Accessing Centnerszwer's quasiracemate-molecular shape controlled molecular recognition

Article abstract of DOI:10.1039/c6ra08131b

M. Centnerszwer's seminal 1899 report investigated the stereochemical relationship between optical antipodes of different substances using melting-point behavior. One intriguing melting-point phase diagram produced from this early investigation combined (+)-2-chlorosuccinic acid [(+)-1] and (-)-2-bromosuccinic acid [(-)-2]. While Centnerszwer's data clearly indicates the formation of a quasiracemic phase-i.e., materials constructed from pairs of isosteric molecules of opposite handedness-at the 1:1 component ratio, this material is energetically less favorable than the chiral counterparts. The consequence of this crystal instability is significant as evident by the absence of literature sited crystal structures for the quasiracemic phase (+)-1/(-)-2 and racemates (±)-1 and (±)-2. This study circumvented this challenge by generating multi-molecular assemblies using additional crystallizing agents capable of complementing the hydrogen-bond abilities of succinic acids 1 and 2. Both imidazole (Im) and 4,4′-bipyridyl-N,N′-dioxide (BPDO) served as tailor-made additives that effectively modified the crystal packing landscape of quasiracemate of (+)-1/(-)-2. Combining imidazole with the quasiracemate, racemate, and enantiopure forms of 1 and 2 resulted in crystal structures characterized as molecular salts with layered motifs formed from highly directional N⁺-H?carboxylate and carboxyl?carboxylate interactions. In contrast to the enantiopure [(+)-1·Im and (-)-2·Im] and racemic [(±)-1·Im and (±)-2·Im] systems, neighboring molecular layers observed in quasiracemate (+)-1/(-)-2·Im are organized by approximate inversion symmetry. Assessment of the crystal packing efficiency for this series of molecular salts via crystal densities and packing coefficients (C_k) indicates imidazole greatly alters the crystal landscape of the system in favor of racemic and quasiracemic crystal packing. A similar desymmetrized crystal environment was also realized for the ternary cocrystalline system of (+)-1/(-)-2·BPDO where the components organize via N⁺-O^-?carboxyl contacts. This study underscores the importance of molecular shape to molecular recognition processes and the stabilizing effect of tailor-made additives for creating new crystalline phases of previously inaccessible crystalline materials.

Full text of DOI:10.1039/c6ra08131b

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Accessing Centnerszwer's quasiracemate –
molecular shape controlled molecular recognition†
Cite this: RSC Adv., 2016, 6, 64921
*
Jacqueline M. Spaniol and Kraig A. Wheeler
M. Centnerszwer's seminal 1899 report investigated the stereochemical relationship between optical
antipodes of different substances using melting-point behavior. One intriguing melting-point phase
diagram produced from this early investigation combined (+)-2-chlorosuccinic acid [(+)-1] and (ꢀ)-2-
bromosuccinic acid [(ꢀ)-2]. While Centnerszwer's data clearly indicates the formation of a quasiracemic
phase – i.e., materials constructed from pairs of isosteric molecules of opposite handedness – at the
1 : 1 component ratio, this material is energetically less favorable than the chiral counterparts. The
consequence of this crystal instability is significant as evident by the absence of literature sited crystal
structures for the quasiracemic phase (+)-1/(ꢀ)-2 and racemates (ꢁ)-1 and (ꢁ)-2. This study
circumvented this challenge by generating multi-molecular assemblies using additional crystallizing
agents capable of complementing the hydrogen-bond abilities of succinic acids 1 and 2. Both imidazole
(Im) and 4,40-bipyridyl-N,N0-dioxide (BPDO) served as tailor-made additives that effectively modified the
crystal packing landscape of quasiracemate of (+)-1/(ꢀ)-2. Combining imidazole with the quasiracemate,
racemate, and enantiopure forms of 1 and 2 resulted in crystal structures characterized as molecular
salts with layered motifs formed from highly directional N+–H/carboxylate and carboxyl/carboxylate
interactions. In contrast to the enantiopure [(+)-1$Im and (ꢀ)-2$Im] and racemic [(ꢁ)-1$Im and (ꢁ)-2$Im]
systems, neighboring molecular layers observed in quasiracemate (+)-1/(ꢀ)-2$Im are organized by
approximate inversion symmetry. Assessment of the crystal packing efficiency for this series of molecular
salts via crystal densities and packing coefficients (Ck) indicates imidazole greatly alters the crystal
landscape of the system in favor of racemic and quasiracemic crystal packing. A similar desymmetrized
crystal environment was also realized for the ternary cocrystalline system of (+)-1/(ꢀ)-2$BPDO where
the components organize via N+–Oꢀ/carboxyl contacts. This study underscores the importance of
molecular shape to molecular recognition processes and the stabilizing effect of tailor-made additives
for creating new crystalline phases of previously inaccessible crystalline materials.
Received 30th March 2016
Accepted 24th June 2016
DOI: 10.1039/c6ra08131b
www.rsc.org/advances
appeal of these materials stems from their ability to form
predictable motifs that mimic the structural patterns observed
in their racemic counterparts.7–9 Crystal structure reports of
small molecule racemates reveal a strong preference (>92%) for
the components to align with inversion symmetry.10–12 While
this penchant for centrosymmetry also translates to all known
quasiracemic structures, quasiracemic assemblies take on
rigorously non-centrosymmetric patterns due to the structural
features of the two chemically distinct components.
