Improvement of the water solubility of tolfenamic acid by new multiple-component crystals produced by mechanochemical methods

Article abstract of DOI:10.1039/c4ce00549j

Tolfenamic acid (HTA) is a drug characterized by very poor water solubility (13.6 nM under acidic conditions) and moderate solubility in ethanol (0.17 M). A series of new multicomponent crystals have been obtained by applying mechanochemical methods (i.e.

Full text of DOI:10.1039/c4ce00549j

CrystEngComm
View Article Online
View Journal | View Issue
PAPER
Improvement of the water solubility of tolfenamic
acid by new multiple-component crystals
produced by mechanochemical methods†
Cite this: CrystEngComm, 2014, 16,
8252
a
a
a
K. Gaglioti, M. R. Chierotti, F. Grifasi, R. Gobetto, a U. J. Griesser,b D. Hasac
*
and D. Voinovichc
Tolfenamic acid (HTA) is a drug characterized by very poor water solubility (13.6 nM under acidic
conditions) and moderate solubility in ethanol (0.17 M). A series of new multicomponent crystals have been
obtained by applying mechanochemical methods (i.e. kneading) to mixtures of HTA with sodium acetate,
sodium carbonate, sodium hydroxide and imidazole. These reactions resulted in two salts (NaTA·0.5H2O
and NaTA HT Form), a co-crystal of salts (NaTA·HTA·0.5NaAc·2H2O) and two salt co-crystals
(NaTA·HTA·H2O/NaHCO3 and IMH-TA·HTA). Due to the lack of suitable crystals for single-crystal X-ray
diffraction analysis, the structural features of the samples have been characterized by solid-state NMR
1
(1H MAS, 13C CPMAS, H-13C FSLG LG-CP HETCOR and 15N CPMAS), IR(ATR) and Raman spectroscopy,
Received 17th March 2014,
Accepted 25th July 2014
VT-XRPD and elemental analysis. The evaluation of thermal stability and dissolution behavior was
performed using thermogravimetry, differential scanning calorimetry and dissolution kinetic tests. The new
solid-state forms show better thermal stability than pure HTA and an improved dissolution rate, which is
most pronounced in NaTA·HTA·H2O/NaHCO3, NaTA HT Form and NaTA·0.5H2O.
DOI: 10.1039/c4ce00549j
www.rsc.org/crystengcomm
barbituric acid.5 The difference between ionic co-crystals and
salt co-crystals is found in the type of co-former used. Ionic
Introduction
Multicomponent crystal formation (salts and co-crystals) is an
effective means to tune the properties of a molecule or, in the
pharmaceutical field, an active pharmaceutical ingredient
(API) by changing its crystal packing.1 A co-crystal is a multi-
component molecular crystal that is a crystalline substance
formed by two or more chemically different molecules.2
Thus, solvates, hydrates and both stoichiometric and non-
stoichiometric lattice inclusion compounds are included in
this definition.3 With respect to the pure polymorphism phe-
nomenon, co-crystal formation provides much more in terms
of possibilities for modulating crystal packing and this is due
to the choice of two parameters: the co-former and the type of
weak interaction.4 The use of inorganic salts as co-formers
has already been explored by our group as an effective
method for producing ionic co-crystals and salt co-crystals of
co-crystals are formed by an organic molecule and an ionic
salt (i.e. alkali halide) in the same asymmetric unit, while salt
co-crystals are formed by an organic molecule and one of its
salts in the same asymmetric unit.6
Tolfenamic acid (HTA, 2-[(2-methyl-3-chlorophenyl)amino]
benzoic acid – Scheme 1) is a drug that belongs to the fena-
mate family. It is a potent, well-tolerated non-steroidal
anti-inflammatory drug (NSAID) that displays low gastro-
ulcerogenicity,7 low overall toxicity and high therapeutic
indices as well as analgesic and antipyretic properties.8
Recent studies have revealed that it is a Cox-2 inhibitor.9 It is
also extensively used in both human and veterinary medicine
for its analgesic and antipyretic properties.10
A common property of fenamates is their general lack
of solubility in water and other common organic solvents
a Dip. di chimica, Fisica d dei materiali, Via P. Giuria, Torino, Italy.
E-mail: roberto.gobetto@unito.it; Fax: 3.9011670786e+011;
Tel: 3.9011670752e+011
b Institute of Pharmacy, Pharmaceutical Technology, University of Innsbruck,
Josef-Moeller-Haus Innrain 52c, Innsbruck, Austria. E-mail: ulrich.griesser@uibk.ac.at;
Fax: +43 512 507 2719; Tel: +43 512 507 5309
c Dept. of Chemical and Pharmaceutical Sciences, University of Trieste, P.le
Europa 1, 34127, Trieste, Italy. E-mail: vojnovic@units.it
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4ce00549j
Scheme 1 HTA structure with atom labelling.
8252 | CrystEngComm, 2014, 16, 8252–8262
This journal is © The Royal Society of Chemistry 2014

