Effect of position isomerism on the formation and physicochemical properties of pharmaceutical co-crystals

Article abstract of DOI:10.1002/jps.21824

The effect of position isomerism on the co-crystals formation and physicochemical properties was evaluated. Piracetam was used as the model compound. Six position isomers, 2,3-, 2,4-, 2,5-, 2,6-, 3,4-, and 3,5-dihydroxybenzoic acid (DHBA), were used as th

Full text of DOI:10.1002/jps.21824

Effect of Position Isomerism on the Formation and
Physicochemical Properties of Pharmaceutical Co-Crystals
XIANGMIN LIAO,1 MOHAN GAUTAM,2 ANDREAS GRILL,1 HAIJIAN JIM ZHU1
1Forest Laboratories, Inc., 49 Mall Drive, Commack, New York 11725
2University of Chicago, Chicago, Illinois 60637
Received 9 January 2009; revised 17 April 2009; accepted 20 April 2009
Published online 5 June 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.21824
ABSTRACT: The effect of position isomerism on the co-crystals formation and
physicochemical properties was evaluated. Piracetam was used as the model compound.
Six position isomers, 2,3-, 2,4-, 2,5-, 2,6-, 3,4-, and 3,5-dihydroxybenzoic acid (DHBA),
were used as the co-crystal formers. Co-crystals were prepared on a 1:1 molar ratio by
crystallization from acetonitrile. The solid-state properties of co-crystals were charac-
terized using X-ray powder diffractometry (XRD), differential scanning calorimetry
(DSC), and Fourier transform infrared (FTIR). All co-crystal formers formed co-crystal
with piracetam except 2,6-DHBA. This failure was possibly due to steric hindrance
of two bulk hydroxyl groups and preference of intra-molecular hydrogen bonding
formation between hydroxyl group and carboxylic acid group. The XRD patterns of
resulting co-crystal indicated that they are highly crystalline and different than
parental compounds. Based on the single crystal data, P_23DHBA is orthorhombic
while P_24DHBA, P_34DHBA, and P_35DHB belong to monoclinlic system. The hydro-
gen bonding network patterns of the co-crystals are also different. DSC data showed
that the melting temperatures of resulting co-crystals are all lower than that of
the starting materials. The melting point rank order of the co-crystals is:
P_24DHBA > P_34DHBA > P_23DHBA > P_25DHBA > P_35DHBA. ß 2009 Wiley-Liss,
Inc. and the American Pharmacists Association J Pharm Sci 99:246–254, 2010
Keywords: co-crystal; position isomer; piracetam; dihydroxybenzoic acid
INTRODUCTION
improvement of properties, such as solubility,
dissolution rate, physical stability and mechanical
properties,2–7 through the utilization of co-crystal
concept. As a result, much like salt and polymorph
screenings, co-crystal screening has become a
routine preformulation process, in the pharma-
ceutical research development world. Although
the benefit of improved API properties makes
Pharmaceutical co-crystals represent a class of
compound in which an active pharmaceutical
ingredient (API) complexes with one or more
molecules (co-crystallizing agents) in the crystal
lattice through noncovalent bonding.1 Recently,
pharmaceutical co-crystals are attracting
a
great deal of attention due to their desirable
chemical, physical and biopharmaceutical proper-
ties. Numerous reports showed the successful
co-crystal forms a rewarding endeavor, the
associated labor-intensive and time-consuming
screening process is a constraint. Therefore, a
better understanding of the mechanism behind
co-crystal formation is essential.
Co-crystals design has been exploited by
crystal engineering strategy and is based on the
understanding of molecular interactions between
Correspondence to: Xiangmin Liao (Telephone: 631-858-
5066; Fax: 631-858-5955; E-mail: xiangmin.liao@frx.com)
Journal of Pharmaceutical Sciences, Vol. 99, 246–254 (2010)
ß 2009 Wiley-Liss, Inc. and the American Pharmacists Association
246
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 1, JANUARY 2010

PHARMACEUTICAL CO-CRYSTALS
247
functional groups, especially hydrogen bonding
interactions. The current practice is to first
evaluate the possibility of forming supramole-
cular synthons between API and co-crystal
formers, such as the carboxylic acid–amide,
carboxylic acid–pyridine, amide–amide, and
alcohol–amide synthons. A Cambridge Structural
Database (CSD) survey can usually provide a
rationale for coformer selection process. However,
co-crystallization is a complex molecular recogni-
tion event that can be influenced by many factors,
including geometry, functional group position and
steric hindrance. Consequently, it is difficult and
at times erroneous to predict the possibility of
co-crystal formation using chemical structure
solely as the basis. A recent finding8 showed that
even for highly analogous compounds with similar
molecular structures, skeletons, and functional
groups, co-crystals formed with different guests
molecules.
