Accepted Manuscript

Powder compression properties of paracetamol, paracetamol hydrochloride,

paracetamol cocrystals and coformers

Ann-Sofie Persson, Hamzah Ahmed, Sitaram Velaga, Göran Alderborn

PII: S0022-3549(18)30185-0
DOI: 10.1016/j.xphs.2018.03.020
Reference: XPHS 1115

To appear in: Journal of Pharmaceutical Sciences

Received Date: 29 December 2017

Revised Date: 26 February 2018
Accepted Date: 22 March 2018

Please cite this article as: Persson AS, Ahmed H, Velaga S, Alderborn G, Powder compression
properties of paracetamol, paracetamol hydrochloride, paracetamol cocrystals and coformers, Journal of
Pharmaceutical Sciences (2018), doi: 10.1016/j.xphs.2018.03.020.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT
Powder compression properties of paracetamol, paracetamol hydrochloride, paracetamol

cocrystals and coformers

Ann-Sofie Persson1*, Hamzah Ahmed2, Sitaram Velaga2 and Göran Alderborn1

PT
1

RI
Department of Pharmacy, Uppsala University, SE-751 23 Uppsala, Sweden
2
Pharmaceutical research, Department of Health Sciences, Luleå University of Technology,

SC
SE-971 87 Luleå, Sweden

Contact details corresponding author*:

U
AN
Tel: +46 18 471 4978
M

Fax: +46 18 471 4223

E-mail: ann-sofie.persson@farmaci.uu.se
D
TE

Keywords: Mechanical properties, Compression, Compaction, Co-crystals, Crystal structure

C EP
AC
 ACCEPTED MANUSCRIPT
Abstract

The objective was to study the relationship between crystal structure, particle deformation

properties and tablet-forming ability for the monoclinic form of paracetamol (PRA), two

cocrystals and a salt crystal of PRA in addition to two coformers (oxalic acid and 4,4’-

bipyridine). Thus, the structure – property – performance relationship was investigated.

PT
Analytical powder compression was used for determination of effective plasticity, as inferred

RI
from the Heckel yield pressure and the Frenning parameter, and the elastic deformation was

determined from in-die tablet elastic recovery.

SC
The plasticity could not be linked to the crystal lattice structure as crystals containing zig-zag

U
layers displayed similar plasticity as crystals containing slip planes. In addition, crystals
AN
containing slip-planes displayed both high and low plasticity.
M

The mechanical properties could neither be linked to the tablet-forming ability as the tablet
D

tensile strength, unexpectedly, displayed a tendency to reduce with increased plasticity

TE

stiffness. Furthermore, the elastic deformation could not explain the tablet forming ability.
EP

It was concluded that no relationship between structure – property – performance for

paracetamol and its cocrystals and salt could be established. Thus, it was indicated that to
C
AC

establish such a relationship an improved knowledge of crystallographic structure and inter-

particle bonding during compaction is needed.

2
 ACCEPTED MANUSCRIPT
1. Introduction

The mechanical properties of particles used in solid dosage forms strongly affect the

possibility to formulate a tablet with expected product profile, e.g. porosity, mechanical

strength and disintegration time. It is thus important to determine key mechanical properties

PT
of particles during tablet formulation development, such as their deformation properties, i.e.

plasticity and elasticity, and their fracture mechanics.1 Among these, the plasticity of particles

RI
is often considered fundamental for the development of the area of inter-particle bonds which

SC
affects the tablet-forming ability. In addition, the generation of fragments and new surfaces

due to fragmentation of brittle materials2 may also set up conditions for increased inter-

U
particulate contact area and a higher tablet tensile strength, i.e. an improved bonding ability.
AN
Furthermore, the tablet tensile strength is affected by the bonding mechanism and thus the

inter-particle bonding strength.2 Due to the importance of particle plastic deformation, the
M

crystallographic properties of particles controlling particle plasticity has been addressed in the
D

literature as well as the question of how the crystal structure can be modified in order to
TE

improve plasticity and thus tablet-forming ability. A common reasoning seems to be that the

presence of parallel planes (commonly referred to as slip planes) in the crystal structure
EP

promotes plane movement during loading due to a relatively low energy barrier along such

planes and thus facilitates plastic flow.3 One approach proposed to facilitate formation of
C
AC

planar structures in the crystal lattice, and thus promote particle plasticity, is the use of

cocrystals.

Cocrystals are non-covalent crystal self-assemblies of two compounds, in pharmaceutical

application typically an active pharmaceutical ingredient (API) and an excipient referred to as

a coformer.4 The tensile strength of tablets formed of cocrystals has been reported to both

3
 ACCEPTED MANUSCRIPT
increase3,5 and decrease6 compared to the tablet tensile strength of the API. This change in

tablet-forming ability may be attributed to both altered plasticity and elasticity of the

cocrystals.

The analgesic drug paracetamol form I (PRA) is known to show poor tablet-forming ability,

PT
explained by the zig-zag shape of the crystal lattice structure.7,8 This crystal structure

generates considerable elastic deformation,9 giving rise to problems of low mechanical

RI
strength and capping. However, it is reported that the metastable polymorphic form of PRA

SC
deforms plastically7,8 and exhibits a reduced elastic recovery7 due to the presence of planar

structures in the crystal lattice. More on, it is also earlier reported by Ahmed et al.,10 that

U
cocrystals and a salt of PRA showed improved tablet-forming ability compared to the PRA
AN
form I particles. This was explained in crystallographic terms by different slip planes with
M

weaker inter-planar bonds in the cocrystals and the salt. In order to better understand the

causative factors behind the improved tablet-forming ability of the cocrystals and the salt
D

compared to the PRA, the consecutive relationship between crystal structure, particle
TE

deformation and tablet-forming ability, i.e. structure – property – performance relationship,

needs to be investigated. The objectives of this study were thus to analyze the compression
EP

properties of powders of PRA and two cocrystals and a salt of PRA as well as the coformers

and to address the question of the structure – property – performance relationships of the
C
AC

crystals. The studied cocrystals were formed from a 1:1 stoichiometric ratio of PRA and the

coformers α-oxalic acid anhydrate (OXA) and 4, 4'-bipyridine (BPY). Analytical powder

compression11 was used to analyze the mechanical properties of the powders in terms of

particle plasticity, as assessed by the Heckel yield pressure, and particle elasticity, as assessed

by the in-die elastic recovery.

