See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/266852027

Preparation and Solid-State Characterization of Dapsone Drug−Drug Co-

Crystals

Article  in  Crystal Growth & Design · August 2014

DOI: 10.1021/cg500668a

CITATIONS READS

68 1,615

6 authors, including:

Jiang Linglei Ying Huang

CNBG-Virogin Chinese Academy of Sciences
16 PUBLICATIONS   405 CITATIONS    215 PUBLICATIONS   5,687 CITATIONS   

SEE PROFILE SEE PROFILE

Qi Zhang Yun Xu
Shanghai institute of Materia Medica Nanjing University
35 PUBLICATIONS   618 CITATIONS    533 PUBLICATIONS   8,134 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Activation and Identification of NC-1: A Cryptic Cyclodepsipeptide from Red Soil-Derived Streptomyces sp. FXJ1.172 View project

outer membrane vesicles (OMVs) View project

All content following this page was uploaded by Jiang Linglei on 17 December 2017.

The user has requested enhancement of the downloaded file.

 Article

pubs.acs.org/crystal

Preparation and Solid-State Characterization of Dapsone Drug−Drug

Co-Crystals
Linglei Jiang, Ying Huang, Qi Zhang, Hongyan He, Yun Xu, and Xuefeng Mei*
Pharmaceutical Analytical & Solid-State Chemistry Research Center, Shanghai Institute of Materia Medica, Chinese Academy of
Sciences, Shanghai 201203, China
*
S Supporting Information

ABSTRACT: Pharmaceutical co-crystals involving two active pharma-

ceutical ingredients are rarely revealed in the literature. In this work,
crystal engineering principles were exercised to guide the design and
synthesis of the Biopharmaceutics Classification System class IV drug
dapsone (DAP). We reported six drug−drug co-crystals of DAP with
sulfanilamide, flavone, luteolin, caffeine in 1:1 stoichiometry, caffeine in
1:2 stoichiometry, and 2(3H)-benzothiazolone. Bioactive coformers
were deliberately selected. The resulting co-crystals were fully
characterized by a range of analytical technologies, including X-ray
powder diffraction, Fourier transform infrared spectroscopy, polarized
light microscopy, differential scanning calorimetry, and thermogravi-
metric analysis, etc. Single-crystal structure analysis reveals that
reoccurring supramolecular synthons are observed in different DAP
co-crystals. Equilibrium solubility and intrinsic dissolution rates were
also compared with those of the parent drug. This work expands the pharmaceutically acceptable solid forms of DAP and
supplements the successful cases of drug−drug co-crystals.

■ INTRODUCTION
Dapsone (DAP) was originally introduced as a sulfur-
rate, stability toward thermal and humid stress, and
compressibility aspects.14−29 As a subset of co-crystals, drug−
containing compound for antibacterial investigation in 1936, drug co-crystals may not only lead to the modification of the
and the drug was studied for the treatment of a totally different physicochemical properties, but also present new opportunities
array of diseases during subsequent decades.1 Currently, it is for the development of combinational therapies. Drug−drug
employed to treat tuberculosis, leprosy, malaria, Kaposi’s co-crystals fulfill the criteria for patent eligibility: novelty, utility,
sarcoma, dermatoses, and AIDS-related pneumonia.1−4 The and nonobviousness.30 Consequently, there has been a growing
story of reinvention of DAP in the clinic has aroused general interest in the study of drug−drug co-crystals reflected in the
interest in drug development history. DAP is classified as a class increasing number of publications and patent applications in
IV compound according to the Biopharmaceutics Classification recent years.31−36
System (BCS) in the literature.5,6 The drug is commercially The pure dapsone drug can exist as an anhydrate, in hydrated
available as an oral tablet and a topical gel. It is slowly absorbed form, and in three solvated forms, including dichloromethane,
with 70−80% bioavailability in an acidic environment.7 Despite 1,4-dioxane, and tetrahydrofuran solvates. Crystal structures of
its therapeutic potential, the low solubility (0.16 mg/mL in the anhydrate and the stable hydrate forms are reported.3,37,38
water) was found to result in a low therapeutic index and high On the basis of a close analysis of all the related single-crystal
microbial resistance.8−10 To overcome the solubility limitation, data of DAP available in the literature, intermolecular hydrogen
several entities such as amino acid derivatives, inclusion bonds between the aromatic amino groups and the oxygen
compounds with β-cyclodextrin, and nanoemulsion formula- atoms of the sulfonyl moieties via synthon I (N−H···OS) are
tions have been developed.8,9,11 In addition, co-crystals of DAP frequently observed in the structures (Figure 1). As the
with 3,5-dinitrobenzoic acid, 1,3,5-trinitrobenzene, 5-nitro- structure of DAP contains two aromatic amino groups, we
isophthalic acid, ε-caprolactam, and 4,4-bipyridine were also envisioned that any compound containing sulfonyl groups
reported.3,4,12,13 Although different DAP co-crystals were could potentially form additional H-bonding with the
synthesized, none of the coformers are pharmaceutically unsatisfied amino groups of DAP molecules. Furthermore,
acceptable and hence cannot be applied directly in the drug similar H-bonding interactions between aromatic amino groups
development.
Over the past decade, pharmaceutical co-crystal screening has Received: May 7, 2014
rapidly developed into a general method for modifying drugs’ Revised: July 25, 2014
physicochemical properties, including the solubility, dissolution Published: August 1, 2014

© 2014 American Chemical Society 4562 dx.doi.org/10.1021/cg500668a | Cryst. Growth Des. 2014, 14, 4562−4573
Crystal Growth & Design Article

Figure 1. Hydrogen-bonding interactions observed in the DAP hydrate form.

Table 1. CSD Statistics of Possible Supramolecular Synthons in DAP with Aromatic Amine Functionalities

Scheme 1. Chemical Structures of DAP and Potential Coformersa

a
Successful selected coformers are labeled in green, and unsuccessful ones are labeled in red.

