Influence of impurities on the crystallisation of 5-X-aspirin and 5-X-aspirin anhydride polymorphs (X = Cl, Br, Me)

Article abstract of DOI:10.1039/c1ce06065a

Crystallisation of 5-X-aspirin (X = Cl, Br) by ambient evaporation from organic solvents commonly yields a mixture of polymorphs I and II for X = Cl, but only polymorph I for X = Br. Addition of 5-X-aspirin anhydride (5-20 mol %) leads to sole production

Full text of DOI:10.1039/c1ce06065a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
CrystEngComm
Cite this: CrystEngComm, 2011, 13, 6991
www.rsc.org/crystengcomm
PAPER
Influence of impurities on the crystallisation of 5-X-aspirin and 5-X-aspirin
anhydride polymorphs (X ¼ Cl, Br, Me)†
*
Katarzyna A. Solanko and Andrew D. Bond.
Received 17th August 2011, Accepted 30th September 2011
DOI: 10.1039/c1ce06065a
Crystallisation of 5-X-aspirin (X ¼ Cl, Br) by ambient evaporation from organic solvents commonly
yields a mixture of polymorphs I and II for X ¼ Cl, but only polymorph I for X ¼ Br. Addition of
5-X-aspirin anhydride (5–20 mol %) leads to sole production of polymorph II for X ¼ Cl, and new
production of polymorph II for X ¼ Br. Slurry experiments show that polymorph I transforms to
polymorph II for X ¼ Cl, while polymorph II transforms to polymorph I for Br. Thus, addition of the
anhydride during crystallisation apparently accelerates the I / II transformation for X ¼ Cl, and
inhibits the II / I transformation for X ¼ Br. 5-X-Aspirin anhydride can be produced as a by-product
during heating of 5-X-aspirin in organic solvents, or during synthesis from 5-X-salicylic acid, and the
duration of the heating step in such a synthesis can therefore change the polymorphic form of the
synthesised product. The 5-X-aspirin anhydrides (X ¼ Cl, Br, Me) are also polymorphic.
Crystallisation of the pure compound from organic solvents yields only polymorph I, while polymorph
II can be obtained by heating in ethanol or acetone. For X ¼ Cl or Br, the heating step produces
significant quantities of 5-X-aspirin in situ, while for X ¼ Me, it is necessary also to add 5-X-aspirin
(5–20 mol %) to obtain polymorph II.
We have studied in detail the polymorphism of aspirin.8–11 The
Introduction
compound is quite unusual in that it can crystallise as an
‘‘intergrown’’ form containing domains with two distinct poly-
morphic structures, and we have reported that domains of the
second aspirin polymorph can be introduced by crystallisation of
aspirin in the presence of aspirin anhydride.10 The anhydride
apparently acts to inhibit solution-mediated transformation of
the metastable form II to the more stable form I. An important
chemical aspect of this observation is that aspirin anhydride is
a common by-product of aspirin synthesis, and it can also be
formed in significant quantities when aspirin is heated in various
organic solvents.12–14 The possibility to create the polymorph-
inducing anhydride during synthesis or under common crystal-
lisation conditions can lead to erratic crystallisation behaviour,
which possibly accounts to some extent for the long-term
Polymorphism is an intensely studied phenomenon that is rele-
vant to any application area of crystalline molecular materials,1
including notably the pharmaceutical industry.2 To establish
control over polymorphism in a general sense, the mechanisms of
nucleation and crystal growth must be better understood at
a molecular level, including specifically the way in which poly-
morphism emerges during these processes. A guiding principle is
Ostwald’s rule of stages,3,4 which states that a system leaving one
state and transforming to another seeks out the next least stable
state rather than the thermodynamically most stable state.
