Journal of Crystal Growth 243 (2002) 345–355

Influence of crystal habit on the compression and densification

mechanism of ibuprofen
Piera Di Martinoa,*, Moira Beccericaa, Etienne Joirisb, Giovanni F. Palmieria,
Anne Gayotb, Sante Martellia
a
Faculty of Pharmacy, Department of Chemical Sciences, University of Camerino, Via S. Agostino, Camerino 62032, Italy
b
Faculty of Pharmacy, Laboratoire de Pharmacotechnie Industrielle, University of Lille II, Rue du Professeur Laguesse, Lille, France

Received 7 May 2002; accepted 30 May 2002

Communicated by R. Kern

Abstract

Ibuprofen was recrystallized from several solvents by two different methods: addition of a non-solvent to a drug
solution and cooling of a drug solution. Four samples, characterized by different crystal habit, were selected: sample A,
sample E and sample T, recrystallized respectively from acetone, ethanol and THF by addition of water as non-solvent
and sample M recrystallized from methanol by temperature decrease. By SEM analysis, sample were characterized with
the respect of their crystal habit, mean particle diameter and elongation ratio. Sample A appears stick-shaped, sample E
acicular with lamellar characteristics, samples T and M polyhedral. DSC and X-ray diffraction studies permit to
exclude a polymorphic modification of ibuprofen during crystallization. For all samples micromeritics properties,
densification behaviour and compression ability was analysed. Sample M shows a higher densification tendency,
evidenciated by its higher apparent and tapped particle density. The ability to densificate is also pointed out by D0 ’
value of Heckel’s plot, which indicate the rearrangement of original particles at the initial stage of compression. This
fact is related to the crystal habit of sample M, which is characterized by strongly smoothed coins. The increase in
powder bed porosity permits a particle–particle interaction of greater extent during the subsequent stage of
compression, which allows higher tabletability and compressibility.
r 2002 Elsevier Science B.V. All rights reserved.

PACS: 61.43.G

Keywords: A1. Characterization; A1. Crystal morphology

1. Introduction

The effect of various powder characteristics on

their compression and densification behaviour has
*Corresponding author. Tel.: +39-737-40-22-89; fax: +39-
been extensively considered in several studies: the
737-637345.
E-mail addresses: dimartin@camserv.unicam.it, crystal form [1], the moisture content [2], the
piera.dimartino@unicam.it (P. Di Martino). original particle size for plastic deforming materials

0022-0248/02/$ - see front matter r 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 0 2 2 - 0 2 4 8 ( 0 2 ) 0 1 5 2 3 - 3
346 P. Di Martino et al. / Journal of Crystal Growth 243 (2002) 345–355

