A PRELIMINARY INVESTIGATION CONCERNING THE EFFECT

OF PARTICLE SHAPE ON A POWDER’S FLOW PROPERTIES

By:
Mark Bumiller John Carson, Ph.D. James Prescott
Malvern Instruments, Inc. Jenike & Johanson, Inc. Jenike & Johanson, Inc.
Vice President President Senior Consultant

Introduction

The particle size distribution and shape of powders and suspensions can affect basic unit
operations including size reduction, classification, mixing, filtration, crystallization, and flow
through hoppers and Y-branches. As a result, particle characterization is often demanded by
regulatory agencies and purchasing contracts. Everything from incoming raw materials to
final products may have size specifications, and once a size specification is written the
material must be carefully analyzed in an analytical laboratory that typically contains one or
more particle characterization techniques for determining the size distribution, shape and
flow characteristics of powders, suspensions and emulsions.

Recently, attention has been given to trying to understand the complex relationship between
particle size, shape, and powder flow (1). Powder flow is a complex subject, influenced by
other factors besides particle size and shape, including density, moisture content,
temperature, consolidation pressure, and time of storage at rest (2). Although shape is
expected to influence powder flow, the relationship between size, shape and flow is not
thoroughly understood, especially in the fine particle regime below 15 µm. To further
complicate the situation, the shape of the particles also influences the size measurement, and
is dependent on the sizing technique used for the analysis.

The effect of particle shape on size measurements

Many different sizing techniques are used and shape affects each differently. Several
researchers have investigated how the shape of particles influences size analysis results (3,4).
The differences are often accentuated at the tails of the distribution – the D10 and D90 more
than at the median, the D50. Since the differences appear at the tails of the distribution, the
spans (defined as (D90-D10)/D50) are often dissimilar. Consider the example of needle shaped
particles measured by sieves and a light scattering system (sometimes called laser
diffraction). Depending on both the aspect ratio and the size of the longest dimension, the
fibers will often tumble through the laser beam at random orientations to the light source. At
times the laser sees the maximum projected area, at other times the minimum projected area
(essentially a sphere with a diameter equal to the thickness of the needle) and every
orientation in-between. Therefore the distribution recorded by this technique could be a
broader distribution than results measured by sieves, microscopy or other techniques.

Using light scattering to measure shape

Light scattering can also provide morphology data for certain specific particle shapes. Several
researchers (5,6) have described techniques for determining a shape characteristic known
often as the chunkiness ratio for plate-like particles. In addition to the size distribution,

1
instruments can take a zero angle turbidity (obscuration) measurement that can be used to
calculate the volume concentration of particle present using the following equation derived
by combining the Mie theory of light scattering with the Beer-Lambert law:

100 loge (1- Obscuration)

where c is the concentration (%), b is the beam length, Vi is the volume in size band I, Qi is
the extinction coefficient of size band I and d i is the mean diameter of size band I. The
extinction coefficient is a measure of how efficient a particle of a particular size is at
scattering light. Equation (1) shows the relationship between volume concentration and the
obscuration measured as part of a normal experiment and is quite accurate for nearly
spherical particles. As particles become less spherical and more plate-shaped, errors in the
calculated volume concentration arise from an increase in obscuration observed from plate
shaped particles when compared to an equal volume of a spherical material. The ratio
between the actual volume concentration and the volume concentration calculated by the
particle size analyzer is used to calculate a three-dimensional shape value known as the
chunkiness ration as shown in equation (2).

Chunkiness ratio = Ds 0.5 (2)

Where Ds0.5 = Median diameter of the distribution based on surface area

The chunkiness ratio and thickness of plate-shaped particles can be determined through a
relatively simple experiment by weighing a sample of known density and adding this sample
to a known volume of diluent in the analyzer. This same experiment can be performed on
more spherical particles as a method to confirm the choice of refractive index for the sample
material. A more accurate choice of refractive index will result in a more accurate calculated
volume concentration.

