Previous Page

Two types of dust measurement have been used here, one concerned with
respirable dust and the other with the appearance of the product. Sample 4
has a large respirable dust content, but it does not appear to be as dusty as
the other samples.
It is clear that within each area of granulation it is possible to control the
physical properties of the product, either by engineering changes, careful
process control or by clever formulation. Comparison of different samples
is, however, made difficult by the wide difference in granule sizes. A major
benefit of the move towards extrusion is that it will allow standardization of
test methods. At the present time progress is being made in this important
area [57].

4.3 Physical properties

One of the primary purposes of formulation is to allow suitable physical

properties to be attained. Active ingredients have a variety of physical
properties of their own and this has led to a range of formulations which can
cope with this. Some of the properties of interest are specific to a particular
manufacturing route and reflect the size and shape of the granules which
form. A good example of this is spray drying, which tends to form highly
spherical, but small, granules. A consequence of the small size is that the
granules do not in themselves carry much momentum. Penetration of the
surface of a spray tank is therefore often difficult for this type and relies to
a large extent on the tap density of the bulk material. Tap density is of less
importance for other, larger granules although it is important that it is
constant during manufacture. This not only ensures room in the packaging
for each dose but also allows volumetric measurement at the site of appli-
cation. Other properties such as granule size distribution can be of impor-
tance, depending on the particular manufacturing method.
The following properties are thought to be universally important:

• dry properties:
• strength;
• friability;
• dustiness;
• wet properties:
• dispersion time;
• dispersion quality (particle size/sieve residue);
• dispersion mechanism.

These physical properties are dependent on several factors. Three main

headings can be used to categorize the areas which are most important:
1. granule size and shape;
2. pore size distribution;
3. quantity and type of binders.

4.3.1 Granule size and shape

The first of these areas, granule size and shape, has been mentioned
already, and is commonly used as a means of controlling properties such as
dispersion time. For example, during the life of the commercial product
'Cymbush', its diameter was changed from 1.2 to 0.6mm, in order to speed
up its dispersion in cold water (results taken from statistical analysis of
production figures over the period 1985-1989 at Zeneca's production plant
at Yalding, Kent, UK; Table 4.6). Figure 4.3 displayed the dependence of
dispersion time on granule diameter for samples made by different pro-
cesses. This included spherical as well as cylindrical granules [58].
Granule size and shape also affect other properties such as dustiness and
strength. In terms of crush strength, the quoted figures for typical pan-
granulated material are often surprisingly low (say 10-2Og) [59, 6O]. De-
spite this, samples of commercially available product do not often appear to
have suffered excessive crush damage. The reason for these apparently
contradictory facts is of course the shape factor. Examination of a single
spherical granule in a crushing device demonstrates the problem. A sphere
cannot distribute the applied load in any way other than straight through
the middle of the granule. Compare this to an extruded shape, where the
granule lies flat, and it will be apparent that the extrudate has an advantage,
i.e. the volume which takes the load is significantly larger (Figure 4.4).
Nevertheless, a collection of spheres held within a volume can pack
efficiently, and in that case the load would be distributed in three dimen-
sions. The result of this is that examination of a single sphere displays an
apparent weakness which is not seen when a gross sample is studied. Com-
parisons between the crush strength of different granule types is therefore
a pointless exercise. However, it is possible within any one granulation type
to measure crush strength and to use this as a criterion for design. In
practice though, the range of granule sizes present in samples, other than

Table 4.6 Average results made on successive production

campaigns on a commercial WG prepared at two different
diameters

1.2mm 0.6mm
0 0
Dispersion (s) (temperature, C) 152 (2O C) 135 (1O0C)
Friability (%) 2.0 0.5
Dustiness (mg/g) 2.2 0.2
Dispersed particle size (^m) 25.0 3.0
extrudates, makes examination of single granules very difficult. Examina-
tion of bulk attrition is considerably easier than examination of bulk
strength, and this has been the more common approach.
Properties such as dustiness are related to the surface of the granules and
reflect the ease with which particles can be scrubbed off the outer layers
[61]. In considering shape, the details of surface texture need to be taken
into account as these can influence this property. One study of a formula-
tion showed that narrow-diameter granules can sometimes form with a
smoother surface than would larger-diameter samples. This can be effective
in reducing the dustiness of a product (Figure 4.5). In this example, granule
size apparently influenced a property, but it was in fact a change to the
surface texture that was responsible. Clearly care is required when analys-
ing results. Spray-dried products, which tend to have smooth surfaces, often
display low surface abrasion dustiness, although the small size of the gran-
ules can lead to other dust-like properties [61].
Where granules are very small, they start to display the properties of dust.
An exact definition of the term dust is open to debate [62]. However, it is
clear that the greater the gravitational pull on the granule, the less likely it
is that dust-like behaviour will be observed. An example of this is shown in
Figure 4.6, which displays light obscuration from different size fractions
taken out of a spray-dried WG sample. The apparent dustiness measured
from the observation of a dust cloud settling under gravity is dependent on
granule size. This can easily be detected for granules up to about 0.3mm in
diameter.
Suitable dry properties tend to be favoured by large granules, whereas a
small granule would be preferred for fast dispersion. The balance between
the two demands, in the case of extrusion, seems to lie most favourably with

Figure 4.4 Uniaxial crushing force applied to a spherical granule and to a cylinder.
Dustiness mg/g

Binder %
Figure 4.5 Dustiness readings taken for WG samples at three diameters, plotted against
binder concentration.

seconds
seconds
seconds
seconds
Obscuration

Sl eve fraction
Figure 4.6 Light obscuration for different sieve fractions. Readings taken over various time
scales after pouring.
a diameter of about 0.6mm, although the ease of manufacture of the prod-
uct may also be a factor which has to be taken into account.
The fact that granule diameter is so closely linked to both the wet and dry
properties has prompted work in the area of diameter control. In order to
produce very narrow diameters it has been found necessary to experiment
with new process technology, and this is at the present time an area of active
research [63]. A product which contains both a small and a large dimension
would appear to go some way towards satisfying both needs, and for this
reason extrusion, as a method of manufacture, is growing in popularity.

4.3.2 Particle assemblies and structures

The second factor mentioned, the pore size distribution within the granule
structure, has also been examined [58]. Where the particle size distribution
is kept constant but the porosity (density) of the granule is varied, it has
been shown that faster dispersion is possible where a more open structure
is present (Table 4.7). Unfortunately this also leads to weaker granules, and
the trade-off was found to be quite poor.
The use of open-granule structures would be complicated by the need to
control the degree of openness, and this, combined with the inherent weak-
ness of such structures, argues strongly against this option for product design.
Extrusion is very good at forming dense granule structures. Where this
method is used, the pore size distribution formed and the internal surface
area present are principally governed by the particle size distribution of the
starting powder mixture (e.g. [64]). Powder samples have previously been
milled to different particle sizes and on these occasions it has been shown
that the finer samples were slower to disperse (Table 4.8; G. A. Bell,

Table 4.7 Physical properties of WG samples with differing

porosities and bonding strengths

Porosity (%) Bonding Dispersion (s) Friability (%)

41.7 Strong 105 0.9

95.5 Strong 110 0.8
52.2 Weak 50 2.0
80.7 Weak 35 11.0

Table 4.8 Influence of particle size on the time taken for gran-
ule dispersion

Particle diameter (\im) Granule dispersion (s)

25.3 15
11.2 25
3.3 50
unpublished results, 1988). The reasons for this are debatable, but possible
explanations are thought to be that with smaller particles, and therefore
smaller pores, the rate of intrusion of water, and the total surface area
which has to be wetted out, combine to slow down the dispersion process.
A water-dispersible granule can be regarded as a cluster of particles. The
cluster will have an external shape which may be quite complicated, and it
will have an internal structure with associated surface area, pore volume
and pore size distribution. Control of these factors should lead to control of
the physical properties of the granule, which is of interest with respect to
product design.
The importance of the pore size distribution has been examined with
reference to the liquid content required for granulation. Apart from the
quantity of liquid required, it also affects the basic rheological parameters
of the wet granule, and this in turn affects the degree of dispersion of the
powder [64] in the wet mass, during manufacture. It has been shown that
starting powders with wide size distributions lead to dense granules, with
small internal pore volumes and narrow pores. Close size distributions lead
to open structures with larger mean pore sizes, assuming the same mean
particle size for the starting powder.
We might expect, then, that a WG prepared with a broad range of
particle size would be different to one made with a narrow range. This
should influence several of the physico-chemical properties of the
formulation.