[3], particle fragmentation [4]. The effect of crystal and densification behaviour, and ibuprofen was
habit, which is the depiction of outer appearance chosen as powder model.
of crystals [5], on mechanical properties of
powders has been already considered, but final
conclusions are difficult to be carried out. In fact 2. Experimental section
the complexity of produced models limits the
possibility to ascribe conclusions to a precise 2.1. Ibuprofen crystals preparation
effect. Some difficulty exists in considering a same
substance in which the only change depends on Ibuprofen USP was kindly supplied by
crystal habit. Besides, it is often difficult to obtain A.C.R.A.F. (Ancona, Italy). Its solubility was
samples of a same material with same particle size, determined by equilibrating the liquid phase with
but different crystal habit. an excess of ibuprofen at 25.070.51C under
The effect of particle shape on tablet strength stirring in several solvents, all of analytical grade.
has been investigated by Alderborn and co-work- Approximately 24-h period was necessary to reach
ers [6,7]. By milling coarse fractions, they achieved the equilibrium. The saturated solution was
variations in shape of materials and they con- filtered with a regenerated cellulose filter syringe
cluded that tablet strength of plastic deforming of pore size of 0.45 mm (Filalbet, Barcelona,
materials is markedly affected by a change in Spain). After an appropriate dilution, the concen-
particle shape, while for fragmenting materials this tration of the filtrate was determined spectro-
properties is independent of particle shape. The photometrically (Cary 1E UV-VIS, Varian, Lein"ı,
limit of this study is that samples were not Italy) at 264 nm.
accurately classified. Moreover, it should be noted Two different procedures were used for ibupro-
that the milling step can affect the crystallinity of fen crystallization:
powders [8].
The effect of particle shape on the mechanical (a) the addition of a non-solvent to a drug
properties of powders has been studied by Wong solution maintained at room temperature
and Pilpel [9]. However, they did not refer to a under continuous stirring;
characteristic shape, but only classify powders in (b) the quench to room temperature of a drug
regular and irregular. They concluded that for solution maintained at 501C under continuous
plastic deforming material there is a large increase stirring.
in compressibility in going from regular to
irregular particles. On the contrary, for fragment- Rapid crystallization kinetic was chosen in order
ing materials there is no influence of particle shape to obtain small and homogeneous crystals. Crys-
on the above property. tals were then filtered and desiccated in presence of
Garekani et al. [10] considered the influence of P2O5. The 53–150 mm sieved granulometric frac-
crystal habit on the compaction behaviour of tion was recovered.
paracetamol. Paracetamol monoclinic form
traditionally shows an elastic behaviour. Polyhe- 2.2. Physical characterization of ibuprofen crystals
dral crystals underwent a greater plasticity during
compression than thin plate-like crystals which Crystals were firstly observed under an optical
were more brittle in nature during compression. microscope (Leitz, Ortholux, Germany). After
Plate-like crystals underwent more elastic defor- powder selection, they were then observed using
mation during compaction than polyhedral crys- a scanning electron microscope (Stereoscan 360,
tals. Besides, Hong-Guang and Ru-Hua [11] Cambridge Instruments, Cambridge, UK). Sam-
evidenciate that paracetamol acicular crystals ples were mounted on a metal stub with a double-
showed greater tendency to the capping. sided adhesive tape and then recovered under
The aim of this study was to investigate the vacuum with a gold layer of 200 A( thickness using
influence of particle morphology on compression a metallizator (Balzer MED 010, Linchestein).
 P. Di Martino et al. / Journal of Crystal Growth 243 (2002) 345–355 347

Particle size was determined by counting the Changes occurring in packing arrangement
Ferret’s diameter of 500 particles using the during the tapping procedure are expressed as
scanning electron microscopy. Elongation ratio the compressibility index which is a measure of the
was also determined by the ratio of the maximum propensity of a powder to consolidate [12],
diameter and the minimum diameter. Compressibility Index ¼ ½V0  Vconst =V0   100:
Samples were analysed using a DSC (Pyris 1,
Perkin Elmer, Co. Norwalk, USA) equipped with ð1Þ
an ethanol cooling system circulating in a refrig-
erator (Cryostat F4-Q, Haake Q, Karlsruhe, 2.3. Compression behaviour of crystals
Germany). A dry purge of nitrogen gas (20 ml/
min) was used for all runs. DSC was calibrated for For the densification study, powders were
temperature and heat flow using a pure sample of compressed with an instrumented Frogerais OA
indium and zinc standards. Sample mass was single-punch tablet machine (Frogerais, Vitry,
about 3–4 mg and aluminium closed pans were France) equipped with 11.3 mm flat-faced
used. Each run was performed in triplicate from punches, by introducing 325 mg samples manually
201C to 1001C at a heating rate of 101C/min. Onset into the prelubricated die. The lubrication was
temperature and enthalpy content of melting peak obtained by introducing in the hopper a mixture of
was calculated. Results are the mean of three Avicel PH 102 (FMC Europe NV, Brussels,
analysis. Belgium) and 1% w/w Mg stearate (FU-USP
The powder X-ray diffraction study was carried grade, ACEF, Fiorenzuola d’Arda, Italy) and by
out to exclude a possible polymorphic modifica- compressing this mixture before each sample,
tion. A Philips PW 1730 (Almelo, Netherlands) according to Lefebvre et al. [13]. Five replicates
was used as X-ray generator for Cu Ka radiation cycles were performed for all substances, corre-
( The experimental X-ray powder
(l ¼ 1:54178 A). sponding to maximal punch pressure of about
patterns were recorded on a Philips PH 8203. The 165 MPa. For a single compression cycle, both the
goniometer supply was a Philips PW 1373 and the compression pressures on the upper and lower
channel control a Philips PW 1390. The data were punches and the displacement of the upper punch
collected in the continuous scan mode using a step were measured and recorded at a frequency of
size of 0.011 2y: The scanned range was 2–501 (2y). 4000 Hz. Correction of the displacement transdu-
Weight loss of all samples was determined until cer data for machine looseness and punch defor-
equilibrium was reached (1 week) maintaining mation was carried out according to Juslin and
them at 201C under vacuum, in presence of P2O5 Paronen [14].
as desiccant (ACEF, Fiorenzuola d’Arda, Italy). The densification behaviour of powders was
Analysis was performed in triplicate. studied using Heckel’s equation [15]:
The determination of residual water on ibupro-
ln 1=1  D ¼ KP þ A; ð2Þ
fen crystals was made according to the titrimetric
method of Karl Fisher (TR85, Schott Ger.ate) after where D is the relative density of the compressed
calibration with sodium tartrate dihydrate and powder bed at applied pressure P: K is the slope of
dissolution of crystals in methanol. The Fisher the straight linear portion of the Heckel plot, and
reagent solutions were from Carlo Erba (Milan, the reciprocal of K is the mean yield pressure (PY ).
Italy). The constant A is the sum of two densification
Apparent particle densities were measured by terms:
using a helium pycnometer (Ultrapycnometer
A ¼ lnð1=1  D0 Þ þ B; ð3Þ
1000, Quantachrome, USA) with a cell of 60 cm3.
Results are the mean of 10 measurements. The where lnð1=1D0 Þ is related to the initial die filling,
apparent densities at the initial stage and at and B is the densification due to the slippage and
constant volume (1250 tapping) were determined rearrangement of both primary and fragmented
by measuring the volumes of 100 g of powder. particles. Constants A and B can be expressed as
348 P. Di Martino et al. / Journal of Crystal Growth 243 (2002) 345–355