The various denitions of quasiracemic compounds found
in the literature are decidedly vague, with each describing
a common prerequisite linked to the selection of pairs of
chemically different molecules of opposite chirality.3,13,14 This
lack of strict descriptors allows for broad exibility in the design
and selection of quasienantiomer components. Coupled with
the organizational control exhibited by quasiracemic materials,
it is foreseeable this method could lend itself to developing new
materials with improved properties. Though several current
science areas have been realized or enhanced by implementing
1 Introduction
Quasiracemates (or quasiracemic materials) – compounds
constructed from equimolar portions of near enantiomers –
pose intriguing technological opportunities to examine the
impact of molecular shape to molecular alignment. These
materials and the quasiracemic approach to supramolecular
assembly continue to attract a broad range of disciplines that
seek to probe and exploit molecular recognition proles. Recent
investigations have successfully applied the quasiracemic
approach to materials and processes such as those focused on
asymmetric assessment,1,2 synthesis,3,4 and separation.5,6 The
Department of Chemistry, Eastern Illinois University, 600 Lincoln Avenue, Charleston,
Illinois, 61920, USA. E-mail: kawheeler@eiu.edu
† Electronic supplementary information (ESI) available: Crystallographic and
hydrogen bond tables for (+)-1, (ꢀ)-2, (+)-1$Im, (ꢀ)-2$Im, (ꢁ)-2$Im,
(+)-1/(ꢀ)-2$Im, and (+)-1/(ꢀ)-2$BPDO. CCDC 1470675–1470681. For ESI and
crystallographic data in CIF or other electronic format see DOI:
10.1039/c6ra08131b
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 64921–64929 | 64921

View Article Online
RSC Advances
Paper
the quasiracemate approach,15,16 topics at the forefront of
materials science such as energy storage and delivery, catalytic
processes of various kinds, and chemical sensors, could all
benet from applications using quasiracemic materials where
the complementary shapes of quasienantiomers direct molec-
ular assembly. Future directions such as these are largely
limited only by the requisite needs and creativity of the
practitioner.
The rich history of quasiracemic materials likely dates back
to 1853 with Pasteur's seminal work on the chirality of tartaric
acids.17 One aspect of this work showed that the deliberate
multicomponent crystallization of ammonium (+)-bitartrate Fig. 1 M. Centnerszwer circa 1934 and quasiracemic phase consisting
of 2-chloro and 2-bromosuccinic acid (portrait reproduced with
permission by Museum Boerhaave, Leiden, Netherlands).
and ammonium (ꢀ)-bimalate resulted in a crystalline ‘combi-
nation compound’ consisting of a 1 : 1 mixture of the two
starting components. Some y years aer Pasteur's report,
studies by Centnerszwer18 and Timmermans19 provided addi-
indistinguishable. Two such reports pursued the crystallo-
tional evidence of quasiracemate formation by examining pairs
graphic examination of Pasteur23 and Karle's24 early quasir-
of succinic acid derivatives. However, unlike Pasteur's assess-
acemates. Though reinvestigating pivotal science from
ment of material solubility and optical activity, these studies
a previous era offers important historical perspectives, a valu-
explored the formation of several quasiracemic phases via
able consequence from such studies is the critical entry point
melting-point phase diagrams. Building on these historic
these discoveries offer to new science directions of current
studies, Fredga and co-workers critically assessed the melting-
point behavior of hundreds of quasiracemic molecular pairs.20
This extensive effort demonstrated the importance of molecular
chirality and isosteric relationships to the successful construc-
tion of these bimolecular systems. Fredga's seminal investiga-
tions are largely responsible for the development of the
quasiracemic approach that has paved the way for more
contemporary reports dedicated to understanding the supra-
molecular tendencies of quasiracemate components. The liter-
ature now contains a sizable database in excess of a hundred
quasiracemate crystal structures.14 Without exception, and
regardless of structural framework and functional group, each
of these crystallographic reports reveals quasiracemic compo-
nents aligned in the crystal with near inversion symmetry.
Because no clear link exists between the observed approximate
centrosymmetric organization and the molecular architectures
interest.
In this study we continue the theme of probing contempo-
rary science from historical perspectives by targeting M. Cen-
tnerszwer's 1899 quasiracemate constructed from L-(ꢀ)-2-
chlorosuccinic acid, (ꢀ)-1, and D-(+)-2-bromosuccininc acid,
(+)-2 (Fig. 1).18 While data from Centnerszwer's structurally
simple compounds suggest the formation of a new quasir-
acemate, this material appears less energetically preferred than
enantiopure 1 and 2. This work assesses the viability of isolating
crystals of Centnerszwer's quasiracemate and the impact of
modifying the packing landscape of this quasiracemate by use
of tailor-made additives. Overall, these ndings contribute to
a growing body of work that examines the fundamental role
topological factors play in the molecular recognition process.
of these quasiracemate components, this collection of struc- 2 Experimental section
tures implies molecular shape oen provides a major contrib-
utor to the driving force for quasiracemate formation. Similar
observations and arguments have been made for purely racemic
materials.10,21,22
2.1 Reagents and materials
All chemicals and solvents were purchased from the Aldrich
Chemical Co. or Acros Chemicals and used as received without
1
further purication. H NMR and 13C NMR spectral data were
Over the last several years our interest with materials
exhibiting quasiracemic behavior has focused on under-
standing the role of topological features of small organics to the
molecular recognition process.8,12 In many of these investiga-
tions, our strategy centered on acquiring complete sets of
crystal structures for each quasiracemic family – i.e. structures
of the quasiracemate, racemates, and homochiral compounds.
This approach compared the racemic and enantiopure struc-
tures to the motifs observed in the related quasiracemate. In
addition to supporting the notion that quasiracemic materials
recorded with a 400 MHz Bruker Ascend spectrometer using
TopSpin v.3.2. The spectra were referenced using a solvent
residual signal as internal standard. The chemical shi values
are expressed as d values (ppm) and the value of coupling
constants (J) in Hertz (Hz). The following abbreviations were
used for signal multiplicities: s, singlet; d, doublet; dd, doublet
of doublets; t, triplet; q, quartet; m, multiplet; and br, broad.
2.2 Sample preparation
mimic the centrosymmetric patterns found in the racemic The synthesis of (S)-(+)- and (R)-(ꢀ)-2-chlorosuccininc acid [(+)-1
compounds, these investigations also reported strong iso- and (ꢀ)-1] was conducted using the following general procedure
structural relationships for many sets of quasiracemic and as previously described.25 Briey, a mixture of L-(+)-aspartic acid
racemic families. With many of these examples, the unit cell (10.0271 g, 0.0753 mol) and urea (1.0607 g, 0.0177 mol) dis-
parameters and crystallographic datasets were nearly solved in hydrochloric acid (12 M, 16 mL) and nitric acid (16 M,
64922 | RSC Adv., 2016, 6, 64921–64929
This journal is © The Royal Society of Chemistry 2016

View Article Online
Paper
RSC Advances
ꢂ
16 mL) was heated at 70 C for 5 hours. The clear-light yellow
solution was allowed to stand at room temperature for an
additional 18 hours. At that time colorless crystals were ltered
and washed with 200 mL of water. The ltrate was neutralized to
a pH ¼ 2–3 using a saturated sodium bicarbonate solution and
extracted with 3 ꢃ 30 mL of ethyl acetate. The organic extracts
were combined, dried over anhydrous magnesium sulfate, and
reduced under vacuo to give (+)-1 as a colorless solid (4.74 g,
Imidazolium hydrogen (ꢁ)-2-bromosuccinate, (ꢁ)-2. Color-
ꢂ
less prisms; method B; ꢀ5 C (5 : 1 methanol : water/acetone);
mp 130–133 ꢂC.