(a) Particle assemblies: wet agglomerates. Paste rheology is a complicated

area, and the relationships between individual particle interactions and the
bulk observation of flow are not well understood. Laboratory observations
have shown that the mean particle size is very important in determining the
quantity of water required, and also the ease of extrusion of a paste. Sam-
ples with fine particles often need large water additions and can display
dilatency, whereas samples with large particles need smaller amounts of
water, but the range over which they extrude can be small. The green body
strength of granules containing fine particles is usually higher as would be
expected from a consideration of Rumpff s equation [65, 66]. Note that //,
the bond strength holding the two particles together, usually has a particle
diameter term (d) depending on the type of bonding present:
= <4i>
•• -^-(l - B]H

where ot is the tensile strength of the agglomerate, e is the porosity of the

granule, d is the diameter of the particles which make up the granule and H
is the bond strength between each pair of particles in the granule.
The work of Newitt and Conway Jones [67] is often cited to explain the
behaviour of wet agglomerates. This compares changes in behaviour of the
 Table 4.9 Quantity of water required to form a paste depends
on the particle size of the powder used

Particle size (|im) Water requirement (%)

2.6 20.1
9.4 12.4
24.2 8.4

particle assembly as the pore volume within the granule becomes filled with
the granulating liquor. Much of the work carried out subsequently has
examined the behaviour of pastes at varying water contents [68-75]. Vary-
ing the amount of water used is one of the easiest ways to control a
granulation process. This is dependent upon the pore spaces between the
particles, and for this reason the particle size distribution of the starting
powder can be an important process parameter.
Studies on pharmaceutical samples of wet pastes have been carried out
with mixer torque rheometers and have shown the value of instrumentation
in assessing paste quality [76-89]. Particle size and the effects of the ingre-
dients are both key aspects in determining paste behaviour, especially
where specialized materials such as microcrystaline cellulose are used. Such
specialized materials are usually avoided in the agrochemical industry be-
cause of the cost. However, recently there have been references to similarly
priced materials in the patent literature. This may indicate that these ma-
terials are becoming commercially viable, especially where agrochemical
tablets are concerned [90-94].
Typical figures for the amount of water required for extrusion are shown
in Table 4.9. In this instance the particle size spread was similar for each
sample. The volume of water required may depend upon the colloidal
forces operating within the wet agglomerate, as well as the degree of mix-
ing, or shear, applied. In addition the soluble components will have an
effect. Consequently, prediction of the required water cut for manufacture
is difficult.
Broad size distributions should permit better packing which in turn
would reduce the internal void volume. We would expect to use lower water
additions under these circumstances.

(b) Particle assemblies: dry structure. When a load is applied through an

extrudate, there will be a distribution of force, and in the case where it is
greater than the strength of the granule, fracture will occur. All fracture is
tensile in nature, in the sense that crack propagation requires the surfaces to
move apart. We would expect granule strength to depend upon
• the nature and amount of the binding material holding the granule
together;
• the size and distribution of particles in the matrix;
• the strength of the particles.
Clearly, the variables outlined above are important and can be altered to
allow a change in the granule strength to be made. Table 4.10 shows some
typical data for extruded samples; the figures relate to the strength of
individual granules, and the quoted materials represent the bulk of the
particles which made up the formulations. As mentioned earlier, it is impos-
sible to quote universally acceptable figures because of the limitations of
granule shape. Note, however, the remarkable strength of single granules
made from inorganic materials.
As the strength of a package of granules will be considerably larger than
the individual strength, there appears to be little to worry about from this
property. The exception would be where large amounts of liquid are
present within the granule formulation, since in that case there would be
more of a challenge. A commercial sample containing 25% of a liquid active
ingredient, however, was examined and found to have a crush strength of
235 g. This appears to be more than adequate for the conditions of use.
A wide particle or pore size range should lead to a granule which is
denser, and therefore stronger, than one formed from a narrow size range.
Consideration of tensile strength [65, 66] or resistance to bending moments
[95] also suggest that a smaller mean particle size will lead to a stronger
granule.

a? = f (4.2)
where otis the tensile strength of the agglomerate, E is the elasticity of the
material, y is the surface energy of the material and d is the particle size.
The situation becomes more complicated, however, when other factors
are taken into consideration. For example, where a finer particle size is used
the surface area of the powder will be larger. Based on a unit mass of
material, there is a reciprocal-law relationship between particle size and
surface area, so that halving the mean diameter will increase the area by a
factor of two. For a similar binder loading there will be less material per unit

Table 4.10 A selection of granule crush strength values;

in each case a binding agent comprises about 10% of the
formulation

Material Crush strength of WG (g)

Bentonite clay 5217

Wetter/salt 3007
Oil/silica 994
Pirimicarb 671
area. This may or may not influence the individual bond strengths within
the granule, depending on whether there is an excess present. At low binder
levels an increase in surface area may lead to a weakening of the structure.
Practical measurements made on model granules to determine the size of
these effects show that they are relatively small, covering a factor of about
two. Table 4.11 shows the results for a range of particles which vary by
about an order of magnitude and are within the practical limitations avail-
able for typical products.
As a complicating factor, particle size distribution can be included. This
makes analysis more complicated, but if we examine a range of mean size
and size distribution, then some idea of the range of crush strength values
can be gained (Table 4.12). This is considerably less than would be expected
from theory and may indicate that the strengths which have been realized
are in fact rather lower than is possible. Note that the particles used for this
series of experiments were very hard (calcium carbonate) and probably do
not take into account fracture of the particles themselves. Individual gran-
ule strengths of about 3kg are, needless to say, quite high.
The dustiness of a granule is known to be influenced by its surface
smoothness, i.e. to the ease of removal of particles from that surface. In the
case of an extrudate this will be strongly influenced by the rheology of
the paste used. Any dust which forms will be related to the particle size of
the starting powder. Finer starting materials may lead to finer dusts [96].

Table 4.11 Crush strength values for calcium carbonate-

dispersible granules are dependent on the particle size of the
starting powder

Sample Crush strength (g) Mean particle size (^m)

1 3747 2.6
2 2887 9.4
3 2690 24.2

Table 4.12 Particle size and size distribution (coefficient of variance) are influencing factors
on crush strength; a range of crush strength values displays the complicated nature of the
relationship

Crush strength (g) Particle size (^m) Coefficient of variation (%)

4280 17.4 135

4008 13.7 160
3747 2.6 123
3376 16.8 158
3328 6.7 112
2887 9.4 93
2690 24.2 105
Table 4.13 Dry properties of a WG are independent of one another

Dustiness Friability Crush strength Particle size Size distribution

(mg/g) (%) (g) dim) (%)

2.9 0.7 2887 9.4 93

2.2 0.7 4008 13.7 160
2.1 0.8 3747 2.6 123
1.8 0.8 3328 6.7 112
2.3 1.0 3376 16.8 158
2.0 1.4 4280 17.4 135
1.6 1.4 2690 24.2 105

Coarse powders, on the other hand, may lead to weak granules which
produce a coarse dust, whereas fine powders may form stronger granules
which, if they do break down, would be considerably more dusty in
appearance. A compromise in terms of particle size may therefore be better
in terms of product design.
Similarly, a broad size distribution may include a fine component which
has the potential to form dust, whereas the breadth of the distribution might
be expected to make the granules stronger (Table 4.13). The ideal distribu-
tion may have to be found experimentally.

