Article

Cite This: Cryst. Growth Des. XXXX, XXX, XXX−XXX pubs.acs.org/crystal

Continuous Manufacturing of Cocrystals Using Solid State Shear

Milling Technology
Published as part of a Crystal Growth and Design virtual special issue Honoring Prof. William Jones and His
Contributions to Organic Solid-State Chemistry
Sachin Korde,†,‡ Sudhir Pagire,†,‡ He Pan,⊥ Colin Seaton,∥ Adrian Kelly,†,§ Yinghong Chen,⊥
Qi Wang,*,⊥ Phil Coates,§ and Anant Paradkar*,†,‡
†
Centre for Pharmaceutical Engineering Science, ‡School of Pharmacy, §Polymer IRC, Faculty of Engineering and Informatics, and
∥
School of Chemistry and Biosciences, University of Bradford, Bradford BD7 1DP, U.K.
⊥
The State Key Laboratory of Polymer Materials Engineering, Sichuan University, Chengdu, China
*
S Supporting Information

ABSTRACT: Solid state shear milling (S3M) is reported as a scalable, continuous, polymer-
assisted cocrystallization technique. A specially designed milling pan was employed to provide
high levels of applied shear, and the addition of a polymeric processing aid enabled generation
of high stress fields. Carbamazepine−salicylic acid cocrystals were produced with 5−25 wt % of
poly(ethylene oxide) (PEO). A systematic study was carried out to understand the effect of
process variables on properties and performance of the cocrystals. S3M offers an important new
route for continuous manufacturing of pharmaceutical cocrystals.

■ INTRODUCTION
A pharmaceutical cocrystal is a multicomponent crystalline
system containing a drug with a hydrogen or halogen bond
forming coformer.1 Solution based cocrystallization has several
drawbacks including the high amount of solvent required,
residual solvent in the product, and low process yields.2 In the
past decade, multidisciplinary teams have developed innovative
technologies in the area of green and scalable technologies for
cocrystal synthesis, including use of ultrasound,3,4 microwaves,5
and hot melt extrusion.6−8 Investigations have also been carried
out to understand challenges in formulation of cocrystals, which
are summarized here.
Jones et al. made a major contribution to the application of
mechanochemistry for cocrystal synthesis.9−11 Mechanochem-
istry provides energy to enable the cocrystallization reaction by
subjecting reactants to forces between sliding/colliding
surfaces;12 for example, grinding together a drug and coformer
in a ball mill has been used to produce cocrystals.9,10,13,14
Cocrystal synthesis using a ball mill can be performed either by
neat grinding or in the presence of a small proportion of
solvent or polymer. Although there have been significant
investigations into cocrystal synthesis using ball mills, it is not
easy to scale up the process. During ball milling, shearing of the
two components occurs when balls impact each other and wall
of the mill. Ball milling is a batch process, where the shear
applied is relatively low, and so it is generally used as a
screening technique.15,16 Andersen and Mack17 successfully
applied the Arrhenius equation to the ball milling process which
will enable conversion of conventional solution based synthesis
to the ball milling process. This approach will address
challenges in scale up and process control in ball milling.
Recently, our group reported continuous thermomechanical
synthesis of cocrystals using a hot melt extruder.7 Significant
advantages have been demonstrated for solvent free continuous
cocrystallization (SFCC)8 using a hot melt extruder compared
to solution cocrystallization and grinding, including high yield
and throughput,18 control of stoichiometry, particle size
control, and suitability for real time monitoring.19 However,
as the material is exposed to shear and high temperature for
relatively long periods of time its stability may be affected in
some cases.
Here we report for the first time a continuous milling
technology for synthesis of cocrystals in the presence of small
amounts of polymers. The solid state shear mill (S3M) (Figure
1, drawn using solid modeling software SOLIDWORKS) is a
pan mill designed by the polymer engineering group at Sichuan
University, China, for tribochemistry.20 The unique design of
the S3M provides extremely high frictional energy at the
Received: December 12, 2017
Revised: March 12, 2018
Published: March 13, 2018

© XXXX American Chemical Society A DOI: 10.1021/acs.cgd.7b01733

Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design
Figure 1. Schematic of continuous cocrystallization using S3M.
tribological interfaces. The high frictional forces cause a rise in
temperature while subjecting the mixture to high shear stresses
for a short duration. S3M has been successfully applied in
nanocomposites, depolymerization, and compatibilization
applications for polymeric systems.20−31 Here, S3M technology
provides a continuous solvent free approach for the
manufacture of cocrystals which may address issues associated
with current technologies.