Imidazolium hydrogen R-(+)-2-chlorosuccinate/imidazolium
hydrogen S-(ꢀ)-2-bromosuccinate, (+)-1/(ꢀ)-2$Im. Colorless
prisms; method A; ꢀ5 ꢂC (methanol); mp 125–127 ꢂC. Colorless
prisms; method A; ꢀ5 ꢂC (1 : 1 methanol : ethanol); mp 125–127
ꢂC. Colorless prisms; method B; ꢀ5 ꢂC (1 : 1 meth-
anol : acetone); mp 125–127 ꢂC. Colorless prisms; method B; ꢀ5
ꢂC (1 : 1 methanol : ethanol/acetone); mp 125–127 ꢂC.
4,40-Bipyridine-N,N0-dioxide R-(+)-2-chlorosuccinic acid/S-
(ꢀ)-2-bromosuccinic acid, (+)-1/(ꢀ)-2$BPDO. Colorless prisms;
method A; ꢀ5 ꢂC (methanol); mp 183–185 ꢂC. Colorless plates,
1
ꢂ
41.4%). Mp 172–174 C. H NMR (acetone-d6): d 4.70 (dd, J ¼
3
7.46, 6.68 Hz, 1H, H–Csp ); 3.16 (dd, J ¼ 17.04, 7.46 Hz, 1H, H–
3
3
C
sp ); 2.95 (dd, J ¼ 17.04, 6.68 Hz, 1H, H–Csp ). 13C NMR
(acetone-d6): 170.9, 170.0, 52.7, 39.9 ppm.
The synthesis of (S)-(+)- and (R)-(ꢀ)-2-bromosuccininc acid
[(+)- and (ꢀ)-2] was achieved via the following general procedure
as previously described.26 Briey, a mixture of D-(ꢀ)-aspartic
acid (2.0091 g, 0.1509 mol) and potassium bromide (8.2546 g,
0.0699 mol) were dissolved in sulfuric acid (18 M, 39 mL). The
vigorously stirred mixture was cooled to ꢀ10 ꢂC and a solution
consisting of sodium nitrite (1.9978 g, 0.0187 mol) dissolved in
4.10 mL of H2O was added over a 10 minute period. The clear-
brown reaction mixture was then allowed to warm to room
temperature and stirred for an additional 5 hours. The crude
reaction mixture was then extracted with 3 ꢃ 30 mL of ethyl
acetate. The organic extracts were combined, dried over anhy-
drous magnesium sulfate, and reduced under vacuo to give
ꢂ
method B; ꢀ5 C (methanol/acetone); mp 183–185 ꢂC.
2.4 Characterization
X-ray Crystallography. Crystallographic details for
compounds (+)-1, (ꢀ)-2, (+)-1$Im, (ꢀ)-2$Im, (ꢁ)-2$Im, (+)-1/
(ꢀ)-2$Im, (+)-1/(ꢀ)-2$BPDO are summarized in Table S2 (ESI†).
X-ray data were collected on a Bruker APEX II diffractometer
˚
(Cu-Ka l ¼ 1.54178 A; graphite monochromator, T ¼ 100(2) K).
Data sets were corrected for Lorentz and polarization effects as
well as absorption – SADABS/multi-scan.27 The criterion for
observed reections is I > 2s(I). Lattice parameters were deter-
mined from least-squares analysis and reection data. Struc-
tures were solved by direct methods and rened by full-matrix
least-squares analysis on F2 using X-SEED28 equipped with
SHELX-2014/7.29 All non-hydrogen atoms were rened aniso-
tropically by full-matrix least-squares on F2 by the use of the
SHELXL29 program. In the case of disordered (+)-1/(ꢀ)-2$BPDO,
the fragment occupancies were rened and constrained to sum
1.0. H atoms for OH and NH groups were located in difference
Fourier synthesis and rened isotropically with restrained O–H
1
(ꢀ)-2 as a colorless solid (2.5075 g, 84.5%). Mp 163–166 ꢂC. H
3
NMR (acetone-d6): d 4.64 (dd, J ¼ 8.68, 6.16 Hz, 1H, H–Csp ); 3.25
3
(dd, J ¼ 17.24, 8.72 Hz, 1H, H–Csp ); 3.01 (dd, J ¼ 17.24, 6.08, 1H,
H–Csp ). 13C NMR (acetone-d6): d 170.3, 169.5, 39.2, 39.0.
3
2.3 Crystal growth experiments
Crystalline materials prepared for this study were grown in
solution using one of two methods. Method A (slow evapora-
tion): the succinic acid or equimolar amounts of imidazole
and succinic acid were dissolved in a single solvent or solvent
mixture. The solution wꢂas le partially covered and allowed to
evaporate slowly at ꢀ5 C or room temperature until crystals
appeared typically aer 2–4 days. Method B (vapour diffu-
sion): equimolar amounts of imidazole and succinic acid were
dissolved in a compatible solvent system. A vial containing
this solution was then placed in a larger enclosed vessel
containing a second precipitating solvent (solvent B) in which
the components were insoluble. Crystal growth via this
method at ꢀ5 ꢂC or room temperature typically occurred aer
1–3 days.
˚
and N–H distances of 0.85(2) A and Uiso ¼ 1.2Ueq of the attached
O and N atom. The remaining H atoms were included in
idealized geometric positions with Uiso ¼ 1.2Ueq of the atom to
which they were attached. Where appropriate, molecular
congurations were compared to the known chirality of the 2-
halosuccinic acid components and estimated Flack parame-
ters.30 Atomic coordinates were inverted to achieve correct
molecular congurations where needed.
