Benefits of Die-Wall Instrumentation For Research and Development in Tabletting

European Journal of Pharmaceutics and Biopharmaceutics 58 (2004) 427–444
www.elsevier.com/locate/ejpb
Review article
Benefits of die-wall instrumentation for research

and development in tabletting
Eric Doelker*, Danielle Massuelle
School of Pharmacy, University of Geneva, Geneva, Switzerland
Received 26 January 2004; accepted in revised form 1 March 2004
Available online 18 May 2004
Abstract
Instrumented presses used in tabletting research and development are normally equipped to measure punch force and displacement. Die-
wall monitoring is rare, probably because instrumentation and calibration are quite difficult. The authors critically examine the tenets of
radial pressure measurement in compression physics. The theoretical background concerning axial to radial stress transmission during the
different phases of the compression cycle is presented. The literature reporting on the use of radial stress measurement to assess the self-
lubricating properties of materials or the effect of lubricants is reviewed. Examples of interpretation of radial pressure cycles to define the
basic material behaviour are given. The influence of particle size and shape as well as that of process and formulation variables on die-wall
response are also discussed. Substantial inconsistencies can be seen in the literature with respect to the interpretation of experimental data,
often because of the poor reliability of results and mostly because powders are essentially not solid, isotropic bodies. There is also a distinct
lack of complementary tabletting parameters that would help understanding their comparative benefits. For this reason, original data on 13
model compounds are presented together with a classification of the materials encountered in pharmaceutical tabletting, based on selected
parameters. In conclusion, none of the determined parameters, including those derived from radial pressure measurement, is able, alone, to
predict the material behaviour under compression. Although die-wall instrumentation contributes little to the development of improved tablet
formulations, it is valuable for characterising the mechanical properties of the materials. This is particularly advantageous given that the
mechanical properties account for variations in tabletting performance to a much greater extent than the magnitude of the interparticulate
attractions. Nevertheless, because of the peculiar nature of powders compared to solid, isotropic bodies, there is a need to develop new
models for analysing their behaviour and to put more emphasis on examination of time-dependent deformation in the later stage of the
compression cycle.
q 2004 Elsevier B.V. All rights reserved.
Keywords: Die-wall; Instrumentation; Tabletting machine; Radial pressure; Long pressure cycles; Tablets; Compacts; Tablets; R&D
1. Introduction (using strain gauges or piezoelectric sensors), the punch

displacement (using, e.g. linear variable differential trans-
Nowadays the instrumentation of presses plays an ducers), the ejection force, the force transmitted radially to
important role in research and development as well as in the die, the so-called ‘adhesion’ force, and the temperature
the production of tablets. Use of instrumented tablet rise during compaction. Fig. 1 presents a typical set of
signals recorded on a single-punch machine which have
machines is essential for basic research in compression
been digitalised and displayed on a monitor [1].
physics, as it facilitates product development, optimisation
In tablet production, instrumentation mainly concerns
and scale up, and enables monitoring and control of
punch force measurement for automatic weight control. In
production.
most R&D situations one mainly deals with punch force and
Tablet machines may be equipped with transducers that
displacement determination to derive parameters useful for
can monitor continuously the force acting on the punches
studying basic material properties (e.g. work of com-
pression, Heckel plot) or tabletting problems (capping or
* Corresponding author. School of Pharmacy, University of Geneva, Quai
Ernest-Ansermet 30, CH-1211 Geneva 4, Switzerland. Tel.: þ 41-22-379-
lamination tendency, tooling wear) due to formulation or
6148; fax: þ 41-22-379-6567. process variables. Many review articles have been published
E-mail address: eric.doelker@pharm.unige.ch (E. Doelker). on the topic in the last decades [2 – 13].
0939-6411/$ - see front matter q 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.ejpb.2004.03.011
428 E. Doelker, D. Massuelle / European Journal of Pharmaceutics and Biopharmaceutics 58 (2004) 427–444
single-punch machines, rotary machines and compaction

simulators were successively equipped to determine die-
wall pressure.
As for axial force measurement, both strain gauges and
piezoelectric transducers were used, but experimental
photoelastic techniques were also employed. The different
settings for die-wall instrumentation have been reviewed in
the literature [1,7,14 – 20]. Table 1 presents some of the
milestones in the field and Fig. 2 shows the most commonly
used arrangements for die instrumentation.
Most of these die designs suffer from deficiencies, render
Fig. 1. Typical digitalised signals obtained using a data acquisition system calibration difficult and thus data interpretation is not very
interfacing an instrumented tabletting machine to a computer. The ordinate reliable. The major concern with the most popular cut-away
is the signal intensity and the abscissa is the time (here, about 200 ms for the
(segmented) die technique (Fig. 2A) is due to non-
compression/decompression cycle). Key: (1) upper punch force, (2) lower
punch force, (3) upper punch displacement, (4) lower punch displacement, symmetrical stress/strain distribution within the die. The
(5) die-wall force (inverted signal), (6) residual lower punch force, and (7) output signal is thus non-linear and depends on the tablet
ejection force (the lower punch force is amplified during ejection for better position in the die and on the tablet height. For these reasons
visualisation) [1]. careful calibration as well as complex mathematical
corrections have to be carried out [31,33,35]. To overcome
In comparison, much less attention has been paid to the these problems, a three-layered die has been described by
measurement of die-wall response. The reason for that is Rippie and Danielson [30] and further developed by Yeh
two-fold: first, die-wall instrumentation is technically et al. [32] (Fig. 2B). By restraining the sensing zone to the
difficult and calibration is crucial; secondly, data reported thin middle layer, this arrangement isolates stresses
in the literature are often contradictory and thus subject to measured around the tablet and uncouples the stress/strain
controversy. distribution from those of the rest of the die. The authors are
The present article was thus written with the aim of of the opinion that this design is the best reported to date.
tentatively explaining sources of the contradictory results Another arrangement is to insert a piston through the die
reported and to draw conclusions about the benefits from wall and link it to a piezoelectric transducer (Fig. 2C), or to
die-wall instrumentation. Most of the results shown have a load cell. However, as the powder to be compressed is
been taken from the literature but original data will also be directly in contact with the piston acting on the transducer,
presented. particles frequently insert at the hole and grip the piston.
Moreover, the system may also be dependent on the tablet
position and height.
2. Die-wall instrumentation Whatever the method, a careful calibration of the die has
to be carried out by applying a set of axial forces on a
Die-wall instrumentation soon followed the installation hydraulic material of different heights and located at
of transducers for measurement of punch force and different positions of the lower punch. Natural or synthetic
displacement, in the mid-1950s. Hydraulic presses, rubber, either a plug or as powder, is used owing to its
Table 1
Some important steps in die-wall instrumentation
Authors Year Press Measuring system
Nelson et al. [21] 1955 Hydraulic Piston inserted through the die and linked to a load cell
Windheuser et al. [22] 1963 Hydraulic Strain gauges attached onto the outside of a cut die
Ridgway [23] 1966 Hydraulic Analysis of fringe pattern observed in a Perspex die
Leigh et al. [24] 1967 Single punch Strain gauges attached onto the outside of a cut die
Marshall [25] 1970 Single punch Piston touching the outside of a cut die and linked to a piezoelectric
transducer
Ridgway and Rosser [26] 1971 Rotary Analysis of fringe pattern observed in a Perspex die
Edwards [27] 1973 Single punch Piston inserted through the die and linked to a piezoelectric transducer
Spang [28] 1973 Rotary Strain gauges attached onto the outside of a cut die
Conte et al. [29] 1977 Single punch Two piezoelectric load washers mounted in a split die
Rippie and Danielson [30] 1981 Rotary Three-layered die with integration of a sensing web (strain gauges) into the
thick middle layer
Cocolas and Lordi [31] 1993 Compaction simulator Four piezoelectric transducers located in a spiral arrangement at different
heights of the die bore
Yeh et al. [32] 1997 Rotary Design of split-web die using finite element analysis
E. Doelker, D. Massuelle / European Journal of Pharmaceutics and Biopharmaceutics 58 (2004) 427–444 429
on the position of the lower punch, the rubber height and

the applied pressure (Fig. 3).
The variability in the die-wall design and the difficulty of
calibrating (including mathematical corrections) may
account for part of the discrepancies observed among the
data reported in the literature for given materials. Tables 2
and 3 list axial to radial stress transmission ratios, h;
reported in the literature for sodium chloride and para-
cetamol, respectively. Of course, much of the differences
are due to the range of maximum applied pressure, the type
of tabletting press, the press tooling, as well as the amount
of powder compressed and its characteristics (size, shape,
crystal defects).
3. Axial to radial stress transmission: theoretical

background
Fig. 2. Some arrangements for die-wall pressure measurement. (A) Cut-
away die with active and compensation strain gauges [33], (B) three-layer
die with strain gauges [32], and (C) die with a piston inserted through the A knowledge of some theoretical background is needed
wall and linked to a piezoelectric transducer [34]. to explain experimental radial stress radially transmitted
may stress data and thus draw benefits from die-wall
supposedly hydraulic fluid behaviour [22 – 24,27 –30,32,33, instrumentation. The radially transmitted force or pressure
(stress) may essentially intervene in two types of analysis:
35 – 39], as well as high density polyethylene [31].
However, it has to be emphasised that rubber materials
models related to axial transmission,
behave like hydraulic fluids when they are solid bodies
models related to radial transmission.
without any pores. Finally, some hysteresis in the die-wall
signal is sometimes observed when the calibration material Fig. 4 shows the stresses operating on a powder under
is positioned at some distance from the strain gauges. This compression. The friction force, Fd ; is the force lost to the
may also depend on the calibration material used [35]. For die wall and represents the difference between the two
this purpose, some authors have used pure hydraulic fluids, vertical forces:
namely water [24,35] or silicone oil [23].
The complexity of calibrating a segmented die Fd ¼ Fa 2 Fb ð1Þ
equipped with strain gauges is exemplified by the work where Fa and Fb are the applied force and the force
of Hölzer and Sjögren [33]. The die-wall signal depends transmitted to the lower punch (in a single-punch press),
Fig. 3. Influence of compact height on the die-wall signal and on signal per cross-sectional area at various pressures. Lower punch positions: A, 12.2 mm; B,
10.2 mm; C, 8.2 mm [33].
Table 2
Values of stress ratio at maximum applied pressure, h; reported in the literature for sodium chloride
Amount of material Particle size Press type Punch Compaction Applied Stress Authors
or compact height (mm) diameter time or speed pressure ratio ðhÞ
(habit) (mm) (MPa)
705 mg 420 –840 Hydraulic 9.5 30 s 57 0.59 Higuchi et al. [40]