View Article Online
CrystEngComm
Paper
(obviously, their solubility is influenced by their polymorphic
form);11 thus, HTA belongs to class II of the Biopharmaceutics
Classification System (high permeability in cells and low
solubility in water).12
[Elem. anal. calc. for C14H12N1O2.5Cl1Na1: C, 57.45; H, 4.13;
N, 4.79; O, 13.67; Cl, 12.11; Na, 7.85%. Found: C, 57.26;
H, 4.29; N, 4.88; O, 13.72; Cl, 12.08; Na, 7.77%.]
NaTA·HTA·H2O/NaHCO3: this is a heterogeneous mixture
between NaTA·HTA and NaHCO3. It was not possible to deter-
mine whether H2O co-crystallizes with NaTA·HTA or NaHCO3.
It occurs as a white powder and was obtained quantitatively
by kneading (ethanol/water 2 : 1, 150 μl) HTA (200 mg)
and Na2CO3 (40.5 mg) with a stoichiometric ratio of 2 : 1 at
room temperature for 20 minutes. It was not possible to
obtain pure NaTA·HTA without NaHCO3 using different ratios
of the starting material either. The stoichiometric formula
NaTA·HTA·H2O/NaHCO3 was confirmed by solid-state NMR,
thermal analysis and elemental analysis. [Elem. anal. calc. for
HTA is known to exist in at least five different polymor-
phic forms.13 The most prominent are anhydrous Form I and
Form II, which show very poor water solubility.14 In order to
improve the water solubility and thus the biopharmaceutical
properties of HTA, we herein report a synthetic strategy
for new HTA co-crystal forms, which is based on mechano-
chemical and solution reactions between HTA, inorganic
reagents [NaOH, Na2CO3 and sodium acetate (NaAc)] and an
organic compound such as imidazole (IM). IM is used as a
precursor for many pharmaceutical compounds, but it is
toxic (LD50 in pigs = 760 mg kg−1) and thus not therapeuti-
cally used.15 Nevertheless, we decided to use IM as a model
co-former in order to study the possibility of an interaction
with the COOH group of HTA.
C
29H28N2O9Cl2Na2: C, 52.34; H, 4.24; N, 4.21; Na, 6.91%.
Found: C, 52.47; H, 4.29; N, 4.08; Na, 6.88%.]
NaTA·HTA·0.5NaAc ·2H2O: this co-crystal of salts was
obtained as a white powder and in maximal yield by kneading
(EtOH 99%, 150 μl) stoichiometric ratios (2 : 1.5) of HTA
(200 mg) and NaAc (90 mg) at room temperature for 20 minutes.
Solid-state NMR, thermal analysis and elemental analysis
indicated that the formula is NaTA·HTA·0.5NaAc·2H2O.
[Elem. anal. calc. for C29H28.5N2O7Cl2Na2: C, 55.96; H, 4.62;
N, 4.50; Na, 7.39%. Found: C, 55.78; H, 4.56; N, 4.69; Na,
7.19%.]
Experimental and methods
HTA (Tokyo Chemical Industry Company and Sigma Aldrich),
Na2CO3, NaOH, IM and NaAc (all from Sigma Aldrich) were
used without further purification.
The HTA commercial batch was characterized by solid-
state NMR analysis as pure Form I. HTA Form II was reproduced
following the method reported in the literature16 only for
comparison.
NaTA HT Form (salt): a white powder was obtained after
heating NaTA·0.5H2O or NaTA·HTA·H2O/NaHCO3 at 170–180 °C
in a muffle oven (ramp 5 °C min−1). [Elem. anal. calc. for
All samples were obtained via mechanochemical [kneading
with a mortar and pestle, i.e. grinding with a catalytic amount
(from 50 to 150 μl, corresponding to two to six drops) of the
solvent] and solvent techniques.2,17 Experiments performed
and products obtained are summarized in Table 1.
The NaTA polymorphic mixture was prepared by refluxing
128 mg of HTA and 1 ml of 0.5 M NaOH (stoichiometric ratio
1 : 1) in 5 ml of absolute EtOH for 3 h.
C14H11N1O2Cl1Na1: C, 59.27; H, 3.91; N, 4.94; O, 11.28;
Cl, 12.50; Na, 8.10%. Found: C, 59.07; H, 4.13; N, 4.70;
O, 11.46; Cl, 12.46; Na, 8.18%.]
IMH-TA·HTA (salt co-crystal): this product (white powder)
was obtained by kneading HTA (200 mg) and IM (26 mg) in a
2 : 1 ratio for 15 minutes in the presence of absolute EtOH
(50 μl). [Elem. anal. calc. for C31H28N4O4Cl2: C, 62.95; H, 4.77;
N, 9.47%. Found: C, 62.87; H, 4.66; N, 9.32%.]
NaTA·0.5H2O (salt): a white powder of this hemihydrate
salt was quantitatively obtained by kneading (H2O, 50 μl) a
NaTA polymorphic mixture at room temperature for 5 minutes.
NaTA·0.5H2O was also obtained by directly kneading (H2O,
100 μl) HTA with NaOH. However, this procedure did not
allow for precise stoichiometric control because the hygro-
scopic nature of NaOH makes correct weighing difficult.
All solid-state NMR spectra were recorded with a Bruker
Avance II 400 instrument operating at 400.23, 100.65 and
1
40.55 MHz for H, 13C and 15N nuclei, respectively. For 13C
and 15N CPMAS (cross-polarization magic angle spinning)
spectra, cylindrical 4 mm o.d. zirconia rotors with a sample
volume of 80 μL were employed. Samples were spun at 12 and
9 kHz for 13C and 15N, respectively. A ramp cross-polarization
Table 1 Summary of the performed experiments involving HTA and additional components (k = kneading, h = heating)
Components
Preparation
Class
Formula/name
Ratio
HTA
NaTA mixture
NaOH
Salification
k (1-BuOH, THF,
DMF, H2O)
Salt
Salt
NaTA polymorphic mixture
NaTA·0.5H2O
1 : 1
—
HTA
Na2CO3
NaAc
k (EtOH/H2O 2 : 1)
Heterogeneous mixture
with a salt co-crystal
Co-crystal of salts
Salt
NaTA·HTA·H2O/NaHCO3
2 : 1
HTA
k (EtOH 99%)
h (170–180 °C)
NaTA·HTA·0.5NaAc·2H2O
NaTA HT Form
2 : 1.5
NaTA·0.5H2O
or NaTA·HTA
2H2O/NaHCO3
HTA
IM
k (EtOH 99%)
Salt
IMH-TA·HTA
2 : 1
This journal is © The Royal Society of Chemistry 2014
CrystEngComm, 2014, 16, 8252–8262 | 8253