(c) Particle assemblies: disintegration in water. In order that a granule can

disperse in water, it is important that the surfaces are initially wetted out.
Simple powder wetting considerations show that this will be related to a
y cosG term, and to the total surface area present [97]. The important
parameter when comparing different distributions of the same material will
therefore be the surface area. Smaller mean sizes or wide distributions
would be expected to take longer to wet out. Similarly the ease of pene-
tration of water into a granule would be expected to be related to the pore
size, which in turn will depend upon the starting particle sizes.
The Washburn-Rideal equation [98, 99] suggests that large pores will
lead to faster penetration, and this points towards a narrow distribution and
a large mean size (e.g. [10O]):
d/ = rycosO ^
dt 41H/
where dl/dt is the rate of penetration of the fluid, T] is the fluid viscosity, / is
the length of the capillary, r is the radius of the capillary, y is the air-liquid
interfacial tension and O is the contact angle between the fluid and the solid.
Dissolution of the binding material is a further complication which is
harder to quantify. For a set loading of binder, it could be argued that
spreading the material as a thinner layer will help to permit faster dissolu-
tion. Halving the mean diameter of the powder particles should in a simple
case allow the binder to be dissolved in half the time. However, this ignores
the fact that water penetration through the pores will be sequential, so there
is the potential for a rate-limiting step. Should the limiting step be dissolu-
tion, then fine particle size will lead to fast granule dispersion, but the
opposite would be true where pore penetration is critical. A real situation
may be a complicated version of both of these scenarios.
Where work has been carried out to determine the importance of size and
distribution on, say, dispersion time, the results have in general followed the
theory, although the scale of the effect was considerably less than expected.
Table 4.14 shows that the dispersion times vary by a factor of just over two.
Unfortunately, in this case, the dispersion times were all quite fast, so the
spread of results was rather narrow compared to the error in the measure-
ment. Other examples of this type of work are not known and, therefore,
the conclusion at the present time must be that particle size does affect
dispersion time, but its effect is relatively small.
In addition to the wetting-out and dissolution steps, there will be the
process of particle separation. This is an area which is not well understood
at the present time, although some work has been done to understand the
contributions of the various viscous, electrical and steric effects. A starting
granule represents a very high concentration of particles, and as these wet
out and disperse, the nature of the paste or slurry that forms will change.
Clearly the interparticle spaces will be important in determining the behav-
iour of the concentrated dispersion that forms (discussed later). After com-
plete dispersion, a consideration of the sedimentation rate of the individual
particles should provide some idea of the ease with which they can be swept
away by any agitation that is supplied. This, however, represents the final
step of the dispersion process, and the area between this and the binder
dissolution should not be ignored:
(]?-& Ap
sedimentation rate a —2—*- (4.4)
T)

Table 4.14 Dispersion time in water is related to particle size,

but size distribution also has an effect

Dispersion time (s) Particle Standard deviation (%)

diameter
(l*m)
25 24.2 104
30 9.4 93
35 17.4 135
35 6.7 111
40 16.8 158
50 13.7 160
60 2.6 122
where d is the diameter of the settling particles, g is the gravitational
constant, Ap is the density difference between the particle and the fluid
and T) is the viscosity of the fluid.

4.3.3 Quantity and type of binders

The quantities and types of binders present in granule formulations are also
major contributors to the measured physical properties (e.g. [101-114]). It
is known, for example, that increasing the level of a dispersant increases the
dispersion time in a roughly linear fashion. A minimum quantity of binder/
dispersant, however, is required to prevent ageing, and to allow for com-
plete dispersion, and there is therefore a balance between this and the need
to reduce the level in order to obtain fast dispersion.
Similarly, the other granule properties, particularly the wet ones, are also
influenced by the dispersant level. The dry properties are less obviously
linked. However, the dispersant influences the rheology of the paste used in
extrusion, and this in turn affects the surface smoothness of the granules.
Surface texture has been shown to influence properties such as dustiness.
Where a combination of dispersants and binders are used, it is clear there
is synergy between the components (see later). Balancing the dry and wet
properties is therefore possible by using the right combination of ingredi-
ents, for example with a binder, dispersant and wetter.
In practice, the formulation is normally altered until the best compromise
has been reached, at which point the diameter of the granule will be exam-
ined to fine tune the product further. This sequence tends to arise because
the formulation will become fixed quite early during the development pro-
cess, whereas the diameter can be selected at a later date.
If we consider the dry properties of a granule, then it is clear that there
are three areas of interest, namely the crush strength, friability and
dustiness.

(a) Dry properties: granule strength. Figure 4.7 shows the crush strength
results for WG samples prepared using both 6 and 10% loadings of different
surfactants and combinations. Obviously the granules become stronger as
greater amounts of binding materials are added. The different binders show
different efficiencies in their ability to impart strength to the WG samples,
each of which has been prepared with the same active ingredient powder
sample. Synergy has often been observed in WG research, so one important
question relates to these possible interactions. In this case the situation is
very simple in that the effects seem to be additive. This is demonstrated
in Figure 4.8 where practical binder mixture results are compared to the
expected results from the pro rata addition of the individual component
effects. Each individual binder result was divided by the concentration used
and this was added to the account for the mixture sample, along with values
 Comparison of Granule Hardness For WGs Containing 6 & 10 %
Surfactant
Granule Hardness (g)

10% Surfactant
6 % Surfactant

Binder Disp. Wetter Disp./ Disp./ Wetter ALLS

Wetter Binder Binder

Sample Content
Figure 4.7 Granule crush strength values from a range of binding agents.

Calculated and Actual Granule Hardness Results for WGs Containing

Various Surfactants

Calculated
Actual
Hardness (g)

EFW and D425 D425 and PVP EFW and PVP D425/EFW/PVP
Granule Components
Figure 4.8 The crush strengths of granules prepared by using mixtures of components can be
calculated by summing the individual contribution from each binder. Generally the mixture is
found to be stronger than expected by 15-20%.
for the other ingredients in the formulation. As can be seen, the fit is very
good, being accurate to within 20% of the real figure. The actual figures are
consistently higher than the calculated ones, and this may be evidence of a
small amount of synergy.

(b) Dry properties: friability. It was thought at one time that granule
hardness, friability and dustiness were all related factors. The surprising
results indicating that they were independent led to confusion until it was
realized that crush strength or hardness was a bulk property, whereas the
other two were related to surface damage. As friability and dustiness tend
to measure different particle size fractions, the fact that these properties are
different is also explainable. The term friability, however, needs to be
defined.
Where large granules are made and sold, there is a concern that they will
break up in transport, and this will lead to small fragments in the packaging.
These fragments may be as small as dust, and have the problems associated
with dust, or they may simply detract from the visual appearance of the
product. A simple sieve test applied after some form of attrition is therefore
common, and typically the size range studied would be below 0.25 or
perhaps 0.15mm (G. A. Bell, 1987, results of a survey carried out in the UK
showing that sieving with a 150 pirn sieve was the most common approach
used, British Agrochemicals Association WG sub-group, unpublished
results). The use of a friability machine is also known, although this has not
been standardized across the industry (several methods are available com-
mercially; for references to nine see [115]).
Examination of the material below 250 pirn in commercial samples shows
that it contains a variety of fragments and primary particles. Whereas the
crush strength of a WG responds directly to the addition of larger amounts
of binder, this is not the case with friability [96]. Figure 4.9 shows some
typical results, in this instance based on granules which have been shaken
gently for 20min. The minimum in a plot of friability against binder level
has been observed for some samples after the application of this test, but
only in cases where very high friabilities were found. It is thought that in this
case the data points were essentially random.

(c) Dry properties: dustiness. Dust can be created during the production
process, in which case it will be present all the way through the life of the
product, up to the point where it is dropped into the spray tank. Alterna-
tively it can form during handling. The term dust is not specific enough on
its own because there are several aspects, or problems, associated with
broken granules or fine material. The dust which is visible is likely to be
very small in size, especially where it has the ability to hang in the air. This
is the worst sort of dust because it represents an inhalation hazard to the
user [62]. Apart from the risks due to toxic ingestion, there is also a consid-
Friability (%)

Binder (%)
Figure 4.9 Variation during sample preparation often outweighs the friability benefits gained
from higher binder loadings.

erable nuisance factor associated with fine dust, for example because it is
visually unattractive. Wetting agents are a common ingredient in the formu-
lation of water dispersible granules and these have a noticeably irritating
effect on the mucous membranes of the inner nostrils. Where dusts form,
these materials often cause problems.
The question as to how fine a dust is can be misleading, because this
concentrates on the size rather than the quantity present. Both factors have
to be taken into consideration, as does the source of the dust. The reason
the source is important is that much of the dust which causes a problem is
formed during pouring and handling, and may not be present until the
package is opened.
The term dustiness is difficult to define per se, other than to define it as a
quantity measured by the use of a particular dust tester. There are three
types of dust tester which are commonly used in the agrochemical industry
[116], and these examine two different size ranges of dust. Heubach and
Cassella dustmeters (Heubach Eng. GmbH, D-3394, Langelsheim,
Germany; the Cassella dust-measuring method has been adopted as CIPAC
method MT 171, apparatus available from Hoechst AG, Abt. Ing-
Forschung, PO Box 80 03 20, D-6230, Frankfurt, Germany) examine the
dust which can be picked up from the surface of a granule in a stream of air
and as such they are commonly thought to represent the inhalation process
and to produce a measure of the risk posed by respiration. The Lorenz dust
tester covers a wider size range of dust and correlates well with the visual
appearance of samples (Lorenz Messgeratebau, Max Planck Strasse 1,3411
Katlenburg-Lindau, Germany). Visual examination of samples is able to
take account of a wide range of particle sizes and, as it is an optical assess-
ment, also measures light scattering efficiency. The use of a laser beam to
measure the visual effect the dust has on the appearance of a WG is
therefore sensible. Note that the Cassella can also be used with an optical
sensor.
Figure 4.10 shows two particle size analyses, one taken from the dust
drawn off from a WG in a Heubach test, and the other by dispersing the
gross sample in water. Clearly the dust is considerably finer than the aver-
age from the starting granule, and this represents some form of particle size
segregation.
Figure 4.11 shows the results of dust collected against time from two
different samples of WG. The two samples were selected from a range, and
were the most, and least, dusty samples available. The two lines on the
graph are straight, which indicates that the dust has formed from the motion
of the granules in the tester. Presumably this represents fine particles being
scraped off the surface as the granules slide past one another.
In order to formulate a granule which is low in dust, we might expect that
higher loadings of binder would be effective because they would allow
greater quantities of glue to be attached to each surface particle. The
problem is, of course, that the amount of glue on the surface is less critical