1000 mg n.m. Hydraulic 12.5 n.m. Up to 150 0.40 Ridgway et al. [41]
4 mma 420 –840 Single punch 12.7b 11 s 237 0.36 Carless and Leigh [42]
4 mma 420 –500 Single punch 12.0b n.m. Obiorah [43]
Dendritic 179 0.39
Cubic 168 0.35
2.62 mm 7 –20 Rotary 8.0b 10 rpm (10 s) 54 0.39 Wiederkehr-von
Vincenz [37]
107 0.60
268 0.84
130 –400 107 0.68
268 0.85
7 –20 31 rpm (0.17 s) 54 0.42
107 0.62
268 0.82
3 mma 270 Single punch 11.3 30 rpm 150 0.46 Hölzer and Sjögren [44]
4.6 mmc n.m. Hydraulic 9.5 5 mm/min 225 0.71 Khossravi et al. [39]
n.m, not mentioned.

a
At zero theoretical porosity.
b
Lubricated die.
c
At maximum applied pressure.
respectively. Fd is sometimes called the frictional force 3.1. Axial transmission

(operating during the compression phase), whereas the
ejection force, Fe ; is the maximum force recorded on the The radial stress generated by pressing axially a body in a
lower punch during expulsion of the tablet out of the die. die is present in the numerous equations proposed to
The radial force, Fr (radial pressure, Pr ) arises as a result of describe die-wall friction. In the simple model of pressing
the horizontal expansion of the mass in response to the axial from one end in a stationary cylindrical die, these equations
compressive force (pressure). relate the pressure applied by the upper punch, Pa ; to the
Here, the term ‘pressure’, P; will be used interchangeably pressure transmitted to the lower punch, Pb : A well known
with that of ‘stress’, s: Note also that forces are actually derivation is that of Unckel [46]:
measured and subsequently converted into pressures or
stresses using the surface area concerned (punch surface, Pa
¼ e4mhL=D ð2Þ
edge or cross-sectional surface for the die). Pb
Table 3
Values of stress ratio at maximum applied pressure, h; reported in the literature for paracetamol
Compact Particle size Press type Punch Compaction Applied Stress Authors
heighta (mm) (mm) diameterb time or speed pressure (MPa) ratio ðhÞ
(mm)
3 14.5 Single punch 12.0 n.m. 112 0.46 Doelker and

Shotton [45]
4 420– 500 Single punch 12.0 n.m. 160 0.36 Obiorah [43]
60 180 0.43
2.62 1–7 Single punch 8.0 10 rpm (10 s) 161 0.57 Wiederkehr-von
Vincenz [37]
170– 300 375 0.68
161 0.56
375 0.67
3 125– 180 Single punch 12.7 Manually Up to 120 0.35 Krycer et al. [38]
n.m., not mentioned.

a
At zero theoretical porosity.
b
Lubricated die.
models is that monitoring of the radial pressure is needed

in order to calculate the coefficient of friction, m: Finally,
mention should be made of the analysis of Mosbab et al.
[51], which allows the estimation of intrinsic properties of
the material compacted, i.e. Young’s modulus and
Poisson’s ratio.
3.2. Radial transmission
Long [52,53] was the first author to elucidate the

