Heterosynthon mediated tailored synthesis of pharmaceutical complexes: A solid-state NMR approach

Article abstract of DOI:10.1039/c0ce00657b

Based on crystal engineering principles, we have explored the predictability of resulting structures of a multi-component pharmaceutical model complex derived from 4-hydroxybenzoic acid (4HBA) and quinidine, an anti-malarial constituent of Cinchona tree bark. Though the obtained complex is stabilized by a slightly different set of charge-assisted heterosynthons as proposed, the applied concept was efficient in predicting the salt formation. The salt 1 crystallizes in a monoclinic space group [P2₁ (no. 4), Z = 8, a = 6.914 A, b = 36.197 A, c = 9.476 A and β = 92.126], where the asymmetric unit is comprised of two quinidine and two 4HBA molecules. In addition, a micro-crystalline, less-defined sample of 1 was obtained from rapid co-crystallization in ethanol and successfully identified via both infrared spectroscopy and multinuclear solid-state NMR. The interpretation of the obtained NMR data was supported by DFT quantum-chemical computations while illustrating options of "NMR crystallography".

Full text of DOI:10.1039/c0ce00657b

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
CrystEngComm
Cite this: CrystEngComm, 2011, 13, 3213
www.rsc.org/crystengcomm
PAPER
Heterosynthon mediated tailored synthesis of pharmaceutical complexes:
a solid-state NMR approach†
*
Mujeeb Khan, Volker Enkelmann and Gunther Brunklaus
Received 17th September 2010, Accepted 10th December 2010
DOI: 10.1039/c0ce00657b
Based on crystal engineering principles, we have explored the predictability of resulting structures of
a multi-component pharmaceutical model complex derived from 4-hydroxybenzoic acid (4HBA) and
quinidine, an anti-malarial constituent of Cinchona tree bark. Though the obtained complex is
stabilized by a slightly different set of charge-assisted heterosynthons as proposed, the applied concept
was efficient in predicting the salt formation. The salt 1 crystallizes in a monoclinic space group [P21
ꢀ
ꢀ
ꢀ
(no. 4), Z ¼ 8, a ¼ 6.914 A, b ¼ 36.197 A, c ¼ 9.476 A and b ¼ 92.126], where the asymmetric unit is
comprised of two quinidine and two 4HBA molecules. In addition, a micro-crystalline, less-defined
sample of 1 was obtained from rapid co-crystallization in ethanol and successfully identified via both
infrared spectroscopy and multinuclear solid-state NMR. The interpretation of the obtained NMR
data was supported by DFT quantum-chemical computations while illustrating options of ‘‘NMR
crystallography’’.
distinct whereas salts are ionized and often less-ordered solids
1. Introduction
where a proton is typically transferred from an acid to a base.9
Though many examples10 of successful co-crystallization11 are
known, which yield solids with hydrogen bonding motifs
expected from the principles of crystal engineering,12 an exact
prediction of resulting three-dimensional structures of solids
obtained from such experiments is still a challenging task,13,14
particularly in the case of less rigid fragments.15
Most pharmaceuticals contain active pharmaceutical ingredients
(APIs) in the form of molecular crystals and are frequently
administered in the solid-state as part of an approved dosage
type such as tablets or capsules.1 Scientists constantly strive to
improve physical properties of APIs, including crystallinity,
stability, taste and (most importantly) solubility.2 Indeed, drug
molecules with limited aqueous solubility are rather challenging
in pharmaceutical development and may pose the risk of insuf-
ficient or inconsistent exposure and thus poor efficacy in patients
upon oral administration.3 Therefore, the basic concepts of
crystal engineering4 have been recently utilized for tailored design
of pharmaceuticals thereby providing a new path for (more)
systematic discovery of a wider range of multi-component
structures such as pharmaceutical co-crystals5,6 and salts with
properly adjusted pharmacokinetic and physical properties.7
Both, co-crystals and salts, constitute multi-component
compounds whose crystal structures are often dominated by the
driving force of building a maximum of hydrogen bonds.8 Co-
crystals, however, comprise neutral molecules that are chemically
Co-crystal formation in principle can be rationalized by
consideration of hydrogen bond donors (D) and acceptors (A)
present in the respective precursor materials. Though hydrogen
bonds have been defined on the basis of interaction geometries
found in crystal structures, certain effects in IR absorption
spectra (e.g. red shift of nDH) or experimental electron density
distributions, a necessary geometric criterion for reliable (rather
strong) hydrogen bonding is a positive directionality preference,
where ‘‘linear’’ D–H/A angles must be statistically favoured
over bent ones.16 Since a distribution of many D–H/A angles
and even hydrogen bonds with more than one acceptor are
feasible (e.g. ‘‘bifurcation’’), a combination of multiple (even
relatively weak) hydrogen bonds may lead to the formation of
highly complex, rather stable molecular aggregates thereby
indicating cooperativity of hydrogen bonding.17 Ionic hydrogen
bonds, however, constitute a particular class of interaction which
is relevant for self-assembly of supramolecular moieties,18
protein folding, many electrolytes and even ionic aggregates.19 In
particular, charge-assisted hydrogen bonds (CAHBs) consist of
hydrogen bond acceptor and donor moieties where each of them
carries an ionic charge that further reinforces the electrostatic
dipole–dipole character of the hydrogen bond, thus rendering
them preferably linear in geometry.20 Notably, CAHBs in
€
Max-Planck-Institut fur Polymerforschung, Postfach 31 48, D-55021
Mainz, Germany. E-mail: brunklaus@mpip-mainz.mpg.de
† Electronic supplementary information (ESI) available: Spectral data
not figured in the manuscript such as simulated powder diffraction
1
1
data, H MAS and H DQ MAS NMR spectra of the ‘‘ethanol phase’’
of salt 1, table of DFT computed chemical shifts, solid-state
a
15N-CPMAS spectra of quinidine, salt 1, infrared absorption spectra of
4HBA, quinidine, acetone and ethanol phases of salt 1, the CIF file of
salt 1 and complete ref. 83. CCDC reference number 804730. For ESI
and crystallographic data in CIF or other electronic format see DOI:
10.1039/c0ce00657b
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 3213–3223 | 3213

View Online
molecular crystals have been classified according to whether they
are negative charge assisted (e.g., Oꢀ/H–O), positive charge
assisted (e.g., O/H–N+), ‘‘neutral charge assisted’’ (e.g., Oꢀ/
H–N+) or even resonance assisted (in this case two oxygens or
nitrogens are connected by a system of p-conjugated double
bonds).21 Moreover, N–H donors with a formal positive charge
tend to form shorter bonds than the corresponding uncharged
N–H groups, while the negatively charged carboxylate ion is
a stronger acceptor than uncharged amides, ketones or carboxyl
moieties.22
Therefore, in the present work, while applying crystal engi-
neering principles,23 we explore the predictability of the resulting
structures of a multi-component pharmaceutical model complex
based on the robust and competitive O–H/N24 heterosynthon
that may be comprised of either neutral (O–H/N) or charge-
assisted (Oꢀ/H–N+) hydrogen bond mediated molecular recog-
nition. In addition, the suitability of solid-state NMR for the
structural characterization of micro-crystalline or rather ill-
defined compounds that may be difficult to identify by other
techniques is highlighted thereby illustrating options of NMR-
based crystallography.25 For this purpose, based on the
compound classification Generally Recognized As Safe
(GRAS),26 we have selected p-hydroxybenzoic acid (4HBA) as the
co-crystal former,27 which is structurally highly related to estab-
lished tablet excipients such as methyl paraben or gentisic acid,
and therefore may also have the potential to be applied as an
excipient. Though 4HBA has been previously used as a co-crystal
former with tetramethylpyrazine (an important active ingredient
of the Chinese herb Ligusticum wallichii Franchat),28 any obtained
formulations with 4HBA still have to be thoroughly evaluated for
long term stability since pharmaceuticals are often prone to solid
state reactions including phase transformations (e.g., into
different polymorphs), dehydration ordesolvation as well as acid–
base or transacylation reactions of the active pharmaceutical
ingredients (APIs) and tablet excipients.29
Scheme 1 Chemical structures and supramolecular homo- and hetero-
synthons offered by pure quinidine, p-hydroxybenzoic acid (4HBA) and
molecular salt 1. (a) acid dimer, (b) acid–pyridine single point neutral, (c)
acid-pyridine two point neutral, (d) acid–phenol single point ionic, (e)
acid–pyridine single point ionic, (f) acid–pyridine two point ionic and (g)
phenol–pyridine single point neutral.
that allows for detailed structural characterization of powdered
materials,34 including pharmaceutical co-crystals24 and
complexes,35 even in the absence of X-ray structural data.