There is a long tradition of interpreting a plot of light intensity versus scattering angle to
obtain shape information on particles (7). The shape of the curve has been used to provide
information on polymer conformation, and derivatives such as the Guiner plot (for globular
structures), the Kratky plot (for Gaussian coils and chains) and the bending rod plot for semi -
flexible chains. The same idea has now been applied to examining the structure of aggregates
to obtain the fractal dimension. Raper and Amal (8,9) have published numerous articles
where light scattering instruments were used to quantify the fractal numbers of floc
structures. The fractal number then comes from the negative slope of the linear portion of a
log-log plot of scattering intensity versus wave number. These studies have proven useful for
characterizing flocculated systems where fractal numbers provide important information in
addition to the size distribution. A loose aggregate will have a low fractal dimension, while a
more compact aggregate will have a higher fractal dimension.

2
 Other particle shape measuring techniques

Although light scattering instruments can provide useful shape information in a few specific
circumstances, this is not a common technique used by industry. Particle shape analysis is
commonly performed using a variety of techniques including microscopy, image analysis and
other automated or semi-automated optical techniques. Just as a variety of methods are used
for particle shape analysis, a plethora of calculations are currently used in attempts to
quantify the shape of particles (10). Unfortunately, there is little harmonization in the world
of particle shape characterization.

Although microscopy can be used for qualitative particle shape analysis, image analysis is
often required to automate the complex task of assigning shape values to large numbers of
particles. Both microscopy and image analysis investigate particle morphology by examining
the two dimensional projection of the particle. As with any particle characterization
technique, sample preparation and presentation to the optical system is a critical element of
the analysis routine. One reason light scattering systems are affected by particle shape is the
fact that particles moving through the optical system in turbulent flow typically arrange
themselves in random orientation to the light source and detection system. This “tumbling”
effect results in particles exposing their minimum and maximum projected areas to the
instrument optics (along with every other conceivable orientation). This random orientation
broadens the reported size distribution.

Since random particle orientation causes difficulties when analyzing particle shape, it follows
that it would be beneficial to align the particle flow through an automated image analysis
system. The flow particle (wet) image analysis data presented in this study was collected on a
Sysmex FPIA-2100 image analysis instrument where the particles are suspended in a liquid
and hyrodynamically focused through a shear cell during data acquisition. Since the particles
are subject to shear forces in the cell, the maximum projected area is exposed to the imaging
camera for analysis. Particles were assigned both a particle size and shape value. The size is
an equivalent spherical diameter and the shape number is the circularity defined in equation
(11) below.

Circularity = perimeter of circle w/same area (11)

Some powders are best analyzed in their natural dry state. The dry image analysis data in this
study was generated using a Pharma Vision 830 system that uses a sample preparation device
that disperses the particles evenly and randomly on a glass slide or plate. T he slide is then
placed on a bench where the objective stages across the sample area and collects the images
for analysis. One of the shape values assigned by this system is defined as roundness.
Roundness is a measurement of the length – width relationship, with a value in the range [0.0,
1.0]. A perfect circle has roundness 1.0, while a very narrow elongated object has roundness
close to 0. Intuitively the roundness is a comparison between the “strength” of the major axis
and the “strength” of the minor axis. For a circle or a square, the “strength” of the major and
minor axis is about the same. However, for a rod shaped object the “strength” of the major
axis is considerably greater than the “strength” of the minor axis.

Formally, this roundness R is defined as:

3
R =  λ 1 − λ 2 (12)
 λ 1 + λ 2

where λ1 and λ2 are the eigenvalues of the covariance matrix M.

 I xx I xy 
M =  
 (13)
 I xy I yy 
I xx = ∑ xi x j (14)
i, j

I xy = ∑ xi y j (15)
i, j

I yy = ∑ yi y j (16)
i, j

where xi and yi are the co-ordinates of pixel pij of the object relative to the center of gravity of
the object. The major axis of the object is the eigenvector of the maximum eigenvalue of
matrix M, and is essentially the principal component of the object.

Powders studied

The materials studied in this investigation are glass spheres, calcium carbonate crystals and
plate-shaped talc powders (shown in Figures 1, 2 and 3). These powders are similar in size
but have quite different morphologies. They also are quite different in particle hardness.
Glass is roughly six times as hard as talc on the moh scale. Calcium carbonate is roughly in
the middle between these two extremes.