theoretical significance of radial vs. axial pressure cycles.
Two types of the so-called Long’s compression cycles were
predicted according to the behaviour of the material
compressed, in addition to that of a perfectly elastic body
(Fig. 5). Let us examine the compression and decompres-
sion profiles obtained with the two types of materials
usually encountered. In both cases, line OA (or OA0 ) is
Fig. 4. Forces and pressures operating on a powder under compression in a
punch and die assembly. Key: Fa ðPa Þ; force (pressure) applied by the upper representative of a perfectly elastic behaviour of a solid
punch; Fb ðPb Þ; force (pressure) transmitted to the lower punch; Fd ; force isotropic plug of material, that is if the axial pressure, Pa ; is
lost to the die (axial frictional force); Fr ðPr Þ; force (pressure) radially released, the radial pressure, Pr ; would return along this to
transmitted to the die wall; Fe ; ejection force; D; die diameter; L; height of zero. The radially transmitted pressure is related to the
the compact.
Poisson ratio n by the relation:
n
Pr ¼ P ð4Þ
where m is the coefficient of die-wall friction; h; the stress 12n a
ratio, defined as the ratio between the radial pressure, Pr ; with a slope equal to n=ð1 2 nÞ for both types of materials.
and the applied pressure, Pa ; L is the compact length, and D When the axial pressure exceeds the elastic limit (yield
is the diameter of the die. pressure) A (A0 ), yield starts and the slope changes
As m is the coefficient of friction between the powder differently according to the failure behaviour of the
mass and the die wall, then: material. For a body with a constant yield stress (strength)
in shear, S; the following equation applies:
Fd ¼ mFr ð3Þ
Pr ¼ Pa 2 2S ð5Þ
Unckel, in his equation, used the length initially occupied by
the powder fill. Other authors have introduced terms to so that Pa 2 Pr remains constant ð¼ 2SÞ and the slope of the
allow for the effect of changes in the bulk density during line AB will equal unity.
compression [47].
All these basic equations describe an exponential decay
of the applied pressure down the compact length and
postulate constant m and h values. However, it has been
suggested that m and h may vary along the length of the
compact according to the conditions of relative interfacial
movement, even though the product mh could remain
constant [48]. This could explain the experimentally
observed uneven stress (and density) distribution in
compacts. In fact, both axial and radial stress gradients
are present although short compacts (like pharmaceutical
tablets) are supposed to be reasonably homogenous along
the compression axis, especially when the die is
lubricated. In contrast, the radial distribution of the axial
pressure on the upper and lower punch shows parabolic
and symmetric patterns about the central axis of the die
[49]. In fact, complex models accounting for both axial
and radial stress gradients have been developed, e.g. that
Fig. 5. Theoretical radial pressure cycles for an ideal elastic body, a body
of Thompson among them[50]. Note that this author with constant yield stress, and a Mohr body. Key: (A, A0 ), yield points; (B,
describes the stress ratio h as a measure of powder B0 ), maximum applied pressures; (C, C0 ), yield points at decompression; (D,
fluidity. The general conclusion from analysing these D0 ), residual die-wall pressures (adapted from Long [52]).
For a Mohr body, i.e. a body with a shear stress porosity continuously decreases and point A (A0 ) is not
depending on the value of the normal stress, Eq. (6) holds: always visible. It is in fact likely that part of plastic flow
is confined to the interparticulate regions [42]. Instead of
ð1 2 mÞPa 2 2S0
Pr ¼ ð6Þ the slope of the OA (OA0 ) segment, it is thus better to
1þm
consider the rest of the cycle where a reasonable degree
where m is the coefficient of internal friction and S0 is the of densification has occurred.
yield stress in pure shear. † The powder yield criterion is not fulfilled when the
After the maximum applied pressure is reached (points B powder yield locus lies above the wall yield locus [56].
and B0 ), decompression occurs and the plug is no longer † If the applied pressure has not been large enough to reach
forced to yield but recovers elastically. The line BC (B0 C0 ) the elastic limit, i.e. point A (A0 ), yield will not take place
will be parallel to OA (OA0 ), with a slope again equal to and the change of slope at point C (C0 ) will not be
n=ð1 2 nÞ: observed. This also affects the magnitude of the residual
At point C (C0 ) the induced radial pressure exceeds the radial pressure.
axial pressure and yield again takes place. In the case of a † Again, if the applied pressure has not been sufficiently
material with a constant yield stress in shear, the line CD high, the part CD (C0 D0 ) may be missing from the cycle,
will be parallel to line AB and the slope is equal to unity. In thus affecting the residual radial pressure.
the case of a Mohr body, the slope of the line C0 D0 is the † Die elasticity also affects the shape of the radial
reciprocal of that of line A0 B0 because of stress inversion. compression cycle [53]. This, of course, may be a
Once the axial pressure has returned to zero but ejection concern when the die thickness is reduced to position the
has not yet occurred, the plug will remain under a residual transducer (cut-die technique). Two different situations
radial pressure, Pr0 : Its value is: have to be considered. First, the cycle is complete and
only the slopes of lines OA (OA0 ) and BC (B0 C0 ) are
Pr0 ¼ 2S ð7Þ reduced. The magnitude of the residual radial pressure is
for a body with constant yield stress in shear and: unchanged. Second, the cycle is incomplete for one or
more of the aforementioned reasons. In this case, line CD
2S0 (C0 D0 ) is absent and line BC (B0 C0 ) continues to intercept
Pr0 ¼ ð8Þ
12m the radial pressure axis, with die elasticity thus causing
for a Mohr body. higher residual radial pressure values.
Table 4 summarises the slope values of the various
All these justify performing compression studies at
segments. It should be noted that somewhat different
sufficiently high applied pressures and comparing different
notation is used by certain authors [54,55].
materials at a given relative density (e.g. 0.90).
In practice, one rarely observes the typical Long’s
Finally, in addition to analysing the shape of the
compression cycles. Several reasons have been put forward
compression cycle, it may be informative to calculate the
to explain such deviations [52,53]:
area of the hysteresis loop OABCD (OA0 B0 C0 D0 ), which
indicates the extent of departure from ideal elastic
† In case of a perfectly hydrostatic body ðn ¼ 0:5Þ; S is
behaviour [7]. Carstensen and Touré [54,57] integrated the
constant and equal to zero, so that A (A0 ) coincides with
various equations related to the compression cycle and
O and C (C0 ) with B (B0 ). The cycle degenerates into a
reached the conclusion that the hysteresis loop area is linear
single straight line defined by Pr ¼ Pa [52].
with the maximum applied pressure in the case of a body
† A compacted powder is obviously not a solid, isotropic
with a constant yield stress in shear, whereas it is
plug and its yield stress cannot be expected to be the
proportional to the square of the maximum applied pressure
same. Thus, segment OA (OA0 ) is often curved as
for a Mohr body.
Table 4
Theoretical slopes and residual die-wall pressure of radial pressure cycles
for a body with constant yield stress in shear and a Mohr body
4. Use of radial stress measurement to assess
Constant yield stress body Mohr body the self-lubricating properties of materials
or the effect of lubricants during tabletting
Segment Slope Segment Slope
Both the friction arising between the material and the die
OA v=ð1 2 vÞ OA0 v=ð1 2 vÞ
AB 1 AB0 ð1 2 mÞ=ð1 þ mÞ (die-wall friction) and that between particles (interparticu-
BC v=ð1 2 vÞ B0 C0 v=ð1 2 vÞ late or internal friction) operate during tabletting. However,
CD 1 C0 D0 ð1 þ mÞ=ð1 2 mÞ internal friction is of significance only at low applied
Pr0 2S Pr0 2S0 =ð1 2 mÞ pressures, i.e. during particle slippage and rearrangement,
Pa ðAÞa 2S=ð1 2 vÞ Pa ðA0 Þa S0 þ mPr
and is definitely not a decisive factor in the tabletting
a
Axial pressure at which yield takes place. process. In contrast, the friction between the powder mass
and the die wall is a real issue, but only when a sufficient At least two coefficients of friction can be defined during
radial pressure has been generated, i.e. beyond a certain the tabletting process [59,61]:
consolidation ratio. Only the powder/die friction will thus
maximum axial frictional force Fd
be considered here. m1 ¼ ð9Þ
maximum radial force Fr
Friction is present and varies at all tabletting phases:
compression, decompression, pre-ejection and ejection. In and
fact, tabletting problems often originate in the compression
ejection force Fe
and decompression phases but become evident at ejection m2 ¼ ð10Þ
[58,59]. residual die-wall force Fr0
Friction phenomena occurring during tabletting are An important limitation has to be noted here.
usually estimated through various parameters calculated Coefficient m2 can be defined as a constant as the
from upper and lower punch force measurements, compact/die-wall surface does not vary during the sliding
eventually with punch displacement measurement. These process. Coefficient m1 is more the result of a formal
parameters include the ratio of the maximum lower punch arithmetic operation. At maximum applied pressure there
force to the maximum upper punch force (referred as to is no movement and the axial frictional force (force lost to
the lubrication ratio, R), or their difference, Fd (Eq. (1)), the die) is dependent on the whole compression phase
the ejection force, Fe ; the residual lower punch force, the where the applied force has constantly changed as did the
work of friction, and the work of ejection [13]. In fact, surface and its texture.
factors influencing the friction conditions at the die also Note that m1 and m2 were referred to as the ‘static’ and
affect the radial transmission of the applied pressure. ‘dynamic’ friction coefficient, respectively, and is not
Thus, lubricants (in the broad sense) reduce both the consistent with the above definitions. Note also that
interparticle friction and friction at the die, and hence several authors refer to the product mh of Eq. (2) as
particles compact more tightly and homogeneously. The ‘apparent coefficient of friction’, thus including the stress
result is a decrease in the coefficient of die-wall friction, ratio [48,62,63].
m; and an increase in radial stress transmission, h: A Apart from the tabletting equipment (type of die and its
consequence may be that the product mh in Eq. (2) is physical state [60,64]) and the product to be compressed,
hardly influenced by the presence of lubricant. The only many manufacturing conditions may in practice influence
way to discriminate between the two effects is to monitor friction coefficient measurements. These include the method
the die-wall pressure in order to measure h and then of tablet production (unique or continuous) and applied
calculate m: It should be noted that m can also be pressure.
experimentally determined using special equipment of a The work of Hölzer and Sjögren [61] nicely illustrates
design very different from those of the pharmaceutical the effect of consecutive tablet production and thus die-wall
tabletting machines [56,60]. conditioning (Fig. 6). When unlubricated sodium chloride
Before considering the various coefficients of friction was used (first 20 tablets), m1 increased to a constant value
related to the tabletting process, it is convenient to make a of 1.4 (Fig. 6, left). Then, 20 tablets of lubricated sodium
clear distinction between static friction—the force to initiate chloride were compressed and m1 rapidly decreased to a
sliding -and dynamic friction—the force to maintain sliding constant value of 0.3. Finally, when returning to unlubri-
between the two surfaces, that are generally characterised cated sodium chloride, more than 30 tablets were needed for
by differing values of friction coefficient. m1 to reach the starting value of 1.4, indicating that a very
Fig. 6. Effect of die conditioning on coefficient of friction at maximum applied pressure, m1 ; between the die-wall and consecutive tablets of sodium chloride
(left) and microcrystalline cellulose (right). Key: A, sodium chloride at 115 MPa; W, sodium chloride with 1% magnesium stearate; K, microcrystalline
cellulose at 60 MPa; P, microcrystalline cellulose at 150 MPa [61].
Fig. 7. Variation of the coefficients of friction at maximum applied pressure, m1 ; and at ejection, m2 ; with the maximum applied pressure, Pa , for various model
materials. Key: A, sodium chloride; W, lubricated anhydrous lactose; K, alprenolol hydrochloride; S, microcrystalline cellulose (adapted from Hölzer and
Sjögren [61]).
resistant film had been formed on the die wall. As for equation:
microcrystalline cellulose, only the unlubricated product
was tabletted, first at 60 MPa, with a constant m1 value F ¼ mFr þ C ð11Þ
reached after 20 tablets (Fig. 6, right). Then, continuing at where Fr is the radial force acting normal to the frictional
60 or 150 MPa did not affect the m1 value, nor die cleaning force F; and C the adhesion of the powder to the die wall.
after 50 tablets. Generally, change in m1 was due to a change Both m and C can be estimated by measuring of Fd or Fe and
in the axial frictional force Fd and not to the maximum Fr by different compressive forces. A good example is given
radial force Fr : by Kikuta and Kitamori [65] with lactose granules
The second parameter that can affect the measured m combined with lubricants, even though the coefficient of
value is the applied compression pressure. An example is friction at ejection was determined using special equipment
again given by Hölzer and Sjögren [61]. Fig. 7 shows the fitted to a hydraulic press (Table 5). Interestingly, a good
profiles of m1 and m2 vs. maximum applied pressure for correlation was observed between the calculated adhesion
different materials. For sodium chloride and lubricated values C and the binding characteristics of the materials on
lactose m1 and m2 values were fairly constant. These
Table 5
materials behaved ‘normally’ as both gave frictional forces,
Relation between coefficient of friction, m2 ; and C values obtained on a
Fd ; and ejection forces, Fe ; proportional to the applied specially equipped hydraulic press and the adhesion tendency during
pressure. In contrast, alprenolol hydrochloride and micro- experimental testing and tabletting on a rotary machine (adapted from
crystalline cellulose exhibited decreasing m1 with increas- Ref. [65])
ing Pa ; but the reasons for that are different for these two Lubricant Concen- Mixing m C Adhesiona
materials: for the former a non-linear increase of Fd and for tration time (kg/cm2)
the latter of Fr with applied pressure can be invoked. As for (%) (min)
m2 it was reasonably constant, but this could be explained by Static Dynamic
a compensation effect between Fe and Fr0 : These two
Mg stearate 0.1 1 0.20 12 2 þþ
materials showed a relaxation of the radial stress above a 10 0.19 8 2 ^
certain applied pressure, due to capping in the case of 30 0.18 4 2 2
alprenolol hydrochloride. It is also worthy of note that for 0.3 1 0.22 5 2 2
this compound the ‘kinetic’ coefficient of friction, m2 ; was 10 0.17 3 2 2
30 0.12 2 2 2
higher than the ‘static’ coefficient of friction, m1 : According
to the authors the reason for this discrepancy could be the Talc 1.0 10 0.48 23 þ þþ
adhesion of material to the die wall. 2.0 10 0.38 14 ^ þþ
3.0 10 0.35 7 2 ^
In fact, mention should be made of the role of adhesion of
a
the material to the die wall according to Coulomb’s Qualitative expression: þ þ, severe; þ, moderate; ^, slight; 2, none.
Table 6 The last example deals with the effect of internal

Frictional parameters of various excipients and drugs compressed at lubrication (i.e. adding a lubricant to the powder mass) vs.
150 MPa (15 kN), both without and with magnesium stearate (adapted from
that of external lubrication (i.e. prelubricating solely the
Ref. [61])
die). In the former case, 0.2% magnesium stearate or 0.5%
Material R Fd h m1 Fe Fr0 m2 stearic acid was added to a direct tabletting caffeine
(kN) (kN) (kN) formulation. In the latter case, the lubricant powder was
dusted on the tools. Tabletting was carried out at 9 and
Microcrystalline cellulose 0.77 2.03 0.42 1.97 0.25 0.19 1.29
Lactose, anhydrous 0.65 5.29 0.39 2.25 3.18 1.19 2.68 15 kN on a rotary press [59]. Both internal and external
Mg stearate 0.10% 0.64 5.44 0.38 2.44 5.10 1.62 3.15 lubrication act favourably on the tabletting process and
Mg stearate 0.25% 0.95 0.81 0.42 0.34 0.23 0.61 0.38 there are only slight differences in m2 values between the
Mg stearate 0.50% 0.95 0.68 0.44 0.26 0.26 0.63 0.42 compression trials with added lubricant and die prelubrica-
Sodium chloride 0.76 3.63 0.49 1.37 0.86 1.04 0.83
Mg stearate 0.25% 0.93 1.03 0.59 0.33 0.12 0.53 0.23
tion (Table 7). This confirms that the reduction of
Mg stearate 0.50% 0.94 0.97 0.59 0.31 0.11 0.47 0.23 interparticulate friction is not decisive and that die-wall
Mg stearate 1.00% 0.96 0.56 0.60 0.17 0.09 0.37 0.23 prelubrication has no deleterious impact on tablet charac-
Alprenolol HCl 0.86 2.03 0.56 0.72 0.75 0.51 1.47 teristics such as disintegration time and mechanical
Aspirin 0.84 2.46 0.55 0.88 0.16 0.14 1.15
Paracetamol 0.80 3.07 0.38 1.55 1.12 0.65 1.72
strength, in contrast to internal lubrication.
both the special testing instrument (static friction) and an

