Structure-stability relationships in cocrystal hydrates: Does the promiscuity of water make crystalline hydrates the nemesis of crystal engineering?

Article abstract of DOI:10.1021/cg901345u

This contribution addresses the role of water molecules in crystal engineering by studying the crystal structures and thermal stabilities of 11 new cocrystal hydrates, all of which were characterized by single crystal X-ray crystallography, powder X-ray diffraction (PXRD), infrared spectroscopy (IR), thermogravimetric analysis (TGA), and differential scanning calorimetry (DSC). The cocrystal hydrates can be grouped into four categories based upon thermal stability: (1) water is lost at <100 °C; (2) water is lost between 100 and 120 °C; (3) water is lost at >120 °C; (4) dehydration occurs concurrently with the melt of the cocrystal. In order to address if there is any correlation between structure and stability, the following factors were considered: type of hydrate (tunnel hydrate or isolated hydrate); number of hydrogen bond donors and acceptors; hydrogen bond distances; packing efficiency. Category 1 hydrates exhibit water molecules in tunnels. However, no structure/stability correlations exist in any of the other categories of hydrate. To complement the cocrystal hydrates reported herein, a Cambridge Structural Database (CSD) analysis was conducted in order to address the supramolecular heterosynthons that water molecules exhibit with two of the most relevant functional groups in the context of active pharmaceutical ingredients, carboxylic acids, and alcohols. The CSD analysis suggests that, unlike cocrystals, there is great diversity in the supramolecular heterosynthons exhibited by water molecules when they form hydrogen bonds with carboxylic acids or alcohols. It can therefore be concluded that the promiscuity of water molecules in terms of their supramolecular synthons and their unpredictable thermal stability makes them a special challenge in the context of crystal engineering.

Full text of DOI:10.1021/cg901345u

DOI: 10.1021/cg901345u
Structure-Stability Relationships in Cocrystal Hydrates: Does the
Promiscuity of Water Make Crystalline Hydrates the Nemesis of Crystal
Engineering?
2010, Vol. 10
2152–2167
Heather D. Clarke, Kapildev K. Arora, Heather Bass, Padmini Kavuru, Tien Teng Ong,
Twarita Pujari, Lukasz Wojtas, and Michael J. Zaworotko*
Department of Chemistry, University of South Florida, CHE205, 4202 East Fowler Avenue,
Tampa, Florida 33620
Received October 28, 2009; Revised Manuscript Received March 4, 2010
ABSTRACT: This contribution addresses the role of water molecules in crystal engineering by studying the crystal structures
and thermal stabilities of 11 new cocrystal hydrates, all of which were characterized by single crystal X-ray crystallography,
powder X-ray diffraction (PXRD), infrared spectroscopy (IR), thermogravimetric analysis (TGA), and differential scanning
calorimetry (DSC). The cocrystal hydrates can be grouped into four categories based upon thermal stability: (1) water is lost at
<100 ꢀC; (2) water is lost between 100 and 120 ꢀC; (3) water is lost at >120 ꢀC; (4) dehydration occurs concurrently with the melt
of the cocrystal. In order to address if there is any correlation between structure and stability, the following factors were
considered: type of hydrate (tunnel hydrate or isolated hydrate); number of hydrogen bond donors and acceptors; hydrogen
bond distances; packing efficiency. Category 1 hydrates exhibit water molecules in tunnels. However, no structure/stability
correlations exist in any of the other categories of hydrate. To complement the cocrystal hydrates reported herein, a Cambridge
Structural Database (CSD) analysis was conducted in order to address the supramolecular heterosynthons that water molecules
exhibit with two of the most relevant functional groups in the context of active pharmaceutical ingredients, carboxylic acids, and
alcohols. The CSD analysis suggests that, unlike cocrystals, there is great diversity in the supramolecular heterosynthons
exhibited by water molecules when they form hydrogen bonds with carboxylic acids or alcohols. It can therefore be concluded
that the promiscuity of water molecules in terms of their supramolecular synthons and their unpredictable thermal stability
makes them a special challenge in the context of crystal engineering.
Crystal engineering,1,2 the understanding of intermolecular
interactions in crystalline solids and their exploitation for the
rational design of new crystal forms with controlled proper-
ties, continues to be a vibrant and growing topic in solid state
chemistry. The relevance of crystal engineering to pharma-
ceutical science is particularly high since crystal forms of
active pharmaceutical ingredients (APIs) are preferred by
industry and regulatory bodies for use in oral dosage for-
mulations of drugs.3 However, crystal forms of APIs have an
Achilles’ heel: their rate of dissolution is directly related to the
thermodynamic (equilibrium) solubility of the crystal form,4 a
physicochemical property that cannot be readily optimized
for a particular clinical need. Indeed, the rate of dissolution
may also be affectedbyparticle size and diffusionphenomena,
but thermodynamic solubility is a fixed property of the crystal
form. Whereas controlled release technologies are available to
modulate the dissolution rate of a highly soluble, highly
permeable API (BCS class I, Scheme 1) or a highly soluble,
poorly permeable API (BCS class III), there is little that can be
done to control a poorly soluble, highly permeable API (BCS
class II) or a poorly soluble, poorly permeable API (BCS class
IV). Hence, the rate of dissolution of an API (BCS II and BCS
IV) critically impacts the pharmacokinetic profile of orally
delivered APIs, and the majority of new chemical entities in
drug development (ca. 80%) suffer from poor solubility.5
Therefore, there is heightened interest in and awareness of
the need to diversify the range of crystal forms exhibited by
APIs, which were traditionally limited to the pure API, its
salts, solvates, and hydrates and polymorphs thereof. Phar-
maceutical cocrystals,6 that is, cocrystals formed via com-
plexation between different molecules, an API and a pharma-
ceutically acceptable “cocrystal former”, have recently gained
interest inthis context as exemplified by bioavailability studies
upon compounds such as fluoxetine hydrochloride (Prozac),7
a sodium channel blocker that was developed by Pfizer,8 and
carbamazepine (Tegretol).9 Pharmaceutical cocrystals can be
highly amenable to design because the very nature of APIs
means that they contain exterior hydrogen bonding groups
that will readily engage in supramolecular synthons10 with
appropriate cocrystal formers. Furthermore, cocrystals that
are highly soluble but form unstable complexes can produce a
“spring-and-parachute” effect with APIs with low bioavail-
ability that can modulate both Tmax and Cmax. In short,
pharmaceutical cocrystals offer innovative approaches to
controlling the bioavailability of APIs in ways that cannot
be readily achieved through traditional approaches.11
Hydrates represent another class of multiple-component
crystal that is of relevance to both crystal engineering and
pharmaceutical science. Water molecules readily incorporate
into crystal lattices because of their small size and their
versatile hydrogen bonding capabilities.12 Indeed, water mole-
cules can even stabilize crystal structures when there is an
imbalance in the number of acceptors and donors13 by form-
ing a diverse arrangement of supramolecular heterosynthons,
and hydrates can be the preferred crystal form of an API14 if
they are more stable to hydration than the corresponding
anhydrate:the antiplateletagentpicotamide15 was foundtobe
*To whom correspondence should be addressed. E-mail: xtal@usf.edu.
Fax: 813-974-3203. Tel: 813-974-3451.
r
pubs.acs.org/crystal
Published on Web 03/30/2010
2010 American Chemical Society