Granule 0(4,3)36.1 |^m

percentage

Dust mean size D(4.3) 5.9 u,m

Particle size (^m)

Figure 4.10 A comparison between the dispersed particle size from a water-dispersible
granule, and the dust particle size extracted by an air stream apparatus during a dustiness
measurement.
than the texture of the surface, hence we find the results shown in Figure
4.12. Here the quantity of binder is relatively unimportant over a range of
3-8% w/w.
The quantity of dust which is being measured in this type of experiment
is usually very small, say 0.1% w/w of the overall formulation. Adding large
quantities of binder in order to tie down this small fraction of particles

Signal from low dust sample X10

High dust
Low dust
Dustiness (mg/g)

Time (min)
Figure 4.11 Dust collected from two WG samples in a Heubach dust meter. The quantity of
dust given off varied by about two orders of magnitude.
Dustiness (mg/g)

Binder (%)

Figure 4.12 Dustiness, which is a surface property, shows little dependence on binder
concentration.
 seems excessive but it is, of course, very important for safety reasons. The
use of oil as a de-dusting agent is known to be effective. However, great
care must be taken so that the benefit is present over the lifetime of the
product, because oil is known to be absorbed into granule structures
with time.
The visual appearance of dust is, perhaps, better analysed using a Lorenz
tester [116]. Small amounts of fine dust will clearly detract from the visual
appearance of a granule and it is clear that effective sieving of the product
is essential. Formulation can alter the surface smoothness of the granules
and this would appear to be the best way to prevent the formation of dust
during handling.

(d) Wet properties: degree of dispersion. Adsorption isotherms carried

out to ascertain the quantity of dispersant required to form a monolayer on
typical WG samples showed that it was about 0.5 to 1 % w/w for a sample
with a particle size of lOjim. Surface coverage by lignosulphonates
and naphthalene sulphonates typically displays strong Langmuir-type
isotherms, and the adsorbed dispersants do not desorb on dilution. Based
on these data, it might be thought that a suitable fine dispersion could be
formed from a WG formulated with this amount of dispersant. This has not
been found to be the case and it is well known that ten times this amount
would be more typical (Figure 4.13).
Examination of a dried granule by sectioning often shows a significant
concentration profile of the dispersant, the centre being leaner than the
exterior. This is a consequence of a chromatographic effect brought about
Particle size (urn)

Dispersant (%)

Figure 4.13 Below a critical threshold of dispersant, aggregation can easily be detected.
 by the high solubility of most dispersants, and the drying mechanism which
siphons the dispersant solution out through the pores between the particles
[117]. Clearly then, a lot of the dispersant is not available for surface
coverage, although this does not explain the basic aggregation phenom-
enon. Where strong isotherms have been observed, it is possible to dry and
redisperse particles without loss of the dispersant to the solution phase.
Further evidence which relates to the problem comes from spray drying,
where it is known that a residual water level is essential for good dispersion,
depending on the particle size of the slurry used. It would seem then that
particle coating by the dispersant is the least of the problem, but that
dispersed particles which come into intimate contact have a habit of aggre-
gating. There are many explanations which could be given to explain this.
However, as it is a common enough phenomenon from other areas of
particle technology, there is no need to do so. A large excess of dispersant
is essential, and part of the reason for this appears to be in order to provide
multiple layers between the particles during drying.
Gross failure of the dispersant system occurs over quite a narrow range of
concentration (Figure 4.13). For this reason, an excess of dispersant is
highly recommended, but there are other aspects to this that should also be
stated. For example, ageing is a very common occurrence and has been
observed with most types of WG product. The ageing of solid formulations
is well known and is often associated with the movement of water soluble
components with time. This is well illustrated by examining salt granules at
different temperatures (Figure 4.14). Most formulations will be required to
have a long shelf life, and this should be used as a criterion for selecting
the optimum dispersant concentration. If it is accepted that a certain degree
of mobility will occur, then the control measure to prevent ageing would be
Dispersion (sec)

Age(days)

Figure 4.14 WG ageing is known to be temperature dependent.

Table 4.15 Effect of dispersant concentration on dispersed particle size

Particle size Dispersant 1 Dispersant 1 Dispersant 2 Dispersant 2

10% 20% 10% 20%

Fine 7.4 3.8 7.3 4.6

Medium 8.3 4.7 14.4 5.2
Coarse 9.0 9.3 9.3 9.3

the concentration used. A less mobile ingredient may suffice at a lower

concentration.
The particle size of the powder used to form the granule is clearly one of
the most important factors in determining the degree of dispersion
achieved. As the size is reduced, the surface area will be increased, and this
will influence the quantity of dispersant required. Table 4.15 should give
some idea of the importance of this effect.
In the case of the coarse powder, 10% of either dispersant was enough,
and little evidence of aggregation could be observed by microscopic exami-
nation. Where excess was added it was not beneficial. A small amount of
aggregation was seen for the medium-sized samples, in this case 20% dis-
persant would be chosen as it reduced this to an acceptable amount. The
sample with a fine particle size was highly aggregated in all cases; this
sample contained submicron primary particles and the problem was related
to poor deaggregation prior to granulation. Although an increase in the
dispersant quantity clearly helped, the problem was in fact related to grind-
ing of the starting sample.

(e) Wet properties: speed of dispersion. Sodium lauryl sulphate has been
reported as an additive for pharmaceutical solid dosage forms, its purpose
being to enhance wetting and improve dissolution rates [118]. Similarly in
the agrochemical industry, wetting agents such as silicone copolymers or
fluoroaliphatic surfactants [119] have been quoted as useful ingredients for
enhancing the rate of dissolution of pesticide water-soluble or dispersible
granules.
The exact effect of a given surfactant on a particle surface will be deter-
mined by the degree and mode (orientation) of its adsorption on the vari-
ous interfaces, and the reversibility of that adsorption [12O]. Some evidence
to support this is available from experiments carried out on pharmaceutical
tablets.
The Washburn equation relates to the rate of penetration (dl/dt) of a
liquid of viscosity TI in a capillary of radius r and length /:
where y is the interfacial surface tension. Thus to facilitate penetration of
the liquid we would want to
• maximize YLV cos 9;
• minimize the viscosity (TJ);
• maximize pore size.
If we consider the granule structure to be a simplified assembly of pores,
then the Washburn equation is potentially a mathematical model by which
to identify the important physical characteristics of the dispersants and
binders which are incorporated into the granule formulation.
The dispersion process covers a range of concentrations, starting with a
solvent-free system and finishing with a very dilute one. The viscosity of the
binders and dispersants present must therefore be examined over a similar
range of concentrations. Some kinematic viscosity data for typical disper-
sants are shown in Figure 4.15.
The types of materials which are commonly selected are obviously those
which dissolve readily, and are usually free from liquid crystal formation,
such as polyelectrolytes. Where polymers that can gel are used, care is taken
to control the amount present such that this does not lead to dispersion
problems. Note, for example, that the grade of PVP shown in Figure 4.15 is
rarely used at levels higher than about 2% w/w, which is the point where its
viscous behaviour becomes important.
It was at one time thought that wetting of the pores of the granule would
be a critical factor in determining the dispersion time. The data in Table
4.16, however, show that this is inconsistent with theoretical models of
wetting which depend on a y cos 6 term (G. A. Bell, presentation made to
the Royal Society of Chemistry, London, 6 April 1997). Instead, the key

viscosity cSt

Morwet D425
Morwet EFW
PVP
D425/EFW(1:1)
D425/EFW/PVP (1:1-1)

solution concentration (%w/w)

