Re-printed from

Cement Analysis –
A Time Critical Solution
Neal Robson,
Oxford Instruments, UK,
introduces cost-effective,
benchtop XRF analysis.

I
n today’s cement manufacturing process the critical
issues facing production are to increase output and
minimise costs while improving quality. From the
quarrying of raw materials to the shipping of finished
product, time-dependent analysis is required in order to
maintain integrity and smooth running. Ensuring that
chemical specifications are met at each key stage will result in
a consistent and predictable product, thus allowing time and
cost optimisation.
XRF analysis is a commonly accepted technique for the
analysis of the raw materials, intermediate and finished
products offering simplicity of sample preparation, speed
and accuracy of analysis.
Benchtop ‘next generation’ XRF instruments, such as
the X-Supreme8000, now provide a low cost solution for
continuous product quality. They offer high performance
analysis with field proven high reliability, and are often
operated by production staff on a 24/7 basis. These systems

Re-printed from [Jul 11] worldcement.com

 can be used both to measure the main elements (magnesium,
aluminium, silicon, calcium and iron) in cement making
materials at key points in the production process and as a
backup to another XRF system. Used either as a stand-alone
or as part of an automation system, XRF instruments
ensure conformance to any ISO specification that requires
compliance with 24/7 continuous product analysis.

XRF instrumentation
The cement industry accepts and respects X-ray spectrometry
as an analytical technique. Instruments fall into two broad
categories: Wavelength Dispersive (WDXRF) and benchtop
Energy Dispersive spectrometers (EDXRF).
Benchtop EDXRF systems are small, low cost, single
sample (e.g. Lab-X3500) or multi-sample (e.g. the
X-Supreme8000 shown in Figure 1) instruments that are
easy to install and are often situated in a variety of locations,
i.e. in a quarry, blending site, grinding plant, at-line, control
room or in the laboratory. They are normally operated by
production shift personnel giving a simple cost-effective
solution to time critical analysis.
Figure 1. Benchtop XRF analyser.
The latest generation of EDXRF instruments now offer
a field proven robust and reliable solution. They employ
matched X-ray tube and solid state high resolution, and
high count-rate detectors for the elemental determination of
low atomic number elements such as sodium in cement and
chlorine in clinker, as well as the standard range of cement
making elements, including Mg, Al, Si, S, K, Ca and Fe.

Principle of operation
XRF occurs when elements in a sample are excited at an
atomic level. As they return to their initial state they emit a
characteristic X-ray photon. Each element emits a unique
X-ray energy so an XRF spectrometer consists of a source of
excitation and selective detection and quantification of the
characteristic elemental X-ray photons. A calibration line is
then used to perform quantitative analysis.
Close attention to detail has produced EDXRF
instrumentation that ensures successful cement analysis.
Examples include the use of low power, high efficiency
Figure 2. Two superimposed X-ray spectra of cement samples X-ray tubes for elemental excitation. One advantage of this
with an X-ray Region of Interest (ROI) for Cl, covering the X-ray method is that X-ray tubes are only switched on when taking
energy range 2.55 – 2.71 keV. a measurement that leads to minimal heat generation with
no requirement for external cooling. This results in highly
stable, reliable X-ray tubes with long lifetimes, typically
greater than ten years. The low running and maintenance
costs are another benefit, giving an overall low cost of
ownership.
Detection of the characteristic X-rays is accomplished
using the latest detection technology: a Silicon Drift Detector
(SDD) providing high spectral resolution. The combination of
selective elemental excitation with optimum detection results
in optimal speed of analysis and stable and reliable results.
Since dust ingress is a major factor in the lifespan of
instrumentation, the design of the X-Supreme means that
cooling air is circulated in a separate compartment – a ‘wind
tunnel’ – from the main spectrometer components. This
feature, combined with the integral industrial grade PC,
which does not use fans, ensures a long operating life for the
main hard disk and spectrometer.
An example of the X-ray spectrum that results from
Figure 3. Standard 40 mm sample holder with cement pellet in using this combination to measure Cl in clinker is shown
place. in Figure 2. In this example two spectra are displayed,

Re-printed from worldcement.com [Jul 11]

representing 0.014% Cl (red spectra) and 0.51% Cl (yellow of Cl in cement, an X-ray ‘Region of Interest’ (ROI) is
spectra) in cement. This example demonstrates that a good defined around the Cl peak and the resulting X-ray intensity
elemental separation from adjacent elements in the periodic is obtained for each calibration standard measured. The
table, in this case S and Cl, can now be achieved, leading to calibration (after X-ray corrections have been applied) is
low detection limits (Table 1). shown in Figure 5. For a comprehensive cement analysis,
the data in Table 2 shows the performance for the remaining
Sample preparation cement elements. In addition, from the X-Supreme8000’s
In routine production control, samples are often measured as instrument repeatability, it can be seen to comply with
pressed pellets: the powder sample is first ground in a swing the precision requirements of the ASTM C114: ‘Standard
mill to ensure a consistent particle size, and then compressed methods of chemical analysis of hydraulic cement’.
using a hydraulic press at 20 t to produce a strong pellet.
A 40 mm sample holder then holds the pressed pellet Quality control
(Figure 3). One strategy that ensures the spectrometer remains within
In some cases an automatic sample preparation system specifications over the long-term is regular measurement of
is used where the pressed pellet is often retained in a steel
ring (Polysius type) measuring 51.5 mm dia. (2 in.). In this
instance, a special magnetic holder is used to load the sample
onto the sample carousel of the instrument (Figure 4). In
both examples the sample is rotated during the measurement
in order to minimise any sample inhomogeneity.
This method is particularly effective during periods
when the main XRF is unavailable, thus allowing analysis to
continue. This has major benefits, as it not only prevents any
interruptions to the production process, but it also means
that ISO criteria on continuous testing will be met.