Although exceptions are known,4 this guideline has provided
a basis to control polymorphism during solution crystallisation
of some compounds. For example, the metastable a polymorph
of L-glutamic acid can be stabilised in a saturated glutamic acid
solution for weeks in the presence of trimesic acid. The a form
crystallises first from solution, then the additive decreases the
rate of transformation to the more stable b form by binding
uncertainty
that
was
associated
with
aspirin
polymorphism.13,15–20
selectively to
b
crystals, thereby slowing their growth.5
Compounds that interact specifically with growing crystals in
this way are referred to as ‘‘tailor-made additives/auxiliaries’’.6,7
Department of Physics and Chemistry, University of Southern Denmark,
5230 Odense, Denmark. E-mail: adb@chem.sdu.dk
† Electronic supplementary information (ESI) available: Details of
PXRD and 13C analyses. CCDC reference numbers 823448, 823449,
823451–823454, 823456 and 823457. For ESI and crystallographic data
in CIF or other electronic format see DOI: 10.1039/c1ce06065a
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6991–6996 | 6991

View Article Online
With the realisation that the polymorphism of aspirin can be
influenced by the presence of aspirin anhydride during crystal-
lisation, we set out to test whether this might also be the case for
other aspirin derivatives. Specifically, we consider here
5-X-aspirin (5-X-A), with X ¼ Cl, Br or Me. Crystal structures
have been reported for these compounds by Hursthouse and co-
workers,21 and we refer to the published structures as 5-X-A form
I. Compounds 5-Cl-A and 5-Br-A are isostructural, while the
structure of 5-Me-A is different. Hursthouse et al. have also
observed polymorphism for 5-Cl-A,22 and we refer to this second
polymorph as form II. Our interest here is to examine the
possible influence of 5-X-aspirin anhydride (5-X-AA) on the
crystallisation of 5-X-A.
Results and discussion
Crystallisation of 5-Cl-A
For 5-Cl-A, polymorphism is evident for crystallisation of the
pure compound by slow evaporation from various organic
solvents under ambient conditions. Commonly, we obtained
mixtures of forms I and II from a given crystallisation trial,
except on crystallisation from 2-propanol and acetonitrile, for
which only form II was obtained. Slurrying experiments establish
that form I transforms to form II in contact with a saturated
solution of 5-Cl-A in various organic solvents, with the trans-
formation typically taking place over a period of hours to days,
and with the fastest transformations occurring in acetonitrile and
2-propanol (see PXRD data in ESI†). Thus, the frequent
observation of mixtures of polymorphs (concomitant poly-
morphism) for 5-Cl-A during crystallisation by slow evaporation
corresponds to interrupted solvent-mediated transformation of
I / II, and the isolation of form II alone from 2-propanol or
acetonitrile is consistent with these solvents facilitating the most
rapid transformation.
We consider principally two routes: (i) specific synthesis and
isolation of 5-X-AA, followed by its controlled addition to
solution crystallisation of 5-X-A; (ii) in situ production of
5-X-AA during heating of 5-X-A in organic solvents, especially in
acetic anhydride during synthesis from 5-X-salicylic acid. The
in situ methods are of particular interest because they represent
circumstances where 5-X-AA might be prepared inadvertently
during common synthesis or crystallisation protocols.
Deliberate addition of 5-Cl-AA to the crystallisations did not
alter the outcome, except that it appeared to accelerate the I / II
transformation. In the absence of 5-Cl-AA, the first-appearing
crystals are always form I, and these crystals transform to form II
within ca 3 days in the solvent under ambient conditions. In the
presence of 5-Cl-AA, however, the first crystals that can be iso-
lated are always form II. Thus, crystallisation conditions that
frequently produced bulk mixtures of forms I and II in the
absence of 5-Cl-AA produced only form II when 5-Cl-AA was
present. Similarly, during synthesis of 5-Cl-A by acetylation of 5-
Cl-salicylic acid in acetic anhydride (Scheme 1), the polymorph
formed on direct precipitation of the synthesis mixture was found
to be dependent on the duration of the synthesis heating step:
heating for 10 mins produced only form I, heating for 20 mins
produced a mixture of forms I and II, and heating for 60 minutes
or longer produced only form II. This reflects an increasing
production of 5-Cl-AA during heating, evident from 13C NMR
(see ESI†). The outcome is also crucially dependent on the
quantity of acid that may be added. Smaller quantities of acid
provide form II, while greater quantities provide form I. This
arises because aspirin anhydride is hydrolysed rapidly under
acidic conditions.