Article
Crystal Growth & Design, Vol. 10, No. 5, 2010 2153
Scheme 1. Cocrystal Formers Used Herein, Abbreviated as Follows: CFA, FER, GAL, INM, NAM, ELA, QUE, HCT, TBR, CAF
Table 1. Crystallographic Data and Structure Refinement Parameters for the 11 Cocrystal Hydrates Presented Herein
FERINM H2O
3
FERNAM H2O
3
ELAINM 1.7H2O
3
GALNAM 0.75H2O
3
GALTBR 2H2O
3
GALINM H2O
3
formula
MW
crystal system
space group
a (A)
C16H18N2O6
334.32
monoclinic
P21/n
6.908 (1)
34.600(6)
7.197(1)
90
116.928(9)
90
1533.7(4)
1.448
C16H18N2O6
334.32
monoclinic
P21/n
6.9389(2)
33.460(1)
7.3179(3)
90
116.652(2)
90
C26H21.4N4O11.7
574.65
triclinic
P1
3.6207(2)
12.4908(6)
13.5227(6)
72.641(3)
89.258(3)
83.979(3)
580.39(5)
1.644
C13H13.5N2O6.75
305.76
C14H16N4O9
386.32
monoclinic
C2/c
27.737(9)
6.664(3)
17.345(8)
90
92.962(9)
90
3202(2)
1.603
C13H14N2O7
310.26
monoclinic
P21/n
17.448(1)
3.709(4)
20.134(2)
90
96.567(5)
90
1294.6(2)
1.592
orthorhombic
P212121
3.7606(1)
15.2613(3)
22.7185(4)
90
b (A)
c (A)
R (deg)
β (deg)
γ (deg)
V/A3
90
90
1518.51(9)
1.462
4
1303.85(5)
1.558
4
Dc/g cm-3
Z
4
1
8
4
2θ range
Nref/Npara
T/K
2.55-63.61
2145/232
100(2)
2.64-67.35
2680/234
100(2)
3.42-65.36
1916/210
100(2)
3.49-68.09
2295/218
100(2)
2.35-25.03
2815/272
173(2)
3.18-65.52
2174/219
100(2)
R1 [I > 2σ (I)]
wR2
GOF
0.0700
0.1527
0.996
0.0336
0.0878
1.026
0.954
0.0439
0.1093
1.037
1.136
0.0323
0.0815
1.085
1.097
0.0563
0.1383
1.076
0.0376
0.0942
1.044
abs coef
0.944
0.136
1.128
GALCAF 0.5H2O
3
C15H17N4O7.5
373.33
monoclinic
C2/c
QUETBR 2H2O
3
C22H24N4O11
518.44
triclinic
P1
ELACAF H2O
3
C22H18N4O11
514.40
triclinic
P1
CFANAM H2O
3
C15H16N2O6
320.30
triclinic
P1
HCTINM H2O
3
C13H16ClN5O6S2
437.88
triclinic
P1
formula
MW
crystal system
space group
a (A)
b (A)
c (A)
19.966(4)
15.800(3)
13.111(3)
90
130.780(3)
90
3131.8(11)
1.584
8
7.0113(3)
11.2981(4)
14.9611(6)
106.545(2)
101.826(3)
94.794(3)
1099.15(8)
1.566
9.705(3)
10.926(5)
11.469(7)
104.74(3)
95.68(3)
113.80(3)
1047.9(9)
1.630
6.2726(2)
7.2730(2)
16.4002(5)
98.430(2)
94.791(2)
95.896(2)
732.48(4)
1.452
7.3100(1)
9.3156(2)
13.5309(3)
74.573(1)
87.442(1)
73.856(1)
852.76(3)
1.705
R (deg)
β (deg)
γ (deg)
V/A3
Dc/g cm-3
Z
2
2
2
2
2θ range
Nref/Npara
T/K
1.86-25.25
2825/291
100(2)
3.17-62.63
3376/341
95(2)
4.09-65.36
3301/349
100(2)
2.74-65.01
2404/243
225(2)
3.39-65.33
2805/268
100(2)
R1 [I > 2σ(I)]
wR2
GOF
0.0383
0.1031
1.044
0.129
0.0429
0.1163
1.037
1.097
0.0549
0.1380
1.013
1.150
0.0342
0.0934
1.050
0.964
0.0349
0.0895
1.064
4.702
abs coef

2154 Crystal Growth & Design, Vol. 10, No. 5, 2010
Clarke et al.
Table 2. Continued
Table 2. Hydrogen Bond Distances and Parameters for Cocrystal Hydrates
hydrogen d (H A)/ D(D A)/
bond
hydrogen d (H A)/ D(D A)/
bond
3 3 3
3 3 3
3 3 3
3 3 3
A
A
—(DHA)
A
A
—(DHA)
FERINM H2O
O-H
O
O
O
O
N
O
O
O
O
O
O
O
N
O
O
N
O
O
O
O
O
O
O
O
N
O
O
O
O
O
O
O
O
O
O
O
O
O
N
O
N
O
O
N
O
O
O
O
O
O
O
O
O
O
N
O
O
O
O
O
O
O
O
N
O
O
O
N
O
O
1.91
2.29
2.681(5)
2.733(5)
2.578(5)
152.0
113.1
161.8
O-H
O
O
O
O
O
O
O
O
O
O
O
O
N
O
O
O
O
O
O
O
O
N
O
1.78
1.93
2.608(3)
2.701(3)
2.706(3)
167.2
152.4
112.6
3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
O-H
O-H
O-H
O-H
N-H
N-H
O-H
O-H
O-H
N-H
N-H
O-H
O-H
O-H
O-H
O-H
N-H
N-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
N-H
N-H
O-H
O-H
O-H
N-H
N-H
N-H
N-H
N-H
N-H
O-H
O-H
1.77
2.27
2.26(3)
1.98(3)
2.10(5)
2.19(5)
1.84
3.037(5) 149(5)
2.835(6) 165(6)
2.997(5) 167(5)
2.991(5) 147(5)
2.08(2)
1.99(2)
1.68(3)
1.75(3)
2.07(2)
2.19(2)
1.61(2)
2.16(2)
2.04(2)
1.90(1)
1.89(2)
2.05(3)
2.48(3)
2.17(3)
2.12(2)
2.08(2)
2.45(2)
2.178(2)
1.94(2)
2.23(2)
2.920(4) 164(4)
2.854(4) 162(4)
2.608(4) 177(2)
2.646(4) 161(2)
2.769(2) 135(2)
2.721(2 118(2)
2.548(2) 166(2)
2.982(2) 157(2)
2.950(2) 171(2)
2.725(4) 160(2)
2.754(3) 166(5)
2.882(6) 159(5)
3.088(2) 130(2)
3.011(3) 161(2)
2.959(2) 164(3)
2.944(2) 168(3)
3.303(2) 168(3)
3.038(2) 174(3)
2.787(2) 163(3)
2.986(2) 148(3)
CFANAM H2O
3
FERNAM H2O
2.624(2)
2.687(1)
155
111
3
2.26
1.66(2)
2.10(2)
2.26(2)
1.91(2)
2.17(2)
1.91
2.587(1) 171(2)
2.992(2) 168(2)
3.010(2) 158(2)
2.777(2) 175(2)
2.973(2) 158(2)
ELAINM 1.7H2O
3
2.749(2)
2.611(2)
2.754(2)
177.1
147.8
112.8
1.86
2.32
HCTINM H2O
3
2.03(3)
2.06(3)
1.98(3)
2.06(3)
1.82(3)
2.04
2.930(3) 174(3)
2.988(3) 173(3)
2.896(2) 169(2)
2.897(2) 157(2)
2.541(3) 145(4)
GALNAM 0.75H2O N-H
3
N-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
2.793(2)
2.732(2)
2.710(2)
2.647(2)
149.4
2.29
1.87
112.9
175.5
169.4
1.82
most stable as the monohydrate, and several drugs such
as cephalexin,16 ampicillin,17 cromolyn sodium (disodium
cromoglycate),18 and sildenafil citrate19 are being developed
or are currently marketed as hydrates. It has been suggested
that ca. 33% of organic compounds form hydrates, whereas
solvates are less prevalent (10%).20 However, a survey of the
Cambridge Structural Database, v5.30 (CSD)21 reveals that
only 8224 of the 159565 molecular organic crystal structures
that have had their 3D coordinates determined contain
discrete water molecules. The corresponding numbers for
organic, organometallic, and coordination compounds sug-
gest that hydrates are more prevalent in general since 49 283
out of the 443 505 structures in the CSD are hydrates. Given
their ubiquity, relevance, and that many hydrates occur from
adventitious water, it should be unsurprising that a number of
researchers have addressed the frequency,22 formation,23 and
the water environment of organic crystalline hydrates,24 and
crystalline hydrates have been classified according to their
structure or energetics.25 Morris and Rodriguez-Hornedo26
proposed a classification system whereby hydrates are sub-
divided into three classes (1) channel hydrates in which water
molecules interact with each other to form tunnels within the
crystal lattice, (2) isolated site hydrates in which water mole-
cules are not directly hydrogen bonded toeach other, (3) metal
ion associated hydrates in which water molecules form strong
interactions with transition metals or alkali metals. Byrn et al.
reported upon the macroscopic behavior of caffeine hydrate
at room temperature and its dehydration, which was ex-
plained by the preferential escape of water molecules along
a channel.27 Other reports have discussed intermolecular
interactions of water molecules in crystalline hydrates with
other water molecules as well as interactions with APIs.28
However, very little published research has addressed the
thermal stability of hydrates in terms of why there is a varia-
tion in the temperature at which the water is lost from
the crystal lattice, and cocrystal hydrates remain a largely
unexplored class of compounds. In this contribution, we
report 11 new cocrystal hydrates and analyze them to address
2.07(2)
1.86(2)
1.91
2.877(3) 160(4)
2.688(4) 163(4)
GALTBR 2H2O
3
N-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
N-H
N-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
N-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
O-H
2.790(3)
2.688(3)
2.747(2)
2.722(5)
2.732(4)
2.671(2)
2.824(3)
174.9
1.88
2.31
161.2
113.0
166.1
158.1
170.5
158.6
1.90
1.93
1.84
2.03
2.50(3)
1.90(4)
2.07(3)
2.13(2)
2.01(3)
2.13(2)
2.17(3)
2.04(3)
1.81
2.862(3) 110(3)
2.807(3) 174(3)
2.902(5) 158(5)
2.975(5) 162(3)
2.853(6) 156(5)
2.961(5) 151(3)
3.013(2) 162(2)
2.891(2) 171(2)
GALINM H2O
3
2.645(2)
2.756(2)
2.832(2)
2.722(2)
2.803(2)
172.8
1.92
2.13
174.9
141.3
113.4
169.2
2.28
1.97
2.12(3)
2.00(3)
1.84(2)
1.96(2)
2.12(2)
2.31(2)
1.76(2)
1.90(2)
1.96
2.888(2) 160(3)
2.853(2) 157(3)
2.703(2) 177(2)
2.750(2) 158(2)
2.857(1) 154(2)
2.712(2) 112(2)
2.705(2) 171(2)
2.752(2) 163(2)
GALCAF 0.5H2O
3
QUETBR 2H2O
2.834(3)
2.722(3)
2.615(2)
2.709(2)
2.705(2)
2.753(2)
2.582(2)
2.822(3)
2.909(3)
2.860(3)
3.115(3)
2.702(4)
2.742(3)
2.708(3)
172.5
3
1.90
1.86
165.2
149.3
113.7
163.0
171.1
147.5
167.4
165.0
160.6
169.9
151.7
113.3
161.3
2.26
1.89
1.92
1.83
1.89
1.97
1.92
2.15
ELACAF H2O
3
1.93
2.30
1.90