relative densities using: grade, PRS, Panreac, Montcada i Reixac, Spain).

DA ¼ 1  e A
; ð4Þ Powder mass was 250 mg. The compression forces
were progressively increased from the lower force
DB ¼ DA  D0 ; ð5Þ able to give tablets of sufficient strength to the
maximal force consistent with the machine char-
To make a clearer distinction between densifica- acteristics. The force at the upper punch was then
tion due to the movements of the original particles recorded. Results for each compression force are
and that due to the brittle fracture, Doelker [16] the mean of five measurements.
proposed modifying the definition of D0 (and Thickness and diameter of intact ejected tablets
subsequently that of DB ) by using a relative were measured with a manual micrometer with an
precompression density D00 term, which also accuracy of 0.001 mm (Mitutoyo, Kanagawa,
includes the initial rearrangement of particles. In Japan) immediately after ejection. Tablet porosity
this conditions, Eqs. (3) and (5) become respec- was calculated from tablet dimensions, mass, and
tively: apparent particle density. Crushing force was
A ¼ lnð1=1  D00 Þ þ B0 ; ð6Þ measured immediately after compression with a
tablet hardness tester (Erweka, TBH 30, Heusen-
D0B ¼ DA  D00 ; ð7Þ stamm, Germany). Tensile strength Q was calcu-
where D0B is this time only representative of the lated according to Eq. (9) [18]:
densification due to fragmentation. D00 corre- Q ¼ 2H=pdt; ð9Þ
sponds to the relative density of the powder at
the moment when the last recorded applied where H is the tablet crushing strength, d the
pressure is still nil. In this study, Heckel’s profiles diameter and t the thickness of the tablet. Powder
were generated by the ‘‘in die method’’ [15] from bed porosity was calculated from die dimensions,
single compression cycles. All samples were upper punch displacement, powder mass, and
compressed approximately at 165 MPa. Para- apparent particle density.
meters PY ; DA ; D00 ; D0B were calculated using a
precompression pressure value of 1.5 MPa, and
data ranging from 50 to 100 MPa for linear
regression analysis. Each value is a mean of five 3. Results and discussion
measurements.
The elastic recovery (ER) was calculated ac- Several studies already pointed out the depen-
cording to Armstrong and Haines-Nutt [17]: dence of crystallization conditions on crystal form
and crystal habit of ibuprofen [19–21].
ERð%Þ ¼ ½ðt2  t1 Þ=t1   100; ð8Þ
Two crystallization methods were used with
where t1 is the minimal thickness of the powder in different solvents in order to obtain ibuprofen
the die when the upper punch displacement was crystals of a characteristic habit. Crystal selection
maximal and t2 is the tablet thickness. was based on the choice of homogeneous crystal
The compression study of all the ibuprofen habit and particle size for each powder. The
crystals was performed using a rotary press (High- optical microscopy analysis was used for a
Tech-Mini, Ronchi, Cinisello Balsamo, Italy) preliminary study. After this, four samples were
equipped with a computerized control system for selected: crystals obtained from acetone by water
detection and analysis of force-signals (pressing addition (sample A); crystals obtained from
force and ejection force). This press is equipped ethanol by water addition (sample E); crystals
with 10 flat punches of 11.28 mm of diameter. obtained from THF by water addition (sample T);
Because of the small quantities of samples, they crystals obtained from methanol by temperature
were manually introduced in only one die. Die and decrease (sample M).
punches were prelubricated with 1% w/v Mg Fig. 1 shows the crystal habit of the selected
stearate slurry in 96% v/v ethyl alcohol (analytical sieved ibuprofen crystals at a magnification of
 P. Di Martino et al. / Journal of Crystal Growth 243 (2002) 345–355 349

