CHAPTER 2

PARTICLE SIZE CHARACTERIZATION

2.1 Introduction to Particle Size and Shape
In many powder handling and processing operations,
particle size and size distribution play a key role in
determining the bulk properties of the powder.
Example of regular-shaped particles.
Shape
Sphere
Cube
Cylinder
Cuboid
Cone

Dimensions
Radius
Side length
Radius and height
Three side lengths
Radius and height

In practice, it is important to use the method of size

measurement, which directly gives the particle size, which is
relevant to the situation, or process of interest.
2.2 Methods of Measurement:
2.2.1

Laser Diffraction

Malvern Mastersizer S (Malvern Instrument Inc.)

8

2.2.2

Sieve

Sieve Shaker (Retsch)

2.2.3

Electrozone Sensing

Electrozone sensing
(Coulter Counter)
9

Coulter counter can measure particle size ranging from

0.4 1200 micron. Its response is unaffected by
particle color, shape, composition or refractive index.

Particles suspended in a weak electrolyte solution are

drawn through a small aperture, separating two
electrodes between which an electric current flows.

The voltage applied across the aperture creates a

"sensing zone". As particles pass through the aperture
(or "sensing zone"), they displace their own volume of
electrolyte, momentarily increasing the impedance of
the aperture.

This change in impedance produces a pulse that is

digitally processed in real time. The Coulter Principle
states that the pulse is directly proportional to the tridimensional volume of the particle that produced it.

Analyzing these pulses enables a size distribution to

be acquired and displayed in volume (m3 or fL) and
diameter (m).

In addition, a metering device is used to draw a known

volume of the particle suspension through the aperture;
a count of the number of pulses can then yield the
concentration of particles in the sample.

2.2.4

Sedimentation

Suitable to be applied in paint and ceramic industry.

Involves instrument that detects the settling rate of

particles in a fluid by monitoring:
o
a change in concentration of particles at a known
depth in the fluid or;
10

the rate at which particles settle out onto the

bottom of the container in which it confined.
Parameter measured are Stokes diameter where
Stokes law is obeyed.
o

Sedimentation by gravity involves particles of size more

than 2 micron

Sedimentation under centrifugal force involve particles

of size less than 2 micron

2.2.5

Scanning Electron Microscope (SEM)

A microscope generating an electron beam scanning

back and forth over a sample.

Due to the interaction between the beam and the

sample, several different signals are produced
providing the user with detailed information about the
surface structure, differences of atomic number within
the sample or information about elemental content.

SEM image of grinded palm kernel shell particles at 21x and

2000x magnifications respectively.

11

3.0 Sampling method

Most laboratory tests use only a small sample and this has
to be taken from a production stream or from an existing,
stored material. This sample has to be the representative of
the whole material.
Powders are unlike fluid where the properties change easily
under applied load.
They may consolidate with time,
segregation may occur in transfer.

and

attrition

and

Sampling is an important element of powder handling such

that it demands careful scientific design and operation of the
sampling system.
The purpose is to collect a manageable mass of material
(sample), which is representative of the total mass of
powder from which it was taken.
This is achieved by taking many small samples from all parts
of total which, when combined, will represent the total with
an acceptable degree of accuracy.
This means that all particles in the total must have the same
probability of being included in the final sample.
To satisfy the above requirements, the following basic
golden rules of sampling should be followed wherever
practicable.
o Sampling should be made preferably from a moving
stream (this applies to both powders and suspensions)
but powder on a stopped belt also can be sampled.

12

o A sample of the whole of the stream should be taken

for many (equal-spaced) periods of time rather than
part of the stream for the whole of the time.

Re-combined primary sample taken from the whole is too

large for most powder test and so that, it need to be subdivided into secondary or tertiary sub-sample.

Allen (1981) reviewed and tested most methods available

for sampling splitting and found the best is equipment
called spinning riffler (Figure 1).

In spinning riffler, the sample is slowly conveyed by a

vibratory feeder from the feed hopper to the rotating
carousel where it is divided into many container ports via a
machined rotary head.

The sub-samples are collected in this depending on how

many samples are required.