benzoic acid and 2,5-dihydroxybenzoic acid
(gentisic acid) were reported.10 Six dihydroxy-
benzoic acids (shown in Fig. 1) were selected as co-
crystal formers. The desired supramolecular
synthon was a carboxylic acid–amide and it was
expected that all six dihydroxybenzoic would
form co-crystals based on a molecular interaction
analysis.
Co-crystal preparation was attempted by crys-
tallization from solution evaporation. Co-crystals
were characterized by single crystal XRD, powder
XRD, DSC, and FTIR. Elucidation of single crystal
data was not provided here since it was not the
focus in this study.
EXPERIMENTAL
Materials
The aim of this study was to evaluate the effect
of position isomerism of co-crystal formers on
the formation and subsequent physicochemical
properties of co-crystals. Due to its structural
simplicity, piracetam was chosen as the model
compound in the present study. Piracetam (2-oxo-
1-pyrrolidine acetamide) is a nootropic drug that
enhances cognition and memory, and improves
Alzheimer’s and dementia.9 As shown in Figure 1,
it is a small molecule with an amide functional
group. Recently, co-crystals of piracetam with
Piracetam and 2,5-dihydroxybenzoic acid (25DHBA,
98%) were purchased from MP Biomedicals, Inc.
(Solon, OH). 2,3-Dihydroxybenzoic acid (23DHBA,
99%), 3,4-dihydroxybenzoic acid (34DHBA, 98%),
3,5-dihydroxybenzoic acid (35DHBA, 97%), and 2,6-
dihydroxybenzoic acid (26DHBA, 98%) were
obtained from Sigma–Aldrich, Inc. (St. Louis, MO).
2,4-Dihydroxybenzoic acid (24DHBA, 98%) was
purchased from Fluka, Inc. (Buchs, Germany). All
materials were used as received without further
purification. The corresponding co-crystals were
Figure 1. Chemical structure of piracetam, 2,3-dihydroxybenzoic acid (23DHBA),
2,4-dihydroxybenzoic acid (24DHBA), 2,5-dihydroxybenzoic acid (25DHBA), 2,6-
dihydroxybenzoic acid (26DHBA), 3,4-dihydroxybenzoic acid (34DHBA)and 3,5-
dihydroxybenzoic acid (35DHBA).
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 1, JANUARY 2010

248
LIAO ET AL.
abbreviated as P_23DHBA, P_24DHBA, P_
25DHBA, P_34DHBA, and P_35DHBA, respectively.
Preparation of Co-Crystals
A 1:1 mixture of piracetam (142.0 mg, 1 mmol)
and co-crystal formers (1 mmol) was added to
10–15 mL of acetonitrile in a 20 mL vial. The
solution was then evaporated in a hood. A crystal
suitable for single crystal XRD analysis was
obtained through slow evaporation over a week.
Figure 2. Overlaid XRD patterns of (A) piracetam—
2,3-dihydroxybenzoic acid, (B) piracetam—2,4-dihy-
droxybenzoic acid, (C) piracetam—2,5-dihydroxy-
benzoic acid, (D) piracetam—3,4-dihydroxybenzoic
acid, and (E) piracetam—3,5-dihydroxybenzoic acid.
X-Ray Powder Diffractometry (XRD)
The powder XRD experiments were performed on
a Rigaku miniflex X-ray diffractometer (Danvers,
MA). The sample was placed on
a
zero
background holder and exposed to Cu Ka radia-
tion (30 kV ꢀ 15 mA). The angular range was
2–408 2u, and counts were accumulated at a
continuous scan rate of 18/min. The data analysis
program used was JADE 8.0.
