X-ray fluorescence spectroscopy is an established technique used in cement plants worldwide. The technique is suitable for quality control all through the cement production process. Energy dispersive X-ray fluorescence (EDXRF) spectrometry is a routinely employed screening and quality control tool used to ensure the right composition of incoming feedstocks, adding gypsum, raw meal mixture balances throughout the manufacturing process. EDXRF analyzers are also commonly deployed as backup instruments for WDXRF spectrometers used for final QC and certification. The empirical approach is used for the analysis of finished Portland cement.
Figure 1.
Instrumentation
Model: |
Rigaku NEX QC+ |
X-ray tube: |
50kV 4W Ag-anode |
Detector: |
SDD |
Sample Type: |
Portland Hydraulic Cement |
Film: |
Prolene |
Analysis Time: |
200 sec |
Environment: |
Helium |
Standard: |
Single position |
Optional: |
Autosampler |
Figure 2.
Sample Preparation
Each sample is prepared by the grinding to a fine, dry, homogeneous powder of less than 200 mesh (less than 75µm particle size) using a ball mill or ring-and-puck shatter box. A sample is prepared by weighing 5 g of sample and making a hydraulically pressed pellet using 20 t pressure for 30 s for measurement.
Calibration – Single Position
Empirical calibrations were built using a set of 8 NIST SRM certified standards: 1880b, 1881a, 1884b, 1885a, 1886a, 1887a, 1888a, 1889a. With the empirical approach, "alpha corrections" are then employed to automatically compensate for changes in X-ray absorption and improvement effects within the sample due to the independent variations in element concentration, thus offering a very accurate model characterizing the cement type.
The single position window ring was used providing the right sensitivity for sodium, magnesium, aluminium and silicon. The 6-position autosampler can be used for its ease of use, enabling multiple calibration or unknown samples to be measured without an operator attending the analyzer. A small change in sensitivity and performance may be seen because of a slight difference in sample height.
A summary of a typical finished Portland cement empirical calibration is shown in Table 1
Table 1.
Component |
Concentration Range (mass%) |
RMS Deviation |
R2 Confidence |
SiO2 |
18.637 – 22.380 |
0.098 |
0.9943 |
Al2O3 |
3.875 – 7.060 |
0.021 |
0.9997 |
Fe2O3 |
0.152 – 3.681 |
0.013 |
0.9999 |
CaO |
57.58 – 67.87 |
0.032 |
0.9999 |
MgO |
0.814 – 4.740 |
0.018 |
0.9998 |
SO3 |
2.086 – 4.622 |
0.046 |
0.9972 |
Na2O |
0.091 – 1.068 |
0.092 |
0.9372 |
Repeatability – Single Position
NIST SRM 1887a was selected from a set of calibration standards to demonstrate repeatability. The measurement of the sample was done in static position for ten repeat analyses using a total analysis time of 200 sec per measurement, with typical results shown in Table 2.
Table 2.
NIST SRM 1887a |
SiO2 |
Al2O3 |
Fe2O3 |
CaO |
MgO |
SO3 |
Na2O |
Standard Value |
18.64 |
6.20 |
2.86 |
60.90 |
2.84 |
4.62 |
0.48 |
Average Value |
18.78 |
6.36 |
2.81 |
61.90 |
2.99 |
4.65 |
0.68 |
Standard Deviation |
0.03 |
0.03 |
0.01 |
0.07 |
0.04 |
0.02 |
0.05 |
For clinker, finished cement and raw meal, separate calibrations must be used. Performance for Raw Meal and Clinker is comparable to the performance for finished cement shown here.
Typical Detection Limits – Single Position
Ten repeat analyses of blank samples were measured and standard deviation calculated in order to determine the lower limit of detection (LLD) using the empirical method, ten repeat analyses of blank sample were measured and the standard deviation calculated. The LLD is then defined as three times the standard deviation. Actual detection limits may change based on analysis time used, combinations of elements present and elemental concentration levels. To simulate detection limits in a high calcium matrix, CaO was chosen as the "blank" material. The detection limits are shown in Table 3 below:
Component |
Lower Limit of Detection |
SiO2 |
0.01 |
Al2O3 |
0.02 |
Fe2O3 |
0.01 |
MgO |
0.11 |
SO3 |
0.05 |
Na2O |
0.21 |
Reporting Results
Analysis results are reported on the main screen and printed. One only needs to scroll down to see the silicon ratio, aluminium ratio and lime saturation factor. Standard formulas for LSF, SR and AR are used.
LSF Finished = (CaO – 0.7SO3)/(2.8SiO2 + 1.2Al2O3 + 0.65Fe2O3)
Clinker SLF = CaO/(2.8SiO2 + 1.2Al2O3 + 0.65Fe2O3)
SR = SiO2/(Al2O3 + Fe2O3)
AR = Al2O3/Fe2O3
Figure 3.
Figure 4.
Conclusion
During the processing cycle and the entire production, oxide composition of the cement material must be reliably monitored to ensure optimal process control, physical characteristics of the cement, as well as profitability. The Rigaku NEX QC+ gives the cement plant a consistent and low-cost system for quality control measurements around the plant, making it an ideal tool throughout the quality control process and as a backup to WDXRF systems. Similar performance as shown here is also applicable to raw meal materials and clinker, and can be used simply to measure gypsum (SO3) in finished cement.
This information has been sourced, reviewed and adapted from materials provided by Rigaku.
For more information on this source, please visit Rigaku.