Directly after its introduction in 1989, UniQuant™
has given a boost to the analytical power of sequential X-Ray
Fluorescence (XRF) spectrometers. In 2008, UniQuant is introduced
on an energy dispersive instrument the Quant'X.
Notably, XRF spectrometry became widely accepted for the analysis
of polymers, oil, catalysts, environmental and forensic samples.
In many research laboratories, UniQuant made the XRF spectrometer
to a more versatile tool.
There are now well over 2500 XRF laboratories using UniQuant.
Position of UniQuant in Materials Analysis
UniQuant represents a completely new
method in XRF analysis and is a commercially available
Product. UniQuant is for standardless semi-quantitative
to quantitative XRF analysis using the intensities
measured by a sequential X-ray spectrometer.
As its name suggests, it Unifies all types of samples
into one and the same analytical program and it is Unique
in that respect. UniQuant is highly effective for analysing
samples for which no standards are available.
Sample preparation is usually minimal or not required at
all. Samples may be of very different natures, sizes and
shapes. Elements from Be up to Am, (or their oxide compounds)
are analysed in samples like a piece of glass, a screw,
metal drillings, lubricating oil, loose fly ash powder,
polymers, phosphoric acid, thin layers on a substrate, soil,
paint, the year rings of trees and in general those samples
for which no standards are available.
The reporting is in weight% along with an estimated error
for each element.
Why to use UniQuant
general terms, UniQuant is used to provide:
quantitative analysis if no standards are available.
quantitative analysis with highest accuracy if standards
About 50% of the current UniQuant users
(about 500) only use UniQuant and do
not use conventional XRF analysis (calibration
by regression). Part of them had
little choice because no standards are available
(Waste materials, Polymers). But
there is also a tendency by many users to
replace the conventional method by
UniQuant using standards to firm it up for ‘families’of
of the % Sulphur present as Sulphide (reported as %S)
the % Sulphur present as Sulphate
(reported as %Sx). For WD only.
of the % Phosphor present as Phosphide (reported as
the % Phosphor present as Phosphate
(reported as %Px). For WD only
analysis of small and/or odd shaped samples.
analysis of a thin composite layer, along with the
mass / area.
The layer may be on a substrate containing
some elements that are also in the layer.
Or, the layer may be on a ‘neutral’ substrate,
like with dust on a filter.
determination of the masses of layers in a multi-layer
structure, usually on a substrate.
samples and detection of unexpected elements.
UniQuant supports 79 elements.
Contrary to the general belief,Wavelength
Dispersive XRF spectrometry +
UniQuant is much faster than Energy Dispersive
XRF spectrometry. The intensities
per analytical line are two orders of magnitude
higher with WD-XRF which also
has far less problems with spectral interferences.
fast pre-analysis of totally unknown samples, prior
to decide on further analyses
chemical analysis to support phase analysis by X-ray diffraction.
The following table shows 79 elements
which may be analysed by UniQuant.
Colored : Analysed Elements (except H and Li, which
may occur as absorbants)
Grey : Not known by UniQuant
Argon is included because it may be found in materials made
under an Argon atmosphere.
The unknown samples may take a great variety of
physical forms such as:
- a solid disk of metal or a synthetic material
- a multi-element mono-layer on a substrate
- a mono-element multi-layer on a substrate
- a small piece of solid sample placed on a supporting film
- a pressed powder that may include a binder
- a very small amount of powder on a supporting film
- a solid solution of a mineral (a glass bead)
- a liquid sample from a small drop to a full cup
- a filter aerosol sample
A totally unknown sample may be measured by the
full set of prescribed measuring channels (about 115 spectral
positions) including one or two lines of up to 79 elements
(Be to Am). The spectrometer time then is about 20
Accuracy for Majors and Minors
Usually, the ultra-light elements are only included in special
Without the ultra-light elements, the spectrometer time is
about 12 minutes.