Notably, NMR not only affords non-invasive, element specific
observation of different nuclei thereby providing outstanding
selectivity for local environments,36 but also facilitates an iden-
tification of chemically distinct sites based on NMR chemical
shifts.37 In particular, protons involved in hydrogen-bonded
structures exhibit well-resolved 1H chemical shifts mainly
between 8 and 20 ppm (depending on their geometry),38
rendering an estimation of hydrogen-bonding strengths
feasible,39 as e.g. documented by established empirical correla-
4HBA provides both a hydroxyl group (OH) allowing for
a neutral O–H/N heterosynthon and a carboxyl group (COOH)
which is often known to engage in an ionized Oꢀ/H–N+ het-
erosynthon with nitrogen bearing compounds such as pyridines
or amines, provided that the DpKa is sufficiently large (>3).30 On
the other hand we considered a representative API that possesses
multiple accessible nitrogen atoms (preferably with different pKa
values) such as quinidine (cf. Scheme 1),24 an anti-malarial
constituent of Cinchona tree bark which can be used as anti-
arrythmatic agent with anti-muscarinic and alpha-adrenoceptor
blocking properties31 or for treatment of neurological disor-
ders.32
1
tions between H isotropic chemical shifts and corresponding
hydrogen bonding strengths, i.e. specified by O/H, N/H or
O/O distances, respectively, by X-ray analysis,40 neutron41 and
even electron diffraction.42 Additional structural insights may be
1
obtained from double-quantum H MAS NMR where homo-
1
nuclear H–1H dipolar couplings correlate protons of different
chemical entities thereby providing both precise information on
proton–proton proximities as well as distances43 (which in prin-
ciple can be very fruitful for unambiguous identification of
Equimolar co-crystallization of 4HBA with quinidine from
acetone has yielded a crystalline molecular salt 1 that is stabilized
via charge-assisted Oꢀ/H–N+ units, whereas co-crystallization
from ethanol produced an apparently ill-defined compound
which was identified as micro-crystalline salt 1 by both infrared
spectroscopy and solid-state NMR. While infrared spectroscopy
is typically applied to monitor changes of the sample identity
with respect to a given reference compound including the
detection of either counterfeit medicines33 or successful co-crystal
formation (particularly when carboxylic acids are involved),2a
modern high-resolution solid-state NMR (at high magnetic fields
and rather fast magic-angle spinning) constitutes a powerful tool
pharmaceutical salts) on length scales of up to 3.5 A44 and proton
ꢀ
positions in arrays of multiple hydrogen bonds.45 Moreover, the
combination of solid-state NMR spectroscopy with density
functional theory (DFT)46 computations not only corroborates
chemical shift assignments but in some cases also provides an
approach to ‘‘NMR crystallography’’,47 which allowed for both
polymorph screening,48 e.g. in hydrochloride pharmaceuticals,49
and even powder structure determination of small drug mole-
cules50 or precursors for synthesis of graphite-like carbon
nitrides, respectively.51 Nevertheless, de novo NMR-based
structure solutions without or only minor support from
3214 | CrystEngComm, 2011, 13, 3213–3223
This journal is ª The Royal Society of Chemistry 2011

View Online
additional X-ray data of functional compounds are still rather
scarce. In many cases, however, merely local complex formation
is of prime interest, which is indeed immediately accessible via
solid-state NMR.
2. Results and discussion
Based on extensive statistical studies of the Cambridge Structural
Database52 (CSD) addressing selected hydrogen-bonding
motifs,30 it has been reported that a variety of supramolecular
heterosynthons are fairly reliable for successful preparation of
co-crystals and often formed preferentially compared to homo-
synthons.14 In particular, the strong and rather specific recogni-
tion among either carboxylic acids and pyridine(s) (COOH/
N)53 or phenols and pyridine(s) (O–H/N)14 has been well
studied. Indeed, many co-crystals have been obtained based on
carboxylic acid and heterocyclic-N hydrogen-bonding,5,53
including the abundantly used co-crystal former isonicotinamide
(a member of the vitamin B group),54 in addition to molecular
salts (i.e., the proton is transferred from the acid to base).9 In
contrast, the interaction of phenolic hydroxyl groups (OH) with
pyridine almost exclusively has produced co-crystals.14,36b
Notably, four variants of the acid–pyridine heterosynthon are
found in the corresponding crystal structures: (i) neutral, two
point; (ii) neutral, single point; (iii) ionic, two point; and (iv)
ionic, single point (Scheme 1).27 The acid–pyridine neutral two
point synthon (cf. Scheme 1c) occurrence probability is around
90% when other competing functional groups are absent.
However, when there is a sufficient pKa (DpKa ¼ [pKa base] ꢀ
[pKa acid] > 3) difference between the COOH and pyridyl group,
a proton transfer will occur resulting in the formation of ionic
hydrogen bonds (Oꢀ/H–N+ (cf. Scheme 1e and f)), thus yielding
molecular salts.9
Scheme 2 proposed model for the synthesis of multi-component phar-
maceutical complex comprised of both charge assisted (Oꢀ/H–N+) and
neutral (O–H/N) hydrogen-bond mediated molecular recognition of
heterosynthons based on crystal engineering principles.
forming a charge assisted Oꢀ/H–N+ heterosynthon, mainly
owing to the pKa difference of more than 3 between the quinu-
clidine N atom and 4HBA, respectively. In addition, the
hydroxyl group (OH, second best donor) of 4HBA should be
engaged in a neutral O–H/N heterosynthon with the quinoline
N atom of quinidine, thus leading to a chain like structure (cf.
Scheme 2). Though considerable efforts have been made to
explicitly define synthons4 and their use in supramolecular
synthesis of pharmaceutical complexes,24 the predictions of the
result of such preparation is still not reliable.13,14 Nevertheless,
there have been some rather promising attempts to predict the
resulting crystal structures of 1 : 1 molecular co-crystals and
solvates using a chemical diagram56 or more recently, computed
crystal energy landscapes,57 including the co-crystal structure of
caffeine with 4HBA.58 Yet, the solid form resulting from exper-
iments targeting molecular complexes comprised of ionic syn-
thons (such as molecular salts) may be even more challenging to
predict due to the tendency of charge assisted complexes to have
alternating lattice compositions.9a This is nicely demonstrated
based on a survey of 85 solids (either salts or co-crystals)
obtained from stoichiometric mixtures of a carboxylic acid with
an N-heterocyclic base, where it was found that 45% of the ‘‘salt
structures’’ were either solvates or had a different stoichiometric
composition than implied from the expected hydrogen bonding
in contrast to only 5% in the case of the co-crystals.9a
Therefore, 4HBA and quinidine have been selected as model
compounds to form a multicomponent molecular complex that
may be stabilized by both neutral (O/H–N) or charge assisted
(Oꢀ/H–N+) heterosynthons. Quinidine consists of three basic
molecular units, a quinoline aromatic ring, a quinuclidine ring (a
tertiary amine), and a methylenic alcohol group linking the two
(cf. Scheme 1). In its pure form, quinidine crystallizes in
ꢀ
a monoclinic space group (P21 (no. 4); Z ¼ 2; a ¼ 11.883 A, b ¼
ꢀ
ꢀ
7.037 A, c ¼ 11.256 A), with the unit cell comprised of two
quinidine molecules that are stabilized by intermolecular
hydrogen-bonding among the C11-hydroxyl group and the N-
atom of the quinuclidine ring while the N-atom of the quinoline
ring remains free.24 According to crystal engineering principles,
nitrogen atoms of both quinuclidine and quinoline rings with
pKa values of 8.4 and 4.25 respectively55 and the carboxyl
(COOH) and hydroxyl (OH) groups of 4HBA with pKa value of
4.48 constitute possible targets for hydrogen-bond mediated
molecular recognition of both neutral and charge assisted heter-
osynthons (cf. Scheme 2).
