Particle Size

Analysis
Overview
White paper
INTRODUCTION A description of various ESDs reported by techniques
— discussed in this document are provided below.

What is a particle? The simplest definition is a minute • DLS: hydrodynamic diameter is the size of a hard
portion of matter. Within the scope of this document sphere that would diffuse at the same rate as the
particles do not include subatomic particles such as measured particles
electrons, protons, neutrons, quarks, etc. Particles
measured by the techniques described here include • Laser diffraction: spheres of this size would
solids (powders), solids in liquids (suspensions), and scatter light in the detected pattern
liquid/liquid emulsions. Not all particles exist as individ-
• Coulter counter: spheres of this size would
ual entities. They have a habit of sticking together to
displace an equivalent volume of electrolyte
form various kinds of clusters, or agglomerates. In
while passing through the orifice
the field of particle technology, we typically define
agglomerates as a loose arrangement of larger • Optical particle counter: spheres of this diameter
structures while aggregates are denser, harder to would obscure/scatter light in the same amounts
disperse collectives. as those detected

Most particles are not ideal spheres, but irregular in To complicate matters several of these techniques
shape. This creates a quandary when defining the also depend on the optical properties such as refractive
size of a particle using a single descriptive value. The index (RI) of both the particles and the dispersing
diameter of some kind of equivalent sphere is the only medium. Add the effects of proprietary designs and
available approach to describe particle size using a algorithms within instruments, sample preparation,
single number. The International Union of Pure and multiple reporting formats, etc., and one begins to
Applied Chemistry (IUPAC) definition1 of the equiva- understand the difficulties of comparing results
lent diameter of a non-spherical particle is equal to a generated by different analyzers.
diameter of a spherical particle that exhibits identical
Different analytical techniques calculate size distribu-
properties (e.g., aerodynamic, hydrodynamic, optical,
tions based on various calculation models. Dynamic
electrical) to that of the investigated non-spherical
light scattering (DLS) reports results based on the
particle. Most particle sizing techniques report results
scattering intensity of the particles – the intensity
as an equivalent spherical diameter (ESD). Figure 1
distribution. Particle counters or any one-at-a-time
shows a particle the shape of California with the ESD
method report the primary result as a number distribu-
and other possible calculated diameters that could
tion. Laser diffraction reports results as a volume-based
define the size.
distribution. The diagram in Figure 2 helps explain the
difference between volume and number distributions.
Longest diameter
Number
1.2
Shortest diameter 1.0
0.8
0.6
Equivalent 0.4
spherical diameter 10 µm 20 µm 30 µm
0.2
0
Three particles: 0 10 20 30 40 50
10, 20 & 30 µm
Sphere with the same Volume
area as California Number: 0.8
equal portions 0.7
Volume: 0.6
Feret diameter max weighted to largest size 0.5
0.4
Number mean = 20 0.3
0.2
Volume mean = 27
Figure 1. Particle the shape of California with several possible size 0.1
0
diameters. Source: Entegris 0 10 20 30 40 50

Figure 2. Number vs. volume distributions. Source: Entegris

2
Particle counters generate the most accurate and Mean = Median = Mode
high-resolution distribution results if enough particles
have been analyzed.2 This is a “model independent”
result since each point of the distribution comes from 68% within
actual measurements. For this reason, converting a 1 standard
deviation (σ)
number to volume distribution is also accurate 95% within
2 standard
without unknown errors. deviation (σ)

On the other hand, ensemble light scattering tech- -2 σ -1 σ Mean +1 σ +2 σ

niques such as DLS or laser diffraction generate model Figure 3. Gaussian distribution. Source: Entegris
dependent results using algorithms and smoothing.
Converting intensity or volume to number distribu-
Mode
tions is much less accurate and only suggested for
comparison purposes. Mean

Differential Volume (%)

Median
Many numerical descriptors can be used to describe
the properties of particle size distributions. The term
average particle size is not typically used in this field
of technology. The mean is a calculated central point
50%
in the distribution. There are various calculated means above
50%
that can be defined 3 including, but not limited to: below
10% 90%
• Intensity mean: the most important below below
D10 D50 D90
DLS result calculation 0.1 1.0 10.0 100.0
Particle Size (µm)
• Number mean: mean based on the
Figure 4. Asymmetric distribution. Source: Entegris
number distribution