View Article Online
Paper
CrystEngComm
pulse sequence was used with contact times of 3–5 (13C) or
Differential scanning calorimetry (DSC) was performed on
a DSC Q200 (TA Instruments). Samples (5–10 mg) were placed
in sealed alumina pans and heated at a rate of 10 °C min−1
(temperature range 30 to 400 °C).
1
4 ms (15N), a H 90° pulse of 3.8 μs, recycle delays of 1–60 s
and 48–4096 transients for 13C and 1598–5956 transients
for 15N. Proton spin–lattice relaxation times, 1H T1, were
obtained using the 13C-detected inversion–recovery cross-
polarization technique [1H(180°-τ-90°)-CP-13C(FID)] where one
measures the 13C magnetization that appears through the CP
process after the relaxation of 1H magnetization over time
τ. 160–1024 transients were acquired and a variable delay
list file (from 22 ms to 42 s) was used. 2D 13C-1H FSLG
(frequency-switched Lee–Goldburg) on- and off-resonance
HETCOR (heteronuclear correlation) spectra were measured
according to the method developed by van Rossum et al. The
MAS rate was set at 12 kHz. The proton rf field strength used
during the t1 delay for FSLG decoupling and during the
acquisition for TPPM (two-pulse phase-modulated) decoupling
was 83 kHz. The two off-resonance pulses with opposite
phases (i.e., +x, −x or +y, −y) used during the FSLG decoupling
were set to 9.8 μs. The contact time used was 100 μs. The
magic angle (54.7°) pulse length for protons was set at 2.0 μs,
while the recycle delay used was 15 s. Quadrature detection
was achieved using the States-TPPI method. All data for 64 t1
increments with 140 scans were collected. A B1(1H) field inten-
sity for the CP of 75 kHz and a mixing period of 2.0 ms were
used for FSLG off-resonance CP (LG-CP) HETCOR NMR. The
1H NMR chemical shift scale in the HETCOR spectra was
corrected by a scaling factor of 1/√3 because the 1H NMR
chemical shift dispersion was scaled by a factor of 1/√3 during
FSLG decoupling. 1H MAS experiments were performed using
a 2.5 mm Bruker probe at a spinning speed of 32 kHz. The
1H MAS spectra were acquired using the DEPTH sequence
(π/2–π–π), which suppresses the probe background signal.19
The 1H 90° pulse length was set to 2.50 μs, the recycle delays
to 1–40 s and 32–64 transients were averaged for all samples.
1H, 13C and 15N chemical shifts were referenced via the reso-
nance of solid adamantane (1H signal at 1.87 ppm), HMB
Thermogravimetric (TGA) measurements were performed
under N2 flow (heating rate 10 °C min−1) using a Q600 SDT
Simultaneous DSC-TGA heat flow analyzer (TA Instruments).
The samples (5–10 mg) were placed in alumina pans. All DSC
and TGA thermograms are reported in Fig. S3 in the ESI.†
X-ray powder diffraction (XRPD) patterns of the samples
under investigation were collected using a PW3050/60 X'Pert PRO
MPD diffractometer (PANalytical), working in Debye–Scherrer
geometry and using, as the source, a high-powered ceramic tube
PW3373/10 LFF with a Cu anode equipped with a Ni filter to
attenuate Kb, which was focused by a PW3152/63 X-ray mirror.
Powdered samples were hosted inside a 0.8 mm boron sili-
cate capillary and mounted on a rotating goniometer head.
Variable temperature (VT)-XRPD studies were carried out
using an X'Pert diffractometer (PANalytical, Cu Kα radiation)
equipped with a hot stage and an environmental chamber
(Anton Paar XRK 900) to analyze the conversion of both
NaTA·0.5H2O and NaTA·HTA·H2O/NaHCO3 to the NaTA HT
Form. XRPD patterns of HTA Form I, Form II, and all multiple-
component crystals are reported in Fig. 1, while XRPD pat-
terns of co-formers used, VT-XRPD patterns, ramps used and
technical details are reported in the ESI† (Fig. S4 and in
Schemes 1 and 2).
Dissolution kinetic tests (DKT) were performed in a pH
7.4 buffer (0.2 M KH2PO4/0.2 M NaOH). 5 mg of HTA (as
a pure compound or a co-crystal and salt) was added to
150 cm3 of the dissolution medium at 37 °C in each experi-
ment. The release test lasted 180 min. In these experiments,
the solution was kept homogeneous using an impeller
(rotational speed 200 rpm) and the determination of HTA
concentration was performed using a fibre optic apparatus
(HELLMA, Milano, Italy) connected to a spectrophotometer
(ZEISS, Germany).21 Measurement was performed at 286.9 nm
(HTA maximum absorption, ESI†). Each sample was tested in
triplicate. The mother solution was obtained by dissolving
100 mg of HTA in 100 ml of methanol. The concentration of
the standard solutions obtained from the mother solution
and the calibration curve are reported in Fig. S5 in the ESI.†
In this dissolution medium, the equilibrium solubility of pure
HTA was 28.5 mg l−1.
(
13C methyl signal at 17.4 ppm) and (NH4)2SO4 (15N signal at
−355.8 ppm with respect to CH3NO2), respectively.
IR and Raman spectra were only used to screen the
products and to monitor reaction conversion. In order to dis-
criminate between the different forms, particular attention
was given to the regions from 3500 to 3300 cm−1 (NH and OH
stretching), 3100 and 3000 cm−1 (CH and CH2 stretching)
and 1700–1600 cm−1 (CO stretching). All IR and Raman
spectra are reported in Fig. S1 and S2 in the ESI.† Fourier-
transform infrared (FTIR) spectra were collected directly on
the sample using a Bruker IFS 25 spectrometer (Bruker
Analytische Messtechnik GmbH, Ettlingen, Germany) connected
to a Bruker IR microscope I (15× Cassegrain objective). The
samples were prepared on ZnSe disks and measured in trans-
mission mode (spectral range 4000–600 cm−1, resolution
4 cm−1, 32 interferograms per spectrum).20 Raman spectra
were measured using a Bruker Vertex 70 spectrometer
equipped with a RAM II using a 1064 nm Nd:YAG source and
a Ge diode detector (laser power 10–50 mW, spectral resolu-
tion 4 cm−1).
Elemental analysis was performed using a CHN analyzer
(CE Instruments, NA2100 Protein), and a Perkin Elmer 3030
flame atomic absorption spectrometer (FAAS) was used for
sodium determination to clarify the stoichiometry of com-
pounds. The data confirm the stoichiometries determined by
solid-state NMR. Data are reported in the Experimental and
methods section.
Results and discussion
The two anhydrous forms of HTA are characterized by the
presence of a dimer, which is connected through the very
8254 | CrystEngComm, 2014, 16, 8252–8262
This journal is © The Royal Society of Chemistry 2014

View Article Online
CrystEngComm
Paper
Scheme 2 Supramolecular synthons and hydrogen bond motif of
anhydrous HTA Forms I and II.
robust R22(8) carboxylic acid homodimeric synthon (Scheme 2),
as observed in the X-ray structures.22 Such a feature gives rise
to a single COOH resonance around 175 ppm in the 13C CPMAS
spectra (Fig. 2a and b). This is an intermediate value between
typical COOH and COO− chemical shifts (~173 and ~178 ppm,
respectively).23 The strength of this interaction most likely
explains the poor water solubility of HTA. Although one example
of a HTA co-crystal,24 five metal complexes25 and a calcium
salt are reported in the literature,26 the R22(8) carboxylic acid
homodimeric synthon is strong enough to prevent the forma-
tion of other co-crystals in which the COOH group must be
accessible.
Indeed, the NH group is not particularly suitable for
hydrogen bond interaction as both the steric hindrance of
Fig. 2 3C (100 MHz) CPMAS spectra with the main assignments of
HTA Forms I (a) and II (b), NaTA·0.5H2O (c), NaTA·HTA·H2O/NaHCO3 (d),
NaTA·HTA·0.5NaAc·2H2O (e), NaTA HT Form (f) and IMH-TA·HTA (g)
recorded at 12 kHz.
Fig. 1 XRPD patterns of HTA and all multiple-component crystals.
This journal is © The Royal Society of Chemistry 2014
CrystEngComm, 2014, 16, 8252–8262 | 8255