5. Die-wall compression cycles
actual rotary tabletting machine (dynamic friction). In
contrast, no correlation was observed between coefficient of
friction and adhesion characteristics of the materials. It is Continuous monitoring of the die-wall pressure as a
also interesting to point out that different values of m would result of applied pressure has found use in the determination
have been calculated according to Eq. (10), i.e. by dividing of the material-related compaction characteristics (body
the ejection force by the residual die-wall force. with constant yield in shear or Mohr body, polymorphism,
The benefits of comparing friction coefficients of pseudopolymorphism, particle size and shape) as well as in
materials compressed at constant pressure have been the evaluation of formulation (moisture content, binder,
shown for both excipients and drugs (Table 6). Note that lubricant) or process (compaction speed) variables.
the authors normalised the frictional force Fd and the The following parameters are employed, either alone or
ejection force Fe by the contact surface area between tablet in combination (see Section 3):
and die-wall, as suggested by Rees and Shotton [66]. This
was not done for the sake of coherence with the other data † slopes of the various segments of Long’s compression
reported in the present review. Determination of the cycle,
coefficients of friction helps to characterise the auto- † axial to radial stress ratio at maximum applied
lubricating properties of the materials tested and to optimise pressure, h;
the lubricant concentration (see Section 5.3). In the present † residual die-wall pressure after load release or at
case, the much higher sensitivity of the m1 parameter ejection Pr0 ;
relative to the lubrication ratio R (axially transmitted † hysteresis loop area of Long’s cycle,
pressure) is also demonstrated. † plot of hysteresis area vs. applied pressure.
Table 7
Influence of internal and die-wall lubrication on the tabletting of a caffeine formulation [59]a
Lubrication Compression h Fe (N) Fr0 (N) m2 Disintegration Crushing force (N)

force (kN) time (min)
– 9 0.25 1000 862 1.16 1.1 135

Mg stearate 0.2% 9 0.41 340 718 0.42 1.8 126
Die wall, Mg stearate 9 0.42 60 646 0.09 1.0 149
Stearic acid 0.5% 9 0.36 160 790 0.33 2.3 101
Die wall, stearic acid 9 0.40 240 718 0.33 0.7 137
– 15 0.33 1160 1384 0.85 4.5 .200
Mg stearate 0.2% 15 0.46 340 1005 0.34 10.2 .200
Die wall, Mg stearate 15 0.50 160 933 0.17 5.3 .200
Stearic acid 0.5% 15 0.42 450 1149 0.39 9.2 167
Die wall, stearic acid 15 0.42 300 1292 0.23 4.3 .200
a
Caffeine formulation: 39% microcrystalline cellulose, 59% anhydrous lactose, 2% caffeine.
Table 8
Parameters from Long’s compression cycles of representative materials
Material Pa (MPa) Segment slopes h Pr0 (MPa) Ref.
OA AB BC CD
Paracetamol 175 0.17 0.45 0.35 3.00 n.a. n.a. [73]

180 0.14 n.a. n.a. n.a. 0.43 8 [43]
Paracetamol DCa 175 0.23 n.a. n.a. n.a. 0.37 20 [43]
Lactose SD 170 0.22 n.a. n.a. n.a. 0.35 25 [43]
235 0.49 n.a. n.a. n.a. 0.71 61 [31]
Dicalcium phosphate 175 0.23 0.45 0.24 1.00 n.a. n.a. [73]
325 0.47 n.a. n.a. n.a. 0.68 103 [31]
Sodium chloride, cubic 166 0.26 0.37 0.04 0.71 0.33 45 [74]
Sodium chloride, dendritic 170 0.33 0.47 0.10 0.23 0.42 52 [74]
n.a., not available.