• Volume mean: mean based on the

number distribution
PARTICLE CHARACTERIZATION TECHNIQUES
Other central points of the distribution used in ­—
common practice include the mode (the highest
There are many analytical techniques available for
point of the frequency distribution) and the median
particle size analysis. Various techniques have different
(50% above and 50% below this diameter). Also in
capabilities and dynamic range as well as assorted
common usage are the diameters shown in Figure 4:
strengths and weaknesses. Before choosing an analytical
• D10: 10% of the distribution lies below this diameter technique, the user should first define what is required
by the instrument and how the data will be used. After
• D50: the median diameter, 50% above, 50% below this is determined it can still be challenging to review
all possible techniques and select the best option for
• D90: 90% of the distribution lies below this diameter
a given requirement. This document is intended to
For a symmetric distribution, such as a Gaussian help scientists new to particle size analysis understand
distribution (Figure 3), the mean, median, and mode some of the available techniques and how to select
are all the same value. As seen in Figure 4, the mean, the proper technique for their samples.
median, and mode can be quite different for an
asymmetric distribution.

3
Questions to be considered when choosing a particle LIGHT SCATTERING TECHNIQUES
size analyzer include: —
• What is the size range of interest? Two common particle size analysis techniques are
dynamic light scattering and laser diffraction. Both
• Is the particle size distribution the only
can be called “ensemble light scattering techniques”
result required?
because they collect the scattered light from all the
• Does the surface charge matter? particles within the measurement zone and then
apply an algorithm to calculate the final result.
• Is it helpful to know the concentration
of the particles? Dynamic Light Scattering
• How much automation is required? Dynamic light scattering is best used for samples
where the mean size is below 1 µm. The upper size
• What format should the results be in?
limit for DLS depends on the sample density. For
– Number distribution? Volume distribution? example; the upper size limit for an emulsion could
be around 5 µm while the upper limit for a heavy
After these questions are understood the technique
metal particle suspension could be around 500 nm.
and then specific instrument selections will become
The lower size limit for DLS is around 1 nm depending
easier.
on the sample concentration and how strongly it
scatters light. DLS can measure samples at low con-
centration such as lysozyme protein at 0.1 mg/mL.
MICROSCOPIC TECHNIQUES The theoretical upper concentration limit is often
— specified quite high – around 40 volume percent,
but in practice lower concentration samples generate
Microscopic inspection is the most direct measurement better results.
for particle size analysis and is therefore considered
the referee technique. Microscopy also provides the The basic principle of DLS is based on the time
additional benefit of offering shape information. A signature of the scattering caused by the Brownian
sophisticated particle characterization lab usually motion of the particles. Smaller particles diffuse
includes a microscope both for the size and shape more quickly while larger particles diffuse more
data provided but also as a reality check when other slowly. For this reason, the fluctuations in light
techniques report conflicting results. Particle size and scattering due to particle diffusion from Brownian
shape analysis is frequently performed using image motion is size dependent. A typical DLS system
analysis software for automating data collection and detects the scattered light at 90o or at some
result computations. Automated image analysis systems other angle and feeds the signal into a digital
are broken into two categories: static image analysis autocorrelator (or correlator), see Figure 6.
where the particles are dispersed onto a slide prior to
Lens
analysis and dynamic image analysis where particle Sample
Laser Diode
images are captured while flowing.

Detector

5 nm AuNPs 20 nm AuNPs C(t)

Figure 5. Gold nanoparticles SEM images.

size t
Source: Entegris applications laboratory
Result Autocorrelator

Figure 6. DLS optical configuration. Source: Entegris

4
The correlator takes this information and determines The zeta potential is measured by applying an electric
the diffusion coefficient of the particles in the sample. field to the sample and then measuring the direction
The diffusion coefficient is then used to calculate the and speed of the particle motion, see Figure 8.
particle size using the Stokes-Einstein equation: Electric field
E
D = kT/6π ŋR
+ + +
+
Where: +
+
+
D = Diffusion coefficient FV = 6πŋv + net + FE = QE
+ + + –
R = Particle radius + +
+ + +
k = Boltzmann’s constant
T = Temperature Kelvin
ŋ= Shear viscosity of the solvent
Negatively charged V
perticle will move Velocity or mobility
The results from a DLS measurement are typically towards the anode
expressed using the intensity mean diameter and the Figure 8. Zeta potential measurement. Source: Entegris
polydispersity index (PI) to quantify the width of the
distribution. DLS systems report results using either The direction indicates if the particles are positively
a Gaussian distribution (Figure 3) for samples with or negatively charged. The speed of the particles
a single peak (mode) or proprietary multi-modal indicates the magnitude of the charge. The velocity
algorithms for samples with more than one peak of the particles can be measured using either phase
(Figures 9 and 16). analysis light scattering (PALS) or using frequency
analysis. PALS is the newer, more sensitive approach
Electrophoretic Light Scattering and is now the preferred method for most samples.