4563 dx.doi.org/10.1021/cg500668a | Cryst. Growth Des. 2014, 14, 4562−4573

Crystal Growth & Design Article

and carbonyl moieties, as depicted in synthon II (N−H···O Preparation of Co-Crystals. Slow-evaporative crystallization
C), could also be utilized as an alternative to the sulfonyl experiments were carried out by dissolving each component in 2 mL
groups in designing DAP co-crystals. A survey of the two of solvent according to Table 2. Crystals suitable for physicochemical
synthons in the Cambridge Structural Database (CSD) was
performed (CSD version 5.35, February 2014). Separate Table 2. Summary of Components in the Crystallization
searches were done for synthons I and II. There are 2031 Experiments
structures containing primary aromatic amino groups. Out of APIa amt coformer amt
those entries, 262 (12.9%) hits form synthon I, suggesting that co-crystal (mg, mmol) (mg, mmol) solvent
the aromatic amino group is likely to form supramolecular DAP·SUL 24.8, 0.1 34.4, 0.2 ethanol−acetone
interactions with sulfonyl groups. There are 376 (18.5%) hits (v/v, 1/1)
containing synthon II, which implies that the interaction DAP·FLA 24.8, 0.1 22.2, 0.1 ethanol−acetone
between the aromatic amino group and carbonyl groups is also (v/v, 1/1)
frequent (Table 1). On the basis of the synthon design DAP·LUT 24.8, 0.1 28.6, 0.1 ethanol−acetone
(v/v, 1/1)
approach, a series of bioactive compounds with either sulfonyl DAP·CAF-1 24.8, 0.1 19.4, 0.1 ethyl acetate
or carbonyl moieties were selected as coformers (Scheme 1). DAP·CAF-2 24.8, 0.1 19.4, 0.1 ethyl acetate−methanol
Interestingly, co-crystallization screening experiments resulted (v/v, 1/1)
in six successful co-crystals with single-crystal structures solved. DAP·HBZ 24.8, 0.1 15.1, 0.1 ethanol−acetone
Unlike the previously reported DAP co-crystals, in this work, (v/v, 1/1)
a
six DAP co-crystals were synthesized with biologically active Active pharmaceutical ingredient.
ingredients as coformers. Sulfanilamide is a chemical analogue
of DAP and exhibits potent antibacterial activity by a characterization were harvested after slow evaporation of the
mechanism similar to that of DAP.39 Flavonoids are generally corresponding solvent in a desiccator. Particularly, single crystals of
recognized as active pharmaceutical ingredients with health- DAP·SUL and DAP·FLA were obtained after 5 days of allowing the
prolonging effects attributed to their antibacterial, antioxidant, solvent to evaporate slowly. Single crystals of DAP·CAF-1 and DAP·
HBZ were harvested after 1 day. However, it took about 1 month to
antitumor, and anti-inflammatory properties.40−42 Caffeine and
obtain single crystals of DAP·LUT with good quality.
its xanthine analogues theophylline and theobromine have been Thermogravimetric Analysis. TGA was carried out on a Netzsch
approved by the FDA as “generally regarded as safe” (GRAS) TG209 F3 instrument using N2 as dry air with a flow rate of 20 mL/
agents that also have activity as anti-inflammatories and min and a scan rate of 10 °C/min.
lipolytics.43 Hence, the development of dapsone−sulfa drug, Differential Scanning Calorimetry. DSC was performed with a
dapsone−flavonoid, and dapsone−caffeine drug−drug co- PerkinElmer DSC 8500 instrument. Samples weighing 3−5 mg were
crystals may open the field for antibacterial combination, and heated in standard aluminum pans at scan rates from 10 to 200 °C/
they also possess the potential to be novel antioxidants and min under a nitrogen gas flow of 20 mL/min. Two-point calibration
anti-inflammatory drugs. 2-Hydroxybenzothizole is well docu- using indium and tin was carried out to check the temperature axis and
mented to have anticonvulsant activity, and a large chemical heat flow of the equipment.
series containing 2(3H)-benzothiazolinone moieties has been Single-Crystal X-ray Diffraction. Single-crystal X-ray diffraction
described for interactions with a number of molecular targets (SCXRD) measurements were conducted on a Bruker Smart Apex II
diffractometer using Mo Kα radiation (λ = 0.71073 Å) with a graphite
such as membrane and nuclear receptors.44−46 The dapsone−
monochromator. Integration and scaling of intensity data were
2(3H)-benzothiazolinone co-crystal may find applications accomplished using the SAINT program. The structures were solved
toward developing combinational anticonvulsant drugs. In by direct methods using SHELXS97, and refinement was carried out
this study, the single-crystal structures of all six co-crystals with the full-matrix least-squares technique using SHELXL97.47The
were revealed. The physicochemical properties were fully hydrogen atoms were refined isotropically, and the heavy atoms were
investigated by X-ray powder diffraction (XRPD), Fourier refined anisotropically. N−H and O−H hydrogens were located from
transform infrared (FT-IR) spectroscopy, polarized light difference electron density maps, and C−H hydrogens were placed in
microscopy (PLM), differential scanning calorimetry (DSC), calculated positions and refined with a riding model. The data were
and thermogravimetric analysis (TGA). Equilibrium solubility corrected for the effects of absorption using SADABS. Crystallographic
and intrinsic dissolution rates were compared with those of the data in CIF format have been deposited at the Cambridge
parent drug as well. This work is the first to report co-crystals of Crystallographic Data Center, CCDC Nos. 995302, 995303, 995304,
DAP with pharmaceutically acceptable compounds and supple- 995306, and 995307 for DAP·SUL, DAP·FLA, DAP·LUT, DAP·CAF-
1, DAP·CAF-2, and DAP·HBZ, respectively. The crystallographic data
ments the successful examples of the design and synthesis of and refinement details are summarized in Table 3.
drug−drug co-crystals.

■
X-ray Powder Diffraction. XRPD was measured on a Bruker D8-
ADVANCE X-ray diffractometer using a Cu Kα radiation (λ = 1.5418
EXPERIMENTAL SECTION Å). The voltage and current were 40 kV and 40 mA, respectively.
Materials. DAP was purchased from Shanghai Demo Biological Samples were measured in reflection mode in the 2θ range of 3−40°
Technological Co. Ltd. Sulfanilamide (SUL), sulfadiazine (SD), and with a scan speed of 15 deg/min (step size 0.025°, step time 0.1 s)
2(3H)-benzothiazolone (HBZ) were purchased from J&K Co. Ltd. using a LynxEye detector. All data were acquired at ambient
Fisetin (FIS), luteolin (LUT), caffeine (CAF), theophylline (THP), temperature (25 °C). The data were imaged and integrated with
and theobromine (THB) were purchased from Aladdin Co. Ltd. RINT Rapid and peak-analyzed with Origin 8.1 software. Calibration
Sulfamerazine (SFM) and flavone (FLA) were purchased from Alfa of the instrument was performed using the Corindon (Bruker AXS
Aesar Co. Ltd. Sulfadimidine (SFD) and genistein (GEN) were Korundprobe) standard.
purchased from TCI Shanghai Co. Ltd. with greater than 99% purity. Polarized Light Microscopy. All PLM examinations were
All analytical grade solvents were purchased from Sinopharm performed on an XPV-400E polarizing microscope and an XPH-300
Chemical Reagent Co. and used without further purification. XRPD hot stage coupled with a JVC TK-C9201 EC digital video recorder
analysis of the commercially available DAP confirmed that it is in the (Shanghai Changfang Optical Instrument Co. Ltd.). Selective PLM
anhydrous form (CSD ref code DAPSUO10).37 photos for DAP co-crystals are presented in Figure 2.