Experimental
Synthesis of 5-X-aspirin
5-X-aspirin (X ¼ Cl, Br, Me) was synthesized from 5-X-salicylic
acid (purchased from Sigma Aldrich) and acetic anhydride in the
presence of H2SO4. 5-X-salicylic acid (0.018 mol) was mixed with
acetic anhydride (0.04 mol) and conc. H2SO4 (0.5 ml), and the
mixture was heated at 60 ꢀC for 20 mins. The solution was then
cooled in an ice bath and H2O (40 ml) was added. The resulting
precipitate was isolated by filtration and recrystallised from
acetone.
Synthesis of 5-X-aspirin anhydride
5-X-aspirin anhydride (X ¼ Cl, Br, Me) was synthesized from
5-X-aspirin and N,N0-dicyclohexylcarbodiimide (DCC). 5-X-
aspirin (0.04 mol) was dissolved in cold acetone (30 ml) and
mixed with DCC (0.02 mol) dissolved in cold acetone (15 ml).
The mixture was stirred for 1 hour in an ice bath, then kept
overnight in a refrigerator at 4 ꢀC. A white precipitate of
dicyclohexylurea was filtered off and the solvent was evaporated
under vacuum to give the crude anhydride product as an oil. A
crystalline product could be obtained from hot ethanol or
methanol.
Crystallisation of 5-Br-A
Crystallisation of 5-Br-AA by evaporation from numerous
organic solvents under ambient conditions always produced only
form I. However, form II could be obtained by addition of 5-Br-
AA (in the range 5–20 mol %) to the crystallisations under
ambient conditions, as confirmed by measured unit-cell param-
eters for single crystals, and PXRD of the bulk (see ESI†). Form
II could also be obtained by heating 5-Br-A in acetic anhydride
to prepare 5-Br-AA in situ. For this system, slurrying experiments
establish that form II undergoes solvent-mediated trans-
formation to form I, opposite to the situation in 5-Cl-A.
Comparative experiments establish also that the transformation
rate for 5-Br-A in a given solvent is faster than for 5-Cl-A. The
isolation of only form I during evaporation crystallisation of
pure 5-Br-A, rather than the concomitant polymorphism
Characterisation methods
Single-crystal X-ray diffraction data were collected on a Bruker-
Nonius X8-APEXII CCD diffractometer using graphite-mono-
ꢀ
chromated MoKa radiation (l ¼ 0.7107 A). Powder X-ray
diffraction data were recorded under ambient conditions on
a Siemens D5000 instrument in Bragg-Brentano geometry using
Ge(111)-monochromated CuKa radiation (l ¼ 1.5406 A). 13C
ꢀ
NMR spectra were recorded in CDCl3 at room temperature with
a Bruker Avance III (100 MHz) instrument, using the solvent as
an internal standard.
6992 | CrystEngComm, 2011, 13, 6991–6996
This journal is ª The Royal Society of Chemistry 2011

View Article Online
reflected with respect to that in form I, the mirror plane being
perpendicular to the c axis of form II (vertical in Fig. 1).
On the mode of action of 5-X-AA
The preceding experimental observations are consistent with
a structure-based mechanism for the action of 5-X-AA. Given
that forms I and II of 5-X-A (X ¼ Cl, Br) comprise two iso-
structural sets, it is reasonable to assume that any preferential
interaction of 5-X-AA with growing crystals of one of the two
polymorphs should be comparable in both systems. The experi-
mental observations are consistent with 5-X-AA inhibiting the
growth of form I. For X ¼ Br, form I appears after solvent-
mediated transformation of form II, and the impurity thereby
facilitates isolation of form II in the manner typical of a tailor-
made additive. Since the rate of the II / I transformation in
these systems is relatively fast compared to the rate of solvent
evaporation, the impurity serves to provide a polymorph that is
not otherwise seen during evaporation crystallisation under
ambient conditions. For X ¼ Cl, the order of appearance of the
polymorphs from solution is reversed so that 5-X-AA inhibits
growth of the crystals that would otherwise appear first. The
result is that form II is isolated alone from crystallisations that
otherwise yield the concomitant presence of forms I and II.
