Particle size analysis

Particle Size Analysis - Evaluating

Laser Diffraction Systems in the
light of ISO 13320-1 – Part 1
With the publication of the International standard IS0/FIDS 13320-1, Particle size analysis – laser diffraction
methods – Part 1: General Principles, it becomes possible for users and potential users of such equipment to
evaluate them within a scientific frame of reference.

INTRODUCTION can see, each approach will yield a different value and none has
This article does not consider all aspects and implications of any greater claim to being more valid than the next.
the standard but concentrates on the parameters most widely So which is the correct result? Given the example above
employed when evaluating such systems. The reader is directed there is not a complete solution to the problem. In the
to review all details of the standard for a complete circumstances outlined the most appropriate answer is the one
understanding. The parameters discussed in this article are as that gives the most relevant information to the analyst, so that
follows: he may best control his process.
• Repeatability With the release of ISO 13320-1 the analyst can reference
• Accuracy results against an authoritative source for a better
• Resolution & Sensitivity understanding of the data obtained.

Part 1 will cover Repeatability and Accuracy, Part 2, to be REPEATABILITY

published in February/March will cover Resolution and There are two forms of repeatability that need to be considered;
Sensitivity. firstly within instrument, i.e. the repeatability of different
analyses of the same sample on the same instrument, and
BACKGROUND secondly instrument to instrument, i.e. the repeatability of
Since the late 1970’s the technique of laser diffraction has been different analyses of the same sample on different instruments.
utilised to calculate a particle size distribution. The technique Whilst taken in isolation within instrument repeatability is
relies upon the fact that particles will act as scattering centres obviously important, of more significance is instrument to
when exposed to electromagnetic radiation. The resultant instrument repeatability. This for example could be because of
scatter pattern can be measured electronically and then the need to compare data between different sites. Many multi-
deconvulated mathematically to infer a particle size national companies will possess a number of instruments
distribution. This is a simplified account of the processes throughout the world; they will need to be assured that results
involved yet gives the most salient features of the technique. If on one unit can be reproduced on another system.
you wish for a more detailed account of Laser Diffraction But how does one assess what is acceptable in terms of
Analysis please contact the author in the first instance. repeatability? ISO 13320-1 has published the following values
Particle size analysis, what ever form it takes, from the very as being acceptable for repeatability:
simple sieving, a direct measurement, to the most complex • Samples with a mean size greater than 10µm.
form of light scattering experiments which are non-direct • Median Size (x50) should have a coefficient of variation of
suffer to some degree in the perceived validity of the result less than 3%. This could also be applied to any chosen
produced. This does not mean that the results reported are central value of the distribution.
scientifically flawed, just that they are often misunderstood or • Values at the sides of the distribution e.g. x10 and x90
misinterpreted. should have coefficient of variation not exceeding 5%.
Why should this be? On one level it is due to the inherent • Samples with a mean size less than 10µm.
difficulty encountered when measuring materials where the • For samples below the 10µm range these values can be
particle shape deviates from that of a perfect sphere with doubled.
known physical properties. For example how does one assign a
single size to an irregular shaped particle? . One person may
give the size as the longest axis, particularly if one was using a PROFICIENCY TESTING
microscope or 2-Dimensional image analysis system, another The Laser Diffraction Proficiency Testing Scheme, essentially
may give an average value based on the sizes of the different an inter-laboratory testing regime, founded and administered
axial lengths, a third may measure the volume of the particle by Beckman Coulter UK Ltd on behalf of an independent
and calculate an Equivalent Spherical Diameter (ESD) . As you steering committee uses the values stated in the standard for

50 International Ceramics • Issue 2 • 2000

 Particle size analysis

evaluation purposes. Whilst a detailed a discussion of the alignment.