4
 ACCEPTED MANUSCRIPT
2. Materials and methods

2.1. Materials

Paracetamol (form I, monoclinic form), α-oxalic acid (anhydrous), and 4,4'-bipyridine

PT
powders were purchased from Sigma-Aldrich Chemie GmbH (Germany). These are

RI
henceforth referred to as PRA, OXA, and BPY. Cocrystals of PRA with OXA and BPY, and a

paracetamol hydrochloride monohydrate salt were prepared and characterized for the solid

SC
state properties as described previously.10 Briefly, cocrystals were produced by stoichiometric

slurrying of pure components in respective solvents while the salt was generated by simple

U
acid-base reaction. These are hereafter denoted PRA-OXA, PRA-BPY, and PRA-HCl. The
AN
materials were dry sieved into three size fractions (<90 µm, 90-360 µm, >360 µm) using a set
M

of precision sieves (VECO, Eerbeek, Holland) and a sieve shaker (Retsch KG, type RV, Haan,

Germany) operating at a relative agitation intensity of 40 for 10 minutes. The size fraction 90-
D

360 µm was selected for further analysis due to its suitability to be used for direct
TE

compression. The materials were stored in a desiccator over a saturated MgCl2 (Sigma-

Aldrich, Germany) solution12 giving a relative humidity of ~33% at room temperature (~ 22

EP

°C) for at least 7 days prior to analysis.

C
AC

PRA and the cocrystals (~1 g of the size fraction larger than 360 µm) were milled for 35

minutes (including a 5 min pause period after 20 minutes) in a planetary ball mill (PM 100,

Retsch, Germany). The milling was done in a 12 cm3 stainless steel container with 50 steel

balls with a diameter of 5 mm. The milled cocrystals showed no alteration in solid state

properties (as inferred from and compared to earlier differential scanning calorimetry

analysis,10 data not shown). The milling of the PRA-HCl crystals altered their solid state

5
 ACCEPTED MANUSCRIPT
properties, as inferred from both smaller and broader peaks in the thermogram, and the milled

powder was consequently not used in any compression experiments.

2.2. Powder characterization

2.2.1. Densities

PT
The apparent particle density (also referred to as the gas pycnometric density in the Ph. Eur.,

 , number of independent measurements, n=3) was determined by helium pycnometry

RI
(AccuPyc 1330, Micrometrics, USA). Each measurement was calculated from an average of

SC
ten repeated cycles. Determined particle densities are reported in Table 1.

U
The poured (unsettled) bulk density (
 , n=2) was measured by manually pouring 10 ml
AN
powder into a graduated measuring cylinder (11.1 mm diameter) and weighed on an analytical
M

balance (Mettler Toledo, AG245, Switzerland). This diameter was selected to enable an

estimation of the in-die poured bulk density and hence enable powder bed height calculations
D

of the powders in-die prior to compression.13 Determined bulk densities are reported in Table
TE

1.
EP

2.2.2. Powder compression

C

A materials testing machine (Zwick Z100, Zwick/Roell Zwick GmbH & Co. KG, Ulm,
AC

Germany) equipped with 11.3 mm diameter flat-faced punches was used for the powder

compression experiments (n=2). These were performed in a humidity controlled room at

~33% relative humidity and at room temperature (~22 °C). In order to reduce the frictional

forces between powder and die wall, the die and the punch faces were lubricated with a 1%

w/w magnesium stearate (Kebo, Sweden) ethanol suspension and air dried prior to

compression. During compression the lower punch was stationary and the upper punch moved

6
 ACCEPTED MANUSCRIPT
with a rate of 10 mm/min up to an applied pressure of 500 MPa. Approximately 400 mg of

each substance was weighed on an analytical balance (Mettler Toledo, AG245, Switzerland)

and poured manually into the die.

The elastic punch deformation was assessed from punch deformation curves obtained by

PT
pressing the upper and the lower punch together up to 500 MPa as described earlier.14 The

RI
punch deformation amounted to approximately 0.4 µm/MPa and was used for correction of

powder bed height during both compression and decompression.

SC
2.2.3. Powder compression analysis

U
The compression analysis during loading was performed according to the classification
AN
protocol suggested by Nordström et al.11 which is based on global compression equations

often used in powder compression analysis, i.e. the Kawakita equation15 and the Heckel
M

equation.16 In addition, the analytically derived Frenning effective medium (EM) equation
D

was used.17
TE

The Kawakita equation15 describes the relationship between engineering strain ( ) and
EP

pressure (
) as,

  
= = ,
C

(1)

AC

where  is the initial powder bed height calculated from
 and  is the powder bed

height during compression.  and   are known as Kawakita parameters and resemble the

maximal engineering strain (  ) at infinite pressure and the pressure required to reach /2,

respectively. The parameters were calculated from the linear form of the Kawakita equation,
  
=  +  , (2)


7
 ACCEPTED MANUSCRIPT
using linear regression in the pressure range 25-500 MPa ( >0.999). The product 

(denoted  -index ) has recently been reported to be indicative of particle rearrangement

during compression.13 An  -index larger than 0.1 suggests extensive particle rearrangement

and classifies the powder as a Class I material.11

PT
The Heckel equation16 describes the change in powder bed porosity (!) with applied pressure

RI
(
) as,

"# $ = %
+ &' , (3)

SC
where ( and &' are the slope and the intercept of the linear region, respectively. 1/( is

U
commonly referred to as the Heckel parameter and considered to represent the yield pressure
AN
(
+ ) of the powder material. The Heckel parameter was calculated by linear regression as

described earlier.18 The pressure range for regression analysis was selected by an Excel macro
M

which identified a minimum derivative as the middle pressure and the first and second

pressure values were identified from a 25% increased derivative in both directions. This
D

derivation was performed in a pressure interval between 0.5 to 500 MPa. For all
TE

compressions, the regression analysis displayed   >0.999.