Approximately 3–8 mg of co-crystal was weighed
in an aluminum pan, sealed hermertically, and
heated from 25 to 3008C at 108C/min heating rate
under nitrogen flow.
Differential Scanning Calorimetry (DSC)
Attenuated Total Reflectance Fourier Transform
Infrared Spectroscopy (ATR-FTIR)
A
differential scanning calorimeter (MDSC,
Model Q1000, TA Instruments, Wilmington,
DE) with a refrigerated cooling accessory was
used. The DSC cell was calibrated using indium.
A
Thermo-Nicolet Nexus-670 unit equipped
with a DTGS detector (Thermo Instrument Co.,
Table 1. Crystallographic Data of Co-Crystals
Compounds
P_23DHBA
P_24DHBA
P_25DHBAꢁ
P_34DHBA
P_35DHBA
Empirical formula
Formula weight 296.28
Crystal system
C13H16N2O6
296.28
Orthorhombic
P21212
10.8745(9)
24.235(2)
5.1710(4)
90
C13H16N2O6
296.28
Monoclinic
C2/c
27.596(4)
5.4611(9)
18.300(3)
90
98.049(3)
90
2730.8(8)
8
C13H16N2O6
296.28
Monoclinic
C2/c
27.896(3)
5.1762(5)
19.7879(18)
90
101.090(2)
90
2803.9(4)
8
C13H16N2O6 C13H16N2O6
296.28
Monoclinic
P21/n
296.28
Monoclinic
P21/n
Space group
˚
a (A)
14.3510(12)
5.7003(5)
17.5002
90
14.7824(14)
5.1242(5)
18.5139
90
˚
b (A)
˚
c (A)
ffi (8)
ffi (8)
ffi (8)
90
90
110.371(1)
90
106.875(1)
90
3
˚
Volume (A )
Z
1362.76(19)
4
1342.07(19)
4
1342.0(2)
4
Density (g/cm3)
ABS coefficient (mmꢂ1
Reflections collected
1.444
0.116
16,134
1848
1670
0.0309
0.0822
1.068
1.441
0.115
9760
3088
1.404
0.112
7622
2873
2417
0.043
0.108
1.4661.466
0.117
11,050
3030
)
0.117
15,455
3043
Independent reflections
Observed reflections
R1 [I >2 ffi (I)]
2213
2366
2564
0.0494
0.1048
1.05
0.0381
0.0906
1.012
0.0351
0.0949
1.095
wR2 (all data)
Goodness-of-fit on F2 wR2 (all data)
*data excerpted from reference 10.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 1, JANUARY 2010
DOI 10.1002/jps

PHARMACEUTICAL CO-CRYSTALS
249
Madison, WI) was used for all spectra acquisition.
Samples were placed under a zinc selenide (ZnSe)
attenuated total reflectance (ATR) crystal acces-
sory and a torque of ꢃ20 cN m was applied to
ensure full contact. Two hundred coadded scans
were collected at a resolution of 2 cmꢂ1 within the
area detector diffractometer for a data collection
at 173(2) K. A preliminary set of cell constants was
calculated from reflections harvested from three
sets of 20 frames. These initial sets of frames were
oriented such that orthogonal wedges of reciprocal
space were surveyed. This produced initial
orientation matrices determined from 59 reflec-
tions. The data collection was carried out using Mo
Ka radiation (graphite monochromator) with a
frame time of 10 s and a detector distance of
4.9 cm. A randomly oriented region of reciprocal
space was surveyed to the extent of one sphere and
region of 4000–600 cmꢂ1
.
Single Crystal X-Ray Powder Diffractometry
A single crystal was placed onto the tip of a 0.1 mm
diameter glass capillary and mounted on a CCD
Figure 3. (a, top) The carboxylic acid—amide heterosynthon in the 1:1 co-crystal
of piracetam and 23DHBA. (Bottom) A portion of the hydrogen bond next work in the
1:1 co-crystal of piracetam and 23DHBA. (b, top) The carboxylic acid–amide hetero-
synthon in the 1:1 co-crystal of piracetam and 24DHBA. (Bottom) A portion of the
hydrogen bond next work in the 1:1 co-crystal of piracetam and 24DHBA. (c, top) The
carboxylic acid–amide heterosynthon in the 1:1 co-crystal of piracetam and 34DHBA.
