i ANIMAL FEED

SCIENCE AND
TECHNOLOGY
;
ELSFYIER Animal Feed Science Technology 6 1 ( 19%)89- 1I2

Physical quality of pelleted animal feed

1. Criteria for pellet quality
M. Thomas *, A.F.B. van der Poe1
Wugeningen Agricultural University, WIAS, Wageningen Institute ofAnimal Science, Department of Animul
Nutrition, Murijkeweg 40, 6709 PG. Wugeningen, Netherhnds

Accepted I December 1995

Abstract

In a series of three articles, the physical quality of pelleted animal feeds is discussed from an
engineer’s point of view. In this first manuscript an overview is given on the binding mechanisms
in pelleted animal feeds. Principles and methods for evaluation of physical quality of pelleted
animal feeds are reviewed with respect to pellet hardness and durability. Methods are outlined
with respect to quality standards both from a pragmatical and scientific point of view.
It is concluded that binding in pellets most probably is due to solubilisation and subsequent
crystallisation of feedstuff components e.g. starch, sugars, fats or ‘liquid necking’. Liquid necking
is a binding mechanism which uses the surface tension of water, in a three-phase system of air,
water and particles to maintain structural integrity of the pellet. Soluble components might be
introduced in the feed mash subjected to pelleting. They are either incorporated in the mixing
phase or result from processing as a function of processing variables during the subsequent stages,
conditioning, pelleting and cooling/drying, of the feed manufacturing process. To evaluate the
physical quality of pelleted feeds, generally a subdivision is made into tests that evaluate
‘hardness’ and tests evaluating ‘durability’ of a given pellet. Several devices measuring fragmenta-
tion strength, and devices determining abrasion strength of pellets are discussed. It is concluded
that some tests presently available evaluate a mixture of hardness (fragmentation) and durability
(abrasion) effects. The feed manufacturer or feed technologist should be aware of the reason for
evaluating the pellet quality and subsequently choose the appropriate, most suitable method, since
no unique test exists that covers all parameters of interest related to physical quality of pelleted
animal feeds.

Keywrk Pelleting; Pellet quality

* Corresponding author.

0377~8401/%/$15.00 Copyright 0 1996 Elsevier Science B.V. All rights reserved.

PII SO377-8401(96)00949-2
90 M. Thomas, A.F.B. vun der Poe1 / Animul Feed Science Technology 61 (1996) 89-112

1. Introduction

Animal feed manufacturing involves the use of a variety of raw materials to produce
compound feeds. The feeds are defined according to certain specifications with regard to
nutritive composition based on specified descriptions for nutritional, hygienic and
physical quality. Together, these specifications require knowledge of a vast number of
different properties of ingredients to optimise processing while maintaining or control-
ling nutritional quality for a given feed form. Therefore, concerted action by the
disciplines ‘nutritional science’ and ‘feed science and technology’ may nowadays be
essential for further progress in livestock production.
Animals which receive pelleted feeds generally have higher performances in terms of
average daily gain and lower feed conversion compared with mash feeds as reported for
pigs (Vanschoubrcek et al., 1971; Pond and Maner, 1984) and poultry (Calet, 1965;
Quemere et al., 1988; Moran, 1989). Feed processing influences rate of degradation and
rate of passage of feed components of ruminants (van der Poe1 et al., 1995). For
instance, pelleting reduces the resistance of starch against ruminal degradation by about
15% (Tamminga and Goelema, 1995). Nielsen (1994) showed that expander processing
reduced the effective protein degradability (EPD) of raw materials by on average 8%.
These effects may lead to differences in milk production and composition. It can
therefore be argued that nutritional value is both influenced by the raw materials used
(van Rooy, 1986) and by processing conditions during operation (Skoch et al., 1981,
Skoch et al., 1983).
Different animal species require different physical properties for their respective
feeds. This means that different quality standards are used. For example, for mouse feed
(Koopmans et al., 1989a, Koopmans et al., 1989b) or broiler feed (Moran, 1989)
different qualities are required. For fish-feeds, additional pellet characteristics such as
flowability, sinking velocity, water absorption and water solubility are important.
Furthermore, the hygienic quality of feeds is important. Hygienic quality involves the
control of microbiological contamination of feeds based on levels of enterobacteriaceae
and salmonella (McCapes et al., 1989).
Feed processing includes the treatment (physical, chemical, thermal) of a feed prior to
consumption by animals (Maier and Bakker-Arkema, 1992). In general, the variability in
processing effects is associated with the choice of equipment, with processing conditions
as well as with the processing system, e.g. the combination and sequence of processing
equipment (Melcion and van der Poel, 1993). Therefore, processing may involve a
simple process such as blending in the form of mash or processing can be much more
complicated, such as (double) pelleting, crumbling or when an extruder or expander is
used (Pipa and Frank, 1989; Veenendaal, 1990; Van Zuilichem and van der Poel, 1993).
It should be noted, however, that a specific piece of equipment per se is limited in its
application and cost-effectiveness. Generally, the specific equipment is used as a
component of a process system where the efficacy of this equipment can only be
optimised by careful consideration (Tran et al., 1991; Das et al., 19931, taking into
account the upstream and downstream processing.
In routine animal feed manufacturing raw materials are blended and ground to obtain
mash feeds that are further subjected to some form of steam and/or water conditioning
 M. Thomas, A.F.B. von der Peel/Animal Feed Science Technology 61 (1996) 89-112 91

before e.g. granulation is applied. Granulation and its products (granulates, pellets)
offers a product form that has many advantages over meal and that is why its application
is widespread for both raw materials, feed additives and animal feeds. Some advantages
(Rumpf, 1958; Friedrich and Robohm, 1969; Vanschoubroek et al., 1971) are:
* Pellets have better flow properties, necessary for good transport in conveying
equipment, and (gravitational) discharging behaviour from silos, than the meal they
were prepared from.
. The bulk density of pellets is generally higher than that of meal, so that more tonnage
can be carried by truck.
* The composition of the pellets as obtained from carefully blending and mixing
ingredients remains fixed, no segregation of e.g. additives occurs.
However, the use of shaping equipment, like pellet presses, requires additional costs
in terms of energy demand and necessary additional equipment; like a boiler to generate
steam, conditioning equipment, a pellet press and a cooler. In light of the fact that after
mixing and blending already a complete diet in the form of mash exists, additional costs
for investment in pelleting equipment should be considered with respect to the gain that
can be achieved due to incorporation of the pelleting process.
Making up mash in the form of pellets originates from almost 60 years ago, when
Patton et al. (1937) defined the nutritional superiority of pelleted poultry feeds over
mashes. Pellets are obtained by a pressing process in animal feed manufacturing. In
former times, the mash was simply pressed between two rollers or in a cake press to
obtain pellets/cakes without any preliminary treatment. Relatively low pressures were
used and the feed did not heat up (Calet, 1965). In modem feed mills the mash is
pelleted in a so-called roller-and-die pellet press, both vertically and horizontally. Before
entering the pellet press, the mash is subjected to some form of pretreatment before
granulation such as mixing with molasses or fats (Beumer, 1978, Beumer, 1980a,
Beumer, 1980b, Beumer, 198Oc, Beumer, 1980d1, conditioning with steam (Skoch et al.,
198 1) or being subjected to the use of an expander (Veenendaal, 1990; Pickford, 19921
to increase temperature or moisture level (Friedrich and Robohm, 1969; Maier and
Bakker-Arkema, 1992). When steam is used, the temperature of pellets after leaving the
die is generally higher in comparison with that of the conditioned meal due to the
frictional heat in the die. Finally, pellets are cooled with ambient air.
The physical quality of feed pellets is important for a number of reasons. First of all.
transportation and handling in both the factory and on the farm require pellets of a
certain integrity without fines produced by attrition stresses. Pellets of high physical
quality must have properties which give a high nutritional quality for example in terms
of higher feed intake and, perhaps, improved nutritional value (Skoch et al., 1983;
Stevens, 1987; Koopmans et al., 1989a, Koopmans et al., 1989bl.
Pellets also need to have a basic form of physical quality in terms of e.g. hardness
and durability to withstand the rigors of transportation. Hardness is the force necessary
to crush a pellet or a series of pellets at a time; durability is the amount of fines
returning from pellets after being subjected to mechanical or pneumatic agitation. Such
quality parameters can also be used to evaluate the effects of diet formulation,
conditioning, expander treatment, pellet binders, die selection, etc. (Pfost, 1963).
For optimisation of product quality in terms of physical characterisation, knowledge
92 M. Thomas, A.F.B. van der Poe1 / Animal Feed Science Technology 61 (19%) 89-112