Scheme 1 Summary of synthesis conditions and polymorphs produced
for acetylation of 5-X-salicylic acid. Addition of acid to the synthesis can
also influence the polymorphic outcome (see text).
frequently observed for 5-Cl-A, is consistent with this more rapid
solvent-mediated II / I transformation. The presence of 5-Br-
AA during crystallisation apparently acts to inhibit the II / I
transformation, thereby allowing form II to be isolated.
Crystallisation of 5-Me-AA
Structures of the 5-X-A polymorphs (X ¼ Cl, Br)
To date, we have not observed polymorphism for 5-Me-A.
However, whilst attempting to prepare the anticipated 5-Me-A
form II by heating in acetic anhydride, we identified two poly-
morphs for the anhydride 5-Me-AA (Table 2). Initially, we iso-
lated single crystals of a monoclinic form (denoted 5-Me-AA
form II). When the reaction mixture was left to stand for several
days, crystals of a triclinic form (denoted form I) appeared.
Forms I and II were readily distinguished by their habits, being
needles and blocks, respectively. The bulk sample after
Crystallographic information for the 5-X-A polymorphs is
summarised in Table 1. Except for 5-Br-A form II, the structures
have been reported previously by Hursthouse and co-workers.21
Form II of 5-Br-A is isostructural with form II of 5-Cl-A, so that
the 5-X-A polymorphs comprise two isostructural sets. The
structures contain comparable 2-D layers of hydrogen-bonded
dimers (horizontal in Fig. 1). The distinction between them is
that the orientation of the neighbouring layer in form II is
Table 1 Crystallographic data for the polymorphs of 5-X-A (X ¼ Cl, Br, Me)
Cl (Form I)a
Br (Form I)a
Me b
Cl (Form II)
Br (Form II)
CCDC
Formula
NUWTOP
C9H7ClO4
214.60
monoclinic
P21/n
120
10.1217(4)
4.7359(1)
19.4447(7)
90
96.219(2)
90
NUWTIJ
C9H7BrO4
259.06
monoclinic
P21/n
120
10.3912(4)
4.7265(1)
19.5116(9)
90
96.383(2)
90
NUWVEH
C10H10O4
194.18
monoclinic
P21/n
120
10.1301(3)
4.8831(1)
18.9682(7)
90
100.696(2)
90
823448
C9H7ClO4
214.60
monoclinic
P21/c
823449
C9H7BrO4
259.06
monoclinic
P21/c
Formula weight
Crystal system
Space group
T/K
150
150
ꢀ
a/A
5.0838(4)
17.9622(15)
10.4409(7)
90
98.445(3)
90
943.09(13)
4
12316
1659
0.030
5.0940(2)
17.8642(7)
10.6736(4)
90
98.935(2)
90
959.51(6)
4
7302
1666
0.023
ꢀ
b/A
ꢀ
c/A
a/ꢀ
b/ꢀ
g/ꢀ
Volume/A
Z
3
ꢀ
926.60(5)
4
952.35(6)
4
921.98(5)
4
Reflns collected
Unique reflns
Rint
Obs refln (I>2s(I))
R1 (I>2s(I))
wR2 (all data)
GOF on F2
1425
0.031
0.086
1.05
1505
0.020
0.052
1.03
a
Data taken from ref. 21. b Polymorphism has not been observed to date for X ¼ Me.
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 6991–6996 | 6993

View Article Online
were obtained on crystallisation from organic solvents under
ambient conditions, while the monoclinic forms (which are not
isostructural, denoted form II) could be obtained by reflux in
acetic anhydride or ethanol, in this case without any requirement
to add 5-X-A. The fact that 5-X-A is generated in situ during
heating is apparent from PXRD of the bulk products, which
clearly contain crystalline 5-X-A (see ESI†). In this case, an
explicit link between the observed polymorphism and the pres-
ence of 5-X-A is less clear, however, since heating cannot be
decoupled from the generation of 5-X-A, i.e. it is not possible to
heat 5-X-AA without introducing significant amounts of 5-X-A.