Figure 1: Glass Spheres Figure 2: CaCO3 Figure 3: Talc

The size distributions were analyzed using light scattering (Malvern Mastersizer2000) and,
both wet (Sysmex FPIA) and dry (Pharma Vision 830) image analysis. Particle shape is
defined by circularity described in equation (11) and roundness as described in equation (12).
The chunkiness ratio and thickness of the talc sample was determined using the Malvern
Mastersizer2000 as described earlier in this paper. Table 1 summarizes the size and shape
characteristics of the powders in this study and Figures 4 and 5 present size and shape
summaries.

4
 Glass Calcium
spheres carbonate Talc
Circularity (wet image analysis) 0.914 0.808 0.626
Roundness (dry image analysis) 0.902 0.677 0.575
Size (D50 ,µm by wet image analysis) 6.44 5.75 4.97
Size (D50 ,µm by dry image analysis) 8.44 5.912 7.79
Size (D50 ,µm by laser diffraction) 6.45 5.1 5.35
Chunkiness ratio 2.44
Thickness 1.88

Table 1
Dv50 by Technique

9
8
7
wet image
6 analysis
5
dry image
4 analysis
3
laser diffraction
2
1
0
Glass CaCO3 Talc

Figure 4

Circularity/Roundness

1
0.9
0.8
0.7 Circularity (wet image
0.6 analysis)
0.5
0.4 Roundness (dry image
0.3
analysis)
0.2
0.1
0
Glass Calcium Talc
spheres carbonate

Figure 5

5
After determining the size and shape distributions of the samples, the powders were exposed
to a series of powder flow tests to investigate any possible relationships between particle
shape and flow properties.

Characterizing powder flowability

The term “powder flowability” can be thought of simply as the ability of a powder to flow.
Using such a definition, powder flowability is sometimes thought of as a one -dimensional
characteristic of a powder, whereby powders can be ranked on a sliding scale from “free-
flowing” to “non-flowing”. Unfortunately, this simplistic view lacks science and
understanding sufficient to address common problems encountered by the formulator and
equipment designer.

Flowability is not an inherent material property, but the result of the combination of material
physical properties that affect material flow AND the equipment used for handling, storing,
or processing the material. Equal consideration must be given to both the material
characteristics and the equipment. The same powder may flow well in one hopper but poorly
in another; likewise, a given hopper may handle one powder well but cause another powder
to hang-up. Therefore, a more accurate definition of powder flowability is the ability of
powder to flow in a desired manner in a specific piece of equipment.

The specific bulk characteristics and properties of a powder that affect flow, which can in
principle be measured, are known as flow properties. These flow properties refer to the
behavior of the bulk material, and arise from the collective forces acting on individual
particles, such as van derWaals, electrostatic, surface tension, interlocking, friction, etc. For
more information on underlying particle properties that contribute to flow behavior, refer to
Rumpf (11), or more recently, Podczeck (12).

Armed with the flow properties of a powder, engineers can optimize the selection of transfer
equipment. These same properties can be used as a basis for retrofitting existing equipment to
correct flow problems. Formulators can use these properties during product development, to
predict flow behavior in existing equipment. The flow properties that are generally of most
interest include:

Cohesive Strength. The consolidation of powder may result in arching and ratholing within
transfer equipment. These behaviors are related to the cohesive strength of the powder, which
is a function of the applied consolidation pressure. In a lab, cohesive strength can be
measured accurately by a direct shear method. The most universally accepted method utilizes
the Jenike shear tester and is described in ASTM standard D 6128 (13). By measuring the
required shear force for various vertical loads, a relationship describing the cohesive strength
of the powder as a function of the consolidating pressure can be developed (2). This
relationship, known as a flow function, can be analyzed to determine the minimum outlet
diameters for bins, press hoppers, blender outlets, etc. to prevent arching and ratholing (14).

Internal Friction. Internal friction values are important when characterizing the flow
properties of a powder. Such friction is caused by solid particles flowing against each other,
and is expressed as an angle of internal friction. This angle can be measured via the cohesive
strength tests described above.