4564 dx.doi.org/10.1021/cg500668a | Cryst. Growth Des. 2014, 14, 4562−4573

Crystal Growth & Design Article

Table 3. Crystallographic Data for the Co-Crystals

DAP·SUL DAP·FLA DAP·LUT DAP·CAF-1 DAP·CAF-2 DAP·HBZ
formula C12H12N2O2S, C12H12N2O2S, C12H12N2O2S, C12H12N2O2S, C12H12N2O2S, C12H12N2O2S,
2(C6H8N2O2S) C15H10O2 C15H10O6, C2H6O C8H10N4O2 2(C8H10N4O2) C7H5NOS
cryst syst monoclinic monoclinic monoclinic orthorhombic orthorhombic monoclinic
space group C2/c P21/n P21/c Pbca Pna21 P21/n
temp (K) 296(2) 296(2) 296(2) 296(2) 296(2) 296(2)
a (Å) 13.2304(2) 11.408(2) 22.293(2) 8.5613(6) 17.231(7) 5.8452(4)
b (Å) 30.2578(5) 11.644(3) 7.7975(7) 15.6676(9) 16.772(7) 14.3835(9)
c (Å) 13.6910(3) 18.448(4) 16.0553(14) 32.2872(19) 20.727(8) 21.7712(13)
α (deg) 90 90 90 90 90 90
β (deg) 102.448(1) 101.324(14) 91.814(5) 90 90 92.205(3)
γ (deg) 90 90 90 90 90 90
cell vol (Å3) 5351.98(17) 2402.7(9) 2789.5(4) 4330.8(5) 5990(4) 1829.1(2)
calcd density 1.471 1.301 1.383 1.357 1.412 1.451
(g/cm3)
Z 8 4 4 8 8 4
no. of independent 4283 2600 2744 3520 7135 3657
reflns
S 0.944 0.935 0.964 1.212 0.969 1.262
R1 0.0405 0.0404 0.0694 0.0588 0.0724 0.0343
Rint 0.0272 0.0311 0.0922 0.0290 0.0801 0.0188
Rw 0.1448 0.1461 0.2169 0.1923 0.2028 0.1477

Figure 2. PLM photographs of DAP co-crystals: (a) DAP·SUL, (b) DAP·FLA, (c) DAP·LUT, (d) DAP·CAF-1, (e) DAP·CAF-2, (f) DAP·HBZ.

Fourier Transform Infrared Spectroscopy. FT-IR spectra were containing a mixture of acetonitrile and 15 mmol/L K2HPO4−H3PO4
collected with a Nicolet-Magna FT-IR 750 spectrometer in the range pH 6.0 buffer. Typically, the gradient started at 5% acetonitrile and
from 4000 to 350 cm−1 with a resolution of 4 cm−1 at ambient 95% K2HPO4−H3PO4 pH 6.0 buffer, which was maintained for 2 min,
conditions. and was changed to 50% acetonitrile and 50% K2HPO4−H3PO4 pH
Solubility and Dissolution Experiments. Equilibrium solubility 6.0 buffer in the following 9 min, which was maintained for 6 min.
was determined by suspending an excess amount of each sample in 1.5 After each gradient analysis, the columns were re-equilibrated with the
mL of distilled water (with 1% Tween 80), pH 2.0 glycine− initial mobile phase (i.e., 5% acetonitrile). The observed retention time
hydrochloric acid buffer, pH 4.6 NaH2PO4−citric acid buffer, and pH points for DAP, SUL, CAF, LUT, FLA, and HBZ are 9.7, 3.4, 7.2, 10.9,
6.8 NaH2PO4−citric acid buffer at 37 °C in a Dragon LAB MS-H-Pro+ 16.0, and 10.9 min, respectively. No overlap between peaks for DAP or
magnetic stirrer for 24 h. Upon equilibration, the solutions were any coformers was observed. Intrinsic dissolution rate (IDR)
filtered through 0.45 μm syringe filters, and the concentrations of the experiments were conducted using a Mini-Bath dissolution apparatus
solutions were determined using an Agilent 1260 series Infinite HPLC equipped with a Julabo ED-5 heater/circulator. In each experiment,
instrument. In all the experiments, the HPLC instrument was approximately 3 mg of sample was compressed into a 0.07 cm2 disk in
equipped with a ZORBAX ECLIPSE-C18 column (4.6 × 150 mm, a rotating disk intrinsic dissolution die using a Mini-IDR press at a
5 μm). An injection volume of 10 μL was used with an eluent flow rate pressure of 40 bar for 1 min. Only one side of the disk was exposed to
of 1 mL/min. The detection wavelength in the UV−vis range was set the dissolution medium. The intrinsic attachment was placed in a jar of
at 270 nm. The samples were gradient eluted with a mobile phase 15 mL of distilled water containing 1% Tween 80 preheated at 37 °C

4565 dx.doi.org/10.1021/cg500668a | Cryst. Growth Des. 2014, 14, 4562−4573

Crystal Growth & Design Article

Figure 3. (a) Major hydrogen-bonding interactions and (b) 3D architecture of DAP·SUL.

and stirred at 75 rpm. At regular time intervals, 200 μL of the sample molecular H-bonding interactions between the sulfonyl and
was withdrawn manually. The collected samples were diluted and amino groups of DAP (synthon I) would be expected in all the
assayed for DAP concentration using the HPLC method described structures. However, by analyzing the single-crystal structures
above.
of DAP, we found that only part of the amino hydrogen atoms