Figure 4.15 The variation of dispersant viscosity with concentration is an important basic
property of the material.
Table 4.16 The term y cos 9 is commonly used in descriptions of capillary penetration; typical
binders, wetters and dispersants used in the agrochemical industry show similar values, and are
not very different to that for water

Ingredient Surface tension (yLv) Contact angle (9) yLV cos 9

MorwetD425 48.3 61 23.4

MorwetEFW 31.2 O 31.2
PVP 35.2 33 29.5
Water (deionized) -79 68 26.2

dissolution time (s)

Geropon T36
Tamol PP
Morwet D42S
Morwet EFW
Disp-SS-Dry
PVP

disc thickness (I)

Figure 4.16 Pressed tablets of typical dispersants dissolve linearly with time.

parameter seems to be the rate at which the binders dissolve. Evidence for
this comes from an examination of pressed discs of the binder materials.
Figure 4.16 shows that the dissolution behaviour is essentially quite simple,
with linear plots of thickness against dissolution time.
If we correlate the dispersion times of larger granules with the relative
dissolution rates of the binders which were used to hold them together, then
a linear relationship is apparent. (One of the problems associated with the
dispersion times is that they are closely related to the size of the granules
used. To improve the accuracy of the test method, it can be convenient to
use larger granules where the size can be more accurately controlled. The
size used in these tests was 3mm.) The effect is shown for three different
starting powders (Figure 4.17). Note that as the powders had different
starting particle sizes, they display different gradients.
 Dispersion time (sj

China day (GTY)

Carbendazim
Talc ATExtra

Rate of Dispersant Dissolution (mm/min)

Figure 4.17 WG samples disperse according to the dissolution rate of the binder used to hold
them together. Three different powder samples were used to display the effect of particle size
and surface hydrophilicity.

It is well known within the agrochemical industry that the addition of

wetting agents can produce faster-dispersing granules. It was once thought
that this was a consequence of faster penetration of the pores. However, as
mentioned earlier, this is now thought to be incorrect. Examination of the
effects produced by wetting agents leads to the conclusion that very small
quantities are required. Figure 4.18 shows the effect of adding varying levels
of alkyl sulphosuccinate to a granule which contains a typical poly-
electrolyte (Morwet D425) as the principal dispersant.
Here the effect is apparent with a one-hundredth part addition. Where
10% of Morwet D425 is used, this would equate to a 0.1% formulation
addition of the wetter, clearly a highly effective way of reducing the disper-
sion time. The rise in dispersion time evident at higher loadings of alkyl
sulphosuccinate is associated with an increase in viscosity. Good wetting
agents, which act at the air-water interface, of necessity show poor, and
slow, water solubility. A solubilizing surfactant is required to enhance their
rate of dissolution and to prevent gelling. It is believed that the product
Morwet EFW acts in a similar fashion, although this has not been confirmed
by the manufacturer (Figure 4.19). In this instance the rate of dispersion has
been doubled by adding a one-tenth part of the wetter to the dispersant,
and it is clear that the gelling problem does not arise at higher loadings.
Morwet EFW is a common ingredient in many WG formulations, and
 dissolution rate (s)

[Aerosol OTB] %w/w

Figure 4.18 Dissolution rates of tablets prepared from a typical binder/dispersant benefit
greatly from the addition of a small amount of wetter, in this case 1 % w/w. Higher loadings
lead to viscosity problems.
Dissolution rate (s)

Morwet EFW cone (%w/w)

Figure 4.19 Dissolution rates of tablets prepared with a wetter which displays low viscosity at
high concentrations.
appears in many patent applications. The exact composition, however,
has not been disclosed, and as a result there has been little published
work which examines the behaviour of this interesting blend of ingredients
[121-126].
Elucidation of the effect brought about by wetting agents was made by
examining the dissolution of dispersant discs in solutions of different
wetting agents (G. A. Bell, presentation made to the Royal Society of
Chemistry, London, 6 April 1997). This showed that the dispersants dis-
solved more rapidly when dynamic surface tension reducing agents were
employed. Figure 4.20 shows a plot of dynamic surface tension against
dispersion time for large water-dispersible granules. A similar plot, with a
better correlation coefficient (0.97), has been measured for the dissolution
of dispersant discs in solutions of the dynamic wetters.
It would appear then that the wetting agent is able to increase the rate at
which the binding agent dissolves. The final piece of evidence which is
available to add to the current state of knowledge about granule dispersion
comes from an examination of the concentration behaviour of non-ionic
wetting agents which display good dynamic activity. Figure 4.21 shows that
the ability of the wetter to dissolve a block of dispersant begins at concen-
trations of about 0.5% w/w. Clearly this is well above the CMC (critical
micelle concentration) of the surfactant and it may be concluded that the
effect is genuinely a dynamic one.
Dispersion time (s)

DST (m N/m)
Figure 4.20 The most efficient wetting agents for a reduction of dispersion time for water-
dispersible granules are those which display low dynamic surface tensions.
 Dissolution rate mnVhr

Surfactant concentration %

Figure 4.21 The concentration of wetting agent required to reduce the dispersion time of a
water-dispersible granule is well above the CMC. This occurs because it is a dynamic effect.

4.4 Design: modern methods

The design of a water-dispersible granule, as with the design of most formu-

lations, is a balancing act to achieve the best properties of each ingredient,
without upsetting those of the other ones (Figure 4.22). The balance insofar
as granule size and particle size are concerned can be shown diagrammati-
cally with respect to four of the key physical properties (although note
should be made that the theoretically ideal position for dustiness has not
been verified practically). Similar conflicting requirements can be specified
for the ingredients used to formulate a WG (this problem was outlined by
H. T. Delli Colli of Westvaco Products, where a three-pointed star was
used to convey the conflict between dry strength, dispersion time and
suspensibility). The properties which are of concern tend to drive the for-
mulation chemist in opposite directions, and there appears to be little that
can be done about this, other than to reach a compromise.
The task of formulating a water-dispersible granule is much the same as
for any other type of product; it requires sample preparation and testing
followed by a good deal of iteration. Methods which have been employed to
try to simplify this problem include the use of techniques such as statistical
experimental design [127-131]. These techniques can be useful, especially
for process examination, but there are problems associated with this ap-
proach when applied to formulations.
 High crush strength Low (lustiness

Granule size

Fine suspension Fast dispersion

Particle size

Figure 4.22 Optimizing a formulation is complicated as different physical properties require

different combinations of granule and particle size.

One problem is that the ingredients in a formulation are not independent

of one another. This leads to the use of triangular diagrams rather than
square ones, which complicates the analysis of any response surfaces which
are found. In addition to this, the number of variables, even for a relatively
simple formulation, is large, which leads to an excessive number of surfaces
which need to be examined. Recently there have been attempts to shorten
this process, and although the results have not yet supplanted the standard
formulation method, they hold out a good deal of hope for the future. One
area in particular, the use of neural networks, is worth mentioning.
Expert systems have been available for many years and have been used
to dramatic effect in many industrial applications (e.g. [132]). One obvious
limitation, though, is that in order to set one up in the first place there is the
need for an expert to supply the rules of thumb and theoretical backing
material. In the area of agrochemicals there are many undoubted experts.
However, the diverse nature of the WG product and the range of ingredi-
ents available are serious limitations which those individuals have to
contend with. In addition to this, the cost and effort required to set up good
expert systems have so far prevented their common adoption.
Neural networks, on the other hand, are inexpensive and easy to set up,
being commercially available as Windows-driven software packages (e.g.
[133]). Mention was made earlier of the common occurrence of synergy.
Although it is often claimed that predictive packages can cope with this, it
is nevertheless an added complication which increases the difficulty of the
task.
Table 4.17 Examples of the ability of a commercial neural network system to predict the
physical properties of a WG formulation

Property Constraint Constraint met Predicted value Actual value

Dispersion time <75 s No 23 s 94 s

Wet sieve residue <0.1% No -0.1% 1.6%
Dispersed particle size <35um Yes 21.9 um 20.3 um
Dustiness <2.0mg/g Yes 1.9mg/g 1.8mg/g
Hardness >450g Yes 466 g 795 g
Tank residue <5% Yes 2.0% 3.6%

There does not appear to have been much work reported in the literature
on WG formulation using neural networks, although some work has been
done to test the system (G. A. Bell and S. E. Skelton, unpublished results,
Zeneca Agrochemicals, 1996). Data relating to the formulation of ten
samples of a WG containing an AI and three further ingredients (wetter,
dispersant and binder) were added to a commercial neural network pack-
age. The package was then asked to optimize the formulation, at which
point the optimized formulation was prepared. Table 4.17 shows the con-
straints which were made prior to computation, the predicted results and
the actual results which were measured practically.
The task which was set in terms of the criteria for success was not
impossible. In fact, one of the formulations fed into the database met all
of the requirements (except for dustiness, which was 2.2mg/g instead
of 2.0mg/g). The neural network formulation was placed roughly in the
middle of the list of formulations in terms of its overall performance. The
fact that the formulation did not meet all of the requirements, however, was
not seen as a total failure. It met six out of nine constraints and it was
remarkably accurate in predicting dispersed particle size and dustiness, two
key properties of the formulation. If network systems such as these can be
used to predict some of the physical properties, then they will undoubtedly
find a use in most laboratories across the world. At the present time,
however, this remains to be proven.