of fundamentals for aggregating particles of a different size, hardness and shape is

needed. Therefore it is important to understand how particles bind and to gain insight
into binding properties and binding mechanisms in granulates or pellets and their
behaviour during transportation and storage. Finally, this will reveal information that can
be used as a basis for the criteria used to evaluate pellet quality.
The objective of this series of studies is to describe the contribution of causative
factors that contribute to the physical quality of animal feed pellets. In this first article,
criteria for pellet quality are evaluated. An overview is given of the binding mechanisms
involved in pelleted feeds. In addition, different devices are discussed that are used in
the compound feed industry to evaluate the hardness and durability of the pellets. Some
general recommendations are given on what should be taken into account when using
the different methods with which pellet quality can be evaluated.
In a second article and third article (Thomas et al., 1996a,b), the physical quality of
pellets is discussed in view of the contribution of raw materials and their components to
the desired properties of pelleted animal feeds. Furthermore, the effect of unit operations
in preparing animal feeds (conditioning, pelleting and cooling/drying) and some of their
inherent process variables are discussed.

2. Binding of particles

Mash feeds are often subjected to granulation (pelleting) in routine feed manufactur-
ing. The subsequent stages in the pelleting process are meal conditioning, pelleting and
drying/cooling. Meal conditioning is a prerequisite for the actual compression of the
meal into pellets and can be controlled by its process variables such as temperature, time
and moisture level. Proper conditioning, therefore, brings adhesive properties on the
surface of meal particles thereby improving pellet quality in terms of a harder pellet or a
reduction of fines produced from pellets during additional handling (Friedrich, 1977;
Skoch et al., 1981).
Various hypotheses have been postulated in order to clarify the factors that determine
the structural integrity in a pellet (Rumpf, 1958; Knacke and Pohl, 1959; Friedrich,
1964b).
In general, causative factors that affect pellet quality are diet ingredient composition
and its properties, process technology and specific pellet binders (Table 1). More
fundamentally, pellet integrity can be examined by studying the binding of particles that
is accomplished through solid-solid bonds between diet ingredient particles, the use of
liquids (e.g. molasses) or the use of specific pellet binders. The best hypothesis to date is
probably the theory of Rumpf (19581, extended by Friedrich (1964b) and Friedrich
(1977), who described causative factors by which feed particles are held together by
various mechanisms (Fig. 1).
The genera1 mechanisms for binding feed particles can be divided into ‘solid-solid’
interactions between particles, capillary forces in a three-phase system of water, air and
solid material, so-called ‘liquid necking’, adhesive and cohesive forces between particu-
lates and binders and interactions between particles due to folding and plying.
‘Solid-solid’ interactions may consist of sintering, recrystallisation or crystal growth
 M. Thomus, A.F.B. eon der Poel/Animol Feed Science Technology 61 (19%) 89-112 93

Table 1
Causative factors affecting pellet quality
Factor Dimension References
Diet ingredient composition
Physical
Part. size (distribution) Rumpf ( 19581,Stevens ( 1987). Mercier and Guilbot ( 19741,
Payne ( 1978)
Specific density kg m-’ Friedrich (1964a)
Bulk density kg m-’ Friedrich ( 1!%%a),
MacMahon and Payne ( 199I ), TeSiC( 1977)
Angle of repose 0 Mohsenin ( 1986). Friedrich and Robohm ( 1969)
Surface area m* Knacke and Pohl ( 1959). Friedrich and Robohm ( 1969)
Chemical
Moisture g kg-’ Skoch et al. (19831, Knacke and Pohl(1959)
Ether extract g kg-’ Salmon ( 19851,Richardson and Day (1976).
Friedrich and Robohm ( 198ICI
Crude fibre g kg-’ TeHiC( 19771,Friedrich and Robohm ( 1981a),
Friedrich and Robohm ( 1981b)
Crude protein g kg-’ Friedrich and Robohm ( 1981a), Friedrich and Robohm ( 1981bl,
Stark (1990)
Ash g kg-’ Friedrich and Robohm (1981a). Friedrich and Robohm ( 1981b)
Functional
Viscosity Pa s Nissinen et al. ( 19931,Keller ( 1983)
Protein solubility % Hermansson ( 1979), Kinsella ( 19791,Wood (1987)
Starch gelatinisation % Wood (19871, Smith (1983), Heffner and Pfost (1973)
Diet ingredients % MacMahon and Payne (19911, Israelsen et al. (1981).
Payne ( 19781
Process technology
Conditioner Leaver (19841
TemperaNte OC Skoch et al. (1983). Winowiski (19881, Stevens (1987)
Time S Israelsen et al. (1981). Mercier and Guilbot (1974).
Beumer ( 1978). Beumer ( 198Oa),Beumer ( 1980b).
Beumer ( 198Oc1,Beumer ( 198Od)
Moisture
steam % Melcion et al. (19741, Maier and Gatdecki (1992).Stevens ( 1987)
Water g kg-’ Friedrich and Robohm ( 1970)
Pellet press die specification
Diameter to length ratio Stevens ( 1987), Schwanghart ( 1%9>, Schwa&art (19701,
TeSif ( 1977). Drevet ( 1972). Hamahan ( 1984)
Cooler/drier
Air speed ms-’ Friedrich and Robohm (1%8)
Time s Maier and Bakker-Arkema ( 1992)
Air moisture level gkg-’ Schwanghart (1970)
Pellet layer thickness m Maier and Bakker-Arkema (1992)

Pellet binders
Adhesion and cohesion between particles Van Zuilichem et al. ( 1979aI. Van Zuilichem et al. (1979b1,
Van Zuilichem et al. (1980). Winowiski (1988).
Brirggeman et al. ( 19641,Eberhardt ( 1964). Payne ( 1978)
94 M. Thomas, A.F.B. vun der Poel/Animul Feed Science Technology 61 (1996) 89-112

Fig. 1. Model figure of general binding forces between two particles (modified after Rumpf, 1958 and
Friedrich, 1977).

of some ingredients, chemical reactions, melting of thermoplastic materials and solidify-

ing afterwards into a crystalline state (Rumpf, 1958). Bonds between particles of the
solid-solid interaction type are established mainly during the drying/cooling process
depending on the applied conditions.
In a porous agglomerate, such as pellets, three separate phases can be distinguished:
water, air and the solid material from the particles. These particles can be held together
by liquid necking. The force of the bond is dependent on the surface tension (Ap) of the
binding liquid y (N m- ’>, and the radius ( r) of the neighbouring particle (m> (Rumpf,
1958). The binding pressure of the bond (Nl is given by the equation of Laplace
Ap = 2 X (y/r), in the case where the radii of the neighbouring particles are equal. In
pelleted feeds this is normally not the case, since milling induces a particle size
distribution (with subsequent larger and smaller particles) of the diet ingredients. From
the above equation it can be easily derived that with decreasing radius of the particles,
binding strength becomes higher. This is in line with the generally accepted principle
that with finer grind, better pellets are produced (Payne, 1978). The binding agent
(water) between the different particles can be redistributed around the particles without
loss of the established bond. Elevated temperatures cause redistribution of water via
evaporation and condensation and this will increase the mobility of water between
particles (Friedrich and Robohm, 1968). In the case where pellets’ moisture level is
decreasing, then the moisture bridges will shrink and total binding forces will decrease
due to a smaller number of moisture bridges. However, force per bond will increase
since water in the larger capillaries is evaporated first. Subsequently, it follows that the
remaining water establishes bonds between smaller particles. From the equation given it
follows that binding force between neighbouring particles is stronger with decreasing
radius. In the case where large quantities of water fill all pores, there will be a two-phase
system of water and particles, with no capillary force present to maintain structure in the
pellet. It is shown by Knacke and Pohl(1959) that an optimum exists in binding strength
of clay agglomerates depending on the amount and surface tension of the water added.
Ingredients with (high) viscous properties will ‘stick’ particles together. Therefore
distances between particles will decrease while at the same time the interacting surfaces
 M. Thomus. A.F.B. van der Peel/Animal Feed Science Technology 61 11996) 89-II2 95