■ RESULTS AND DISCUSSION

Single-Crystal Structure Analysis. The dapsone molecule
and sulfonyl oxygen atoms participated in the H-bonding
formation. We envisioned that if other appropriate H-bonding
donor/acceptor moieties are present, the hydrogen-bonding
possesses a considerable potential for intermolecular inter- formation capability for the amino group and sulfonyl groups
actions with coformer molecules via −NH2 and −SO2− could be more satisfied. This opens new opportunities to
functional groups. Our strategy is to use these two important accommodate additional H-bonding guest molecules, such as
functional groups as interaction sites in the design and synthesis sulfamide, flavonoids, and alkaloids. A variety of coformers were
of multicomponent DAP co-crystals. It should be noted that the selected for a co-crystallization screening experiment. Prefer-
sulfonyl group is an ideal H-bonding acceptor to form H- ences were given to those coformers that have potential to form
bonding interactions with amino groups (Table 1). Intra- H-bonding interactions with amino moieties (Scheme 1). Out
4566 dx.doi.org/10.1021/cg500668a | Cryst. Growth Des. 2014, 14, 4562−4573
Crystal Growth & Design Article

Figure 4. (a) Tetramer unit constructed by DAP and FLA. (b) 2D structure view along the a axis pf DAP·FLA. (c) Sheet structure along the b axis of
DAP·FLA. Only the carbonyl oxygen atoms in FLA are depicted for clarity in (b) and (c).

of 12 potential coformers, 5 of them were found to form co- multiple H-bonding interactions between DAP and SUL result
crystals with DAP. The crystallographic data for co-crystals are in the 3D packing pattern depicted in Figure 3b.
summarized in Table 3, and the hydrogen bond information is Interestingly, although different crystallization methods were
listed in Table S1 (Supporting Information). employed, attempts to synthesize DAP co-crystals with other
DAP·SUL (1:2) Co-Crystal. As a structural analogue to sulfanilamide derivatives, such as SD), SFM, or SFD, were
DAP, sulfanilamide contains both sulfonamide and aromatic unsuccessful (Table S2, Supporting Information). The
amino functionalities. The DAP·SUL co-crystal was designed sulfonamide−sulfonamide interactions were found to be a key
on the basis of the analogue−analogue co-crystal formation factor in the formation of the DAP·SUL co-crystal. If one
principle. DAP and SUL were crystallized in 1:2 stoichiometry, hydrogen atom on the amino group of the coformers is
and the single-crystal structure was solved and refined in the replaced with other bulky substitutions, the one-dimensional
monoclinic system with the C2/c space group and one DAP sulfonamide H-bonding chain structure cannot be formed due
molecule and two SUL molecules in the asymmetric unit. to the steric hindrance, and hence, the co-crystal structure
Hydrogen bonding in the crystal structure of DAP·SUL is very cannot be built.
complex. There are multiple potential H-bonding donors/ DAP·FLA Co-Crystal. Encouraged by the successful result
acceptors in both DAP and SUL molecules. Each of these for the DAP·SUL example, we then extended our studies to
groups can participate in the formation of at least two hydrogen compounds containing a carbonyl moiety. DAP and FLA were
bonds. Major H-bonding interactions can be identified in the crystallized in 1:1 stoichiometry. The crystal structure was
complementary sulfonamide−sulfonamide interactions of SUL solved and refined in the monoclinic system with the P21/n
to form a one-dimensional chain structure (Figure 3a). space group and one DAP molecule and one FLA molecule in
Specifically, the sulfonamide group of SUL was connected the asymmetric unit. As expected, the carbonyl groups of FLA
with the other two SUL molecules by intermolecular N5−H27··· formed H-bonding interactions with the amino groups of DAP
O3 (2.23 Å) and N5−H28···O4 (2.49 Å) hydrogen bonds to give to form a tetrameric aggregation as depicted in Figure 4a. Two
a one-dimensional chain structure along the a axis. The amino host molecules and two guest molecules are linked by hydrogen
group of DAP was also found to form a H-bonding interaction bonding between the carbonyl and the amine groups via
with the sulfonamide oxygen from SUL via synthon I (N6− synthon II (N2−H10B···O4, 2.11 Å; N1−H4A···O4, 2.08 Å) to
H15···O1, 2.54 Å). The sulfonyl group of DAP, on the other give a tetrameric structure. Along the b axis, the host molecules
hand, formed a H-bonding interaction with the aromatic amino are also connected by hydrogen bonding via synthon I (N2−
group of SUL also via synthon I (N1−H4B···O4, 2.20 Å). The H10A···O1, 2.16 Å) to form a one-dimensional chain (Figure
4567 dx.doi.org/10.1021/cg500668a | Cryst. Growth Des. 2014, 14, 4562−4573
Crystal Growth & Design Article