3 Results and discussion
3.1 Centnerszwer's quasiracemate
S-(ꢀ)- and R-(+)-2-chlorosuccinic acid, (ꢀ)- and (+)-1. Color- M. Centnerszwer's quasiracemate provides a structurally simple
less plates; method A; room temperature (acetone); mp 172–174 system with components that differ by Cl and Br substituents.
ꢂC.
Given the resources available at the end of the 19th century, it is
S-(ꢀ)- and R-(+)-2-bromosuccinic acid, (+)-2. Colorless plates; not surprising this early investigation pursued the effects of co
method A; room temperature (acetone); mp 161–163 ꢂC.
crystallization via melting-point data. This method involved
Imidazolium hydrogen R-(+)-2-chlorosuccinate, (+)-1$Im. generating a series of solid mixtures that differed incrementally
Colorless prisms; method A; room temperature (1-propanol); by the ratio of components. Collecting melting-point data for
mp 125–128 ꢂC.
each sample then provided a useful tool to probe the thermal
Imidazolium hydrogen S-(ꢀ)-2-bromosuccinate, (ꢀ)-2$Im. properties of these materials. Fig. 2 shows several multi-
Colorless prisms; method A; ꢀ5 ꢂC (1-propanol); mp 121– component phase diagrams consisting of various combina-
ꢂ
123 C.
tions of 2-chlorosuccinic acid (1) and 2-bromosuccinic acid (2).
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 64921–64929 | 64923

View Article Online
RSC Advances
Paper
than the enantiopure compounds. Second, the Cambridge
Structural Database (CSD)32 reports the crystal structures of
(ꢀ)-1 (ref. 33) and (+)-2,34 but those corresponding to the
racemic forms are suspiciously absent.
Our investigation began with the asymmetric synthesis of 1
and 2 using previously reported procedures.25,26 These synthetic
transformations involved diazotization of either D- or L-aspartic
acid with NaNO2/HX. This approach followed the in situ prep-
aration of diazonium intermediates that underwent further
double substitution reactions to give the desired 2-halosuccinic
acids with retained stereochemistry as compared to the starting
aspartic acid.
Attempts to grow crystals of Centnerszwer's quasiracemate
involved combining equimolar portions of (+)-1 and (ꢀ)-2.
Crystal growth experiments using this mixture ranged from
standard solution techniques employing a variety of solvents to
crystal growth from the melt. Materials retrieved from these
studies consisted of either the starting homochiral Cl- or Br-
succinic acids or amorphous conglomerate solids as deter-
mined by melting point or X-ray diffraction data. Previously
examined room-temperature crystal structures of (ꢀ)-1 (ref. 33)
and (+)-2 (ref. 34) were re-determined at 100 K for this study.
Crystallographic assessment of these materials are isostructural
with closely related unit cell parameters and crystal alignment.
These structures consist of molecules linked by R22(8)
carboxyl/carboxyl contacts to give chain motifs propagating
along the a-axis (Fig. 3A). Adjacent inversion related motifs to
these chains organize with effective close-packing as compared
to other small organic carboxylic acids (Fig. 3B).35 The crystal
densities [(ꢀ)-1, 1.740 g mLꢀ1; (+)-2, 2.168 g mLꢀ1] and packing
coefficients (Ck, a measure of free space in crystals (ref. 36 and
37)) [(ꢀ)-1, 81.6%; (ꢀ)-2, 81.7%] suggest relatively high packing
Fig. 2 Centnerszwer's melting-point phase diagrams18 showing (A)
the quasiracemate of (+)-1 and (ꢀ)-2, (B) racemate (ꢁ)-1, and (C)
a mixture of (ꢀ)-2 and (ꢁ)-1.
The melting-point data from Centnerszwer's materials provides
an intriguing example that supports the presence of quasir-
acemate (+)-1/(ꢀ)-2 (Fig. 2A). The diagram reveals two eutectic
points at component ratios of ꢄ40 : 60 and 60 : 40 with the
appearance of quasiracemic phase nearly 20 degrees lower than
the homochiral counterparts. A similar melting-point landscape
was also recorded for (ꢁ)-1 (Fig. 2B). The melting-point behavior
of the ternary system (ꢁ)-1/(ꢀ)-2 shows the effect of introducing
(ꢀ)-1 to the quasiracemic system. The result is an ill-dened
diagram complicated by the assortment of interactions arising
from this three-component system (Fig. 2C).
The substantially lower melting points associated with the
quasiracemic and racemic systems occur infrequently in the
literature and likely originates from the differences in packing
forces that exist in the crystal lattice. Unlike the melting-point
behavior of the majority of racemic compounds, racemates of
the type shown in Fig. 2B melt at similar temperatures to that of
the 1 : 1 conglomerates. The thermodynamic consequence of
this observation on the recognition process is signicant.
Recrystallization of racemates of this class typically result in
mechanical separation via a Pasteurian-type resolution process
to give distinctly different crystals containing either the (+) or
(ꢀ) antipodes. Because quasiracemates are close structural
relatives to racemic compounds, it is not surprising sponta-
neous resolution has also been reported for these materials.31
Two lines of evidence suggest crystal growth of the quasir-
acemic system consisting of 1 and 2 also undergoes separation.
First, Centnerszwer's melting point diagrams indicate racemic
and quasiracemic phases with melting points signicantly less
Fig. 3 Crystal structure projections of isostructural (ꢀ)-1 and (+)-2
depicting (A) R22(8) carboxyl/carboxyl motifs, (B) close-packing of
unit cell, and (C) fingerprint plots.
64924 | RSC Adv., 2016, 6, 64921–64929
This journal is © The Royal Society of Chemistry 2016

View Article Online
Paper
RSC Advances
efficiency. Perhaps another indication of the compact align- carboxy-2-(imidazol-3-ium-1-yl)propanoate.44–47 To avoid this
ment of these materials comes from comparison to the crystal unwanted transformation and to promote crystal growth of the
densities of the structurally related achiral 2-chloro and 2-bro- desired multi-molecular compounds, all crystallization studies
mofumaric acids.38,39 While it is foreseeable these species could involving bromo 2 in the presence of imidazole were conducted
adopt planar geometries that translate to efficient crystal at ꢀ5 ꢂC. Crystalline samples retrieved from this study were
stacking, their densities are markedly less (Cl, 1.66 g mLꢀ1; Br, initially examined by melting point (mp) determinations.