Fig. 1. Scanning electron microscopy of ibuprofen samples at a magnification of 100  .

100  and Fig. 2 shows the same crystals at higher sample A, which also shows a greater elongation
magnifications (700  ). Crystals of sample A ratio. On the contrary, sample M shows a lower
appear stick-shaped and particle surface is quite elongation ratio.
regular as shown at higher magnification. Sample Weight loss at desiccation, water content and
E crystals are characterized by an acicular habit DSC results are reported in Table 2. Weight loss
and it is possible to distinguish a lamellar aspect at percentage is very similar for samples A and E,
higher magnifications. Sample T crystals appear and just lower for samples M and T. The very low
irregular with a quasi-cubic shape and higher water content indicates that weight loss is preva-
magnification shows the presence of smooth coins. lently due to the presence of organic solvents.
Sample M crystals are polyhedral with strongly Crystals were then allowed to dry at room
smoothed coins. Faces are quite irregular. temperature until no mass loss was measured by
Dimensions of sieved crystals were presented in further desiccation test.
Table 1. The mean particle size is very similar for The melting onset temperature and enthalpy are
samples E, M and T, while it is far higher for approximately the same for all the samples. This
350 P. Di Martino et al. / Journal of Crystal Growth 243 (2002) 345–355

Fig. 2. Scanning electron microscopy of ibuprofen samples at a magnification of 700  .

fact permits to predict that a same crystal form is higher one is referred for sample M. It is
involved, with a similar crystallinity degree. undoubtedly related to the crystal habit and
In order to confirm that no polymorphic surface, which allow particle arrangement during
modification occurs during crystallization process, filling and packing. The compressibility index
an X-ray diffraction study was carried out. diminish in the order of samples A>E>M>T.
Interreticular distances are the same, confirming So, unexpected, the propensity to consolidate
that a same crystal form is involved (Fig. 3). This seems higher for sample T and it is probably due
analysis also permits to evaluate that the crystal- to the fact that there is a minor change in volume
linity degree is approximately the same for all the during the packing.
samples, as it is evidenciated by the similar peak The densification study was carried out on a
intensities. single punch tablet machine. Typical Heckel’s
Table 3 reports some physical properties of the profiles are presented in Fig. 4. It can be observed
powders. Particle densities, mainly affected by that this profile is in good agreement with a
molecule arrangement in the crystal, are very plastic–elastic compression behaviour, as de-
similar. Apparent and tapped powder bed densi- scribed by Marshal et al. [22], but that a noticeable
ties, mainly affected by outer appearance of fragmentation occurs at low compression force, as
crystals, are lower for samples A and E. The indicated by the initial curvature of Heckel’s plots.
 P. Di Martino et al. / Journal of Crystal Growth 243 (2002) 345–355 351

Table 1
Morphological characteristics and dimensions of ibuprofen samples of different crystal habit

Crystal habit Mean particle diameter Elongation ratio (mm)b

(normal distribution) (mm)a

Sample A Stick-shaped 71.39744.61 2.5571.66

Sample E Acicular-laminar 44.66716.85 2.2671.18
Sample M Polyedric 32.12710.98 1.0370.44
Sample T Polyedric 42.60721.07 1.9771.35
a
Determined by counting Ferret’s diameter of 500 particles on SEM micrograph. Standard deviations are indicated.
b
Determined by the ratio of the maximum Ferret’s diameter and the minimum Ferret’s diameter of 500 particles. Standard
deviations are indicated.