(Bottom) A portion of the hydrogen bond next work in the 1:1 co-crystal of piracetam and
34DHBA. (d, top) The carboxylic acid–amide heterosynthon in the 1:1 co-crystal of
piracetam and 35DHBA. (Bottom) A portion of the hydrogen bond next work in the
1:1 co-crystal of piracetam and 35DHBA.
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 1, JANUARY 2010

250
LIAO ET AL.
˚
to a resolution of 0.77 A. Four major sections of
frames were collected with 0.308 steps in v at four
different f settings and a detector position of ꢂ288
in 2u. The intensity data were corrected for
absorption and decay. Final cell constants were
calculated from the actual data collection after
integration.
RESULTS AND DSICUSSION
All co-crystal formers formed co-crystals with
piracetam except the 26DHBA. The powder XRD
patterns of co-crystals displayed sharp diffraction
peaks with high intensity and a flat baseline
(Fig. 2). The co-crystals all exhibited unique
patterns which are distinguished from both
piracetam and their corresponding co-crystal
formers. The structure of each co-crystal was
determined by single crystal X-ray diffraction. A
summary of crystallographic data are displayed in
Table 1. PXRD patterns of the co-crystals also
matched well with the simulated patterns derived
from single crystal structure. Close examination
of crystallographic data shows that only
P_23DHBA belongs to orthorhombic system while
the remaining co-crystals are monoclinic. More-
over, among monoclinic systems, P_24DHBA and
P_25DHBA belong to C2/c space group while
P_34-DHBA and P_35DHBA are in P21/n space
group. However, the hydrogen bonding network is
the focus of analysis in this study. It is evident
from Figure 3 hydrogen bonds formed between
carboxylic acid and amide moieties in all co-
crystals. For P_23DHBA, P_24DHBA and
P_25DHBA, the 2-hydroxy group not only forms
an intramolecular hydrogen bond with the carbo-
nyl group of the carboxylic acid, but also acts an
acceptor to the anti-oriented NH of the primary
amide. The other hydroxyl group, either in the 3,
Figure 5. (a) FTIR spectra of (A) piracetam, (B) P_
23DHBA, (C) P_24DHBA, (D) P_25DHBA, (E) P_
34DHBA, and (F) P_35DHBA between 1500 and
1750 cmꢂ1 region. (b) FTIR spectra of (A) P_35DHBA,
(B) P_34 DHBA, (C) piracetam, (D) P_25DHBA, (E)
P_24DHBA, and (F) P_23DHBA between 3000 and
3500 cmꢂ1 region.
4, or 5 position, serves as a donor to the ring
carbonyl of piracetam. AS to P_34DHBA and
P_35DHBA, there is no intramolecular hydrogen
bond due to the lack of 2-hydroxy group. The anti-
oriented NH of the primary amide connects with
Figure 6. Overlaid DSC profiles of (A) P_35DHBA,
(B) P_25DHBA, (C) P_23DHBA, (D) P_34DHBA, and
(E) P_24DHBA.
Figure 4. Proposed intermolecular hydrogen bond-
ing formation of 26DHBA.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 1, JANUARY 2010
DOI 10.1002/jps

PHARMACEUTICAL CO-CRYSTALS
251
the carbonyl of the carboxylic acid of another
DHBA, thus forming tetramer. Between
Table 2. Melting Point of Starting Materials and
Corresponding Co-Crystals
a
P_34DHBA and P_35DHBA, some differences
exist in how the hydroxyl groups connect. In
P_35DHBA, one hydroxyl group binds to the ring
carbonyl of piracetam while the other interacts
with one hydroxyl group from another 35DHBA
molecule. As for P_34DHBA, 3-hydroxy group
acts as the hydrogen bond donor to both the 3- and
4-hydroxy group of another 34DHBA molecule
while the 4-hydroxy group donates hydrogen to
the ring carbonyl of piracetam. Based on the above
structural analysis, 26DHBA co-crystal formation
may be inhibited by two factors: First, both 2- and
6-hydroxy groups have a steric hindrance on
the approach of piracetam. This steric effect
was observed in several studies.11,12 Second,
both the 2- and 6-hydroxy groups may form an
intramolecular hydrogen bond with 1-carboxylic
acid group (Fig. 4). This reduces intermolecular
hydrogen bonding ability,11 thus eliminate
the chance of carboxylic acid–amide synthon
formation.