It is also quite difficult to obtain pure samples of 5-X-AA (X ¼
Cl, Br), since some decomposition to 5-X-A is invariably
observed. For 5-Me-AA, the situation is different because it is the
most stable of the anhydrides in solution, so that the amount of
5-Me-A generated on heating is much less than the amounts of 5-
Cl-A or 5-Br-A generated from 5-Cl-AA or 5-Br-AA under the
same conditions.
Structures of the 5-X-AA polymorphs (X ¼ Cl, Br, Me)
In form I of 5-Me-AA (Fig. 2), the molecules adopt a confor-
mation where the planes of the aromatic rings are essentially
perpendicular. Face-to-face intermolecular interactions between
rings link the molecules into columns along the c axis. Between
columns, centrosymmetric C–H$$$O motifs exist between acetyl
groups, comparable to those in aspirin form I.23 In 5-Me-AA
form II, the molecules adopt a flatter conformation, in which the
angle between the planes of the aromatic rings is ca 70ꢀ. The rings
form edge-to-face contacts between molecules, and layers can be
envisaged in the bc plane (horizontal in Fig. 2). There are two
distinct interlayer regions (central in Fig. 2, and between the
upper and lower boundaries of the unit cell). Both contain C–
H$$$O contacts formed from the aromatic rings to the acetyl
groups, while one of them (central in Fig. 2) also contains cate-
meric C–H$$$O motifs between acetyl groups, comparable to
those in aspirin form II.9
Fig. 1 Crystal structures of 5-X-A (X ¼ Cl, Br) forms I and II. The
structures contain comparable layers of hydrogen-bonded dimers
(horizontal).
evaporation to dryness was shown by PXRD to contain
a significant quantity of 5-Me-A (see ESI†). For pure samples of
5-Me-AA prepared by independent synthesis (using the method
described in the Experimental section), crystallisation from
various organic solvents yielded only form I, therefore suggesting
that form II might be prepared only in the presence of 5-Me-A.
Systematic addition of 5-Me-A did not in fact have any influence
on the crystallisation of 5-Me-AA under ambient conditions, but
form II could be obtained reproducibly by refluxing 5-Me-AA in
ethanol or acetone with the concurrent addition of 10 mol % of 5-
Me-A. Reflux without addition of 5-Me-A yielded only form I
(i.e. the varying crystallisation outcome is not solely attributable
to temperature), while addition of larger amounts of 5-Me-A
yielded form II together with crystalline 5-Me-A after evapora-
tion to dryness. Slurrying experiments confirmed that form II of
5-Me-AA transforms rapidly to form I in organic solvents, so
that the presence of 5-Me-A during reflux in ethanol or acetone
serves apparently to stabilise the first-appearing 5-Me-AA poly-
morph II.
For 5-Cl-AA and 5-Br-AA, form I is isostructural. The
molecular conformation is comparable to that in 5-Me-AA form
I, except that one of the acetyl groups points in the opposite
direction with respect to the remainder of the molecule. The
structure contains face-to-face contacts between aromatic rings
of neighbouring molecules along the b axis, but these are dis-
torted to accomodate other interactions, particularly that
between the acetyl group and the central O atom of a neigh-
bouring molecule. The columns along the b axis can be envisaged
to form 2-D sections in the ab planes, with neighbouring mole-
cules forming C–H$$$O contacts between the aromatic rings and
the carbonyl groups of the anhydride unit, and with one C–Cl
bond directed towards the centroid of a neighbouring aromatic
ring. This description of layers in the ab planes highlights simi-
larity with 5-Cl-AA form II (Fig. 4), which contains closely
comparable 2-D sections. In both 5-Cl-AA polymorphs, the
principal contacts between layers are C–H$$$O contacts between
the aromatic rings and the acetyl groups. The distinction between
the polymorphs is that the orientation of the neighbouring layer
in form II is reflected with respect to that in form I, the mirror
plane being perpendicular to the a axis. The structures therefore
look essentially identical in projection along a (Figs. 3 and 4).
Crystallisation of 5-X-AA (X ¼ Cl, Br)
Extension of these results to 5-X-AA with X ¼ Cl and Br also
yielded two polymorphs for these compounds (Table 2).‡ The
orthorhombic forms (which are isostructural, denoted form I)
‡ Crystallisation from acetonitrile produced needle crystals of a solvate
5-X-AA$CH3CN, for X ¼ Cl and Br. The crystals lose solvent rapidly
and to date we have obtained only poor quality X-ray crystal structures.