4b). These chains are further linked by hydrogen bonding via crystallization conditions. Interestingly, the inclusion of ethanol
synthon II depicted previously to form a two-dimensional solvent was also found to be vital in stabilizing the co-crystal.
network structure. As the amino group can serve as both a H- When the co-crystal sample was subjected to thermal stress,
bonding donor and a H-bonding acceptor, the bc layers were ethanol molecules could be released at 170 °C under reduced
further extended through N1−H4B···N2 (2.43 Å) hydrogen pressure. However, the resulting solid was found to change
bonding along the a axis to generate its 3D structure (Figure back to a physical mixture of DAP and LUT after the ethanol
4c). molecules were removed from the crystal lattice (Figure S1,
DAP·LUT Co-Crystal. The dapsone−luteolin co-crystal was Supporting Information). Notably, attempts to grow DAP co-
crystallized in the monoclinic system with the P21/c space crystals with structural analogues of LUT, such as FIS and
group. The asymmetric unit contains one dapsone, one luteolin, GEN, were unsuccessful under similar crystallization conditions
and an ethanol molecule (Figure 5a). As LUT is also a (Table S2, Supporting Information). These findings further
confirmed that two phenolic hydroxyl groups (−O17H17 and
−O3′H3′) play important roles in construction of the unique H-
bonding interactions between DAP and LUT molecules. When
either of the phenol groups is absent, such as FIS and GEN, a
co-crystal with DAP cannot be synthesized. The two hydroxyl
groups perform a pivotal function in the formation of the two-
dimension network in the DAP·SUL co-crystal.
DAP·CAF Co-Crystal. Depending on the crystallization
solvent mixture utilized, DAP and CAF were found to
crystallize in 1:1 and 1:2 stoichiometries to give DAP·CAF-1
and DAP·CAF-2 co-crystals, respectively. The crystal structure
of DAP·CAF-1 was solved and refined in the orthorhombic
system with the Pbca space group. One DAP molecule and one
CAF molecule are included in the asymmetric unit. Notably, a
cyclic tetramer structure constructed by four DAP molecules
connected via consecutive H-bonding interactions via synthon I
(N1−H4B···O1, 2.17 Å; N2−H10A···O2, 2.13 Å) (Figure 6a).
Caffeine molecules were found to be anchored on the amino
group of each DAP molecule via synthon II (N2−H10B···O3,
2.02 Å). These tetramers further extended into the 2D H-
bonding network via synthon I along the a-axis and b-axis to
form a grid-sheet structure (Figure 6b).
Compared with the crystal structure of DAP·CAF-1, that of
the DAP·CAF-2 co-crystal presents obviously different unit cell
parameters and a much more complex hydrogen-bonding
network. DAP·CAF-2 was solved in the orthorhombic crystal
system with the Pna21 space group and two DAP molecules
and four CAF molecules in the asymmetric unit. The CAF
molecules present both complex hydrogen-bonding formations
and complex packing patterns. These CAF molecules can be
divided into four types. Each type of CAF molecule shows
different H-bonding and packing patterns via synthon II
(Figure 7a). When viewed along the c axis, the DAP molecules
Figure 5. (a) Major crystal interactions. (b) 2D sheet running along and CAF molecules are constructed into two types of 2D sheet
the b axis of DAP·LUT. structures. In the type A layer, DAP molecules are
intermolecularly H-bonded via synthon I (N1−H4B···O2, 2.49
flavonoid coformer, the crystal structure of DAP·LUT presents Å), forming infinite chains of DAP. The CAF molecules are
some similarities to that of the DAP·FLA co-crystal. The DAP connected by nonclassic auxiliary C−H···O H-bonding
host molecules are connected by an intermolecular H-bonding interactions (C220−H68···O23, 2.36 Å) to generate chains of
interaction between the amino group and sulfonyl oxygen via CAF molecules along the a axis. These chains interweave with
synthon I (N2−H10B···O1, 2.50 Å) to form one-dimensional each other via synthon II (N1−H4A···O24, 2.31 Å), forming a
chains along the c axis. Between the DAP chain structures, 2D network structure (Figure 7b). In the type B layer, zigzag
luteolin molecules also arrange in chains parallel to the DAP CAF chains are connected via nonclassic hydrogen-bonding
chain and form H-bonding interactions with the amino groups interactions (C119−H112···O3, 2.58 Å; C20-H20···O14, 2.41 Å).
from DAP (N1−H4A···O17, 2.37 Å; N1−H4A···O3′, 2.16 Å; O4′- Between the 1D infinite chains, the CAF molecules are linked
H4′···N1, 2.12 Å) to form a two-dimensional network structure via synthon II (N2−H10A···O13, 2.17 Å; N2−H10B···O4, 2.13 Å)
(Figure 5b). Isolated ethanol molecules were found to fill into to give a 2D H-bonding network (Figure 7c). These two types
the void space and form H-bonding interactions with the of layers assemble in an “AAB” fashion along the c axis to
phenol groups of LUT. It was found that the inclusion of construct the 3D structure of DAP·CAF-2 (Figure S2,
ethanol molecules is important in facilitating the formation of Supporting Information).
the DAP·LUT co-crystal. When other solvents were utilized, no Interestingly, each CAF molecule contains three N-
co-crystal of DAP·LUT could be obtained even with the same substituted methyl groups in the structure. It was found that
4568 dx.doi.org/10.1021/cg500668a | Cryst. Growth Des. 2014, 14, 4562−4573
Crystal Growth & Design Article

Figure 6. (a) Supramolecular tetramer unit of DAP·CAF-1. (b) 2D structure of DAP·CAF-1 viewed along the c axis.