between particles increases. Solid-solid interactions between different particles may

come into effect when the distance between particles is sufficiently small, for instance,
when pressure during pelleting is applied. Under such conditions, binding agents cannot
be redistributed around the particles. When these bonds are present in large amounts in
the feed, time will affect the binding between particulates and the structural integrity of
the pellets may change with storage time (Pfost and Young, 1973).
‘Van der Waals’ forces may play a role in building up structural integrity in the pellet
when particles are sufficiently small ( < 50 pm, de Jong, 1995; < 60 km, Rumpf, 1958
and Friedrich, 1964b). In general, in animal feeds the fraction of these particles is too
negligibly small to substantially increase binding strength in pellets. When particle size
increases, the effect of ‘Van der Waals’ forces diminishes and other binding mecha-
nisms (capillary forces and binder materials) become more pronounced.
Electrostatical forces are negligible between particles due to repulsional forces
between materials. On theoretical grounds the magnitude of the binding forces between
particles is 0.2905 smaller than with ‘Van der Waals’ forces (Rumpf, 1958).
Finally, particles and fibres may be folded due to the pelleting process and thus plied
around each other (Rumpf, 1958). This phenomenon will also aid in building structure in
the pellet. However, up until now no theoretical framework has been developed to
quantify this type of binding (Rumpf, 1958). In Thomas et al. (1996a) some of the
effects of raw material components, for instance starch, protein and fibre, are discussed
with respect to their effects on physical pellet quality.
‘The above proposed mechanisms are useful to study the structural integrity of pellets.
Which of the different above mentioned effects is the most causative factor for pellet
integrity, depends on the combination of raw materials used and the different
operating/processing variables during pelleting, cooling and storage. All these variables
affect pellet integrity, the magnitude of the different binding mechanisms being depen-
dent on the type of feed produced with its constituents and physical properties.
For quality control reasons, pellet quality devices have been developed to measure
pellet integrity and, again, these devices can be further used to examine the effects of
diet formulation and of process operation conditions on quality of pelleted animal feeds.

3. Pellet quality criteria and devices

When forces are involved that are statical or are present as a consequence of
movements (dynamic forces), handling of cooled pellets causes attrition. Static forces
for example may occur during bin storage while dynamic forces are present during
screw or pneumatic transportation or during the filling of a bin. Attrition of feed pellets
comprises two phenomena, fragmentation and abrasion, both of which have conse-
quences for its particle size distribution (Fig. 2) and bulk density.
Fragmentation involves the fracture of pellets into smaller particles and fines at the
fracture area. Fragmentation causes a small shift in the particle size distribution but will
cause only a limited increase in bulk density or specific surface (de Jong, 1993). The
stresses required to overcome the cohesive and adhesive forces in the cracktip and the
amount of energy necessary, is dependent on the direction of the applied stresses, length
96 M. Thomas, A.F.B. vun der Poe1 / Animal Feed Science Technology 61 (1996) 89-l 12

Diameter Diameter

Fig. 2. ‘Theoretical effects of attrition of pellets @ragmentation and abrasion) on the particle size distribution
curves (after de Jong, 1993).

of the crack and material dependent characteristics. The relation between these quantities
has been described in detail by Luyten et al. (1992).
Abrasion involves the fracture on the edges or surface-unevenesses of particles. This
type of attrition affects the particle size distribution. The fines produced are smaller than
the original pellet size and form an additional peak in the particle size distribution curve
(Fig. 2). Depe nd’ mg on the amount of fines produced, bulk density may increase, since
small particles occupy space in the large voids in between pellets, thereby decreasing
pellet porosity.
Fragmentation of inhomogeneous materials always starts near the point of inhomo-
geneities due to the fact that local stresses and strains are always highest near such
imperfections. Larger defects cause fracture to occur at lower overall stresses and
strains. In homogeneous materials such imperfections can be errors in crystalline
structure, or small holes in or just below the surface (Luyten et al., 1992; de Jong,
1993). When dealing with pellets, by nature of the material, particles of different sizes,
hardnesses and various shapes are aggregated. It is at these cracktips that stresses will
accumulate. When the local stress at the tip of a crack becomes higher than the cohesive
or adhesive stresses, the crack (defect) in the material starts to grow, fracture starts and
ultimately the material will fall apart. Fracture will propagate spontaneously as the
deformation energy released is, at least, equal to the energy needed to create new
surfaces.
Attrition can be brought about by tension, compression and shear stresses, compres-
sion being most important when related to animal feeds. Shear and tension tests may
give much information but are often more difficult to perform (Luyten et al., 1992). In
practice, therefore, more emphasis has been put on compression devices. The actual
 M. Thomus. A.F.B. van der Poel/Animol Feed Science Technology 6/ (1996) 89-112 97

assessment, however, when pellets are subjected to attrition by means of different

devices, is derived from properties of the more fundamentally desired quantity ‘energy’
needed for fracture or abrasion of pellets. Actual routine assessments are often expressed
just as ‘kg force’ after simulation of impact devices (mainly fragmentation) or as
‘percent fines returned’ after simulation of transport (mainly abrasion).
Several types of empirical devices for the routine evaluation of pellet quality have
been developed (Melcion and Delort-Lava], 1981). In general, these routine devices test
the hardness of pellets or their durability or resistance to attrition stresses (Pfost, 1963).
Some important devices are discussed in the following paragraphs with the often used
quantities ‘hardness’ and ‘durability’ as pellet quality parameters.

3.1. Hardness

Hardness is a quantity which is important for the nutrition of animals since hardness
may play a role with preference of animals (Skoch et al., 1983). Also, availability of
nitrogenous components for intestinal absorption has been reported to be affected by
hardness (CuperloviC, 1973; CuperloviC et al., 1973).
In the case of statical pressure one may simulate the forces on the pellets. The statical
pressure is caused for example by the weight of the pellets on top of the lower ones. In
wide silos where a part of the pressure cannot be relieved to the walls of the silo,
breakage of the pellets on the bin bottom may occur.
Studies of Knacke and Pohl (1959) using clay minerals showed effects of water
content on hardness of the agglomerates. The hardness of the agglomerates is propor-
tional to the surface tension of the pelleting fluid and inversely related to the particle
size distribution. They showed that with an increase in the filling of the pores with
water, an optimum value for hardness exists. This optimum is dependent on the surface
tension of the pelleting fluid.
Hardness is determined by using equipment which measures the force needed to
fragment a pellet. In general, one can distinguish between tension, compression and
impact based devices in which the compression component is the most important one.
Test devices for hardness have been developed for both scientific objectives and for
on-line application (Anonymous, 1992) in routine animal feed manufacturing.
Early devices for testing hardness were developed by McCormick and Schellenberger
(1960) and Young (1962). Young used a ‘model handling system’ in which pellets could
be subjected to typical actions encountered in the handling of pellets. This model system
was further used as a control for comparison of various other test devices (see Pfost,
1963) such as the Farmhand tester, tumbling can device and the Stokes hardness tester.
Nowadays, several devices are available for the evaluation of product hardness, each
of which have different attrition or operating mechanisms (Table 2).
A first and common device used in industry to test pellet hardness is the ‘Kahl’
device (Fig. 3), analogous to the early developed, manually used Stokes tester. In the
Kahl device, a pellet is inserted between two bars, and by increasing statical pressure
applied by means of a spring, the force needed to crack the pellet is determined. The
average of ten measurements is referred to as the ‘Kahl-hardness’ of the pellet.
Table 2
Generally applied test devices for pellet hardness, their attrition mechanisms and operation
Principal force applied at Attrition mechanism Operation a Force Device example Reference
One point Fragmentation Compression(axial tension) Static Kahl tester Melcion and Delort-Lava1 (I 981)
One point Fragmentation Compression(axial tension) Dynamic Schletmiger Beumer and Vooijs (I 993)
One point or whole pellet Fragmentation Impact Dynamic Pendulum Jindal and Mohsenin ( I9761
Variable Fragmentationand/or shear Compression(radial tension) Dynamic lnstron Luyten et al. (I 992)
Variable Fragmentationand/or shear Compression/shear Dynamic Kramer shear press Anonymous ( 1970)
(axial or radial tension)

’ Depending on standardconditions.
 M. Thomus, A.F.B. uan der Peel/Animal Feed Science Technology 61 (1996) 89-112 99

+Sample pellet

Force direction and sample-

pellet position

Fig. 3. Kahl pellet hardness tester.

