Multicomponent solids of lamotrigine with some selected coformers and their characterization by thermoanalytical, spectroscopic and X-ray diffraction methods

Article abstract of DOI:10.1039/c1ce05458a

The present study investigates the structural and pharmaceutical properties of different multicomponent crystalline forms of lamotrigine (LTG) with some pharmaceutically acceptable coformers viz. nicotinamide (1), acetamide (2), acetic acid (3), 4-hydroxy

Full text of DOI:10.1039/c1ce05458a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
CrystEngComm
Cite this: CrystEngComm, 2011, 13, 6271
www.rsc.org/crystengcomm
PAPER
Multicomponent solids of lamotrigine with some selected coformers and their
characterization by thermoanalytical, spectroscopic and X-ray diffraction
methods†
Renu Chadha, a Anupam Saini,a Poonam Arora,a Dharamvir Singh Jain,b Archi Dasguptac and T. N. Guru Rowc
*
Received 18th April 2011, Accepted 19th July 2011
DOI: 10.1039/c1ce05458a
The present study investigates the structural and pharmaceutical properties of different
multicomponent crystalline forms of lamotrigine (LTG) with some pharmaceutically acceptable
coformers viz. nicotinamide (1), acetamide (2), acetic acid (3), 4-hydroxy-benzoic acid (4) and saccharin
(5). The structurally homogeneous phases were characterized in the solid state by DSC/TGA, FT-IR
and XRD (powder and single crystal structure analysis) as well as in the solution phase. Forms 1 and 2
were found to be cocrystal hydrate and cocrystal, respectively, while in forms 3, 4 and 5, proton transfer
was observed from coformer to drug. The enthalpy of formation of multicomponent crystals from their
components was determined from the enthalpy of solution of the cocrystals and the components
separately. Higher exothermic values of the enthalpy of formation for molecular complexes 3, 4 and 5
suggest these to be more stable than 1 and 2. The solubility was measured in water as well as in
phosphate buffers of varying pH. The salt solvate 3 exhibited the highest solubility of the drug in water
as well as in buffers over the pH range 7–3 while the cocrystal hydrate 1 showed the maximum solubility
in a buffer of pH 2. A significant lowering of the dosage profile of LTG was observed for 1, 3 and 5 in
the animal activity studies on mice.
The multicomponent approach that includes salts as well as co-
Introduction
crystals results in a set of structural variations of the same API. Salt
Considerable time and funds are required for the research and
formation indeed is a widely accepted approach to modify the
development and formulation of any API that is to be delivered in
physical properties of APIs,12–17 however, pharmaceutical cocrys-
a solid form.1 Unfortunately, some potentially useful APIs
tallization is a relatively recent technology which offers an alter-
administered as solids may never realize their maximum potential
native platform to improve the physicochemical properties of
due to unsatisfactory bioavailability, limited by their solubility.2
active pharmaceutical ingredients (APIs), such as the melting point,
The enhanced solubility critically impacts the pharmacokinetic
solubility, stability, and dissolution rate.18–26 Pharmaceutical coc-
profile of orally delivered APIs resulting in better absorption in the
rystals, multiple component solids, also represent a broad patent
gastrointestinal tract (GIT) and reduced dosage-level require-
space since they are clearly new chemical entities, and their design
ments.3 Thus, a number of approaches have been pursued for
and preparation involve several elements of non-obviousness and
optimizing the solubility of APIs such as preparation of solid
they generally have novel and useful properties.27
dispersions,4 micronization,5 the use of surfactants,6 polymorphs,7
Lamotrigine (LTG) [3,5-diamino-6-(2,3-dichlorophenyl)-
self-emulsifying formulations,8 inclusion complexation with cyclo-
1,2,4-triazine] is an anticonvulsant drug having poor solubility in
dextrins,9 nanocrystals10 and the use of multicomponent molecular
water (0.17 mg mLꢀ1 at 25 ꢁC)28 which limits its absorption and
crystals.11 Recently there is heightened interest and awareness of
dissolution rate and thusdelaysthe onsetof action.29 Thus, there is
the need to diversify the range of crystal forms exhibited by APIs.
a need to develop alternative forms of LTG with improved solu-
bility which can significantly enhance the oral absorption of this
drug in GIT. The LTG framework is comprised of four acidic
aUniversity Institute of Pharmaceutical Sciences, Panjab University,
Chandigarh, 160014, India. E-mail: renukchadha@rediffmail.com; Tel:
amino hydrogen bond donors along with two basic hydrogen
bond acceptors i.e. amino-pyridine nitrogen atoms giving rise to
a variety of hydrogen bonding donor/acceptor sites for an
approaching coformer to bind, thus making it a potential target
for both cocrystal and salt formation. These features of LTG
along with the availability of a vast number of coformers of
GRAS (Generally Regarded AsSafe) status30 for cocrystallization
+91-9316015096
bDepartment of Chemistry, Panjab University, Chandigarh, 160014, India
cSolid State and Structural Chemistry Unit (SSCU), Indian Institute of
Science, Bangalore, 560012, India
† CCDC reference numbers 791931, 819591 and 819592. For
crystallographic data in CIF or other electronic format see DOI:
10.1039/c1ce05458a
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6271–6284 | 6271

View Article Online
persuaded us to explore this multicomponent approach in an
attempt to enhance the solubility of this anticonvulsant drug.
Various groups of researchers have worked on LTG crystal forms
and have reported solvates,31–37 salts38–42 and cocrystals42 of this
drug. One of the recent reports is by Galcera and Molins41 on LTG
salts with four different counter-ions such as succinic acid,
fumaric acid, DL-tartaric acid and saccharin, of which only sac-
charinate and DL-hemitartrate dimethylsulfoxide solvates
exhibited higher aqueous solubility than the pure drug. However,
the use of dimethylsulfoxide as a solvent is not safe for human
consumption. Beside this, Cheney et al.42 reported cocrystals of
LTG with methylparaben and nicotinamide and salts with
saccharin, adipic acid, malic acid and nicotinic acid in 2010. The
solubility and pharmacokinetic studies on only a few of these
forms have been reported and revealed that only the saccharinate
salt of LTG exhibited substantial improvement in these targeted
properties. In the present work, multicomponent forms of LTG
with nicotinamide, acetamide, acetic acid, 4-hydroxy-benzoic acid
and saccharin are reported. Most of these coformers have been
selected based on their pKa value with an intention to obtain
intermolecular hydrogen bonded complexes with LTG. Except
for the cocrystal hydrate of LTG with nicotinamide and the salt
of LTG with saccharin, none of the other forms described in this
work have been reported earlier. Despite the fact that LTG-
saccharinate had already been reported41 this coformer was
selected in the present work because of its additional advantages
such as a sweetening agent, low toxicity and appreciable water
solubility.15 Although the first characterization of LTG-nico-
tinamide monohydrate and LTG-saccharin salt has already been
reported by Cheney et al.,42 we incorporated their findings in the
present work for comparison and found our results to be in
good agreement with their reports. In addition, a few additional
parameters such as the enthalpy of solution as well as stability
studies which are missing in the literature have been included in
the present study. The characterization was done using differ-
ential scanning calorimetry (DSC), thermogravimetric analysis
(TGA), Fourier transform infra-red spectroscopy (FT-IR) and
X-ray diffraction (XRD) analysis, which confirmed that two of
these forms are cocrystals and the remaining three are salts.
Beside this, the technique of solution calorimetry has also been
utilized to give an insight into the breaking of the lattice of these
multicomponent crystals upon dissolution. Further, the solu-
bility, stability and animal activity studies using the mouse
maximal electroshock (MES) model43 were performed on these
multicomponent forms.
50 ꢁC for 2 hours. The clear solution was evaporated rapidly
under vacuum, and then the crystalline solid obtained was
scraped from the walls of the flask and stored in an airtight glass
vial. The attempts to produce single crystals of 1 as described by
Cheney et al.42 resulted in the re-crystallization of the starting
components individually, may be due to different laboratory
temperature conditions.
LTG-acetamide (2). The co-crystallization was performed by
the reaction crystallization method by adding LTG (0.2 g,
0.78 mmol) to 4 mL of presaturated isobutanol solution of acet-
amide and keeping at room temperature for slow evaporation.