Samples belonging to a known family may be measured by a smaller
sub set of the full set of measuring channels. For
example, the analysis of routine waste disposal samples may
be limited to say 55 measuring channels with extra long measuring
times for relevant traces. The spectrometer time can then
be as low as 5 minutes.
UniQuant is intended to cover the widest possible
concentration ranges while using one single set of calibration
data. Here we are not thinking about a wide range of alloys
or of oxide samples.The range that we mean includes samples
like oils, polymers, beads, thin layers and all types of alloys!
Errors are smallest for thick full-area homogeneous samples
and are quite acceptable for less favorable physical conditions.
For specific applications, where very high accuracy is required,
UniQuant may use specially calibrated data sets, for example
one for Alloys and one for Beads or Glass. Then international
or own standards are used to firm up the calibration. This
way of working may lead to the same high accuracy as with
conventional analysis using regression analysis of standards.
Although using specialised data sets has not been the primary
philosophy behind UniQuant, its application allows replacing
conventional methods by the UniQuant method with far less
specialised analytical programs. Several UniQuant users have
indeed done so.
We here introduce ‘Reliability’. An analysis is not Reliable
if once and again one or more concentrations are completely
out of range. We have seen this happening with semi-quantitative
methods (based on scanning). In one such case, 23 %Fe was
calculated whilst in fact there was less then 100 ppm Fe in
The sample contained a high % of Lead and the 2nd order of
one of the Pb lines strongly overlapped FeKa. The program
then cleverly diverted to FeKb which however was not well
calibrated by the standards in the ‘library’. Hence, the gross
error for Fe resulted. UniQuant can hardly make such enormous
errors. UniQuant gives reliable (plausible) results.
Precision (reproducibility) of the analysis of a given
sample specimen depends only on counting statistics. For each
analysed element, UniQuant reports the Standard Deviation
(Sigma) in ppm.
For large (full area) samples which are not or not highly
diluted, the Sigma’s are surprisingly small. For example 1
or 2 ppm for measuring times of 4 or 10 seconds per analytical
line. The Sigma is smallest with lighter matrix samples, for
higher atomic numbers and with longer measuring times.
Trace elements (with Z>20) in heavier matrices can be well
determined from 20 ppm onward. For light matrix samples like
polymers, this value is 5 ppm or even lower accuracy.The
accuracy for traces is depending on the quality of the corrections
well done by UniQuant.
- spectrometer’s spectral impurities
- spectral line overlaps
solved by UniQuant. Very
- matrix effects
solved by FP (Fundamental
has special ways to compensate for
certain physical effects.
Thin Layer Samples
UniQuant can calculate the sample mass along with its associated
standard deviation. At the same time the composition of the
layer is calculated. If the layer is on a substrate with elements
that are also in the layer, UniQuant can take their effect
Multi-layer Samples UniQuant may calculate the masses
of the layers in a multi-layer structure.
UniQuant has been designed for a maximum of interactivity.
A pre-condition is speed of entering data and speed of calculation.
The user interface has been designed for an absolute minimum
of key strokes or mouse operations.The need for fast interactivity
is illustrated by the following example:
A totally unknown powder sample is evaluated by UniQuant in
a first calculation (5 seconds) for which the analyst assumes
that a mineral sample consists of oxides. The results however
may show a very high content of Sulphur. The analyst concludes
that the sample is a Sulphide ore. Elements like Pb, Zn, Fe,
Mo are as sulphide whilst elements like Si and Al occur as
oxides. The original assumption would assume most
elements to be as oxides, even Sulphur. The sum of concentrations
would end up at a level higher then 100%. Now, the analyst
just changes Oxides to Sulphides in the General Data and starts
a second calculation. All this is a matter of a few seconds
In order to short-cut a lot of work at the PC,
Batch Modes are provided. These may be used for evaluation
of calibration (set-up) samples as well as for evaluating
a suite of unknown samples. Samples are ‘tagged’ in a directory
and the ‘process’ is started.