EP

The Frenning effective medium equation17 describes the relationship between pressure and in-
C

die powder bed height () during compression as

AC

 456

−
 = &- ./ ln 2 3 − ln 2 37, (4)
 456

or in linear form,

(9 ) :;<(456 )/( 456 )=

= &- – ?. (5)
:;(/ ) :;(/ )

The parameter &- , referred to as the Frenning parameter, is the slope of the linear part of the

profile and is earlier proposed to be proportional to the yield stress of the particles18 and

8
 ACCEPTED MANUSCRIPT
represents thus an indication of the particle plasticity. &- was determined by linear regression

in a similar manner as the Heckel parameter, however, the middle pressure value was obtained

from a maximum derivative and the first and the second pressure values from a 25% decrease.

Powder bed heights exceeding 10 mm were excluded in the analysis (  >0.998).  was

calculated from the unsettled bulk density of each powder determined using the glass cylinder

PT
described above and the corresponding pressure
 was thus set at 0 MPa. @A was

RI
calculated based on a compact with a porosity of zero.

SC
2.2.4. In-die elastic recovery

Calculations of the in-die elastic recovery (B@AC@D ) were performed from the complete

U
decompression curves (n=2) according to Eq. 7,19,20
AN
E
B@AC@D = , (6)
E
M

where  is the powder bed height in-die during decompression and F is the height of the

powder bed at maximum load, i.e. 500 MPa. The elastic punch deformation during
D

decompression was set equal to the punch deformation during compression and corrected for
TE

∗
in the calculations of  as described in Sec. 2.2.2. The maximum elastic recovery (B@AC@D )
EP

was used as an indication of the total elastic recovery occurring after decompression in-die
∗
prior to ejection. The B@AC@D was determined from the last pressure recorded by the
C

materials tester that was < 0.06 MPa for all materials except for PRA for which the last
AC

recorded pressure was 0.5 MPa.

An elastic parameter or modulus, denoted BH , was calculated from the reciprocal of the slope

of the relationship between B@AC@D and the pressure difference (

 −
) using a pressure

range of 100-400 MPa. The   of the linear regression was > 0.989 and was regarded as

highly linear although not perfectly linear.

9
 ACCEPTED MANUSCRIPT
2.3. Statistical analysis

Results are given as mean values and where applicable (i.e. in Tables) together with standard

deviations. Statistical differences have been investigated by analysis of variance (ANOVA)

followed by Tukey post hoc tests using the MINITAB®15 statistical software. A p-value <

PT
0.05 was considered as significant.

RI
3. Results and discussion

SC
3.1. Powder compression properties and particle plasticity

U
AN
In order to get a broad description of the compression properties of the powders their
M

pressure-engineering strain relationships were derived (Fig. 1a) and analyzed according to an

earlier published11 classification protocol. In general, the pressure-engineering strain profiles

D

of the cocrystals and the salt crystals had intermediate character compared to the starting
TE

materials i.e., PRA, BPY, and OXA. The pressure-engineering strain relationships were

transformed into linear relationships using the Kawakita equation (Fig. 1b) and the Kawakita
EP

parameters  and   were derived. The compressibility, as assessed by the  parameter

indicating the compressibility at infinite pressure, is given in Table 2. The compressibility of

C
AC

the substances either decreased (PRA) or increased (BPY and OXA) when self-assembled

into cocrystals. In addition, the compressibility of the PRA-HCl decreased in comparison to

PRA. The initial part of the engineering strain-pressure profiles (Fig. 1a) displayed a rapid

increase in with the lowest gradient for OXA. This was reflected by the   parameter,

which was higher for OXA (about 11 MPa) compared to the other substances (Table 2) which

showed a small spread in   parameter ranging from 3 to 6 MPa. The product of  and  

10
 ACCEPTED MANUSCRIPT
i.e., the  -index (Table 2), was for OXA lower than 0.1, i.e. indicating a limited degree of

particle rearrangement during compression. OXA was thus classified as a Class II material.13

The other powders displayed  values > 0.1 and were accordingly classified as Class I

materials displaying extensive particle rearrangement. Similar results have previously been

reported for PRA.21

PT
For all materials, the Heckel yield pressure (
+ )22 was used as an indication of the particle

RI
plasticity. In general, the Heckel profiles (Fig. 2) were similar with respect to the overall

SC
curve shape for all powders except for OXA. All profiles displayed an initial bending that was

followed by a linear region during which plastic particle deformation was the dominating

U
compression mechanism and thus used for calculation of
+ . Finally, a non-linear region at
AN
high compression pressure was obtained for all materials except OXA at which plastic particle
M

deformation became limited and the compact started to deform elastically.23 Based on the

calculated
+ (Table 2) the materials were categorized according to Roberts and Rowe.22 Both
D

the BPY and the PRA-BPY were categorized as soft materials as the
+ of BPY was not
TE

significantly lower than 40 MPa, which was the limit between soft and very soft materials.

PRA-OXA, PRA-HCl, and PRA displayed

+ values between 80-200 MPa and was thereby
EP

classified as moderately hard materials. The very high

+ of OXA reflects the fact that a very
C

low slope of the Heckel profile was obtained over an extended pressure range (Fig. 2). A large
AC

part of the profile was non-linear and the

+ was calculated in a pressure range of 400 to 500

MPa. This may indicate that the particles fractured over a large pressure range or it may be

indicative of a strain hardening of the crystals during compression. Thus, OXA was classified

as a hard material and also brittle.