The crystals of compound 2 of size suitable for single crystal XRD
were obtained, filtered, air dried and stored in airtight glass vials.
LTG-acetic acid (3). Single crystals of 3 were obtained by
dissolving LTG in an excess of acetic acid with heat and evap-
orating slowly.
LTG-4-hydroxy-benzoic acid (4). LTG (0.128 g, 0.49 mmol)
was added to 3 mL of ethanol and dissolved by warming. To this
clear solution, an ethanolic solution of 4-hydroxybenzoic acid
(0.069 g, 0.49 mmol) was added producing instantaneous
precipitation of a fine powder of compound 4 which was filtered,
air dried and stored in a vial. Single crystals of 4 were obtained by
slow evaporation of the filtrate.
LTG-saccharin (5). LTG (0.758 g, 2.9 mmol) was added to 50
mL solution of saccharin (0.549 g, 2.9 mmol) in acetonitrile and
dissolved with slight warming until the dissolution was complete.
The clear solution was allowed to slowly evaporate at room
temperature. Single crystals of 5 were obtained within a few
hours.
Characterization of crystals in the solid state
Differential scanning calorimetry (DSC). Differential scanning
calorimetry of all the samples was conducted using a DSC Q20
(TA Instruments, USA). The samples (3–5 mg) were placed in
sealed non-hermetic aluminium pans and were scanned at a rate
of 5 C minꢀ1 in the range of 25–300 C under a dry nitrogen
atmosphere (flow rate 50 mL minꢀ1). The data were managed by
TA Q series Advantage software (Universal analysis 2000).
ꢁ
ꢁ
Thermogravimetric analysis (TGA). TGA was performed using
a Mettler Toledo TGA/SDTA 851e instrument. Approximately 5
mg sample was heated from 25 to 300 ꢁC in an open alumina pan
at the rate of 10 ꢁC minꢀ1 under nitrogen purge at a flow rate of
50 mL minꢀ1. The data were managed by STAR software (9.00).
Experimental
LTG was obtained as complimentary sample from Rantus
Pharma Pvt. Ltd. (India). The guest compounds and crystalli-
zation solvents were of AR grade and purchased from various
commercial suppliers. All of these were used as received.
Hot stage microscopy. Melting points and physical changes
were visually examined at 50ꢂ magnification by hot stage
microscopy. The study was carried out using an optical/polarized
hot stage microscope (Leica DMLP, Leica, Germany) equipped
with a controlled heating and cooling stage (LTS350, Linkam)
and an imaging system (VTO 232, JVC-Digital camera and
Linksys 32 imaging software, Linkam, England). The powder
Sample preparation
LTG-nicotinamide monohydrate (1). Compound
1 was
prepared by the solvent evaporation method. A stoichiometric
amount of LTG (0.256 g, 1 mmol) and nicotinamide (0.122 g,
1 mmol) were added to 5 mL methanol followed by stirring at
ꢁ
ꢁ
sample was mounted in air and heated from 25 C to 230 C at
a rate of 5 C minꢀ1
ꢁ
.
6272 | CrystEngComm, 2011, 13, 6271–6284
This journal is ª The Royal Society of Chemistry 2011

View Article Online
X-Ray powder diffraction (XRPD). XRPD patterns were
liquid samples like acetic acid, the enthalpy of solution in buffer
was determined using the titration mode of the Micro Reaction
Calorimeter. The reference and sample vials filled with an equal
volume of buffer were placed in a calorimetric block set at 37 ꢁC.
A 100 ml syringe loaded with acetic acid was mounted on the
sample vial. After baseline stabilization, 50 ml of the acid was
injected into the sample vial and the heat output was recorded
and integrated to calculate the enthalpy of solution.
collected using an X’Pert PRO diffractometer system (Pan-
ꢀ
alytical, Netherlands) with a Cu Ka radiation (1.54060 A). The
tube voltage and current were set at 45 kV and 40 mA respec-
tively. The divergence slit and anti-scattering slit settings were set
at 0.48ꢁ during illumination on the 10 mm sample size. Each
sample was packed in an aluminium sample holder and measured
by a continuous scan between 5 and 50ꢁ in 2q with a step size of
0.017ꢁ. The experimental XRPD patterns were refined using
X’Pert High Score software.
Equilibrium solubility studies. Solubility of the LTG free base
and compounds 1–5 has been determined in water at various time
points to ensure that the solution has reached equilibrium. For
this study, the starting solids were sieved using a Gilson mesh
sieve to provide samples with an approximate particle size of 150
mm. In each experiment, a flask containing 50 mL of water was
equilibrated at 37 ꢁC in a constant temperature bath. An excess
of solid phase (ca. 50 mg) was added to the flask and the resulting
slurry was shaken at 200 rpm. An aliquot of slurry was with-
drawn at multiple time points, filtered through a 0.45 mm
membrane filter, diluted suitably and the concentration was
determined spectroscopically by measuring the absorbance at
305 nm with a Lambda 25 UV/VIS spectrometer. The extinction
coefficient of LTG was obtained through calibration experiments
carried out in pure water (3305 ¼ 6.8 ꢂ 103 Mꢀ1 cmꢀ1). For
experiments involving compounds 1–5, the absorbance values
were related to solution concentrations using their respective
calibration curves in their respective medium. This was possible
since none of the coformers absorb at 305 nm and therefore did
not interfere with determination of the concentration of LTG.
After the last aliquot was collected, the remaining solids were
filtered, air dried and analyzed by XRPD. The equilibrium
solubility of LTG free base and compounds 1–5 has also been
determined in phosphate buffers of pH 7, 5, 3 and 2. As the pKa
of LTG is 5.7, there is a possibility that after dissolution it may be
present in the solution in either its protonated or unprotonated
form depending upon the pH of the solution. The concentration
was determined at l corresponding to an isosbestic point of LTG
(288 nm).
Fourier transform-infrared spectroscopy (FT-IR). A Spectrum
RX I FT-IR spectrometer (Perkin Elmer, UK) was employed in
the KBr diffuse-reflectance mode (sample concentration 2 mg in
20 mg of KBr) for collecting the IR spectra of samples. The
spectra were measured over the range of 4000–400 cmꢀ1. Data
were analyzed using Spectrum software.
Single crystal X-ray diffraction. The X-ray diffraction dataset
for compound 2 was collected on an Oxford Xcalibur (Mova)
diffractometer44 equipped with an EOS CCD detector using Mo-
ꢀ
Ka radiation (l ¼ 0.71073 A) at room temperature. X-Ray
diffraction datasets for compounds 3 and 4 were collected on
a Bruker AXS Kappa Apex CCD diffractometer using Mo-Ka
radiation at 90 K. All structures were solved by direct methods
using SHELXS-97 and refined against F2 using SHELXL-97.45
The hydrogen atoms in carboxylic acid groups and amide groups
associated with the formation of either a salt or a cocrystal of
lamotrigine were located based on the difference Fourier map
and were refined isotropically. All other hydrogen atoms were
placed geometrically and refined with an isotropic displacement
parameter fixed at 1.2 times Uq of the atoms to which they were
attached. The WINGX package (version1.70.01)46 was used for
refinement and production of data tables and ORTEP-347 for
structural visualization. All ORTEP representations were made
using POV-Ray48 showing ellipsoids at the 50% probability level.
Analysis of the H-bonding and other non-covalent interactions
was carried out using PARST95 and PLATON49 for all the
structures. Packing diagrams were generated using
Mercury-2.2.50
Stability studies. Accurately weighed samples (approximately
100 mg) placed in looseꢁly capped glass vials were kept in
a stability chamber at 25 C/60% RH for one month and then
analyzed by XRPD.