S3M Design. The S3M (SI 1, Figure 1) consists of two in-
laid pans, a rotor, and a stator made up of wear-resistant
materials with specially treated surfaces. The stator is fixed on
the equipment case and a rotor pan is mounted on the axis
attached to a motor. Engraved diametrical lines divide the pan
surfaces into equal sectors. Within each sector a number of
bevels are present with ridges parallel to the dividing
diametrical lines. The ridges and the bevels of the pan surfaces
are placed facing each other to form a unit cell CDEF-CHGB
(Figure 2),20 ABCD represents its projection on the level
Figure 2. (a) Pattern of moving pan on stationary pan, (b) three
dimensional scheme for unit cell CDEF-CHGB.20
surface. A 3D scissoring action is provided by the change in
shape, volume, and position of the unit cell when the moving
pan surface runs over the stationary pan. The change in volume
of the unit cell results in localized high pressure and shear
forces. The movement of the unit cell in both circular and radial
direction causes continuous contraction and division of material
B DOI: 10.1021/acs.cgd.7b01733
Article
from its entrance at the center of the pan to its exit at the outer
edge.
S3M, characterized by high shear, low residence time, and
the ability to operate without solvents, is already proven in
mechanochemical polymeric applications. The high level of
shear at tribological interfaces coupled with the associated rise
in temperature could also provide sufficient energy to
synthesize cocrystals, with low residence times to minimize
degradation. The objective of this work was to investigate the
potential of S3M as a continuous technology for cocrystal
manufacturing.
■ EXPERIMENTAL SECTION
Materials. Carbamazepine (CBZ, purity: 99.0%, mol. wt. 236.26 g
(g)/mol) was purchased from Taj pharmaceuticals, India (Lot no.
TPL/CARB/002), salicylic acid (SAL, purity: 99.0%, Mol. Wt. 138.12
g/mol) was purchased from Sigma-Aldrich (UK), poly(ethylene
oxide) (PEO N80 grade, mol. wt. 200 000 g/mol) was purchased
from Colorcon (UK). All the organic solvents used for UPLC analysis
were purchased from Sigma-Aldrich (UK).
S3M Processing. Accurately weighed CBZ (236.26 g) and SAL
(138.12 g) in a 1:1 molar ratio were mixed well, and to this PEO N80
was added in the concentration of 5%, 10% 15%, and 25% w/w. This
mixture was blended using a Turbula mixer for 15 min. The resultant
homogeneous mixture was fed to the S3M through the hopper via
central feeding port. The proof of concept mill does not have feed
control so manual feeding was used. The motor speed was maintained
at 500 rpm, and the gap between stator and rotor was kept to the
minimum setting (<1 mm). An agglomerated thread like product was
obtained approximately 10−15 s after addition of the blend to the feed
hopper. The product was collected in a vessel with a thermocouple so
as to measure product temperature. The agglomerated product was
characterized by a range of analytical techniques and dissolution.
Three different S3M designs S3M1, S3M2, and S3M3 having different
pan radii, bevel angles, and slot top widths were used for preliminary
study (SI Table 1). S3M1 and S3M2 were discontinued after
preliminary studies as cocrystal were not obtained using these two
mills.
Solution Crystallization. CBZ/SAL 1:1 pure cocrystals were
prepared by a solution crystallization method. 11.8075 g of CBZ and
6.91 g of SAL were accurately weighed and dissolved in 180 mL of
HPLC grade acetonitrile with simultaneous stirring and heating. The
resultant solution was allowed to cool and stand at room temperature
for 24 h. The resultant mixture was filtered and dried. Cocrystal yield
was 87.6%.
Differential Scanning Calorimetry (DSC). A TA Instruments
Q2000 differential scanning calorimeter, equipped with an RSC90
cooling unit (Crawley, UK), was used to record thermograms of CBZ,
SAL, and processed cocrystal. An indium standard was used for
calibration of the instrument. Around 1.5−3 mg of each sample was
loaded into an aluminum pan, and an accurately weighed blank
aluminum pan was used as reference. All samples were recorded using
a suitable temperature profile set according to the melting behavior of
the API (25−200 °C). A heating rate of 10 °C/min or 20 °C/min was
used and a nitrogen flow rate was kept at 50 mL/min to maintain an
inert environment.