Infrared spectroscopy is a very sensitive tool
used to detect molecular interactions. It has been
used to monitor the secondary structure of
peptides,13 hydrogen bonding structures14 and
others.15 N–H stretching bands is extremely
sensitive in terms of both position and intensity
to changes in hydrogen bonding structures.15 As
mentioned in the previous XRD analysis, hydro-
gen bonds are formed between amide (CO–NH)
and carboxylic acid (COOH) groups. Therefore,
it is expected that the vibration frequency of
CO and NH will be affected by the formation
of hydrogen bonds. Bands between 1500 and
1750 cmꢂ1(mainly CO group stretching mode)
of co-crystals and piracetam are shown in
Figure 5a. The bands at 1690 and 1640 cmꢂ1 of
Starting
Materials
Melting
Point (8C)
Melting
Point (8C)
Co-Crystals
Piracetam
2,3-DHBA
2,4-DHBA
2,5-DHBA
2,6-DHBA
3,4-DHBA
3,5-DHBA
152.1
207.7
224.8
201
172.6
201.4
237.9
P_2,3-DHBA
P_2,4-DHBA
P_2,5-DHBA
133.9
150.5
126.1
P_3,4-DHBA
P_3,5-DHBA
144.1
120.6
piracetam correspond to the symmetric and anti-
–
–
symmetric C O stretching modes, respectively.
In the co-crystals, the band at 1640 cmꢂ1 shifted to
–
–
higher wavenumbers. Since C O stretching mode
of ionized carboxylic acids are below 1600 cmꢂ1
,
this suggests the presence of un-ionized carboxylic
acids and thus new co-crystal phases. In addition,
an analysis of hydrogen bonding data confirmed
that the distance between acid proton and the O
´
˚
atom of the carboxylic acid was constant at 0.88 A
for all co-crystals (Tab. 3). In the 3000–3500 cmꢂ1
N–H stretching band region, piracetam showed
two bands, at 3165 and 3330 cmꢂ1, which were
assigned to intermolecular hydrogen bonded N–H
to C–O. A small shift in both bands is observed
in the FTIR spectra of all co-crystals. This shift
is caused by the change from the hydrogen
bond between two piracetam molecules to the
hydrogen bond between one piracetam molecule
and one dihydroxybenzoic acid molecule. How-
ever, there is a peak at approximately 3430 cmꢂ1
in P_34DHBA and P_35DHBA but not in the other
samples. This high frequency band is assigned
to the O–H stretching frequency of hydroxyl
groups.
Table 3. Selected Hydrogen Bond Parameters of Co-Crystals
´
˚
´
˚
´
˚
Co-Crystals
P_23DHBA
D–Hꢄ ꢄ ꢄA
d(D–H)/A
d(Hꢄ ꢄ ꢄA)/A
d(Dꢄ ꢄ ꢄA)/A
u(D–Hꢄ ꢄ ꢄA)/8
N2–H2Cꢄ ꢄ ꢄO3
O4–H4Cꢄ ꢄ ꢄO2
N2–H2Aꢄ ꢄ ꢄO3
O4–H4Aꢄ ꢄ ꢄO2
N2–H2Aꢄ ꢄ ꢄO3
N2–H2Bꢄ ꢄ ꢄO3#3
O4–H4Cꢄ ꢄ ꢄO2
N2–H2Cꢄ ꢄ ꢄO3
N2–H2Dꢄ ꢄ ꢄO3#1
O4–H4Cꢄ ꢄ ꢄO2
0.88
0.84
0.88
0.84
0.88
0.88
0.84
0.88
0.88
0.84
2.08
1.74
2.06
1.76
2.12
2.29
1.76
2.07
2.22
1.78
2.933(2)
2.5595(19)
2.914(2)
2.5810(18)
2.9731(16)
2.9946(15)
2.5859(15)
2.9233(14)
2.9256(14)
2.6103(12)
163
165.4
163.9
166.3
162.2
137.2
165.1
163.5
136.8
168.9
P_24DHBA
P_34DHBA
P_35DHBA
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 1, JANUARY 2010

252
LIAO ET AL.