the two methyl groups (designated as Me-A and Me-B) play The sensitivity of mid-IR spectra to the solid-state form has
crucial roles in the architecture of the corresponding crystal rendered them a useful tool for characterization and study of
structures of the CAF co-crystals (Figure 8). Particularly, in the co-crystal formation. The FT-IR spectra of DAP and its co-
DAP·CAF-1 crystal structure, Me-A participates in the crystals are compared in Figure S4 (Supporting Information).
formation of a one-dimensional chain structure with DAP Because of the changes in H-bonding patterns and interactions,
molecules via C13−H13A···N6 (2.72 Å) hydrogen bonding and significant differences were observed between the co-crystals
Me-B forms a strong C−H···π interaction (C19−H19C···π, 3.01 and DAP in the vibration frequency corresponding to the
Å) with the DAP molecules (Figure 8a). In DAP·CAF-2, Me-A aromatic amino moiety and sulfonyl groups ranging from 3300
is connected with the host molecule via C13−H13A···O2′ (2.56 to 3500 cm−1 and from 1100 to 1300 cm−1, respectively.
Å) hydrogen bonding and Me-B is connected to another CAF Thermal Analysis. All the solids were analyzed by TGA
molecule via C19−H19B···O14 (2.36 Å) hydrogen bonding and DSC to probe the thermal stability and purity. The
(Figure 8b). Noticeably, a DAP co-crystal could not be formed decomposition temperature and melting point of each co-
when analogues of CAF, such as THP and THB, were used crystal along with the corresponding data for the pure
under similar crystallization conditions (Table S2, Supporting components are summarized in Table 4. The results show
Information). that the decomposition temperatures of these co-crystals are
DAP·HBZ Co-Crystal. The single-crystal structure of DAP· correlated to their coformers. Generally speaking, the higher
HBZ was solved and refined in the monoclinic system with the the thermal stability of the coformer, the higher the thermal
P21/n space group. One DAP molecule and one HBZ molecule stability of the corresponding co-crystal. The DSC thermo-
are included in its asymmetric unit. The DAP·HBZ co-crystal grams of DAP·FLA, DAP·CAF-1, and DAP·HBZ present sharp
resembles an inclusion complex (Figure 9a). Four DAP endotherms for melting, confirming the bulk purity and
molecules are combined by intermolecular hydrogen bonding homogeneity (Figures S5−S7, Supporting Information). The
between two amino groups (N1−H4A···N2, 2.33 Å) and H- DSC curve of DAP·LUT presents a pronounced endothermic
bonding between the amino and sulfonyl groups via typical peak at 172 °C, corresponding to the liberation of ethanol
synthon I (N2−H10B···O2, 2.41 Å) to form tetrameric molecules, and such an endotherm signal is also echoed in the
structures. These tetramers are extended via the R44(20) ring weight loss step in the corresponding TGA diagram (Figure S8,
motif to form a two-dimensional network structure (Figure 9b). Supporting Information). A phase transition was detected in
Two guest molecules are interconnected via homomeric the DSC ramp of the DAP·SUL co-crystal (Figure S9,
synthon R22(8) to form a dimer, and such dimers are Supporting Information). To study the thermodynamic
incorporated into the DAP tetramer cages. Nonclassic contacts, property of DAP·SUL, hyper-DSC was employed to investigate
including C14−H14···π (2.60 Å), C15−H15···π (2.66 Å), C16− the unstable polymorphic form transformation behavior under
H16···O2 (2.64 Å), and S2···π (3.55 Å), further stabilize the accelerated heating rates. The phase transformation can be
incorporation of guest HBZ into this structure (Figure 9c). almost fully inhibited when the scan rate reaches 150 °C/min
Powder X-ray Diffraction and FT-IR Spectroscopy. The or higher (Figure S10, Supporting Information).
XRPD patterns of the six co-crystals and the parent compound Solubility and Dissolution. Equilibrium solubility was
DAP are compared (Figure S3, Supporting Information). The determined by slurrying an excess amount of materials under
results show that the XPRD patterns of the co-crystals are various buffer conditions, including pH 2.0, 4.6, 6.8 buffers and
significantly different from that of DAP itself. All the 1% Tween 80 aqueous solution. The residual solids were
experimental XRPD patterns of the bulk samples closely examined by XRPD, and the solubility results were compared
match the patterns simulated from the corresponding single- with those of the pure DAP drug (Table 5; Figures S11−S16,
crystal diffraction data. The excellent agreement confirmed the Supporting Information). The solubility of DAP at pH 4.6 and
high purity and homogeneity of the bulk co-crystals. Notice- 6.8 was determined to be 0.17 mg/mL, which agrees well with
ably, the bulk solid of DAP·CAF-2 was found to be a mixture of the results from the literature.9 No significant solubility
DAP·CAF-1 and DAP·CAF-2. Surprisingly, attempts to prepare improvement was observed for all the co-crystals studied in
pure DAP·CAF-2 powder were unsuccessful even with a 1% Tween 80 solution. However, DAP·CAF-1 exhibits about
significant effort to change the crystallization method. 1.8 times higher equilibrium solubility than the DAP pure form
4569 dx.doi.org/10.1021/cg500668a | Cryst. Growth Des. 2014, 14, 4562−4573
Crystal Growth & Design Article

Figure 7. (a) Major interaction between the four types of CAF and host molecules in DAP·CAF-2. (b) Type A layer in DAP·CAF-2. (c) Type B
layer in DAP·CAF-2. Only the amino groups are depicted for clarity.

in both pH 4.6 and pH 6.8 media. No solid form DAP, DAP·FLA, and DAP·LUT, solid-phase transformation
transformation was observed with DAP·SUL, DAP·CAF-1, was observed during the equilibrium process; consequently, the
and DAP·HBZ during the solubility experiments. However, for equilibrium solubility data cannot represent the true solubility
4570 dx.doi.org/10.1021/cg500668a | Cryst. Growth Des. 2014, 14, 4562−4573
Crystal Growth & Design Article

Figure 8. (a) Nonclassic contacts of the methyl groups in DAP·CAF-1. (b) Nonclassic contacts of the methyl groups in DAP·CAF-2.

Figure 9. (a) Tetramer unit constructed by DAP and the attachment of the HBZ dimer to DAP. (b) 2D network of DAP·HBZ. (c) Nonclassic
contacts for the stabilization of the coformer of DAP·HBZ.

Table 4. Decomposition Temperature and Melting Onset values of the corresponding starting materials. A close
(°C) of DAP and Its Co-Crystals examination of the residual solids revealed that both DAP
and the DAP·LUT co-crystal were converted to the DAP
decomp temp decomp temp mp (°C) mp (°C)
(°C) of (°C) of the of of the hydrate form and DAP·FLA was converted to a new co-crystal
compd API/CCF co-crystal API/CCF co-crystal polymorph, which was characterized by NMR, XRPD, and
DAP·SUL 281/233 245 176/153 141 thermal analysis (Figures S17−S20, Supporting Information). It
DAP·FLA 281/142 184 176/94 103 is generally expected that when no phase transformation occurs,
DAP·LUT 281/328 285 176/328 158 solubility is limited by both the lattice energy and solution
DAP·CAF-1 281/162 162 176/− 172 energy (hydrophobicity of the solute). The melting point can
(sublimation) be a critical parameter to access the lattice energy of the crystal.
DAP·HBZ 281/157 140 176/130 148
However, due to the involvement of different coformers in the
solution, the solution energies cannot be directly compared.
According to the solubility results, the inverse correlation
4571 dx.doi.org/10.1021/cg500668a | Cryst. Growth Des. 2014, 14, 4562−4573
Crystal Growth & Design