2.07 g mLꢀ1) than (ꢀ)-1 and (+)-2. We reasoned that our Unlike Centnerszwer's report of the mp trend for the family of
unsuccessful attempts to prepare crystals of quasiracemate 1/2, succinic acids 1 and 2 [i.e., (+)-1, (ꢀ)-2, (ꢁ)-1, and (+)-1/(ꢀ)-2;
combined with the packing efficiency of (ꢀ)-1 and (+)-2, offers Fig. 2], the effect of combining imidazole with these succinic
further support that the thermodynamic stability of these acids sufficiently alters the crystalline landscapes to promote
enantiomers components outweighs the benet of crystallizing crystal growth of the racemic materials. The mp values for
the racemic or quasiracemic mixtures.
molecular salts (+)-1$Im and (ꢀ)-2$Im are 125–128 ꢂC and 121–
ꢂ
123 C, respectively. By contrast, those mp values for (ꢁ)-1$Im
(134–137 ꢂC, ref. 43) and (ꢁ)-2$Im (130–133 ꢂC) are higher
suggesting the homochiral imidazolium salts take on destabi-
lized crystalline phases compared to their racemic counterparts.
3.2 Multi-molecular assemblies of Centnerszwer's
quasiracemate
Moving forward, our attention then turned to pursuing Cen- In turn, the consequence of this thermal property trend is that it
tnerszwer's quasiracemate by creating multi-molecular assem- poses a potential scenario for the formation of quasiracemate
blies of 1 and 2 with the aid of the cocrystallization method. (+)-1/(ꢀ)-2 involving imidazole as an additional crystallizing
This strategy exploits the complementary structural features agent.
and non-bonded contacts between chemically distinct compo-
Inspection of the crystal structures generated for this series
nents. The successful use of tailor-made additives to engender of compounds reveals several noticeable similarities with the
desired chemical functions on multi-molecular systems has molecular and supramolecular structures. In each case, the
experienced considerable attention in the literature.40,41 The asymmetric units consists of a molecular salt derived from
motivation for applying this approach to the current study rests monohydrogen succinate and imidazolium components. These
with the opportunity to create new crystalline phases with ions link via N+–H/carboxylate interactions with N+/Oꢀ and
different thermodynamic properties to that of the starting H/Oꢀ bond lengths ranging from 2.67–2.74 A and 1.87–1.94 A,
˚
˚
˚
components. Since the molecular framework for succinic acids respectively, with average distances of 2.72 and 1.90 A. The
1 and 2 consists of a pair of carboxyl groups, it was initially nature of these interactions is quite directional as evident from
important to consider secondary molecules that take advantage an average N–H/O angle of 167ꢂ (154 to 177ꢂ). These motifs
of the hydrogen-bond ability of one or both CO2H groups. The link with translationally related neighbors to give molecular
non-bonded behavior of carboxylic acids in the presence of chains.
a variety of functional groups is well documented.42 Because
A second distinct interaction involves the succinate compo-
these reports provide a substantial list of secondary molecules nent participating in carboxyl/carboxylate contacts. The role of
that create persistent assemblies with carboxylic acids, the O–H/ꢀO interaction to crystal stabilꢀization is signicant
˚
a potential obstacle facing this study focused on narrowing the with short hydrogen-bond contacts [O/ O, 2.54 and 2.60 A
ꢀ
˚
˚
˚
possible selection of components to be paired with succinic (2.57 A (average)); H/ O, 1.69–1.77 A (1.73 A (average)) and
acids 1 and 2. A search of the CSD for entries containing multi- nearly linear O–H/ꢀO angles [170–176ꢂ (173ꢂ (average))]. This
component systems with either 1 or 2 offered valuable insight to supramolecular synthon connects adjacent monohydrogen
the crystallization preferences of these succinic acids. One succinate entities to form molecular chains. Because the two
critical entry, the structure of imidazolium (ꢁ)-2-chlor- primary hydrogen-bond contacts (i.e., carboxylate/imidazo-
osuccinate by MacDonald et al., gave direct support to the lium and carboxyl/carboxylate) observed for this family of
accessibility of a racemic crystalline phase involving component imidazolium salts propagate in orthogonal directions, the
1 and imidazole.43
result is the construction of robust molecular layers. This 2D
With the aid of this structural information, the investigation motif provides a common feature inherent to these structures
then turned to preparing a series of compounds that contain with variations due to the orientation of neighboring layers.
imidazole (Im) and succinic acids 1 and 2. To understand the
Quasiracemic crystals for this investigation were prepared by
supramolecular preferences of combining these components, low temperature slow evaporation of a 1 : 1 : 2 mixture of the
this study included not only the quasiracemate (ꢀ)-1/(+)-2$Im, (+)-1, (ꢀ)-2, and imidazole components. Crystallographic
but also the structures of the racemic and enantiopure assessment of a crystal retrieved from these experiments indi-
compounds – 5 crystal structures total. The asymmetric units cated molecules organized in non-centrosymmetric space group
and crystal packing views of this family of structures are P1 with X-ray reection data and unit cell parameters nearly
provided in Fig. 4A–E. Crystal growth experiments of these indistinguishable to those observed for the racemic structures
ꢀ
multi-molecular systems utilized both slow evaporation and (space group P1). Unlike the homochiral (Fig. 4A and E) and
vapor diffusion techniques. An unanticipated outcome was the racemic (Fig. 4B and D) materials, the asymmetric unit of the
discovery crystallization of (ꢁ)-2 or (ꢀ)-2 with imidazole under quasiracemate consists of a ternary system with components
ambient conditions gave crystals of the substitution product 3- (+)-1 and (ꢀ)-2 aligned with near inversion symmetry due to
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 64921–64929 | 64925

View Article Online
RSC Advances
Paper
Fig. 4 Crystal structure projections and fingerprint plots for (A) (+)-1$Im, (B) (ꢁ)-1$Im,43 (C) (+)-1/(ꢀ)-2$Im, (D) (ꢁ)-2$Im, and (E) (ꢀ)-2$Im. Except
for (ꢁ)-1$Im, the asymmetric units are depicted with atomic displacement parameters at the 50% level.