11
 ACCEPTED MANUSCRIPT
In addition to the Heckel parameter, the Frenning &- parameter was used as an indication of

particle plasticity. The shapes of the EM profiles could as for the Heckel profiles be divided

into three regions, i.e an initial nonlinear, a linear, and finally a bending towards a plateau

(Fig. 3). The shape of the curves was expected with regard to the classification as Class I and

Class IIB materials.18 The derived &- parameters are reported in Table 2 and ranges between

PT
43.0 MPa for BPY and 1016.3 MPa for OXA. A good correlation between &- and
+ was

RI
obtained18 but the values of &- were about a 1.1 factor higher than
+ . The use of the EM

equation supports the classification according to the plasticity scale and the Heckel yield

SC
pressure is henceforth used to represent the material plasticity.

U
AN
In summary, the cocrystals showed different compression properties compared to the starting

materials, i.e. PRA and coformers. Also the salt crystals showed a change in compression
M

properties compared to PRA. In terms of particle plasticity, cocrystals and salt crystals were

more plastic than PRA. The range in plasticity was however relatively low for PRA, PRA-
D

HCl, PRA-OXA, PRA-BPY and three of four materials were classified as moderately hard
TE

and only one material (PRA-BPY) as soft, i.e. PRA-BPY was softer than PRA. In a

compilation of yield pressure values from the literature,1 a range of values from 5 MPa to
EP

nearly 1000 MPa have been reported. Hence, the differences in both &- and
+ reported here
C

indicate a relatively limited change in plasticity while forming the salt crystals and the
AC

cocrystals of PRA.

It is earlier reported that the

+ of a certain substance is influenced by particle size.24-26 Thus,

a consistent sieve fraction of powder was used for all materials in this study. However, since a

broad sieve cut was used, a variation in particle size distribution between the materials may

have existed and affect the obtained variation in

+ values. In order to further compare the

12
 ACCEPTED MANUSCRIPT
plasticity of the materials and their classification, milled powders of the materials were

prepared and studied. The yield pressures of the milled PRA, PRA-OXA and PRA-BPY

powders (Table 3) increased with an average factor of 1.3 and thus, the milled and the finer

particles seemed less plastic than the original particles. However, the relative ranking of the

materials based on their yield pressures as well as the plasticity classification of the materials

PT
was unaltered post milling (Table 3). Thus, this supports the classification of the materials as

RI
moderately hard (PRA and PRA-OXA) and soft (PRA-BPY).

SC
3.2. Tablet elastic recovery and particle elastic deformation

U
A common means to derive a measure of the elasticity of a tablet is to calculate the ratio
AN
between the height of the powder bed in-die during maximal loading and the height of the

powder bed in-die when the powder bed is completely unloaded (e.g. Sun and Hou3 and Sun
M

and Grant18), as described in Eq. (7). Such a value of elastic recovery has been referred to as
D

the immediate axial recovery27 and are reported in Table 4 for the materials used in this study.
TE

∗
The B@AC@D ranged from 4.1% for OXA to 9.6% for PRA (Table 4). The low elastic recovery

for OXA (Table 4) is consistent with other reports on the properties of fragmenting
EP

materials.27,28 In an earlier paper on the compaction of the same materials as used in this

study,10 PRA, PRA-OXA, and PRA-BPY gave tablets which showed lamination and capping
C

∗
while no such tendencies were reported for PRA-HCl, OXA, and BPY. The highest B@AC@D
AC

(Table 4) among the materials were obtained for PRA and PRA-PBY which may partly
∗
explain their capping propensity. However, the B@AC@D for PRA-OXA was significantly

lower than for BPY and equal to PRA-HCl, i.e. two materials which did not give tablets with

visible cracks or caps. Hence, other factors are also of importance for the capping propensity,

such as the degree of tablet mechanical anisotropy.29

13
 ACCEPTED MANUSCRIPT

The Young’s modulus is a measure of elastic deformation and is commonly determined from

bending, indentation or compression testing1 but rarely from the decompression phase.30

Bending and indentation tests are frequently done using compacts, often referred to as beams,

and give thus an indication of the elastic deformation of the compact. It is however argued

PT
that for a compact of zero porosity, the measured modulus may represent the elastic

deformation of the material.1 For relative comparison between materials, compacts of

RI
consistent porosity have also been used.31 In this paper, a parameter was calculated from the

SC
in-die decompression profile at which the compact recover by elastic strain.30,32 Such a

modulus, here referred to as a modulus of elastic deformation (BH ), is considered to give a

U
better indication of the elastic deformation of the material in comparison to a single value of
AN
∗
tablet elastic recovery, i.e. B@AC@D . The parameter (BH ) was calculated from a series of
M

consecutive elastic recovery values (B@AC@D ) as a function of the reduction in upper punch

pressure during unloading,33 i.e. the slope of the relationship between B@AC@D and
 −
.
D

The decompression relationship between B@AC@D and

 −
is presented in Fig. 4.
TE

All materials showed during the initial part of the decompression phase a negative recovery
EP

which is indicative of a time-dependent viscous deformation of the particles, i.e. a creeping of

C

particles under nearly constant loading. It is unclear if this is a plastic or an elastic time-
AC

dependent deformation and it is thus referred to as a viscous deformation. The viscous

deformation was most pronounced for OXA and its cocrystal PRA-OXA.

After the initial viscous stage, the B@AC@D increased nearly linearly over a considerable range

of pressures up to a
 −
of about 450 MPa. Thereafter, a marked bending of the

relationship occurred followed by high rate of elastic recovery during the final stage of

14
 ACCEPTED MANUSCRIPT
unloading. The final stage accounted for the majority of the elastic recovery of all materials as

the resistance towards elastic recovery of the tablet was low due to reduction of the upper

punch pressure.

The elastic modulus (BH ) was calculated for all materials in the pressure range 100-400 MPa

PT
and are reported in Table 4 together with some tablet porosity data from the unloading phase.

For the materials used in this study, the BH ranged from 8.48 GPa (BPY) to 14.6 GPa (OXA).