Analysis in the solution phase
Solution calorimetry studies. The enthalpy of solution of the
drug, coformers and the multi-component forms was determined
using a Micro Reaction Calorimeter (a power compensation
system) obtained from Thermal Hazards Technology (UK) in
a phosphate buffer at pH 7.0 and 37 ꢁC. Two experimental vials
(reference and sample) filled with an equal volume of buffer were
placed in a calorimetric block. A solid sample of about 1 to 2 mg,
accurately weighed (Sartorius Model CP225D), was loaded into
a cylindrical glass tube (solid sample insert) covered with par-
afilm on one side and submerged into the sample vial. After
baseline stabilization at 37 ꢁC (ꢃ0.0005 ꢁC), the sample was
released into the sample vial by means of a plunger. The heat
output was recorded and integrated to calculate the enthalpy of
solution. The precision of any individual measurement was better
than 0.02 kJ molꢀ1, for three consecutive experiments and agreed
with the standard value within ꢃ0.03 kJ molꢀ1. In the case of
Animal activity studies. Male mice (Balb/C, 20–30 g) were used.
Animals were weighed and placed in standard cages with free
access to food and tap water. After 7 days of adaptation to
laboratory conditions, the animals were randomly assigned to
experimental groups each comprising 6 mice. LTG free base and
compounds 1–5 were prepared as suspensions in 0.5% carboxy-
methylcellulose in saline and administered via oral gavage. After
15 minutes of drug dosing, electroconvulsions were produced by
current (fixed current 30 mA, stimulus duration 0.2 s) delivered
to saline-wetted eyes via corneal electrodes from an electroshock
apparatus (IMCORP, India). The criterion for the occurrence of
seizure activity was the tonic hind limb extension (HLE i.e., the
hind limbs of animals outstretched 180ꢁ to the plane of the body
axis). The protective activity of LTG free base and its multi-
component forms was determined as the median effective dose
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6271–6284 | 6273

View Article Online
(ED50 value in mg kgꢀ1) against maximal electroshock (MES)
induced seizures. Sufficient animals were tested over the range of
four different doses of each compound (2.5, 5.0, 7.5 and 10.0 mg
kgꢀ1 of LTG or its equivalent) to provide data for calculation of
ED50 values. The data were processed using the Microsoft Office
Excel 2007 software package. This experiment is approved by the
Institutional Animal Ethical Issue Committee and has been
conducted according to Indian National Science Academy
(INSA) guidelines for use and care of experimental animals.
Results and discussion
Scheme 1 Molecular structures of Lamotrigine and coformers.
LTG has the ability to form both salts and cocrystals due to its
relatively basic nature (pKa ¼ 5.7). The pKa difference between
the drug and coformer (DpKa) can give an idea about the
formation of salt or cocrystal. In general, a higher DpKa (greater
than 3) will result in salt formation while a lower DpKa (less than
0) will almost exclusively result in cocrystal formation.51
However, in the range 0 to 3, the complexes between acids and
bases can be salts or cocrystals or may contain shared protons or
mixed ionization states that cannot be assigned to either cate-
gory. The pKa values of coformers involved in this study are
summarized in Table 1. Based on the DpKa values, acetamide has
a high probability of forming a cocrystal, saccharin is expected to
give a salt only, while the remaining three coformers may result
in cocrystal or salt formation with LTG. Thus, the forms
obtained in these cases have to be ascertained by various tech-
niques. Melting points and shapes of crystals of various phases
Table 1 pKa Values of LTG and coformers and the resulting DpKa
values for the LTG multicomponent forms
DpKa (pKa base
ꢀ pKa acid)
Compound
pKa
5.70
3.35
17.00
4.80
LTG (free base)
Nicotinamide
Acetamide
Acetic acid
4-Hydroxy-benzoic acid
Saccharin
—
2.35
ꢀ11.3
0.90
1.22
3.38
Fig. 1 Optical micrographs of compounds 1–5.
4.48
2.32
are given in Table 2. The difference in melting points of these
compounds and their individual components strongly indicates
the formation of phases that are unique to their starting
components. These solid phases, formed with nicotinamide 1,
acetamide 2, acetic acid 3, 4-hydroxy-benzoic acid (4-HBA) 4
and saccharin 5, were characterized in both the solid state and the
solution form. The molecular structures of LTG as well as all the
coformers are given in Scheme 1 and the morphology of all these
forms is shown in Fig. 1.
Table
compounds 1–5
2 Melting point and morphology of LTG free base and
Compound
Mp/ꢁC
Morphology of crystals
LTG
1
Nicotinamide
216–217
169–173
128–131a
159–160
78–81a
149–170
16.64 (mp)a,
118 (bp)a
232–234
213–216a
Rectangular blocks
Irregular shape
—
Rectangular blocks
—
Rectangular blocks
—
2
Acetamide
3
Acetic acid
Characterization in the solid state
4
Rhombus shape
—
Thermal analysis utilizing DSC, TGA and hot stage micros-
copy. DSC and TGA scans of LTG and compounds 1–5 are
shown in Fig. 2 and 3 respectively. The DSC theꢁrmogram of
LTG showed a single melting endotherm at 217.1 C. A broad
endotherm was observed in the DSC thermogram of 1 with an
onset at 80.7 ꢁC followed by a sharp endothermic peak at
4-Hydroxy-
benzoic acid
5
255–257
228–229a
Rectangular blocks
—
Saccharin
a
Obtained from Merck Index.
6274 | CrystEngComm, 2011, 13, 6271–6284
This journal is ª The Royal Society of Chemistry 2011

View Article Online
173.9 ꢁC which is different from melting peaks of both LTG and
ꢁ
nicotinamide (133.2 C). The DSC scan of a binary mixture of
the starting components exhibited two different endotherms
corresponding to their melting thus ruling out the probability of
any eutectic formation. A TGA scan of 1 showed a weight loss of
3.58% (in the range of 80–100 ꢁC) which correlates to the loss of a
single water molecule (theoretical weight loss of 4.54%). This
suggests that 1 is a monohydrate. This is further confirmed by
comparison of our results with those of Cheney et al.42 which
showed a slight variation in the desolvation peak (onset at
90.0 ꢁC42) whereas the melting endotherm (174.8 ꢁC42) is
approximately the same. The difference in the onset of the des-
olvation peak may be due to the difference in the rate of scan-
ning. Similarly the % loss in TGA is also in good agreement
(3.74%42). These thermoanalytical results supported by those of
Cheney et al.42 suggest 1 to be the monohydrate cocrystal of LTG
with nicotinamide.
Two major endothermic events were observed in the DSC scan
of 2. A broad endotherm at 160.5 ꢁC depicting the melting of
cocrystal is associated with a weight loss in the TGA. This is
probably due to sublimation of the coformer (acetamide) from
the melt as the TGA scan of pure acetamide also shows a similar
weight loss in its melting range. Interestingly, this first broad
peak is followed by a sharp endothermic peak at 216.9 ꢁC
depicting the melting of pure drug which might have appeared
due to recrystallization of free base after the loss of coformer
from the cocrystal melt. To have an insight into the thermal
events taking place during the melting of this cocrystal resulting
Fig. 2 DSC thermograms of LTG free base and its multicomponent
forms 1–5.
Fig. 4 Hot stage microscopy micrographs of (A) cocrystal 2 and (B) salt
solvate 3.
Fig. 3 TGA curves of compounds 1, 2, 3 and acetamide.
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6271–6284 | 6275

View Article Online
in the emergence of drug, hot stage microscopy was performed.
coformers indicating the generation of new solid phases. These
unique peaks corresponding to each of the molecular complexes
are shown in Fig. 5A. The experimental powder pattern of
compound 1 was compared with the simulated XRPD pattern
generated from the previously determined single crystal data of
both LTG-nicotinamide cocrystal monohydrate as well as the
LTG-nicotinamide cocrystal (anhydrous).42 It is clear from
Fig. 5B that the XRPD pattern of compound 1 correlates nicely
to the simulated pattern of LTG-nicotinamide cocrystal mon-
ohydrate rather than its anhydrous form, thus, confirming 1 to
be identical to the known hydrate form. Similarly, the powder
pattern of compound 5 is compared with the simulated pattern
derived from the known crystal structure of LTG-saccharin
salt42 and the two are found to be identical (Fig. 5B), thus,
confirming compound 5 to be LTG-saccharin salt. No prior
data are available for comparison of XRPD patterns of
compounds 2–4. Although formation of new solid phases in 2, 3
and 4 has been identified by a combination of DSC/TGA and
XRPD analysis, the transfer of protons in these solids was
ascertained only by FTIR spectroscopy and single crystal X-ray
diffraction studies.