Up-to-date devices for Kahl hardness imply the use of an automated version to generate
the statical pressure.
The Schleuniger test apparatus comprises a steel moving ram with a width of
approximately 2 mm of the measuring part. Individual pellets are placed between the
moving ram and a flat anvil. The moving ram will be pushed against the pellet by an
electrically driven spindle with an increasing force. The force needed to fracture the
pellet is recorded by a force transducer and registered. Measurement of Schleuniger
hardness is normally performed ten times and is expressed in (kilojpounds.
A further device which can be used to determine pellet characteristics is the
pendulum (Fig. 4). The pendulum is an impact resistance testing device, by which the
amount of energy needed to fracture a pellet can be investigated. The basic principle of a
pendulum impacting device consists of a swingarm with a sample holder. From a known
height this arm with the clamped on material of interest is dropped against a non-defor-
mable and heavy block. The physical characteristics of the material under investigation
that are recorded, are:
- Angular position of the pendulum arm before and after impact.
- Peak deceleration or force due to impact.

Turning point

_ Pendulum beaming

Pendulum rod

Hammer

Sample pellet

Fig. 4. Pendulum pellet hardness test device.

100 h4. Thomas, A.F.B. van der Poe1 / Animal Feed Science Technology 61(19%) 89-112

._(->..____.).pad cell

05
Fig. 5. Universal compression test device.
.

?? The duration of the contact.

. The area of contact.
Using a pendulum type impacter several damage resistance characteristics of the used
materials can be determined such as hardness, dynamic yield pressure and dynamic
elasticity modulus (Jindal and Mohsenin, 1976). Not much research has been conducted
in which physical quality of feed pellets is determined using a pendulum. The pendulum
impact test device was used by Van Zuilichem and Stolp (1976) to asses the hardness
(mechanical strength) of extruded pellets of maize grits and extrusamyl; a potato starch
derivate.
Other devices use both compression and tension, and comprise so-called ‘universal
tension and compression’ apparatus (Instron; overload dynamics). Their main advantage
is the accuracy with which the different materials can be subjected to the different tests
known in the field of feed and food engineering. These devices consist of a fixed plate
containing a load cell and a moving bar with variable speed. On this instrument
measuring bodies of a various geometry (either knife or plate for example) can be fitted
and uniaxial compression, uniaxial tension, three-point bending and cutting experiments,
respectively, can be performed (Luyten et al., 1992). For the testing of pellets, uniaxial
testing (plate) and cutting (knife) experiments are most appropriate. The force units
needed to break the pellet (or a series of pellets) is recorded as a function of time. Both
single (see example Fig. 5, plate) and multiple sample pellet(s) can be tested.
The Kramer shear press (Anonymous, 1970) consists of a hydraulic ram on which a
measuring body with various blades is mounted (Fig. 6). The test material is inserted at
the grid bottom of the shear box. When the ram comes down, the material is compressed
and sheared between the intermeshing blades and grid. The force units are recorded as a
function of time and from this graph the maximum force can be read and the energy
necessary for shear can be calculated. The force exerted on the material is recorded by
means of a proving ring and transducer. The speed of the hydraulic ram can be adjusted
over a relatively limited range and is in the same order of magnitude as with the Instron.
In Table 3, some data on hardness and durability of barley pellets are given as
determined using different devices. Pellets of barley were manufactured using a Labora-
tory CPM pellet press and two dies. Different temperature conditions, by the aid of
 M. Thomas, A.F.B. van der Peel/Animal Feed Science Technology 61 (1996) 89-112 101

Force direction
pellet position:
and sample

nib Hydraulic cylinder

r
1.

Fig. 6. Kramer shear press for measurement of pellet hardness.

steam, were used. From these data it is clear that Kahl hardness shows considerable
variation as can be gathered from the high coefficient of variation (CT). Using the
Instron or Kramer shear press reduces the CV. However, with these devices more pellets
are sheared during one test which will decrease the CV since it is calculated as the

Table 3
Effects of processing conditions on durability and hardness of a barley pellet according to different devices
(means and coeflcient of variation; A.F.B. van der Poe], G. Heijnen, V. Walk and H. Beumer, unpublished
results, 1993)
Temperature Hardness Durability
Kahl (kg) * Instron Kramer Holmen (%,) d Pfost (%I d
(Newton, N) b (pounds, lb) ’
Die5X25mm
70°C 6.6 (52.7) I25 (22.9) I63 (5.8) 75.7 (1.1) 91.8+2, (0.2)
80°C 6.6 (21.8) 161 (6.4) 190 (5.0) 91.3 (0.2) %.7+z4 (0.1)
90°C 7.6 (10.8) 166(4.6) 192(3.1) 92.8 (0.2) 97.4+24 (0.0)
Die 5 X 35mm
50°C 7.6cl3.9) I55 (9.4) 192 (6.2) 94.1 (0.1) 97.8+*, (0.0)
60°C 7.8 (15.4) I64 (6.7) 206 (5.0) 94.3 (0.1) 97.9+,,(0.1)
70°C 7.8 ( 15.6) I72 (6.9) 208 (3.8) 94.0 (0.1) 98. I + ?,, (0.0)
80°C 7.8 (16.4) I75 (6.2) 203 (6.8) 94.5 (0.2) 98.2+24 (0.0)
90°C 8.0 (23.7) I76 (5.0) 204 (6.5) 94.0 (0.8) 98.2+,,(0.1)

’ Single pellet measurement, ten replicates.

b Average of five pellets each, ten replicates.
’ 15 g of pellet sample, five replicates.
d Percentage of pellets returned, two replicates.
102 M. Thornus, A.F.B. van der Peel/Animal Feed Science Technology 61 (1996) 89-112

variation in multiple tests. It is noted that under the conditions of this experiment the CV
increases when hardness is decreasing.

3.2. Durability

Durability is another physical quality parameter of feed pellets. The devices used for
the determination of durability measure the amount of fines returned from a batch of
feed pellets under standard&d conditions.
In practical feed manufacturing, pelleted feeds are subject to shearing and abrasing
actions during transportation. This induces fines in the feed. For the purpose of feeding
ease, pellets need to have a certain resistance against the stresses exerted on them during
transportation and distribution to the animals. In the past decennia, the mechanical
transport of feed changed from relative low speeds to high speed pneumatic transport,
thereby changing the order and magnitude of forces on feed pellets.
Pellets are sensitive to shearing actions at the places where they are cut off after
leaving the die. Through this impact new surfaces are created which are sensitive to
further deterioration. Improper cooling may increase this sensitivity to further deteriora-
tion. Pellets that are not properly cooled can have a reduced durability due to stresses in
the pellet between the (cooled) outer layer and the (still) warmer centre. Up to a certain
level, relative cool air takes up moisture and heat from the pellets during the cooling
process. In a steady state, the same amount of moisture (and latent heat) is transported
through capillaries from the inner part to the surface. When air speed is increased, more
water and heat will be removed from the pellet surface than can be delivered by the
capillaries; a brittle outer layer emerges with physical properties differing from those of
the inner kernel, the latter being warmer and more viscous. These differences in physical
properties create stresses in the pellet which cause the outer layer to crack under less
optimal conditions. These cracks will allow for an easier formation of fines.
Robohm and Apelt (1985) and Robohm and Apelt (1986) conducted a simulation
experiment for pneumatic transport to elucidate the effect of air speed (m s- ’), air load
(kg pellets per kg air) and pellet diameter on the amount of fines appearing. They found
that increasing the air speed from 10 to > 30 m s- ’ increased the amount of fines by
approximately 3.5%. On the other hand, higher air load (more kg pellets per kg air for
transport) decreased the amount of fines. Furthermore, the authors showed that pellets
with small diameters (3 mm) proved to be more susceptible to breakage than those with
larger diameters (6 mm).
Several instruments are nowadays available for the evaluation of durability of pellets.
In principal, these devices bring about attrition stresses that are exerted to pellets, due to
either mechanical or pneumatic transportation (Table 4). Similar to hardness devices, the
test devices for durability have been developed for both scientific objectives and on-line
application in the factory. Already in the early 196Os, a rather complete model handling
system was used (Young, 1962) followed by the development of tumbling can devices
(Gutekunst, 1962; Pfost and Allen, 1962).
According to the procedure of Pfost (Pfost and Allen, 1962; Pfost, 1963) durability is
determined by inducing fines through an abrasing action of pellets shearing over each
other and over the wall of drums (tumbling can device: Fig. 7). The procedure is
 M, Thomas, A.F.B. uan der Peel/Animal Feed Science Technology 61 (19%) 89-112 103