View Article Online
Paper
CrystEngComm
the methyl and COOH groups and the formation of an intra-
molecular hydrogen bond with the COOH moiety hamper it.
This is probably the main reason why it does not form classic
organic co-crystals with malonic acid, succinic acid, sodium
malonate, glycine, L-alanine, aspartic acid, sodium glycinate,
sodium acetate (NaAc), adenine, thymine, guanine and resor-
cinol. The ionic co-crystal approach27 using inorganic salts
(KI, KF, KBr, NaCl, KBr, KHCO3, K2CO3, NaHCO3, AlCl3 and
FeCl3) was also unsuccessful. For this reason, we decided to
use co-formers, which are either able to react quantitatively
with HTA to form the corresponding salt, or to partially salify
HTA and form a salt co-crystal such as salts and bases NaOH,
Na2CO3, NaAc and IM.
All samples were obtained via simple kneading. Crystalli-
zation with different solvents resulted in either a NaTA poly-
morphic mixture or a pure HTA Form I or II. Different forms
were present in the NaTA polymorphic mixture, as shown by
the 13C CPMAS spectrum (Fig. S6 in the ESI†). Several
attempts to promote the selective crystallization of single
forms using various solvents (EtOH, 1- BuOH, THF, DMSO,
DMF, H2O or CH3CN) failed. Thus, in this case, mechano-
chemistry allowed us to obtain pure compounds. The chemi-
cal shifts in the solid-state NMR spectra of all crystal forms
are listed with assignments in Table 2.
salt (NaHCO3) in a different domain. Any other attempt to
change either the stoichiometric ratio or the grinding condi-
tions yielded a heterogeneous mixture of components. The
13C CPMAS spectrum (Fig. 2d) shows two sets of signals,
which indicate the presence of a HTA and a NaTA molecule
in the unit cell. The carboxylic region of the spectrum is char-
acterized by three peaks at 179.3, 170.1 and 164.5 ppm. The
two former signals are attributed to the COO− (179.3 ppm,
NaTA) and COOH (170.1 ppm, HTA) groups. The latter peak
(at 164.5 ppm) is assigned to the bicarbonate formed during
1
the reaction. Interestingly, H T1 measurements indicate that
HTA and TA− signals belong to the same phase, which is in
agreement with the formation of a salt co-crystal, while the
NaHCO3 resonance belongs to a different domain.30 The
1H MAS spectrum (Fig. 3d) is characterized by two signals in
the hydrogen bond region (14.0 and 12.8 ppm, integral ratio
1 : 2), which are attributed to the COOH of HTA and the NH
1
of HTA and TA−. TGA and H integral values suggest the pres-
ence of one molecule of H2O, which is in agreement with
the elemental analysis. However, it was not possible to estab-
lish whether the water molecule is associated with the salt
co-crystal or NaHCO3.
NaTA·HTA·0.5NaAc·2H2O was obtained as a pure compound
via kneading (EtOH 99%) HTA with NaAc. The carboxylic/
carboxylate region of the 13C solid-state NMR spectrum
(Fig. 2e) reveals the number of independent molecules in the
NaTA·0.5H2O, prepared in a pure form by kneading stoi-
chiometric amounts of HTA and NaOH, shows the splitting
of almost all resonances in the 13C CPMAS spectrum (Fig. 2c),
indicating the presence of two independent molecules in the
unit cell. The splitting is especially marked for atoms C3, C8
and C14. Salification was confirmed by the shift of the car-
boxylic resonance from 175.3 ppm (typical for the cyclic dimer-
ization of the COOH moieties in pure HTA) to 178.3 ppm,
which is characteristic of the COO− group. The possible for-
mation of a 50 : 50 mixture between two polymorphs was
unit cell and thus reveals the presence of one HTA (δCOOH
=
175.9 ppm), one NaTA (δCOO− = 179.0 ppm) and one NaAc
(δCOO− = 184.7 ppm) in a 1 : 1 : 0.5 ratio. The signals due to the
NaAc are only slightly shifted from the pure compound values,
indicating a weak interaction with NaTA·HTA (184.7 and
22.8 ppm in NaTA·HTA·0.5NaAc·2H2O with respect to 183.5
1
and 26.7 ppm in pure NaAc). The H MAS spectrum (Fig. 3e)
shows two resonances in the weak hydrogen bond region at
10.8 and 8.6 ppm (ratio 2 : 1), which are assigned to the COOH
and NH protons and to the other NH proton (overlapped with
aromatic resonances), respectively. By evaluating the 1H inte-
gration, we can assume the presence of two molecules of
water in the structure, as confirmed by TGA and elemental
analysis. The final composition is NaTA·HTA·0.5NaAc·2H2O.
The same 1H T1 value for all of the protons confirms that the
best definition for describing such a structure is a co-crystal
of salts (NaTA·HTA·0.5NaAc·2H2O) because the solid is
composed of a salt co-crystal (NaTA·HTA) and a salt (NaAc).
1H-13C FSLG on- and off-resonance CP HETCOR experiments
were performed in order to obtain a deeper understanding of
the spatial proximities and to complete the assignment of the
1
ruled out by H T1 measurements, which provided the same
1
value for all H atoms. Indeed, according to a simplified but
1
reliable approach,28 the spin diffusion process averages H T1
when protons belong to the same phase or in the case of
homogeneous samples on a nanometer scale.29 On the other
hand, protons with different relaxation times belong to differ-
ent domains with an average linear dimension of greater
than about 100 Å and vice versa.30 In this case, all resonances
possess the same T1 values; thus, two molecules are present
in the asymmetric unit. The presence of half a water mole-
cule in the structure has been predicted by combining data
from TGA (see Fig. S3 in the ESI†) and integral values from
the 1H MAS spectrum (Fig. 3c). This value agrees with the ele-
mental analysis. The NH proton chemical shift suggests the
presence of the NH⋯COO− intramolecular interaction that
characterizes both HTA Forms I and II and all of the crystal
forms presented in this paper.
1
strongly overlapped H resonances. The spectra are reported
in Fig. S7 (ESI†) and Fig. 4, respectively. These experiments
allow intra- and intermolecular heteronuclear correlations to
be obtained.18,31 By performing 2D experiments with short
or long contact times (100 and 2000 μs, respectively), it is pos-
sible to discriminate between short- (single bond) and long-
range spatial proximities, respectively. In particular, the use
of off-resonance CP (LG-CP) allows for a more detailed analy-
sis of long-range spatial 1H–13C contacts. This is achieved
NaTA·HTA·H2O/NaHCO3 was obtained by kneading HTA
and Na2CO3 (2 : 1) with an EtOH/H2O (2 : 1) mixture used as
the mediating solvent. This heterogeneous mixture consists
of a salt co-crystal (concomitant presence of both neutral
HTA molecule and NaTA salt; NaTA·HTA) and an inorganic
8256 | CrystEngComm, 2014, 16, 8252–8262
This journal is © The Royal Society of Chemistry 2014