Table 2
Results from DSC analysis and mass loss

Mass loss (%)a Water content (%)b Onset melting temperature (T1C)c DH (J/G)c

Sample A 6.5170.22 0.03 75.4370.17 123.870.31

Sample E 6.2770.11 0.04 75.1170.17 121.871.24
Sample M 3.6970.08 0.07 75.0270.06 124.071.31
Sample T 4.0470.30 0.02 74.7770.41 121.970.26
a
Mass loss was determined after 1 week at 201C under vacuum, and by drying in presence of P2O5 as desiccator until equilibrium was
reached. Analysis was performed in triplicate. Standard deviations are indicated.
b
Water content was measured by Karl Fisher’s titrimetric method. Assays were performed in triplicate.
c
Assays were performed in triplicate. Standard deviations are indicated.

Results for Heckel’s parameters and elastic recov- curvature is reported for the sample T. This is
ery of the four parameters are reported in Table 4. probably due to a light defect in the correction of
D00 is far higher for sample M, indicating a better displacement data which is more evident at so low
particle rearrangement during die filling, related to porosity values.
the typical particle shape. D0B ; which is affected by From the densification study, it is possible to
the fragmentation propensity of a material, is very conclude that crystal habit only affect the initial
similar for samples A, E and T, while it is particle rearrangement. On the contrary, the
markedly lower for sample M. It is difficult to intrinsic compressibility of powders is not influ-
ascribe this effect to the crystal habit or to intrinsic enced by crystal habit.
resistance of the material. Probably, because Tabletability, which is the capacity of a pow-
sample M reduces in volume preferably by particle dered material to be transformed into a tablet of a
slippage, densification by brittle fracture does not specified strength under the effect of compression
occur. DA and PY values are similar for all pressure [1], and compressibility, which is the
samples, indicating that small differences observed ability of a material to be reduced in volume as a
in the initial stage of compression are leveled out result of an applied pressure [1], are presented in
once a higher pressure range is applied. Elastic Figs. 5 and 6. Samples A and E show very low
recovery is slightly higher for samples A and E increment in tablet tensile strength when applied
than for the two polyedric samples M and T, pressure is increased, swiftly reaching a plateau
probably resulting from a lower interparticulate value. In contrast, sample T and more particularly
bonding. sample M show greater increase in tablet tensile
Apart of values already considered, it must be strength, which is accompanied by greater decrease
pointed out that at the maximum compression in tablet porosity. This explains why at low
pressure all cycles are curved, and the maximum pressures (under 100 MPa), the four samples
352 P. Di Martino et al. / Journal of Crystal Growth 243 (2002) 345–355

quite constant whatever the compression force,

while sample M tablets exhibit greater tensile
strength and densification.
Undoubtedly crystal habit influences the com-
pression behaviour of a powder. Powder compres-
sion behaviour is a result of different components:
particle rearrangement and solid deformability
which influence solid consolidation. Particle re-
arrangement at the initial stage of compression,
which is a time-dependent phenomenon, is fa-
voured by a polyhedral crystal habit. A high
densification in this stage allows a particle–particle
interaction of a greater extent, which can explains
better tabletability and compressibility [23].
Beside, the elastic recovery tendency is not only
affected by the intrinsic deformability of a solid,
but it is a balance between this characteristic and
the ability of a material to form stronger bonds.
Lower porosities of powder bed favours particle–
particle interactions with the constitution of more
strong bonds.