■
Article

Table 5. Summary of the Equilibrium Solubility of DAP and CONCLUSIONS

Its Co-Crystals As an alternative approach to improve the physicochemical and
apparent solubility (mg·mL−1) pharmacokinetic behaviors of drug substances, co-crystals
arouse general interest in both academia and industry. Interest
1% residue after
material Tween pH 2.0 pH 4.6 pH 6.8 equilibrium in the design and synthesis of co-crystals has explosively grown
DAP 0.85a 0.40a 0.17a 0.17a DAP hydrate
with an increasing number of publications on the subject in
DAP·SUL 0.92 0.38 0.19 0.18 co-crystal recent years. However, it is still a challenging task to precisely
DAP·FLA 0.92a 0.47a 0.20a 0.18a new polymorph predict, rationally design, and successfully synthesize any co-
DAP·LUT 0.26a 0.51a 0.20a 0.20a DAP hydrate crystal even with the simplest chemical structures. In this work,
DAP·CAF-1 0.81 0.65 0.30 0.31 co-crystal on the basis of an in-deep analysis of the characteristic crystal
DAP·HBZ 0.49 0.22 0.13 0.12 co-crystal structure and functionalities of DAP, sulfonyl- and carbonyl-
a containing compounds were deliberately selected as coformers
The values correspond to the equilibrium solubility of the residue
solids.
to prepare DAP co-crystals. The interactions between the
aromatic amino groups of DAP and sulfonyl (synthon I) or
between the melting point and equilibrium solubility was not carbonyl (synthon II) oxygen were utilized in designing and
effectively observed in the DAP co-crystals. directing the synthesis of DAP co-crystals. Six drug−drug co-
The IDRs of DAP and its co-crystals were also studied under crystals of DAP were successfully prepared, and the structures
various buffer conditions (Table 6). The IDR trend of the co- were revealed by single-crystal X-ray diffraction analysis.
Noticeably, in each series of structural analogues, not all
coformers can successfully form co-crystals with DAP.
Table 6. Summary of Intrinsic Dissolution Rates of DAP and
Reoccurring packing patterns and H-bonding interactions via
Its Co-Crystals
synthons I and II were observed in the various DAP co-crystals.
IDR (μg·cm−2·min−1) The unsubstituted sulfonamide group in SUL was important in
material 1% Tween pH 2.0 pH 4.6 pH 6.8 formation of the DAP·SUL co-crystal; the two unique phenol
DAP 54.9 98.9 44.9 31.1
groups in DAP·LUT and the two indispensable methyl groups
DAP·SUL 63.6 70.6 36.1 13.7
in DAP·CAF-1 and DAP·CAF-2 perform key roles in the
DAP·FLA 24.6 46.6 35.6 19.8
architecture of the corresponding co-crystals. If such moieties
DAP·LUT 22.2 105.1 52.5 36.4
of great importance are removed or replaced, the co-crystal
DAP·CAF-1 66.2 94.9 51.6 38.5
structure cannot be built. In addition, the physicochemical
DAP·HBZ 37.6 52.4 12.1 18.8
properties of the DAP co-crystals were investigated by thermal
analysis (TGA and DSC), PLM, XRPD, and Raman and FT-IR
spectroscopy. Their equilibrium solubility and intrinsic
dissolution rates were compared with those of the parent
crystals in 1% Tween is DAP·CAF-1 (66.2) > DAP·SUL (63.6) drug as well. The results show the influence of co-crystallization
> DAP·HBZ (37.6) > DAP·FLA (24.6) > DAP·LUT (22.2), on the properties of DAP. This work also expands the
and this trend is directly correlated to the order of solubility of pharmaceutically acceptable solid forms of DAP and supple-
these co-crystals. ments the successful cases of drug−drug co-crystals.

■
Physical Stability. Pharmaceutical solids often contact
water during the manufacturing and processing procedures, and ASSOCIATED CONTENT
they may also be exposed to humid air during storage. Hence,
water may be absorbed to form new hydrate forms. Such *
S Supporting Information

transformation always results in products with drastically Characterization data and crystallographic data files (CIF
different physicochemical properties. Therefore, monitoring format). This material is available free of charge via the Internet
at http://pubs.acs.org.

■
and controlling the physical stability is an important element
during the course of whole drug development. Previous
investigations have found that both DAP and caffeine can be AUTHOR INFORMATION
converted to their corresponding hydrate forms when exposed Corresponding Author
to high relative humidity (RH), e.g., >75% RH or higher.38,48 In *E-mail: xuefengmei@simm.ac.cn. Phone: 001-86-021-
this study, accelerated stability experiments were conducted to 50800934.
determine the physical stabilities of DAP and its co-crystals. Notes
The results show that DAP anhydrate was converted to its
The authors declare no competing financial interest.
hydrate form when stored at 40 °C/75% RH for 4 months
(Figure S21, Supporting Information) while the solid forms for
all the co-crystals remained unchanged (Figures S22−S26,
Supporting Information). Similar experiments were also
■ ACKNOWLEDGMENTS
We thank the National Natural Science Foundation of China
performed using caffeine and its co-crystal DAP·CAF-1. After (Grant 81273479) and Chinese Academy of Science for
the materials were exposed to 98% RH for only 1 day at room funding.
temperature, caffeine was found to be converted to its hydrate
form (Figure S27, Supporting Information) and the solid-state
form of the DAP·CAF-1 co-crystal was found to remain
■ REFERENCES
(1) Barr, J. J. Hist. Med. Allied Sci. 2011, 66 (4), 425−467.
unchanged. These results indicate that co-crystallization can be (2) Wozel, V. E. G. Dermatol. Clin. 2010, 28 (3), 599−610.
used as an important technique to improve the physical stability (3) Lemmer, H.; Stieger, N.; Liebenberg, W.; Caira, M. R. Cryst.
of particular drug substances. Growth Des. 2012, 12 (3), 1683−1692.