RI
SC
In Table 4, porosities of tablets at some selected loading states of the decompression phase are

reported. The minimum recorded porosities were negative for all materials except for OXA.

U
An accurate determination of the porosities during the transition between compression and
AN
decompression is difficult but negative porosities have been reported earlier.33 The negative
M

values indicate that the materials were densified at the highest applied pressures due to an

elastic compression of the crystals. In the range of decompression pressures used in

D

calculations of BH (i.e. 100-400 MPa), the porosities of the tablets were generally low and
TE

close to zero porosity and the rates of elastic recovery were low in comparison to the rates of

recovery in the final recovery phase. It is thus proposed that in the decompression pressure
EP

range used for calculation of BH , the tablet elastic recovery was controlled by the elastic strain

of the crystals. This supports that the calculated values of BH represent indications of the
C
AC

elastic deformation of the crystals with the possible exception for OXA. At the final unloading

stage, the porosities increased quickly, giving significant increases in tablet porosities. This

suggests that during the final decompression stage and eventually also during ejection, pores

are formed in the compact which also may be associated with the breakage of inter-particulate

bonds in the compact.

15
 ACCEPTED MANUSCRIPT
Values of Young’s modulus of paracetamol have previously been reported to be approximately

10-11 GPa, derived from bending1 and decompression30 testing. The value obtained by beam

testing was derived by extrapolation of a series of moduli to zero beam porosity. The reported

moduli compare favorably to the estimated value of 11.33 GPa for PRA derived in this paper

(Table 4), further supporting that the calculated values of BH represents the elastic

PT
deformation of the crystals. For the salt, the BH and thus the elastic deformation was the same

RI
as for PRA while for the cocrystals, the moduli were similar but slightly lower. This indicates

that the cocrystals displayed a somewhat lower elastic deformation than PRA. In a

SC
compilation of Young’s moduli from the literature,1 a range of values from 3 GPa to 88 GPa

have been reported. Hence, the small differences in BH reported in this work indicate nearly

U
the same elastic deformation for PRA, the salt and the cocrystals.
AN
M

3.3. Relationships between crystal structure and particle deformation (structure-property-

relationships)
D
TE

As previously mentioned, it is commonly conjectured that the presence of slip planes within

the crystal lattice structure facilitates particle plastic deformation during loading. Slip planes
EP

can be identified by visual inspections of crystal structure and by crystal energy calculations.34
C

The lattice structure of the crystals investigated in this work has been described in a preceding
AC

paper, hence for details regarding the crystal structure and potential slip planes the reader is

referred to Ahmed et al.10 In summary, no slip planes were identified in PRA and PRA-BPY

crystals while slip-planes were identified for OXA, BPY, PRA-OXA and PRA-HCl crystals.

Thus, there was no general relationship between the presence of slip planes and the effective

plasticity of the materials as assessed by the Heckel and the Frenning parameters, i.e.

materials with both high and low plasticity had slip planes in the crystal lattice structure. OXA

16
 ACCEPTED MANUSCRIPT
exhibited a low plasticity despite the presence of slip planes in the crystal structure which may

be explained by the presence of weak inter-planar hydrogen bonds that prevent plane

slipping.34 In addition, based on the results presented by Bansal and co-workers34 it may be

conjectured that both the shape of the lattice structure and the presence of inter-planar

bonding influences the plasticity. The PRA-HCl and PRA powders which displayed

PT
intermediate plasticity have both zig-zag layers in the crystal structure. It seems thus that

RI
relative movement of crystal layers may be possible also in crystals of a zig-zag structure. For

PRA-HCl, the presence of water molecules in the crystal lattice may facilitate the relative

SC
movement of crystal layers and thus affects the deformation properties.35

U
The mode and direction of force application influence the response of the crystals in terms of
AN
plastic deformation. Crystal hardness has often been determined by nano-indentation

techniques, e.g.6,36-38 using relatively well-controlled force application. However, during

M

powder compression, particles are stochastically oriented in the die and subjected to both
D

axial and radial forces, approaching an isostatic loading condition. The net-effect may be that
TE

simple slip-plane movement is difficult or even prohibited due to a complex mode of force

application and a transmission of forces at inter-particulate contacts along chains of particles.

EP

Hence, the plasticity determined using various techniques may be incomparable.

C
AC

It is thus concluded that a relationship between plastic deformation expressed during powder

compression and the presence of slip planes in a crystal is difficult to establish. The

identification of slip planes visually or by calculation of energy attachment may be inadequate

and other approaches to identify potential slip planes may be needed, such as the use of

energy vector diagrams.39 Another complication is that a straightforward relationship between

slip planes and plastic deformation of particles expressed during powder compression may not

17
 ACCEPTED MANUSCRIPT
exist due to the complex loading situation during powder compression. The value of crystal

structure analysis for the identification of structure-powder compression mechanics

relationships may thus be questionable and deserves to be further studied.

Regarding the elastic deformation, the two materials with the highest and lowest BH (OXA

PT
and BPY respectively) both had slip planes while materials showing intermediate BH (PRA

and PRA-BPY) had a zig-zag layered crystal lattice structure.10 As pointed out above, all

RI
materials could be described as of the same elastic deformation. Thus, there seemed to be no

SC
relationship between crystal structure and elastic deformation of the studied crystals.

U
AN
3.4. Relationships between particle deformation and tablet-forming ability (property-
M

performance relationships)
D

It has been reported10 that the cocrystals and the salt (PRA-BPY, PRA-OXA and PRA-HCl)
TE

used in this study could be compacted into tablets in contrast to the monoclinic form of PRA

(form I) which did not form coherent compacts. However, capping was reported for tablets of
EP

the cocrystals giving problems in determination of tablet tensile strength.10 In addition, the co-
C

formers OXA and BPY gave coherent tablets.