References

1. Kanehara, M. and Suzuki, M. (1960) Readily dispersible insecticide granules, Japanese

Patent JP 60 258101, American Cyanamid Company.
2. Littler, C. (1958) Herbicidal Pellets, US Patent 2992090, E.I. DuPont et de Nemours and
Company.
3. Martin, H. (ed.) (1968) Pesticide Manual, 1st edn, British Crop Protection Council.
4. Martin, H. (ed.) (1974) Pesticide Manual, 4th edn, British Crop Protection Council.
5. Belli Colli, H.T. (1988) Water dispersible granules and WDG technology, p. 48, paper
presented to ICI Agrochemicals as part of a series of seminars on lignosulphonates.
Copies available from the author c/o Westvaco Corporation, South Carolina, USA.
 6. Croze, E. and Gautier, M. (1993) Agrochemical compositions in the form of dispersible
granules, European Patent Application EP 0661000 Al, Rhone-Poulenc.
7. Graber, G. and Bossy, A. (1993) Solid agricultural concentrates containing two active
materials being a mixture of granules each containing one active component, French
Patent FR 2704387 Al, Rhone-Poulenc.
8. Gouge, S.T. and Shue, J. (1992) Soluble container for agricultural chemicals - comprises
outer soluble bag containing one chemical and inner bag containing different chemical,
International Patent WO 92/17381 Al, Rhone-Poulenc.
9. Dutzman, S., Feucht, D., Feyen, P. et al (1992) Solid agrochemical formulations for
plant protection - contain polyethoxyisotridecyl alcohol, alkylarylpolyethoxy ethanol
phosphate ester, W-alkyl lactam or tertiary amide, German Patent DE 4120694 A, Bayer
AG.
10. Chasin, D.G. and Davis, R.I. (1993) Solid free flowing adjuvant composition for use with
agrochemicals, California Patent CA 2093377A.
11. Biernger, H., Hackert, E., Henrich, R. et al (1992) Synergistic herbicide mixtures -
contain alkyl polyglycol ether sulphate surfactant and a leaf effective herbicide, German
Patent DE 4029304 A, Hoechst AG.
12. Fu, E. and Narayanan, K. (1993) Free flowing water dispersible agriculturally active
granules, International Patent WO 95/12974, ISP Investments Inc.
13. Lo, R. and Huang, C.W. (1994) Solid herbicidal compositions containing N-
phosphonomethyl glycine, US Patent 5514641, Zeneca Agrochemicals.
14. Kuchikata, M., Prill, E., Richardson, R. et al. (1988) Granular composition containing
glyphosate and a liquid surfactant, Australian Patent AU 635514, Monsanto Australia
Ltd.
15. Kuchikata, M., Prill, EJ., Richardson, R.O. et al (1988) Improved glyphosate formula-
tions, European Patent EP 378985 Al, Monsanto Company.
16. Arnold, K.A. (1992) Improved glyphosate herbicide formulation, European Patent EP
582561 Al, Monsanto Company.
17. Schaefer, T., Taagegaard, B., Thomsen, LJ. and Kristensen, H.G. (1990) Melt
pelletisation in a high shear mixer v. effects of apparatus variables. Eur. J. Pharm. ScL,
1(3), 133-41.
18. Schaefer, T., Taagegaard, B., Thomsen, LJ. and Kristensen, H.G. (1993) Melt
pelletisation in a high shear mixer iv effects of process variables in a laboratory scale
mixer. Eur. J. Pharm. ScL, 1(3), 125-31.
19. Sandell, L. and Wysong, R. (1993) Water dispersible agricultural compositions made by
heat extrusion, International Patent WO 95/08265 Al, DuPont et de Nemours & Co.
20. Sandell, L. (1992) Water dispersible pesticide granular formulations made by heat extru-
sion, European Patent EP 0501798 Al, DuPont et de Nemours & Co.
21. Albert, R.E. and Weed, G.B. (1970) Water dispersible pesticide aggregates, US Patent
3920442, DuPont et de Nemours & Co.
22. Geigle, W., Sandell, L. and Wysong, R. (1989) Water dispersible pesticide granular
formulations from heat activated binders, International Patent WO 91/13546 Al, DuPont
et de Nemours & Co.
23. Sandell, L. (1991) Water dispersible granular agricultural compositions made by heat
extrusion, International Patent WO 92/15197 Al, DuPont et de Nemours & Co.
24. Freeman, R. and Sandell, L. (1992) Water dispersible granular agricultural compositions,
International Patent WO 93/25074, DuPont et de Nemours & Co.
25. Freeman, R. (1993) Prep of melt granulation of compositions using dielectric heating,
International Patent WO 95/09044 Al, DuPont et de Nemours & Co.
26. Hill, A.C. (1993) Solid formulation, European Patent EP 0638235 Al, Shell
International.
27. Jaeger, H.-U., Boeckh, D., Broeckel, U. et al (1996) Granulation of amphiphilic polymer
for use as soil release agent in laundry detergent, German Patent DE 4441189S, BASF
AG.
28. Jang, C. and Pasarela, N. (1979) Solid composition of a pyrazolium salt, urea and a liquid
surfactant, US Patent 4289525, American Cyanamid Company.
29. France, P.A.R.G., Schmitt, J.C. and Wilkinson, L.P.D. (1993) Granular detergent compo-
sitions, European Patent EP 643130, Procter and Gamble Company.
30. Trinh, T. (1990) Use of polyalkyleneglycol compounds as carrier materials, US Patent
5384186 A, Procter and Gamble Company.
31. Fujimoto, M., Yasui, K. and Okochi, T. (1992) Stabilised agrochemical formulation,
Japanese Patent JP 06 263602 A, Sankyo Company Ltd.
32. Tsuji, K., Ogawa, M. and Shigenori, T. (1989) Preparation of emulsifiable pesticidal solid
compositions, UK Patent GB 2230954, Sumitomo Chemical Co.
33. Ogawa, M., Inoe, M. and Ootsubo, T. (1994) Herbicide compositions containing water-
insoluble solid organic compounds and polyoxyalkylene glycols, Japanese Patent JP 08
81302, Sumitomo Chemical Co.
34. Ogawa, M., Inoe, M. and Ootsubo, T. (1994) Microbiocide compositions containing
water-insoluble solid organic compounds and polyoxyalkylene glycols for paddy fields,
Japanese Patent JP 08 81303, Sumitomo Chemical Co.
35. Tagami, M., Ogawa, M. and Otsubo, T. (1992) Production of agricultural chemical
composition, Japanese Patent JP 06 345604, Sumitomo Chemical Co.
36. Nagai, H., Mizuno, T. and Tanaka, N. (1992) Pesticide granules containing surfactants
and water soluble binding materials, Japanese Patent JP 04 273801 A, Nippon Soda Co.
Ltd.
37. Kuroda, M. (1992) Sustained release composition for agrochemical containing oil absorb-
ing material, hydrophobic agent and active agent, useful for water soluble insecticide,
International Patent WO 95/09532 Al, Nippon Soda Co. Ltd.
38. Sawamura, M., Murai, J. and Fujita, S. (1994) Granules containing low-melting point
pesticides, Japanese Patent JP 08 92,004, Takeda Chemical Industries Ltd, Kumiai
Chemical Industry Co.
39. Kimmler, J. (1992) WG herbicide composition comprising dinitroaniline herbicide,
montmorillonite carrier and water swellable polymer, US Patent 5294594, American
Cyanamid Company.
40. Kimmler, J. and Kubisch, R. (1993) Water dispersible granular herbicidal compositions
comprising dinitroaniline and imidazolinone herbicides with montmorillonite carriers,
US Patent 5393731, American Cyanamid Company.
41. Doms, J.R.P., Gillis, M.J.E.G. et al (1993) Bentonite containing particulate fabric soften-
ing compositions, European Patent EP 570237 A2, Colgate Palmolive Company.
42. Kikuchi, H. and Sato, M. (1998) Granule preparations dispersible in water, Japanese
Patent JP 2 111620 A, Kunimine Kogyo KK.
43. Horie, H., Abe, S. and Fukutome, S. (1994) Manufacture of granular non-ionic detergent
compositions, Japanese Patent JP 07 286197, Lion Corp.