As can be seꢁen in Fig. 4A, there is no change in the crystals of 2
ꢁ
up to 132.0 C. The melting of the cocrystal began at 134.3 C
and the recrystallization of LTG from this melt started at
155.0 ꢁC. These crystals of LTG continued to grow up to
a temperature of 185.0 ꢁC and finally started melting at 219.0 ꢁC.
The occurrence of both the endothermic event due to melting of
the cocrystal and the exothermic event due to recrystallization of
the pure drug from this melt in the same temperature range (134–
185 ꢁC) resulted in an overall broad endothermic peak in this
region in the DSC thermogram.
The DSC scan of 3 showed two broad endotherms followed by
a sharp melting endotherm. The first two broad endotherms are
accompanied by a weight loss in TGA in temperature ranges of
75–110 ꢁC (26.9%) and 115–170 ꢁC (13.7%) respectively. The two
step weight loss in TGA is attributed to evaporation of overall
three molecules of aceticꢁacid released in two steps. The broad
peak in the DSC at 102 C is attributed to desolvation and the
stoichiometric calculation (theoretical weight loss of 27.5%)
shows the release of two molecules of acetic acid. The second
broad endotherm at higher temperature (170 ꢁC) indicates the
phase transition leading to the melting of the multicomponent
crystal with simultaneous release of one molecule of acetic acid
from the crystal lattice as suggested by the theoretical weight loss
of 13.8%. The emergence of sharp melting endotherm at 217.0 ꢁC
suggests the transformation of the original crystal lattice of 3
after the evaporation of three acetic acid molecules to pure LTG.
This was supported by various events observed during hot stage
microscopy. Fig. 4B shows that the crystals became opaque in
the temperature range of 75–102 ꢁC suggesting the loss of acetic
acid molecules present as solvent in these crystals as indicated by
DSC and TGA. These desolvated crystals started melting at
FT-IR. FTIR spectroscopy is an excellent technique to char-
acterize and distinguish cocrystals from salts, especially when
a carboxylic acid is used as a coformer. LTG shows an IR
absorption frequency for a primary amine N–H stretch at 3448
and 3316 cmꢀ1, a C–H aromatic stretch at 3209 cmꢀ1 and
a primary amine N–H bend at 1620 cmꢀ1 (Fig. 6). In the FTIR
spectra of 1, the amino N–H stretch of LTG shifted to 3413 and
3325 cmꢀ1 while the N–H and C]O stretches of the amide group
of nicotinamide at 3160 and 1680 cmꢀ1 shifted to 3145 and 1692
cmꢀ1, respectively suggesting that both these components are
participating in certain kinds of interactions. The IR spectrum of
compound 1 is in good agreement with that reported by Cheney
et al.42 By combining these results with DSC/TGA and XRPD
findings, compound 1 can be identified as a cocrystal hydrate of
LTG with nicotinamide. In the FTIR spectra of 2, the N–H
stretch of the amine base of LTG at 3448 cmꢀ1 and the amide
function of acetamide at 3347 cmꢀ1 shifted to 3424 and 3327 cmꢀ1
respectively while the C]O stretch of the acetamide shifted from
1676 to 1650 cmꢀ1 demonstrating interactions between these
groups of LTG and acetamide. Similarly, these changes were also
observed in multicomponent crystal 3. The coformer acetic acid
being an aliphatic carboxylic acid absorbs at 1715 cmꢀ1 corre-
sponding to the C]O stretch. This absorption moves to a lower
wavenumber if the oxygen atom attached to carbonyl is ionized
and this is evident in the IR spectra of 3 which exhibits a band at
1687 cmꢀ1 corresponding to carboxylate stretching in addition to
the significant shift of the NH stretch of the amine base of LTG
suggesting it to be the LTG salt. Compounds 4 and 5 also showed
characteristic broad bands between 3300 and 2000 cmꢀ1 corre-
sponding to N–H+ signals of amine hydrogen bonded salts
(absent in spectra of drug) suggesting these to be salts of LTG.
Further, the shifting of carboxylic acid C]O stretch from 1677
and 1721 cmꢀ1 to lower wavenumbers of 1664 and 1678 cmꢀ1 in 4
and 5, respectively clearly indicates salt formation. In addition,
the IR spectrum of 5 was found to be identical to that of LTG
saccharinate provided by M. L. Cheney et al.,52 thus confirming
the two to be the same.
ꢁ
142.0 C. A sequence of images shows simultaneous recrystalli-
zation of LTG from the melt. Upon further heating, the size of
melt droplets decreased while the crystals of LTG continued
growing up to 175.0 ꢁC with no further change until final melting
ꢁ
at 220.3 C. Thus, we can say that the exothermic peak due to
recrystallization of LTG has merged with the broad endothermic
peak due to melting of desolvated crystals of 3 in the DSC scan.
Therefore, DSC results suggest 3 to be a multicomponent crystal
of LTG and acetic acid with 1 : 3 stoichiometry.
DSC scans of 4 and 5 showed single, sharp melting endotherms
at 234.5 and 257.3 ꢁC respectively, which are at a higher
temperature than LTG as well as the respective coformers (4-
HBA ¼ 214.9 ꢁC and saccharin ¼ 230.5 ꢁC). These also indicate
the absence of any eutectic formation as eutectics melt at
temperatures between those of the melting points of the starting
components. The melting peak of 5 has been found to be in good
agreement with that of theꢁLTG-saccharin salt reported by
Galcera and Molins41 (256.0 C), however, there is slight varia-
tion from that reported by Cheney et al.42 (252.5 C). Further,
ꢁ
TGA of both 4 and 5 did not reveal any weight loss indicating the
existence of anhydrous and stable phases unique to their starting
components. Thus, DSC/TGA has shown the existence of
multicomponent crystals of LTG with various coformers which
are present in different stoichiometry.
XRPD analysis. Compounds 1–5 displayed unique crystalline
XRPD patterns in comparison to LTG and their respective
6276 | CrystEngComm, 2011, 13, 6271–6284
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Crystal structure analysis. The single crystal structures of 1
(LTG-nicotinamide hydrate) and 5 (LTG-saccharin salt) have
been published previously by M. L. Cheney et al.42 The crystal
structure of LTG42 is built from hydrogen bonded dimeric units
of the drug molecule showing two dominant supramolecular
synthon motifs, the aminopyridine dimer (motif 1) and the
amine–aromatic nitrogen synthon (motif 2). The LTG dimer
homosynthon is conserved in complexes 2–4 as well except for
slight deviations in the geometry of LTG dimer which is formed
through three different parts of the LTG molecule, N3 and N2
Fig. 5 (A) XRPD patterns of LTG free base, coformers and their multicomponent forms 1–5. (B) Experimental XRPD patterns (red) of compounds 1–
5 compared with their Rietveld refined simulated XRPD patterns (blue/ black).
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6271–6284 | 6277

View Article Online
drug. However, in all these three complexes (2–4), the incorpo-
ration of complementary coformer breaks motif 2. Fig. 7 repre-
sents the four different hydrogen bond motifs. Crystallographic
data, Dc–o distances (i.e. carbon oxygen distances in the acid
carbonyl group) and hydrogen bond geometries of compounds
2–4 are given in Tables 3, 4 and 5 respectively. The simulated
XRPD patterns of compounds 2–4 generated from their single
crystal data are compared with their experimental XRPD
patterns and found to be identical (Fig. 5B).