Equimolar co-crystallization of quinidine and 4HBA in
acetone has yielded the salt 1 (Fig. 1) rather than the anticipated
molecular complex shown in Scheme 2. It crystallizes in
ꢀ
a monoclinic space group [P21 (no. 4), Z ¼ 8, a ¼ 6.914 A,
ꢀ
ꢀ
b ¼ 36.197 A, c ¼ 9.476 A and b ¼ 92.126] with significantly
larger b-axis, where the asymmetric unit comprises two quinidine
and two 4HBA molecules respectively. Indeed, the quinuclidinic
N-atom in 1 is part of a charge assisted single point Oꢀ/H–N+
heterosynthon, where the proton is transferred from the carboxyl
group of 4HBA to the N-atom of the quinuclidine ring, just as
predicted. However, the N-atom of the quinoline ring did not get
involved in any type of hydrogen-bonding and remained ‘‘free’’
in 1, similarly to pure quinidine. Hence, no neutral O–H/N
In our previous work, co-crystal formation of quinidine and
methyl paraben based on the neutral O–H/N heterosynthon has
been successfully demonstrated,24 where the quinuclidine
nitrogen was hydrogen-bonded to a hydroxyl group (OH) of
methyl paraben. However, when methyl paraben is replaced by
4HBA, we expected that the COOH group (best donor)28 of
4HBA could be hydrogen-bonded to the same nitrogen thereby
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 3213–3223 | 3215

View Online
Indeed, the refined single crystal X-ray structure of the 1 : 1
complex of quinidine and 4HBA indicates successful formation
of a salt: the measured C–O bond distances of the carboxylate
ꢀ
ꢀ
ꢀ
group (COO ) of 4HBA in 1 are 1.261 A and 1.231 A with an
average ratio of 1.024, which is very close to the reported ratio of
salts.9a Moreover, the measured C–N–C angles of the aliphatic
quinuclidine ring in 1 (109.63ꢂ, 109.17ꢂ and 112.26ꢂ) are slightly
larger than the corresponding C–N–C angles of the quinuclidine
ring in the ‘‘neutral’’ l : 1 co-crystal of quinidine and methyl-
paraben (107.97ꢂ, 107.23ꢂ, 111.13ꢂ),24 in good agreement with the
expected trend in the case of salt formation.
As stated earlier, equimolar co-crystallization of both quini-
dine and 4HBA from acetone yielded colorless well-defined
prism-shaped crystals of 1 while a similar attempt of rapid co-
crystallization from ethanol produced a rather glassy substance
sticking to the wall of the container, which turned into a white
sticky powder upon scratching. Since the physical states of both
samples were quite different we acquired the respective powder
X-ray diffraction patterns (Fig. 2) to possibly identify other
phases or impurities. Notably, the diffractogram of the ‘‘acetone
phase’’ with its sharp and well-resolved peaks clearly indicated
a crystalline nature of the sample. The pattern is almost identical
to a simulated powder pattern of 1 (cf. ESI†), whereas the large
‘‘bump’’ with smaller and blurred (highly unresolved) peaks in the
powder pattern of the ‘‘ethanol phase’’ point towards a reduced
bulk crystallinity of the sample (e.g., ill-defined or rather micro-
crystalline). Nevertheless, at this point, it is not clear whether the
compound from ethanol reflects the molecular salt 1 or another
polymorph.
Fig. 1 Portion of the molecular packing of pharmaceutical salt 1
obtained from quinidine and 4HBA depicting the total number of
hydrogen bonds responsible for stabilizing the structure. All hydrogen
atoms are omitted for clarity except for those involved in hydrogen
bonding (dotted lines represent the hydrogen bond).
heterosynthon24 is formed between the OH group of 4HBA and
the N-atom of the quinoline ring in 1. Rather, the hydroxyl group
(OH) of 4HBA is hydrogen-bonded to the carboxylate group
(COOꢀ) of an another 4HBA molecule within the asymmetric
unit of 1 reflecting a charge assisted O–H/Oꢀ homosynthon
(Scheme 1d). Notably, the crystal structure of pristine 4HBA
comprises the commonly found hydrogen bonding motif of
a carboxylic acid dimer (Scheme 1a),59 while there is no direct
hydrogen-bonding between the 4-hydroxyl (OH) and carboxylic
acid groups. Clearly, such COOH dimer formation among
4HBA molecules is absent in 1. Rather, one of the oxygen atoms
in the carboxylate group of 4HBA acts as hydrogen-bond
acceptor for hydroxyl groups (OH) of both quinidine and 4HBA
(cf. Fig. 1). In summary, the crystal structure of 1 is stabilized by
three types of charge assisted synthons, one ionic single point
Oꢀ/H–N+ heterosynthon and two O–H/Oꢀ homosynthons,
while interestingly, upon co-crystallization of 4HBA with pyri-
dine, a neutral compound can be obtained.60
Upon close inspection of the respective powder diffractograms
of both samples (crystalline and apparently ill-defined), the
rather weak diffraction peaks on top of the large bump (‘‘ethanol
phase’’) clearly coincide with the more prominent peaks of the
‘‘acetone phase’’ hence indicating a very similar molecular
packing (Fig. 2). Since the identity of a considered pharmaceu-
tical formulation (and in some cases even the amount of an active
ingredient within a tablet) can be established via infrared (IR)
spectroscopy,64 we recorded the corresponding IR absorption
spectra of quinidine, 4HBA as well as both the acetone and
ethanol phases of salt 1 (ESI†), which indeed exhibit a number of
distinct differences in the fingerprint and high frequency region.
While this clearly indicates changes in the nature of the respective
sample, IR spectroscopy may be used as a primary method only
if artificial calibration samples with the same composition and
structure as the samples to be analyzed are available. Neverthe-
less, the absence of a strong C]O stretching band around 1700
cmꢀ1 in the ‘‘acetone phase’’ hints at formation of carboxylate
(COOꢀ) anions,65 and the observable high energy shift of the OH
stretching band from around 3380 cmꢀ1 in the case of 4HBA to
3422 cmꢀ1 in salt 1 may reflect changes of the hydrogen bonding
pattern upon complex formation. Apart from these features,
however, the IR absorption spectrum of 1 is highly unresolved
(e.g. due to strong peak overlapping of both 4HBA and quinidine
resonances) rendering structural assignments and identification
of the formed complex rather difficult.
Though selective hydrogen-bonding constitutes the preferred
interaction in crystal engineering studies,61 which is known to
significantly contribute to physical properties or reactivity36b of
molecular complexes or supramolecular aggregates,62 its char-
acterization by X-ray analysis is difficult. In most cases,
hydrogen atoms are calculated with standard distances into
a refined structure of heteroatoms and are practically left unre-
fined.63 Therefore, characteristic distances between donor–
acceptor pairs (such as N/O, O/O) are more reliable and thus
quite useful to distinguish the formation of either a salt or
‘‘neutral’’ co-crystal. Other important factors in known struc-
tures of co-crystals and salts, such as different C–O bond lengths
in carboxylate and carboxylic acid,36a as well as sensitivity of the
C–N–C angle of pyridine to protonation at its nitrogen atom
may also serve to determine the nature of a compound. Notably,
the average ratio of C–O (long) to C–O (short) bond lengths is
ꢁ1.027 for salts and ꢁ1.081 for co-crystals. Corresponding C–
N–C angles on neutral pyridine-based molecules are reported to
be in the range of 117.7–118.5ꢂ, while protonated analogues can
have bond angles over 120ꢂ.9a
In contrast, solid-state NMR provides outstanding selectivity
for local environments, even in the case of ill-defined, powdered
compounds, often facilitating unambiguous identification of
chemically distinct sites based on the corresponding NMR
3216 | CrystEngComm, 2011, 13, 3213–3223
This journal is ª The Royal Society of Chemistry 2011

View Online
Fig. 2 Powder X-ray diffraction pattern of the compounds obtained from co-crystallization of quinidine and 4HBA using either acetone or ethanol
reflecting the well-defined crystalline and apparently ill-defined nature of the samples.
chemical shifts (even in the absence of X-ray data).25,40 In
particular, the sensitivity of 13C (or 15N) chemical shifts to small
changes in the local environment can be considered as a finger-
print to rather quickly identify whether co-crystallization has
produced an intact co-crystal (or salt) or eventually a random
mixture of co-crystal formers, which in the case of simpler
mixtures may be also feasible via IR spectroscopy.33 In the
absence of X-ray data, an NMR-based structure determination
would at first require identification of the asymmetric unit, which
in principle can be deduced from the corresponding 13C-CPMAS
spectra (Fig. 3), provided that sufficient spectral resolution can
be achieved. Notably, for molecular compounds with low space
group symmetry, the integrant unit (which is the first multiple of
the asymmetric unit leading to integer number of atoms reflecting
the stoichiometry of the considered crystal formula) is often
equivalent to the molecule comprising the crystal.
In the case of 1, well-resolved peaks reflecting individual
signals, i.e., due to either quinidine or 4HBA, were found,
thereby tentatively suggesting that the asymmetric unit may be
comprised of one quinidine and one 4HBA molecule, though the
asymmetric unit of 1 as derived from X-ray analysis contains two
quinidine and two 4HBA molecules, respectively. Their local
environments, however, are rather similar so that merely one set
of 13C peaks is observed for both quinidine and 4HBA. Similarly,
this applies to pristine quinidine, thus illustrating that for
molecular crystals, sometimes integer multiples of the initially
guessed number of molecules comprising the asymmetric have to
be considered. The carbon signals of 4HBA in 1 experienced
significant shifts compared to those of the pure form, where in
particular, C49 and C50 shifted from 171.7 to 174.2 ppm and
123.3 to 132.5 ppm, respectively, indicating both different
hydrogen-bonding environments and changes in the ionization
Fig. 3 Solid-state 13C-CPMAS spectra of (a) salt 1, (b) quinidine,24 and (c) p-hydroxybenzoic acid (4HBA). All 13C-CPMAS spectra were collected at
75.5 MHz using a Bruker Avance-II 300 machine with contact time of 2 ms, co-adding 8196 transients. The experiments were carried out using
a standard 4 mm double-resonance MAS probe spinning at 12 kHz, typical p/2 pulse length of 4 ms, and a recycle delay of 5 s. All spectra were acquired
at room temperature, while the given peak assignments are based on DFT computations (see ESI†).