View Article Online
CrystEngComm
Paper
Table 2 1H, 13C and 15N chemical shifts with assignments for all obtained compounds, pure IM and HTA Forms I and II
HTA
HTA
NaTA·HTA·H2O/
NaTA
Note
Form I
Form II
NaTA·0.5H2O
NaHCO3
NaTA·HTA·0.5NaAc·2H2O
HT Form
IMH-TA·HTA
IM
13C CPMAS
C1
175.4
109.4
175.5
108.2
178.3
118.4
179.3
170.1
114.2
111.4
136.3
132.5
130.5
120.2
117.9
115.7
179.0
175.9
114.0
178.9
111.7
175.7
174.3
113.4
112.2
133.9
133.1
131.3
127.4
126.1
117.6
115.8
—
—
C2
CHAr/CHIM
136.2
135.0
130.1
129.4
127.8
116.6
112.1
132.8sh
125.0
123.6
122.9
117.8
116.6
111.9
134.6
127.0
125.4
122.8
120.2
133.3
128.0
117.1
116.7
133.4
132.5
128.4
127.9
127.1
125.4
121.9
120.2
118.7
128.4
or 127.1
134.7
137.0
127.3
115.7
C9
133.6
125.9
136.3
146.9
138.7
14.6
131.3
137.8
134.0
127.7
139.4;
138.2
150.5
130.4
124.0
120.6
139.9
136.2
148.8
146.9
141.4
—
—
—
—
—
C10
C3
137.0
136.1
150.4
136.2;
135.5
151.7sh
151.1
142.4sh
141.9
19.3
147.3
145.6
141.9
140.5
15.4
147.0
140.5
14.7
C8
138.9
15.6
147.5
C14
16.4
15.6
14.7
13.9
13.2
—
13.4
16.0
Ca
—
—
—
—
164.5b
184.7c
22.8
—
—
—
15N CPMAS
NH
66.5
72.2
70.8
66.1
60.4
—
64.6
71.7
67.6
62.5
59.1sh
NH+IM/NIM
NHIM
—
—
—
—
—
—
—
—
—
—
—
—
153.2d
145.5
220.5e
149.2
NHIM
1H MAS
COOH
12.8
12.9
—
14.0
12.8
14.0
12.8
6.6
10.8
—
15.4
10.4
—
—
7.2
—
NH
8.8sh
8.9sh
9.3
10.8
8.6sh
7.5
6.7
1.3
9.8
6.4
1.7
CHAr/CHIM
CH3
7.1
6.0
1.4
7.0
6.7
1.6
7.5
6.4
1.7
8.5
6.5
1.4
1.5
1.2
0.9
H2O
—
—
—
—
4.2
—
6.7
—
1.3
—
—
—
—
2.8
—
NH+
20.2
19.5
15.4
IM
NHIM
—
—
—
—
—
—
14.7
a
b
c
e
Carbon atoms of either molecules in other phases or of solvents or of co-former molecules. NaHCO3. NaAc. d NH+
.
IM
NIM.
using a Lee–Goldburg (LG) spin-lock (LG-CP), which efficiently
suppresses unwanted 1H–1H spin exchange, leading to fake
correlation peaks that reflect 1H–1H homonuclear rather than
1H–13C heteronuclear dipolar coupling.18,32 All 1H–13C proxim-
ities are depicted in Scheme 3. The 1H-13C FSLG LG-CP
HETCOR experiment (Fig. 4) allowed us to establish the pres-
ence of an intramolecular NH⋯O–C hydrogen bond in the NaTA
molecule (correlation between the 1H NH signal at 8.6 ppm and
the 13C COO− at 179 ppm). The lack of a similar correlation
between signals assigned to the HTA molecule indicates a
nonplanar conformation of the two rings for this molecule. On
the other hand, the HTA OH proton (10.8 ppm) correlates with
the COO− carbon of NaTA (179.0 ppm), indicating that the two
groups are involved in an O–H⋯O hydrogen bond. Other cor-
relations were found for the NH hydrogen of NaTA (8.6 ppm)
with the NaAc methyl and carboxylate carbons (22.8 and
184.7 ppm, respectively). Finally, polarization transfer is observed
from the 1H HTA OH (10.8 ppm) to the NaAc methyl (22.8 ppm).
NaTA HT Form – It is possible to obtain an anhydrous
high temperature form, called the NaTA HT Form, which
This journal is © The Royal Society of Chemistry 2014
CrystEngComm, 2014, 16, 8252–8262 | 8257

View Article Online
Paper
CrystEngComm
Scheme 3 1H–13C correlations between HTA, NaTA and NaAc in
NaTA·HTA·0.5NaAc·2H2O as deduced from 1H-13
C FSLG LG-CP
HETCOR.
HT Form) involves the dehydration of NaTA·0.5H2O (release
of water and formation of NaTA) and the complete salification
of the HTA molecule in NaTA·HTA·H2O/NaHCO3 after the
bicarbonate decomposes to CO2 and H2O. A combination of
13C and 15N CPMAS data made it possible to reveal the pres-
ence of only one NaTA molecule in the unit cell. The shift to
178.9 ppm in the 13C spectrum (Fig. 2f) confirms the presence
of COO− groups, while the small NH shift (from 66.5, HTA, to
64.6 ppm, NaTA) indicates a slight rearrangement of the weak
intramolecular hydrogen bond, which involves the NH. This
is in agreement with the 1H MAS spectrum (Fig. 3f) where the
NH group falls at 9.8 ppm.
The solid-state conversions of NaTA·0.5H2O and
NaTA·HTA·H2O/NaHCO3 to NaTA HT Form were also followed
by VT-XRPD (Fig. S4,† Schemes 1 and 2). The 5–40° 2θ range
was used to evaluate the most significant changes in the pat-
terns. In particular, NaTA·0.5H2O is characterized by five
principal reflections at 9.58, 12.57, 29.76°, which all disap-
pear at 60 °C, and at 15.77 and 19.45°, which vanish at
190 °C. The appearance of three principal new reflections at
14.70, 17.63 and 20.51° at the same temperature leads to
the formation of the NaTA HT Form. On the other hand,
NaTA·HTA·H2O/NaHCO3 is characterized by reflections at
12.42, 17.38 and 18.34° plus the signals attributed to NaHCO3
(at 30.43 and 34.39°). Upon increasing the temperature to
190 °C, the reflection at 17.38° decreases in intensity while all
of the others disappear, giving the characteristic signals of
the NaTA HT Form. The disappearance of the NaHCO3 signals
agrees with the salification of HTA by the bicarbonate.
Fig. 3 1H (400 MHz) MAS spectra of HTA Forms I (a) and II (b),
NaTA·0.5H2O (c), NaTA·HTA·H2O/NaHCO3 (d), NaTA·HTA·0.5NaAc·2H2O
(e), NaTA HT Form (f) and IMH-TA·HTA (g) recorded at 32 kHz.
IMH-TA·HTA was obtained by kneading (EtOH 99%) HTA
and IM in a 2 : 1 ratio. Two TA− and two HTA independent
molecules are present in the unit cell, as shown by the
13C CPMAS spectrum (Fig. 2g), which is characterized by the
splitting of the C1 (175.7 and 174.3 ppm in a 3 : 1 ratio) and
C14 (14.7, 13.9 and 13.2 ppm in a 1 : 2 : 1 ratio) signals. This
is also confirmed by 15N CPMAS (Fig. 5) where four peaks,
caused by the NH group (71.7, 67.6, 62.6 and 59.1sh ppm),
are clearly observable. The number of IM molecules is deter-
mined, in this case, by 15N CPMAS because the IM 13C reso-
nances fall under the aromatic CH of the TA−/HTA molecules.
Fig.
4
1H-13
C
FSLG
LG-CP
HETCOR
spectrum
of
NaTA·HTA·0.5NaAc·2H2O acquired with a contact time of 2000 μs with
relevant assignments. Intramolecular and intermolecular correlations
are indicated in red and light blue, respectively.
is also stable at room temperature, via heating both
NaTA·0.5H2O and NaTA·HTA·H2O/NaHCO3 in a oven at around
180 °C. However, heating NaTA·HTA·0.5NaAc·2H2O results in
a mixture of phases, as observed by 13C CPMAS NMR (see
Fig. S8 in the ESI†). The formation of the new form (NaTA
8258 | CrystEngComm, 2014, 16, 8252–8262
This journal is © The Royal Society of Chemistry 2014