Feed rate is controlled by varying the gap under the hopper

and varying the vibration of the feeder.

13

Spinning Riffler

Sample Splitter

14

4.0 Types Of Particle Diameters, Mean Size And Size

Distribution.
4.1 Types of particle diameters
Some definitions of terms used:

dp

Sieve size, the width of the minimum square

aperture through which the particle will pass

dv

Volume diameter, the diameter of a sphere

having the same volume as the particle.

dsv

Surface/volume diameter, the diameter of a

sphere having the same external surface
area/volume ratio as the particle.

ds

Surface diameter, the diameter of a sphere

having the same surface as the particle.

Some of the diameters are related through Waddells

sphericity factor, , defined as;

= surface area of equivalent volume sphere

surface area of the particle
Some common diameters used in microscopy analysis are:
(Figure 1.1, page 2)
Equivalent circle diameter
Martins diameter
Ferret diameter
Shear diameter

d
= v
ds

(2.1)

15

or

d
= sv
dv

(2.2)

, dv and dsv can be calculated exactly for geometrical

shapes such as cuboids, rings etc.

According to Geldart (1989), in practice, value of for a

specific material can be obtained as follows;
o Sieve out a narrow size fraction of the powder, usually
the middle of the size range or the coarser material, so
as to obtain at least 0.5 1 kg.
o Split this down using a riffler to obtain a few hundreds
particles.
o Count the exact number (n) and weight them (M).
Knowing the particle density, p, calculate the
equivalent sphere volume diameter, dv from

6M
dv =

n
p

1/ 3

(2.3)

o Put the 0.5 1 kg of powder in a circular tube 50 75

mm diameter and measure the pressure drop, P,
across the bed at a variety of low flowrates.
o The Carman-Kozeny equation relates pressure drop to
particle size, dsv, voidage and bed depth, L at low
Reynolds number.
2

P 150(1 ) U
=
L
d sv2
3
16

(2.4)

o Thus, dsv can be calculated from pressure drop data.

o From equation above, calculate .

For irregular particles:

o Particles more than 75 m: sieving
o Particles less than 75 m: Laser diffraction technique:
e.g. Malvern Mastersizer, Coulter Counter etc.

Different measurement techniques give different sizes if the

particles are non-spherical, which is usually the case.

dsv and dv are generally considered to be the most useful

sizes where fluid/particle interactions are involved. (eg. Flow
thru fixed and fluidized bed, pneumatic conveying etc.) and
they are related to each other thru the particle sphericity, .

Malvern Master Sizer S: it measures dv if sphericity is

known, than dsv can be estimated.

Other way to find dsv is by using Ergun equation.

According to Abrahamsen & Geldart (1980):

o For quartz: = 0.8
o
dv = 1.13dp

(2.5)

The average sphericity for regular figures:

dsv
= 0.773
dv

(2.6)

Thus, for non-spherical particles,

dsv 0.87dp

17

(2.7)

For spherical or near-spherical particles;

dv = dsv = dp

(2.8)

ASSIGNMENT
1. Calculate the equivalent volume sphere diameter dv and the
surface volume equivalent sphere, dsv of a cuboid particle of
side length 1, 2 and 4 mm.
2. Particles with density of 2000 kg/m3 with sphericity of are
poured into a container. A sample of 2000 particles are taken
from the powder weighed 1000 mg. determine the volume
diameter, dv and surface volume diameter, dsv of the particles.
4.2

Mean size and size distribution

No industrial powder is monosized and it is usually

necessary to characterize the powder by both the size
distribution and a mean size.

If a powder of mass M has a size range consisting of Np1 of

size d1, Np2 of size d2 and so on, the mean surface/volume
size is expressed by:

d sv =

d sv =

Np1d13 + Np 2 d 32 + Np 3 d 33 + ...
Np1d12 + Np 2 d 22 + Np 3 d 32 + ...

(2.9)

x
x d

(2.10)

where x is the weight fraction of particles in each size range.

18

When sieving is used, d1, d2, are replaced by the

arithmetic averages of adjacent sieve apertures, dpi and the
equation becomes:
dp =

1
xi d pi

(2.11)

and dp is mean particle size

and not directly related
to the dsv.
Mean particle size, dp: emphasis to the important influence
which small proportions of fines have.