AC

It is often assumed that the tablet-forming ability is related to the contact area available for

inter-particle bonding and the inter-particle bonding strength.2 It is further assumed that

particle plastic deformation is fundamental for the contact area development during

compaction. Thus, the contact area is typically larger for particles with high plasticity

compared to particles of low plasticity. In addition, an increased contact area may be observed

18
 ACCEPTED MANUSCRIPT
after particle fracture. Both the cocrystals and the salt used in this study showed a slightly

increased plasticity compared to PRA as inferred from both the Heckel and the Frenning

plastic parameters (Table 2) which may contribute to the difference in tablet-forming ability.

In order to study the relationship between particle plasticity and tablet-forming ability for the

PT
crystals used in this study, the tensile strength of tablets formed at compaction pressures of ~

100 MPa and ~140 MPa (from Ahmed et al.10) was plotted as a function of
+ (Fig. 5). The

RI
main image of Fig. 5 displays the relationship between compaction pressure and
+ for all

SC
studied materials whereas the inserted image displays the relationship between PRA,

cocrystals and salt crystals. The compaction pressures of 100 MPa and 140 MPa were used in

U
the comparison in order to have at least three independent measurements of tablet tensile
AN
strength also for the capping prone crystals. There was no general relationship between tensile
M

strength and particle plasticity. For the powders forming tablets, i.e. the coformers, the

cocrystals and the salt, the tablet tensile strength tended to decrease with a decreased
+ which
D

is opposite to an expected relationship. This is applicable also to the four materials identified
TE

to have slip-planes in their crystal structure (OXA, BPY, PRA-OXA and PRA-HCl). This was

further confirmed from a similar relationship between the tablet tensile strength and the
EP

Kawakita   parameter which have been proposed also to be an indication of particle

C

plasticity.26 Thus, the variation in particle plastic deformation as expressed during powder
AC

compression between the materials is not the primary explanation for the variation in tablet-

forming ability.

The formation of pores during the final decompression stage, i.e. the stage where the tablet

porosities increased rapidly (Table 4), may induce breakage of inter-particle bonds in the
∗
compact. The B@AC@D was reported to be highest (7.8%) for PRA-BPY which may affect the

19
 ACCEPTED MANUSCRIPT
∗
low tablet tensile strength (0.61 MPa) obtained for this powder. The B@AC@D was similar for

PRA-OXA (5.8%) and PRA-HCl (5.9%) and they gave tablets of similar tensile strength after

compaction at 100 MPa (1.18 MPa vs 1.17 MPa) but different tensile strength after

compaction at 140 MPa (1.30 MPa vs 1.64 MPa). Since the particle elastic deformation was

similar for all crystals studied, elastic deformation and tablet elastic recovery could not

PT
explain the different tablet-forming ability of the crystals.

RI
The findings reported in this study thus indicate that other factors than difference in particle

SC
plasticity and elastic deformation contribute to or are the main explanation for the earlier

reported differences in tablet-forming ability. This conclusion seems valid also for the

U
improved tablet-forming ability of the cocrystals and the salt compared to PRA. Possible
AN
factors are particle fragmentation and particle surface energy, i.e. properties that may affect
M

the inter-particle bonding strength rather than the inter-particle contact area.
D

4. Conclusions
TE

In this work, the effective particle plasticity and elastic deformation of paracetamol (PRA)
EP

and two cocrystals (PRA-OXA and PRA-BPY) and a salt (PRA-HCl) of PRA in addition to

the coformers (OXA and BPY) were analyzed and related to the crystal structure and powder
C
AC

tablet-forming ability.

The studied crystals displayed similar effective particle plasticity and elastic deformation

irrespective of the presence of slip-planes or zig-zag layers in the crystal lattice structure.

However, the tablet-forming ability of the cocrystals and the salt was improved compared to

20
 ACCEPTED MANUSCRIPT
the PRA although an obvious relationship between particle deformation properties and tablet

tensile strength could not be observed.

To conclude, the results in this paper show that crystal engineering can be used to modifiy the

mechanical properties and tablet-forming ability of PRA. However, the establishment of a

PT
structure-property-performance relationship for a series of crystals is intricate and there is a

RI
need to improve our understanding of the importance of crystallographic features for the

deformation properties of crystals and the subsequent effect on their tablet-forming ability. In

SC
addition, other factors controlling the evolution in inter-particle bond strength during

compression deserves to be investigated further.

U
AN
References
M

1. Rowe RC, Roberts RJ. Mechanical properties. In: Alderborn G, Nyström C, eds.
Pharmaceutical Powder Compaction Technology, New York: Marcel Deeker. 1996:283-322.
2. Nyström C, Alderborn G, Duberg M, Karehill PG. Bonding surface-area and bonding
mechanism-two important factors for the understanding of powder compactibility. Drug Dev
D

Ind Pharm 1993; 19(17-18):2143-2196.

3. Sun CC, Hou H. Improving Mechanical Properties of Caffeine and Methyl Gallate Crystals
TE

by Cocrystallization. Crystal Growth & Design 2008; 8(5):1575-1579.

4. Vishweshwar P, McMahon JA, Bis JA, Zaworotko MJ. Pharmaceutical co-crystals. J Pharm
Sci 2006; 95(3):499-516.
EP

5. Karki S, Friščić T, Fábián L, Laity PR, Day GM, Jones W. Improving Mechanical
Properties of Crystalline Solids by Cocrystal Formation: New Compressible Forms of
Paracetamol. Adv Mater 2009; 21(38-39):3905-3909.
6. Chattoraj S, Shi L, Sun CC. Understanding the relationship between crystal structure,
C

plasticity and compaction behaviour of theophylline, methyl gallate, and their 1 : 1 co-crystal.
CrystEngComm 2010; 12(8):2466-2472.
AC

7. Joiris E, Martino P, Berneron C, Guyot-Hermann A-M, Guyot J-C. Compression Behavior

of Orthorhombic Paracetamol. Pharm Res 1998; 15(7):1122-1130.
8. Nichols G, Frampton CS. Physicochemical characterization of the orthorhombic polymorph
of paracetamol crystallized from solution. J Pharm Sci 1998; 87(6):684-693.
9. Carless JE, Leigh S. Compression characteristics of powders: radial die wall pressure
transmission and density changes. J Pharm Pharmacol 1974; 26(5):289-297.
10. Ahmed H, Shimpi MR, Velaga SP. Relationship between mechanical properties and
crystal structure in cocrystals and salt of paracetamol. Drug Dev Ind Pharm 2016; 1-9.
11. Nordström J, Klevan I, Alderborn G. A protocol for the classification of powder
compression characteristics. Eur J Pharm Biopharm 2012; 80(1):209-216.