6994 | CrystEngComm, 2011, 13, 6991–6996
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Table 2 Crystallographic data for the polymorphs of 5-X-AA (X ¼ Cl, Br, Me)
a
Cl (Form I)
Br (Form I)
Me (Form I)
Cl (Form II)
Br (Form II)
Me (Form II)
CCDC
Formula
823451
C18H12Cl2O7
411.18
orthorhombic
P212121
150
10.5412(7)
12.4530(8)
13.5129(10)
90
90
90
1773.8(2)
4
18370
3363
0.040
2943
0.030
0.065
1.01
ꢁ0.01(5)
823453
C18H12Br2O7
500.10
orthorhombic
P212121
150
10.5579(5)
12.5176(7)
13.7487(8)
90
823456
C20H18O7
370.34
triclinic
ꢁ
P1
823452
C18H12Cl2O7
411.18
monoclinic
P21/c
823454
C18H12Br2O7
500.10
monoclinic
P21/c
823457
C20H18O7
370.34
Monoclinic
P21/c
Formula weight
Crystal system
Space group
T/K
150
150
150
150
ꢀ
a/A
9.9176(9)
9.9757(10)
11.5419(17)
95.485(7)
109.352(7)
117.882(4)
908.32(18)
2
10529
3189
0.042
1997
10.7838(3)
14.2206(4)
12.3774(4)
90
107.635(3)
90
1808.90(9)
4
7319
3766
0.021
16.9944(7)
7.9432(4)
14.0330(7)
90
105.825(2)
90
1822.52(15)
4
24574
3205
0.039
20.749(2)
8.3818(8)
10.6736(10)
90
97.440(5)
90
ꢀ
b/A
ꢀ
c/A
a/ꢀ
b/ꢀ
g/ꢀ
Volume/A
Z
90
90
3
ꢀ
1817.02(17)
4
17330
3427
0.030
3211
0.018
0.039
0.99
0.005(5)
1840.7(3)
4
Reflns collected
Unique reflns
Rint
Obs refln (I>2s(I))
R1 (I>2s(I))
wR2 (all data)
GOF on F2
Flack parameter
47624 a
5172
0.065
a
2966
0.040
0.102
1.03
2629
0.023
0.050
1.01
3242
0.040
0.086
0.90
0.045
0.113
1.00
a
The crystal was integrated as a non-merohedral twin with two components related by 180ꢀ rotation around a*. The total number of reflections and Rint
value correspond to all single-component and overlapped reflections. Refinement performed using SHELXL HKLF-5 format: refined BASF ¼ 0.476(1).
Fig. 3 Crystal structure of 5-X-AA form I (X ¼ Cl, Br).
The structure of 5-Br-AA form II is distinct. It resembles 5-Me-
AA form I, in that it contains face-to-face contacts between the
aromatic rings, but the molecule has a flatter conformation (ca
45ꢀ between ring planes), and there are also clear Br$$$Br
ꢀ
contacts (3.5352(3) A).
Conclusions
We have shown that the outcome of solution crystallisations for
the polymorphic 5-X-aspirin (X ¼ Cl, Br) can be influenced by
the presence of 5-X-aspirin anhydride, in a manner similar to
that shown for aspirin itself.10 For X ¼ Cl, the effect is subtle
because the impurity inhibits growth of form I, which in any case
transforms to form II on standing in the crystallisation solvent.
The presence of the anhydride can influence whether or not
Fig. 2 Crystal structures of the 5-Me-AA polymorphs.
This journal is ª The Royal Society of Chemistry 2011
concomitant polymorphism is observed under given
CrystEngComm, 2011, 13, 6991–6996 | 6995

View Article Online
example, the polymorphic outcome can depend critically on the
duration of heating during synthesis of the compound. For 5-Cl-
A in particular, a difference of only a few minutes during the
heating step in Scheme 1 is sufficient to change the crystallisation
outcome completely. Such circumstances can lead to erratic and
apparently irreproducible crystallisation results, especially if the
impurities remain undetected.