complementary molecular shapes of the succinic acid adducts. (ꢁ)-1$Im and (ꢁ)-2$Im. In each structure, the hydrogen-bonded
As shown in Fig. 4B–D, the crystal packing motifs for quasir- layers align in the (011) plane. For the quasiracemic phase,
acemate (+)-1/(ꢀ)-2$Im imitate those observed for racemates these molecular layers alternate with motifs constructed from
64926 | RSC Adv., 2016, 6, 64921–64929
This journal is © The Royal Society of Chemistry 2016

View Article Online
Paper
RSC Advances
(+)-1$Im and (ꢀ)-2$Im arrays. The observed translational non- to be controlled by N+–Oꢀ/H–O interactions. Fig. 5 shows the
crystallographic symmetry existing between the layers of the outcome from cocrystallizing (+)-1, (ꢀ)-2, and BPDO in a 1 : 1 : 2
quasiracemic phase is the direct result of pairing the quasie- ratio. Crystallographic assessment of samples from this exper-
nantiomeric forms of 1 and 2. As with all known quasiracemic iment revealed that the BPDO and succinic acid components
crystal structures, the assembly of (+)-1 and (ꢀ)-2 combined link by the anticipated N+–Oꢀ/carboxyl contacts in space
with imidazole is predisposed to adopt near inversion symmetry group P1. This non-bonded interaction creates molecular
with a high degree of structural mimicry to the local and global chains with alternating BPDO and (+)-1 or (ꢀ)-2 motifs propa-
crystal packing features of the ‘true’ racemic compounds.
gating along the [011] axis. These assemblies further align in the
In contrast to the racemic and quasiracemic phases, the crystal with neighboring (+)-1$BPDO and (ꢀ)-2$BPDO motifs
crystal structures of homochiral (+)-1$Im and (ꢀ)-2$Im display related by approximate inversion symmetry. This desymme-
signicantly different crystal packing with components orga- trized arrangement closely mimics the centrosymmetric space
ꢀ
nized in space groups P1 and P21, respectively (Fig. 4A and E). group P1 where the deviation from true symmetry arises from
The structural implication of the imposed space group the chemical variation of the Cl and Br groups.
symmetry is that neighboring hydrogen-bonded layers are
Though the crystal structures of the ternary quasiracemates
either related by translational symmetry or by a two-fold screw (+)-1/(ꢀ)-2$Im and (+)-1/(ꢀ)-2$BPDO are markedly different,
axis. As such, the overall patterns from these structures are both take advantage of the complementary structural features
distinctly non-centrosymmetric.
of molecular shape and strong hydrogen-bond interactions. The
Evaluating the packing efficiency of this series for structures structural consequence of these design elements is the
provided further evidence of the driving force for quasiracemate construction of robust molecular assemblies that closely
formation. This was accomplished by comparison of the crystal imitate inversion symmetry.
densities (r) and packing coefficients (Ck). These related iden-
tiers of crystal compactness follow a similar trend for this
3.4 Hirshfeld surface analysis
structural family. The crystal structures corresponding to race-
mates (ꢁ)-1$Im (r ¼ 1.572 g mLꢀ1, Ck ¼ 81%) and (ꢁ)-2$Im (r ¼
The Hirshfeld surface model for assessing atomic interactions
in crystals came about in the late 1970's.51 By partitioning the
electron density of atomic fragments in a molecule, this method
1.883 g mLꢀ1, Ck ¼ 83%) are appreciably closer packed than
those of enantiomers (+)-1$Im (r ¼ 1.552 g mLꢀ1, Ck ¼ 80%) and
(ꢀ)-2$Im (r ¼ 1.801 g mLꢀ1, Ck ¼ 80%). These trends are
effectively generates weighted functions for mapping atomic
surfaces that are highly sensitive to the immediate crystal
consistent with Wallach's Rule48 and the companion study by
environment. Because the Hirshfeld surface is dened by the
Brock et al.,49 where comparison of sets of racemic/chiral pairs
molecule and the proximity of its nearest neighbours, the
generated surface can effectively relay information about the
crystal environment and individual intermolecular interactions.
This method is sensitive to seemingly small crystal anomalies
and thus offers an important structural tool for comparing
families of related structures.
A particularly useful application of the Hirshfeld surface has
been incorporated into M. A. Spackman's soware platform
Crystal Explorer.52,53 In addition to offering a practical means to
graphically access 3D Hirshfeld surfaces from input molecules,
this package also contains a wide-array of structural tools for
indicated racemic crystals are, on average, about 1% more
tightly packed than the their chiral counterparts. Our quasir-
acemate also benets from the propensity of racemic crystals to
align with greater density. Because the crystal structure of (+)-1/
(ꢀ)-2$Im is isostructural with racemates (ꢁ)-1$Im and (ꢁ)-2$Im,
it is not surprising the crystal density (r ¼ 1.572 g mLꢀ1) and Ck
(81%) values reside between those for the racemates. This
information surmises that tighter crystal packing of the racemic
and quasiracemic systems reects greater thermodynamic
stability. Such a structural bias is signicant and suggests
packing of the imidazolium succinate racemates or quasir-
assessing the properties associated with crystalline assemblies.
acemate outweighs the potential mechanical separation of
One such tool targets the parameters di and de – dened as the
components as in the case of the bimolecular (ꢁ)-1, (ꢁ)-2, or
(+)-1/(ꢀ)-2 systems.
3.3 Bipyridine-N,N0-dioxide quasiracemate
Considering the success of the ternary quasiracemate (+)-1/
(ꢀ)-2$Im, we wondered if this study could be further developed
by substituting imidazole for other additives. Given the initial
design strategy of adopting co-former molecules that comple-
ment the hydrogen-bond abilities of succinic acids 1 and 2, 4,40-
bipyridyl-N,N0-dioxide (BPDO) seemed a viable candidate for
creating a second ternary system with quasiracemate (+)-1/(ꢀ)-2.
Many cocrystallization investigations have demonstrated the
merit of BPDO as a reliable ditopic hydrogen-bond acceptor.50
Fig.