(2) LTG-acetamide cocrystal (1 : 1). Complex 2 crystallizes in
ꢁ
the space group P1 with the asymmetric unit consisting of one
molecule of LTG and one molecule of acetamide (Fig. 8). Both
LTG and acetamide remain neutral in the crystal lattice of 2 and
hence complex 2 is a cocrystal. The basic supramolecular unit is
comprised of a LTG dimer homosynthon (N3–H3A/N4, N/N
ꢁ
ꢀ
ꢀ
3.070(3) A, N/H 2.256(2) A, N–H/N 157.9(2) ) and an acet-
ꢀ
amide dimer homosynthon (N6–ꢁH6B/O1, N/O 2.962(5) A,
ꢀ
O/H 2.25(5) A, N–H/O 170(6) ). The two dimeric motifs are
linked together through the N–H/O hydrogen bond (N1–
ꢀ
ꢀ
H1B/O1, N/O 2.870(4) A, O/H 2.146(3) A, N–H/O 141.6
(2)ꢁ) resulting in a chain along the c-axis. Additionally, the chains
are held together by type I trans chlorine–chlorine interactions
(Cl/Cl 3.354(2) A, q1 ¼ 147.1 and q2 ¼ 147.0 )53 (Fig. 9A and
B). It is of interest to note that the LTG homosynthon dimer in
compound 2 involves motif 3 unlike motif 1 in the reported
crystal structure of LTG.42
ꢁ
ꢁ
ꢀ
Fig. 6 FT-IR spectra of LTG free base and its multicomponent forms
1–5.
Table 3 Crystallographic data for multicomponent crystals 2–4
(motif 1), N4 and N3 (motif 3) or N1 and N2 (motif 4). The
possible reason for these geometry changes is that the other
hydrogen bonding interactions that come up due to incorpora-
tion of the coformer are stronger as compared with the original
Parameters
Formula
2
3
4
C9H7Cl2N5,
C2H5NO
C9H8Cl2N5,
2(C2H4O2),
C2H3O2
1 : 3
Monoclinic
P21/c
7.0795(3)
17.7124(7)
15.9227(7)
90.000(0)
94.618(2)
90.000(0)
1990.14(14)
1.456
C9H8Cl2N5,
C7H5O3
Stoichiometry
Crystal system
Space group
1 : 1
Triclinic
ꢁ
P1
1 : 1
Monoclinic
P21/c
ꢀ
a/A
7.7752(3)
9.6284(4)
10.4511(3)
106.390(3)
107.730(3)
101.093(3)
680.98(5)
1.537
15.5808(10)
9.6890(6)
11.1057(7)
90.000(0)
91.312(3)
90.000(0)
1676.10(18)
1.562
ꢀ
b/A
ꢀ
c/A
a/deg
b/deg
g/deg
Volume/A
Calc. density/g cmꢀ3
3
ꢀ
Z
2
5.46
293(2)
4
3.01
90(2)
4
3.71
90(2)
R-Factor (%)
T/K
Table 4 Distribution of C–O bond lengths for multicomponents formed
with carboxylic acids
ꢀ
Dc–o/A
ꢀ
DDc–o/A
Compound Coformer
3
4
Acetic acid
1.310(2), 1.220(2)
(Dc10–o3, Dc10–o4)
1.311(2), 1.222(2)
(Dc12–o1, Dc12–o2)
1.258(2), 1.261(2)
(Dc14–o5, Dc14–o6)
0.090
0.099
0.003
0.022
4-Hydroxybenzoic acid 1.283(2), 1.261(3)
(Dc16–o1, Dc16–o2)
Fig. 7 Scheme representing all the four hydrogen bonding motifs.
6278 | CrystEngComm, 2011, 13, 6271–6284
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Table 5 Geometrical parameters of hydrogen bonds in compounds 2–4
:D–H/A/ꢁ
ꢀ
ꢀ
ꢀ
r (D–H)/A
D–H/A
Symmetry
r (H/A)/A
r (D/A)/A
Compound 2
N1–H1B/O1
N6–H6B/O1
N3–H3A/N4
N6–H6A/N4
Compound 3
ꢀx + 1, ꢀy + 1, ꢀz + 1
ꢀx, ꢀy + 1, ꢀz + 1
ꢀx + 1, ꢀy, ꢀz ꢀ 1
ꢀx + 1, ꢀy + 1, ꢀz
2.146(3)
2.25(5)
2.256(2)
2.54(6)
2.870(4)
2.962(5)
3.070(3)
3.383(6)
0.860(2)
0.72(5)
0.860(2)
0.87(6)
141.6(2)
170(6)
157.9(2)
164(5)
N1–H1A/O2
O3–H3O/O5
N1–H1B/O4
N4–H4A/O5
N3–H3A/N2
O1–H1O/O6
Compound 4
x ꢀ 1, +y, +z
1.969(1)
1.69(3)
2.040(1)
1.87(2)
2.284(1)
1.47(3)
2.808(2)
2.553(2)
2.840(2)
2.719(2)
3.137(2)
2.509(2)
0.860(1)
0.87(3)
0.860(1)
0.89(2)
0.860(1)
1.04(3)
164.5(1)
167(3)
154.4(1)
159(2)
171.6(1)
175(3)
x + 1, +y, +z
x, y, z
x,ꢀy + 1/2, +z ꢀ 1/2
ꢀx + 2, ꢀy, ꢀz + 1
ꢀx + 1, +y ꢀ 1/2, ꢀz + 1/2 + 1
N4–H4A/O1
N3–H3A/O1
O3–H3O/O2
N1–H1A/N2
N1–H1B/O1
N3–H3B/O2
ꢀx + 1, ꢀy + 1, ꢀz + 1
ꢀx + 1, +y + 1/2, ꢀz + 1/2
ꢀx ꢀ 1, +y ꢀ 1/2, ꢀz + 1/2
ꢀx + 1,ꢀy + 2,ꢀz + 1
1.79(3)
2.194(1)
1.830(1)
2.063(2)
2.427(1)
1.879(2)
2.692(2)
3.001(2)
2.642(2)
2.910(3)
2.977(2)
2.713(2)
0.90(3)
0.860(2)
0.820(2)
0.860(2)
0.860(2)
0.860(2)
176(3)
156.1(1)
170.1(1)
168.1(1)
122.3(1)
163.0(1)
ꢀx + 1, ꢀy + 1, z + 1
x, ꢀy + 1/2 + 1, +z + 1/2
the asymmetric unit (Fig. 10). An acetic acid trimeric homo-
ꢀ
ꢀ
synthon ((O3ꢁ–H3O/O5, O/O 2.553(2) A, O/H 1.69(3) A, O–
ꢀ
H/O 167(3) ), (O1ꢁ–H1O/O6, O/O 2.509(2) A, O/H 1.47(3)
ꢀ
A, O–H/O 175(3) )) generates a layer motif perpendicular to
the c axis (Fig. 11A). In this trimer, one of the acetic acid
molecules transfers a proton to LTG, which is seen from the
nearly equal bond distances associated with the C–O moiety.
ꢀ
Indeed, the DDc–o of 0.003 A for this acetic acid molecule (Table
4) is much smaller when compared to the values of DDc–o of
Fig. 8 ORTEP diagram of the asymmetric unit of cocrystal 2 with 50%
thermal ellipsoid probability.
ꢀ
ꢀ
0.090 A and 0.099 A respectively for the other two acetic acid
molecules. An atom (H4A) is transferred from the carboxylic
acid group (atom O5) to the triazine (atom N4) ring. In fact, the
C9–N4–N5 angle of the triazine ring in the crystal structure of 3
is 123.3ꢁ which correlates with the previously reported values of
protonated LTG and with the trend of protonated amino-
pyridines, that is a larger angle than those of neutral pyri-
dines.42,54 Thus an acid LTG heterosynthon is formed through
the N+H/Oꢀ charge assisted hydrogen bonding (N4+H4A/O5,
+
ꢀ
N/O 2.719(2) A, O/H 1.87(2) A, N H/O 159(2)ꢁ). One
trimer is further connected to another by another acid LTG
heterosynthon consisting of N–H/O hydrogen bondꢁs ((N1–
ꢀ
ꢀ
ꢀ
ꢀ
H1A/O2, N/O 2.808(2) A, O/H 1.969(1) A, 164.5(1) ), (Nꢁ1–
ꢀ
ꢀ
H1B/O4, N/O 2.840(2) A, O/H 2.040(1) A, 154.4(1) ))
(Fig. 11B). A LTG dimer homosynthon present in the original
drug (motif 1) is retained in the LTG layer (N3H3A/N2, N/N
ꢁ
ꢀ
ꢀ
3.137(2) A, N/H 2.284(1) A, N–H/N 171.6(1) ) which is
sandwiched between two acid layers (Fig. 11C).