View Article Online
CrystEngComm
Paper
formation. In particular, COOH groups are characterized by ν
and δ around 1680 and 890 cm−1, respectively. On the other
hand, νaCOO− and νsCOO− are at about 1610 and 1385 cm−1
,
respectively. Other differences are found in the NH and aro-
matic CH stretching regions (3500–3050 and 3000–2800 cm−1
,
respectively). The frequencies of some characteristic bands in
the solid-state MIR spectra of the polymorphs and multiple-
component crystals are listed in Table 3.
Differential scanning calorimetry (DSC)
and thermogravimetric (TGA) analysis
All compounds were investigated using TGA and DSC in
order to characterize the thermal behavior and to determine
the water molecule ratios in the samples. The results are
listed in Table 4, while all DSC and TGA figures are displayed
in Fig. S3 in the ESI.†
Heating both NaTA·0.5H2O and NaTA·HTA·H2O/NaHCO3
in an oven at 180 °C yielded the same polymorphic form
(NaTA HT Form). In the case of NaTA·HTA·0.5NaAc·2H2O,
water release (around 160 °C) is followed by the loss (subli-
mation) of one HTA molecule (at 182.6 °C) that leads to mix-
tures of anhydrous phases of NaTA and NaAc, as confirmed
by 13C CPMAS spectra (see Fig. S7 in the ESI†).
It is worth noting that the melting and/or decomposition
point of all samples, except IMH-TA·HTA, are found between
320 and 339 °C (pure HTA sublimates around 212 °C followed
by decomposition), indicating that multicomponent HTA
crystals exhibit significantly improved thermal stability over
single-component forms of HTA.
Fig. 5 15N (40.56 Hz) CPMAS of HTA Form I, IMH-TA·HTA and pure IM
recorded at 9 kHz.
The estimated number of independent IM molecules is two
(
15N δNH = 153.2 and 145.5 ppm). The observed shifts for the
IM NH resonances (from 220.5 and 149.2 ppm to 153.2 and
145.5 for N and NH, respectively) suggest that the IM is pres-
ent in the protonated form (imidazolium, IMH+), as reported
in the literature.33 This implies that the proton transfer
occurs from one HTA (becoming TA−) to the IM (becoming
IMH+). The similarity between COOH and COO− 13C chemical
shifts may suggest that an interaction between these two
groups is able to delocalize the negative charge. It is possible
to note, from 1H MAS (Fig. 3g), that the IMH+ NH+ groups
form very strong hydrogen bonds (20.2 and 19.5 ppm), while
the IMH+ NH and the HTA COOH hydrogen atoms, both
involved in hydrogen bonding, fall at the same chemical shift
1
(15.4 ppm). For HTA, as in the H MAS spectra of the other
samples, the NH hydrogen is located at 10.4 ppm.
Dissolution kinetic tests (DKT)
In order to confirm that the co-crystallization process is a
successful strategy for the improvement of the dissolution
properties of HTA, the in vitro dissolution performance of
pure HTA and various salts and co-crystals were evaluated.
Due to its acidic character, the solubility of HTA depends on
the pH of the aqueous medium. In fact, it is reported that the
solubility of this drug in a pH 1.2 buffer is about 13.6 nM.34 It
is very difficult to detect such extremely low concentrations of
Infrared and Raman spectroscopy
The Raman and FT-IR spectra of all prepared samples are
shown in Fig. S1 and S2 in the ESI.†
The Raman spectra show clear differences in almost all
spectral regions. The most important differences can be found
in the COOH/COO− frequency range (1750–1550 cm−1), reflecting
the differences in hydrogen bonding interactions and salt
Table 3 Wavenumbers (cm−1) of characteristic MIR vibrations with assignments for all multiple-component crystals and two HTA polymorphs
HTA
HTA
NaTA·HTA·H2O/
NaTA
Form I
Form II
NaTA·0.5H2O
NaHCO3
NaTA·HTA·0.5NaAc·2H2O
HT Form
IMH-TA·HTA
NH
3340, s
3326, s
3343, 3255,
broad
3279, m
3312, s
3367.5, s 3243.7,
broad
3290, s
Ar–CH
ν(CO)OH
3083–2870, w
3078–2914, w
2733–2487, w
3078–2859, w,
broad
—
2676–2388, w,
broad
—
3075–2934, w
2740–2591, w
—
3110–2586,
broad
—
2736–2493,
w, road
900, s
1660, s
—
—
δ(CO)OH
νCO(OH)
νaCO(O−)
νsCO(O−)
886, m
1668, s
—
—
—
1610, s
1388, s
—
884, m
1775, w 1727, vs
—
—
—
—
—
—
1689, s
1611, s
1391, 1378,
broad
1608.4, s
1379.2, s
1579.7, s
1385.4, w
—
—
CCl
778, 751,
743, m
756, m
758, m
750, m
752.8, m
752.1, m
This journal is © The Royal Society of Chemistry 2014
CrystEngComm, 2014, 16, 8252–8262 | 8259

View Article Online
Paper
CrystEngComm
Table 4 TGA and DSC data for all achieved samples. TH O: dehydration
temperature, Ttrans: transition temperature, Tfus: melting temperature,
very low in the first 30 minutes of analysis and increases line-
arly (due to the protonation phenomenon) in the following part
of the dissolution test. Fig. 6 shows that the dissolution profile
of IMH-TA·HTA is very similar to the dissolution profile of pure
HTA. Better results are obtained by NaTA·HTA·0.5NaAc·2H2O,
especially in the first part of the analysis (up to 90 minutes),
where HTA solubilizes in higher amounts. In the case of
NaTA·0.5H2O and the NaTA HT Form, instantaneous HTA
solubilization is observed. In fact, HTA solubilizes completely
in the first minute of the dissolution test and high concentra-
tions are maintained throughout. Moreover, no precipitation
phenomenon is observed in the dissolution profiles of these
forms. It is also worth noting that NaTA·HTA·H2O/NaHCO3
is completely dissolved in only 15 minutes of analysis, similar
to the NaTA salts (Fig. 6). Table 5 lists the data extrapolated
after a dissolution time of 1 and 15 minutes. These data
highlight how the active component is liberated from the
multicomponent samples and that this process is strongly
accelerated not only in the salts but also in NaTA·HTA·H2O/
NaHCO3.
Subsequently, FDA bioequivalence limits were adopted
and used for the in vitro solubilization experiments in order
to make a comparison between the performed samples. In
more detail, FDA guidance suggests that two compounds/
formulations can be considered bioequivalent if the ratio of
their area under the curve (AUC) is in the 0.80–1.25 interval.35
This evaluation parameter is proposed for in vivo studies, but
it has been successfully applied to compare AUC values
obtained from in vitro data.21
2
Tdec: decomposition temperature
Sample
TH O
Ttrans
Tfus
Tdec
2
HTA
—
—
210–214
NaTA·0.5H2O
158.5
67.3, 92.3,
191.4a
136.6,
160.6b
182.6c
109.7
320.9
NaTA·HTA·H2O/NaHCO3
—
247.5
339.1
NaTA·HTA·0.5NaAc·2H2O
IMH-TA·HTA
158.0
—
334.1
181–213
a
b
Solid–solid transitions (see the ESI). CO2 and H2O release due to
HTA salification. c Release of HTA.
HTA using the UV method. Therefore, in the present study, a
pH 7.4 buffer (where HTA is more soluble) is used as the dis-
solution medium. The solubility profiles of pure HTA and
selected composites are reported in Fig. 6 (see Fig. S9 in the
ESI† for the complete profile).
Pure HTA shows the typical dissolution profile of a poorly
soluble drug: the amount of the solubilized drug remains
For each sample (pure HTA Form I, HTA salts and HTA
co-crystals), AUC was calculated from time 0 to the last mea-
sured point (AUC0–180) using the classical trapezoidal method.
The AUC of HTA was then used as the reference and indicated
as AUCref. Furthermore, the ratio (indicated as Frel) between
the AUC of each formulation (AUCsample) and the AUCref was
calculated as follows:
AUCsample
Frel 
(I)
AUCref
The associated uncertainty (indicated as dFrel) for all Frel
ratios was calculated using the classic formula for the relative
error of a ratio:36
Fig. 6 Dissolution rate profiles of the samples measured in the first
15 minutes: HTA (red line with red circles), NaTA·0.5H2O (blue line
with blue squares), NaTA HT Form (black line with yellow circles),
IMH-TA·HTA (green line with green triangles), NaTA·HTA·H2O/NaHCO3
(black line with black triangles) and NaTA·HTA·0.5NaAc·2H2O (black
line with white circles).
dFrel  Frel dAUCsample /AUCsample 2  dAUCref /AUCref
2
(II)