Powder X-ray Diffraction (PXRD). All the cocrystal samples were
crushed to reduce their particle size by milling or grinding in a mortar
and pestle and then subjected to powder X-ray diffraction analysis. A
Bruker D8 diffractometer (Coventry, UK) was used, having an X-ray
wavelength 0.154 nm and a 40 kV Cu source with filament emission
40 mA. Scanning range for all the samples was kept between 2 and 30°
(2θ) using a 0.01° step width and 1 s time count, while the scatter slit
and receiving slit were 0.2° and 1° respectively. A silica sample holder
was used in cases where the available sample size was limited.
Near Infrared Spectroscopy (NIR). NIR analysis for CBZ, SAL,
and cocrystals was carried out using an FT-NIR analyzer (Antaris II,
Thermo Scientific, UK). Powdered samples were placed in glass vials
Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design
with a transparent base and placed on the NIR integrating sphere. A
spectral scan was performed every 30 s in the wavenumber range of
4500−9000 cm−1, averaged over 32 scans with a resolution of 16 cm−1.
Samples of pure components of drug and polymer were also analyzed
by NIR. Data processing was performed using TQ Analyst Operation
software, Thermo Scientific.
Fourier Transform Infrared Spectroscopy (FT-IR). All samples
were analyzed by FT-IR spectroscopy using a PerkinElmer Spectrum
100 FT-IR spectrometer. A small amount of powder or extrudate
sample was placed on a diamond crystal detector. Powdered material
or extrudate was pressed against the diamond crystal to a
predetermined pressure. Once the sample was placed properly,
measurements were recorded. For all samples the scanning scale was
kept in the range of 600−4000 cm−1 with 16 scans per sample. The
assembly involves an electronically stabilized source and detector for
accurate results with high repeatability. The obtained spectrum was
processed using GRAMS/AI software. All samples were analyzed by
FT-IR and compared with spectra of the pure drug and the polymer.
Scanning Electron Microscopy (SEM). SEM was used to study
the morphology and structural features of the produced cocrystals.
Self-adhesive carbon mounts were used to mount samples on
aluminum pin stubs (Agar Scientific, Stansted, UK). SEM images of
the mounted samples were collected using an FEI Quanta 400
scanning electron microscope (Cambridge UK).
Computational Methodology. Molecular models of carbamaze-
pine and salicylic acid were extracted from the known crystal structures
and optimized at the DFT level (TPSS/def2-TZVPPD)32 using the
program ORCA.33 Atomic point charges for each atom were
determined by fitting to the calculated electron density. A PEO
model was constructed using the polymer builder in Materials Studio
(10 repeat units), the conformation was optimized using COMPASSII
force field in Forcite. Atomic point charges for this model were then
calculated in ORCA at the same level of theory as the other molecules.
Intermolecular interactions between selected molecules were
minimized using Drift-Bias Free Differential Evolution (DBF-DE)34
to optimize the intermolecular interaction energy calculated using the
AA-CLP force field35 in the in-house programmer dimer. The DBF-
DE control parameters (P, F, Gmax, Np) were set to 0.5, 0.75, 5000, 60
respectively and each system was run independently 20 times.
In-Vitro Dissolution Study. Dissolution studies for cocrystals were
performed using USP 2 dissolution test apparatus (Copley Scientific,
Type NE4-COP, Serial No. 14355). The dissolution study was carried
out in 1% sodium lauryl sulfate. The dissolution medium volume used
during testing was 900 mL at 37 ± 2 °C and 100 rpm for 1 h. Samples
size was maintained at 100 mg, and samples were analyzed by UPLC
method (SI 4).
■
RESULTS AND DISCUSSION
Cocrystal Preparation. Carbamazepine (CBZ) and
salicylic acid (SAL) in a 1:1 molar ratio were chosen as a
model cocrystal pair. CBZ degrades to form the yellow-colored
impurity iminostilbene at high temperature. Screening of the
product was performed using PXRD, the appearance of the
characteristic peaks of CBZ/SAL cocrystal36 (CSD reference-
MOXWAY) at 2θ values of 6.48, 8.84, 13.08, 16.44°, and a
reduction in relative intensities or disappearance of the
characteristic peaks of CBZ indicated formation of the
cocrystal.