Table 4
Generally applied test devices for pellet durability and their mechanisms
Principal Attrition mechanism Operation a Device example References
Mechanical resistance Abrasion Surface flow Pfost tumbling can Pfost ( 1963)
Vibration Abrasion/fragmentation Surface flow/impact Sieve de Jong (1993)
Pneumatic resistance Abrasion Impact Holmen Major ( 1984)
Fragmentation Surface flow

a Depending on standard conditions.

standard&d by using a drum with specified dimensions, in which 500 g of sieved

pellets are inserted. After tumbling for 10 min at 50 rpm, the pellets are subsequently
sieved and the amount of fines passing a sieve with a grid size “just smaller than the
nominal pellet diameter” (Pfost, 1963) is determined. In practice, grid-sizes are used of
0.8 X pellet diameter. Durability is then expressed as the ratio of the weight after
tumbling over the weight before tumbling, multiplied by 10. The resultant figure
(between 0 and IO) should be rounded off to two significant digits. Nowadays, however,
durability is normally expressed as a percentage which gives the amount of fines
returned or the amount of pellets recovered. In the original description, a subscript is
added depicting the time of physical analysis of pellets after their manufacturing in
hours. Subscript + 24 indicates durability results from pellets out of the cooler for more
than 1 day, - 1 indicates results from warm pellets. In scientific publications, however,
the subscript (an important indication for comparison reasons) is not always given.
Nowadays, automated devices exist to monitor on-line the durability of pellets during
their manufacture (Anonymous, 1984). Often these devices incorporate a fast cooling
procedure to quickly obtain physical quality figures that can be used to optimise process
parameters of that specific batch.

Fig. 7. ffost tumbling can device for measurement of pellet durability.

104 M. Thomas, A.F.B. uan der Poe1/ Animal Feed Science Technology 61 (19%) 89-112

.
- inlet

_ valves

- outlet

Fompressor

0
3
b
i sieve
fines collect bin

Fig. 8. Holmen pellet durability tester.

A sieve is a further device for measuring durability. Although not all standard
procedures for sieving are suitable, those which comprise the use of balls between the
screens can be used. By vibrating a pellet sample in a sieve apparatus, abrasion and
perhaps fragmentation will occur through the action of (rubber) balls. Variables in this
device are the sieve openings, the number of balls, the amplitude, screening time and
eventually the use of interval sieving. Particle size distribution obtained by dry sieving
may add useful information when the influence of sieving (under standardised condi-
tions) for example is expressed in terms of the modulus of fines or the modulus of
uniformity before and after sieving (Pfost and Headley, 1976).
The ‘Holmen’ pellet tester (Major, 1984; Fig. 8) simulates a more rigorous treatment
of pellets by pneumatic handling. In this apparatus, a 100-g sieved sample of pellets is
introduced in a stream of air. For a standard time (0.5 to 2 min) this air together with
pellets is circulated through right-angled bends, impinging repeatedly on hard surfaces.
Pellet attrition will then occur. After treatment, the product is sieved again using a sieve
with an opening of approximately 80% of the pellet diameter. Standards have been
developed for the testing time in relation to the pellet diameter under investigation.
Many people refer to the Holmen pellet tester as a device for measuring durability of
pellets. Closer examination of pellets under investigation reveals that fines produced by
the Holmen tester comprise a ‘mixture’ of fines derived from both abrasion and
fragmentation.
 M. Thomas, A.F.B. uan der Peel/Animal Feed Science Technology 61 (1996) 89-112 105

Not much information can be found on the correlation between measurements

conducted on all of the types of equipment mentioned, except for the Holmen durability
tester and Pfost tumbling Can. Data from McKee (19901, indicated that for the Holmen
pellet tester there was a linear decrease in durability, ranging from approximately 95%
to 60% with testing time, up to 5 min. The F’fost tumbling can showed a curvilinear
decrease, ranging from approximately 98 to 91% with time, up to 20 min. From these
results it can be concluded that the Holmen pellet tester gives results in a wider range
and in a shorter timespan than Pfosts tumbling can.
Wood (1987) found in his research a log-linear relationship (R = 0.94) between
hardness tested with the Kahl device and durability tested with a Holmen pellet tester.
These relations may be found within certain diet formulations. However, because of the
fact that hardness and durability are two distinct quantities, differences in relationships
may be found between different diet formulations. These differences may become more
distinct when new types of equipment are used to manufacture pelleted feed, e.g.
expanders.

3.3. Other criteria

In the former paragraphs some of the most widespread measures of physical quality
from a manufacturers point of view are evaluated. However, other criteria exist to
evaluate feed physical properties. These criteria are for instance water absorption index
(WAI), water solubility index (WSI), and sinking velocity (SV), all criteria being useful
in e.g. fish feeding. Other criteria well related to properties of the raw materials from
which feeds are manufactured are the angle of repose, bulk density and specific weight.
These criteria can be used to characterise the different feeds or feedstuffs. Some of these
criteria can be related to complete feeds as well. For a more fundamental description of
both methods and theories on physical properties related to characterising plant and
animal materials, the reader is referred to Mohsenin (1986). Since these criteria are not
specific for pellets as such, and thus not within the scope of this study, they are not
discussed.

4. Discussion

Different methods for the evaluation of the attrition behaviour of pellets supply
different information. It is obvious that it is not possible to use each type of device with
every type of feed pellets. The choice for a certain method therefore is partly determined
by the objective of the measurement, related to handling or nutritional purposes or to
study simulated production of fines caused by either fragmentation or abrasive [Link].
The described methods for the evaluation of physical quality of pellets have been
shown to be tests that use stresses brought about by bending/tension or compression. In
the context of animal feed pellets, bending tests and tensile tests in general may give
more information compared with compression tests (Luyten et al., 1992) but are of less
value due to the type of attrition that feed pellets undergo between manufacturing in the
feed mill and the animal feeding trough.
106 M. Thomas, A.F.B. van der Peel/Animal Feed Science Technology 61 (19%) 89-112