Table 5 Dissolution data: amount (mg l−1) of dissolved or released HTA calculated after 1, 15 and 180 minutes with the respective percentage of the
solubilized sample
Sample
After 1′
%
After 15′
%
After 180′
%
HTA
NaTA·0.5H2O
NaTA·HTA·H2O/NaHCO3
NaTA·HTA·0.5NaAc·2H2O
NaTA HT Form
IMH-TA·HTA
0.0
33.0
26.1
3.7
32.1
0.3
0.0
7.3
33.3
32.3
19.7
33.5
5.3
21.8
99.8
96.8
59.1
99.8
15.8
28.5
33.4
33.3
25.7
33.2
24.1
85.4
100
100
77.2
99.4
72.4
98.9
78.4
11.4
96.2
1
8260 | CrystEngComm, 2014, 16, 8252–8262
This journal is © The Royal Society of Chemistry 2014

View Article Online
CrystEngComm
Paper
Table 6 Concentration C (mg l−1), AUC values (mg l−1 h−1) and Frel of all
the samples reported for 180, 70 and 15 minutes
The peculiarities of these HTA derivatives are the
following:
- Heating NaTA·0.5H2O and NaTA·HTA·H2O/NaHCO3
results in the same NaTA HT Form, as confirmed by solid-
state NMR and RAMAN spectroscopy as well as by VT-XRPD
measurements. The transformation of the salt hydrate only
involves a dehydration process while the transformation of
the NaTA·HTA·H2O/NaHCO3 to the NaTA HT Form includes
a salification reaction between the HTA of NaTA·HTA and
NaHCO3, which releases CO2 and H2O.
Sample
AUC0–180
S.D.
Frel
dFrel
HTA
NaTA·0.5H2O
NaTA·HTA·H2O/NaHCO3
NaTA·HTA·0.5NaAc·2H2O
NaTA HT Form
IMH-TA·HTA
48.6
99.8
98.6
54.6
99.7
42.9
3.7
2.0
3.0
13.8
4.0
5.2
—
—
2.1
2.0
1.1
2.1
0.9
0.2
0.5
0.3
0.2
0.1
-
NaTA·HTA·0.5NaAc·2H2O behaves differently upon
where dAUCsample and dAUCref indicate the S.D. of the
AUCsample and AUCref values, respectively. The calculated
Frel dFrel values are listed in Table 6.
heating and loses HTA to yield a NaTA + NaAc mixture, as
characterized by solid-state NMR data.
- All samples show increased thermal stability and
improved solubility behavior (especially the NaTA HT Form,
NaTA·HTA·H2O/NaHCO3 and NaTA·0.5H2O). All of the sam-
ples showed this trend for the first 10 minutes, while
NaTA·HTA·0.5NaAc·2H2O and IMH-TA·HTA were found to be
less soluble than pure HTA but only after 180 minutes.
This investigation has also demonstrated that the crystal
engineering approach makes it possible to change, and in
these cases, improve the macroscopic properties (dissolution
profile and thermal behavior) of an active pharmaceutical
ingredient. Dissolution studies confirm that co-crystallization
is an efficient and innovative technique to improve the bio-
pharmaceutical performance of poorly soluble drugs.
It can be seen from Table 6 that IMH-TA·HTA and
NaTA·HTA·0.5NaAc·2H2O are bioequivalent with pure HTA
in vitro. The other products are probably more bioavailable
than pure HTA (Table 6). These results confirm that co-
crystallization is an efficient and innovative technique to
improve the biopharmaceutical performance of poorly solu-
ble drugs.
Conclusions
Three new multicomponent crystals of tolfenamic acid
(NaTA·HTA·H2O/NaHCO3, NaTA·HTA·0.5NaAc·2H2O and IMH-
TA·HTA) and two sodium salts (NaTA·0.5H2O and NaTA HT
Form) have been prepared using mechanochemical tech-
niques (see Scheme 4). The stoichiometries of the samples
were successfully elucidated using a combination of solid-
state NMR and TGA data due to a lack of single-crystal XRD
analysis. The results are consistent with elemental analysis
data. This study thus highlights the great potential that solid-
state NMR techniques possess when analyzing the composi-
tion and structural characteristics of complex multicomponent
crystals,37 for which it may be difficult or impossible to
acquire single-crystal structure data. Indeed, important struc-
tural information (asymmetric unit, salt or co-crystal forma-
tion, weak interactions, molecule proximities, hydration
degree, etc.) about the samples was provided by 1D and 2D
solid-state NMR experiments.
Acknowledgements
We are indebted to Prof. L. Celi for the elemental analyses
and Dr. A. Castellero and G. Fiore for the VT-XRPD experiments.
Notes and references
1 P. Vishweshwar, J. A. McMahon, J. A. Bis and M. J. Zaworotko,
J. Pharm. Sci., 2006, 95, 499–516.
2 T. Friščić and W. Jones, Cryst. Growth Des., 2009, 9,
1621–1637.
3 J. D. Dunitz, CrystEngComm, 2003, 5, 506–506.
4 D. Braga, E. Dichiarante, G. Palladino, F. Grepioni,
M. R. Chierotti, R. Gobetto and L. Pellegrino, CrystEngComm,
2010, 12, 3534–3536.
5 D. Braga, F. Grepioni, L. Maini, S. Prosperi, R. Gobetto and
M. R. Chierotti, Chem. Commun., 2010, 46, 7715–7717.
6 M. R. Chierotti, K. Gaglioti, R. Gobetto, D. Braga, F. Grepioni
and L. Maini, CrystEngComm, 2013, 15, 7598–7605.
7 O. Eskerod, Pharmacol. Toxicol., 1994, 75, 44–48.
8 P. A. Insel, in Goodman and Gilman's, The Pharmacological
Basis of Therapeutics, McGraw-Hill, New York, 1996, p. 27.
9 (a) T. Kawamori, C. V. Rao, K. Seibert and B. S. Reddy,
Cancer Res., 1998, 58, 409–412; (b) H. M. Tucker, R. E. Rydel,
S. Wright and S. Estus, J. Neurochem., 1998, 71, 506–516.
10 S. Cafaggi, E. Russo, G. Caviglioli, B. Parodi, R. Stefani,
G. Sillo, R. Leardi and G. Bignardi, Eur. J. Pharm. Sci., 2008,
35, 19–29.
Mechanochemistry is a successful method for producing
new solid-state forms. One big advantage that this technique
provides is the fact that it gives pure products that do not
require further purification.
Scheme 4 Representation of the preparation paths (k = kneading,
Δ = heating).
This journal is © The Royal Society of Chemistry 2014
CrystEngComm, 2014, 16, 8252–8262 | 8261