44. Matsumoto, N., Koko, T., Suzuki, K. and Shirai, Y. (1991) Herbicide WG from extrusion
of mineral fine powder, Japanese Patent JP 05 017303 A, Nissan Chemical Industries Ltd.
45. Jaeger, K.-F., Kilbride, T. Jr and Lisa, R.E. (1993) Solid containing mepiquat chloride,
International Patent WO 95/02963 Al, BASF AG.
46. Ploss, H. and Ingwersen, W. (1985) Water dispersible granulate containing a water
insoluble active ingredient, German Patent DE 3333872 Al, Norddeutsche Affinerie
AG.
47. Yoshimoto, H., Mizuguchi, M. and Yoshikawa, H. (1993) Inorganic porous
microparticles, Japanese Patent JP 07 173452 A, Suzuki Yushi Kogyo KK.
48. Ogawa, M. (1989) Water dispersible granules, UK Patent GB 2234678 A, Sumitomo
Chemical Co.
49. Sherrington, PJ. and Oliver, R. (1981) Granulation, Heyden and Son Ltd, London.
50. Bell, G.A. (1989) Herbicide granules - revue of processes and products, in Brighton Crop
Protection Conference - Weeds, 2, 745-52.
51. Holley, C.A. (1985) Water dispersible granule development process and equipment, in
Pesticide Formulations and Application Systems: Fourth Symposium (eds T.M. Kaneko
and L.D. Spicer), American Society for Testing and Materials, Philadelphia, PA, ASTM
STP 875, pp. 144-57.
52. Rettig, G. (1965) Granulation, its place in chemical engineering. Br. Chem. Eng., 10, 608.
53. Belcher, D.W., Smith, D.A. and Cook, E.M. (1963) Design and use of spray driers (I),
principles and application. Chem. Eng., 68, 83-8.
54. Belcher, D.W., Smith, D.A. and Cook, E.M. (1963) Design and use of spray driers (II),
design and costs. Chem. Eng., 68, 201-8.
55. Raff, A.M., Robinson, MJ. and Svedres, E.V. (1961) Spray-drying of tablet granulations
I; a preliminary report. /. Pharm. ScL, 50, 76-9.
56. Scott, M.W., Liebermann, H.A., Rankell, A.S. and Battista, J.V. (1964) Continuous
product of tablet granulations in a fluidized bed I; theory and design considerations. /.
Pharm. ScL, 53, 314-20.
57. FAO (1996) Manual on the development and use of FAO specifications for plant protec-
tion products, 4th edn.
58. Bell, G.A. (1990) The structure/physical property relationships of a model water
dispersible granule. Pesticide ScL, 29, 467-73.
59. Fu, E., Narayanan, K.S., Hall, F.R. and Downer, R.A. (1995) in Pesticide formulations
and Application systems; Vol. 14 (eds F.R. Hall, P.D. Berger and H.M. Collins), ASTM
STP 1234, Philadelphia, pp. 179-89.
60. Fu, E. (1992) Free flowing nondusting water dispersible granules, International Patent
WO 94/10841, ISP Investments Inc.
61. Bell, G.A., Formstone, C.A. and Skelton, S.E. (1994) The dustiness of water dispersible
granule formulations, poster presented at the 8th International Congress on Pesticide
Chemistry, Washington.
62. Dustiness estimation methods for dry materials, British Occupational Hygiene Society,
Technical Guide No. 4, Science Reviews Ltd.
63. Fujimoto, T., Nakayama, M., Ohta, Y. and Uesugi, H. (1992) Screw type granulator
having domed die plate - with openings through which moistened powder is forced by
curved extrusion blades, European Patent EP 491238 Al, Fuji Paudal Co. Ltd.
64. Benbow, J. and Bridgewater, J. (1993) Paste Flow and Extrusion Clarendon Press,
London.
65. Rumpf, H. (1958) Basic principles and methods of granulation I and II. Chemie-Ing
Techn., 30, 144-58.
66. Rumpf, H. (1970) Zur Theorie der Zugfestigkeit von Agglomeraten bei
Kraftubertragung an Kontaktpunkten. Chemie-Ing Techn., 42, 538-40.
67. Newitt, D.M. and Conway-Jones, J.M. (1958) A contribution to the theory and practice
of granulation. Trans. Inst. Chem. Eng., 36, 422-42.
68. Kiortsis, S. and Malamatarls, S.T. (1995) Powder wettability and agglomeration in the
granulation process. Biopharm. Pharm. Technol., 1, 139-40.
69. Wellm, A.B., Seville, J.P.K., Blewett, RJ. and Pomphrey, D.M. (1996) Factors influenc-
ing agglomeration in a high shear mixer, /. Chem. E. Res. Event, Eur. Conf. Young Res.
Chem. Eng., 2,1018-20.
70. Kristensen, H.G., Holm, P. and Schaefer, T. (1985) Mechanical properties of moist
agglomerates in relation to granulation mechanisms part 1 deformability of moist
densified agglomerates. Powder Technol., 44, 227-37.
71. Greissle, W., Graczyk, J. and Buggisch, H. (1993) Rheological investigation of suspen-
sions and ceramic pastes: characterisation of extrusion properties. Kona, 11,125-37.
72. Ghaly, E., Ruiz, S. and Nanette, R. (1996) Compressibility characteristics of matrices
prepared with ethyl cellulose aqueous dispersion. Drug. Dev. Ind. Pharm., 22(2), 91-
5.
73. Bohm, H. and Blackburn, S. (1994) Effect of mixing procedure on fine alumina paste
extrusion. Br. Ceram. Trans., 93(5), 169-77.
74. Fielden, K.E., Newton, J.M. and Rowe, R.C. (1993) The influence of moisture content on
spheronisation of extrudate processed by a ram extruder. Int. J. Pharm., 1-3, 79-92.
75. Wan, L.S.C., Heng, P.W.S. and Ling, B.L. (1995) Fluidised bed granulation with PVP K90
and PVP K120. Drug. Dev. Ind. Pharm., 21(7), 857-62.
76. Rowe, R.C. (1995) The rheological properties of lactose/microcrystalline cellulose/water
mixes - measurement using mixer torque rheometry. Pharm. ScL, 1, 547-9.
77. Landin, M., Rowe, R.C. and York, P. (1996) A comparison of the wet massing properties
of high and low density microcrystalline celluloses. Pharm. ScL, 2, 125-6.
78. Parker, M.D., Rowe, R.C. and Upjohn, N.G. (1990) Mixer torque rheometry: a method
for quantifying the consistency of wet granulation. Pharm. Technol. Int., September, 50-
64.
79. Cliff, M.U. and Parker, M.D. (1990) Scale-up of mixer granulators, in Proceedings of the
12th Interphex Conference, pp. 5/12-5/32.
 80. Raines, C.L., Newton, J.M. and Rowe, R.C. (1990) Pharmaceutical and biological ma-
terials with general rheology, in Rheology of Food (ed. R.E. Carter), Elsevier Applied
Science, London, pp. 248-57.
81. Rowe, R.C. (1996) Mixer torque rheometry: further advances. Pharm. Technol. Europe,
September, 38-48.
82. Baylac, G., Delalonde, M., Bataille, B. et al (1995) Development of the compresso-
rheometer for the granulation and shaping of moist powders, STP Pharma Prat., 5(6),
495-501.
83. Landin, M., Rowe, R.C. and York, P. (1995) Characterisation of wet powder masses with
a mixer torque rheometer. 3 Non-linear effects of shaft speed and sample weight. /.
Pharm. ScL, 84(5), 557-60.
84. Baylac, G., Delalonde, M., Bataille, B. et al (1996) The rheology of wet powders: a
measuring instrument, the compresso rheometer. Int. J. Pharm., 130(1), 147-51.
85. Rowe, R.C. (1996) Characterisation of wet powder masses with a mixer torque
rheometer. 4 Effect of blade orientation. Int. J. Pharm., 133(1-2), 133-8.
86. Rowe, R.C. (1996) Mixer torque rheometry further advances. Pharm. Technol. Europe,
September, 38-48.
87. Baert, L. and Remon, J.P. (1993) Influence of the amount of granulation liquid on the
drug release rate from pellets made by extrusion spheronisation. Int. J. Pharm., 95(1-3),
135-41.
88. Ritala, M. and Virtanen, S. (1991) Effect of binder solution quantity and lactose particle
size on granule properties. Ada Pharm. Nord, 3(4), 229-34.
89. Bleier, A. and Westmoreland, C.G. (1991) Effects of pH and particle size on the micro-
