietveld refinement. The results are shown in Table 1. The XRF, as mainly dependent on the calibration, has the uncertainty of around +/0.1 wt%. The XRD-Rietveld depends much more on the

Method for calculation of phase assemblage

Thermodynamic simulations and mass-balance were used to calculate the quantities of hydrates. As the input for the two methods we used the amount of clinker reacted measured by XRD-Rietveld and the amount of SCMs reacted measured by SEM-EDS. To eliminate scatter in the input, the amount of unreacted clinker phases by XRD-Rietveld refinement and of the amorphous SCM glasses by SEM-EDS image analysis as a function of time were smoothed by fitting a curve:y=aexpblnx0.1c+d

An example of the

Reaction of clinker

The degrees of reaction of clinker in the systems studied are shown in Fig. 6. At early hydration times the presence of SCMs enhanced the reaction of clinker similar to that of the inert filler. This is due to filler effect, i.e. more water and space available per gram of clinker [38]. At later ages the rate of clinker reaction in the substituted systems is reduced compared to the system with quartz substitution as the hydration of the SCMs decreases the space available for clinker hydration

Conclusions

This study shows the potential and the drawbacks of thermodynamic modelling and mass-balance calculations to determine the phase assemblage of hydrated composite cements, here with 45 wt% slag and calcareous fly ash. The input is comprised of the degree of hydration of clinker measured by XRD-Rietveld refinement, that of SCMs measured by electron microscopy full element mapping and the chemical composition of the C-S-H obtained by SEM-EDS spot analysis. Due to considerable uncertainties linked

Acknowledgements

The authors would like to acknowledge HeidelbergCement AG for the financial support.

References (46)

  • J.E. Rossen et al.

    Optimization of SEM-EDS to determine the C–A–S–H composition in matured cement paste samples

    Mater. Charact.

    (2017)
  • R.S. Barneyback et al.

    Expression and analysis of pore fluids from hardened cement pastes and mortars

    Cem. Concr. Res.

    (1981)
  • T. Matschei et al.

    Thermodynamic properties of Portland cement hydrates in the system CaO–Al2O3–SiO2–CaSO4–CaCO3–H2O

    Cem. Concr. Res.

    (2007)
  • B. Lothenbach et al.

    Thermodynamic modelling of the effect of temperature on the hydration and porosity of Portland cement

    Cem. Concr. Res.

    (2008)
  • D.A. Kulik

    Improving the structural consistency of C-S-H solid solution thermodynamic models

    Cem. Concr. Res.

    (2011)
  • B.Z. Dilnesa et al.

    Fe-containing phases in hydrated cements

    Cem. Concr. Res.

    (2014)
  • E. Berodier et al.

    Evolution of pore structure in blended systems

    Cem. Concr. Res.

    (2015)
  • K.L. Scrivener et al.

    Advances in understanding hydration of Portland cement

    Cem. Concr. Res.

    (2015)
  • B. Lothenbach et al.

    Supplementary cementitious materials

    Cem. Concr. Res.

    (2011)
  • A.C.A. Muller et al.

    Influence of silica fume on the microstructure of cement pastes: new insights from 1H NMR relaxometry

    Cem. Concr. Res.

    (2015)
  • N.V.Y. Scarlett et al.

    Quantification of phases with partial or no known crystal structures

    Powder Diffract.

    (2006)
  • I.C. Madsen et al.

    Description and survey of methodologies for the determination of amorphous content via X-ray powder diffraction

    Zeitschrift Fur Krist.

    (2011)
  • P.T. Durdziński et al.

    Outcomes of the RILEM round robin on degree of reaction of slag and fly ash in blended cements

    Mater. Struct.

    (2017)
  • Cited by (96)

    View all citing articles on Scopus
    View full text