results from the scheme are beyond the scope of this article it is The change in distribution is quite clear to see and this was
fair to say that the degree of freedom allowed for by the caused with only a minor misalignment of the system. This
coefficient of variation values has led to excellent results when clearly illustrates the importance of having a robust alignment
laboratories have followed the analytical protocols supplied. If procedure that is operator independent.
anyone wishes for further information regarding this scheme Having discussed the importance of alignment how can
please contact the author. we verify that such procedures allow us to recover accurate
Fundamental to the success of establishing repeatability data. Remember that as the technique is constrained to equate
will be the need to ensure that the sample being evaluated can and measure spherical systems (non-spherical systems will be
be sampled consistently to produce homogeneous aliquots and represented as an equivalent spherical diameter) it is
when added to the sample module produce a stable dispersion. important that we use spherical materials to evaluate accuracy.
To this end a sample that is representative of the typical In addition we must be able to accurately input the correct
material that will normally be analysed is an ideal candidate as optical constants i.e. the real and imaginary part of the
a control sample. complex refractive index, if the Mie theory (see below –
It is also important that different sizes of materials are adequacy of the optical model) is to be applied in the data
evaluated to ensure consistency over the entire dynamic analysis.
measuring range. The type of materials for the evaluation of accuracy is
Manufacturers should also be able to provide test data to entirely at the discretion of the instrument operator, but the
prove repeatability. For example reproduced below are two following should be borne in mind as outlined in the standard.
examples of test data showing the repeatability of the Beckman The certified or reference material should consist of a known
Coulter LS‘ 230. distribution having a range of spherical particles over one
decade of size. They should be certified to a mass percentage by
NOMINAL 570µM SAMPLE; MEAN DIAMETER an absolute technique, if available, and used in conjunction
We have for this sample a 0.46% coefficient of variation on 120 with a detailed protocol.
units. The response of the laser diffraction instrument is
considered to meet this standard if the mean value of the x50
NOMINAL 0.305µM SAMPLE; MEAN DIAMETER coming from at least three independent measurements deviates
This sample shows a coefficient of variation of 0.43% on 106 less than 3% from then certified range of values specified for
units. the material, i.e. the mean value together with its standard
In the field we see typical values of repeatability of less than deviation; the mean values for the x10 and x90 should deviate
1%. Indeed this is borne out by the Proficiency testing Scheme. less than 5% from the certified range of values.
Good quality control procedures will call for the analysis on a The following distributions show examples of recovered
regular basis of a control material, this should be available in mean size values from a number of certified reference
sufficient quantity to ensure measurements can be made over materials, polystyrene latex (Duke Scientific, Palo Alto, CA,
an extended period of time. USA) from an LS‘ 230.

ACCURACY ATTRIBUTED MEAN SIZE VALUES

Because laser diffraction systems are based on first principles, (nominal size) Assayed values LS™ 230
albeit with constrained ideals, calibration in the strict sense is 100nm 102nm ± 3nm 108nm
not required. However, as with any optical system, alignment 150nm 155nm ± 4nm 151nm
will be required. Thus it is very important to have some easy 220nm 220nm ± 6nm 211nm
and reproducible way of ensuring alignment. On the Beckman 300nm 300nm ± 5nm 306nm
Coulter LS‘ 230 this is provided by means of a reticule. This 404nm 404nm ± 6nm 400nm
reticule, essentially a metal plate with a precision drilled hole,
acts as an ideal particle. The scattering produced by this ‘ideal This approach allows one too effectively evaluate the
particle’ is used to align the laser onto the detector array. This instrument for accuracy using the coefficient of Variation
entire procedure is fully automatic and requires no operator values referred to in the standard. ■
intervention, a crucial element to ensure reproducibility in the
alignment method. Ian Ley,
This alignment procedure is particularly crucial for large European Marketing Manager,
particles, this is because they scatter with characteristic, Beckman Coulter UK Ltd,
discrete, well defined minima at low angles in relation to the Oakley Court,
optical path (see resolution & sensitivity section for more Kingsmead Business Park,
detail). Without this ‘angular calibration’ the recovery of an London Road,
accurate distribution would be impossible. Reproduced below High Wycombe,
is an example of the error caused by misalignment (created HP11 1JU,
deliberately). Glass beads, nominal size 560µm, have been Bucks,
analysed and acquired with a system in alignment and out of UK.

International Ceramics • Issue 2 • 2000 51