Performance and results

For high quality production control the instrument is first Figure 4. Pressed cement pellets in 51.5 mm (2 in.) steel rings are
calibrated, which involves measuring a range of well analysed loaded into the magnetic holder and placed onto the 10-position
samples of known elemental content. In the determination sample carousel of the X-Supreme8000 instrument.

Table 1. Results for the calibration for Cl, shown in Figure 4

Analyte Range (% m/m) Standard error of Guaranteed limit Mid-range precision Measurement time
calibration (% m/m) of detection (3σ) (95% confidence) (seconds)*
(% m/m) (% m/m)

Cl 0.014 – 0.51 0.006 0.005 0.0035 200

* The measurement time is a simultaneous time measuring Na, Mg, Al, Si, S and Cl.

Table 2. Typical calibration performance for full cement analysis (XSMET-03B)

Analyte Range (% m/m) Standard error of Guaranteed limit Mid-range precision Measurement time
calibration (% m/m) of detection (3σ) (95% confidence) (minutes)
(% m/m) (% m/m)
Na2O 0.02 – 1.07 0.04 0.021 0.012
MgO 0.81 – 4.48 0.06 0.015 0.03
Al2O3 3.9 – 7.1 0.1 n/a 0.03
SiO2 18.6 – 22.4 0.2 n/a 0.07
P2O5 0.02 – 0.31 0.009 0.005 0.003
SO3 2.1 – 4.6 0.1 n/a 0.011
K2O 0.09 – 1.23 0.04 0.005 0.011
~7
CaO 57.6 – 67.9 0.5 n/a 0.08
TiO2 0.08 – 0.37 0.006 0.003 0.004
Cr2O3 0.002 – 0.06 0.003 0.001 0.001
Mn2O3 0.007 – 0.26 0.006 0.001 0.002
Fe2O3 0.15 – 3.1 0.06 0.003 0.008
ZnO 0.001 – 0.11 0.001 0.0006 0.001
SrO 0.02 – 0.64 0.004 0.002 0.001
The precision was calculated from 10 repeat measurements of NIST standards. The standards were chosen so that the analytes’ concentration
matches the calibration mid-range.

Re-printed from [Jul 11] worldcement.com

 measurement. This gives a rapid assessment of elemental
content, allowing confirmation of product quality.
If multi-samples are being measured, the red/yellow/
green labels for each sample show those samples already
measured and those awaiting analysis.
At the end of each measurement, customer-specified
‘cement moduli calculations’ are also performed on the
final result. Examples include lime saturation factor (LSF),
silica ratio (SR) and alumina ratio (AR), and these are
calculated and displayed at the end of each measurement,
allowing optimisation of the production process
(Figure 6).

Additional software
In addition to the elemental concentrations and cement
moduli, a software program called SMARTCHECK allows
Figure 5. Calibration regression for Cl in cement. one or a series of ‘logic calculations’ to be performed on
the final results. The output from the program can be
instructions to the production staff on optimisation of
the production process. The program can be applied to
the value of the LSF. As an example, the specification for
clinker could be 90 – 92, so the manager would specify a
SMARTCHECK calculation to ensure that every clinker
sample measured has an LSF between these values.
If the production is within tolerance then no warning
message will be displayed. However, if a result is outside
the tolerance, then a customer-defined warning message
can be displayed alerting the production operator to
a potential problem. These checks can be applied to
individual elements, ratios and calculations, etc. In
all cases, this level of checking on key cement process
parameters ensures that continued high quality analysis is
carried out by production staff.

Data storage and output

Once the results are accepted, they can be printed and/
or stored as specified by the manager in the analytical
Figure 6. Results and cement calculations displayed on the
X-Supreme8000. method. Results can be accessed remotely by other
PC-based systems for subsequent data storage or as part of
a centralised quality control system.
a quality control (QC) sample. If the QC sample results are
within customer-defined specifications, no further action is Remote diagnostics
required. Correction (restandardisation) is only required if In addition to the inherent high reliability of EDXRF,
the QC sample is outside specification. remote diagnostics can be performed, allowing an ongoing
The latest generation of EDXRF instruments have health check, using in-built samples located underneath
inherent stability, typically meaning that restandardisation the sample carousel. This aid to preventive maintenance
is required infrequently. When it is required, it involves ensures the instrument is continuing to perform to
the simple measurement of a few ‘Setting Up Samples’ specification. Any servicing requiring additional parts is
(SUS), which are provided with the instrument, and after then readily available to minimise downtime.
measurement redefines the instrument performance.
A full new calibration is typically only performed every Conclusion
12 – 18 months. This is because an ISO procedure specifies a Cost-effective XRF analysis with field proven reliability
maximum time limit before a full calibration is accomplished, can now be achieved with instruments located on or
not because the SUS cannot compensate for any slight near a production site, and operated by production staff
change in long-term instrument performance. giving 24/7 analysis coverage, ensuring consistent product
quality.
Routine analysis
On a 24/7 basis, production staff can load from one to Bibliography
ten samples onto the sample carousel; enter the sample • BERTIN, E.P., Introduction to X-Ray Spectrometric Analysis, (New York and
data either by mouse, bar code reader, or touch screen; London, 1978).
and the measurement starts. For a single measurement • HORNUD, I., Anwendertreffen Rontgenfluoreszenz und
Fukenemissionsspektrometrie, (Dortmund, 1995).
(in simultaneous analysis mode), provisional results
• PRICE, B.J. and BRADY, M.A., ‘The rapid analysis of cement to
on key elements are displayed after just 5 seconds of ASTM:C-114’, WORLD CEMENT, Vol. 23, No. 8, pp. 24 - 26, (August 1992).

Re-printed from worldcement.com [Jul 11]