21
 ACCEPTED MANUSCRIPT
12. Nyqvist H. Saturated salt solutions for maintaining specified relative humidities.
International Journal of Pharmaceutical Technology and Product Manufacture 1983; 4(2):47-
48.
13. Nordström J, Klevan I, Alderborn G. A Particle Rearrangement Index Based on the
Kawakita Powder Compression Equation. J Pharm Sci 2009; 98(3):1053-1063.
14. Nordström J, Welch K, Frenning G, Alderborn G. On the physical interpretation of the
Kawakita and Adams parameters derived from confined compression of granular solids.
Powder Technol 2008; 182(3):424-435.
15. Kawakita K, Lüdde K-H. Some considerations on powder compression equations. Powder

PT
Technol 1971; 4(2):61-68.
16. Heckel RW. Density-pressure relationships in powder compaction. T Metall Soc AIME
1961; 221(4):671-675.

RI
17. Frenning G, Mahmoodi F, Nordström J, Alderborn G. An effective-medium analysis of
confined compression of granular materials. Powder Technol 2009; 194(3):228-232.
18. Mahmoodi F, Klevan I, Nordström J, Alderborn G, Frenning G. A comparison between

SC
two powder compaction parameters of plasticity: The effective medium A parameter and the
Heckel 1/K parameter. Int J Pharm 2013; 453(2):295-299.
19. Sun C, Grant DJW. Influence of crystal shape on the tableting performance of L-lysine
monohydrochloride dihydrate. J Pharm Sci 2001; 90(5):569-579.

U
20. Armstrong NA, Haines-Nutt RF. Elastic recovery and surface area changes in compacted
powder systems. Powder Technol 1974; 9(5-6):287-290.
AN
21. Klevan I, Nordström J, Tho I, Alderborn G. A statistical approach to evaluate the potential
use of compression parameters for classification of pharmaceutical powder materials. Eur J
Pharm Biopharm 2010; 75(3):425-435.
M

22. Roberts RJ, Rowe RC. The compaction of pharmaceutical and other model materials - a
pragmatic approach. Chem Eng Sci 1987; 42(4):903-911.
23. Sun CQ, Grant DJW. Influence of elastic deformation of particles on Heckel analysis.
Pharm Dev Technol 2001; 6(2):193-200.
D

24. Vromans H, Lerk CF. Densification properties and compactibility of mixtures of

pharmaceutical excipients with and without magnesium stearate. Int J Pharm 1988; 46(3):183-
TE

192.
25. Fell JT, Newton JM. Effect of particle size and speed of compaction on density changes in
tablets of crystalline and spray-dried lactose. J Pharm Sci 1971; 60(12):1866-1869.
EP

26. Patel S, Kaushal AM, Bansal AK. Effect of particle size and compression force on
compaction behavior and derived mathematical parameters of compressibility. Pharm Res
2007; 24(1):111-124.
27. Haware RV, Tho I, Bauer-Brandl A. Evaluation of a rapid approximation method for the
C

elastic recovery of tablets. Powder Technol 2010; 202(1-3):71-77.

AC

28. Picker KM. Time dependence of elastic recovery for characterization of tableting
materials. Pharm Dev Technol 2001; 6(1):61-70.
29. Mullarney MP, Hancock BC. Mechanical property anisotropy of pharmaceutical excipient
compacts. Int J Pharm 2006; 314(1):9-14.
30. Dwivedi SK, Oates RJ, Mitchell AG. Estimation of elatic recovery, work of
decompression and Young's modulus using a rotary tablet press. J Pharm Pharmacol 1992;
44(6):459-466.
31. Hancock BC, Clas S-D, Christensen K. Micro-scale measurement of the mechanical
properties of compressed pharmaceutical powders. 1: The elasticity and fracture behavior of
microcrystalline cellulose. Int J Pharm 2000; 209(1–2):27-35.
32. Rippie EG, Danielson DW. Viscoelastic stress/strain behavior of pharmaceutical tablets:
Analysis during unloading and postcompression periods. J Pharm Sci 1981; 70(5):476-481.

22
 ACCEPTED MANUSCRIPT
33. Adolfsson A, Nyström C. Tablet strength, porosity, elasticity and solid state structure of
tablets compressed at high loads. Int J Pharm 1996; 132(1-2):95-106.
34. Khomane KS, Bansal AK. Weak Hydrogen Bonding Interactions Influence Slip System
Activity and Compaction Behavior of Pharmaceutical Powders. J Pharm Sci 2013;
102(12):4242-4245.
35. Sun C, Grant DJW. Improved Tableting Properties of p-Hydroxybenzoic Acid by Water of
Crystallization: A Molecular Insight. Pharm Res 2004; 21(2):382-386.
36. Chattoraj S, Shi LM, Chen M, Alhalaweh A, Velaga S, Sun CC. Origin of Deteriorated
Crystal Plasticity and Compaction Properties of a 1:1 Cocrystal between Piroxicam and

PT
Saccharin. Crystal Growth & Design 2014; 14(8):3864-3874.
37. Krishna GR, Shi LM, Bag PP, Sun CC, Reddy CM. Correlation Among Crystal Structure,
Mechanical Behavior, and Tabletability in the Co-Crystals of Vanillin Isomers. Cryst Growth

RI
Des 2015; 15(4):1827-1832.
38. Mishra MK, Ramamurty U, Desiraju GR. Mechanical property design of molecular solids.
Curr Opin Solid State Mater Sci 2016; 20(6):361-370.