a
Direct compression paracetamol containing 4% gelatine hydrolysate.
5.1. Material behaviour Table 8 lists some slope values as well as stress ratios and
residual die-wall pressures of representative materials. As a
A knowledge of the shear failure of a given material is matter of fact, values of unity for slopes AB and CD were
pharmaceutically important in many areas: successful never observed for sodium chloride, known to deform
tabletting, sensitivity to lubricant, tablet disintegration or plastically. Reasons for discrepancies have already been
dissolution. Materials with constant shear stress (plastic or discussed, in particular the fact that these materials are not
ductile materials) are generally presented as forming models of solid, isotropic bodies. The present authors are
coherent tablets whereas Mohr bodies (brittle materials) also convinced that these inconsistencies result in many
are prone to capping and lamination. This simplistic view cases from defect in die calibration. It comes out that
has to be modulated by taking into consideration the elastic classifying materials based on slope calculation is not
and viscoelastic properties of the materials Moreover, reliable.
Hiestand [58,67 –69] is of the opinion that possible particle Another way of gaining insight from the radial –axial
fragmentation during the compression phase is of little pressure cycle is to look at the area of the hysteresis loop,
importance as the strength of the compact depends on the which reflects the extent of departure from ideal elastic
behaviour of the material on unloading. behaviour. This approach has been followed by Touré [57]
when analysing the effect of moisture content on the
Many investigators have attempted to relate the shape of
compactibility of starch granules. A good correlation was
the compression cycles of model materials to their ability to
observed between hysteresis area and tablet mechanical
(or not to) form good compacts. For the present discussion
strength. However, here a single material was investigated.
we will consider representative materials of the various
A recent work by Khossravi and Morehead [39]
types reported in the literature [37,71,72].
compared the hysteresis areas of the 1st and 5th com-
Regarding paracetamol, known to cap at ejection, Leigh
pression cycles, representing the quasi-equilibrium state
et al. [24] have concluded that its behaviour resembles that
(Table 9). Considering only the hysteresis areas of the 1st
of a Mohr body, while that of paracetamol granulated with
cycle, the values are quite similar except for the plastically
3% povidone (which does not cap) is similar to the
deforming sodium chloride. In particular the non-tablettable
behaviour of a body with constant yield stress. A
classification with respect to constant stress was also
proposed for sucrose and sucrose granules. It is noticeable Table 9
that the authors rid themselves from the requirement for the Hysteresis areas for the 1st and 5th compression cycles along with the
percent change of representative materials compressed at 225 MPa [39]
slope of segments AB and CD to be unity and only rely on
the equality of both slopes (see Fig. 5 and Table 4). As for Material Hysteresis area (MPa2)
sodium chloride, it was declared as a body with constant
1st cycle 5th cycle Percent change
yield stress at low applied pressures, with a Mohr behaviour
at high pressures [24]. Microcrystalline cellulose 4860 2434 69.9
Later, Shotton et al. [43,45,70,73] found that paraceta- Sodium chloride 12,725 2795 78.0
mol, both with and without a binder, exhibited compression Aspirin 4300 3357 21.9
cycles resembling a Mohr body. Obiorah and Shotton [43, Lactose monohydrate 5707 1490 73.9
Sucrose 6340 1108 82.5
74] also classified sodium chloride as a Mohr body, in
Paracetamol 6290 2720 56.8
contrast to Long [52].
paracetamol displays a hysteresis area noticeably higher correlation was found between the hysteresis loop areas of
than that of microcrystalline cellulose, confirming that the lower punch and of die-wall pressures versus applied
value calculated for the first compression –decompression pressure [57,75].
cycle is not an indicator of material compactibility, but of Long’s compression cycles and axial to radial stress
the extent of elastic behaviour. Comparing now the areas for ratios are in fact not much influenced by the type of material
the 1st and 5th cycles, it is not quite clear that they are considered, because the measured radial pressure contains
comparable for plastic materials (sodium chloride and an elastic component. In contrast, residual die-wall pressure
aspirin), in contrast to the case of brittle materials (lactose reflects irreversible deformation (except axial expansion).
monohydrate, sucrose, paracetamol), as stated by Khossravi The first investigation into the nature of the residual die-wall
and Morehead [39]. pressure was that by Higuchi et al. [40]. First, the authors
Another type of investigation of the consolidation observed a correlation between the ejection force and the
behaviour of materials is to simply consider the axial to residual die-wall pressure. Second, they could show that this
radial stress ratio, h; relying on the observation of Wind- lateral die-wall response actually decays with time, reaching
heuser et al. [22] that materials which permit good (under the conditions used) equilibrium within roughly
conversion of axial pressure to radial pressure tend to 2 min, at a rate dependent on crystal hardness and
form a good compact. Since then, a similar conclusion has interparticle friction. Later, Shotton and Obiorah [43,73,
been drawn by many authors but seems essentially valid 74] found that materials forming satisfactory tablets
when considering the effect of an additive—most often a displayed high residual die-wall pressures whereas values
binding agent—on the compression cycle of a given were low for materials likely to cap or laminate. This would
material (see Section 5.2). This is not the case when suggest that the Pr0 parameter is a measure of the compact
materials of different types are compared [33,37 – 39,42,43]. capability to withstand the radial force exerted by the die.
Interestingly, it has been observed by Higuchi et al. [40] and Unfortunately, viscoelastic materials like microcrystalline
by Ridgway et al. [41] that axial to radial transmission cellulose or dextrates were not involved in the above-
increases as the material hardness decreases, probably due mentioned studies, and in fact such materials form strong
to the lower yield strength allowing increased area of tablets although exhibiting Pr0 values as low as for instance
contact and thus stress transmission. the poorly compactible paracetamol [37,61]. This indicates
Long’s compression cycles have been interpreted in that the compact had recovered axially, either elastically or
terms of hysteresis area as a function of applied pressure in by brittle failure. In fact, conflicting results were reported on
order to distinguish bodies with constant yield stress from elastic recovery of microcrystalline cellulose compacts [37,
Mohr bodies [54,57]. The data of Leigh et al. [24] were re- 38], possibly because the mechanism of compaction varies
analysed with the conclusion that sodium chloride behaves with applied pressure as shown for the elastic recovery as
like a Mohr body [54,57]. As noted by Krycer and Pope [5], measured after tablet ejection (Fig. 8, left) and for the
this does not resolve the conflict with independent studies residual die-wall pressure (Fig. 8, right). Note that Krycer
showing that sodium chloride consolidates by plastic et al. [38] also measured residual die-wall pressure after
deformation rather than by brittle fracture. A Mohr multiple compression but this apparently did not reveal
behaviour was also observed for a starch granulation using more information regarding the compaction mechanism.
the aforementioned analysis [54,57]. In addition, a good In another report the same investigators warned that
Fig. 8. Residual die-wall pressure, Pr0 ; vs. maximum applied pressure, Pa ; plot (left) and elastic recovery, ER, vs. lower punch work plot (right) for various
model materials. Key: X, microcrystalline cellulose; W, D.T. sucrose; O, mannitol; K, paracetamol [38].
the capping tendency cannot always be predicted from the residual die-wall pressure was much higher for fine
residual die-wall pressure and proposed a capping index to particles that resulted in coherent tablets whereas capping
be calculated from the gradient of elastic recovery vs. Pr0 was noticed for coarse material [37]. Generally, compression
plot [76]. A noteworthy point is that microcrystalline cycle profiles were not very much affected by particle shape,
cellulose was not included in the second paper. In fact, either for the plastically deforming sodium chloride [38,40]
elastic recovery was measured after tablet ejection and may or the brittle material lactose [70]. The only significant
not be indicative of the expansion in the die, where failure difference was a reduced axial transmission of the applied
occurs. At the least, comparative data on more materials are pressure and thus an increased coefficient of friction in case
needed to evaluate the benefit of measuring residual die- of dendritic sodium chloride compared to cubic sodium
wall pressure. In contrast, it is a meaningful indicator of chloride. This was also evident for more irregular lactose
possible friction at ejection. particles that also induced a higher ejection force and thus
Finally, it is of interest to notice that differences in die- friction coefficient. As for the crushing strength of lactose
wall response (shape of the pressure cycle, stress ratio and/or tablets, it was superior for regular particles probably because
residual radial pressure) were observed for different poly- of better densification. In any case, one has to recognise that
morphs of sulfathiazole or barbital [77] and for pseudopo- reports on the effect of particle size and shape are too few to
lymorphs of lactose [43,70] or magnesium stearate [78]. formulate general conclusions. Finally, mention should
again be made of the work of Summers et al. [77] who
5.2. Influence of particle size and shape obtained significant differences in the Long’s compression
cycles for various aspirin crystal habits.
Very few studies have been published on the effect of
particle size or particle shape (crystal habit) on the 5.3. Formulation and process variables
parameters derived from die-wall pressure measurement
[24,37,43,70,74,77]. For the plastically deforming sodium Investigations dealing with the effect of formulation
chloride changing particle size did not significantly affect variables on die-wall response are mainly concerned with
the residual die-wall pressure and ejection force and thus two types of additives, namely binders and lubricants, as
coefficient of friction m2 : However, coarser particles well as with the moisture content of the mass to compress.
generated better axial to radial transmission at maximum With the exception of dicalcium phosphate [45], all
pressure, and weaker tablets. This contrasts with the similar studies involving binders report on the compactibility of
product potassium chloride (parallel increase in h and tablet paracetamol for obvious reasons [24,38,43,45,70,73,76].
mechanical strength) [37]. Regarding the brittle material They generally show an improvement of the tablet
sucrose, the residual die-wall pressure did not vary with mechanical strength, but the effect on the radial trans-
particle size, whereas a decrease in ejection force was mission of the applied pressure is rather controversial. An
observed for coarse particles, and thus a reduced m2 value. example is given by the work of Doelker and Shotton [45]
During compression, the radially transmitted pressure was for both paracetamol and dicalcium phosphate granulated
not influenced by particle size, but weaker tablets were with 4% povidone, maize starch or methylcellulose. For
produced with coarser crystals [24,37]. A similar pattern paracetamol, the radial conversion during compression was
was obtained with the brittle, but poorly compactible lower in presence of a binder, whereas residual die-wall
paracetamol. The only remarkable difference was that pressure increased (Fig. 9, left). An opposite trend was
Fig. 9. Pressure cycles for (A) paracetamol powder and granules, and (B) dicalcium phosphate powder and granules. Key: B, powder; K, povidone granules; P,
maize starch granules; A, methylcellulose granules [45].
observed for dicalcium phosphate: increased radial trans- adhesion to the die wall occurred. However, it is our
mission and slightly reduced residual die-wall pressure opinion that similar conclusions could also be drawn if only
(Fig. 9, right). Fd (or R) and Fe had been considered.
Other reports on paracetamol are generally in line with The moisture content of the mass often influences
the above-mentioned study [43,70,73], but work by Carless compression characteristics. In fact, a small proportion of
et al. [24,42] displays increased h values for direct moisture is generally present in the formulation and, in some
compression paracetamol. cases, this is necessary to form a tablet. Thus, paracetamol
It is evident from most studies that a good radial granules are satisfactorily compressed only if the moisture
transmission of pressure alone is not sufficient to explain content is above 1% [2,79].
compactibility, as exemplified by paracetamol. On the other An investigation dealt with the effect of moisture content
hand, for paracetamol again, tablet formation was related to of sodium chloride [80,81]. At low level (0.02%) die-wall
a higher residual die-wall pressure, but the strongest tablets friction was increased probably because of an increase in the
(with povidone as a binder) were not associated with the radial to axial stress ratio. At moisture contents above
highest Pr0 value [45]. Probably, the effect of particle size 0.55% friction was reduced at all pressures owing to the
and that of material hardness play a prominent role besides lubricating effect of water. Studies on the effect of moisture
that of the binder on the axial to radial pressure content on tabletting are actually scant and those reporting
transmission. die-wall monitoring even rarer. We have already commen-
As already mentioned, for measuring the effect of the ted on the work of Touré [54,57,75]. That of Obiorah [43,
type and concentration of lubricant during compression 73] relates to the influence of moisture content on the
most authors rely on the axial frictional force, Fd ; the ratio behaviour of the two poorly compactible products para-
of lower to upper punch force, R (the so-called lubrication cetamol and phenacetin, both as powders and granules.
ratio), or eventually the work of friction if the press is Adding water yielded coherent tablets and led to a slight
equipped with displacement transducers, the residual force increase in the OA slope values of Long’s cycles, the
on lower punch before ejection, Fb0 ; and the ejection force, radially transmitted pressures and the residual die-wall
Fe : Generally, good correlations are obtained with the pressures.
performance of the lubricant, especially with Fe : To A final factor likely to affect radial response is the speed
evaluate the possible benefit of monitoring die-wall of compression, through its influence on densification and
pressure, a look at the literature led only to the investi- permanent deformation. Actually, when increasing the
gations by Hölzer and Sjögren [61] and Schrank-Junghäni compression speed of the rotary machine by 60 (contact
et al. [59]. The latter work on the effect of magnesium time of 0.17 instead of 10 s), Wiederkehr-von Vincenz [37]
stearate and stearic acid on the tabletting performance of a did not observe significant changes in the stress ratio and
caffeine formulation has already been commented upon in residual die-wall pressure of sodium chloride, potassium
terms of coefficient of friction. It can be added that the chloride and sucrose.
lubricants (and die-wall prelubrication) drastically
increased the axial to radial stress transmission and the
residual die-wall pressure. As for Hölzer and Sjögren [59], 6. Which parameters are of importance?
they have fully reported (except the residual lower punch
force) on the various parameters resulting from the tablet- When looking into the literature reporting on radial
ting of anhydrous lactose and sodium chloride combined pressure measurement, we are confronted with both
with magnesium stearate (Table 6). Regarding friction at inconsistencies and lack of a full complement of parameters
maximum applied pressure, the axial to radial transmission that would help to understand their comparative benefits. In
ratio slightly increased in presence of magnesium stearate, particular, we were interested in gaining insight into the
in agreement with the known fact that lubricants favour effect of the material characteristics on the parameters
densification. In parallel, the friction force (as assessed related to Long’s compression cycles and not so much to
through R or Fd ) decreased for lubricated materials. It is that of lubrication. We thus decided to carry out a complete
interesting to note that addition of 0.1% magnesium stearate investigation of the compression characteristics of 13
was not sufficient to prevent adhesion of lactose to the die materials, representative of the two broad categories as
wall. Friction coefficient also diminished because the effect reported in the literature from independent studies:
of lubricant was much more pronounced on the frictional
force than on the radial transmission. † 8 materials known to generate high tablet strength:
The same pattern was true when the friction at ejection sodium chloride, potassium chloride, paracetemol DC
was considered, but the effect of the lubricant on the radial with 5% gelatine (Mallinckrodt), powdered cellulose
transmission was more marked. A point raised by the (Elcemaw P 100), microcrystalline cellulose (Avicelw
authors is that the friction coefficient, m1 ; was 0.2– 0.4 for PH 101), pregelatinised starch (Starch w 1500),
well-lubricated materials, 0.7 – 2 for poorly lubricated dicalcium phosphate dihydrate (Emcompressw) and
materials but non-sticking materials, and above 2 if nitrofurantoin,
† 5 materials known to yield low tablet strength: maize the symbols defined previously [13]. The assignment of the
starch, hexamine, sodium bicarbonate, phenacetin, and materials to one class or the other is based on the work of
paracetamol. Nyström et al. [71,72].
It is logical to relate the energy brought by the press (as
A 100– 200 mm size fraction was generally used, except mechanical work done by the upper punch, Wu ) and the
for the proprietary products. Each material was compressed tablet strength, Fc : However, it has to be recognised that the
at 150 MPa on a single punch machine (Korsch EK-0) compaction work is used to (i) cause packing and
instrumented as described by Rime et al. [82]. In addition, rearrangement of primary and subsequently of broken
the radial pressure was measured using a cut-away die particles, (ii) induce elastic deformation, plastic defor-
instrumented and calibrated in a similar way to that of mation and/or brittle fracture of the material, and (iii) to
Hölzer and Sjögren [33]. Polynomials were used to account overcome friction between particles and friction between
for the effects of applied pressure, compact height and lower particles and the die. The most decisive factor (ii) is
punch position. The weight of each material was calculated material-dependent (particle rearrangement is not highly
from its true density to provide a 2-mm high tablet at zero energy demanding and, here, external friction is low
porosity using a 12-mm flat-faced punch and die set. In because the die was prelubricated). In this respect, materials
order to elucidate the densification characteristics of the exhibiting strong attractions and high surface bonding are
materials, the die was prelubricated by dusting with more difficult to densify. In this respect, powdered and
magnesium stearate powder. Thus, friction related par- microcrystalline cellulose, where considerable hydrogen
ameters will not be reported here (R . 0:91; Fe , 600 N). bonding is present, show high WU values compared to the
For each material, both single and double compression materials yielding unsatisfactory tablets, even though all
cycles were performed in order to determine the net work of materials had been compressed at the same pressure. The
compression. The data are summarised in Table 10, using work of expansion, WEXP ; is the recoverable part of the
Table 10
Compression characteristics of representative materials tabletted at 150 MPa (unpublished data)
Material type WU WEXP PL1 PL2 Py Segment slopes h Pr0 A1 A2 1c 1 ER0 ER Fc