Preliminary trials involving neat grinding of an equimolar
mixture of CBZ and SAL in the S3M showed that the mixture
did not flow in the radial direction over the pan but instead
accumulated in the center of the plates. The accumulated mass
softened due to excessive heating, and as a result, pressure on
the motor exceeded the machine’s threshold and caused the
drive to trip. The final mass was yellow in color indicating
formation of iminostilbene. Recently, there were reports of the
use of polymers to assist mechanochemical synthesis of
C DOI: 10.1021/acs.cgd.7b01733
Article
cocrystals37 by milling10 and hot melt extrusion techniques.
The polymer acts in a similar manner to the addition of
solvents in ball milling experiments, facilitating the trans-
formation process. The polymer can act as a flow aid by
providing a lubricating action. Thus, effective material flow over
the pan surface was achieved by inclusion of 5% of
poly(ethylene oxide) (PEO), a pharmaceutically acceptable
polymer to the mixture. Experiments were designed to
understand the effect of equipment variables and percentage
of polymer in the mixture. The effect of equipment variables
such as the radius of the stator-rotor and bevel angel were
studied using two different mills. A mixture containing a
stoichiometric amount of CBZ and SAL with 5% PEO was fed
to the S3M. The mixtures were fed separately to S3Ms with pan
radii of 10 and 30 cm (SI Table 1). The residence time of the
material in both mills was observed to be less than 1 min. The
product obtained from the large diameter S3M were
agglomerates found to contain pure CBZ: SAL cocrystals.
The product temperature at the exit was between 38 and 45 °C.
The product obtained from the smaller S3M was in powdered
form with a temperature of about 35 °C without any indication
of cocrystal formation. It is a common practice during milling
processes to achieve the desired product properties by
subjecting material to multiple cycles. Therefore, the powdered
product from the small diameter S3M was reprocessed 10 times
and analyzed after each cycle; however results showed that
cocrystals had not been formed.
Another trial was conducted to study effect of bevel angle on
the engraved pan channels, using the large diameter S3M with
low and high bevel angles. The residence time of the material in
the mill with high bevel angle (S3M2) was found to be too
short and resulted in generation of negligible levels of shear.
The S3M with large pan diameter and lower bevel angle
(S3M3) provided sufficient residence time and shear to convert
the material into pure cocrystals.
These preliminary studies indicated that the material did not
receive sufficient shear in the small diameter S3M (S3M1) or
the large diameter S3M equipped with high bevel angles
(S3M2). Failure to achieve cocrystallization even after multiple
cycles in the S3M1 indicates that a reaction did not take place,
suggesting that the threshold shear had not been reached. The
shear applied per unit time even in the S3M1 might be
significantly higher than the manual mortar and pestle grinding
or a laboratory ball mill but was insufficient to impart the
threshold energy level to cause mechanochemical trans-
formation. The typical grinding time in a laboratory ball mill
is between 20 min to 2 h, while it was measured as less than a
minute in the S3M1. Continuous application of low shear for
longer duration is more effective than very short period of high
but subthreshold shear milling. Xu et al. proposed that the
actual distance traveled by the particle in the mill is significantly
greater than the pan radius.19 The material follows a spiral path
between the sliding pans from entrance to exit in the mill.
Though the path followed by the material is much longer than
the radius of the mill, the shear received and temperature rise
caused in the small diameter mill was not enough to generate a
cocrystal phase. The large bevel angle lead to large unit cell size,
which generates less intense levels of shear stress. Further
studies were continued using S3M3, which is termed as S3M
here onward.
Polymer was added primarily to aid the material flow, but it
will also affect the process by dissolving at least part of the
components of the cocrystal pair. The miscibility of the
Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design
cocrystal pair components with the polymer may in turn
plasticize the polymer. DSC thermograms of the physical
mixtures of the CBZ/SAL with 5%, 10%, 15%, and 25% PEO
showed a reduction in the glass transition temperature (Tg) of
the system (SI2 Table 2). The difference between in the
solubility parameters of PEO and CBZ is 5.2, while between
PEO and SAL it is 9.73, suggesting that SAL has higher
miscibility with PEO compared to CBZ.18
The absence of a melting endotherm for CBZ after
processing in S3M (Figure 3) confirmed that there was
Figure 3. DSC thermograms for carbamazepine/salicylic acid 1:1
cocrystals with PEO. [A: Carbamazepine, B: CBZ:SAL 1:1 PURE CC,
C: S3M CBA:SAL 1:1 5% PEO, D: S3M CBZ:SAL 1:1 10% PEO, E:
S3M CBZ:SAL 1:1 15% PEO, F: S3M CBZ:SAL 1:1 25% PEO].