structure of alumina zirconia composites. /. Am. Ceram. Soc., 74(12), 3100-11.
90. Fersch, K. (1992) Granular water soluble or hygroscopic agricultural formulations and
methods of making them, International Patent WO 94/09627, BASF AG.
91. Lahr, W. (1993) Polymethacrylate microparticles containing active ingredient, German
Patent DE 4328069, Henning Berlin GmbH.
92. Woodburn, J. and Wedlock, D. (1993) Solid insecticidal formulation, European Patent
EP 0659341 Al, Shell International Research.
93. Ogawa, M., Tagami, M., Otsubo, T. and Tsuji, K. (1993) Agricultural chemical tablet with
pesticide, carbonate and solid acid, Japanese Patent JP 06 293603 A, Sumitomo Chemical
Co.
94. Ogawa, M., Tagami, M. and Ootsubo, T. (1993) Product of foaming type agricultural
chemical composition, Japanese Patent JP 07 010706 A, Sumitomo Chemical Co.
95. Kingery, W.D., Bowen, H.K. and Uhlmann, D.R. (1976) Introduction to Ceramics, 2nd
edn, Wiley, New York.
96. Bell, G.A., Formstone, C.A. and Skelton, S.E. (1994) The dustiness of water dispersible
granule formulations, poster presented at the 8th International Congress on Pesticide
Chemistry, Washington.
97. Parfitt, G.D. (ed.) (1973) Dispersion of Powders in Liquids, with Special Reference to
Pigments, 2nd edn, John Wiley, New York.
98. Washburn, E.W. (1921) Dynamics of capillary flow. Phys. Rev., 12, 374-5.
99. Rideal, E.K. (1922) The flow of liquids under capillary flow. Phil Mag., 44, 1152-9.
100. Zografi, G. and Yang, Y.-W. (1986) Use of the Washburn-Rideal equation for studying
capillary flow in porous media. /. Pharm. ScI, 75(7), 719-21.
101. Utz, C.G., Sekmistrz, J.M. and Hollis, R.P. (1995) Effects of non-ionic surfactants on
properties of water dispersible granule formulations, in Pesticide Formulations and Ap-
plication Systems, 14th volume (eds F.R. Hall, P.D. Berger and H.M. Collins), ASTM STP
1234, Philadelphia, pp. 191-200.
102. Parker, M.D., York, P. and Rowe, R.C. (1990) Binder substrate interactions in wet
granulation 1: the effect of binder characteristics. Int. J. Pharm., 64, 207-16.
103. Parker, M.D. and Rowe, R.C. (1991) Source variation in the wet massing (granulation) of
some microcrystalline celluloses. Powder Technol, 65, 273-81.
104. Parker, M.D. and Rowe, R.C. (1991) Binder substrate interactions in wet granulation 2:
the effect of binder molecular weight. Int. J. Pharm., 72, 243-9.
105. Eijiro, H., Akio, K., Kazumi, D. and Akinobu, O. (1995) Effect of binder characteristics
on the strength of agglomerates prepared by the wet method II. Chem. Pharm. Bull,
43(3), 488-91.
106. Sheskey, PJ. and Dasbach, T.P. (1995) Evaluation of various polymers as dry binders in
the preparation of an immediate release tablet formulation by roller compaction. Pharm.
Technol, 19(10), 98-112.
107. Cheong Wan, L.S. and Jeyabalan, T. (1987) Effect of binders on the formation of pellets
II. Polyvinylpyrrolidone. Chem. Pharm. Bull, 35(4), 1624-8.
108. Danjo, K., Kozaki, K.S. and Hisakazu Otsuka, A. (1994) Influence of the molecular
weight of binding agents on the physical properties of granules and tablets. Chem. Pharm.
Bull, 42(10), 2121-5.
109. Symecko, CW. (1995) The functional evaluation of binders in the wet granulation
process. Dissertation Abstracts International, 55(12), 5299B.
110. Narayanan, K.S., Ianniello, R.M., Benning, MJ. and Goehner, R. (1995) Value added
binders for solid formulations via polybasic acid complexation with vinyl pyrrolidone
homo and copolymers, in Fourth International Symposium on Adjuvants for
Agrochemicals (ed. R.E. Gaskin), 3-6 October 1995, Melbourne, pp. 42-7.
111. Rowe, R.C. (1992) Interactions in powders and granules - a reappraisal. Int. J. Pharm.,
79, 257-61.
112. Morse, T. (1979) Handbook of organic additives for use in ceramic body formulation,
Montana Energy and MHD Research and Development Institute, Montana.
113. Symecko, C.W. and Rhodes, C.T. (1995) Binding performance of pregelatinised starches.
Biopharm. Pharm. Technol., 1,145-6.
114. Schroeder, M. and Kleinebudde, P. (1995) Structure and pharmaceutical properties of
pellets, effect of the granulation liquid. Biopharm. Pharm. Technol., 1, 341-2.
115. Schultz, P. and Kleinebudde, P. (1995) Determination of pellet friability by use of an air
stream apparatus. Pharm. Ind., 57(4), 323-8.
116. Reuter, K.A. and Baum, S. (1988) Evaluation of various dust determining methods,
poster presented at the 7th International Congress on Pesticide Chemistry, Hamburg.
117. Gladden, L.F. (1993) Trans. Inst. Chem. Eng., 71(A), 657-74.
118. Attwood, D. and Florence, A.T. (1983) Surfactant Systems; Their Chemistry, Pharmacy
and Biology, Chapman & Hall, London, p. 390.
119. Kajikawa, A. (1990) Improved formulations having enhanced water dissolution, Euro-
pean Patent EP 394 211 Al, Monsanto Company.
120. Myers, D. (1988) Surfactant Science and Technology, VCH Publishers, New York.
121. Ogawa, M., Tagami, M. and Ootsubo, T. (1993) Granular agrochemical chemical compo-
sition - AI which is solid at RT, sulphosuccinic diester type surfactant and other
surfactant, Japanese Patent JP 07 126106 A, Sumitomo Chemical Co.
122. Daratech Pty Ltd (1992) Solid water dispersible trifluralin formulations - supported
inside porous carrier particles to give stability of size and orange polymorphic form,
International Patent WO 94/21121.
123. Johnson, J.L. and Kimmler, J. (1993) Herbicidal extruded granular compositions, Euro-
pean Patent EP 0623281 Al, American Cyanamid Company.
124. Kimmler, J. and Kubisch, R. (1993) Water dispersible granular herbicidal compositions
comprising dinitroaniline and imidazolinone herbicides with montmorillonite carriers,
US Patent 5393731, American Cyanamid Company.
125. Fersch, K. (1992) Granular water soluble or hygroscopic agricultural formulations and
methods of making them, International Patent WO 94/09627, BASF AG.
126. Sandell, L. and Wysong, R. (1993) Water dispersible agricultural compositions made by
heat extrusion, International Patent WO 95/08265 Al, DuPont et de Nemours & Co.
127. Kristl, A., Mrhar, A. and Kozjec, F. (1994) Physicochemical evaluation and statistical
assessment of wet granulation. Drugs Made in Germany, 37(3), 78-82.
128. Ogawa, S., Kamajima, T., Miyamoto, Y. et al. (1994) A new attempt to solve the scale up
problem for granulation using response surface methodology. /. Pharm. Sd., 83(3), 439-
43.
129. Iskandrani, B., Clair, J.H., Patel, P.S. et al (1993) Simultaneous optimisation of capsule
and tablet formulation using response surface methodology, Drug Dev. Ind. Pharm.,
19(16), 2089-101.
130. Rowe, R.C. (1996) Using experience in product formulation. Manufacturing Chemist,
March, 20-1.
131. Maggi, L., Bonfanti, P., Santa, P. et al (1996) The suitability of a small scale high-shear
mixer for powder pelletizaton. Pharm. Technol Europe, 8, 82-90.
132. Lay, S., Podazeck, F., Newton, J.M. and Daumesnil, R. (1996) An expert system to aid the
development of capsule formulations. Pharm. Technol Europe, 8, 60-8.
133. Colbourn, E.A. and Rowe, R.C. (1996) Modelling and optimisation of a tablet for-
mulation using neural networks and genetic algorithms. Pharm. Technol. Europe, 8, 46-
55.