Article
Crystal Growth & Design, Vol. 10, No. 5, 2010 2155
Figure 1. (a) Illustration of the water molecule in FERINM H2O that links three FER2INM2 tetramers. (b) Crystal packing in FERINM H2O
reveals 2 þ 2 acid-amide hydrogen bonded tetramers that H-bond to water molecules so as to form H-bonded chains that pack as undulating
3
3
sheets because of CH O hydrogen bonds between adjacent chains.
3 3 3
whether or not there is any correlation between their structure
and thermal stability.
Ellagic Acid Isonicotinamide Hydrate, ELAINM 1.7H2O.
3
3
Ellagic acid dihydrate (20.0 mg, 0.06 mmol) was dissolved in
2 mL of N,N-dimethylacetamide, and isonicotinamide (290
mg, 2.37 mmol) was dissolved in 1 mL of deionized water.
The resulting solutions were mixed and left for slow evapo-
ration. Yellow plates were harvested after one month. m.p.:
300 ꢀC.
Experimental Section
General. All reagents (Scheme 1) were purchased from com-
mercial vendors and used as received. Solvents were obtained
from commercial sources and distilled before use. Single crystals
were obtained via slow evaporation of stoichiometric amounts
of starting materials in an appropriate solvent. Cocrystals
were characterized by infrared spectroscopy (IR), X-ray powder
diffraction (PXRD), differential scanning calorimetry (DSC),
thermogravimetric analysis (TGA), and single crystal X-ray
analysis.
Gallic Acid Nicotinamide Monohydrate, GALNAM 0.75H2O.
3
3
Gallic acid (59.5 mg, 0.348 mmol) and nicotinamide (49.9 mg,
0.405 mmol) were dissolved in 5 mL of ethanol and allowed to
slowly evaporate at room temperature. Colorless needles were
harvested after 3 days. m.p.: 210 ꢀC.
Gallic Acid Theobromine Dihydrate, GALTBR 2H2O. Gallic
3
3
acid (50.5 mg, 0.297 mmol) and theobromine (53.3 mg, 0.296
mmol) were dissolved in 6 mL of water/ethanol (1:1 volume ratio)
by heating and allowed to slowly evaporate at room temperature.
Colorless prisms were harvested after 6 days. m.p.: 270 ꢀC.
Gallic Acid Isonicotinamide Monohydrate, GALINM H2O. Gal-
Ferulic Acid Isonicotinamide Monohydrate, FERINM H2O.
3
3
Ferulic acid (19.9 mg, 0.1 mmol) and isonicotinamide (61.3 mg,
0.5 mmol) were dissolved in 3 mL of ethanol. The solution was left to
slowly evaporate at room temperature and needles of FER-
3
3
lic acid (59.5 mg, 0.348 mmol) and isonicotinamide (49.9 mg,
0.405 mmol) were dissolved in 5 mL of ethanol until a clear solution
was obtained. The solvent was allowed to slowly evaporate at room
temperature. Colorless needles were harvested after 3 days. m.p.:
203 ꢀC.
INM H2O were harvested after 7 days. m.p.: 153 ꢀC.
3
Ferulic Acid Nicotinamide Monohydrate, FERNAM H2O. Feru-
3
3
lic acid (19.9 mg, 0.1 mmol) and nicotinamide (61.3 mg, 0.5 mmol)
were dissolved in 3 mL of ethanol. The solution was left to slowly
evaporate at room temperature. Colorless needles of FER-
Gallic Acid Caffeine Hemihydrate, GALCAF 0.5H2O. Gallic
3
3
acid (50.3 mg, 0.296 mmol) and caffeine anhydrous (57.4 mg,
NAM H2O were harvested after 7 days. m.p.: 127 ꢀC.
3

2156 Crystal Growth & Design, Vol. 10, No. 5, 2010
Clarke et al.
0.296 mmol) were dissolved in 5 mL of methanol and allowed to
slowly evaporate at room temperature. Colorless prisms were
harvested after 14 days. m.p.: 242 ꢀC.
0. 0999 mmol) were dissolved in 5 mL of ethanol and 10 mL of
1:1 mixture of water and ethanol respectively by heating. Resulting
solutions were filtered together and placed in the refrigerator for
slow evaporation. Colorless needles were harvested after 4 days.
m.p.: 293 ꢀC.
Quercetin Theobromine Dihydrate, QUETBR 2H2O. Quercetin
3
3
dihydrate (33.8 mg, 0.0999 mmol) and theobromine (18.0 mg,
Ellagic Acid Caffeine Monohydrate, ELACAF H2O. Ellagic
3
3
acid dihydrate (10.0 mg, 0.0296 mmol) and caffeine (29.0 mg,
0.148 mmol) were dissolved in 5 mL of 1,2-propanediol/ethanol
(6:4 volume ratio) by heating, filtered, and allowed to slowly
evaporate at room temperature. Yellow plates were harvested after
a week. m.p.: 299 ꢀC.
Scheme 2. The Water H-Bond Environments That Have Been
Observed in Cocrystal Hydrates (D: donor; A: acceptor)a
Caffeic Acid Nicotinamide Monohydrate, CFANAM H2O.
3
3
Caffeic acid (9 mg, 0.04995 mmol) and nicotinamide (30 mg
(0.1000 mmol) were dissolved in 3 mL of ethanol/water (1:1
volume ratio) by heating and allowed to slowly evaporate at room
temperature. Colorless needles were harvested after two days. m.p.:
108 ꢀC.
Hydrochlorothiazide Isonicotinamide Monohydrate, HCTINM
3
3
H2O. Hydrochlorothiazide (15 mg, 0.05 mmol) and isonicotin-
amide (6.1 mg (0.049 mmol) were dissolved in 4 mL of ethyl acetate
by heating, filtered, and allowed to slowly evaporate at room
temperature. Colorless plates were harvested after two days. m.p.:
127 ꢀC.
a Only strong H-bonds were included (i.e., no C-H O H-bonds).
3 3 3
Figure 2. Crystal packing in FERNAM H2O exhibits 2 þ 2 acid-amide tetramers that are cross-linked by H-bonded water molecules.
3
Figure 3. Crystal packing of ELAINM 1.7H2O reveals molecular tapes of INM dimers with lateral H-bonding to phenol and carbonyl
moieties of ELA. Water molecules form channels perpendicular to the aromatic rings and donate H-bonds to the phenol moieties of ELA.
3