formation of phase pure cocrystals with PEO added as
polymeric aid. PEO acted as a molten solvent matrix during
the S3M processing due to the ultrahigh shear exerted by the
S3M stator and rotor pans. It has been reported that the 3D
scissoring action of S3M extends or breaks down high
molecular weight polymer chains of polymers,24 a similar
kind of behavior might have been exerted on PEO. PEO melts
at 65−70 °C and has a glass transition temperature (Tg) of −67
°C,38 and it is suggested that the shear in S3M generated
sufficient heat to reach the melting temperature of PEO. This
melting/softening of the polymer chains allowed it to solubilize
the CBZ and SAL and initiate the cocrystallization mechanism
within the high stress S3M environment.
PXRD patterns for CBZ/SAL 1:1 with 5%, 10%, and 15%
PEO showed identical 2θ values to the pure cocrystal (Figure
4) supporting the findings of the thermal analysis. It was noted
that in the system with 25% PEO, the PXRD pattern revealed
reduced or null 2θ intensities. From this observation, it can be
concluded that the 25% w/w concentration of PEO produced
an amorphous multicomponent phase. At higher concen-
trations, the solubilizing capacity of PEO is enhanced under the
high shear conditions allowing it to completely solubilize the
CBZ and SAL.
FT-IR analysis of S3M processed cocrystals showed identical
vibrations to pure cocrystals1,15,39 (3504 cm−1 for −NH
stretching, 1661 cm−1 −CO stretching, Figure 5). The
presence of the PEO did not adjust the peak positions.
Pure cocrystals of CBZ:SAL 1:1 showed NIR absorption
spectra in the second overtone region at 1448 and 1474 nm
which differed from the starting components (Figure 6). This
D DOI: 10.1021/acs.cgd.7b01733
Article
Figure 4. PXRD for carbamazepine: salicylic acid 1:1 cocrystals with
PEO processed by S3M. [A: Carbamazepine, B: CBZ:SAL 1:1 PURE
CC, C: S3M CBZ:SAL 1:1 5% PEO, D: S3M CBZ:SAL 1:1 10% PEO,
E: S3M CBZ:SAL 1:1 15% PEO, F: S3M CBZ:SAL 1:1 25% PEO].
Figure 5. FT-IR spectra for carbamazepine: salicylic acid 1:1 cocrystals
with PEO processed by S3M. [A: Carbamazepine, B: Salicylic acid, C:
CBZ:SAL 1:1 PURE CC, D: S3M CBZ:SAL 1:1 5% PEO, E: S3M
CBZ:SAL 1:1 10% PEO, F: S3M CBZ:SAL 1:1 15% PEO, G: S3M
CBZ:SAL 1:1 25% PEO].
distinct change in NIR absorption peaks may due to hydrogen
bond formation between −NH of carbamazepine17,40 and
−COO of salicylic acid. In the case of CBZ/SAL 1:1 cocrystals
with polymer aids, there was no change in the NIR absorption
in the second overtone region at 1448 and 1474 nm.
The nature of agglomerates obtained varied significantly with
the amount of polymer added during processing (Figure 7).
The solution crystallized product formed unagglomerated fine
particulate cocrystal, whereas 5% PEO agglomerates were of
smaller size but harder in nature. The agglomerated cocrystals
containing 10% and 15% PEO were formed as extrudates, the
extrudates containing 15% PEO being denser compared to
those containing 10% PEO. The agglomerates containing 25%
PEO took the form of irregular granules, which were soft at the
time of collection (when product was warm) but hardened
upon cooling.
SEM images revealed that cocrystals obtained by solution
crystallization were in the form of long needles (Figure 7A−E)
compared to crystals with plate morphology from S3M
processing (Figure 7A−E). At lower polymer concentrations,
the structure was more compact, whereas more elongated
Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design
Figure 6. NIR spectra for carbamazepine: salicylic acid 1:1 cocrystals with PEO processed by S3M.