6
Wall Friction. Used in a continuum model, wall friction (or particles sliding along a surface)
is expressed as the wall friction angle or coefficient of sliding friction. The lower the
coefficient of sliding friction, the less steep hopper or chute walls need to be for powder to
flow along them. This friction coefficient can be measured by sliding a sample of powder in a
test cell across a stationary wall surface using a shear tester (2, 13). The coefficient of sliding
friction is the ratio of the shear force required for sliding to the normal force applied
perpendicular to the wall material coupon. A plot of the measured shear force as a function of
the applied load generates a relationship known as the wall yield locus. This flow property is
a function of the powder handled and the wall surface in contact with it. Variations in the
material, or the wall surface finish, can have a dramatic effect on the resulting friction
coefficient (15). Wall friction can be used to determine the hopper angles required to achieve
mass flow.

Bulk Density. The bulk density of a given powder is not a single or even a dual value, but
varies as a function of the consolidating pressure applied to it (2,14). By measuring the
degree to which a powder compacts as a function of the applied pressure and expressing the
results as a straight line on a log-log plot, the slope of this line represents the material’s
“compressibility”. The resulting data can be used to accurately determine capacities for
storage and transfer equipment, as well as to provide information to evaluate wall friction and
feeder operation requirements.

Permeability. Flow rate limitations may occur when handling fine powders, due to the
expansion and contraction of voids during flow creating air pressure gradients within the
powder bed. The permeability of a powder, or its ability to allow air to pass through it, will
have a controlling effect on the discharge rate that can be achieved. Permeability is measured
as a function of bulk density (2). The method typically employed involves measuring the
flow rate of air at a specific pressure drop through a sample of known density and height.
Once this relationship is determined it can be used to calculate critical powder discharge rates
that will be achieved for steady flow conditions though various orifice sizes.

Test results and analysis

All tests were run at room temperature on samples at their as received moisture content (0.3%
for talc, 0.1%. for the other two). Only continuous flow conditions were analyzed. Wall
friction was measured on a coupon of 2B finish stainless steel sheet. The flow tests results
can be summarized in many different ways. Cohesive strength can be used to calculate the
minimum outlet diameter, BC, in a conical mass flow hopper to prevent arching. A powder’s
bulk density/pressure relationship can be described either by the range of bulk density values
or by the slope, β, of this relationship on a log-log plot. Wall friction can be represented by
the recommended hopper wall angle θc (measured from vertical) to achieve mass flow in
hopper fabricated from the tested wall material and for a particular outlet size (chosen at 12
in. for Table 2). A powder’s permeability can be used in conjunction with its bulk
density/pressure relationship to calculate critical steady state flow rates through the outlet of a
mass flow hopper. An outlet size of 2 ft. and an Effective Head of 5 ft. were used to tabulate
the values in Table 2. Finally, a powder’s internal friction can be represented by its effective
angle of internal friction, δ. Using these means to characterize the flow properties of the
powders studied, the results are shown in Table 2.

7
 Talc Calcium Carbonate Glass Spheres
BC (ft.) 0.4 0.6 0.1
Bulk density
range [pcf] 13 to 43 35 to 75 53 to 84
β 0.16 0.11 0.06
θc [deg] 4* 12 23
Critical flow rate [tph] 0.3 9.1 2.8
δ [deg] 35 38 36
*Flow questionable along any sloping hopper surface
Table 2

Looking at this data in an attempt to find possible correlations between flow property values
and particle size or shape, several hypotheses can be developed:

1. The higher the mean circularity the lower the wall friction angle, and hence the larger θc.
This relationship holds for the three materials tested on 2B finish stainless steel sheet. It
seems reasonable to expect that more spherical particles are, in general, less frictional.
However, particle hardness may also be affecting this correlation, since the hardness of
these three materials increases as mean circularity increases. Since harder particles
deform less under load, their surface area in contact with the wall surface may be less than
with softer particles. If so, this would result in lower wall friction and in turn, larger
values of θc.

2. The higher the mean circularity the lower the value of β. This relationship holds for all
three powders studied. While there is some reason to expect that more spherical particles
may be less compressible, a stronger driving force is probably particle hardness.