4. Conclusion

Ibuprofen samples were selected with respect to

their different crystal habits: stick-shaped, acicu-
lar-laminar, polyhedral. All samples are very
similar in mean particle diameter, in water content,
in crystallinity degree, in crystal form, so differ-
ences can only be attributed to their different
crystal habit. Differences were pointed out with
the respect of densities and packing properties.
Sample M, obtained by crystallization from
methanol by temperature decrease, shows the
higher apparent particle density.
Heckel’s analysis confirms that ibuprofen is a
plastic–elastic material, but, during this study, a
certain fragmentation propensity has also been
pointed out, the extent of which could be affected
by the crystal habit. However, it has also been
Fig. 3. X-ray diffraction patterns of ibuprofen samples of pointed out that the densification by brittle
different crystal habit. fracture is responsible only to a limited degree to
the tabletability and compressibility of ibuprofen.
remain quite undistinguishable, with a progressive In fact, samples that show greater fragmentation
increase in tensile strength and a corresponding propensity show very low increase in tablet tensile
decrease in porosity. Above 100 MPa, for samples strength by increase in compression pressure. A
A, E and M, tensile strength and porosity remain more important role in tabletability is played by
 P. Di Martino et al. / Journal of Crystal Growth 243 (2002) 345–355 353

Table 3
Physical properties of ibuprofen particles after sieving

Sample A Sample E Sample M Sample T

Particle density (g cm3)a 1.1172 (0.0009) 1.1167 (0.0006) 1.1189 (0.0005) 1.1173 (0.0008)
Apparent density (g cm3)b 0.2163 0.2158 0.4251 0.3561
Tapped density (g cm3)c 0.3194 0.2908 0.5668 0.4452
Compressibility indexd 32.26 25.78 25.00 20.00
a
Determined by using a helium pycnometer. Results are the mean of ten measurements. Standard deviations are indicated.
b
Determined by measuring the volume of 100 g of powder.
c
Determined by measuring the volume of 100 g of powder after 1250 tapping.
d
Determined by the following equation: ½V0 Vconst =V0   100; where V0 is the volume of the powder bed at the initial stage and
Vconst the volume after tapping at constant volume.

6.0 Sample A 6.0 Sample E

5.0 5.0

4.0 4.0
-ln porosity

-ln porosity

3.0 3.0

2.0 2.0

1.0 1.0

0.0 0.0
0 50 100 150 200 0 50 100 150 200
Compression pressure (MPa) Compression pressure (MPa)

6.0 6.0
Sample M Sample T
5.0 5.0

4.0 4.0
-ln porosity

-ln porosity

3.0 3.0

2.0 2.0

1.0 1.0

0.0 0.0
0 50 100 150 200 0 50 100 150 200
Compression pressure (MPa) Compression pressure (MPa)

Fig. 4. Heckel’s plots of ibuprofen samples. Plots are obtained from a single compression cycle. Straight lines correspond to linear
regression analysis of data ranging from 50 to 100 MPa.

densification which takes place during the initial sibility, thanks to its polyhedral aspect, with
stage of compression and which depends on the strongly smoothed faces, which concurs to this
particle slippage and rearrangement. In fact, arrangement. This fact permits to reach a higher
sample M shows better tabletability and compres- powder densification and a lower powder bed
354 P. Di Martino et al. / Journal of Crystal Growth 243 (2002) 345–355

Table 4
Heckel’s parameters and elastic recovery for ibuprofen samples. 95% confidence intervals are also indicated

Sample A Sample E Sample M Sample T

D00 0.55470.022 0.55770.008 0.64870.011 0.58470.03

D0B 0.29270.028 0.27370.014 0.18870.007 0.24970.026
DA 0.84670.008 0.83070.007 0.83670.008 0.83370.010
PY 58.974.8 58.973.2 54.372.3 54.874.2
ER (%) 4.7170.31 5.4970.52 4.2070.31 4.2770.3

porosity. The greater extent of particle–particle

1.0
interaction on the subsequent stage of compres-
sion could explain the lower elastic recovery and
Tablet tensile strength (MPa)

the greater tabletability and compressibility ob-

served for this sample during the compression
study. Influence of crystal habit on the mechanical
0.5 properties of ibuprofen remains relatively low. It
can come from the low melting point of this
Sample E
Sample M substance, which can induce partial particle fusion
Sample A during compression, as previously reported [24].
Sample T
0.0
0 50 100 150 200 250 300
Upper punch compression pressure (MPa) Acknowledgements
Fig. 5. Tensile strength-pressure profiles of ibuprofen samples.
The authors gratefully acknowledge the finan-
Tablet tensile strength is quoted as a function of maximal upper
punch pressure. Each point is the mean of five measurements. cial support of the Italian MURST—‘‘Fondi 40%
Standard deviations are indicated. Progetto Nazionale Tecnologie Farmaceutiche.’’
In addition, they would like to thank Dr. G.
Cantalupo for his kind help in the SEM studies.