The compression test methods more generally employed in the routine evaluation of
pellet attrition more or less reflect a mixture of measuring fines produced by fragmenta-
tion and abrasion rather than what is normally distinguished by the terms ‘hardness’ and
‘durability’, respectively. There may be a relationship between hardness and durability
(Wood, 1987) and between the results obtained by different devices for durability
(McKee, 1990). However, these relations may only hold for a given feed composition.
The relationship is strongly influenced by the diet ingredient composition of the pellet
and the pelleting conditions used. With all devices applied, pellets are subjected to a
change in particle size distribution. Therefore, the addressed question should not be
“what is the hardness or durability of feed pellets” but from a scientific point of view
“which combination of fragmentation or abrasion reflects at best the way feed pellets
are handled, conveyed or stored”. The subsequent choice for a type of device also
depends on the outline of the factory, whether mechanical or pneumatic transport is
used, the form and shape of storage bins and, last but not least, the ultimate consumer’s
demand (whether man or animal) for a certain physical quality of pellets. For instance,
Skoch et al. (1983) found differences in feeding preference of pigs, when feeds in
different forms were given in a ‘choice-feeding’ experiment. Pigs preferred pellets over
meal but softer pellets over harder pellets.
The large variation that exists within results of e.g. hardness devices (Table 3) is
partly due to the fundamental misunderstanding of breaking behaviour of agglomerates
of biological origin. Analysis of pellet hardness figures with data from five experiments
confirmed this large variation (Table 5). As shown by Fell and Newton (19701,
empirical modification would permit the decrease of the large variation that can be
found in fragmentation-type measurements that employ diametrical compression such as
the Kahl test, Schleuniger test and (partly) the Kramer shear press. Careful application
of the standard conditions under which the device is operated is therefore recommended.
It may be relevant to adjust these standard conditions, in particular for those devices
which show a large variation in pellet quality results.
Current routine devices used for durability are a compromise between measurements
that give a practical value for pellet quality (on-line level) and measurements that have
scientific value. For example, durability measured according to the tumbling can method
(Pfost, 1963) is an actual measurement of abrasion. The pneumatic resistance test
(Holmen tester), however, is an actual measurement of both fragmentation and abrasion
(Robohm, 1987).
Since transport and handling involve both fragmentation and abrasion phenomena, it
would therefore relate more closely to quantities measured with the Holmen pellet tester
than with the tumbling can device, since the first device acts as an ‘in between’
simulator. For fundamental directed studies, some devices may give no valid estimate of
pellet quality as a function of various processing conditions (insufficient resolution) but
are still useful as a quick, routine device for both processing control and/or quality
control in the factory. The latter means that the use of a certain device method appears
suitable for process control within a certain factory when used under standard condi-
tions.
In the authors’ point of view, it should be recognised that the time of physical quality
measurement after manufacturing of pellets took place should always be indicated, as
 M. Thomas, A.F.B. van der Peel/Animal Feed Science Technology 61 (1996) 89-112 107

Table 5
Test criteria for pellet quality measuring devices
Device Test criteria
Repeatability ’ Validity
Fragmentation Abrasion
-
Compression devices
Kahl -- Yes No
Schleuniger -/+ Yes No
Pendulum + Yes No
Universal tension compression b + Yes No
Kramer-shear press + Yes Yes/no

Attrition devices
Pfost +++ No Yes
Sieving ? Yes/no Yes
Holmen +++ Yes Yes

a Repeatability on the basis of coefficient of variation (CV). The CV ranges between (M. Thomas, unpublished
results, 1994): + + +, O-l; + +, l-5; +, 5-10; -, 10-20, - --,> 20.b Examples: lnstron and overload
dynamics.

included in the original description for durability (Pfost, 1963). In fact, for each
measurement it is important irrespective of the device used. In addition, (specified)
standard conditions need not be similar to guideline recommendations. While the
individual feed manufacturer should set standards for his own situation, science asks for
a standardised approach for comparative reasons.
In the case where a number of pellets is examined, the CV of data is commonly
lowered by excluding the highest and lowest value. This procedure is rather commonly
applied by for example the Kahl and Schleuniger tests. From a scientific point of view,
this procedure, however, is highly questionable.
It can be concluded that only general descriptions can be given on the magnitude of
the different binding mechanisms. It appears that crystallisation processes taking place in
the pellet during cooling/drying and capillary forces account for the largest part of
binding particles (Schwanghart, 1970). However, other non-discussed phenomena, such
as glass-formation, may play a role as well in creating structure within agglomerates.
Within the scope of animal feed manufacturing these are still poorly understood. More
research is needed to gain knowledge on the effects of different processing conditions
and the magnitude of the different binding mechanisms.
The devices used in animal feed manufacturing most of the time yield a figure which
merely reflects a measurement which consists of a mixture of fragmentation and
abrasion, for instance the Holmen pellet tester. For practical purposes the aim should be
to identify the method which at best reflects the mixture of hardness and durability
which is most closely related to the figures derived from the specific plant at which the
feeds have been produced. For this purpose empirical equipment has been developed
which at present is commonly used in animal feed production: Pfost’s tumbling can, the
Holmen pellet tester and the Kahl hardness tester. Recently, on-line test devices have
108 M. Thomas, A.F.B. van der Poe1 / Animal Feed Science Technology 61 (I 996) 89-1 I2

been developed for use during the manufacturing stage. Therefore variation in pellet
quality during a production run can be evaluated and subsequently used to adjust
pelleting conditions during manufacturing.
For scientific evaluation, equipment is needed with which measurements can be
performed under standardised conditions for one unique characteristic, for instance
hardness. Still, using standard&d conditions, results found with ‘scientific’ equipment,
for instance Instron, overload dynamics or Kramer shear press, are often no better than
those found with empirical tests. This is due to the inherent variability of the pellets
produced in the factory and to limited knowledge on breaking behaviour of materials of
biological origin. The main advantage so far in using ‘scientific’ equipment is its use via
standard&d conditions and its measurement of one property, whether fragmentation or
abrasion. This will ensure pellet quality measurements which are comparable between
and within pellet manufacttuing systems.

Acknowledgements

The authors are greatly indebted to Dr ir T. van Vliet for his valuable comments on
the manuscript during preparation.

References

Anonymous, 1970. Operating instructions for L.E.E.-Kramer shear press. Al10 Precison Metals Engineering
Inc.. Rcckville, MD, USA. 17 pp.
Anonymous, 1984. Automating pellet testing the Swiss way. Feed Int., 5: 18-23.
Anonymous, 1992. Snelwerkende on-line hardheidstester. De Molenaar, 95: 1105-l 107.
Beumer, H., 1978. De invloed van enkele wijaen van mengselvoorbereiding op het persmsultaat. De Molenaar,
81: 431-440.
Beumer, H., 1980a. De invloed van enkele wijzen van mengselvoorbereiding op het perstesuhaat, De
Molenaar, 83: 1688- 1690.
Beumer, H., 1980b. De invloed van enkele wijzen van mengselvoorbereiding op het persresuhaat, De
Molenaar, 83: 1716-1718.
Beumer, H.. 198Oc. De invloed van enkele wijren van mengselvoorbereiding op het persresuhaat, De
Molenaar, 83: 1756- 1758.
Beumer, H.. 198Od. De invloed van enkele wijzen van mengselvoorbereiding op het persmsultaat, De
Molenaar. 83: 1792- 1794.
Beumer, H. and Vooijs, A.J., 1993. Optimaliseren van het pelleteren van veevoeders. Fase II. Dee1 II.:
experiment& ondetzcek. TNO-report number B 93.327, 155 pp.
Briiggeman, J., Niesar, K.H. and Walter. U., 1964. EinfluS von PreOhilfsstoffen auf PreSleismng und
Pellethbrte im GroMechnischen Versuch. Kmftfutter, 47(8): 414-420.
Calet, C.. 1%5. The relative value of pellets versus mash and grain in poultry nutrition, World Poult, 21:
23-52.
Cuperlovic. M., 1973. Intestinal loads of lysine and histidine and their concentration in systemic blood plasma
of pigs after fecdmg, Res. Vet. Sci., 14: 398-400.
Cuperlovif, M., Cmiljanif, R. and Boban. Z., 1973. The effect of pelleting a corn-sunflower oil meal type diet
supplemented with synthetic lysine on post-prattdial plasma free lysine and histidme levels in pigs, Act.
Vet. Beograd, 23(S): 251-258.
 M. Thomas, A.F.B. van der Peel/Animal Feed Science Technology 61 (1996) 89-112 109