(J) (J) (%) (%) (MPa) (MPa) (MPa2) (MPa2) (%) (%) (%) (%) (N)
OA AB BC CD
High tablet strength

Plastic materials
Sodium chloride 7.4 1.1 84 90 73 0.27 0.47 0.19 0.26 0.37 19.2 2905 2076 5.8 7.0 0 1.1 194
Potassium chloride 6.6 0.7 90 88 44 0.28 0.58 0.26 0.71 0.44 25.5 3696 2149 0.9 4.0 0 3.4 192
Paracetamol DC 7.3 1.0 86 79 76 0.28 0.37 0.22 0.42 0.32 12.3 2036 1505 5.0 9.5 2.4 4.7 111
Viscoelastic materials
Powdered cellulose 10.5 1.3 88 79 75 0.29 0.42 0.26 0.37 0.35 5.3 1545 1275 4.1 13.8 4.7 10.3 147
Microcrystalline cellulose 11.3 1.1 90 85 55 0.29 0.47 0.26 0.41 0.35 4.3 1197 1130 1.1 7.9 2.0 7.0 426
Pregelatinised starch 8.2 1.5 82 67 41 0.29 0.54 0.34 0.57 0.40 3.7 2264 1241 2.2 13.5 3.8 11.4 76
Brittle, plastic materials
Dicalcium phosphate 7.1 0.6 92 85 215 0.25 0.31 0.19 0.34 0.30 10.3 1528 1571 17.2 19.0 0.8 2.1 75
Nitrofurantoin 7.4 0.8 90 83 124 0.33 0.47 0.23 0.38 0.39 11.2 1994 1668 11.5 15.2 2.2 4.3 108
Low tablet strength

Plastic materials
Maize stach 9.1 1.1 87 75 58 0.28 0.56 0.31 0.52 0.39 4.5 2357 1302 2.0 13.0 2.4 10.8 16
Hexamine 4.7 0.7 85 83 41 0.31 0.54 0.24 0.89 0.44 9.5 3707 2361 0 3.0 0 4.0 25
Sodium bicarbonate 6.9 0.7 89 86 117 0.26 0.33 0.18 0.44 0.30 11.3 1762 1546 11.0 10.8 0 2.2 34
Brittle, elastic materials
Phenacetin 3.3 0.5 86 62 53 0.33 0.50 0.23 0.46 0.40 8.5 2274 1566 1.4 C/La 2.2 C/La C/La
Paracetamol 3.6 0.8 78 57 69 0.28 0.37 0.22 0.42 0.33 6.6 2095 1947 1.2 C/La 2.4 C/La C/La
WU ; upper punch work; WEXP ; work of expansion (upper punch); PL1 (plasticity index 1) ¼ 100(WU 2 WEXP)/WU; PL2 (plasticity index 2) ¼ WN(net
work)/WL (lower punch work, 1st compression cycle), where WN ¼ WL 2 WL (lower punch work, 2nd compression cycle, or work of elastic deformation);
Py ; yield pressure (stress); h; axial to radial stress ratio; Pr0 ; residual die-wall pressure; 1c ; compact porosity at maximum compression pressure (150 MPa); 1;
porosity after compact ejection; ER0, elastic recovery within the die (compact height when pressure is returned to zero); ER, elastic recovery after compact
ejection; A1 ; hysteresis area of the Long compression cycle (radial pressure–upper punch pressure cycle); A2 ; hysteresis area of lower punch pressure–upper
punch pressure cycle; Fc ; compact crushing force.
a
Not measurable because compact capping and/or lamination.
compression work done by the powder compact on the compounds and shows that classifying material behaviour
receding upper punch, and thus incompletely represents the on slope calculation is not reliable.
elastic deformation. Moreover, elastic recovery also takes Two other parameters can be derived from Long’s
place after decompression, both within the die and after pressure cycles. A high axial to radial stress transmission at
tablet ejection. It is, however, a rough indicator used by maximum pressure was originally associated with resistant
some authors to estimate the elastic component, alterna- compacts [22]. Many studies have in fact invalidated this
tively to the plasticity index 1 derived from WEXP : Double statement on comparing different materials and, here again,
compression is frequently performed to quantitate better the this criterion was not fulfilled. However, when examining h
work of elastic deformation and calculate the net work of values for hexamine, potassium chloride and sodium
compression and the plasticity index 2 [13]. The higher the chloride, the pressure transmitted radially appeared to
PL2 value, the more work received by the compact. In this decrease with increasing hardness as described by Ridgway
sense, low values were recorded for the very poorly et al. [41].
compactible phenacetin and paracetamol, but it has to be Much attention has been given to the residual die-wall
stressed that for many materials plastic deformation still pressure as it relates to the irreversible deformation of the
takes place beyond two compression cycles [38]. material during compaction. High Pr0 values were recorded
Mean yield pressure, as calculated from Heckel plot, is by Shotton and Obiorah [43,70,73,74] for materials yielding
also a measure of the extent of plastic deformation. Thus, sound tablets in contrast to materials prone to capping or
the plastic and viscoelastic materials showed relatively low laminating, supposedly because of pronounced axial
Py values, in line with published data, with the noticeable recovery by brittle fracture in the decompression phase.
exception of sodium bicarbonate for which a high yield Later, this was not observed for viscoelastic materials [37,
value (together with a very low compact strength) was 38,61]. Our findings confirm literature results, particularly
recorded. In fact, it is confirmed that several so-called in that viscoelastic materials are able to withstand axial
plastically deforming materials do not possess adequate elastic recovery.
plasticity to develop large contacting areas [71,72]. Note The significance of the residual die-wall pressure for
that the validity of the Py values listed here is somehow tablet strength is tentatively evaluated by considering the
restricted by the fact that the relative density of the compact two models proposed by Hiestand [67] for tablet formation.
was determined using an in-die method [83], so that an The ‘weak model’ (brittle mechanism) assumes only elastic
elastic component was included in the Py values that may be deformation during unloading while the ‘strong model’
falsely too low. (ductile mechanism) includes plastic deformation during
With respect to the parameters inferred from die-wall elastic deformation. Both ductile extension and viscoelas-
pressure monitoring, it is confirmed that no material ticity contribute to the strength of the tablet. This can be
displays a radial vs. axial pressure profile similar to the correlated with the fact that the residual die-wall pressure
theoretical cycles described by Long [52,53]. Significant will be determined by the shear strength of the compact if
differences between materials are not discernible, as already remaining intact during decompression. Our results show
quoted for instance by Krycer et al. [38]. In particular, that the residual radial pressure, alone, does not provide
values of the slopes of the segments OA (OA0 ) and BC sufficient predictive information for tablet mechanical
(B0 C0 ) were too low to calculate relevant Poisson ratios. strength.
Thus, values of 0.21, 0.22, 0.22 and 0.24 were obtained for A final parameter calculated from the radial vs. axial
sodium chloride, potassium chloride, microcrystalline pressure cycles is their surface area. Areas of the loops were
cellulose and hexamine, respectively, in disagreement generally larger for plastically deforming materials, indi-
with those reported in the literature (0.25 for sodium cating the departure from ideal elastic behaviour, compared
chloride [84], 0.27 for potassium chloride [84], 0.30 for with viscoelastic and brittle materials. It would be
microcrystalline cellulose [85] and 0.18 for hexamine [86]). preferable to consider successive compressions in order to
Slopes of segment AB or CD were never equal to unity for discern the elastic component. Following the suggestion of
materials known to deform by plastic flow (see Table 4), Touré and Carstensen [54,57] these data were also analysed
confirming their unsuitability in classifying compaction in terms of hysteresis areas of lower punch pressure vs.
behaviour. Among the various reasons already discussed for upper punch pressure cycles. A rather good correlation was
this failure (see Section 3.2), it can be argued that plastic observed as can be seen in Fig. 10.
flow beyond pressure A (A0 ) of the cycle could be in fact As already mentioned, the degree of success in tablet
mainly confined to the interparticular regions and total forming is primarily dictated by the extent of survival
plastic flow of the material itself has not occurred [42]. This during decompression of the bonding area created during
is probably why the same group [24] proposed to consider the compression phase [58]. Two extreme situations can be
only the requirement of equality of slopes AB and CD, in envisaged during decompression: elastic and plastic defor-
contrast to slopes A0 B0 and C0 D0 (see Table 4). Values mations. Elastic deformation can also operate after tablet
reported in Table 10 do not permit clarification of the ejection. It is therefore of utmost interest to quantitate the
inconsistencies reported in the literature for the same in-die and out-of-die expansion. This has been done by
are able to characterise the mechanical properties of the

materials of the various categories defined.
7. Concluding remarks
Interest in die-wall instrumentation is continuing as

recognised at the 2003 AAPS Annual Meeting [87].
However, when examining the literature dealing with radial
pressure measurement during compression of pharmaceu-
tical materials, one is confronted with a lot of inconsis-
tencies in the authors’ conclusions. One reason is the
technical difficulties of die instrumentation and calibration
Fig. 10. Plot of the hysteresis areas of the lower punch vs. hysteresis areas that lead to discrepancies in the results reported for similar
die wall for five classes of model materials (see Table 10). Key: High tablet products. Another reason is that the models used for
strength: O, plastic materials; B, viscoelastic materials; X, brittle, plastic
materials. Low tablet strength: K, plastic materials; W, brittle, elastic
interpreting data have been proposed mostly for solid,
materials. isotropic bodies and not for powders.
Based on literature examination and on some original
results presented here, the authors are of the opinion that
Carless and Leigh [42] and by Wiederkehr-von Vincenz most parameters derived from die-wall pressure measure-
[37] for similar model compounds. They observed that ment do not bring much to the tablet formulation scene
elastic recovery occurred mainly after tablet ejection, but when considering the challenges of low friction during
these authors were of the opinion that tablet capping and/or compression and of yielding strong tablets. In fact, it
lamination, if present, most probably had taken place during appears that optimisation of many factors, in particular the
decompression or just at ejection. Long [52] actually type and concentration of additives such as binders and
considers that axial separation is the result of relief of the lubricants, can be achieved solely through instrumentation
residual pressure exerted by the die wall. We have reached of the punches.
the same conclusion, noticing that the change in porosity Die-wall instrumentation is, in contrast, of help for basic
and thus in elastic recovery is generally more pronounced compaction studies aimed at determining the mechanical
for brittle materials. characteristics of materials, considered as the main factor
Finally, based on our results and those from the literature, accounting for the differences in tabletting behaviour. It is
in particular those of Wiederkehr-von Vincenz [37] and of also useful for elucidating the friction phenomena during
Nyström et al. [71,72], we propose a set of parameters that compaction and the related tabletting problems (capping,
could help classifying the various materials used in lamination, tooling wear). In this respect, the axial to radial
pharmaceutical tabletting and to understand their behaviour pressure transmission ratio as well as the slopes and
(Table 11). It is apparent that none of these parameters is hysteresis surface area of the so-called Long pressure cycles
able, alone, to predict the strength of the tablets produced are of limited value compared to the residual die-wall
(actually, models still remain to be set up), but overall they pressure. In particular, slope values do not permit to classify
Table 11
Comparative levels of compression parameters in relation to various classes of materials
Parameter High tablet strength Low tablet strength
Plastic materials Viscoelastic materials Brittle, plastic materials Plastic materials Brittle, elastic materials
Work of compression, Wu Medium High Low Medium Low