Many DLS systems are built to measure both the The zeta potential is a function of the specific surface
particle size and the zeta potential of the sample. chemistry in the condition as analyzed. Conditions that
The zeta potential is a measure of the charge on the can affect zeta potential include pH, salt concentration,
surface of a particle. The technical definition of the surfactant type, surfactant concentration, and other
zeta potential is the charge a short distance away chemical conditions. Therefore, many zeta potential
from the surface of the particle, see Figure 7. This results are reported including other conditions such as
distance is known as the slipping plane. Inside the pH. Note that zeta potential is only valid for suspensions
slipping plane the ions near the particle surface move – there is no zeta potential value for powders.
with the particle. Outside of the slipping plane the
ions remain randomly dispersed. Combining particle size by DLS and zeta potential
can be a powerful tool for formulators creating new
-
suspension products. The zeta potential is an indicator
+ - -
-
of dispersion stability. The magnitude (not the sign of
-
-
- -
-
the charge) of the zeta potential provides information
- -
+ + + on the electrostatic repulsion between the particles
- + + +
-
- or emulsion droplets in a suspension. A higher zeta
+ + +
- - potential value indicates the suspension should be
- + + -
- more stable than a value near zero. When the zeta
-
- + +
+ + Slipping plane potential is near zero the suspension is likely to de-
+ - -
+ + +
- + + stabilize; particles may agglomerate and settle, and
- -
-
-
-
emulsions may phase separate. An example of this is
-
+
- shown in Figure 9. A zinc oxide powder was dispersed
- Distance from surface
- in water in different surface chemistry conditions. The
blue result reports a single peak at pH = 9 where the
Stern layer
Zeta potential
zeta potential = -31 mV. The pink result reports a similar
potential in mV

first peak plus a peak of agglomerates at about 500 nm

at pH =11.2 where the zeta potential = 0.7 mV.

Figure 7. Zeta potential of negatively charged particle.

Source: Entegris

5
 Although laser diffraction is an older, established
Agglomerates technique several shortcomings make generating
Zeta potential = 0.7mV
accurate, reliable results challenging. As mentioned
above, selecting the proper RI values is a critical
Zeta potential = -31 mV requirement that is difficult to accomplish for many
samples. This is particularly true for active pharma-
ceutical ingredients (APIs) since the RI value is often
unknown. Size peaks can appear and disappear when
guessing improper RI values, making data interpretation
very difficult. Another problem with laser diffraction
analyzers is that since the algorithms are proprietary
Figure 9. ZnO dispersed in H2 O at pH 9 and pH 11.2.
Source: Entegris and unique for each vendor and model results from
one analyzer can be significantly different than results
from any other laser diffraction analyzer. Since laser
diffraction is an inherently low-resolution technique
LASER DIFFRACTION the results may underestimate or completely miss
— peaks separated from the main distribution (tails).
Laser diffraction is an ensemble light scattering
technique that can be used for many kinds of samples
including powders, suspensions, and sprays. The
PARTICLE COUNTING
dynamic range for laser diffraction varies by model —
but is typically from about 100 – 3000 µm. The basic
principle of laser diffraction is that smaller particles Coulter Counter
scatter at wider angles while larger particles scatter
The first particle counter/size analyzer is known as
at lower angles. As seen in Figure 10 3 the optics of a
electrical sensing zone or the Coulter counter. With
laser diffraction analyzer include one or more laser
this technique particles must be suspended in an
light sources, multiple detectors at a range of angles
electrolytic solution. As particles flow through an
(forward, side, and back scattering), and a sampler
aperture a change in impedance is proportional to
and cell to transport the sample through the optics.
the volume of the particle, Figure 11. The physics
The scattering from all the detectors is input into
of the aperture geometry and electronics limits the
proprietary algorithms based on either Fraunhofer
dynamic range to an approximately 30:1 ratio, for
or Mie light scattering theories. Mie theory provides
example a 140 µm aperture can measure from
more accurate results and can measure at smaller
2.8 – 84 µm.
sizes but requires refractive index (RI) values for the
sample. The RI values required include both the real
component (which can be determined for many
samples) and an imaginary component that models V
the light absorption. There are no good techniques
for measuring the imaginary component of sample,
so this often leads to difficulties in data interpretation
and inaccurate results.

8
8 9 11 Anode
10 –
7 Ɵ Ɵ Cathode
8 r +
/(r)
(/) Particle Aperture
1
/(Ɵ)
3 2
4 Electrolyte
6 5 Ɵ

Key
1 Obscuration detector 6 Particle ensemble Figure 11. Coulter counter principle.
2 Scatter beam 7 Light source laser Source: Entegris
3 Direct beam 8 Beam processing unit
4 Fourier lens 9 Working distance of lens 4 While the Coulter counter is still used in hematology,
5 Scattered light not 10 Multi-element detector
collected by lens 4 11 Focal distance of lens 4 it is now seldom used for general particle size analysis
due to the limited dynamic range and practical difficul-
Figure 10. Laser diffraction optics. ties with particles clogging the orifices.
Source: ISO 13320-1, from reference 4
6
Optical Particle Counters The increase in concentration range is important
because to avoid coincidence errors all counting
The original optical particle counting systems started
techniques can only have one particle at a time in
with the sensor based on light obscuration, or extinc-
the measurement zone. This error occurs when
tion. As particles pass through a flow cell they block
two particles pass through the measurement zone
(obscure) some of the incident light. The amount of
simultaneously, resulting in both a size error and a
light obscured is related to the particle size through a
count error – the two particles are counted and sized
calibration curve. Sensors based on light obscuration
as one larger particle. When measuring low concen-
only have a dynamic range of about 1.5 – 150 µm.
tration contamination samples this concern is not
Advances in sensor technology, sampling fluidics, and important. But when measuring higher concentration
counter electronics created the technique known as suspensions, the sample is often diluted to below the
single particle optical sizing (SPOS). The first sensor coincidence error level before the analysis begins.
advancement was to combine a scattering detector
and an extinction detector, widening the dynamic
range to 0.5 – 400 µm, see Figure 12.
LABORATORY VS. ONLINE ANALYSIS
LE400 sensor
0.5 – 400 µm
Scattering —
Particles detector VScattering

Laser Most particle size analysis measurements are per-

diode
VExtinction
formed in the laboratory. Some, but not all, particle
B sizing techniques can and have been adapted to
C A A = 1000 µm Extinction
B = 400 µm detector the process environment for online measurements.
Flow cell C = 35 µm
Compression Online image analysis can be used for both powders
∑
and liquids. DLS with automatic dilution has been
Summation
used to monitor size reduction operations such as
Two interactions
Apply calibration Pulse height added to form a
homogenization. Laser diffraction has been more
Result
curve analyzer composite pulse successfully adapted to free-flowing powders than
suspensions due to the challenges in keeping optical
Figure 12. SPOS sensor. Source: Entegris windows clean. Liquid particle counters are often used
online for contamination control of pure DI water
Pulses from the sensor are input into a pulse height
and chemical applications. Online SPOS systems
analyzer, or counter. Pulses are converted to a particle
are used in fabs around the world to monitor large
size using a calibration curve. The reported result is
particle counts (LPCs) in CMP slurries, Figure 14.
both the particle size distribution and the concentration
in particles/mL broken into up to 1024 size channels. Tote Mixer Day tank
A higher powered, focused laser beam (Figure 13) can
Filter Manifold
be used to extend the lower size limit down to 0.15 µm
and extend the concentration limit up to one million POU filter

particles per mL.

Flow direction
Analyzer Analyzer
Polishing
tools
Flow channel

Scattering Figure 14. Online SPOS system. Source: Entegris

Laser lens
diode
Light
scattering
Laser detector
beam

Particles in
flow channel

Particles in
Light obscuration flow channel
detector

Figure 13. Focused laser beam sensor. Source: Entegris

7
RESULT COMPARISONS
—
A 1 µm emulsion sample was analyzed using several
techniques including DLS, laser diffraction, and SPOS.
The size of 1 µm was chosen since this lies within the
dynamic range of all three technologies.

The DLS results using both the Gaussian unimodal

and Nicomp® multi-modal algorithm are shown in
Figures 15 and 16. Figure 17. 1 µm SPOS number distribution result.
Source: Entegris

The result shown in Figure 17 was then converted

to the volume distribution and is shown in Figure 18.
The conversion from number to volume distributions
with the AccuSizer system is accurate and justifiable.
Now the first peak remains at 1 µm and a larger peak
is reported at 6 µm.