Equation (2.11) should not be used if the powder has an

unusual distribution, for example bi- or tri-modal or has an
extremely wide size range.

This type of powder will not behave in a homogeneous way

and cannot be characterized by a single number.

Refer to Table 1 for example.

The British Standard Sieve series is arranged in multiples of

2, and this is used as a basis in Table 2.

This will give an idea of relative spread as judged from the

number of sieves. (Refer Table 2).

It is always advisable to first plot the size distribution of

powder as a weight fraction, or percentage in a size range,
against the average size, i.e. x vs. dpi because a plot of
cumulative percentage undersize can conceal peculiarities
of distribution.

19

Table 1: Size Distribution of sieved sand

Sieve aperture,
m

Size dpi,
m

Wt % in
range, xi

Cum. %
undersize

-600 + 500
-500 + 420
-420 + 350
-350 + 300
-300 + 250
-250 + 210
-210 + 180
-180 + 150
-150 + 125

550
460
385
325
275
230
195
165
137

0.50
11.60
11.25
14.45
20.80
13.85
12.50
11.90
3.15

100
99.5
87.9
76.65
62.2
41.4
27.55
15.05
3.15

Table 2: Width of Size Distributions Based on Relative Spread

Number of sieves on
which the middle 70%
(approx.) of the powder is
found
1
2
3
4
5
6
7
9
11
> 13

Type of distribution

d pm
0
0.03
0.17
0.25
0.33
0.41
0.48
0.60
0.70
> 0.80

20

Very narrow
Narrow
Fairly narrow
Fairly wide

Wide

Very wide

Extremely wide

Other method of characterizing powder, the median, dpm:

corresponds to the 50% value on the graph of cumulative
percentage undersize versus size.

There is no universally agreed way of comparing the width

of the size distribution of two powders having different mean
sizes, nor of defining how wide a distribution is.

In order to compare the width of the size distribution of two

powders having different mean size or defining how wide a

distribution is; relative spread, d

pm

from cumulative

percentage undersize plot is used as suggested by Geldart

(2003).

d84% d16%
=
2

(2.12)

See Figure 1 for example.

In the example given, = 105 m and /dpm = 0.39.

21

Figure 1:

Size distribution of a sand characterized in various

ways

22

5.0

Particle Size Analysis

5.1 Equivalent of Means

A population of particles is described by a particle size
distribution.
Particle size distributions may be expressed as frequency
distribution curves, f(x) or dF/dx or cumulative curves, F(x).
The distributions can be by number, surface, mass or
volume (if particle density does not vary, then the mass
distribution is the same as the volume distribution).
In most practical applications, the population of powders
must be described by a single number.
Some of the central tendency expression which depending
on the particular application are as below:
o Mode: The most frequently occurring size in the
sample. The mode has no practical significance as a
measure of central tendency and is rarely used in
practice.
o Median: Easily read from the cumulative distribution as
the 50% size, which splits the distribution into two
equal parts. Also has no special significance as a
measure of particle size.

Many different means can be defined for a given size

distribution. All the means below can be described by the
equation;

23

g(x )dF
g(x ) =

0
1

but

dF

(2.13)

dF

= 1

(2.14)

and thus,

g ( x ) = g ( x )dF
0

(2.15)

where x is the mean and g is the weighting function, which

is different for each mean definition (refer Table 3).
Table 3: Definition of means
g(x)
Mean and notation
dp
Arithmetic mean, dp,a
2
dp
Quadratic mean, dp,q
3
dp
Cubic mean, dp,c
Log dp
Geometric mean, dp,g
1/ dp
Harmonic mean, dp,h

Equation (2.15) tells us that the mean is the area between

the curve and the cumulative distribution, F(x) axis in a plot
of F(x) versus the weighting function g(x) (refer Figure 2).

Graphical determination of mean is always recommended

because the distribution is more accurately presented as a
continuous curve.

24

Figure 2: Plot of cumulative frequency vs. weighting function,

g(x).

25