Das, H.K., Hattula, M.T., Myllym~i, O.M. and Miilkki, Y., 1993. Effects of formulation and processing
variables on dry fish feed pellets containing fish waste, J. Sci. Food Agric., 61: 18l- 187.
de Jong, J.A.H.. 1993. Slijtage van deeltjes: harde realiteit vraagt zachte aanpak. Proces Technologie, Juli:
30-34.
de Jong, J.A.H., 1995. Gruisvorming bij transport en opslag. Optimale mengvoederproduktie anno 2000, IV
ontwerp en inrichting van de produktie in de praktijk. Studiedag voor de mengvoedetindustrie, 22 Februari
1995. Ede, The Netherlands.
Drevet, S., 1972. Untersuchung einiger physikalischer Stoffeigenschaften von Mischfutter. Mbhle Mischfut-
tertech., lO9(4): 50-54.
Eberhardt, W., l%4. Futterverpressung und Pre8hilfsstoff auf Ligninsulfonatbasis, Kraftfutter, 47(l): 16-20.
Fell, J.T. and Newton, J.M., 1970. Determination of tablet strength by the diametral-compression test, J.
Pharmac. Sci.. 69(5): 688-691.
Friedrich, W., 1964a. Zur Technologie des Verpressens von Mischfutter. Grundlagen, Versuchsergebnisse und
ihre Bedeutung Iiir den Mischfutterbetrieb, 1, Teil. Krafttimer. 47(6): 3 10-3 14.
Friedrich, W., 1%4b. Zur Technologie des Verpressens von Mischfutter. Gnmdlagen, Versuchsergebnisse und
ihre Bedeutung fiir den Mischfutterbetrieb, 2, Teil. Kraftfittter, 47(7): 352-358.
Friedrich, W., 1977. Futtermitteltechnologie. In: M. Kling and W. Wiihlbier (Editors), Handelsfuttermittel 1.
Band 1, Gesamtschau, Futtermittel tier&her Herkunft, Fette und 61e, Zusatzstoffe. Verlag Eugen Uhner
Stuttgart, pp. 113-143.
Friedrich, W. and Robohm, K.F., 1%8. Die Abriebfestigkeit von Pellets und ihre Abhangigheit vom
PreSprozeB insbesondere der Kiihlung, Kraftmtter, 5 l(2): 59-64.
Friedrich, W. and Robohm, K.F., 1969. Vergleichende Pregversuche mit unterschiedlichen Stoffeigenschaften
auf der Ring- un Scheibenmatrize, Kmftfutter, 52 (8): 390-395.
Friedrich. W. and Robohm, K.F., 1970. PreBversuche mit Ligninsulfonat-Bindemitteln. Kraftfutter, 53(7):
344-349.
Friedrich, W. and Robohm. K.F., 1981a. PreBmigkeit von Komponenten und Mischungen, Teil I: Wirhnmg
von Getteide - MeStechnik und Auswertung, Krafttittter, 64(4): 158-168.
Friedrich, W. and Robohm, K.F., PreSFahigkeit von Komponenten und Mischungen, Teil II: Wirkung von
Ghlmlihlennachprodukten. Kraftfutter, 64(g): 434-444.
Friedrich, W. and Robohm, K.F., l981c. PreSffiigkeit von Komponenten und Mischungen, Teil III: Wirkung
von Komponenten mit vorherrschenden Inhaltsstoffen, Kraftfutter, 64(1 I): 564-569.
Gutekunst, G., 1962. Development of a tumbling device for measuring pellet durability. Proc. Feed Production
School, Midwest Feed Manuf. Assoc., Kansas City, MO, p. 29.
Ham&an, T.J., 1984. Effect of pellet size and pellet quality on pig performance. Anim. Feed Sci. Technol. 10:
277-283.
Heffner. L.E. and Pfost, H.B., 1973. Gelatinization during pelleting. Feedstuffs, 45(23): 32.
Hermansson, A.M., 1979. Methods of studying functional characteristics of vegetable proteins. J. Am. Gil
Chem. Sot., 56: 272-279.
Israelsen. M., Busk, J. and Jensen, J., 1981. Pelleting properties of dairy compounds with molasses,
alkali-treated straw and other byproducts. Feedstuffs, 7 Dec.: 26-28.
Jindal, V.K. and Mohsenin. N.N.. 1976. Analysis of a simple pendulum impacting device for determining
dynamic strength of selected food materials, Trans. ASAE, 19(4): 766-770.
Keller, J., 1983. Food starches and their uses. In: D.L. Downing (Editor) Gum and Starch Technology
Eighteenth Annual Symposium. New York State Agricultural Experiment Station, Geneva, pp. l-9.
Kinsella, J.E.. 1979. Functional properties of soy proteins, J. Am. Oil Chem. Sot., 56: 242-258.
Knacke, 0. and Pohl, H., 1959. Experimenteller Beitrag zu den Grundlagen des Pelletisierens, Chemie-Ing.-
Tech., 31( 1): 50-54.
Koopmans, J.P., Scholten, P.M., Roeleveld, P.C., Velthuizen, Y.W.M. and Beynen, A.C., 1989a. Hardness of
diet pellets and its influence on growth of pre-weaned and weaned mice, Z. Versuchstierkd., 32: 71-75.
Koopmans, J.P., Scholten, P.M. and Beynen, A.C., 1989b. Hardness of diet pellets and growth of pm-weaned
mice: separation of direct effects on the young and indirect effects mediated by the lactating females, 2.
Versuchstierkd., 32: 257-260.
Leaver, R., 1984. Feedstock conditioning. Feed Int., Nov.: 37-44
110 M. Thomas, A.F.B. van der Peel/Animal Feed Science Technology 61 (19%) 89-112

Luyten, H., van Vliet, T. and Walstra, P., 1992. Comparison of various methods to evaluate fracture
phenomena in food materials, J. Text. Stud., 23: 245-266.
MacMahon. M.J. and Payne, J.D., 1991. ‘lbe Pelleting Handbook. Borregaard Lignotech, Sarpsborg Norway,
64 PP.
Mater, D.E. and Bakker-Arkema, F.W., 1992. The counterflow cooling of feed pellets, Agric. Eng. Res., 53:
305-3 19.
Maier, D.E. and Gardecki, J.. 1992. Feed mash conditioning field case studies. Paper no. 92-1541, ASAE
International Winter Meeting, 15- 18 December, Nashville.
Major, R., 1984. Testing of physical pellet quality: the pneumatic method. Feed Manage., June: 20-26.
McCapes, R.H., Ekperigin, H.E., Cameron, W.J., Ritchie, W.L., Slagter, J., Stangeland, V. and Nagaraja,
K.V., 1989. Effect of a new pelleting process on the level of contamination of poultry mash by Escherichia
coli and Salmonella, Avian Dis., 33: 103- I I 1.
McCormick, F.J. and Schellenberger, J.A., 1960. Tests to determine hardness, Feedstuffs, March, 12: 38-40.
McKee, J., 1990. Vergelijkend onderzoek naar slijtvastheid van pellets, De Molenaar, 93: 273-275.
Melcion, J.P. and Delort-Laval. J., 1981. MCthodes objectives d’appreciation des caract&istiques physiques
des aliments compods, In: INRA Publications (Editors), Prevision de la Valeur Nutritive des Aliments des
Ruminants. Versailles, pp. 307-323.
Melcion, J.P. and van der Poe.1, A.F.B., 1993. Process technology and antinutritionaI factors: principles,
adequacy and process optimization. In: A.F.B van der Poel, I. Huisman and H.S. Saini (Editors), Recent
Advances of Research in Antinutritional Factors in Legume Seeds. Proceedings of the Second International
Workshop on Antinutritional Factors (ANF’s) in Legume Seeds, Wageningen, The Netherlands. pp.
419-434.
Melcion, J.P., Vaissade, P., Valdebouze, P. and Viroben, G., 1974. Influence des conditions d’agglomeration
sur quelques caract&istiques physico-chimiques d’un aliment pour porcelet. Ann. Zootech., 23(2): 149- 160.
Mercier, C. and Guilbot, A., 1974. Influence des conditions du mdis sur les caracteristiques physico-chimiques
de son amidon, Ann. Zootech., 23(3): 241-251.
Mohsenin, N.N., 1986. Physical Properties of Plant and Animal Materials. Gordon and Breach, New York,
891 pp.
Moran, E.T.. 1989. Effect of pellet-quality on the performance of meat birds. In: W. Haresign and D.J.A. Cole
(Editors), Recent Advances in Animal Nutrition. Butterworths, London, pp. 87-108.
Nielsen, I., 1994. Effect of expanding on protein degradability in different raw materials and mixtures.
Intercoop Feedstuffs Congress, 23-26 June 1994, Lofoten, Norway. 3 pp.
Nissinen, V.J., Peisker. M. and Liebert, F., 1993. Futtermittelbearbeitung und Enzymzusatz bei Broilem,
Kraftfutter, 76(9): 364-367.
Patton, J.W., Buskirk, H.H. and Rams, L.A., 1937. A study of the relative merits of pellets and mash poultry
feeds, Vet. Med., 32: 423-427.
Payne, J.D., 1978. Improving quality of pellet feeds, Milling Feed Fertil., 162: 34-41.
Pfost, H.B., 1963. Testing the durability of pelleted feed, Feedstuffs. March, 23: 66-68.
Pfost, H.B. and Allen, R.N., 1962. A standard method of measuring pellet durability. Proc. Feed. Prod. School,
Midwest Feed Manuf. Assoc., Kansas City, 12-14 Nov., pp. 25-29.
Pfost, H. and Headley, V., 1976. Methods of determining and expressing particle size. In: H.B. Pfost and D.
Pickering (Editors), Feed Manufacturing Technology. pp. 512-520.
Pfost, H.B and Young, L.R., 1973. Effect of colloidal binders and other factors on pelleting. Feedstuffs,
45(49): 2 I-22.
Pickford, J.R., 1992. The potential and use of expanders and extruden in animal feed production. Animal Feed
Production for the 1990s. Feed Expo Ireland.
Pipa, F. and Frank, G., 1989. High-pressure conditioning with annular gap expander. A new way of feed
processing, Adv. Feed Technol., 2: 22-30.
Pond, W.G. and Maner, J.G., 1984. Swine Production and Nutrition Animal Science. Textbook Series AVI
Publ. Co., Inc., Westport. CT, USA, 646 pp.
Quemere, P., Castaing, J., Chastanet, J.P., Grosjean, F., Latimier, P., Saulnier, J. and Willequet, F., 1988.
Influence of the physical form of the feed on the performance of bacon pigs, J. Rech. Porcine France, 20:
351-360.
 M. Thomas, A.F.B. van der Poe//Animal Feed Science Technology 61 (1996) 89-112 III