Figure 7. Surface morphology by SEM of carbamazepine/salicylic acid 1:1 cocrystals processed by S3M. [A - Solution cocrystal, B - S3M 5% PEO, C
- S3M 10% PEO, D - S3M 15% PEO, E - S3M 25% PEO].
crystals were produced as polymer content increased.41
Variations in structural morphology and polymer concentration
can affect the dissolution behavior of the cocrystals. Another
major role of the polymer would be to form a continuum,
which allowed generation of stress fields which are critical for
tribochemical transformation to occur in the S3M. The stress-
field set at the tribological surfaces of the stator and rotors
induce strain in the polymer which is passed on to the
crystalline structures evolving in the matrix. The generated
stress varies with the amount of polymer and in turn the strain
induced which not only affects particle/agglomerate size but
will also contribute the morphological changes in the crystals.
■ COMPUTATIONAL STUDY
To understand the interactions between the component species
and to explain the behavior of the system, a molecular modeling
study was carried out. The interaction energy between all pairs
of molecules (CBZ/SAL, CBZ/PEO, SAL/PEO) as calculated
using the AA-CLP force field was minimized using the Drift-
Bias Free Differential Evolution global optimization algorithm.
Each calculation was undertaken 20 times and in all cases a set
of low energy minima were located (SI3 Table 3). The lowest
energy interaction was located between CBZ and SAL, followed
by CBZ-PEO and then SA-PEO (SI3 Table 3). The CBZ/SAL
interaction (Figure 8a) is not the motif observed in the crystal
structure (a R22(8) acid···amide dimer), although this
E DOI: 10.1021/acs.cgd.7b01733
Article
interaction was located as a higher energy solution. The other
solutions display a mixture of hydrogen boding and dispersion
interactions (Figure 8). Optimization of CBZ, SAL, and PEO
molecules together retains the interactions displayed by the
individual optimizations (Figure 8c), although the total energy
is lower than that of the sum of these interactions (Δ = 12 kJ
mol−1) suggesting a slight reduction in the strength of
interaction.
Optimization of molecules with two strands of PEO polymer
displays an intertwining of the PEO molecules with the
cocrystal components filling within the gaps. The strong CBZ/
SAL hydrogen bond is retained, suggesting that pair may
continue to exist within the bulk of the polymer. This supports
the potential for dissolution of the molecular pairs into the
PEO bulk during the processing and promotion of dimer
formation leading to cocrystallization.
In Vitro Dissolution Study. CBZ/SAL cocrystal showed a
significantly higher dissolution rate than to CBZ. More than
90% CBZ was dissolved in 30 min in the case of pure CBZ/
SAL cocrystals produced by solution crystallization. The
dissolution rate of CBZ from cocrystals containing different
amounts of PEO was found to be significantly different. The
dissolution study was conducted on agglomerates in sieve
fraction between 105 and 297 μm. The S3M cocrystal
containing 10% and 25% PEO showed drug release comparable
to solution cocrystal. More than 95% drug was dissolved in 1 h.
Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design Article

Figure 8. Molecular configurations for the lowest energy solutions for (a) CBZ−SAL, (b) CBZ−PEO, (c) SAL−PEO, (d) CBZ−SAL−PEO, (e)
CBZ−2PEO, and (f) CBZ−SAL−2PEO.

Figure 9. Dissolution profile for carbamazepine/salicylic acid 1:1 cocrystals with PEO.

The batch containing 5% PEO showed the slowest dissolution
rate followed by 15% PEO; in 3 h only 80% drug was released
from these batches. The major factor responsible for difference
in dissolution rate was found to be the amount of PEO,
morphology of the cocrystals, and compactness of agglomer-
ates. At 5%, PEO content was sufficient to aid the flow of
powder within the mill but not sufficient to avoid breaking of
growing crystal and melt welding at the contact points of the
crystal under high shear. Dhumal et al. reported similar melt
F DOI: 10.1021/acs.cgd.7b01733
agglomerate formation, for ibuprofen−nicotinamide cocrystal
by hot melt extrusion.6 The cold welding phenomena is also
observed during tabletting due to localized heating at the
contact points between the powder particles. The drug release
rate depends on strength and number of welding spots. The
SEM image also suggests broken crystals and compacted
agglomerate plates which is responsible for slow dissolution
from these cocrystals. At 10% concentration PEO content was
sufficient to avoid shear induced damage and cold welding of
Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design
the cocrystals and enough to act as binder providing the
extrudate structure. PEO is hydrophilic polymer which swells
forming a low viscosity gel layer around the dissolving
agglomerate. At 10% the gel layer was very thin and eroded
quickly, while 15% PEO containing cocrystal agglomerates were
compact which gelled and eroded very slowly causing slower
drug release. At 25% PEO concentration the drug release was
faster which may be attributed to well grown relatively
noncompacted crystal which might have formed on the surface
of the agglomerates during cooling and storage. The high
amorphous content of the 25% PEO batch would also be
responsible for faster drug release.