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 3213–3223 | 3217

View Online
state of the carboxylic acid group of 4HBA (i.e., formation of the
carboxylate ion due to proton transfer from 4HBA to quinidi-
ne)35a upon complex formation, in good agreement with the IR
data. In pure 4HBA,27,28 the carboxyl group is engaged in
a dynamic hydrogen-bonding36a typically found for an acid dimer
(Scheme 1a) while in 1 such interaction is clearly absent. Rather,
the proton of the carboxyl group has been transferred from
4HBA to the quinuclidine nitrogen (N24) resulting in salt
formation. Moreover the hydroxyl group of 4HBA (which is
engaged in a hydrogen-bonded charge assisted O–H/Oꢀ
homosynthon with the carboxyl group of another 4HBA mole-
cule in 1) is not involved in any kind of hydrogen bonding in its
pure form. Most 13C signals of quinidine within 1 displayed
minor changes of 1–2 ppm owing to a slightly different molecular
packing, except the signals of C7 (shifted from 113.8 to 118.0
ppm) and C9 (shifted from 107.8 to 99.3 ppm) of the quinoline
ring that have shifted z7 ppm (Fig. 3), reflecting the changed
conformation of the methoxy group (C10) in 1.24 In addition,
significant shifts of 4–5 ppm were also observed for some
aliphatic carbons of the quinuclidine ring possibly owing to
conformational changes of its C]C group.
Fig. 4 13C-CPMAS spectra of both well-defined crystalline (‘‘acetone
phase’’) and rather micro-crystalline (‘‘ethanol phase’’) compounds
obtained from co-crystallization of quinidine and 4HBA in either acetone
or ethanol. Note the slightly increased linewidth and the presence of
smaller ‘‘bumps’’ at the shoulder of some signals in the 13C-CPMAS
spectrum of the ‘‘ethanol phase’’.
even the previously reported co-crystal of quinidine and methyl
paraben.24 While it is obvious that a dimer formation of
carboxylic acid units in 1 can be excluded, the 1H signal at 12.06
ppm cannot be readily assigned. The distance of heavy atoms
within the charge assisted Oꢀ/H–N+ heterosynthon in the salt 1
The 13C chemical shift assignment (Fig. 3) is given based on
DFT computations (at B3LYP/6-311 + G** level of theory)36 of
an optimized cutout from the crystal structure of 1 that includes
representative hydrogen bonding environments (where all non-
hydrogen atoms were frozen on their crystallographic positions)
while ignoring packing effects.66 In particular, we applied the
recently introduced multi-standard (MSD) approach67 which is
sufficiently accurate (particularly for molecular systems) but
computationally less demanding than highly electron-correlated
approaches68 or even full consideration of crystal lattice effects.69
After confirming that the compound obtained from co-crys-
tallization of quinidine and 4HBA in acetone is salt 1 (‘‘acetone
phase’’), an unambiguous identification of the microcrystalline
compound (‘‘ethanol phase’’) was considered. Clearly, the cor-
responding IR spectra of both phases (ESI†) are indistinguish-
able hence indicating an identical nature of the two compounds
even though the powder diffractogram of the ‘‘ethanol phase’’
was highly unresolved (unlike the well-defined ‘‘acetone phase’’).
Likewise, its 13C-CPMAS spectrum exhibited sharp and well-
defined peaks, yet with slightly broader lines than the ‘‘acetone
phase’’ (Fig. 4), where the presence of small ‘‘bumps’’ at the
shoulder of some signals may be caused by more disordered
fractions of the sample. This clearly demonstrates that solid-state
NMR (and to some extent IR spectroscopy) can be efficiently
applied to characterize even those micro-crystalline pharmaceu-
tical co-crystals or salt that cannot be (unambiguously) identified
using X-ray powder diffraction analysis.70
ꢀ
(d(N–O) ¼ 2.661 A) is only slightly larger than the corresponding
distance within the neutral O/H–N heterosynthon (d(N–O) ¼
2.620 A) of the quinidine/methyl paraben co-crystal,24 but in the
ꢀ
latter case, the NH proton resonates at 13.45 ppm. Notably, all
hydrogen-bonds considered here fall within the range of ‘‘clas-
sical’’ moderately strong hydrogen-bonds,62 where the distance
of the heavy atoms is less than the sum of their van-der-Waals
71
ꢀ
ꢀ
radii (N/O 3.22 A, O/O 3.04 A). While moderate ‘‘stretch-
ing’’ of the distance between heavy atoms embracing a hydrogen-
bonded proton typically leads to reduced 1H chemical shifts (i.e.,
indicating weaker bonds),40,72 the introduction of positive
charges often result in strongly increased 1H chemical shifts, and
hence the opposite.45b Within the neutral O/H–N hetero-
synthon of pristine quinidine, a heavy atom distance of d(N–O)
ꢀ
¼ 2.760 A and an NH proton peak at 9.34 ppm were observed.
Since the convenient case of rather linear hydrogen bonding is
considered, we may be tempted to also assume a linear depen-
dence of the 1H chemical shift of the NH proton with respect to
ꢀ
d(N–O), at least in the given range of 2.62 to 2.76 A, although
commonly, a fixed heavy atom distance is applied where the
proton approaches the respective hydrogen-bond acceptor.37
ꢀ
Nevertheless, considering Dd(N–O) ¼ 0.14 A and Dd ¼ 4.01 ppm
ꢀ
(yielding an apparent increment of z0.286 ppm per 0.01 A,
1
Furthermore, H MAS NMR spectra were recorded in order
ignoring charge effects), we can ‘‘predict’’ the 1H chemical shift of
Oꢀ/H–N+ in 1 at 12.16 ꢃ 0.2 ppm, which on one hand is
remarkably close to the experimental value of 12.06 ppm but on
the other hand could be purely fortuitous.
to in detail analyze the local hydrogen bonding environments
and characteristic signatures of the salt identity of both
compounds, which is particularly beneficial in view of NMR-
based structure determination of related compounds in the
possible absence of X-ray data. Similarly to the IR and 13C-
CPMAS spectra, both phases revealed almost identical 1H MAS
NMR spectra including observable linewidths (ESI†). The peak
at 12.06 ppm in the corresponding 1H MAS NMR spectrum of 1
(Fig. 5) is clearly indicative of a different hydrogen-bonding
pattern in 1 compared to both pristine 4HBA or quinidine and
Though the additional 1H MAS NMR signals at 8.3 ppm and
8.78 ppm may be tentatively assigned to the protons within
charge assisted O–H/Oꢀ homosynthons formed between the
hydroxyl groups of both quinidine and 4HBA with a carboxylate
group of 4HBA, possible peak overlap with resonances from
aromatic protons cannot be excluded. However, a DFT 1H
chemical shift computation (at B3LYP/6-311 + G** level of
3218 | CrystEngComm, 2011, 13, 3213–3223
This journal is ª The Royal Society of Chemistry 2011

View Online
Fig. 5 1H MAS NMR spectra of (a) quinidine, (b) 4-hydroxybenzoic acid (4HBA), and (c) salt 1, acquired at 850.1 MHz using a commercially available
Bruker 2.5 mm double resonance MAS probe at a spinning frequency of 29762 Hz, typical p/2 pulse lengths of 2 ms, and a recycle delay of 5–10 s,
co-adding 32 transients.
theory)36 based on an optimized model cutout from the crystal
structure of 1 predicts H chemical shifts of 12.9 ppm (Oꢀ/H–
dynamics (with respect to the timescale of the experiment).
Hence, in order to provide further evidence of co-compound
1
N+), 11 ppm (O–H/Oꢀ, quinidine–OH, COOꢀ of 4HBA) and
11.6 ppm (O–H/Oꢀ, 4HBA–OH, COOꢀ of another 4HBA),
respectively, with ꢃ1.5 ppm for each value, rendering all
hydrogen-bonded protons barely distinguishable. Aromatic
protons were computed at 7–9 ppm, in reasonable agreement
with the experimental data.
formation and peak assignments of the corresponding H MAS
1
NMR spectrum of 1, fast MAS 1H DQ spectra of both the
crystalline ‘‘acetone phase’’ and micro-crystalline ‘‘ethanol phase’’
(cf. ESI†) of salt 1 were recorded at two different recoupling
times (TR/2 ¼ 16.8 ms and TR ¼ 33.6 ms, respectively) (Fig. 6).