Acknowledgements
This work was supported by the Danish Council for Independent
Research | Natural Sciences. We are grateful to Ismail Cumar
(Dept. of Physics and Chemistry, University of Southern
Denmark) for collecting 13C NMR data.
Notes and references
1 J. Bernstein, Polymorphism in Molecular Crystals, Oxford University
Press, New York, 2002.
2 R. Hilfiker, ed. Polymorphism in the Pharmaceutical Industry, Wiley-
VCH, Weinheim, 2006.
3 W. Ostwald, Z. Phys. Chem., 1897, 22, 289.
4 T. Threlfall, Org. Process Res. Dev., 2003, 7, 1017.
5 R. J. Davey, N. Blagden, G. D. Potts and R. Docherty, J. Am. Chem.
Soc., 1997, 119, 1767.
6 I. Weissbuch, R. Popovitz-Biro, M. Lahav and L. Leiserowitz, Acta
Crystallogr., Sect. B: Struct. Sci., 1995, B51, 115.
7 I. Weissbuch, M. Lahav and L. Leiserowitz, Cryst. Growth Des., 2003,
3, 125.
8 A. D. Bond, R. Boese and G. R. Desiraju, Angew. Chem., Int. Ed.,
2007, 46, 615.
Fig. 4 Crystal structures of 5-X-AA form II (X ¼ Cl, Br). For 5-Cl-AA,
similarity with the structure of 5-X-AA-form I (Fig. 3) is seen more clearly
by shifting the molecules ca 1/4 unit along the b axis.
9 A. D. Bond, R. Boese and G. R. Desiraju, Angew. Chem., Int. Ed.,
2007, 46, 618.
10 A. D. Bond, K. A. Solanko, S. Parsons, S. Redder and R. Boese,
CrystEngComm, 2011, 13, 399.
11 S. Varughese, M. S. R. N. Kiran, K. A. Solanko, A. D. Bond,
U. Ramamurty and G. R. Desiraju, Chem. Sci., 2011, DOI:
10.1039/C1SC00430A.
12 A. L. De Weck, Int. Arch. Allergy Immunol., 1971, 41, 393.
13 H. Bundgaard and C. Bundgaard, J. Pharm. Pharmacol., 1973, 25,
593.
14 H. Bundgaard, J. Pharm. Pharmacol., 1974, 26, 535.
15 A. G. Mitchell and D. J. Saville, J. Pharm. Pharmacol., 1967, 19, 729.
16 R. Tawashi, Science, 1968, 160, 76.
17 A. G. Mitchell and D. J. Saville, J. Pharm. Pharmacol., 1969, 21, 28.
18 R. Tawashi, J. Pharm. Sci., 1971, 60, 1420.
19 M. P. Summers, J. E. Carless and R. P. Enever, J. Pharm. Pharmacol.,
1970, 22, 615.
20 B. Jerslev and U. Lund, Arch. Pharm. Chem. Sci. Ed., 1981, 9, 61.
21 M. B. Hursthouse, R. Montis and G. J. Tizzard, CrystEngComm,
2010, 12, 953.
22 M. B. Hursthouse, private communication.
23 P. J. Wheatley, J. Chem. Soc., 1964, 6036.
crystallisation conditions. For X ¼ Br, the influence of the
anhydride is more clear because it serves to yield a polymorph
that is not otherwise seen under comparable crystallisation
conditions.
We have also found that the 5-X-aspirin anhydrides (X ¼ Cl,
Br, Me) are polymorphic and that the crystallisation outcome for
these compounds appears also to be influenced by the presence of
5-X-aspirin. The link in this case is less clear, because it is difficult
to prepare pure samples of the anhydride for X ¼ Cl and Br, and
heating seems to be required together with the additive for
polymorphism to be observed.
The fact that impurities can influence crystallisation outcomes
is of course well known. The principal significance of this work
probably lies in an illustration of polymorph control induced by
specific impurities that can be generated as by-products during
synthesis or a common crystallisation procedure such as heating
in an organic solvent. In several of the cases described here, for
6996 | CrystEngComm, 2011, 13, 6991–6996
This journal is ª The Royal Society of Chemistry 2011