It shows that S3M based cocrystal agglomerates provide an
opportunity to tailor drug release profiles by selection of
suitable processing variables and polymer concentration.
■
CONCLUSION
S3M is a scalable technology for continuous polymer assisted
manufacturing of cocrystals. Pan diameter, bevel angles, and
polymer concentration were found to be important variables
affecting cocrystallization and performance of the cocrystal−
polymer composite. The polymer played a critical role in the
technology including material processability, stress field set up,
cocrystal morphology, and drug release. The S3M technology
can also be used as a method to achieve continuous solvent and
plasticizer free manufacturing of amorphous solid dispersions.
■
ASSOCIATED CONTENT
*
S Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.cgd.7b01733.
SI 1: S3M mill types and details or dimensions; SI 2:
DSC findings of CBZ/SAL 1:1 physical mixtures. SI 3:
Computational study. SI 4: CBZ/SAL 1:1 cocrystal
dissolution study and analysis (PDF)
■
AUTHOR INFORMATION
Corresponding Authors
*(A.P.) E-mail: A.Paradkar1@Bradford.ac.uk.
*(Q.W.) E-mail: qiwang@scu.edu.cn.
ORCID
Colin Seaton: 0000-0003-4094-720X
Anant Paradkar: 0000-0003-1704-9858
Notes
The authors declare no competing financial interest.
■ ACKNOWLEDGMENTS
Authors would like to acknowledge funding support from
Engineering and Physical Sciences Research Council (EPSRC),
UK (EP/K004204/1, EP/L027011/1 and EP/L020572/1).
■
REFERENCES
(1) Schultheiss, N.; Newman, A. Cryst. Growth Des. 2009, 9, 2950−
2967.
(2) Eddleston, M. D.; Patel, B.; Day, G. M.; Jones, W. Cryst. Growth
Des. 2013, 13, 4599−4606.
(3) Apshingekar, P. P.; Aher, S.; Kelly, A. L.; Brown, E. C.; Paradkar,
A. J. Pharm. Sci. 2017, 106, 66−70.
(4) Aher, S.; Dhumal, R.; Mahadik, K.; Paradkar, A.; York, P. Eur. J.
Pharm. Sci. 2010, 41, 597−602.
(5) Pagire, S.; Korde, S.; Ambardekar, R.; Deshmukh, S.; Dash, R. D.;
Dhumal, R.; Paradkar, A. CrystEngComm 2013, 15, 3705−3710.
G DOI: 10.1021/acs.cgd.7b01733
Article
(6) Dhumal, R.; Kelly, A.; York, P.; Coates, P. D.; Paradkar, A.
Pharm. Res. 2010, 27, 2725−33.
(7) Kulkarni, C.; Wood, C.; Kelly, A.; Gough, T.; Blagden, N.;
Paradkar, A. Cryst. Growth Des. 2015, 15, 5648−5651.
(8) Paradkar, A.; Kelly, A.; Coates, P.; York, P. WO2010013035A1,
2010.
(9) Hasa, D.; Rauber, G. S.; Voinovich, D.; Jones, W. Angew. Chem.,
Int. Ed. 2015, 54, 7371−7375.
(10) Hasa, D.; Carlino, E.; Jones, W. Cryst. Growth Des. 2016, 16,
1772−1779.
(11) Hasa, D.; Jones, W. Adv. Drug Delivery Rev. 2017, 117, 147−161.
(12) Tan, D.; Loots, L.; Frišcǐ ć, T. Chem. Commun. 2016, 52, 7760−
7781.
(13) Majumder, M.; Buckton, G.; Rawlinson-Malone, C.; Williams,
A. C.; Spillman, M. J.; Shankland, N.; Shankland, K. CrystEngComm
2011, 13, 6327−6328.
(14) Chow, P. S.; Lau, G.; Ng, W. K.; Vangala, V. Cryst. Growth Des.
2017, 17, 4064−4071.
(15) Li, S.; Yu, T.; Tian, Y.; McCoy, C. P.; Jones, D. D.; Andrews, G.