5 Crystal structure packing diagram (atomic displacement
Building on these previous reports, we anticipated molecular
parameters, 50%) and fingerprint plot for quasiracemate (+)-1/
architectures constructed from BPDO and succinic acids 1 or 2 (ꢀ)-2$BPDO.
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 64921–64929 | 64927

View Article Online
RSC Advances
Paper
distance from the Hirshfeld surface to an internal (di) and N+–H/carboxylate and carboxyl/carboxylate interactions. In
external (de) nuclei. By plotting di versus de the result is an contrast to the enantiopure [(+)-1$Im and (ꢀ)-2$Im] and
informative ngerprint diagram with features highly sensitive racemic [(ꢁ)-1$Im and (ꢁ)-2$Im] systems, neighboring molec-
to the immediate environment of the molecule.
ular layers observed in quasiracemate (+)-1/(ꢀ)-2$Im are orga-
Applying this method to the current study offers important nized by approximate inversion symmetry. Assessment of the
insight to the structural similarities and variations of this family crystal packing efficiency for this series of molecular salts via
of succinic acids. For isostructural succinic acids (+)-1 and crystal densities and packing coefficients (Ck) indicates imid-
(ꢀ)-2, the ngerprint plots are nearly indistinguishable azole effectively alters the crystal landscape of the system in
(Fig. 3C). A dominant contact is depicted as sharp spikes located favor of racemate and quasiracemate formation. A similar
at the bottom le of the diagrams and signies an O/H desymmetrized crystal environment was also observed for the
interaction. This combined with the diffuse region (indicating crystal structure of (+)-1/(ꢀ)-2$BPDO where the components are
H/H contacts) located between the spikes represents the cyclic organized via N+–Oꢀ/carboxyl contacts. This study under-
R22(8) carboxyl/carboxyl interactions. A second principal scores the importance of molecular shape to molecular recog-
contact recognized in Fig. 3 is the Cl/Br/H interaction as nition processes and the stabilizing effect of tailor-made co-
identied as a pair spikes with greater separation and at much former molecules for creating new crystalline phases of previ-
larger values of di and de. Unlike the plot features corresponding ously inaccessible crystals.
to the carboxylic acid dimer, these spikes are less pronounced
due to the ill-dened nature of these interactions. Visually
comparing these plots to those depicted in Fig. 4 for the imid-
Acknowledgements
azole molecular salts highlights the common features of strong This work was supported by the National Science Foundation
directional O/H interactions. In the case of the ve structures (DMR-1505717 and CHE-0722547) and Eastern Illinois Univer-
shown in Fig. 4, each lack the diffuse patterns between the pair sity CFR (KAW) and URSCA (JMS) awards.
of O/H spikes. This feature is consistent with sets of acyclic
carboxylate/carboxyl contacts present in these structures. The
topology represented in the ngerprint plots corresponding to
Notes and references
¨
(ꢁ)-1$Im, (ꢁ)-2$Im, and quasiracemate (+)-1/(ꢀ)-2$Im are
decidedly similar owing to the isostructurality of these racemic
and quasiracemic phases. One area of variation in the nger-
print diagrams relates to the Cl/Br/H interactions. While this
contact is present in each structure, it is more pronounced in
(ꢁ)-2$Im (Fig. 4D). This same variation is seen with the homo-
chiral (+)-1$Im and (ꢀ)-2$Im structures. The overall differences
observed with the patterns in the plots for these homochiral
structures are even more distinct and result from the variation
in space groups and stacking of molecular sheets constructed
from imidazole and (+)-1$Im and (ꢀ)-2$Im. Finally, as expected
the ngerprint plot for (+)-1/(ꢀ)-2$BPDO exhibits unique
features as compared to the other plots presented in this study.
1 F. Bachle, I. Fleischer and A. Pfaltz, Adv. Synth. Catal., 2015,
357, 2247.
`
2 S. Piovesana, R. Samperi, A. Lagana and M. Bell, Chem.–Eur.
J., 2013, 19, 11478.
3 Y. Lu, A. J. Bolokowicz, S. A. Reeb, J. D. Wiseman and
K. A. Wheeler, RSC Adv., 2014, 4, 8125.
4 E. A.-H. Yeh, E. Kumli, K. Damodaran and D. P. Curran, J.
Am. Chem. Soc., 2013, 135, 1577.
5 F. Cardona, D. Lalli, C. Faggi, A. Goti and A. Brandi, J. Org.
Chem., 2008, 73, 1999.
6 N. A. Shaye, S. Chavda, E. Coulbeck, J. Eames and
Y. Yohannes, Tetrahedron: Asymmetry, 2011, 22, 439.
7 C. H. Gorbitz and P. Karen, J. Phys. Chem. B, 2015, 119, 4975.
8 J. T. Cross, N. A. Rossi, M. Seran and K. A. Wheeler,
CrystEngComm, 2014, 16, 7251.
Fig.
5 shows distinct patterns for Cl/Br/H and p/p
interactions.
9 F. T. Martins, R. S. Correa, A. A. Batista and J. Ellena,
CrystEngComm, 2014, 16, 7013.
10 J. Jacques, A. Collet and S. H. Wilen, Enantiomers, Racemates
and Resolutions, Wiley-Interscience, New York, 1981.
4 Conclusions
Our results suggest that combining succinic acids (ꢀ)-1 and
(+)-2 is generally destabilizing compared to forming crystals of 11 B. Dalhus and C. H. Gorbitz, Acta Crystallogr., Sect. B: Struct.
the starting optically active components. This observation is
consistent with Centnerszwer's 1899 melting-point data that 12 M. Hendi, P. Hooter, V. Lynch, R. E. Davis and K. A. Wheeler,
reports a nearly 20 degree depression in melting-points for the Cryst. Growth Des., 2004, 4, 95.
racemate of 1 and quasiracemic phases. To circumvent this 13 Q. Zhang and D. P. Curran, Chem.–Eur. J., 2005, 11, 4866.
Sci., 2000, B56, 715.