According to the hydrogen-optimized model cutout of the crystal
structure, the shortest distance of the NH+ proton to both
1H–1H double-quantum (DQ) NMR is in general a highly
useful and selective approach to identify close contacts or spatial
proximities of structural moieties and can be used to reveal
changes of hydrogen-bonding environments, i.e. upon successful
formation of pharmaceutical co-compounds or salts. In such
a two-dimensional experiment, double-quantum coherences
(DQC) due to pairs of dipolar coupled protons are correlated
with single-quantum coherences resulting in characteristic
correlation peaks. Double-quantum coherences between so-
called like spins appear as a single correlation peak on the
diagonal (‘‘auto-peak’’) while a pair of cross-peaks that are
symmetrically arranged on either side of the diagonal reflect
couplings among unlike spins. DQ peaks appear at the sum
frequency of the two coupled spins and therefore often allow for
an increased spectral resolution.
ꢀ
aromatic and aliphatic protons amounts to d ¼ 2.216 A (Dij ¼
ꢀ
11.04 kHz) and d ¼ 2.366 A (Dij ¼ 9.07 kHz), respectively, while
the quinidine–OH proton has a comparable proximity to
ꢀ
aromatic protons (d ¼ 2.206 A, Dij ¼ 11.19 kHz) but a longer
ꢀ
distance to aliphatic protons (d ¼ 2.658 A, Dij ¼ 6.41 kHz). In
contrast, the 4HBA–OH proton is close to two different aromatic
ꢀ
ꢀ
protons (d ¼ 2.204 A, Dij ¼ 11.22 kHz; d ¼ 2.279 A, Dij ¼ 10.15
ꢀ
kHz) but further away from aliphatic protons (d ¼ 2.904 A, Dij ¼
4.91 kHz). Since the dipolar couplings among the protons are
very sensitive with respect to the distance (cf. static dipolar
couplings Dij given in brackets), this should indeed be reflected
by the observed DQ signal intensities. In particular, the presence
of the DQ peak at 15.36 ppm (12.06 ppm + 3.3 ppm) strongly
1
suggests that the H peak at 12.06 ppm reflects the NH+ proton
In addition, we exploit the fact that observable double-
ꢀ6
which is involved in salt formation (i.e., the proton is transferred
from the carboxyl group of 4HBA to the quinuclidine nitrogen
(N24) of quinidine), as the signal at ꢁ3.3 ppm (in good agree-
2
quantum signal intensities are proportional to Dij or rij
,
respectively (Dij is the homonuclear dipolar coupling constant, rij
the internuclear distance), at least in the limit of short dipolar
recoupling times (i.e., 16.8–33.6 ms). Strong signal intensities in
the corresponding double-quantum spectrum therefore reveal
protons in rather close spatial proximity (i.e. distances up to
1
ment with the DFT H chemical shift computation) represents
aliphatic protons of the quinuclidine ring of quinidine. This is
further supported by a rather strong cross-peak at 18.76 ppm
(12.06 ppm + 6.67 ppm) reflecting the contact of the NH+ proton
with its closest CH moiety (the bridging CHOH). In contrast, the
cross-peak at 20.51 ppm (11.74 ppm + 8.77 ppm) indicates the
3.5 A).73 In contrast, rather weak DQ signals reflect either long-
ꢀ
distance contacts or the presence of fast local molecular
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 3213–3223 | 3219

View Online
Fig. 6 1H–1H DQ MAS NMR spectrum of the salt 1 at 850.1 MHz and 29762 Hz MAS, acquired under the following experimental conditions: s(exc.)
¼
16.8 ms (or 33.6 ms), 64 t1 increments at steps of 33.6 ms, relaxation delay 60 s, 16 transients per increment. Sixteen positive contour levels between 10%
and 100% (or 15% and 100%) of the maximum peak intensity were plotted. The F2 projection is shown on the top; the most important DQ cross-peaks
are highlighted.
presence of an additional hydrogen-bonded proton site whose
1
resonance in the regular H MAS NMR spectrum most likely
either so-called DQ spinning sideband pattern75 or DQ signal
build-up curves76 are generated. Since the corresponding 1H DQ
MAS spectra of both the ‘‘acetone phase’’ and ‘‘ethanol phase’’ of
1 are virtually identical while corroborating the presence of NH+
(hence indicating salt formation), we refrained from a full
determination of the hydrogen substructures of the charge-
assisted synthons. Nevertheless, the hydrogen bonding parame-
ters obtained from the partially optimized representative cutout
from the crystal structure for the corresponding synthons
is buried under the fairly broad (full width at half
height z1.1 ppm) but slightly asymmetric peak centered at 12.06
1
ppm. Based on the H chemical shift computation, this site is
assigned to the 4HBA–OH protons (computed at 11.6 ppm),
whose closest aromatic protons are computed at 8.9 ppm. Two
further moderately strong cross-peaks present in the 1H DQ
MAS spectrum of 1 applying the shortest possible dipolar
recoupling time of TR/2 (in our case 16.8 ms) at 18.76 ppm (8.77
ppm + 9.99 ppm) and (8.35 ppm + 10.41 ppm), however, cannot
be unambiguously assigned. While at first glance it may appear
feasible to attribute the ‘‘obtained’’ peaks at either 9.99 ppm or
10.41 ppm to the quinidine–OH proton (computed at 11 ppm),
ꢀ
+
ꢀ
ꢀ
O /H–N (d(N/O) ¼ 2.661 A, d(N–H) ¼ 1.067 A, <(NHO) ¼
161.2ꢂ), O–H/Oꢀ (4HBA–OH, COOꢀ of another 4ꢂHBA;
ꢀ
ꢀ
d(O/O) ¼ 2.618 A, d(O–H) ¼ 0.989 A, <(OHO) ¼ 173.0 ) and
ꢀ
ꢀ
ꢀ
O–H/O (quinidine–OH, COO of 4HBA; d(O/O) ¼ 2.642 A,
ꢂ
ꢀ
d(O–H) ¼ 1.005 A, <(OHO) ¼ 169.7 ) are fairly similar to
previously reported data of binary co-crystals and salts,53b indi-
cating that the respective proton within the O–H/Oꢀ synthons
is involved in almost ideally linear hydrogen bonding (bond
angle z170 to 180ꢂ).
1
such resonances cannot be identified in the corresponding H
MAS NMR spectrum of 1. Therefore, those signals possibly
reflect multi-spin effects such as ‘‘DQ relay’’ where an initially
formed DQ coherence couples with at least a third spin yielding
apparent DQ peaks (i.e., at the sum of three chemical shifts).
This, however, requires sufficiently strong dipolar couplings
among the spins and a suitable geometry (hence spatial prox-
imity), which in principle could be exploited for NMR-based
structure determination, but is rarely applied to organic samples.
Further insight into the structural environment of both crys-
talline ‘‘acetone phase’’ and micro-crystalline ‘‘ethanol phase’’ of
1 was obtained from 15N CPMAS NMR. The 15N chemical shift
is rather sensitive to packing or coordination effects and benefits
from a larger chemical shift range and pronounced anisotropic
properties thereby providing superior resolution, particularly in
cases where nitrogen atoms are partially protonated. Indeed,
a reliable interpretation of such data requires comparison with
similar known compounds or quantum-chemical shift compu-
tations of model structures.77 In the case of 1, two resonances
at ꢀ74.2 ppm and ꢀ337.0 ppm with rather narrow linewidths
about 60 Hz were observed in the 1D 15N CPMAS spectrum
1
Nevertheless, multi-spin effects could be further probed by H
triple-quantum MAS NMR,74 which is beyond the scope of this
work.
In favourable cases, particularly in the case of dipolar coupled
clusters (i.e., triple- or quadruple hydrogen-bonded moieties),
selected internuclear proton–proton distances (derived from
1H–1H dipolar couplings) may be quantitatively determined if
3220 | CrystEngComm, 2011, 13, 3213–3223
This journal is ª The Royal Society of Chemistry 2011

View Online
(ESI†), indicating a high degree of crystallinity42 and magnetic
equivalence of the two quinidine molecules contained in the
asymmetric unit of 1, whereas the corresponding 15N CPMAS
spectrum of pure quinidine displays signals at ꢀ73.6 ppm
and ꢀ344.0 ppm. Notably, the minor shift of 0.6 ppm of the
signal at 74.2 ppm in 1 (assigned to the quinoline ring)24 with
respect to pure quinidine clearly reveals that the nitrogen atom
(N23) of the quinoline ring in 1 does not engage in hydrogen
bonding, similarly to its pure form. In contrast, the remaining
signal at ꢀ337.0 ppm (assigned to N24 atom of quinuclidine ring
in 1)24 is shifted ꢁ7 ppm to higher ppm compared to the chemical
shift of N24 (ꢀ344.0 ppm) in pure quinidine. Notably,
the computed 15N chemical shifts of ꢀ66.8 ppm (N23) and
ꢀ343.4 ppm (N24), respectively, are in reasonable agreement
with the experimentally observed values (within ꢃ7.5 ppm),
particularly the marginal upfield shift upon co-crystallization.