Plasticity index, PL2 High Variable Medium Medium Low
Yield pressure, Py Low Low High Medium Medium
Initial curvature of Heckel plota Absent Absent Quite long Absent Present
Stress ratio, h Variable Variable Variable Variable Variable
Residual die-wall pressure, Pr0 High Low Medium Medium Low
Hysteresis of the Long cycle, A1 High Low Medium Medium Medium
Elastic recovery, ER0 Low Medium Low Low Low
Elastic recovery, ER Low High Low High –b
a
After deduction of the densification due to particle slippage and rearrangement.
b
Not relevant.
the powdered materials as bodies with constant stress in [16] G. Schwartz, The instrumented tablet press: uses in research and
shear or as Mohr bodies. The residual die-wall pressure is of production, Pharm. Technol. 5 (9) (1981) 102 –132.
[17] K. Marshall, Instrumentation of tablet and capsule filling machines,
greater importance because the success of tablet formation Pharm. Technol. 7 (3) (1983) 68 –82.
is determined by the decompression and ejection phases, but [18] P. Ridgway Watt, Tablet Machine Instrumentation in Pharmaceutics:
more emphasis should probably be placed on studying the Principles and Practice, Ellis Horwood, Chichester, 1988, pp. 288 –
kinetics of time-dependent deformation, again especially in 298.
[19] K. Marshall, Monitoring punch forces and punch movements as an aid
the later period of the compression cycle. Finally, equations
to developing robust tablet formulations, Drug Dev. Ind. Pharm. 15
taking into account the material compressibility (volume (1989) 2153– 2176.
reduction under pressure) have been proposed, but as for the [20] H.S. Thacker, Instrumentation of tablet machines, in: N.G. Stanley-
relation between the tablet strength (material compactibil- Wood (Ed.), Enlargement and Compaction of Particulate Solids,
ity) and the compression parameters inferred from die-wall Butterworths, London, 1983, pp. 227– 240.
[21] E. Nelson, L.W. Busse, T. Higuchi, The physics of tablet
monitoring, models still need to be developed. compression. VII. Determination of energy expenditure in the tablet
compression process, J. Am. Pharm. Assoc. Sci. Ed. 44 (1955)
223–225.
[22] J.J. Windheuser, J. Misra, S.P. Eriksen, T. Higuchi, Physics of tablet
References compression. XIII. Development of die-wall pressure during com-
pression of various materials, J. Pharm. Sci. 52 (1963) 767– 772.
[1] J. Barra, E. Doelker, Instrumentation of an eccentric tablet press, in: [23] K. Ridgway, The use of photoelastic techniques in the measurement of
A. Muños-Ruiz, H. Vromans (Eds.), Data Acquisition and Measure- die-wall stress in tabletting, J. Pharm. Pharmacol. 18 (1966)
ment Techniques, Interpharm Press, Inc, Buffalo Grove, USA, 1998, 176S–181S.
pp. 189– 238. [24] S. Leigh, J.E. Carless, B.W. Burt, Compression characteristics of
[2] E. Shotton, J.A. Hersey, P.E. Wray, Compaction and compression, in: some pharmaceutical materials, J. Pharm. Sci. 56 (1967) 888–892.
L. Lachman, H.A. Lieberman, J.L. Kanig (Eds.), The Theory and [25] K. Marshall, PhD thesis, University of Bradford, 1970.
Practice of Industrial Pharmacy, 2nd ed., Lea & Febiger, Philadelphia, [26] K. Ridgway, P.H. Rosser, The application of photoelastic techniques
PA, 1976, pp. 296– 320. to a rotary tabletting machine, J. Pharm. Pharmacol. 23 (1971)
[3] E. Doelker, Physique de la compression: Intérêt et limite des machines 202S–209S.
instrumentées pour l’optimisation de la formulation, Pharm. Acta [27] N.J. Edwards, PhD thesis, University of London, 1973.
[28] R. Spang, PhD thesis, E.T.H., Zürich, 1973.
Helv. 53 (1978) 182 –188.
[29] U. Conte, G. Colombo, A. La Manna, C. Caramella, La mesure de la
[4] T.M. Jones, Preformulation studies to predict the compaction
force radiale à l’aide d’une matrice divisée. Applications aux
properties of materials used in tablets and capsules, Acta Pharm.
problèmes pharmaceutiques, Il Farmaco 11 (1977) 551 –559.
Technol. Suppl. 6 (1978) 141 –159.
[30] E.G. Rippie, D.W. Danielson, Viscoelastic stress/strain behavior of
[5] I. Krycer, D.G. Pope, The interpretation of powder compaction data—
pharmaceutical tablets: analysis during unloading and post-com-
a critical review, Drug Dev. Ind. Pharm. 8 (1982) 307 –342.
pression periods, J. Pharm. Sci. 70 (1981) 476– 482.
[6] E. Doelker, Developments in compression: compressional tests as an
[31] H.G. Cocolas, N.G. Lordi, Axial to radial pressure transmission of
aid in tablet formulation, in: D.D. Breimer, P. Speiser (Eds.), Topics in
tablet excipients using a novel instrumented die, Drug Dev. Ind.
Pharmaceutical Sciences, 1983, pp. 371– 386, Amsterdam, The
Pharm. 19 (1993) 2473– 2497.
Netherlands.
[32] C. Yeh, S.A. Altaf, S.W. Hoag, Theory of force transducer design
[7] K. Marshall, Compression and consolidation of powdered solids, in:
optimization for die wall stress measurement during tablet compac-
L. Lachman, H.A. Lieberman, J.L. Kanig (Eds.), The Theory and
tion: optimization and validation of split-web die using finite element
Practice of Industrial Pharmacy, 3rd ed., Lea & Febiger, Philadelphia, analysis, Pharm. Res. 14 (1997) 1161–1170.
PA, 1986, pp. 66–99. [33] A.W. Hölzer, J. Sjögren, Instrumentation and calibration of a single
[8] E. Doelker, Recent advances in tableting science, Boll. Chim. Farm. punch press for measuring the radial force during tableting, Int.
127 (1988) 37– 49. J. Pharm. 3 (1979) 221 –230.
[9] E.G. Rippie, Compression of solids and compressed dosage forms, in: [34] J. Millet, J. Paris, D. Duchêne, F. Puisieux, Etude de comprimés. XII.
J. Swarbrick, J.C. Boylan (Eds.), Encyclopedia of Pharmaceutical Mise au point d’un nouvel équipement pour la mesure des forces au
Technology, Marcel Dekker, New York, 1990, pp. 149 –165. niveau des différents organes d’une machine à comprimer alternative,
[10] P.E. Wray, The physics of compaction revisited, Drug Dev. Ind. Pharm. Acta Helv. 50 (1975) 109 –115.
Pharm. 18 (1992) 627–658. [35] P.D. Huckle, M.P. Summers, The use of strain gauges for radial stress
[11] M. Celik, Overview of compaction data analysis techniques, Drug measurement during tableting, J. Pharm. Pharmacol. 37 (1985) 722–725.
Dev. Ind. Pharm. 18 (1992) 767–810. [36] D. Duchêne, A. Djiane, F. Altlas, J. Clerc, F. Puisieux, Etude des
[12] A.F. Rime, E. Doelker, Caractéristique de compression et comprim- comprimés. V. Equipement d’une machine à comprimer alternative
abilité des poudres de polymères à usage pharmaceutique, STP pour la détermination des forces intervenant lors de la compression,
Pharma Sci. 3 (1993) 109 –129. Ann. Pharm. Fr. 30 (1972) 127–138.
[13] E. Doelker, Assessment of powder compaction, in: D. Chulia, M. [37] C. Wiederkehr-von Vincenz, PhD thesis, E.T.H. Zürich, 1979.
Deleuil, Y. Pourcelot (Eds.), Powder Technology and Pharmaceutical [38] I. Krycer, D.G. Pope, J.A. Hersey, An evaluation of the techniques
Processes, Elsevier, Amsterdam, 1994, pp. 403– 471. employed to investigate powder compaction behaviour, Int. J. Pharm.
[14] K. Marshall, Some observations on the elucidation and control of die 12 (1982) 113 –134.
compaction processes, in: A.S. Goldberg (Ed.), Proceedings of the [39] D. Khossravi, W.T. Morehead, Consolidation mechanisms of
First International Conference on Compaction and Consolidation of pharmaceutical solids: a multi-compression cycle approach, Pharm.
Particulate Matters, The Powder Advisory Centre, London, 1972, pp. Res. 14 (1997) 1039–1045.
173–181. [40] I. Higuchi, T. Shigamoto, S.P. Eriksen, T. Yashiki, Physics of tablet
[15] D. Sixsmith, Instruments for tablet technology, Manuf. Chem. compression. XIV. Lateral die wall pressure during and after
Aerosol News 48 (1) (1977) 17–21. compression, J. Pharm. Sci. 54 (1965) 111–118.
[41] K. Ridgway, J. Glasby, P.H. Rosser, Effect of crystal hardness on [67] H.E.N. Hiestand, D.P. Smith, Indices of tableting performance,
radial pressure at the wall of a tableting die, J. Pharm. Pharmacol. 21 Powder Technol. 38 (1984) 145 –159.
(1969) 24S–29S. [68] H.E.N. Hiestand, Dispersion forces and plastic deformation in tablet
[42] J.E. Carless, S. Leigh, Compression characteristics of powder: radial bond, J. Pharm. Sci. 74 (1985) 768–770.
die-wall pressure transmission and density changes, J. Pharm. [69] H.E.N. Hiestand, Principles tenets and notions of tablet bonding and
Pharmacol. 26 (1974) 289– 329. measurements of strength, Eur. J. Pharm. Biopharm. 44 (1997)
[43] B.A. Obiorah, Possible prediction of compression characteristics from 229 –242.
pressure cycle plots, Int. J. Pharm. 1 (1978) 249 –255. [70] E. Shotton, B.A. Obiorah, Effect of physical properties on
[44] A.W. Hölzer, J. Sjögren, Evaluation of some lubricant by the compression characteristics, J. Pharm. Sci. 64 (1975)
comparison of friction coefficients and tablet properties, Acta Pharm. 1213–1215.
Suec. 18 (1981) 139–148. [71] C. Nyström, G. Alderborn, M. Duberg, P.-G. Karehill, Bonding
[45] E. Doelker, E. Shotton, The effect of some binding agents on the surface area and bonding mechanism—two important factors for the
mechanical properties of granules and their compression character- understanding of powder compactability, Drug Dev. Ind. Pharm 19
istics, J. Pharm. Pharmacol. 29 (1977) 193 –198.
(1993) 2143–2196.
[46] H. Unckel, Vorgänge beim Pressen von Metallpulvern, Arch.
[72] C. Nyström, P.-G. Karehill, The importance of intermolecular
Eisenhüttenwesen 18 (1945) 161 –167.
bonding forces and the concept of bonding surface area, in:
[47] D. Train, J.A. Hersey, Some fundamental studies in the cold
G. Alderborn, C. Nyström (Eds.), Pharmaceutical Powder
compaction of plastically deforming solids, Powder Metall. 6
Compaction Technology, Marcel Dekker, New York, 1996,
(1960) 20–35.
pp. 17–53.
[48] D. Train, J.N. Carrington, J.A. Hersey, The pressing of polymers in
cylindrical dies, Ind. Chem. 38 (1962) 77 –80. [73] B.A. Obiorah, E. Shotton, The effect of waxes, hydrolyzed gelatin and
[49] R.F. Lammens, T.B. Liem, J. Polderman, C.J. De Blaey, Evaluation of moisture on the compression characteristics of paracetamol and
force–displacement measurements during one-sided powder compac- phenacetin, J. Pharm. Pharmacol. 28 (1976) 629–632.
tion in a die—the influence of friction with the die on upper and lower [74] E. Shotton, B.A. Obiorah, The effect of particle shape and crystal
punch pressure, Powder Technol. 26 (1980) 169–185. habit on properties of sodium chloride, J. Pharm. Pharmacol. 25
[50] R.A. Thompson, Mechanics of powder pressing: I. Model for powder (1973) 37P–43P.
densification, Am. Ceram. Soc. Bull. 60 (1981) 237–243. [75] J.T. Carstensen, P. Touré, D. Duchêne, F. Puisieux, Correlation
[51] P. Mosbah, D. Bouvard, E. Ouedraogo, P. Stutz, Experimental between hysteresis loop areas of lower punch and of die pressure
techniques for analysis of die pressing and ejection of metal powder, versus upper punch pressures, Drug Dev. Ind. Pharm. 7 (1981)
Powder Metall. 40 (1997) 269 –277. 645 –648.
[52] W.M. Long, Radial pressure in powder compaction, Powder Metall. 6 [76] I. Krycer, The prediction of paracetamol capping tendencies, J. Pharm.
(1960) 73–86. Pharmacol. 34 (1982) 802 –804.
[53] W.M. Long, Die design and related questions in powder compaction, [77] M.P. Summers, R.P. Enever, J.E. Carless, The influence of crystal
Spec. Ceram. 17 (1962) 327– 340. form on the radial stress transmission characteristics of pharmaceu-
[54] J.T. Carstensen, P. Touré, Compression cycles in tableting, Powder tical materials, J. Pharm. Pharmacol. 28 (1976) 89–99.
Technol. 26 (1980) 199–204. [78] J. Barra, R. Somma, Influence of the physicochemical variability of
[55] E.L. Parrott, Compression, in: H.A. Lieberman, L. Lachman, J.B. magnesium stearate on its lubricant properties: possible solutions,
Schwartz (Eds.), Pharmaceutical Dosage Forms: Tablets, vol. 2, Drug Dev. Ind. Pharm. 22 (1996) 1105–1120.
Marcel Dekker, New York, 1990, pp. 201–243. [79] P.I. Seth, Münzel, Der Einfluss des Feuchtigkeitsgehalt eines
[56] S. Strijbos, P. Rankin, R.J. Klein Wassink, J. Bannink, G.J. Granulates auf die Pressbarkeit und die Eigenschaften der Tabletten,
Oudemans, Stresses occurring during one-sided die compaction of Pharm. Ind. 21 (1959) 9–12.
powders, Powder Technol. 18 (1977) 187– 200. [80] J.E. Rees, E. Shotton, Effect of moisture in compaction of particulate
[57] P. Touré, PhD thesis, Université Paris-Sud, 1979. material, J. Pharm. Sci. 60 (1971) 1704–1708.
[58] E.N. Hiestand, J.E. Wells, C.B. Peot, J.F. Ochs, Physical process of [81] E. Shotton, E. Ress, The compaction properties of sodium chloride in
tableting, J. Pharm. Sci. 66 (1977) 510–519. the presence of moisture, J. Pharm. Pharmacol. 18 (Suppl) (1966)
[59] H. Schrank-Junghäni, H.P. Bier, H. Sucker, The measurement of die 160S–167S.
wall forces to determine the minimum concentration of lubricant
[82] A.F. Rime, F. Lescure, D. Mordier, R. Gurny, E. Doelker, A new high
needed for tablet formulations, Acta Pharm. Technol. 30 (1984)
performance data acquisition system interfacing an instrumented
224–234.
tablet machine to a microcomputer, Acta Pharm. Technol. 36 (1999)
[60] S. Strijbos, Friction between a powder compact and a metal wall, Sci.
264 –268.
Ceram. 8 (1976) 415–428.
[83] P. Humbert-Droz, D. Mordier, E. Doelker, Méthode rapide de
[61] A.W. Hölzer, J. Sjorgen, Friction coefficient of tablet masses, Int.
détermination du comportement à la compression pour les études de
J. Pharm. 7 (1981) 269– 277.
[62] M.-Al. Shammat, D.N. Travers, T.C. Buttery, Die wall reaction and préformulation, Pharm. Acta Helv. 57 (1982) 136–143.
friction during compaction of some direct compression base, J. Pharm. [84] G. Simmons, H. Wang, Single Crystal Constants and Calculated
Pharmacol. 31 (1979) 76P. Aggregate Properties: A Handbook, 2nd ed., MTI Press, Cambridge,
[63] V.T. Laich, T. Kissel, Axiales Matrizen Wandkraftminimum, Pharm. 1971.
Ind. 57 (1995) 174 –182. [85] R.J. Roberts, R.C. Rowe, P. York, The Poison’s ratio of microcrystal-
[64] S. Strijbos, Powder-wall friction: the effects of orientation of wall line cellulose, Int. J. Pharm. 105 (1994) 177– 180.
lubricants, Powder Technol. 18 (1977) 209– 214. [86] R.J. Roberts, R.C. Rowe, P. York, The relationship between Young’s
[65] J. Kikuta, N. Kitamori, Evaluation of the die wall friction tablet modulus of elasticity of organic solids and their molecular structure,
ejection, Powder Technol. 35 (1983) 195–200. Powder Technol. 65 (1991) 139 –146.
[66] J.E. Rees, E. Shotton, The effect of dimensions on the compaction [87] S.W. Hoag, Die wall instrumentation: design and application in
properties of sodium chloride, J. Pharm. Pharmacol. 21 (1969) compaction, AAPS Annual Meeting and Exposition, Salt Lake City,
731–743. 2003.