P. Mol. Pharmaceutics 2016, 13, 3054−3068.
(16) Fischer, F.; Lubjuhn, D.; Greiser, S.; Rademann, K.; Emmerling,
F. Cryst. Growth Des. 2016, 16, 5843−5851.
(17) Andersen, J. M.; Mack, J. Chem. Sci. 2017, 8, 5447−5453.
(18) Kelly, A. L.; Gough, T.; Dhumal, R.; Halsey, S.; Paradkar, A. Int.
J. Pharm. 2012, 426, 15−20.
(19) Agrawal, A. M.; Dudhedia, M. S.; Zimny, E. AAPS PharmSciTech
2016, 17, 133−47.
(20) Xu, X.; Wang, Q.; Kong, X.; Zhang, X.; Huang, J. Plastic Rubber
Composites Processing Applications 1996, 25, 152−158.
(21) Xia, H.; Wang, Q.; Li, K.; Hu, G. H. J. Appl. Polym. Sci. 2004, 93
(1), 378−386.
(22) Wang, G. Polym. Phys. 2008, 46, 807−817.
(23) Wang, Q.; Cao, J.; Huang, J.; Xu, X. Polym. Eng. Sci. 1997, 37,
1091−1101.
(24) Shao, S.; Chen, Y.; Gu, Y.; Wang, Q. Mater. Manuf. Processes
2006, 21, 173−179.
(25) Shao, W.; Wang, Q.; Wang, F.; Chen, Y. Carbon 2006, 44,
2708−2714.
(26) Shao, W.; Wang, Q.; Wang, F.; Chen, Y. J. Polym. Sci., Part B:
Polym. Phys. 2006, 44, 249−255.
(27) Shao, W.; Wang, Q.; Li, K. Polym. Eng. Sci. 2005, 45, 451−457.
(28) Shao, W.; Wang, Q. J. Appl. Polym. Sci. 2006, 101, 1806−1809.
(29) Liu, Y.; Li, J.; Wang, Q. Mater. Manuf. Processes 2008, 23, 284−
288.
(30) Zhu, J.; Zhang, X.; Liang, M.; Lu, C. J. Polym. Res. 2011, 18,
533−539.
(31) Zhang, W.; Liang, M.; Lu, C. Cellulose 2007, 14, 447−456.
(32) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7, 3297.
Rappoport, D.; Furche, F. J. Chem. Phys. 2010, 133, 134105. Weigend,
F. Phys. Chem. Chem. Phys. 2006, 8, 1057. Hellweg, A.; Hattig, C.;
Hofener, S.; Klopper, W. Theor. Chem. Acc. 2007, 117, 587. Libint: A
library for the evaluation of molecular integrals of many-body
operators over Gaussian functions, Version 2, Edward F. Valeev,
http://libint.valeyev.net/.
(33) Neese, F. WIREs Comput. Mol. Sci. 2012, 2, 73−78. Neese, F.
WIREs Comput. Mol. Sci. 2018, 8, e1327.
(34) Price, K. V. in Advances in Differential Evolution; Chakraborty, U.
K., Ed.; Springer: UK, 2008; 33.
(35) Gavezzotti, A. New J. Chem. 2011, 35, 1360.
(36) Groom, C. R.; Bruno, I. J.; Lightfoot, M. P.; Ward, S. C. Acta
Crystallogr., Sect. B: Struct. Sci., Cryst. Eng. Mater. 2016, 72, 171−179.
(37) Boksa, K.; Otte, A.; Pinal, R. J. Pharm. Sci. 2014, 103, 2904−
2910.
(38) Sigma-Aldrich, Polyethylene Oxide, 181986.
(39) Childs, S.; Rodriguez-Hornedo, N.; Reddy, L. S.; Jayasankar, A.;
Maheshwari, C.; McCausland, L.; Shipplett, R.; Stahly, B. C.
CrystEngComm 2008, 10, 856−864.
(40) Wood, C.; Alwati, A.; Halsey, S.; Gough, T.; Brown, E.; Kelly,
A.; Paradkar, A. J. Pharm. Biomed. Anal. 2016, 129, 172−181.
Cryst. Growth Des. XXXX, XXX, XXX−XXX
Crystal Growth & Design Article

(41) Rahman, Z.; Agarabi, C.; Zidan, A. S.; Khan, S. R.; Khan, M. S.
AAPS PharmSciTech 2011, 12, 693−704.

H DOI: 10.1021/acs.cgd.7b01733
Cryst. Growth Des. XXXX, XXX, XXX−XXX