SC
39. Upadhyay PP, Bond AD. Crystallization and disorder of the polytypic alpha1 and alpha2
polymorphs of piroxicam. CrystEngComm 2015; 17(28):5266-5272.

U
AN
M
D
TE
C EP
AC

23
 ACCEPTED MANUSCRIPT

Table legends

Table 1: Particle and powder densities. Standard deviations are given in parentheses.

Table 2: Compression parameters and mechanical classification of the powders (standard

deviations are given in parentheses).

Table 3: Yield pressure (

+,@DC ) and relative yield pressure (
+,@DC /
+ ) for milled PRA

PT
and cocrystals. Standard deviations are given in parentheses.
∗
Table 4: Maximal in-die elastic recovery (B@AC@D ), elastic modulus (BH ) and compact

RI
porosity (!) calculated from decompression data (n=2). Standard deviations are given in
parentheses.

U SC
AN
M
D
TE
C EP
AC

24
 ACCEPTED MANUSCRIPT
Figure legends

Figure 1: Typical representations of the non-linear (a) and the linear (b) Kawakita
relationships for the powders.

Figure 2: Representative Heckel profiles for the studied powders in the complete pressure
range (a) and the initial pressure range (b).

Figure 3: Typical effective medium plots for the studied powders.

PT
Figure 4: Typical plots of the in-die axial elastic recovery. Note that the pressure is
normalized with respect to maximum applied pressure (
 −
).

RI
Figure 5: Correlation between tablet tensile strength for tablets compacted at ~100 MPa
(squares) and ~140 MPa (dots) (n ≥ 3, adapted from Ahmed et al.10) and yield pressure (
+ )
for the studied materials. The inserted image displays the relationship for PRA, cocrystals and

SC
salt crystals.

U
AN
M
D
TE
C EP
AC

25
 ACCEPTED MANUSCRIPT
Tables

Table 1: Particle and powder densities. Standard deviations are given in parentheses.

3 3
Substance
 a (g/cm )
 b (g/cm )
PRA 1.287 (0.001) 0.26 (0.001)
PRA-OXA 1.481 (0.002) 0.31 (0.001)
PRA-BPY 1.260 (0.001) 0.34 (0.001)
PRA-HCl 1.420 (0.000) 0.43 (0.002)

PT
OXA 1.903 (0.001) 0.63 (0.002)
BPY 1.232 (0.003) 0.46 (0.000)
a
Apparent particle density (n=3), obtained from Ahmed et al.10

RI
b
Poured bulk density (n=2), obtained from Ahmed et al.10

U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
Table 2: Compression parameters and mechanical classification of the powders (standard
deviations are given in parentheses).
a
 b   c  d  e
Substance (-) (MPa) (-) (MPa) (MPa) Classification
0.811 2.95 0.27 105.7 115.4
PRA (0.000) (0.01) (0.001) (0.6) (0.3) Moderately hard
0.804 3.39 0.24 85.4 93.3
PRA-OXA (0.000) (0.03) (0.002) (1.0) (5.7) Moderately hard
0.746 4.26 0.18 61.2 68.9

PT
PRA-BPY (0.000) (0.01) (0.001) (0.9) (1.2) Soft
0.714 5.98 0.12 91.0 104.5
PRA-HCl (0.000) (0.06) (0.001) (1.1) (7.4) Moderately hard

RI
0.643 11.43 0.06 898.0 1016.3
OXA (0.000) (0.11) (0.001) (3.6) (9.3) Hard
0.651 4.68 0.14 39.0 43.0

SC
BPY (0.000) (0.05) (0.001) (1.0) (1.0) Soft
a
Kawakita parameter (n=2).
b
Kawakita parameter (n=2).
c

U
Rearrangement index (n=2).
d
Heckel yield pressure (n=2).
e
AN
Frenning effective medium parameter (n=2).
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
Table 3: Yield pressure ( , ) and relative yield pressure ( , /  ) for milled PRA
and cocrystals. Standard deviations are given in parentheses.
Substance , (MPa) , /  (-)
PRA 146.3 (18.8) 1.37
PRA-OXA 106.4 (1.4) 1.22
PRA-BPY 79.5 (1.3) 1.28

PT
RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
∗
Table 4: Maximal in-die elastic recovery (
 ), elastic modulus ( ) and compact
porosity () calculated from decompression data (n=2). Standard deviations are given in
parentheses.
∗

   a  b  c  d
Substance (%) (GPa) (%) (%) (%) (%)
9.6 11.33 -4.1 -3.8 -1.1 5.2
PRA (0.0) (0.04) (0.01) (0.07) (0.05) (0.00)
5.8 10.51 -4.2 -3.8 -0.88 1.8
PRA-OXA (0.1) (0.03) (0.14) (0.12) (0.08) (0.18)

PT
7.8 9.59 -5.8 -5.3 -2.1 1.9
PRA-BPY (0.1) (0.02) (0.09) (0.03) (0.05) (0.02)
5.9 11.29 -3.0 -2.8 -0.06 2.9

RI
PRA-HCl (0.1) (0.01) (0.04) (0.02) (0.03) (0.01)
4.1 14.57 9.9 10 12 14
OXA (0.1) (0.14) (0.05) (0.05) (0.07) (0.00)

SC
7.2 8.48 -6.7 -6.1 -2.5 0.66
BPY (0.0) (0.00) (0.07) (0.04) (0.08) (0.08)
a
Minimum detected in-die compact porosity.
b
In-die compact porosity at = 400 MPa corresponding to  − = 100 MPa.

U
c
In-die compact porosity at = 100 MPa corresponding to  − = 400 MPa.
d
In-die compact porosity at ≈ 0 MPa corresponding to  − = 500 MPa i.e. when fully
AN
unloaded but prior to ejection.
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC