Next: Conclusions Up: Main Previous: Stereological analysis

Cement example

The Cement and Concrete Reference Laboratory (CCRL) portland cement proficiency sample program for inter-laboratory testing distributed its first sample in 1936 for chemical and physical testing. A program is underway at NIST to examine performance properties relative to compositional and physical parameters of these cements. These images are also utilized within the Virtual Cement and Concrete Testing Laboratory (VCCTL) where computer-based simulation models are used to predict cement and concrete performance properties [19].

CCRL cement #142 was selected for analysis in this example. A graphical display of phase abundance averages versus the number of fields shown in Fig. 8. For this cement, it is apparent that four fields provide a good estimate of phase abundance.

Fig. 8. Cumulative mass fraction averages versus number of fields shows the estimates stabilize after four fields.

Fig. 9 compares three separate phase composition estimates based upon SEM imaging and X-ray powder diffraction (QXRD). Table 3 provides individual field estimates of the SEM data and that by the Bogue calculation as provided in ASTM C-150. The SEM estimates of both the bulk and surface phase fractions are expressed as mass fractions for comparison. The QXRD data are based upon a bulk cement and salicylic acid-methanol extraction data refined using Rietveld analysis [20].

Table 3
Mass fraction estimates of CCRL 142 cement based upon SEM imaging of six fields of view at 500x magnification. ASTM C-150 estimates of bulk phase abundance are provided in the bottom row.
  AliteBeliteFerriteTricalcium
aluminate
Alkali sulfateGypsumPericlaseQuartz
Field 163.29.613.61.82.2 6.13.50.0
Field 249.624.97.68.61.1 2.65.50.1
Field 365.615.86.60.50.5 4.66.40.0
Field 463.613.35.87.60.6 4.14.60.4
Field 564.8 15.35.49.20.71.43.00.1
Field 659.719.25.78.90.5 2.93.20.0
Average61.116.37.56.10.9 3.64.30.1
Std. Dev.6.05.33.13.90.7 1.71.40.2
 
ASTM C-15050.919.78.37.7  

 

Compared to the XRD data, the Bogue estimates (Table 3) are low for alite and high for belite, aluminate and ferrite. They do not provide an estimate for periclase, alkali sulfate or gypsum. In contrast, in Fig. 9 it is seen that the SEM bulk and the QXRD data generally show agreement within the error bars of their estimates.

The surface phase fractions are of interest in early-age cement hydration as they are often quite different from the bulk phase composition. SEM imaging and image analysis provide a means to make these measurements on cements. Comparison of bulk versus surface area mass fractions is provided in Fig. 9. Another means of examining this relationship is via a surface/bulk ratio plot as shown in Fig. 10. The difference in surface to bulk phase fraction reflects a combination of clinker texture and grinding characteristics. For cement #142, alite surface area is substantially lower relative to the bulk, belite is about equally represented, and the phases tricalcium aluminate, periclase, alkali sulfate and calcium sulfate exhibit greater surface area relative to the bulk. These relationships are the subject of additional study in a project to characterize the CCRL cements. The ability to better describe cement compositional and textural characteristics promises to make cements a more predictable material. Many of the cements analyzed are available online through VCCTL at http:// vcctl.cbt.nist.gov/ .

Fig. 9. Comparison of mass fraction estimates for CCRL 142 based upon bulk SEM imaging, quantitative X-ray powder diffraction, and surface area fraction by SEM.

Fig. 10. Surface area to mass fraction ratio plot for averaged values of the six fields shows that the silicates have less surface area relative to the interstitial phases for this cement.


Next: Conclusions Up: Main Previous: Stereological analysis