3. The higher the mean diameter the higher the permeability, and hence the higher the
critical steady state flow rate. This relationship does not hold well for all three materials.
While the talc has the smallest mean diameter and the lowest critical flow rate, the glass
spheres have the highest mean diameter but a lower critical flow rate than the calcium
carbonate. The calcium carbonate has an intermediate value of mean particle diameter but
the highest critical flow rate. Undoubtedly, there are other variables besides mean
diameter that affect critical flow rate, e.g. the complete particle size distribution,
compressibility, cohesiveness, and more. It is also important to bear in mind that the
differences in mean particle diameter between the three powders studied are quite small.

4. The higher the mean circularity the higher the critical steady state flow rate. As with
hypothesis 3, this works well for the talc (lowest mean circularity, lowest critical flow
rate), but not for the other two.

5. The higher the mean circularity the lower the cohesive strength, and hence the smaller the
value of BC. This works for the glass spheres (highest circularity, smallest BC), but does
not work for the other two powders that were analyzed. Undoubtedly, there are other
significant factors besides mean circularity that affect a powder’s cohesive strength.

8
6. The higher the mean particle diameter and the lower the particle diameter standard
deviation, the higher the critical flow rate. There is too little data to evaluate this
hypothesis.

Conclusions

This study has focused on three powders having approximately the same particle size but
quite different shapes. We have found a possible correlation between mean circularity and
two measures of powder flowability: wall friction angle and compressibility parameter β.
Correlation with other measures of flowability were either very weak or the data was too
limited.

Clearly, much more data of this type must be generated before any firm conclusions can be
drawn. However, it is encouraging to see the possible correlations that have been developed
so far.

Acknowledgements

The authors wish to thank Specialty Minerals of Bethlehem, PA for donating the calcium
carbonate used in this study and Luzenac of Englewood, CO for donating the talc sample.

References

1. F. Podczeck, Y. Miah, “The influence of particle size and shape on the angle of internal
friction and the flow factor of unlubricated and lubricated powders”, Int. J. Pharm., 144
(1996)
2. J.W. Carson and J. Marinelli, “Characterize Bulk Solids to Ensure Smooth Flow,”
Chemical Engineering, 101 (4), 78-90 (April 1994).
3. M. Naito et. al, “Effect of particle shape on the particle size distribution measured with
commercial equipment”, Powder Technology, 100 (1998)
4. B. Kaye, D. Alliet, L. Switzer, C. Turbitt-Daoust, “The effect of shape on intermethod
correlation of techniques for characterizing the size distribution of powder”, Part. Part.
Syst. Charact., 14 (1997)
5. Baudet, G., M. Bizi, J. Rona, “Estimation of the average aspect ratio to lamellae-shaped
particles by laser diffractometry”, Part. Sci. & Technol., 11 (1993)
6. Lips, A., Hart P., Evans D., Proceedings to the fifth European symposium in particle
characterization (1992)
7. Denkinger, P., Burchard, W., J. Polymer Science, Part B, 29
8. Jung, S., Amal, R., Raper, J., Part. Part. Syst. Charact., 12 (1995)
9. Bushell, G., Amal, R., Raper, J., Part. Part. Syst. Charact., 15 (1998)
10. F. Podcek, “The Shape of Powder particles and its Influence on Powder Handling, IFPRI
Review Report (1999)_
11. H. Rumpf, and W. Herrmann, “Properties, Bonding Mechanisms and Strength of
Agglomerates,” Processing Preparation, 11 (3), 117-127, (1970).
12. F. Podczeck, Particle-particle Adhesion in Pharmaceutical Powder Handling, (Imperial
College Press. London, 1998).
13. Standard Shear Testing Method for Bulk Solids Using the Jenike Shear Cell, ASTM
Standard D6128-97, American Society for Testing and Materials (1998).

9
14. A.W. Jenike, Storage and Flow of Solids (Bulletin 123 of the Utah Engineering
Experimental Station), 53 (26), (1964, Revised 1980).
15. J.K. Prescott, D.A. Ploof and J.W. Carson, “Developing a Better Understanding of Wall
Friction,” Powder Handing and Processing, 11 (1), 27-35 (January/March 1999).

10