4572 dx.doi.org/10.1021/cg500668a | Cryst. Growth Des. 2014, 14, 4562−4573

 Crystal Growth & Design Article

(4) Martins, I.; Martins, M.; Fernandes, A.; Andre, V.; Duarte, M. T. (38) Yathirajan, H. S.; Nagaraja, P.; Nagaraj, B.; Bhaskar, B. L.;
CrystEngComm 2013, 15 (40), 8173−8179. Lynch, D. E. Private communication, 2004.
(5) Grebogi, I. H.; Tibola, A.; Barison, A.; Grandizoli, C.; Ferraz, H. (39) Hanafy, A.; Uno, J.; Mitani, H.; Kang, Y.; Mikami, Y. Jpn. J. Med.
G.; Rodrigues, L. N. C. J. Inclusion Phenom. Macrocyclic Chem. 2012, 73 Mycol. 2007, 48 (1), 47−50.
(1−4), 467−474. (40) Cushnie, T. P. T.; Lamb, A. J. Int. J. Antimicrob. Agents 2011, 38
(6) Cao, X. H.; Gibbs, S. T.; Fang, L. Y.; Miller, H. A.; Landowski, C. (2), 99−107.
P.; Shin, H. C.; Lennernas, H.; Zhong, Y. Q.; Amidon, G. L.; Yu, L. X.; ́
(41) Sowa, M.; Slepokura, K.; Matczak-Jon, E. CrystEngComm 2013,
Sun, D. X. Pharm. Res. 2006, 23 (8), 1675−1686. 15 (38), 7696−7708.
(7) Hughes, W. T. Clin. Infect. Dis. 1998, 27 (1), 191−204. (42) Cushnie, T. P. T.; Lamb, A. J. Int. J. Antimicrob. Agents 2005, 26
(8) Monteiro, L. M.; Lione, V. F.; do Carmo, F. A.; do Amaral, L. H.; (5), 343−356.
da Silva, J. H.; Nasciutti, L. E.; Rodrigues, C. R.; Castro, H. C.; de (43) Daly, J. Cell. Mol. Life Sci. 2007, 64 (16), 2153−2169.
Sousa, V. P.; Cabral, L. M. Int. J. Nanomed. 2012, 7, 5175−5182. (44) Ucar, H.; Van derpoorten, K.; Cacciaguerra, S.; Spampinato, S.;
(9) Calderini, A.; Martins, M. H.; Pessine, F. B. T. J. Inclusion Stables, J. P.; Depovere, P.; Isa, M.; Masereel, B.; Delarge, J.; Poupaert,
Phenom. Macrocyclic Chem. 2013, 75 (1−2), 77−86. J. H. J. Med. Chem. 1998, 41 (7), 1138−1145.
(10) Chougule, M.; Padhi, B.; Misra, A. AAPS PharmSciTech 2008, 9 (45) Roger, G.; Dollé, F.; De Bruin, B.; Liu, X.; Besret, L.; Bramoullé,
(1), 47−53. Y.; Coulon, C.; Ottaviani, M.; Bottlaender, M.; Valette, H. Bioorg. Med.
(11) Pochopin, N. L.; Charman, W. N.; Stella, V. J. Int. J. Pharm. Chem. 2004, 12 (12), 3229−3237.
1995, 121 (2), 157−167. (46) Foulon, C.; Duhal, N.; Lacroix-Callens, B.; Vaccher, C.; Bonte, J.
(12) Smith, G.; Wermuth, U. D. Acta Crystallogr., Sect. E: Struct. Rep. P.; Goossens, J. F. Eur. J. Pharm. Sci. 2007, 31 (3−4), 165−171.
Online 2012, 68 (3), 669. (47) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr.
(13) Smith, G.; Wermuth, U. D. J. Chem. Crystallogr. 2013, 43 (12), 2007, 64 (1), 112−122.
664−670. (48) Trask, A. V.; Motherwell, W. S.; Jones, W. Cryst. Growth Des.
(14) Vishweshwar, P.; Nangia, A.; Lynch, V. M. Cryst. Growth Des. 2005, 5 (3), 1013−1021.
2003, 3 (5), 783−790.
(15) Trask, A. V.; Motherwell, W. D. S.; Jones, W. Cryst. Growth Des.
2005, 5 (3), 1013−1021.
(16) Jones, W.; Motherwell, W.; Trask, A. V. MRS Bull. 2006, 31
(11), 875−879.
(17) Bethune, S. J.; Schultheiss, N.; Henck, J. O. Cryst. Growth Des.
2011, 11 (7), 2817−2823.
(18) Aher, S.; Dhumal, R.; Mahadik, K.; Ketolainen, J.; Paradkar, A.
Pharm. Dev. Technol. 2013, 18 (1), 55−60.
(19) Vangala, V. R.; Chow, P. S.; Tan, R. B. CrystEngComm 2011, 13
(3), 759−762.
(20) Sanphui, P.; Goud, N. R.; Khandavilli, U. R.; Nangia, A. Cryst.
Growth Des. 2011, 11 (9), 4135−4145.
(21) Goud, N. R.; Gangavaram, S.; Suresh, K.; Pal, S.; Manjunatha, S.
G.; Nambiar, S.; Nangia, A. J. Pharm. Sci. 2012, 101 (2), 664−680.
(22) Hickey, M. B.; Peterson, M. L.; Scoppettuolo, L. A.; Morrisette,
S. L.; Vetter, A.; Guzmán, H.; Remenar, J. F.; Zhang, Z.; Tawa, M. D.;
Haley, S. Eur. J. Pharm. Biopharm. 2007, 67 (1), 112−119.
(23) Thakuria, R.; Delori, A.; Jones, W.; Lipert, M. P.; Roy, L.;
Rodríguez-Hornedo, N. Int. J. Pharm. 2013, 453 (1), 101−125.
(24) Tao, Q.; Chen, J.-M.; Ma, L.; Lu, T.-B. Cryst. Growth Des. 2012,
12 (6), 3144−3152.
(25) Shan, N.; Zaworotko, M. J. Drug Discovery Today 2008, 13 (9),
440−446.
(26) Babu, N. J.; Nangia, A. Cryst. Growth Des. 2011, 11 (7), 2662−
2679.
(27) Good, D. J.; Rodríguez-Hornedo, N. Cryst. Growth Des. 2009, 9
(5), 2252−2264.
(28) Bolla, G.; Sanphui, P.; Nangia, A. Cryst. Growth Des. 2013, 13
(5), 1988−2003.
(29) Aitipamula, S.; Wong, A. B. H.; Chow, P. S.; Tan, R. B. H.
CrystEngComm 2014, 16, 5793−5801.
(30) Sekhon, B. S. Daru, J. Pharm. Sci. 2012, 20, 45.
(31) Bhatt, P. M.; Azim, Y.; Thakur, T. S.; Desiraju, G. R. Cryst.
Growth Des. 2008, 9 (2), 951−957.
(32) Aitipamula, S.; Chow, P. S.; Tan, R. B. CrystEngComm 2009, 11
(9), 1823−1827.
(33) Grobelny, P.; Mukherjee, A.; Desiraju, G. R. CrystEngComm
2011, 13 (13), 4358−4364.
(34) Brittain, H. G.; Felice, P. V. US 8304404 B2, 2012.
(35) Veverka, M.; Simon, P.; Gallovic, J.; Jorik, V.; Veverkova, E.;
Dubaj, T. Monatsh. Chem. 2012, 143 (10), 1405−1415.
(36) Vitorino, G. P.; Sperandeo, N. R.; Caira, M. R.; Mazzieri, M. R.
Cryst. Growth Des. 2013, 13 (3), 1050−1058.
(37) Alleaume, M. Doctoral Thesis, University of Bordeaux, France,
1967.

4573 dx.doi.org/10.1021/cg500668a | Cryst. Growth Des. 2014, 14, 4562−4573

View publication stats