Figure 15. 1 µm emulsion Gaussian DLS intensity result.

Source: Entegris

Number Area Volume

10 0.774 0.859 0.890

50 0.973 1.030 1.093

90 1.209 1.427 3.939

Figure 18. 1 µm SPOS volume distribution result. Source: Entegris

The volume distribution shows the 1 µm main peak

Figure 16. 1 µm emulsion multi-modal DLS intensity result. but also shows how larger droplets make up some
Source: Entegris
of the volume.
The Gaussian result indicates a high Chi Squared value
The same sample was analyzed using a Horiba LA-960
of 744.1, strongly suggesting multiple peaks and that
laser diffraction system. The result shown in Figure 19
the multi-modal algorithm result should be used for
reports the volume distribution based on an RI value
this data. The multi-modal algorithm result reports a
of the dispersant oil = 1.45.
first peak at 950 nm (0.95 µm) and a larger peak at
10433.4 (10.43 µm).

The same sample was analyzed using the AccuSizer®

AD SPOS system. This instrument uses exponential
dilution to reduce the concentration to the optimum
level to avoid coincidence error (more than one particle
at a time in the measurement zone). The LE400 sensor
has a dynamic range of 0.5 – 400 µm. The counts/mL,
or number distribution, is shown in Figure 17 where
the main peak is centered near 1 µm.

Figure 19. Laser diffraction result, RI = 1.45. Source: Entegris

8
This study highlights both the similarities and differ- ABOUT ENTEGRIS
ences between results generated by several particle —
size analysis techniques. None of the results match
Entegris is a leader in specialty chemicals and
each other perfectly. That is the real-world situation
advanced materials solutions for the microelectronics
when comparing results analyzed using different
industry and other high-tech industries. Entegris is
technologies. Each technique has advantages and
ISO 9001 certified and has manufacturing, customer
disadvantages. DLS is the most popular technique
service, and/or research facilities in the United States,
for sub-micron samples, but data interpretation can
China, France, Germany, Israel, Japan, Malaysia,
become challenging and not all multi-modal algo-
Singapore, South Korea, and Taiwan. Additional
rithms are created equally. The SPOS technique
information can be found at www.entegris.com
provides high-accuracy, high-resolution results and
has the best sensitivity to tails of distributions. But
References
SPOS is not well suited for powders or sprays. Laser
McNaught, A. and Wilkinson, A., Compendium of Chemical
1
diffraction provides quick, easy measurements across Terminology, Oxford, 1997
a wide dynamic range. But the lack of resolution
broadens the peaks; both at the small and large end
2
Entegris Technical Note – Number of Particles Analyzed

of the distribution as seen in Figure 19. The best 3

ISO 9276-2:2014, Representation of results of particle size
approach for most chemists is to analyze their analysis – Part 2: Calculation of average particle sizes/diameters
specific samples using a range of techniques and and moments from particle size distributions, available at
https://www.iso.org/standards.html
base their selection decision on which technique
generates the best results for their requirements. 4
From ISO 13320-1, Particle size analysis – Laser Diffraction
Methods, Part 1: General Principles

FOR MORE INFORMATION

Please call your Regional Customer Service Center today to learn what Entegris can do for you.
Visit entegris.com and select the Contact Us link to find the customer service center nearest you.

CONTENT AND LIABILITY DISCLAIMER

Entegris believes the information in this document is accurate as of its publication date. Any and all specifications and designs
are subject to change without notice. Entegris is not liable for errors or omissions in this document. Entegris undertakes no
obligation to update the information presented in this document. You may not use or facilitate the use of this document in
connection with any infringement or other legal analysis concerning Entegris products described herein. You agree to grant
Entegris a non-exclusive, royalty-free license to any patent claim thereafter drafted which includes subject matter disclosed
herein. No license, express or implied, by estoppel or otherwise, to any intellectual property rights is granted by this document.

Corporate Headquarters Customer Service

129 Concord Road Tel +1 952 556 4181
Billerica, MA 01821 Fax +1 952 556 8022
USA Toll Free 800 394 4083

Entegris®, the Entegris Rings Design®, and other product names are trademarks of Entegris, Inc. as listed
on entegris.com/trademarks. All third-party product names, logos, and company names are trademarks or
registered trademarks of their respective owners. Use of them does not imply any affiliation, sponsorship,
or endorsement by the trademark owner.
©2019 Entegris, Inc. | All rights reserved. | Printed in the USA | 7127-10785ENT-1019

www.entegris.com