View Article Online
Paper
CrystEngComm
11 (a) S. Romero, B. Escalera and P. Bustamante, Int. J. Pharm.,
1999, 178, 193–202; (b) X. M. Chen, T. L. Li, K. R. Morris and
S. R. Byrn, Mol. Cryst. Liq. Cryst. Sci. Technol., Sect. A, 2002,
381, 121–131.
12 H. Istanbullu, S. Ahmed, M. A. Sheraz and I. urRehman,
BioMed Res. Int., 2013, 2013, 1–8.
25 D. K. Demertzi, D. H. Litina, A. Primikiri, M. Staninska,
C. Kotoglou and M. A. Demertzis, Chem. Biodiversity,
2009, 6, 948–960.
26 F. Atassi and S. R. Byrn, Pharm. Res., 2006, 23, 2405–2412.
27 D. Braga, F. Grepioni, L. Maini, D. Capucci, S. Nanna,
J. Wouters, L. Aerts and L. Quéré, Chem. Commun., 2012, 48,
8219–8221.
28 J. Brus, H. Petrícková and J. Dybal, in Potential and
Limitations of 2D 1H–1H Spin-Exchange CRAMPS Experiments
to Characterize Structures of Organic Solids in Current Devel-
opments in solid-state NMR Spectroscopy, ed. N. Muller and
P. K. Madhu, Springer-Verlag, Wien, 2002, pp. 107–137.
29 I. Ando and T. Asakura, Solid-state NMR of Polymer, Elsevier,
Amsterdam, 1998 (and references therein).
13 V. López-Mejías, J. W. Kampf and A. J. Matzger, J. Am. Chem.
Soc., 2009, 131, 4554–4555.
14 (a) A. Mattei and T. Li, Int. J. Pharm., 2011, 418, 179–186; (b)
A. Mattei and T. Li, Pharm. Res., 2012, 29, 460–470.
15 P. H. Stahl and C. G. Wermuth, Handbook of Pharmaceutical
salts: properties, selection and use, Wiley-VCH, Freiburg im
Breisgau, 2008, pp. 19–40.
16 R. K. Gilpin and W. Zhou, J. Pharm. Biomed. Anal., 2005, 37,
509–515.
17 (a) D. Cinčić, T. Friščić and W. Jones, J. Am. Chem. Soc.,
2008, 130, 7524–7525; (b) T. Friščić, Chem. Soc. Rev.,
2012, 41, 3493–3510.
18 B. J. van Rossum, C. P. de Groot, V. Ladizhansky, S. Vega
and H. J. M. de Groot, J. Am. Chem. Soc., 2000, 122, 3465–3472.
19 (a) O. N. Antzutkin, S. C. Shekar and M. H. Levitt, J. Magn.
Reson., Ser. A, 1995, 115, 7–19; (b) O. N. Antzutkin, Y. K. Lee
and M. H. Levitt, J. Magn. Reson., 1998, 135, 144–155.
20 D. E. Braun, T. Gelbrich, V. Kahlenberg, R. Tessadri,
J. Wieser and U. J. Griesser, J. Pharm. Sci., 2009, 98(6),
2010–2026.
21 (a) D. Hasa, D. Voinovich, B. Perissutti, A. Bonifacio,
M. Grassi, E. Franceschinis, S. Dall'Acqua, M. Speh, J. Plavec
and S. Invernizzi, J. Pharm. Sci., 2011, 100, 915–932; (b)
D. Hasa, B. Perissutti, M. R. Chierotti, R. Gobetto,
I. Grabnar, A. Bonifacio, S. Dall'Acqua, S. Invernizzi and
D. Voinovich, Int. J. Pharm., 2012, 436, 41–57.
22 K. V. Andersen, S. Larsen, B. Alhede and N. Gelting, J. Chem.
Soc., Perkin Trans. 2, 1989, 1443–1447.
23 (a) M. Sardashti and G. E. Maciel, J. Phys. Chem., 1988, 92,
4620–4632; (b) K. M. N. Burgess, M. N. Kevin, X. Yang,
C. M. Leclerc and D. L. Bryce, J. Phys. Chem. A, 2013, 117,
6561–6570.
30 M. Geppi, G. Mollica, S. Borsacchi and C. A. Veracini, Appl.
Spectrosc. Rev., 2008, 43, 202–302.
31 (a) M. R. Chierotti, L. Ferrero, N. Garino, R. Gobetto,
L. Pellegrino, D. Braga, F. Grepioni and L. Maini, Chem. –
Eur. J., 2010, 16, 4347–4358; (b) M. R. Chierotti, R. Gobetto,
L. Pellegrino, L. Milone and P. Venturello, Cryst. Growth
Des., 2008, 8, 1454–1457.
32 (a) J. Brus and A. Jegorov, J. Phys. Chem. A, 2004, 108,
3955–3964; (b) M. R. Chierotti and R. Gobetto,
CrystEngComm, 2013, 15, 8599–8612.
33 (a) R. M. Claramunt, C. López, M. A. García, G. S. Denisov,
I. Alkorta and J. Elguero, New J. Chem., 2003, 27, 734–742;
(b) S. Li and M. Hong, J. Am. Chem. Soc., 2011, 133,
1534–1544.
34 C. A. S. Bergström, C. M. Wassvik, K. Johansson and
I. Hubatsch, J. Med. Chem., 2007, 50, 5858–5862.
35 Food and Drug Administration, Guidance for Industry,
Bioavailability and Bioequivalence Studies for Orally
Administered Drug Products - General Considerations, 2003.
36 J. R. Taylor, An Introduction to Error Analysis: The Study of
Uncertainties in Physical Measurements, University Science
Books, Sausalito, CA, USA, 1997, pp. 1–464.
37 (a) R. K. Harris, NMR crystallography, John Wiley & Sons,
Chichester, 2009; (b) R. K. Harris, Analyst, 2006, 131,
351–373; (c) E. Benedicte and L. Emsley, J. Am. Chem. Soc.,
2005, 127, 9140–9146.
24 L. Fábián, N. Hamill and K. S. Eccles, Cryst. Growth Des.,
2011, 11, 3522–3528.
8262 | CrystEngComm, 2014, 16, 8252–8262
This journal is © The Royal Society of Chemistry 2014