Benefits of Die-Wall Instrumentation For Research and Development in Tabletting

Uploaded by

Copyright:

Available Formats

Benefits of Die-Wall Instrumentation For Research and Development in Tabletting

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Benefits of Die-Wall Instrumentation For Research and Development in Tabletting

Uploaded by

Copyright:

Available Formats

European Journal of Pharmaceutics and Biopharmaceutics 58 (2004) 427–444

Benefits of die-wall instrumentation for research

1. Introduction (using strain gauges or piezoelectric sensors), the punch

single-punch machines, rotary machines and compaction

Authors Year Press Measuring system

on the position of the lower punch, the rubber height and

3. Axial to radial stress transmission: theoretical

705 mg 420 –840 Hydraulic 9.5 30 s 57 0.59 Higuchi et al. [40]

n.m, not mentioned.

respectively. Fd is sometimes called the frictional force 3.1. Axial transmission

3 14.5 Single punch 12.0 n.m. 112 0.46 Doelker and

n.m., not mentioned.

models is that monitoring of the radial pressure is needed

3.2. Radial transmission

Long [52,53] was the first author to elucidate the

Table 6 The last example deals with the effect of internal

both the special testing instrument (static friction) and an

Lubrication Compression h Fe (N) Fr0 (N) m2 Disintegration Crushing force (N)

– 9 0.25 1000 862 1.16 1.1 135

Material Pa (MPa) Segment slopes h Pr0 (MPa) Ref.

Paracetamol 175 0.17 0.45 0.35 3.00 n.a. n.a. [73]

n.a., not available.

Material type WU WEXP PL1 PL2 Py Segment slopes h Pr0 A1 A2 1c 1 ER0 ER Fc

High tablet strength

Low tablet strength

are able to characterise the mechanical properties of the

Interest in die-wall instrumentation is continuing as

Parameter High tablet strength Low tablet strength

Work of compression, Wu Medium High Low Medium Low

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.