Article
Crystal Growth & Design, Vol. 10, No. 5, 2010 2157
Figure 4. (a) Crystal packing in GALNAM 0.75H2O reveals molecular tapes of heterodimers of GAL and NAM with disordered water
molecules cross-linking each dimer. (b) Lateral view of heterodimers. (c) Crystal packing shows heterodimers forming an interwoven 3D
network.
3
Differential Scanning Calorimetry (DSC). Thermal analysis was
carried out employing a TA Instruments DSC 2920 differential
scanning calorimeter. Aluminum pans were used for all samples
and temperature calibrations were made using indium as a standard.
An empty pan, sealed in the same way as the sample, was used as a
reference. The thermograms were run at a scanning rate of 2-10 ꢀC/
min from 30 ꢀC on 3-5 mg crystalline samples.
Thermogravimetric Analysis (TGA). TGA was performed with a
Perkin-Elmer STA 6000 simultaneous thermal analyzer. Open
alumina crucibles were used for analysis in the temperature range

2158 Crystal Growth & Design, Vol. 10, No. 5, 2010
Clarke et al.
Figure 5. Crystal packing of GALTBR 2H2O reveals that GAL:TBR dimers are connected by water molecules so as to form H-bonded sheets.
3
30 ꢀC to the appropriate temperature at 2-10 ꢀC/min scanning rate
under nitrogen stream.
one molecule of each component. The crystal structure
is sustained by a 2-point supramolecular heterosynthon
between the carboxylic acid moiety of FER and the amide
moiety of INM (O O: 2.578(5) A, N O: 2.997(6) A).
Infrared Spectroscopy (FT-IR). All crystalline samples were
characterized by infrared spectroscopy using a Nicolet Avatar
320 FT-IR instrument. 2-3 mg of material was used for all samples.
Spectra were measured over the range of 4000-400 cm-1 and data
were analyzed using EZ Omnic software.
3 3 3
3 3 3
The acid-amide supramolecular heterosynthon35 further
dimerizes to form an R2 (8) tetramer sustained by N-H
O
hydrogen bonding (N O: 2.991(5) A, Figure 1a,b). The
4
3 3 3
Powder X-ray Diffraction (PXRD). Bulk samples were analyzed
by X-ray powder diffraction with a Bruker AXS D8 powder
diffractometer. Experimental conditions: Cu KR radiation (λ =
3 3 3
water molecules exhibit DDA H-bonding (Scheme 2): they
link tetramers by donating bifurcated hydrogen bonds to the
methoxy moieties (O O: 3.037(5) A) and the pyridyl
ꢀ
1.54056 A); 40 kV; 30 mA; scanning interval 3-40ꢀ 2θ; time per step
3 3 3
0.5 s. The experimental PXRD patterns and calculated PXRD
patterns from single crystal structures were compared to confirm
the composition of materials.
moieties (OH N: 2.835(6) A) and accepting OH
O
hydrogen bonds from phenolic moieties (O O: 2.681(5)
3 3 3
3 3 3
3 3 3
3 3 3
Single-Crystal X-ray Data Collection and Structure Determina-
tions. Crystalline products were examined under a microscope, and
suitable crystals were selected for single crystal X-ray crystallogra-
A, Figure 1a). CH O hydrogen bonds (C O: 3.608 A)
3 3 3
cross-link the resulting zigzag chains to form an undulating
sheet. Adjacent water molecules lie in tunnels parallel to the c
axis, but they are not hydrogen bonded to each other.
FERNAM H2O. FERNAM H2O crystallizes in the
monoclinic space group P21/n, and it exhibits a crystal
packing similar to that of FERINM H2O. Indeed, the
versatility of the water molecule compensates for the differ-
ent orientations of the pyridyl moieties in NAM and INM,
and FERNAM H2O is isostructural with FERINM H2O.
Water molecules form tunnels along the c axis and are
hydrogen bonded in the DDA environment (Figure 2).
phy. Data were collected on single crystals of GALCAF 0.5H2O
and GALTBR 2H2O on a Bruker-AXS SMART APEX CCD
3
3
diffractometer with monochromatized Mo KR radiation (λ =
0.71073 A), whereas data for single crystals of the hydrates of
FERINM H2O, FERNAM H2O, ELAINM 1.7H2O, GALNAM
0.75H2O, GALINM H2O, QUETBR 2H2O, ELACAF H2O,
CFANAM H2O, and HCTINM 2H2O were collected on a Bruker-
3
3
3
3
3
3
3
3
3
3
3
3
AXS SMART APEX 2 CCD diffractometer with monochroma-
tized Cu KR radiation (λ = 1.54178 A). Both diffractometers were
connected to a KRYO-FLEX low temperature device. Indexing was
performed using SMART v5.62529 or using APEX 2008v1-0.30
Frames were integrated with SaintPlus 7.5131 software package.
Absorption corrections were performed by a multiscan method
implemented in SADABS.32 The structures were solved using SHE-
LXS-97 and refined using SHELXL-97 contained in SHELXTL
v6.1033 and WinGX v1.70.0134 program packages. All non-hydrogen
atoms were refined anisotropically. Hydrogen atoms were placed in
geometrically calculated positions or found in the Fourier difference
map and included in the refinement process using riding model or
without constraints. Occupancy factors were refined for disordered
water molecules of ELAINM 1.7H2O and GALNAM 0.75H2O. For
3
3
ELAINM 1.7H2O. The crystal structure reveals that
ELAINM 1.7H2O contains one ELA molecule, two INM
3
3
molecules, and disordered water molecules in the unit
cell. ELAINM 1.7H2O exhibits hydrogen bonding bet-
3
ween the pyridyl group of INM and the phenolic hydrogen
atoms of ELA (O N: 2.611(2) A). The neutral nature of
3 3 3
INM is supported by the C-N-C angle of 118.07ꢀ in
the pyridyl moiety. Two INM molecules further interact
through amide-amide supramolecular homosynthons
3
3
ELAINM 1.7H2O, it was not possible to locate all hydrogen atoms
attached to disordered water molecules.
3
(N O: 2.988(3) A). The phenol-pyridine supramolecular
3 3 3
heterosynthon36 and the amide dimer supramolecular
homosynthon37 generate a tape (Figure 3). The amide
dimer also acts as both a hydrogen bond donor and
Crystallographic data are presented in Table 1. Hydrogen bond
parameters are given in Table 2.
acceptor toward the carbonyl moiety (NH O: 2.930(3)
3 3 3
Results
A) and phenolic hydrogen atom of ELA (OH O:
3 3 3
FERINM H2O. FERINM H2O crystallizes in the mono-
3
2.749(2) A) to form an R33(10) trimer that links the INM-
ELA chains. Disordered water molecules form channels
3
clinic space group P21/n, and the asymmetric unit contains

Article
Crystal Growth & Design, Vol. 10, No. 5, 2010 2159
Figure 6. (a) R34(8) tetrameric motif involving two INM molecules, a GAL molecule, and a water molecule; (b) R44(8) tetrameric motif
involving phenolic moieties of two GAL molecules and two water molecules. (c) The crystal packing of GALINM H2O reveals sheets sustained
by COOH N supramolecular heterosynthons. (d) Crystal packing is that of a 3D network with water molecules (green) parallel to the b axis.
3
3 3 3
parallel to the crystallographic a axis and exist in DDA
environments.
GALNAM 0.75H2O. GAL and NAM crystallize with a
3
disordered water molecule in the orthorhombic space group