This in turn implies that the DFT optimized hydrogen-bonding
environment taken from the crystal structure is rather
was almost efficient in predicting the salt formation. The
obtained salt crystallizes in a monoclinic space group [P21 (no.
ꢀ
ꢀ
ꢀ
4), Z ¼ 8, a ¼ 6.914 A, b ¼ 36.197 A, c ¼ 9.476 A and b ¼ 92.126]
with significantly larger b-axis, where the asymmetric unit is
comprised of two quinidine and two 4HBA molecules. In addi-
tion, a micro-crystalline, less-defined sample of salt 1 was
obtained from rapid co-crystallization in ethanol, and success-
fully identified via IR spectroscopy and multinuclear solid-state
NMR. The results were discussed with respect to ‘‘NMR-based
crystallography’’ of structurally less-defined co-compounds,
where an interpretation of the obtained NMR data was sup-
ported by DFT quantum-chemical computations, thereby illus-
trating the great potential of solid-state NMR for
complementary application in the fast screening and structural
analysis of pharmaceutical complexes79 (including co-crystals
and salts) obtained under similar conditions (i.e., fast evapora-
tion).
representative. Similar to the H and 13C NMR spectra, the 15N
1
4. Experimental section
CPMAS spectra of the crystalline (‘‘acetone phase’’) and micro-
crystalline (‘‘ethanol phase’’) compound obtained from co-crys-
tallization of quinidine and 4HBA are rather comparable except
for slightly increased linewidths (about 70 Hz rather than 60 Hz)
in the latter case.
Both quinidine ((9S)-60-methoxycinchonan-9-ol and p-hydrox-
ybenzoic acid (4HBA) were purchased from Aldrich and used as
obtained. Pharmaceutical salt of 4HBA and quinidine were
prepared by dissolving 1 mmol of quinidine (324.4 mg) and
1 mmol of 4HBA (138.12 mg) in 50 ml acetone. The solution was
left for slow evaporation in an open container (50 ml crystallizing
dish). After two days, colourless well-defined prism-like crystals
were obtained and subsequently ground to powder for structural
characterization via powder X-ray diffraction and solid-state
NMR. In order to obtain sufficiently large crystals suitable for
single crystal X-ray analysis, the same solution was left for slow
evaporation in a test tube. However, when ethanol was used
under similar conditions except fast evaporation (within 1–
2 days) of the solution in a crystallizing dish, colourless glassy
substance sticking to the wall of the dish was obtained. Upon
scratching a white, ill-defined rather micro-crystalline powder (as
identified by powder X-ray diffraction) could be retrieved which
was later identified as salt 1 via solid-state NMR. Nevertheless,
when the same solution of quinidine and 4HBA (in ethanol) was
left for much slower evaporation over a period of 2–3 weeks in an
open glass tube, then well-defined white micro-crystalline powder
(confirmed by powder X-ray diffraction data) similar to the one
obtained by acetone were formed. However, the quality of those
crystals was not good enough to be analyzed by single crystal X-
ray diffraction.
According to a recent report,78 for a heterocyclic and fairly
basic nitrogen, an upfield shift (i.e., larger negative ppm values)
of about 20 up to 40 ppm has been observed in the case of strong
hydrogen-bonding, while upfield shifts of even 80 ppm or more
may occur if a proton is transferred from a donor (such as
carboxylic acid) to an acceptor nitrogen. Notably, similar trends
were also observed in the reported co-crystal of quinidine and
methylparaben,24 i.e., a marginal upfield shift (ꢁ2.5 ppm) was
found when replacing a moderate hydrogen bond to a compara-
tively strong hydrogen bond. However, in the case of 1, the
quinuclidine nitrogen (N24) is rather protonated, as revealed by
its single crystal and NMR analysis. This finding may be
attributed to two facts: at one hand the presumed upfield shift
of ꢁ80 ppm49 may be preferably observed in cases where
a previously free nitrogen atom is protonated. The quinuclidinic
nitrogen (N24), however, was not free even in its pure form (N24
is hydrogen-bonded to the hydroxyl group of another quinidine
molecule and exhibits an 15N chemical shift of ꢀ344 ppm). On the
other hand, observable upfield shifts tend to increase upon
contraction of the corresponding N–O distances, thus yielding
stronger hydrogen bonds in neutral compounds (co-crystals),
while the opposite trend is found for charged species (salts).49
Since the N–O distance in the case of pure quinidine amounts to
Solid-state NMR methods
ꢀ
ꢀ
2.76 A, which shrinks to 2.66 A upon co-crystallization with
4HBA, the shift of ꢁ7 ppm to higher ppm could well be
explained.
Proton solid-state NMR data were recorded at 850.1 MHz
employing a Bruker Avance III spectrometer, while 13C-CPMAS
spectra were recorded at 75.5 MHz using a Bruker Avance-II 300
machine. Most experiments were carried out using a commer-
cially available Bruker 2.5 mm double-resonance MAS probe at
a spinning frequency of 29762 Hz, typical p/2-pulse lengths of
2 ms and recycle delays of 5–10 s. The spectra were referenced
with respect to tetramethylsilane (TMS) using solid adamantane
as the secondary standard (1.63 ppm for 1H and 29.456 ppm for
13C). In addition, 15N-CPMAS spectra were recorded at 30.4
MHz using a Bruker Avance-II 300 machine and referenced to
3. Conclusion
In this work, we have explored the predictability of resulting
structures of a multi-component pharmaceutical model complex
based on 4-hydroxybenzoic acid (4HBA) and quinidine, an anti-
malarial constituent of Cinchona tree bark. Though the salt is
stabilized by a slightly different set of heterosynthons as
proposed based on crystal engineering principles, the concept
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 3213–3223 | 3221

View Online
solid 15NH4Cl (ꢀ341.0 ppm). If not stated otherwise, all spectra
were collected at room temperature. The back-to-back (BaBa)80
recoupling sequence was used to excite and reconvert double-
quantum coherences, applying States-TPPI81 for phase sensitive
detection. Further details are given in the figure captions of the
respective 2D-spectra.
4 (a) G. R. Desiraju, Angew. Chem., Int. Ed., 2007, 46, 8342–8356; (b)
G. R. Desiraju, Angew. Chem., Int. Ed. Engl., 1995, 34, 2311–2327;
(c) G. R. Desiraju, Nat. Mater., 2002, 1, 77–79; (d) G. R. Desiraju,
Nature, 2001, 412, 397–400.
5 (a) 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; (b) 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; (c)
L. S. Reddy, N. B. Jagadeesh and A. Nangia, Chem. Commun.,
2006, 1369–1371.
DFT-based chemical shift calculations
6 (a) S. L. Childs and M. J. Zaworotko, Cryst. Growth Des., 2009, 9,
4208–4211; (b) N. Shan and M. J. Zaworotko, Drug Discovery
Today, 2008, 13, 440–446; (c) D. R. Weyna, T. Shattock,
P. Vishweshwar and M. J. Zaworotko, Cryst. Growth Des., 2009, 9,
1106–1123; (d) J. A. Bis, P. Vishweshwar, D. Weyna and
M. J. Zaworotko, Mol. Pharmaceutics, 2007, 4, 401–416.
7 (a) S. H. Jeong, Y. Takaishi, Y. Fu and K. Park, J. Mater. Chem.,
2008, 18, 3527–3535; (b) 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; (c) A. V. Trask, Mol. Pharmaceutics, 2007, 4, 301–309.
The proton positions of a selected fragment of the crystal
structure reflecting representative hydrogen-bonding environ-
ments were optimized with all heavy atoms fixed at the crystal-
lographic positions via DFT quantum chemical calculations,
applying the B3LYP functional and 6-311G82 split valence basis
set augmented with diffuse and polarization functions. Subse-
quently, H, 13C and 15N chemical shifts with respect to either
1
1
tetramethylsilane (TMS, H), benzene (13C) and methanol (13C)
or nitromethane (15N) were computed at B3LYP/6-311 + G**
level of theory with the GIAO approach as implemented in the
Gaussian03 program.83 Note that the recently introduced multi-
standard approach is applied in the case of 13C.67
€
8 (a) C. B. Aakeroy and K. R. Seddon, Chem. Soc. Rev., 1993, 22, 397–
407; (b) J. Bernstein, M. C. Etter and L. Leiserowitz, Struct. Correl.,
1994, 2, 431–507; (c) K. Biradha, CrystEngComm, 2003, 5, 374–384;
€
(d) C. B. Aakeroy, J. Desper and M. E. Fasulo, CrystEngComm,
2006, 8, 586–588.