Thermo behavior of the resulting co-crystals
was also characterized by DSC, as shown
in Figure 6. All samples displayed a sharp
melting endotherm, indicating highly crystalline
materials. For P_35DHBA, the first endotherm
(ꢃ1058C) could possibly be due to polymorphic
transition and the last endotherm (ꢃ1408C) could
be decomposition after melting. The meting points
of co-crystals and starting materials are tabulated
in Table 2. Their melting points of the co-crystals
Figure 7. (a) Hydrogen bonding network patterns in the 1:1 co-crystal of piracetam
and 2,3-dihydroxybenzoic acid. The arrows show the direction of hydrogen bond from the
donor to the acceptor. (b) Hydrogen bonding network patterns in the 1:1 co-crystal of
piracetam and 2,4-dihydroxybenzoic acid. The arrows show the direction of hydrogen
bond from the donor to the acceptor. (c) Hydrogen bonding network patterns in the
1:1 co-crystal of piracetam and 2,5-dihydroxybenzoic acid. The arrows show the direction
of hydrogen bond from the donor to the acceptor. (d) Hydrogen bonding network patterns
in the 1:1 co-crystal of piracetam and 3,4-dihydroxybenzoic acid. The arrows show the
direction of hydrogen bond from the donor to the acceptor. (e) Hydrogen bonding network
patterns in the 1:1 co-crystal of piracetam and 3,5-dihydroxybenzoic acid. The arrows
show the direction of hydrogen bond from the donor to the acceptor.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 1, JANUARY 2010
DOI 10.1002/jps

PHARMACEUTICAL CO-CRYSTALS
253
range from 120 to 1508C although all co-crystals
are position isomers. The melting point of
co-crystals was all lower than that of starting
materials. The rank order of melting point for
of positions also causes different hydrogen bond-
ing network patterns, which affects the melting
point. Therefore, position isomerism should be
taken into consideration in designing co-crystals
with desirable properties.
co-crystals
was:
P_24DHBA > P_34DHBA >
while
P_23DHBA > P_25DHBA > P_35DHBA,
the rank order of dihydroxybenzoic acids was:
35DHBA > 24DHBA > 23DHBA > 34DHBA ꢅ
ACKNOWLEDGMENTS
25DHBA.
Generally, there is a correlation
between melting point and density. Higher
density means the molecules pack closer in the
crystal lattice and thus leads to higher melting
point. However, a few anomalies are noted that
P_24DHBA has the second lowest density but the
highest melting point. Further, P_35DHBA has
the highest density but the lowest melting point.
While P_34DHBA has the same density as
P_35DHBA, it has the second highest melting
temperature. When closely examining the hydro-
gen bonding network of all co-crystals, some
difference between them are observed. Here the
focus is the hydrogen bond linking the co-crystal
molecular (the arrows pointed in Fig. 7). As shown
in Figure 7b and d, P_24DHBA and P_34DHBA
adopt a regular double chain-antiparallel for-
mation while P_23DHBA, P_25DHBA, and
P_ 35DHBA show a regular double chain-parallel
formation. According to a recent study,16 the
strength of antiparallel formation is higher than
that of parallel formation, which may explain both
P_24DHBA and P_34DHBA have a higher melt-
ing point. However, it is difficult to build a
quantitatively correlation between melting point
and structure since many factors, such as hydro-
gen bonding strength and numbers of hydrogen
bonding, have to be taken into consideration. As
demonstrated by this study, the unexpected large
variation in melting point (308C) underscores the
complexity in correlation between structure and
prediction of chemical and physical properties.
On the other hand, modification of melting
point by co-crystallization may provide additional
pharmaceutical processing pathways, such as
melt crystallization and hot melt extrusion,
without compromising the chemical stability of
API.