Richardson, W. and Day, E.J., 1976. Effect of varying levels of added fat in broiler diets on pellet quality,
Feedstuffs, 4800): 24.
Robohm, K.F., 1987. Kwaliteitsgarantie van mengvoederpellets door bepaling van fysische eigenschappen van
de materie (2), De Molenaar, 90: 649-653.
Robohm, K.F. and Apelt, J., 1985. EinfluD der pneumatischen Tankwagenentleerung auf Bruch- und
Mehlbildung bei Mischfutterpellets, Teil 1, Kraftfutter. 68(12): 416-420.
Robohm, K.F. and Apelt, J., 1986. Einflu6 der pneumatischen Tankwagenentleerung auf Bruch- und
Mehlbildung bei Mischfutterpellets, Teil 2, Kraftfutter, 6% 1): 15- 16.
Rumpf, H., 1958. Grundlagen und Methoden des Granulierens, Chemie-Ing.-Tech., 30(3): 144- 158.
Salmon, R.E., 1985. Effects of pelleting, added sodium bentonite and fat in a wheat-based diet on performance
and carcass characteristics of small white turkeys, Anim. Feed Sci. Technol., 12: 223-232.
Schwanghart, H., 1969. Messung und Betechnung von Druckverh”altnissen und Durchsatz in einer Ringkoller-
Strangpresse, AuIbereit.-tech., lOfl2): 713-722.
Schwanghart, H., 1970. Festigkeiten von Futtermittelpreglingen. Aufbemit.-tech., 1l(4): 192- 199.
Skoch, E.R., Behnke, K.C., Deyoe, C.W. and Binder, S.F., 1981. The effect of steam-conditioning rate on the
pelleting process, Anim. Feed Sci. Technol., 6: 83-w.
Skoch, E.R., Binder, SF., Deyoe, C.W.. Allee, G.L. and Behnke, KC., 1983. Effects of pelleting conditions
on performance of pigs fed a corn-soybean meal diet, J. Anim. Sci., 57(4): 922-928.
Smith, P.S., 1983. Food starches and their uses. In: D.L. Downing (Editor), Gum and Starch Technology
Eighteenth Annual Symposium. New York State Agricultural Experiment Station, Geneva, pp. l-9.
Stark, C.R., 1990. Evaluation of pelleted soybean meal for domestic and international markets. [Link]. Thesis,
Kansas State University, USA, 92 pp.
Stevens, C.A., 1987. Starch gelatinization and the influence of particle size, steam pressure and die speed on
the pelleting process. Ph.D. Thesis Kansas State University, USA, 86 pp.
Tamminga, S. and Goelema, J.O., 1995. The significance of rate and site of starch digestion in ruminants.
Carbohydrates in feeds for ruminants. Proc. SCI, 28 February 1995, London, UK.
TeSiC, M., 1977. Das Verdichten unter Ntirstoffaufschluss von Futterpflanzen in Matrizenptessen. Ph.D.
Thesis, Georg-August Universidt Gottingen, Germany, 181 pp.
Thomas, M., van Vliet, T. and van der Poel, A.F.B., 1996a. Physical quality of pelleted animal feeds. 2.
Contribution of feedstuff components. Anim. Feed Sci. Technol., in press.
Thomas, M., van Zuilichem, D.J. and van der Poel, A.F.B., 1996b. Physical quality of pelleted animal feeds.
3. Contribution of processes and its conditions. Anim. Feed Sci. Technol., in press.
Tran V., Morat, J.L. and Bertrand, D., 1991. A new serial method for searching me optimum operating
conditions of a technological process, J. Chemom., 5: 73-84.
van der Poel, A.F.B., Verstegen, M.W.A. and Tarnminga, S., 1995. Chemical, physical and nutritional effects
of feed processing technology. In: 16th Western Nutrition Conference, 13- 14 September 1995, Saskatoon
Sask., Canada. 266 pp.
van Rooy, C., 1986. De invloed van technologische bewerkingen, in het bijzonder van pelleteren. op de
vcrteerbaarheid van mengvoeders bij varkens. [Link]. Thesis, Wageningen Agricultural University, Wa-
geningen, The Netherlands.
Vanschoubrock, F., Coucke, L. and van Spaendonck. R., 1971. The quantitative effect of pelleting feed on the
performance of piglets and fattening pigs, Nutr. Abstr. Rev., 41: 1-9.
Van Zuilichem, D.J. and Stolp, W., 1976. Theoretical aspects of the extrusion of starch based products in
direct extrusion cooking process. In: Proc. International Snack Seminar, 18-21 October, German Confec-
tionary Institute, Solingen.
Van Zuilichem, D.J. and van der Poel, A.F.B., 1993. The comparative costs and benefits of expanders and
double pelleting in animal feed production. In: hoc. Feed Expo Ireland, I l-12 May, Dun Laoghaire,
Ireland. 102 pp.
Van Zuilichem, D.J., Moscicki, L. and Stolp, W., 1979a. Nieuwe bindmiddelen verbeteren brokkwaliteit, De
Molenaar, 82: 1572.
Van Zuilichem, D.J., Moscicki. L. and Stolp, W., 1979b. Nieuwe bindmiddelen verbetemn brokkwaliteit, De
Molenaar, 82: 16l8- 1622.
Van Zuilichem, D.J., Moscicki, L. and Stolp, W., 1980, Nieuwe bindmiddelen verbetemn brokkwaliteit, De
Molenaar, 83: I2- 18.
112 M. Thomas, A.F.B. van der Peel/Animal Feed Science Technology 61 (1996) 89-112

Veenendaal, J., 1990. Extrusion in the compound feed industry, Adv. Feed Technol., 3: 60-72.
Winowiski, T., 1988. Problems in optimizing pellet quality. Feed Manage., Sep.: 36-42.
Wood, J.F., 1987. The functional properties of feed raw materials and their effect on the production and
quality of feed pellets, Anim. Feed. Sci. Technol., 18: 1- 17.
Young, L.R., 1%2. Mechanical durability of feed pellets. [Link]. Thesis, Kansas State University, KS, USA.