16
Sample E
14 Sample M References
Sample A
12 Sample T
Tablet porosity (%)

[1] E. Joiris, P. Di Martino, C. Berneron, A.M. Guyot-

10
Hermann, J.C. Guyot, Pharm. Res. 15 (1998) 1122.
8 [2] A. Nokhodchi, J.L. Ford, P.H. Rowe, M.H. Rubinstein,
J. Pharm. Pharmacol. 48 (1996) 1122.
6 [3] M. Eriksson, G. Alderborn, Int. J. Pharm. 113 (1995) 199.
4 [4] M. Eriksson, G. Alderborn, Pharm. Res. 12 (1995) 1031.
[5] P. York, Int. J. Pharm. 14 (1983) 1.
2 [6] G. Alderborn, C. Nystrom, Acta Pharm. Suec. 19 (1982)
0 147.
0 50 100 150 200 250 300 [7] G. Alderborn, E. Borjesson, M. Glazer, C. Nystrom, Acta
Upper punch compression pressure (MPa) Pharm. Suec. 25 (1988) 31.
[8] R. Huttenrauch, S. Friecke, P. Zielke, Pharm. Res. 6
Fig. 6. Compressibility of ibuprofen samples. Tablet porosity is (1985) 302.
quoted as a function of maximal upper punch pressure. Each [9] L.W. Wong, N. Pilpel, Int. J. Pharm. 59 (1990) 145.
point is the mean of five measurements. Standard deviations are [10] H.A. Garekani, J.L. Ford, M.H. Rubinstein, A.R. Rajabi-
indicated. Siahboomi, Int. J. Pharm. 187 (1999) 77.
 P. Di Martino et al. / Journal of Crystal Growth 243 (2002) 345–355 355

[11] W. Hong-Guanh, Z. Ru-Hua, Drug Dev. Ind. Pharm. 21 [17] N.A. Armstrong, R.F. Haines-Nutt, Powder Technol. 9
(1995) 863. (1974) 287.
[12] R.L. Carr, Chem. Eng. 72 (1965) 163. [18] J.T. Fell, J.M. Newton, J. Pharm. Sci. 5 (1970) 688.
[13] C. Lefebvre, F. Guillaume, R. Bouch!e, R. Buoaziz, J.C. [19] N. Udupa, Drug Dev. Ind. Pharm. 16 (1990) 1591.
Guyot, Etude de diff!erentes formes polymorphes du [20] P. York, Drug Dev. Ind. Pharm. 18 (1992) 677.
meprobamate, Proceeding of the Fifth International [21] V. Labhasetwart, S.V. Deshmukh, A.K. Dorle, Drug Dev.
Conference On Pharmaceutical Technology, Paris, 1989, Ind. Pharm. 19 (1993) 631.
p. 211. [22] P.V. Marshall, P. York, J.Q. Maclaine, Powder Technol.
[14] M.J. Juslin, T.P. Paronen, J. Pharm. Pharmacol. 32 (1980) 74 (1993) 171.
796. [23] P. Di Martino, M. Scoppa, E. Joiris, G.F. Palmieri, C.
[15] R.W. Heckel, Trans. Metall. Soc. AIME 221 (1961) 661. Andr"es, Y. Pourcelot, S. Martelli, Int. J. Pharm. 213 (2001)
[16] E. Doelker, Assessment of powder compaction, in: D. 209.
Chulia, M. Deleuil, Y. Pourcelot (Eds.), Powder Technol- [24] A. Nokhodchi, M.H. Rubinstein, H. Larhib, J.C. Guyot,
ogy and Pharmaceutical Process, Elsevier, Amsterdam, Int. J. Pharm. 120 (1995) 13.
1994, p. 403.