The authors would like to thank Dr. Raj Surya-
narayanan (University of Minnesota), Dr. Ritesh
Sanghvi (Forest Laboratories, Inc.), and Leanne
Grafmuller (University of Pennsylvania) for
reviewing the manuscript. Dr. Victor G. Young,
Jr. is acknowledged for providing single crystal
X-ray diffraction data.
REFERENCES
1. Vishweshwar P, McMahon JA, Bis JA, Zaworotko
MJ. 2006. Pharmaceutical co-crystals. J Pharm Sci
95:499–516.
2. Fleischman SG, Kuduva SS, McMahon JA, Moulton
B, Bailey Walsh RD, Rodriguez-Hornedo N, Zawor-
otko MJ. 2003. Crystal engineering of the composi-
tion of pharmaceutical phases: Multiple-component
crystalline solids involving carbamazepine. Cryst
Growth Des 3:909–919.
3. Remenar JF, Morissette SL, Peterson ML, Moulton
B, MacPhee JM, Guzman HR, Almarsson O. 2003.
Crystal engineering of novel cocrystals of a triazole
drug with 1,4-dicarboxylic acids. J Am Chem Soc
125:8456–8457.
4. Trask AV, Motherwell WDS, Jones W. 2005. Phar-
maceutical cocrystallization: Engineering a remedy
for caffeine hydration. Cryst Growth Des 5:1013–
1021.
5. Trask AV, Motherwell WDS, Jones W. 2006.
Physical stability enhancement of theophylline
via cocrystallization. Int J Pharm 320:114–123.
6. McNamara DP, Childs SL, Giordano J, Iarriccio A,
Cassidy J, Shet MS, Mannion R, O’Donnell E, Park
A. 2006. Use of a glutaric acid cocrystal to improve
oral bioavailability of a low solubility API. Pharm
Res 23:1888–1897.
7. Sun CC, Hou H. 2008. Improving mechanical prop-
erties of caffeine and methyl gallate crystals by
cocrystallization. Cryst Growth Des 8:1575–1579.
8. Takata N, Shiraki K, Takano R, Hayashi Y, Terada
K. 2008. Cocrystal screening of stanolone and mes-
tanolone using slurry crystallization. Cryst Growth
Des 8:3032–3037.
CONCLUSIONS
Position of function groups has indeed a signi-
ficant impact not only on the formation of
co-crystals, but also on the physicochemical
properties of the subsequent co-crystals. Change
9. Winblad B. 2005. Piracetam: A review of pharma-
cological properties and clinical uses. CNS Drug
Rev 11:169–182.
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 1, JANUARY 2010

254
LIAO ET AL.
10. Vishweshwar P, McMahon JA, Peterson ML,
Hickey MB, Shattock TR, Zaworotko MJ. 2005.
Crystal engineering of pharmaceutical co-crystals
from polymorphic active pharmaceutical ingredi-
ents. Chem Commun 4601–4603.
14. Dalterio R, Huang XS, Yu K-L. 2007. Infrared and
nuclear magnetic resonance spectroscopic study of
secondary amide hydrogen bonding in benzoyl
PABA derivatives (retinoids). Appl Spectrosc 61:
603–607.
11. Abraham MH, Dearden JC, Bresnen GM. 2006.
Hydrogen bonding, steric effects and thermody-
namics of partitioning. J Phys Org Chem 19:242–248.
12. Dearden JC, Bresnen GM. 2005. Thermodynamics of
water-octanol and water-cyclohexane partitioning of
some aromatic compounds. Int J Mol Sci 6:119–129.
13. Rozenberg M, Shoham G. 2007. FTIR spectra of
solid poly-L-lysine in the stretching NH mode
range. Biophys Chem 125:166–171.
15. Iogansen AV. 1999. Direct proportionality of
the hydrogen bonding energy and the intensifica-
tion of the stretching n(XH) vibration in infrared
spectra. Spectrochim Acta Part
1612.
A 55A:1585–
16. Simperler A, Watt SW, Bonnet PA, Jones W,
Motherwell WDS. 2006. Correlation of melting
point of inositols with hydrogen bonding patterns.
CrystEngComm 8:589–600.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 1, JANUARY 2010
DOI 10.1002/jps