2160 Crystal Growth & Design, Vol. 10, No. 5, 2010
Clarke et al.
Figure 7. (a) Crystal packing of GALCAF 0.5H2O tapes cross-linked by water molecules. (b) The water molecules sustain bilayers that pack
3
ABAB.
P212121. GAL and NAM form a 1:1 complex sustained
by acid-amide supramolecular heterosynthons35 (N O:
consists of GAL and TBR molecules that form acid-imide
supramolecular heterosynthons (N O: 2.790(3) A, O
O: 2.671(2) A) interconnected by ordered water molecules to
form 6-component assemblies that are further connected by
disordered water molecules into sheets (Figure 5). The GAL:
TBR dimers are also connected through water molecules to
3 3 3
3 3 3
3 3 3
2.896(2) A, O...O: 2.647(2) A, Figure 4a). The heterodimers
link to one another via OH Narom H-bonds (2.710(2) A) to
3 3 3
adjacent heterodimers that are at dihedral angles of 46.3 ꢀC.
In addition, the carbonyl moiety of GAL accepts a hydrogen
bond from the NH2 moiety of an adjacent heterodimer at a
GAL:TBR dimers that lie above and below the sheets (O
3 3 3
dihedral angle of 46.83 ꢀC (N O: 2.897(2) A). The dis-
O: 2.902(5) A) and by π-π stacking so that the overall crystal
packing can be described as a 3D network. The disordered
water molecules exist in a DDAA environment, whereas the
ordered water molecules H-bond in a DDA environment.
GALINM H2O. The crystal structure of GALINM H2O
3 3 3
ordered water molecule is in a DDA environment and
connects three heterodimers as follows: accepting a bifur-
cated H-bond from the phenolic moieties of contiguous
heterodimers; donating an H-bond to the carbonyl moiety
of an adjacent heterodimer in an offset manner (Figure 4b);
H-bonding with phenolic moieties so as to forms tunnels
parallel to the a-axis. The crystal packing is that of an
interwoven 3D H-bonded network (Figure 4c).
3
3
reveals a 1:1 cocrystal monohydrate that crystallizes in
space group P21/n with one molecule of each component in
the asymmetric unit. The crystal structure is sustained by the
COOH Narom supramolecular heterosynthons,38 but the
amide-amide supramolecular homosynthon is absent. In-
stead, the carbonyl moiety of INM H-bonds to the amide
moiety of an adjacent molecule of INM (N O: 2.891(2) A),
3 3 3
GALTBR 2H2O. The crystal structure of GALTBR
3
3
2H2O reveals a 1:1 cocrystal dihydrate which crystallizes in
the monoclinic space group C2/c. The crystal structure
3 3 3

Article
Crystal Growth & Design, Vol. 10, No. 5, 2010 2161
Figure 8. (a) Crystal packing in QUETBR 2H2O is sustained by dimers of QUE and TBR molecules linked by pairs of water molecules. (b)
Lateral view of the sheets that are sustained by R66(12) motifs.
3
a water molecule and a GAL molecule to generate a tetrameric
TBR molecules (O O: 2.753(2) A). The QUE and TBR
3 3 3
ring motif (Figure 6a). A centrosymmetric R4 (8) tetrameric
ring motif occurs between phenol moieties of two different
GAL molecules and two water molecules, (O O = 2.853(2)
homodimers alternate to form molecular tapes (Figure 8a).
The two crystallographically independent water molecules
(DDA and DDAA environments) cross-link the homodi-
4
3 3 3
and 2.833(2) A, Figure 6b). Overall, the crystal packing can be
described as a 3D H-bonded network with water molecules in
DDAA environments that form tunnels parallel to the b axis.
mers through an R6 (12) ring motif involving four water
6
molecules and the phenolic moieties of two QUE molecules
(Figure 8b).
GALCAF 0.5H2O. The crystal structure of GAL-
3
ELACAF H2O. Analysis of the crystal structure reveals
3
CAF 0.5H2O reveals a 1:1:0.5 ratio of gallic acid, caffeine,
3
that ELACAF H2O contains two ELA molecules, two CAF
3
and water molecules. GAL and CAF molecules form H-
N
38 supramo-
molecules, and two water molecules in the unit cell. The
phenolic hydrogen atoms of ELA act as H-bond donors to
basic nitrogen and carbonyl moieties of CAF molecules at
2.702(4) and 2.708(3) A so as to form straight chains of
alternating ELA and CAF molecules. Adjacent chains are
bonded tapes that are sustained by COOH
3 3 3
lecular heterosynthons (O N: 2.705(2) A) and OH
supramolecular heterosynthons formed between adjacent
O
3 3 3
3 3 3
phenolic moieties of GAL molecules and carbonyl moieties
of adjacent CAF molecules (O O: 2.703(2), 2.750(2) A).
linked via OH (ELA) O = C(CAF) H-bonds (2.708(3) A)
3 3 3
3 3 3
These tapes lie parallel to the a axis and are cross-linked by
water molecules that H-bond with the third phenolic moiety
of each GAL molecule (O O: 2.857(1) A, Figure 7a). The
and water molecules at (2.609(3) and 2.854(4) A) that
H-bond to two ELA molecules, thereby forming a sheet
(Figure 9a). Adjacent sheets are cross-linked by H-bonds
between water molecules and carbonyl oxygen atoms of ELA
(2.920(4) A). Water molecules are isolated and sit in DDA
environments (Figure 9b).
3 3 3
neutral nature of the molecules is supported by the C-O
bond distances, 1.320 and 1.224 A and the C-N-C bond
angle, 103.49ꢀ. The water molecules exist in DDAA environ-
ments and generate bilayers that stack in an ABAB manner
sustained by π-π interactions.
CFANAM H2O. The crystal structure of CFANAM
3
3
H2O reveals that CFA and NAM molecules form carboxy-
lic acid-amide supramolecular heterosynthons (O O:
QUETBR 2H2O. The crystal structure of QUETBR
3
3
3 3 3
2H2O reveals supramolecular homosynthons of QUE di-
mers (O O: 2.615(2) A) and TBR dimers (N O: 2.834(3)
A). These dimers interact via donation of H-bonds from the
2.548(2) A; N O: 2.950(2) A) that self-assemble to form
3 3 3
tapes rather than 2 þ 2 tetramers. These tapes are cross-
3 3 3
3 3 3
linked by phenolic dimer supramolecular homosynthons
(O O: 2.769(2) A, Figure 10). Disordered water molecules
phenolic moiety of QUE molecules to the carbonyl moiety of
3 3 3

2162 Crystal Growth & Design, Vol. 10, No. 5, 2010
Clarke et al.
Figure 9. (a) Crystal packing in ELACAF H2O reveals H-bonding between ELA and CAF molecules via OH N supramolecular
3 3 3
3
heterosynthons. (b) View along the c axis illustrates isolated water molecules.
Figure 10. Molecular tapes of CFANAM H2O cross-linked by phenol-phenol interactions.
3
exist in DDAA environments as they H-bond with the
phenolic dimers via R6 (12) motifs and form tunnels parallel
to the b axis.
to the pyridyl moieties of INM molecules and the sulfona-
mide groups of HCT molecules (O N: 2.787(2) A,
6
3 3 3
O
O: 2.986(2) A, Figure 11). Water molecules are in
3 3 3
HCTINM H2O. The crystal structure of HCTINM H2O
3
DDAA environments since they accept H-bonds from
two different HCT molecules and they form tunnels in
which they are not H-bonded to each other. The overall
crystal packing consists of H-bonded sheets that pack via
π-π stacking.
3
reveals that it is a monohydrate of the 1:1 cocrystal of HCT
and INM. INM molecules form amide-amide supramole-
cular homosynthons (N O: 3.038(2) A) that are linked
to HCT molecules by water molecules that donate H-bonds
3 3 3