The crystal structure of 3 further explains the two step weight
loss observed in TGA. The two neutral acetic acid molecules
which are bound by the hydrogen bonds in the trimer are released
earlier at 75–110 ꢁC (first step weight loss) while the acetate
molecule which is bound to the LTG molecule by the stronger
ionic hydrogen bond is released at 115–170 ꢁC (second step
weight loss).
Fig. 9 (A) Hydrogen bonding scheme of 2. (B) Packing diagram of 2
(showing hydrogen bonding and Cl/Cl interactions).
(3) LTG-acetic acid salt solvate (1 : 3). Complex 3 crystallizes
in a monoclinic space group P21/c. The asymmetric unit consists
of one LTG cation, two molecules of acetic acid and one acetate
anion. Hence this complex can be called a 1 : 1 salt of LTG and
acetic acid along with two neutral acetic acid solvent molecules in
(4) LTG-4-hydroxy-benzoic acid (1 : 1). Complex 4 crystal-
lizes in the monoclinic space group P21/c. The asymmetric unit
consists of one LTG cation and one 4-hydroxybenzoate anion.
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6271–6284 | 6279

View Article Online
Fig. 10 ORTEP diagram of the asymmetric unit of salt solvate 3 with
50% thermal ellipsoid probability.
Fig. 12 ORTEP diagram of the asymmetric unit of salt 4 with 50%
thermal ellipsoid probability.
Hence this complex can be called a 1 : 1 salt of LTG and 4-
hydroxy benzoic acid (Fig. 12). The LTG dimer homosynthon is
ꢀ
present in the salt structure (N1–H1A/N2, N/N 2.910(3) A,
ꢁ
ꢀ
N/H 2.063(2) A, N–H/N 163.1(1) ), but instead of motif 1 in
a pure LTG, motif 4 is involved in the formation of the dimeric
unit (Fig. 13). The drug dimers are further connected to acid
ꢀ
ꢀ
dimers (O3–H3O/O2, O/O 2.642(2) A, O/H 1.830(1) A, O–
H/O 170.1(1)ꢁ) through N+H/Oꢀ charge assisted hydrogen
bonding (N4 H4A/O1, N/O 2.692(2) A, O/H 1.79(3) A,
N+H/Oꢀ 176(3)ꢁ) apart from several other N–H/O interac-
tions. Proton transfer is evidenced by the C–O bond distances of
+
ꢀ
ꢀ
Fig. 11 A) Acetic acid trimer homosynthon layer. (B) Drug dimer
homosynthon and drug acid heterosynthons. (C) LTG dimers sand-
wiched between acetic acid layers.
Fig. 13 Basic supramolecular unit showing the 4-HBA dimer and LTG
dimer with various weak interactions.
6280 | CrystEngComm, 2011, 13, 6271–6284
This journal is ª The Royal Society of Chemistry 2011

View Article Online
ꢀ
the carboxylate group (with DDc–o ¼ 0.022 A) (Table 4) and the
the molar enthalpies of solution of the drug and the coformer
respectively.
geometry of the LTG triazine ring. The C9–N4–N5 angle of the
triazine ring in the crystal structure of 4 is 122.6ꢁ which correlates
with the previously reported values of protonated LTG.42 This
basic supramolecular unit extends in three dimensions through
various other N–H/O hydrogen bonds and Cl/O halogen
Comparison of the experimentally determined enthalpy of
solution of multicomponent forms with their calculated molar
enthalpy of solution shows that these forms behave more endo-
thermically when compared to the calculated values. The driving
force for the formation of these multicomponents is presumably the
introduction of additional stabilization by hydrogen bonding and
also of Coulombic interactions in the salts. The breaking of these
hydrogen bonds, which is an endothermic process, is responsible
for their higher endothermic enthalpy of solution than the calcu-
lated values. At this point, it is worth mentioning that although the
enthalpy of solution for the molecular complexes involves not only
the crystal lattice but also the hydration of the components, in the
present study, other terms such as primary and secondary solvation
are the same for the individual components as well as multicom-
ponent forms. This is because all the experiments have been per-
formed at a constant pH of 7.0, and it is expected that the same
molecular species exist for individual components as well as the
components after dissolution of molecular complexes. We can
determine the enthalpy of formation (DrH) of the cocrystals/salts
from their components using eqn (2).
ꢀ
bonded (2.986(2) A) interactions (Fig. 13).
Analysis in solution form
Enthalpy of solution. The enthalpy of solution is a direct
measurement of heat evolved or taken up when the crystal lattice
breaks down in a given solvent.55 Thus, it is an important inherent
thermodynamic parameter to characterize a particular crystal.
Scheme 2 explains the enthalpy change accompanying the disso-
lution of multicomponent crystals. The enthalpy of solution for
LTG, all the coformers and their multicomponent forms was
determined in a phosphate buffer (pH 7.0) at 37 ꢁC and the results
are presented in Table 6. The buffer of pH 7 is selected to ensure
that drug (pKa 5.7) remains in its neutral (unprotonated) form
upon dissolution. The enthalpy of solution of LTG and all the
coformers except acetic acid is endothermic. It can be seen from
Table 6 that dissolution of cocrystals as well as salts also exhibit
endothermic behaviour. The molar average enthalpy of the
physical mixtures (with the stoichiometry of the multicomponent
crystal forms) was calculated utilizing the individual molar
enthalpy of solution for the drug and coformer using eqn (1):
DrH ¼ DsolH(cal) ꢀ DsolH3
(2)
where DsolH3 is the experimentally determined molar enthalpy of
solution of the cocrystals/salts. The exothermic value of DrH
supports the fact that hydrogen bonds between complementary
functional groups of different molecules are more favoured than
between like molecules of either compound suggesting that these
multicomponent forms are thermodynamically favoured. The
packing and hydrogen bonding diagrams of various multicom-
ponent forms have been described in the previous section.
Table 6 shows the absolute value of DrH at a maximum for
compound 3 where one drug molecule is associated with three
acetic acid molecules. Higher exothermic enthalpies of formation
(DrH) for 3, 4 and 5 indicate these salts of LTG are more stable
than their cocrystals 1 and 2.
D
solH(cal) ¼ n1(DsolH1) + n2(DsolH2)
(1)
where n1 and n2 are the number of moles of the drug and the
coformer in the cocrystal/salt. DsolH(cal) is the calculated molar
enthalpy of solution of physical mixture. DsolH1 and DsolH2 are
Equilibrium solubility studies. The solubility profiles for all the
ꢁ
new phases at various time points in water at 37 C are shown
in Fig. 14 with a comparison to LTG free base. In the free
base sample, LTG reached its maximum concentration
Scheme 2 Thermodynamic cycle representing enthalpies of solution.
Table 6 Molar enthalpy of solution of LTG free base, coformers and their multicomponent forms in phosphate buffer pH 7.0 at 37 ꢁC and the values
of DrH
Solubility (s) in phosphate
buffer, pH 7.0 (mol Lꢀ1) ꢂ 10ꢀ3
Compound
Molar ratio
D
solH/kJ molꢀ1
D
solH(cal)/kJ molꢀ1
DrH/kJ molꢀ1
LTG
Nicotinamide
1
Acetamide
2
Acetic acid
3
4-Hydroxy-benzoic acid
—
—
1 : 1 : 1
—
1 : 1
—
1 : 3
—
12.37
25.53
52.98
7.14
34.19
ꢀ1.70
32.52
19.28
55.39
4.83
—
—
37.90
—
19.50
—
7.25
—
—
—
1.02
3904.72
2.81
7809.45
1.29
—
20.66
38.54
2.23
13.43
3.20
ꢀ15.08
—
ꢀ14.68
—
ꢀ25.27
—
ꢀ23.74
—
4
Saccharin
5
1 : 1
—
1 : 1
31.65
—
17.19
36.44
ꢀ19.25
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6271–6284 | 6281

View Article Online
Fig. 14 Solubility studies of LTG free base and its multicomponent forms at various time points in water at 37 ꢁC.
(0.40 mg mLꢀ1) after a time interval of 2 hours while the highest
solubility was observed for salt solvate 3 (5.81 mg mLꢀ1) that was
attained within 5 minutes followed by a decline over the
remaining period of study. This type of solubility profile is dis-
played by both the cocrystals and solvated salt of acetic acid as
cocrystals 1 and 2 also reached the maximum solubility in less
than 30 min. However, the salt forms 4 and 5 achieved their
maximum concentration in approximately 60 min. The pH of
each sample solution was observed during the experiment and
has been found to decrease in each case except 2 in contrast with
LTG free base (Table 7). In the case of compound 3, the decrease
in pH is significant (from 6.7 to 4.4), which can be attributed to
the dissolution of two neutral molecules of acetic acid present as
solvent in the crystal structure of 3. This explains the initial sharp
increase in the solubility of this compound. XRPD and DSC
analysis of solid residues remaining after the solubility experi-
ment showed that LTG free base, compounds 1 and 2 converted
to the LTG hydrate while salt solvate 3 converted to its free salt
form upon slurrying in aqueous medium. This explains the
decrease in concentration of the drug after an initial rise. These
are not isolated cases; there are ample records of cocrystals/
solvates reverting back to their free base or hydrates during
aqueous solubility studies.56,57 The salt forms 4 and 5 have been
found to be stable during solubility experiments as indicated by
XRPD analysis. The solubility (s) of LTG in all these multi-
components measured after 24 h of equilibration is given in
Table 7. Solubility is a complex parameter that depends upon the
enthalpy of fusion, the temperature of solvent and the melting
point of the solid.58 In the present study, we have tried to
establish a correlation between melting points and log (s) of
compounds 3, 4 and 5 (Table 7). Cocrystals 1 and 2 have not been
included in this correlation analysis as these forms exhibit low
solubility despite their lower melting points owing to their
instability in water. In order to make the correlation more reli-
able, we included the data from other known complexes of
LTG41 such as LTG hemisuccinate dimethylsulfoxide (DMSO)
solvate (A), LTG hemifumarate DMSO solvate (B) and LTG D,
L-hemitartrate DMSO solvate (C). A general trend where the
higher melting point complexes show lower log (s) values is seen,
however, this appears as a cluster with one outlying point (3)
(Fig. 15). Further data in the low melting point region are
therefore required to draw any firm conclusions.
Equilibrium solubility studies of LTG in its free base sample
and compounds 1–5 were also performed in phosphate buffer at
pH 7, 5, 3 and 2 with an aim of determining the effect of pH on
the solubility (Fig. 16). As the pH decreases to 2, the solubility of
LTG increases up to 4.60 mg mLꢀ1 and a similar pattern was
observed for cocrystals 1 and 2 (6.67 and 4.63 mg mLꢀ1). Inter-
estingly, the salt forms 3, 4 and 5 showed an increase in the
concentration of drug as the pH is lowered from pH 7 to 5 but
started declining as the pH is further decreased to 2 (5.99, 1.30
and 1.74 mg mLꢀ1 respectively). These differences in the solu-
bility with varying pH might be due to differences in the sum of
free energy of hydration of the species involved in the solubili-
zation process (i.e. anions, cations and neutral species that can be
present when a salt is dissolved in the aqueous medium).
From the crystal engineering point of view, all the three
compounds 2, 3 and 4 show breaking of motif 2 and exhibit
a higher solubility at pH 2 as compared to that in water.
However, salt solvate 3 that retained motif 1 exhibited higher
solubility than pure LTG while compound 2 that contained motif
3 showed solubility comparable to pure LTG and compound 4
Table 7 Melting points, solubility and potential conversion of cocrystal/salt during solubility studies in water upon equilibration at 37 ꢁC
Compound
Melting peaks/ꢁC
Solubility (s) in water/mg mLꢀ1
log (s)
pH of the solution at equilibrium
Conversion during experimenta
LTG
1
2
3
4
217.1
173.9
160.5
170.0
234.5
257.3
250.0
275.0
243.0
255.2
546.1
320.1
8414.7
534.8
696.7
156.2
110.1
673.5
2.41
2.74
2.51
3.93
2.73
2.84
2.19
2.04
2.83
6.7
6.6
6.9
4.4
5.2
4.5
—
Yes (LTG hydrate)
Yes (LTG hydrate)
Yes (LTG hydrate)
No
No
No
—
5
A41
B41
C41
—
—
—
—
a
Conversion assessed by XRPD and DSC. (Compounds A, B and C are added in these data from ref. 41 in order to make the correlation between
solubility and melting point more reliable.)
6282 | CrystEngComm, 2011, 13, 6271–6284
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Fig. 15 Log (s) as a function of cocrystal/salt melting point.
Fig. 17 ED50 values (mg kgꢀ1) of LTG in a pure commercial sample and
its multicomponents.
Conclusion
The work discussed herein illustrates the successful and
methodical implementation of the multicomponent approach
(cocrystals and salts) for improving the physicochemical prop-
erties of LTG. The formation of new solid phases in each case
was indicated by DSC, whereas the preliminary differentiation
between cocrystals and salts was performed by FT-IR analysis.
The single crystal structure for the cocrystal with acetamide and
salts with acetic acid and 4-hydroxy-benzoic acid was deter-
mined. In the cocrystal 2, LTG exhibits an aminopyridazine
dimer homosynthon (motif 3) which is linked to the acetamide
dimer homosynthon through a N–H/O hydrogen bond. In the
crystal structures of salts 3 and 4, the aminopyridine dimer of
LTG is retained but with slight deviations in the geometry in the
case of compound 4 (motif 4). In all these three structures motif 2
was broken. Solution calorimetry suggested compounds 3, 4 and
5 to be thermodynamically more stable forms as compared to 1
and 2. The cocrystals 1 and 2 were found to convert into LTG
hydrate, salt solvate 3 converted to its pure salt form while salts 4
and 5 are found to be stable during solubility studies. Maximum
enhancement in aqueous solubility has been observed in salt
solvate 3, which is found to be stable and has significantly low-
ered the dosage profile of LTG in animal activity studies. Thus,
this multicomponent form 3 can be considered for further
development. Some other coformers are also under investigation
in our laboratory for exploring more cocrystals and salts of this
drug and will constitute our future communication.
Fig. 16 Solubility of LTG and multicomponent forms 1–5 measured
after 24 h of equilibration in phosphate buffers of various pH at 37 ꢁC.
exhibiting motif 4 has been found to be less soluble than pure
LTG. These observations are in agreement with Cheney et al.
who concluded that LTG cocrystal structures that break motif 2
but retain motif 1 are more soluble than pure LTG under acidic
conditions while LTG salts that break motif 1 are more soluble
than pure LTG in aqueous solutions.42
Stability studies. The physical stability of multicomponent
forms of LTG was investigated at 25 ꢁC/60% RH for one month.
The DSC of these samples did not show any additional peaks
indicating these to be stable which is further confirmed by the
absence of any significant changes in their XRPD patterns.
Animal activity studies. In the MES test, LTG and all the
compounds tested produced a dose dependent abolition of hind
limb extension in mice. Results are presented as median effective
doses (ED50 in mg kgꢀ1) required to protect 50% of animals
tested against MES-induced seizures. ED50 values determined in
mice (Fig. 17) following oral administration showed compounds
1, 3 and 5 to be effective even at a lower dose of 2.5 mg kgꢀ1 while
compound 4 achieved similar potency at a dose of 5.0 mg kgꢀ1 as
compared to compound 2 and LTG free base which abolished
the HLE only at a dose of 7.5 mg kgꢀ1 and above. All these
compounds showed a pattern of absorption quite similar to that
observed in the solubility studies. Thus, this study illustrates that
pharmaceutical cocrystals and salts can significantly alter the
dosage profile of parent drug in animal activity studies.
Acknowledgements
We gratefully acknowledge the funding provided by the UGC
(Research Fellowship for Meritorious Students), New Delhi,
India.
References
1 S. Datta and D. J. W. Grant, Nat. Rev. Drug Discovery, 2004, 3, 42–
57.
2 C. B. Aakeroy, M. E. Fasulo and J. Desper, Mol. Pharmaceutics,
2007, 4, 317–322.
3 P. V. Arnum, Pharm. Technol., 2010, 2, 50–56.
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6271–6284 | 6283

View Article Online
4 C. Leuner and J. Dressman, Eur. J. Pharm. Biopharm., 2000, 50, 47–
60.