€
9 (a) C. B. Aakeroy, M. E. Fasulo and J. Desper, Mol. Pharmaceutics,
2007, 4, 317–322; (b) S. L. Childs, G. P. Stahly and A. Park, Mol.
Pharmaceutics, 2007, 4, 323–338.
Single crystal structure analysis
€
10 (a) C. B. Aakeroy, A. M. Beatty and B. A. Helfrich, Angew. Chem.,
Int. Ed., 2001, 40, 3240–3242; (b) C. B. Aakeroy and D. J. Salamon,
Crystal parameters of 1 are reported as follows: colourless prism-
€
like crystals, which were crystallized in a monoclinic space group
ꢀ
CrystEngComm, 2005, 7, 439–448.
ꢀ
ꢀ
[P21 (no. 4), Z ¼ 8, a ¼ 6.914 A, b ¼ 36.197 A, c ¼ 9.476 A and
b ¼ 92.126]. Data collection at 120 K was done on a Nonius
11 (a) M. Khan, V. Enkelmann and G. Brunklaus, CrystEngComm,
2009, 11, 1001–1005; (b) L. R. MacGillivary, J. L. Reid and
J. A. Ripmeester, J. Am. Chem. Soc., 2000, 122, 7817–7818.
12 (a) O. Almarsson and M. J. Zaworotko, Chem. Commun., 2004, 1889–
1896; (b) 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.
13 P. G. Karamertzanis, A. V. Kazantsev, I. Nizar, G. W. A. Welch,
C. S. Adjiman, C. C. Pantelides and S. L. Price, J. Chem. Theory
Comput., 2009, 5, 1432–1448.
ꢀ
KCCD diffractometer (MoKa (l ¼ 0.71073 A)), equipped with
a graphite monochromator. The intensity data were corrected for
Lorentz and polarization effects, while structure solution and
refinement were performed employing the SHELXS8684 and
CRYSTALS85 software packages. All non-hydrogen atoms were
refined in the anisotropic approximation against F of all
observed reflections. The hydrogen atoms were refined in the
riding mode with fixed isotropic temperature factors; 1: R-factor
(%) ¼ 3.74.
14 M. Khan, V. Enkelmann and G. Brunklaus, Cryst. Growth Des., 2009,
9, 2354–2362.
15 K. Chow, H. H. Y. Tong, S. Lum and A. H. L. Chow, J. Pharm. Sci.,
2008, 97, 2855–2877.
16 T. Steiner and G. R. Desiraju, Chem. Commun., 1998, 891–892.
17 I. Rozas, Phys. Chem. Chem. Phys., 2007, 9, 2782–2790.
18 J. M. Lehn, Supramolecular Chemistry: Concepts and Perspectives,
Wiley, 1995.
19 M. Meot-Ner, Chem. Rev., 2005, 105, 213–284.
20 M. D. Ward, Struct. Bonding, 2009, 132, 1–23.
21 (a) P. Gilli, V. Bertolasi, V. Ferretti and G. Gilli, J. Am. Chem. Soc.,
1994, 116, 909–915; (b) S. J. Grabowski, Annu. Rep. Prog. Chem.,
Sect. C, 2006, 102, 131–165.
Infrared spectroscopy
Infrared absorption spectra were obtained at a resolution of
4 cmꢀ1 using a Perkin Elmer FTIR BXII model Fourier trans-
form infrared spectrometer using potassium bromide pellets of
the compounds.
22 R. Taylor and O. Kennard, Acc. Chem. Res., 1984, 17, 320–326.
€
Acknowledgements
23 C. B. Aakeroy and N. Schultheiss, in Making Crystals by Design,
Wiley-VCH, Weinheim, 2007, pp. 209–240.
Financial support from the Deutsche Forschungsgemeinschaft
(DFG) through the SFB 625 in Mainz is gratefully acknowl-
edged.
24 M. Khan, V. Enkelmann and G. Brunklaus, J. Am. Chem. Soc., 2010,
132, 5254–5263.
25 NMR Crystallography, ed. R. K. Harris, R. Wasylishen and M. Duer,
Wiley, Chichester, 2009.
26 The GRAS list, http://www.fda.gov/Food/FoodIngredientsPackaging/
GenerallyRecognizedasSafeGRAS/default.htm.
27 B. Sarma, N. K. Nath, B. R. Bhogala and A. Nangia, Cryst. Growth
Des., 2009, 9, 1546–1557.
28 B. R. Sreekanth, P. Vishweshwar and K. Vyas, Chem. Commun.,
2007, 2375–2377.
29 S. R. Byrn, W. Xu and A. W. Newman, Adv. Drug Delivery Rev.,
2001, 48, 115–136.
30 T. R. Shattock, K. D. Arora, P. Vishweshwar and M. J. Zaworotko,
Cryst. Growth Des., 2008, 8, 4533–4545.
31 (a) A. M. Wahbi, M. S. Moneed, I. I. Hewala and M. F. Bahnasy,
Chem. Pharm. Bull., 2008, 56, 787–879; (b) S. Kashino and
References
1 S. Dutta and D. J. W. Grant, Nat. Rev. Drug Discovery, 2004, 3, 42–
47.
2 (a) N. Schultheiss and A. Newman, Cryst. Growth Des., 2009, 9, 2950–
2967; (b) S. Byrn, R. Pfeiffer, M. Ganey, C. Hoiberg and
G. Poochikian, Pharm. Res., 1995, 12, 945–954; (c) Z. Ma and
B. Moulton, J. Chem. Crystallogr., 2009, 39, 913–918.
3 (a) N. Blagden, M. De Metas, P. T. Gavan and P. York, Adv. Drug
Delivery Rev., 2007, 59, 617–630; (b) L. F. Huang and W. Q. Tong,
Adv. Drug Delivery Rev., 2004, 56, 321–324.
3222 | CrystEngComm, 2011, 13, 3213–3223
This journal is ª The Royal Society of Chemistry 2011

View Online
M. Haisa, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1983,
39, 310–312; (c) B. Pniewska and A. Suszko-Purzycka, Acta
Crystallogr., Sect. C: Cryst. Struct. Commun., 1989, 45, 638–642.
32 R. A. Smith, Expert Opin. Pharmacother., 2006, 7, 2581–2591.
33 (a) A. C. Moffat, S. Assi and R. A. Watt, J. Near Infrared Spectrosc.,
2010, 18, 1–15; (b) A. J. O’Neill, R. D. Jee, G. Lee, A. Charvill and
A. Moffat, J. Near Infrared Spectrosc., 2008, 16, 327–333.
34 (a) D. Reichert, Annu. Rep. NMR Spectrosc., 2005, 55, 159–203; (b)
S. E. Ashbrook and M. E. Smith, Chem. Soc. Rev., 2006, 35, 718–
735; (c) S. P. Brown, Prog. Nucl. Magn. Reson. Spectrosc., 2007, 50,
199–251.
35 (a) F. G. Vogt, J. S. Clawson, M. Strohmeier, A. J. Edwards,
T. N. Pham and S. A. Watson, Cryst. Growth Des., 2009, 9, 921–
937; (b) F. G. Vogt, J. A. Vena, M. Chavda, J. S. Clawson,
M. Strohmeier and M. E. Barnett, J. Mol. Struct., 2009, 932, 16–30.
36 (a) M. Khan, G. Brunklaus, V. Enkelmann and H. W. Spiess, J. Am.
Chem. Soc., 2008, 130, 1741–1748; (b) M. Khan, V. Enkelmann and
G. Brunklaus, J. Org. Chem., 2009, 74, 2261–2270; (c)
A. M. Orendt and J. C. Facelli, Annu. Rep. NMR Spectrosc., 2007,
62, 115–178.
S. A. Bourne, Acta Crystallogr., Sect. B: Struct. Sci., 2008, 64, 780–
790; (c) P. Vishweshwar, A. Nangia and V. M. Lynch, Cryst.
Growth Des., 2003, 3, 783–790.
55 D. C. Warhurst, J. C. Craig, I. S. Adagul, D. J. Meyer and S. Y. Lee,
Malar. J., 2003, 2, 1–14.
56 (a) A. J. C. Cabeza, G. M. Day and W. Jones, Chem.–Eur. J., 2008,
14, 8830–8836; (b) A. J. C. Cabezy, G. M. Day,
W. D. S. Motherwell and W. Jones, J. Am. Chem. Soc., 2006, 128,
14466–14467.
57 S. L. Price, Acc. Chem. Res., 2009, 42, 117–126.
58 M. Habgood and S. L. Price, Cryst. Growth Des., 2010, 10, 3263–
3272.
59 B. M. Kariuki, C. L. Bauer, K. D. M. Harris and S. J. Teat, Angew.
Chem., Int. Ed., 2000, 39, 4485–4488.
60 L. H. Wei, Acta Crystallogr., Sect. E: Struct. Rep. Online, 2006, 62,
4506–4507.