Article
Crystal Growth & Design, Vol. 10, No. 5, 2010 2163
Figure 11. The crystal packing of HCTINM H2O is sustained by amide-amide supramolecular homosynthons between INM molecules that
are in turn H-bonded to HCT molecules via water molecules.
3
Discussion
Group 2. The Group 2 hydrates exhibit dehydration
temperatures from 109 to 118 ꢀC. Tunnels of water molecules
Table 3 tabulates selected structural features of the 11
cocrystal hydrates reported herein and 8 cocrystal hydrates
involving zwitterions that have been reported elsewhere.39
The following structural features were addressed: local water
environment in terms of strong H-bond donors and acceptors,
whether water molecules are isolated or in tunnels, and
packing efficiency index. Table 3 also provides parameters
related to thermal stability: temperature at which water loss
occurs, melting point, and result of dehydration. In terms of
thermal stability, the cocrystal hydrates in Table 3 can be
grouped as follows: (1) water is lost at <100 ꢀC for loosely
held water or channel hydrates; (2) water is lost between 100
and 120 ꢀC for water molecules exhibiting the hydrogen
bonding patterns of bulk water; (3) water is lost at >120 ꢀC
for tightly bound crystalline water molecules; (4) dehydra-
tion occurs concurrently with the melt of the cocrystal. This
is grouped according to the classification of hydrates pro-
posed by Morris and Rodriguez-Hornedo26 where the
dehydration temperature is related to the degree of how
tightly the water molecule is bound, that is, whether the
water molecule is an integral part of the crystal structure or
loosely held. Each group is discussed individually and then
collectively.
not H-bonded to each other are exhibited in GALNAM
0.75H2O, GALINA H2O, GALNAC H2O, and GALTBR
2H2O, whereas in PCAINA H2O and GALNAC 1.5H2O
3
3
3
3
3
3
isolated water molecules are observed. The water molecules
are H-bonded in DDA environments in all Group 2 cocrystal
hydrates. However, in GALTBR 2H2O and GALNAC
3
3
1.5H2O, DDAA environments are also exhibited by a second
crystallographically independent water molecule. Packing
indices range from 68.5 to 75.9. Dehydration of Group 2
hydrates afforded crystal forms with new crystal packing
according to PXRD patterns.
Group 3. With the exceptions of CITINA 2H2O, QUEI-
3
NA H2O, and GALINM H2O, crystal structures in Group 3
3
3
contain waters of crystallization that lie in isolated sites.
GALINM H2O, QUEINA H2O, and CITINA 2H2O exhi-
3
3
3
bit water loss at 124, 175, and 136 ꢀC, respectively. These
crystal structures displayed tunnels of water molecules not
hydrogen bonded to each other. The crystal structures of
Group 3 hydrates reveal DDA water environments except
GALCAF 0.5H2O, CITINA 2H2O, and QUETBR 2H2O,
3
3
3
which display DDAA environments, and BTNRES H2O,
3
which display a DD environment. The packing indices range
from 70.4 to 75.6. PXRD of the dehydrated crystal forms
reveals new crystal packing for BTNRES H2O, GALINM
Group 1. Inspection of group 1 reveals that all the cocrystal
hydrates in this group exhibit water molecules that lie in
tunnels and they are not hydrogen bonded with each other
3
3
H2O, CITINA 2H2O, ELACAF H2O, and QUEINA H2O.
3
3
3
However, dehydration of GALCAF 0.5H2O and QUETBR
3
3
except for ELAINM 1.7H2O, in which water molecules
2H2O resulted in isostructural anhydrates.
Group 4. In Group 4, water loss occurs concurrently with
3
form H-bonded chains parallel to the a axis. The dehydration
of Group 1 cocrystal hydrates occurs at temperatures ran-
ging from 58 to 98 ꢀC. Group 1 cocrystal hydrates exist in the
melt of the cocrystal. The water molecules in CFANAM
H2O and HCTINM H2O sit in DDAA environments,
3
3
DDA environment except for CITINA 2H2O, which con-
whereas in FERBAC 0.5H2O it sits in a DDA environment.
3
3
tains water molecules that lie in both DDA and DDAA
environments. Interestingly, in CITINA 2H2O the dehydra-
Tunnels of water molecules not hydrogen bonded to each
other are observed in CFANAM H2O, FERBAC 0.5H2O,
3
3
3
tion of the water molecules occurs in two steps, at tempera-
tures of 93 and 136 ꢀC. These water molecules exhibit a D2
water motif,40 which indicates that there are two water
molecules connected by hydrogen bonding in a discrete
chain. The packing indices calculated using Platon41 were
in the range 71.7-74.0. The dehydration of these cocrystal
and HCTINM H2O and packing indices range from 68.8
3
to 73.3.
In summary, there appears to be little or no correlation
between structure and thermal stability in the 11 new co-
crystal hydrates reported herein other than for tunnel hy-
drates, which exhibit a tendency to lose water at or around
the boiling point of water. Isolated water molecules that
form at least three H-bonds were observed to exhibit greater
thermal stability, and it might be anticipated that the pre-
sence of strong hydrogen bonding (short hydrogen bond
hydrates revealed new crystal forms for FERINM H2O,
3
FERNAM H2O, and CITINA 2H2O, whereas an isostruc-
3
3
tural anhydrate occurred following dehydration of ELAINM
1.7H2O.
3

2164 Crystal Growth & Design, Vol. 10, No. 5, 2010
Table 3. Structural and Thermal Stability Properties of Selected Cocrystal Hydratesa
Clarke et al.
Group 1: Thermal Loss of Water of Crystallization T < 100 ꢀC
energy of
dehydrationper
stoichiometric disordered thermal
ratio water loss
(T:CF:H2O) molecule of water cocrystal
melting
point of
water
environment
packing mole of water
index (J/g, kJ/mol)
result after
dehydration
cocrystal
type of hydrate
CITINA 2H2O
FERINM H2O
1:2:2
1:1:1
no
no
93
58
187
153
DDA, DDAA tunnels D2 motif
72.7
74.0
new crystal form
new crystal form
3
DDA
DDA
DDA
tunnels of water
molecules not H-bonded
tunnels of water
3
FERNAM H2O
1:1:1
1:1:2
no
98
127
300
71.7
new crystal form
new crystal form
3
molecules not H-bonded
channel
ELAINM H2O
3
yes
<100
Group 2: Thermal Loss of Water of Crystallization 100 > T < 120 ꢀC
energy of
dehydrationper
packing mole of water
stoichiometric disordered thermal melting
ratio water loss point of
(T:CF:H2O) molecule of water cocrystal environment
water
result after
dehydration
cocrystal
type of hydrate
index
(J/g, kJ/mol)
GALNAM
0.75H2O
GALINA H2O
1:1:1
1:1:1
1:1:2
no
no
yes
109
115
115
210
234
270
DDA
DDA
tunnels of water molecules 74.4
not H-bonded
92.58, 37.74
new crystal form
new crystal form
new crystal form
3
tunnels of water
molecules not H-bonded
73.3
87.25, 27.15
283.8, 54.81
3
GALTBR 2H2O
DDA, DDAA tunnels of water
molecules not H-bonded
isolated
3
PCANAC H2O
3
1:1:1
2:2:3
no
no
115
118
224
211
DDA
68.5
75.9
106.2, 31.35
236.2, 50.42
new crystal form
new crystal form
GALNAC 1.5H2O
3
DDAA, DDA isolated, tunnels of water
molecules not H-bonded
D2-motif
Group 3: Thermal Loss of Water of Crystallization T > 120 ꢀC
cocrystal
stoichiometric disordered thermal melting
loss of
water
water
environment
type of hydrate
packing
energy of
result after
dehydration
ratio
water
(T:CF:H2O) molecule
point of
cocrystal
index dehydrationper
mole of water
(J/g, kJ/mol)
BTNRES H2O
GALINM H2O
2:1:1
1:1:1
no
yes
∼120
188
203
DD
DDA
isolated
70.4
73.3
new crystal form
new crystal form
3
124
tunnels of water
molecules not H-bonded
isolated
3
GALCAF 0.5H2O
CITINA 2H2O
1:1:0.5
1:2:2
1:1:2
2:1:1
1:1:1
no
no
no
no
no
135
136
142
175
196
242
187
293
299
282
DDAA
74.7
72.7
73.1
75.6
74.2
69.7, 52.04
isostructural
new crystal form
3
DDA, DDAA tunnel D2-Motif
DDA, DDAA isolated D2 motif
3
QUETBR 2H2O
176.0, 45.62 isostructural
124.8, 64.20 new crystal form
new crystal form
3
ELACAF H2O
QUEINA H2O
DDA
DDA
isolated
tunnels of water
molecules not H-bonded
3
3
Group 4: Thermal Loss of Water of Crystallization Occurs at Melt of the Cocrystal
stoichiometric ratio disordered water
(T:CF:H2O)
thermal loss
of water
melting point
of cocrystal
water
environment
cocrystal
molecule
type of hydrate
packing index
71.7
CFANAM H2O
1:1:1
yes
occurs at melt
occurs at melt
occurs at melt
108
156
127
DDAA
DDA
tunnels of water molecules
not H-bonded
tunnels of water
3
FERBAC 0.5H2O
1:1:0.5
1:1:1
no
no
68.8
73.3
3
molecules not H-bonded
tunnels of water
molecules not H-bonded
HCTINM H2O
3
DDAA
a Energies of dehydration are only reported for compounds with a well-resolved endotherm in the DSC thermogram.
distances) would afford the high degree of thermal stability.
However, this does not seem to be the case since the same
environment from a structural perspective can exhibit a
dramatically different thermal stability. Indeed, even the
number of hydrogen bonds between the host molecules
and the water molecules is not necessarily a good predictor
of thermal stability. Likewise, packing indices show no
correlations with thermal stability. In addition, the enthalpy
of dehydration was calculated for the eight cocrystal hy-
drates that exhibited a well-defined DSC endotherm asso-
ciated with dehydration. The range of values is wide,
27.15-64.20 kJ/mol, and in general the enthalpy of dehy-
dration increases with thermal stability of the hydrate.
However, there is a great deal of variability from compound
to compound. That we are in a sense “comparing apples with
oranges”, that is, the cocrystal hydrates reported herein
exhibit variable hydrate stoichiometry, multiple hydrate
packing environments, and several possible outcomes of
dehydration (i.e., melting, isostructural anhydrate, anhy-
drate is a new crystal form), makes it hard to correlate
thermal stability of the cocrystal hydrate and its enthalpy
of dehydration.