31 B. Sridhar and K. Ravikumar, Acta Crystallogr., Sect. E: Struct. Rep.
Online, 2005, 61, o3805–o3807.
5 J. C. Chaumeil, Exp. Clin. Psychopharmacol., 1998, 20, 211–215.
6 K. Kawakami, N. Oda, K. Miyoshi, T. Funaki and Y. Ida, Eur. J.
Pharm. Sci., 2006, 28, 7–14.
7 D. Singhal and W. Curatolo, Adv. Drug Delivery Rev., 2004, 56, 335–
347.
€
8 A. Abdalla and K. Mader, Eur. J. Pharm. Biopharm., 2007, 66, 220–
226.
9 K. Uekama, F. Hirayama and T. Irie, Chem. Rev., 1998, 98, 2045–
2076.
10 J. Hecq, M. Deleers, D. Fanara, H. Vranckx and K. Amighi, Int. J.
Pharm., 2005, 299, 167–177.
11 S. L. Morissette, O. Almarsson, M. L. Peterson, J. F. Remenar,
M. J. Read, A. V. Lemmo, S. Ellis, M. J. Cima and C. R. Gardner,
Adv. Drug Delivery Rev., 2004, 56, 275–300.
12 A. T. M. Serajuddin, Adv. Drug Delivery Rev., 2007, 59, 603–616.
13 S. M. Berge, L. D. Bighley and D. C. Monkhouse, J. Pharm. Sci.,
1977, 66, 1–19.
32 R. W. Janes, J. Chem. Crystallogr., 1999, 29, 163–167.
33 R. B. Parthasaradhi, R. K. Rathnakar, R. R. Raji,
R. D. Muralidhara, and R. K. S. Chander, US. Pat., 2005/0119265
A1, 2005.
34 B. Sridhar and K. Ravikumar, Acta Crystallogr., Sect. E: Struct. Rep.
Online, 2006, 62, o4752–o4754.
35 R. W. Janes, J. N. Lisgarten and R. A. Palmer, Acta Crystallogr.,
Sect. C: Cryst. Struct. Commun., 1989, 45, 129–132.
36 M. Kubicki and P. W. Codding, J. Mol. Struct., 2001, 570, 53–
60.
37 B. Sridhar and K. Ravikumar, Mol. Cryst. Liq. Cryst., 2007, 461,
131–141.
38 D. A. Sawyer and F. C. Copp, US. Pat., 4847249, 1989.
39 B. Potter, R. A. Palmer, R. Withnall, M. J. Leach and
B. J. Chowdhry, J. Chem. Crystallogr., 1999, 29, 701–706.
40 B. Sridhar and K. Ravikumar, Acta Crystallogr., Sect. C: Cryst.
Struct. Commun., 2009, 65, o460–o464.
14 A. T. M. Serajuddin and M. Pudipeddi, Salt Selection Strategies, in
Handbook of Pharmaceutical Salts: Properties, Selection and Use,
2002, pp. 135–160.
15 R. Banerjee, P. M. Bhatt, N. V. Ravindra and G. R. Desiraju, Cryst.
Growth Des., 2005, 5, 2299–2309.
16 L. D. Bighley, S. M. Berge and D. C. Monkhouse, Salt Forms of
Drugs and Absorption, in Encyclopedia of Pharmaceutical
Technology, 1996, vol. 13, pp. 453–499.
17 W. Q. Tong and G. Whitesell, Pharm. Dev. Technol., 1998, 3, 215–223.
18 N. Schultheiss and A. Newman, Cryst. Growth Des., 2009, 9, 2950–
2967.
19 K. Shiraki, N. Takata, R. Takano, Y. Hayashi and K. Terada, Pharm.
Res., 2008, 25, 2581–2592.
20 A. Bak, A. Gore, E. Yanez, M. Stanton, S. Tufekcic, R. Syed,
A. Akrami, M. Rose, S. Surapaneni, T. Bostick, A. King,
S. Neervannan, D. Ostovic and A. Koparkar, J. Pharm. Sci., 2008,
97, 3942–3956.
21 J. Lu and S. Rohani, Org. Process Res. Dev., 2009, 13, 1269–1275.
22 D. J. Good and N. Rodriguez-Hornedo, Cryst. Growth Des., 2009, 9,
2252–2264.
41 J. Galcera and E. Molins, Cryst. Growth Des., 2009, 9, 327–334.
42 M. L. Cheney, N. Shan, E. R. Healey, M. Hanna, L. Wojtas,
M. J. Zaworotko, V. Sava, S. Song and J. R. Sanchez-Ramos,
Cryst. Growth Des., 2010, 10, 394–405.
43 A. A. Miller, P. Wheatley, D. A. Sawyer, M. G. Baxter and B. Roth,
Epilepsia, 1986, 27, 483–489.
44 Oxford Diffraction, CrysAlis PRO CCD and CrysAlis PRO RED,
Oxford diffraction Ltd, Yarnton, England, 2009.
45 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2008,
64, 112–122.
46 L. J. Farrugia, J. Appl. Crystallogr., 1999, 32, 837–838.
47 L. J. Farrugia, J. Appl. Crystallogr., 1997, 30, 565.
48 POV-Ray, version 3.6, Chemical Crystallography Laboratory,
Australia.
49 A. L. Spek, J. Appl. Crystallogr., 2003, 36, 7–13.
50 C. F. Macrae, I. J. Bruno, J. A. Chisholm, P. R. Edgington,
P. McCabe, E. Pidcock, L. Rodriguez-Monge, R. Taylor, J. van de
Streek and P. A. Wood, J. Appl. Crystallogr., 2008, 41, 466–470.
51 S. L. Childs, G. P. Stahly and A. Park, Mol. Pharmaceutics, 2007, 4,
323–338.
23 M. K. Stanton and A. Bak, Cryst. Growth Des., 2008, 8, 3856–3862.
24 J. F. Remenar, S. L. Morissette, M. L. Peterson, B. Moulton,
J. M. Macphee, H. R. Guzman and O. Almarsson, J. Am. Chem.
Soc., 2003, 125, 8456–8457.
52 M. Hanna, N. Shan and M. L. Cheney, Thor Pharmaceutics, WO
Pat., 2009/061513 A1, 2009.
53 V. R. Hathwar and T. N. Guru Row, J. Phys. Chem. A, 2010, 114,
13434–13441.
25 S. Basavoju, D. Bostrom and S. P. Velaga, Pharm. Res., 2008, 25,
530–541.
54 R. A. Palmer, B. S. Potter, M. J. Leach and B. Z. Chowdhury,
J. Chem. Crystallogr., 2007, 37, 771–777.
26 N. Shan and M. Zaworotko, Drug Discovery Today, 2008, 13, 440–
446.
55 P. G. Royall and S. Gaisford, Curr. Pharm. Biotechnol., 2005, 6, 215–
222.
27 A. V. Trask, Mol. Pharmaceutics, 2007, 4, 301–309.
28 The Internet Drug Index, www.rxlist.com/cgi/generic/lamotrigine.htm,
accessed 13 November, 2009.
56 S. L. Childs, L. J. Chyall, J. T. Dunlap, V. N. Smolenskaya,
B. C. Stahly and G. P. Stahly, J. Am. Chem. Soc., 2004, 126,
13335–13342.
29 K. R. Parmar, K. A. Patel, S. R. Shah and N. R. Sheth, J. Inclusion
Phenom. Macrocyclic Chem., 2009, 65, 263–268.
30 (a) GRAS Notices, http://www.cfsan.fda.gov/rdb/opa-gras.html; (b)
Food additive status list, http://www.cfsan.fda.gov/ꢄrdb/opa-appa.
html.
57 D. P. McNamara, S. L. Childs, J. Giordano, A. Iarriccio, J. Cassidy,
M. S. Shet, R. Mannion, E. O’Donnell and A. Park, Pharm. Res.,
2006, 23, 1888–1897.
58 A. Martin, J. Newburger and A. Adjei, J. Pharm. Sci., 1980, 69, 487–
491.
6284 | CrystEngComm, 2011, 13, 6271–6284
This journal is ª The Royal Society of Chemistry 2011