61 (a) R. A. Weatherhead-Kloster, H. D. Selby, W. B. Miller, III and
E. A. Mash, J. Org. Chem., 2005, 70, 8693–8702; (b) K. K. Arora
and V. R. Pedireddi, J. Org. Chem., 2003, 68, 9177–9185; (c)
L. R. MacGillivray, J. Org. Chem., 2008, 73, 3311–3317.
62 N. Tosa, A. Bende, R. A. Varga, A. Terec, I. Bratu and I. Grosu,
J. Org. Chem., 2009, 74, 3944–3947.
63 T. Steiner, Angew. Chem., Int. Ed., 2002, 41, 48–76.
64 E. Bouveresse, C. Casolino and C. de la Pezuela, J. Pharm. Biomed.
Anal., 1998, 18, 35–42.
37 R. K. Harris, Solid State Sci., 2004, 6, 1025–1037.
38 H. Szatylowicz, J. Phys. Org. Chem., 2008, 21, 897–914.
39 (a) H. H. Limbach, in Hydrogen Transfer Reactions, ed. J. T. Hynes, J.
P. Klinman, H. H. Limbach and R. L. Schowen, Wiley-VCH, 2007;
(b) S. P. Brown and H. W. Spiess, Chem. Rev., 2001, 101, 4125–
ꢁ
4155; (c) R. Gobetto, C. Nervi, E. Valfre, M. R. Chierotti,
65 (a) H. G. Brittain, Cryst. Growth Des., 2009, 9, 3497–3503; (b)
€
C. B. Aakeroy, D. J. Salmon, M. M. Smith and J. Desper, Cryst.
Growth Des., 2006, 6, 1033–1042.
D. Braga, L. Maini, F. Grepioni, R. K. Harris and P. Y. Ghi,
Chem. Mater., 2005, 17, 1457–1466.
40 (a) R. K. Harris, in Encyclopedia of Nuclear Magnetic Resonance,
Wiley, Chichester, 1996, vol. 5, p. 3314; (b) R. K. Harris,
Y. Phuong, B. Robert, C. Y. Ma and K. J. Roberts, Chem.
Commun., 2003, 2834–2835.
41 T. Emmler, S. Gieschler, H. H. Limbach and G. Buntkowsky, J. Mol.
Struct., 2004, 700, 29–38.
42 T. Gorelik, G. Matveeva, U. Kolb, T. Schleuß, A. F. M. Kilbinger,
J. van de Streek, A. Bohle and G. Brunklaus, CrystEngComm,
2010, 12, 1824–1832.
66 (a) P. Lazzeretti, Prog. Nucl. Magn. Reson. Spectrosc., 2000, 36,
1–88; (b) J. A. N. F. Gomes and R. B. Mallion, Chem. Rev., 2001,
101, 1349.
67 A. M. Sarotti and S. C. Pellegrinet, J. Org. Chem., 2009, 74, 7254–
7260.
68 J. Gauss and J. F. Stanton, Adv. Chem. Phys., 2002, 123, 355–422.
69 D. Stueber, Concepts Magn. Reson., Part A, 2006, A28, 347–368.
70 K. D. M. Harris and E. Y. Cheung, Chem. Soc. Rev., 2004, 33, 526–
538.
43 J. P. Bradley, C. Tripon, C. Filip and S. P. Brown, Phys. Chem. Chem.
Phys., 2009, 11, 6941–6952.
44 M. Schulz-Dobrick, T. Metzroth, H. W. Spiess, J. Gauss and
I. Schnell, ChemPhysChem, 2005, 6, 315–327.
71 P. A. Frey, Magn. Reson. Chem., 2001, 39, S190–S198.
72 M. R. Chierotti and R. Gobetto, Chem. Commun., 2008, 1621–
1634.
73 J. P. Bradley, C. Tripon, C. Filip and S. P. Brown, Phys. Chem. Chem.
Phys., 2009, 11, 6941–6952.
€
45 (a) I. Schnell, B. Langer, S. H. M. Sontjens, R. P. Sijbesma,
M. H. P. van Genderen and H. W. Spiess, Phys. Chem. Chem.
Phys., 2002, 4, 3750–3758; (b) I. Bolz, C. Moon, V. Enkelmann,
G. Brunklaus and S. Spange, J. Org. Chem., 2008, 73, 4783–4793.
46 (a) P. Geerlings, F. De Proft and W. Langenaeker, Chem. Rev., 2003,
103, 1793–1873; (b) R. G. Parr and W. Yang, Density Functional
Theory of Atoms and Molecules, Oxford University Press, Oxford, 1989.
47 (a) B. Elena, G. Pintacuda, N. Mifsud and L. Emsley, J. Am. Chem.
Soc., 2006, 128, 9555–9560; (b) C. J. Pickard, E. Salager,
G. Pintacuda, B. Elena and L. Emsley, J. Am. Chem. Soc., 2007,
129, 8932–8933; (c) F. Taulelle, Solid State Sci., 2004, 6, 1053–1057;
(d) J. Senker, L. Seyfarth and J. Voll, Solid State Sci., 2004, 6,
1039–1052.
74 I. Schnell, A. Lupulescu, S. Hafner, D. E. Demco and H. W. Spiess,
J. Magn. Reson., 1998, 133, 61–69.
75 G. P. Holland, B. R. Cherry and T. M. Alam, J. Magn. Reson., 2004,
167, 161–167.
76 L. Seyfarth and J. Senker, Phys. Chem. Chem. Phys., 2009, 11, 3522–
3531.
77 P. Lorente, I. G. Shenderovich, N. S. Golubev, G. S. Denisov,
G. Buntkowsky and H. H. Limbach, Magn. Reson. Chem., 2001,
39, 18–29.
78 Z. J. Li, Y. Abramov, J. Bordner, J. Leonard, A. Medek and
A. V. Trask, J. Am. Chem. Soc., 2006, 128, 8199–8210.
79 C. A. Lepre, J. M. Moore and J. W. Peng, Chem. Rev., 2004, 104,
3641–3675.
80 (a) J. Geen, J. Titman, J. Gottwald and H. W. Spiess, Chem. Phys.
Lett., 1994, 227, 79–86; (b) J. Gottwald, D. E. Demco, R. Graf and
H. W. Spiess, Chem. Phys. Lett., 1995, 243, 314–323; (c)
48 (a) R. K. Harris, J. Pharm. Pharmacol., 2007, 59, 225–239; (b)
R. K. Harris, Analyst, 2006, 131, 351–373; (c) J. Thun, L. Seyfarth,
J. Senker, R. E. Dinnebier and J. Breu, Angew. Chem., Int. Ed.,
2007, 46, 6729–6731.
€
K. Saalwachter, R. Graf and H. W. Spiess, J. Magn. Reson., 2001,
148, 398–418.
49 H. Hamaed, J. M. Pawlowski, B. F. T. Cooper, R. Fu, S. H. Eichhorn
and R. W. Schurko, J. Am. Chem. Soc., 2008, 130, 11056–11065.
50 E. Salanger, R. S. Stein, C. J. Pickard, B. Elena and L. Emsley, Phys.
Chem. Chem. Phys., 2009, 11, 2610–2621.
81 D. Marion, M. Ikura, R. Tschudin and A. Bax, J. Magn. Reson.,
1989, 85, 393–399.
51 L. Seyfarth, J. Sehnert, N. E. A. El-Gamel, W. Milius, E. Kroke,
J. Breu and J. Senker, J. Mol. Struct., 2008, 889, 217–228.
52 F. H. Allen, Acta Crystallogr., Sect. B: Struct. Sci., 2002, 58, 380–388.
53 (a) M. Sharmarke, D. A. Tocher, M. Vickers, P. G. Karamertzains
and S. L. Price, Cryst. Growth Des., 2009, 9, 2881–2889; (b)
82 R. Krishnan, J. S. Binkley, R. Seger and J. A. Pople, J. Chem. Phys.,
1980, 72, 650–654.
83 M. J. Frisch, et al., Gaussian 03 (Revision D.02), Gaussian, Inc.,
Wallingford, CT, 2004.
84 G. M. Sheldrick, SHELXS-86, Program Package for Crystal
€
C. B. Aakeroy, I. Hussain, S. Forbes and J. Desper,
CrystEngComm, 2007, 9, 46–54.
€
Structure Solution and Refinement, Univeristat Gottingen, Germany,
€
1986.
85 P. W. Betteridge, J. R. Carruthers, R. I. Cooper, K. Prout and
D. J. Watkin, J. Appl. Crystallogr., 2003, 36, 1487–1487.
€
54 (a) C. B. Aakeroy, A. M. Beatty and B. A. Helfrich, Cryst. Growth
Des., 2003, 3, 159–165; (b) A. Lemmerer, N. B. Bathori and
ꢂ
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 3213–3223 | 3223