Article
Crystal Growth & Design, Vol. 10, No. 5, 2010 2165
Chart 1. Supramolecular Heterosynthons Observed in Hydrates of Carboxylic Acids That Contain No Additional H-Bonding Groupsa
a The number given under each scheme indicates the times of occurrence.
Table 4. CSD Statistics of Polymorphs of Single and Multicomponent
(Cocrystals and Hydrates) Organic Structuresa
Chart 2. Supramolecular Heterosynthons Observed in Hydrates
of Phenols That Contain No Additional H-Bonding Groupsa
% of organic
structures
no. of entries
196575b
141350
organic crystal structures
single component molecular
organic structures
single component polymorphic
structures
100e
71.9e
2319c
1.6f
hydrates
15938
6351
65
8.1e
3.2e
1.0h
molecular organic hydratesi
polymorphic molecular organic
hydratesi
cocrystalsi
2415d
48
1.2e
1.9g
polymorphic cocrystalsi
a CSD May 09 update, 483 021 entries. b Structures containing Na, Li,
K, As were excluded. c Only one refcode was counted for each poly-
morphic compound. d Cocrystals sustained by strong hydrogen bonds.
e Percentages with respect to organic crystal structures only. f Percen-
tages with respect to 141 350 subtotals. g Percentages with respect to 2415
subtotals. h Percentages with respect to 6351 entries only. i Conquest
1.11, only organics, 3D coordinates known, R e 7.5, no ions.
amenable to crystal engineering studies. For example, car-
boxylic acids and phenols both form persistent supramolecular
synthons with pyridines37 even in the presence of competing H-
bond acceptors. However, as revealed by Charts 1-3, water
molecules are rather promiscuous in terms of their H-bond
interactions with carboxylic acids and phenols. These charts
summarize a CSD analysis of the supramolecular heterosyn-
thons that exist when at least one water molecule is present in a
crystal that contains a carboxylic acids or a phenol in the
absence of competing functional groups. There are just 23
hydrates of simple carboxylic acids and remarkably 10 supra-
molecular heterosynthons exist between the carboxylic acid
moieties and the water molecules (Chart 1). Supramolecular
heterosynthons between water molecules and phenols are also
diverse with 7 different supramolecular heterosynthons out of
a The number given under each schematic indicates the times of
occurrence in the CSD.
CSD Analysis. The persistence of supramolecular hetero-
synthons is a feature of cocrystals that makes them particularly

2166 Crystal Growth & Design, Vol. 10, No. 5, 2010
Chart 3. Supramolecular Heterosynthons Observed in Hydrates of Carboxylic Acids and Phenolsa
Clarke et al.
a The numbers given under each schematic indicate the times of occurrence in the CSD.
just 17 crystal structures (Chart 2). There are only 10 crystal
structures in which a carboxylic acid moiety, a phenol moiety
and a water molecule are involved in the absence of competing
H-bonding groups, and remarkably nine different supramole-
cular heterosynthons occur in this subset (Chart 3). These
structural observations beg the question of whether or not the
promiscuity in supramolecular synthons and polymorphs ex-
hibited is general for multicomponent crystals or an artifact of
the water molecules of hydration and their promiscuity. A
CSD analysis (Table 4) reveals that structurally characterized
polymorphs remain rare in single-component molecular crys-
tals (2319/141350, 2.0%) and in multicomponent crystals such
as hydrates (65/6351, 1.0%) and cocrystals (48/2415, 2.0%).
However, there are enough compounds for which 3D co-
ordinates are known on both polymorphic cocrystals (38
compounds) and polymorphic hydrates (44 compounds) to
address polymorphism in multicomponent crystals. Analysis
of the 38 polymorphic cocrystals reveals that in 35 poly-
morphic pairs polymorphism is the result of conformational
flexibility and/or structural changes in the overall packing; that
is, the hydrogen bonded supramolecular synthons are persis-
tent as had been previously suggested.38 The remaining three
cocrystals exhibit different supramolecular heterosynthons:
carbamazepine and saccharin, ref codes: UNEZAO and UN-
EZAO01; bis(4-hydroxybenzoic acid) and 2,3,5,6-tetramethyl-
pyrazine, ref codes: ODOBIT and ODOBIT01; urea and
barbituric acid, ref codes: EFOZAB, EFOZAB 01 and EFO-
ZAB02. However, this is not the case with polymorphic
hydrates: 28 out of 44 polymorphic hydrates for which the
3D coordinates are known exhibit different supramolecular
synthons with respect to the water molecules.
In conclusion, there are numerous structural features
that might influence the thermal stability of organic crystal-
line hydrates, and the promiscuity of water molecules in