´
´
supramolecular difficulty, the inclusion of additional molecules 14 S. P. Kelley, L. Fabian and C. P. Brock, Acta Crystallogr., Sect.
capable of complementing the hydrogen-bond ability of suc- B: Struct. Sci., 2011, B67, 79.
cinic acids 1 and 2 were pursued and utilized in the study. Both 15 Z. Hayouka, N. C. Thomas, D. E. Mortenson, K. A. Satyshur,
imidazole (Im) and 4,40-bipyridyl-N,N0-dioxide (BPDO) served as
ternary components that altered the crystal packing landscape
B. Weisblum, K. T. Forest and S. H. Gellman, J. Am. Chem.
Soc., 2015, 137, 11884.
of 1 and 2. Combining imidazole with the quasiracemate, 16 C. K. Wang, G. J. King, S. E. Northeld, P. G. Ojeda and
racemate, and enantiopure forms of 1 and 2 resulted in crystal D. J. Craik, Angew. Chem., Int. Ed., 2014, 53, 11236.
structures identied by molecular layers formed from strong 17 L. Pasteur, Ann. Chim. Phys., Ser., 1853, 38, 437.
64928 | RSC Adv., 2016, 6, 64921–64929
This journal is © The Royal Society of Chemistry 2016

View Article Online
Paper
RSC Advances
18 M. Z. Centnerszwer, J. Phys. Chem., 1899, 29, 715.
19 J. Timmermans and M. Dumont, Bull. Soc. Chim. Belg., 1931,
40, 689.
20 A. Fredga, Bull. Soc. Chim. Fr., 1973, 1, 173.
21 A. Gavezzotti and L. L. Presti, Cryst. Growth Des., 2015, 15,
3792.
37 Molecular volumes were calculated using CPK model
functions in Spartan10 (Wavefunction, Inc.). For the
imidazolium succinate structures, the volumes of the 1, 2,
and imidazolium components were determined by
calculating Vmol for each molecular species and averaging
over the family of crystal structures.
`
22 R. Centore, S. Fusco, F. Capone and M. Causa, Cryst. Growth 38 A. Wong, K. J. Pike, R. Jenkins, G. J. Clarkson, T. Anupold,
Des., 2016, 16, 2260.
23 K. A. Wheeler, R. C. Grove, R. E. Davis and W. S. Kassel,
Angew. Chem., Int. Ed. Engl., 2008, 47, 78.
24 M. E. Breen, S. L. Tameze, W. G. Dougherty, W. S. Kassel and
K. A. Wheeler, Cryst. Growth Des., 2008, 8, 3483.
25 J. E. Robinson and M. A. Brimble, Org. Biomol. Chem., 2007,
5, 2572.
A. P. Howes, D. H. G. Crout, A. Samoson, R. Dupree and
M. E. Smith, J. Phys. Chem. A, 2006, 110, 1824.
39 A. Fischer, Acta Crystallogr., Sect. E: Struct. Rep. Online, 2006,
62, o4190.
ˇ
40 M. B. J. Atkinson, S. V. S. Mariappan, D.-K. Bucar,
ˇˇ ´
J. Baltrusaitisa, T. Friscic, N. G. Sinada and
L. R. MacGillivray, Proc. Natl. Acad. Sci. U. S. A., 2011, 108,
26 B. J. Naysmith and M. A. Brimble, Org. Lett., 2013, 15, 2006.
10974.
´ ´
27 G. M. Sheldrick, SADABS — Program for Area Detector 41 J. Wouters and L. Quere, Pharmaceuticals Salts and Co-
¨
¨
Absorption Corrections, University of Gottingen, Gottingen,
Germany, 2013.
crystals, RSC Drug Discovery Series No.16, Royal Society of
Chemistry, Cambridge, UK, 2012, vol. 330.
28 L. J. Barbour, J. Supramol. Chem., 2001, 1, 189.
29 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr.,
2008, 64, 112.
42 F. H. Allen, W. D. S. Motherwell, P. R. Raithby, G. P. Shields
and R. Taylor, New J. Chem., 1999, 25–34.
43 J. C. MacDonald, P. C. Dorrestein and M. M. Pilley, Cryst.
Growth Des., 2001, 1, 29.
30 H. D. Flack, Acta Crystallogr., 1983, 39, 876.
31 M. Hendi, M. R. E. Davis, V. Lynch and K. A. Wheeler, Cryst. 44 S. A. Reeb, M. C. Shields and K. A. Wheeler, Acta Crystallogr.,
Eng., 2001, 4, 11. Sect. E: Struct. Rep. Online, 2009, 65, o1424.
32 F. H. Allen, Acta Crystallogr., Sect. B: Struct. Sci., 2002, B58, 45 Y. X. Zhi, J. Hu, J. Y. Chen and J.-Z. Chen, Z. Kristallogr.–New
380. Cryst. Struct., 2009, 224, 489.
33 L. Kryger, S. E. Rasmussen and J. Danielsen, Acta Chem. 46 A. Briceno, D. Leal and G. D. de Delgado, New J. Chem., 2015,
Scand., 1972, 26, 2339. 39, 4965.
34 J. F. Britten, H. E. Howard-Lock, D. Leblanc and C. J. L. Lock, 47 J. M. Xiao, Acta Crystallogr., Sect. E: Struct. Rep. Online, 2009,
Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1993, 49,
65, o1157.
1222.
48 O. Wallach, Liebigs Ann. Chem., 1895, 286, 90.
35 A search of the CSD (version 5.37) for organic carboxylic acid 49 C. P. Brock, W. B. Schweizer and J. D. Dunitz, J. Am. Chem.
structures containing #20 atoms with only C, H and one X Soc., 1991, 113, 9811.
group (X ¼ Cl or Br) resulted in 27 (Cl) and 22 (Br) entries 50 B.-Q. Ma, S. Gao, H.-L. Sun and G.-X. Xu, CrystEngComm,
with average crystal densities of 1.543 and 1.824 g mLꢀ1
respectively.
,
2001, 3, 147.
51 F. L. Hirshfeld, Theor. Chim. Acta, 1977, 44, 129.
36 The packing coefficient is dened as the ratio of the volume 52 M. A. Spackman and D. Jayatilaka, CrystEngComm, 2009, 11,
of the molecules in the unit cell to the volume of the unit 19.
cell, Ck ¼ Vmol/Vcell: A. I. Kitaigorodsky, Molecular Crystals 53 M. A. Spackman and J. J. McKinnon, CrystEngComm, 2002, 4,
and Molecules, Academic Press, New York, 1973.
378.
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 64921–64929 | 64929