Article
Crystal Growth & Design, Vol. 10, No. 5, 2010 2167
terms of supramolecular synthons mean that their composi-
tion, structure, and thermal stability remain largely unpre-
dictable with the exception that channel hydrates tend to
exhibit low thermal stability. It would therefore be appro-
priate to assert that water is indeed in many ways the nemesis
of crystal engineering.
(20) Henck, J. O.; Griesser, U. J.; Burger, A. Pharm. Ind. 1997, 59, 165–
169.
(21) (a) Allen, F. H.; Kennard, O. Chem. Des. Automat. News 1993, 8,
31–37. (b) Allen, F. H. Acta Crystallogr. 2002, B58, 380–388.
ꢀ
(22) Infantes, L.; Fabian, L; Motherwell, W. D. S. CrystEngComm
ꢀ
2007, 9, 65–71.
(23) (a) Khankari, R. K.; Grant, D. J. Thermochim. Acta 1995, 248
61–79. (b) Morris, K. R. Polymorphism in Pharmaceutical Solids;
Brittain, H. G., Ed.; Marcel Dekker, Inc.: New York, 1999; Vol. 95,
pp 125-181.
Supporting Information Available: X-ray crystallographic infor-
mation in CIF format, IR, DSC, TGA, PXRD are available free of
charge via the Internet at http://pubs.acs.org.
(24) (a) Infantes, L.; Chisholm, J.; Motherwell, S. CrystEngComm 2003,
5, 480–486. (b) Hickey, M. B.; Peterson, M. L.; Manas, E. S.; Alvarez,
€
J.; Haeffner, F.; Almarsson, O. J. Pharm. Sci. 2007, 96, 1090–1099.
References
(c) Kiang, Y.-H.; Xu, W.; Stephens, P. W.; Ball, R. G.; Yasuda, N. Cryst.
Growth Des. 2009, 9, 1833–1843. (d) Bak, A.; Cooke, M.; Kleppe, M.
Am. Pharm Rev. Sep/Oct 2008.
(1) (a) Pepinsky, R. Phys. Rev. 1955, 100, 971. (b) Schmidt, G. M. J. Pure
Appl. Chem. 1971, 27, 647–678. (c) Desiraju, G. R. Crystal Engineer-
ing: The Design of Organic Solids; Elsevier: Amsterdam, 1989. (d)
Etter, M. C. J. Am. Chem. Soc. 1982, 104, 1095–1096.
(2) (a) Moulton, B.; Zaworotko, M. J. Chem. Rev. 2001, 101, 1629–
1658. (b) Zaworotko, M. J. Cryst. Growth Des. 2007, 7, 4–9.
(3) Vippagunta, S. R.; Brittain, H. G.; Grant, D. J. W. Adv. Drug
Delivery Rev. 2001, 48, 3–26.
(25) (a) Falk, M.; Knop, O. Water: A Comprehensive Treatise; Franks,
F., Ed.; Plenum Press: New York, 1973; Vol. 2, pp 55-113. (b)
Authelin, J. R. Int. J. Pharm. 2005, 303, 37–53.
(26) Morris K. R.; Rodriguez-Hornedo, N. Encyclopedia of Pharma-
ceutical Technology; Swarbrick, J., Boylan, J., Eds; Marcel Dekker,
Inc: New York, 1993; Vol. 7, pp 393-440.
(4) Noyes, A. A.; Whitney, R. R. J. Am. Chem. Soc. 1897, 19, 930–934.
(5) Amidon, G. L.; Lennernas, H.; Shah, V. P.; Crison, J. R. Pharm.
Res. 1995, 12, 413–420.
(6) (a) Vishweshwar, P.; McMahon, J. A.; Bis, J. A.; Zaworotko, M. J.
J. Pharm. Sci. 2006, 95, 499–516. (b) Special issue on Pharmaceutical
Cocrystals Mol. Pharmaceutics 2007, 4, 299-486. (c) Shan, N.;
Zaworotko, M. J. Drug. Discov. Today 2008, 13, 440–446. (d)
Meanwell, N. A. Annu. Rep. Med. Chem. 2008, 43, 373–404. (e)
Schultheiss, N; Newman, A. Cryst. Growth Des. 2009, 9, 2950–2967.
(f) Stanton, M. K.; Bak, A. Cryst. Growth Des. 2008, 8, 3856–3862. (g)
(27) Byrn, S. R.; Lin, C. T. J. Am. Chem. Soc. 1976, 98, 4004.
(28) (a) Reutzel-Edens, S. M.; Bush, J. K.; Magee, P. A.; Stephenson,
G. A.; Byrn, S. R. Cryst. Growth Des. 2003, 3, 897–907. (b) Britain,
H. G.; Morris, K. G.; Boerrigter, S. X. M. Polymorphism in Pharma-
ceutical Solids; Brittain, H. G., Ed.; Informa Healthcare, Inc.:
New York, 2009; Vol. 192, pp 233-279.
(29) SMART-V5.625. Data Collection Software; Bruker-AXS: Madison,
Wisconsin, USA, 2001.
(30) APEX2 (Version 2008.1-0); Bruker: Madison, Wisconsin, USA,
2008.
€
Stahly, G. P. Cryst. Growth Des. 2009, 9, 4212–4229. (h) Aakeroy,
C. B.; Desper, J.; Fasulo, M.; Hussain, I.; Levin, B.; Schultheiss, N.
CrystEngComm 2008, 10, 1816–182.
(31) SAINT-V7.51A, Data Reduction Software, Bruker-AXS: Madison,
Wisconsin, USA, 2008.
(32) Sheldrick, G. M. SADABS. Program for Empirical Absorption
(7) Childs, S. L.; Chyall, L. J.; Dunlap, J. T.; Smolenskaya, V. N.; Stahly,
B. C.; Stahly, G. P. J. Am. Chem. Soc. 2004, 126, 13335–13342.
(8) McNamara, D. P.; Childs, S. L.; Giordano, J.; Iarriccio, A.;
Cassidy, J.; Shet, M. S.; Mannion, R.; O’Donnell, E.; Park, A.
Pharm. Res. 2006, 23, 1888–1897.
(9) Hickey, M. B.; Peterson, M. L.; Scoppettuolo, L. A.; Morrisette,
S. L.; Vetter, A.; Guzman, H.; Remenar, J. F.; Zhang, Z.; Tawa,
M. D.; Haley, S.; Zaworotko, M. J.; Almarsson, O. Eur. J. Pharm.
Biopharm. 2007, 67, 112.
Correction; University of Gottingen: Germany, 1996.
(33) Sheldrick, G. M. SHELXTL, v. 6.10; Bruker-AXS Madison:
Wisconsin, USA, 2000.
(34) (a) Farrugia, L. J. Appl. Crystallogr. 1999, 32, 837–838. (b)
Sheldrick, G. M. Acta Crystallogr. 2008, A64, 112–122. (c) Sheldrick,
G. M. Acta Crystallogr. 1990, A46, 467–473.
(35) (a) Fleischman, S. G.; Kuduva, S. S.; McMahon, J. A.; Moulton,
B.; Walsh, R. B.; Rodriguez-Hornedo, N.; Zaworotko, M. J. Cryst.
€
Growth Des. 2003, 3, 909–919. (b) Aakeroy, C. B.; Beatty, A. M.;
(10) Walsh, R. D. B.; Bradner, M. W.; Fleischman, S.; Morales, L. A.;
´
Moulton, B.; Rodrıguez-Hornedo, N.; Zaworotko, M. J. Chem.
Helfrich, B. A. Angew. Chem., Int. Ed. 2001, 40, 3240. (c)
Reddy, L. S.; Nangia, A.; Lynch, V. M. Cryst. Growth Des. 2004, 4,
Commun. 2003, 186.
€
89–94. (d) Aakeroy, C. B.; Desper, J.; Scott, B. M. T. Chem.
(11) (a) Almarsson, O.; Zaworotko, M. J. Chem. Commun. 2004, 1889–
1896. (b) McNamara, D. P.; Childs, S. L.; Giordano, J.; Iarriccio, A.;
Cassidy, J.; Shet, M. S.; Mannion, R.; O'Donnel, E.; Park, A. Pharm.
Res. 2006, 23, 1888–1897. (c) Variankaval, N.; Wenslow, R.; Murry, J.;
Hartman, R.; Helmy, R.; Kwong, E.; Clas, S. D.; Dalton, C.; Santos, I.
Cryst. Growth Des. 2006, 6, 690–100.
Commun. 2006, 13, 1445. (e) Vishweshwar, P.; Nangia, A.; Lynch,
V. M. Cryst. Growth Des. 2003, 3, 783. (f) Videnova-Adrabinska, V.;
Etter, M. C. J. Chem. Crystallogr. 1995, 25, 823–829. (g) Leiserowitz,
L.; Nader, F. Acta Crystallogr., Sect. B: Struct. Sci. 1977, 33, 2719–
2733.
(36) (a) Bis, J. A.; Vishweshwar, P.; Weyna, D.; Zaworotko, M. J. Mol.
Pharmaceutics 2007, 4, 401–416. (b) Vishweshwar, P.; Nangia, A.;
Lynch, V. M. CrystEngComm 2003, 5, 164–168. (c) Papaefstathiou,
G. S.; MacGillivray, L. R. Org. Lett. 2001, 3, 3835–3838.
(37) (a) McMahon, J. A.; Bis, J. A.; Vishweshwar, P.; Shattock, T. R.;
McLaughlin, O. L.; Zaworotko, M. J. Z. Kristallogr. 2005, 220,
340–350. (b) Jagadeesh Babu, N.; Sreenivas Reddy, L.; Nangia, A.
Mol. Pharmaceutics 2007, 4, 417–434. (c) Leiserowitz, L.; Hagler,
A. T. Proc. R. Soc. London Series A 1983, 388, 133–175.
(12) Gillon, A. L.; Feeder, N.; Davey, R. J.; Storey, R. Cryst. Growth
Des. 2003, 3, 663–673.
(13) Desiraju, R. G. J. Chem. Soc., Chem. Commun. 1991, 426–428.
(14) (a) Stahl, H. P. In Towards Better Safety of Drugs and Pharmaceu-
tical Products; Braimar, D. D., Ed.; Biomedical Press: Elsevier North
Holland, 1980; pp 265-280. (b) Jeffrey, G. A. An Introduction to
Hydrogen Bonding; Oxford University Press: New York, 1997.
(15) Bettinetti, G.; Mura, P.; Sorrenti, M.; Faucci, M. T.; Negri, A.
J. Pharm. Sci. 1999, 88, 1133–1139.
(16) Stephenson, G. A.; Groleau, E. G.; Kleeman, R. L.; Xu, W.;
Rigsbee, D. R. J. Pharm. Sci. 1998, 87, 536–542.
(17) Khankari, R. K.; Law, D.; Grant, D. J. W. Int. J. Pharm. 1992, 82,
117–127.
(18) Chen, L. R.; Young, V. G., Jr.; Lechuga-Ballesteros, D.; Grant,
(38) (a) Shattock, T. R.; Arora, K. K.; Vishweshwar, P.; Zaworotko,
€
M. J. Cryst. Growth Des. 2008, 8, 4533–4545. (b) Aakeroy, C. B.;
Beatty, A. M.; Helfrich, B. A. J. Am. Chem. Soc. 2002, 124, 14425. (c)
Vishweshwar, P.; Nangia, A.; Lynch, V. M. J. Org. Chem. 2002, 67,
556–565.
(39) Kavuru, P.; Aborayes, D.; Arora, K. K.; Clarke, H. D.; Kennedy,
A.; Marshall, L.; Ong, T. T.; Pujari, T.; Wojtas, L.; Zaworotko,
M. J.; Perman, J. A.; Cryst. Growth Des. submitted for publication.
(40) Infantes, L.; Motherwell, S. CrystEngComm 2002, 4 (75), 454–461.
(41) Spek, A. L. PLATON, Molecular Geometry Program; University of
Utrecht: The Netherlands, May, 1997.
D. J. W. J. Pharm. Sci. 1999, 88, 1191–1200.
(19) (a) Apperly, D. C.; Basford, P. A.; Dallman, C. I.; Harris, R. K.;
Kinns, M.; Marshall, P. V.; Swanson, A. G. J. Pharm. Sci. 2005, 94,
516–523. (b) Zegarac, M.; Mestrovic, E.; Dumbovic, A.; Devcic, M.;
Tudja, P. WO 2007/080362 A1.