Solution mediated phase transformations between co-crystals

Article abstract of DOI:10.1039/c2ce26801a

A solution mediated transformation between two co-crystal phases has been observed for the p-toluensulfonamide-triphenylphosphine oxide co-crystal system. This system has two known co-crystals with 1 : 1 and 3 : 2 stoichiometry respectively, and the ternary phase diagram (TPD) for the system has been determined in acetonitrile previously. By manipulating the solution composition in this solvent to a region of the TPD where the 1 : 1 co-crystal is stable, the 3 : 2 co-crystal could be observed to convert to the 1 : 1 co-crystal. The corresponding transformation was true for the 1 : 1 co-crystal in a region of the TPD where the 3 : 2 co-crystal is stable; the 1 : 1 co-crystal converted to the 3 : 2 co-crystal.

Full text of DOI:10.1039/c2ce26801a

CrystEngComm
View Article Online
View Journal | View Issue
COMMUNICATION
Solution mediated phase transformations between co-
crystals
Cite this: CrystEngComm, 2013, 15, 2044
Denise M. Croker,*ac Roger J. Davey,b Åke C. Rasmusona and Colin C. Seaton*a
Received 2nd November 2012,
Accepted 14th December 2012
DOI: 10.1039/c2ce26801a
www.rsc.org/crystengcomm
A
solution mediated transformation between two co-crystal
system.18 The carbamazepine–nicotinamide system has been
investigated to study both the competing crystallisation kinetics
of the differing phases in regions where two solid phases are
stable11 and the solution mediated transformation from carba-
mazepine to the co-crystal through the introduction of enough
solid nicotinamide to shift the composition into a different region
of the TPD.14 In the carbamazepine–nicotinamide system only a
single co-crystal phase is known and so only the conversion
between the starting materials and the co-crystal are possible.
While a number of systems exhibit co-crystals of differing
stoichiometry and the identification of the regions of the TPD
where the differing phases are stable has been reported,16,19,20 the
conversion process between these differing co-crystal phases has
not been fully investigated.
It has been shown for some co-crystal systems that the relative
stability of co-crystals with differing stoichiometries can be altered
through the choice of solvent. For example, the benzoic acid–
isonicotinamide pair has two co-crystals: one with a 1 : 1
composition and one with a 2 : 1 composition.21 Both phases
have been shown to be stable in different aqueous solution
compositions but only the 1 : 1 co-crystal is stable in ethanolic
solutions. This was demonstrated by the addition of crystals of the
2 : 1 phase to a saturated solution, which lead to the complete
dissolution of the 2 : 1 crystals in 20 minutes, followed by
crystallisation of the 1 : 1 phase after 10 hours.13 Understanding
the balance of stability between the differing phases and the
choice of solvent is thus required to design suitable conditions for
the controlled growth of selected phases. Such understanding was
recently applied to the successful growth of crystals suitable for
single crystal structure determination of the 1 : 2 malic acid and
caffeine co-crystal. This malic acid–caffeine system exists in two
stoichiometries (1 : 1 and 1 : 2) and both phases are readily
obtainable through solid-state grinding. However, only crystals of
the 1 : 1 phase suitable for single crystal structure solution could
be obtained by slow evaporation of dichloromethane solution.5 In
contrast, only poorly diffracting crystals of the 1 : 2 co-crystal
could be produced through anti-solvent addition of cyclohexane to
a saturated 1 : 2 solution in 15 : 1 chloroform–methanol mixture.
Investigations into the relative solubilities of the components
phases has been observed for the p-toluensulfonamide–triphe-
nylphosphine oxide co-crystal system. This system has two known
co-crystals with 1 : 1 and 3 : 2 stoichiometry respectively, and the
ternary phase diagram (TPD) for the system has been determined
in acetonitrile previously. By manipulating the solution composi-
tion in this solvent to a region of the TPD where the 1 : 1 co-
crystal is stable, the 3 : 2 co-crystal could be observed to convert
to the 1 : 1 co-crystal. The corresponding transformation was true
for the 1 : 1 co-crystal in a region of the TPD where the 3 : 2 co-
crystal is stable; the 1 : 1 co-crystal converted to the 3 : 2 co-
crystal.
The creation of co-crystals has recently gained increased interest
within the solid-state community as a route for the modification of
the physico–chemical properties of active pharmaceutical ingre-
dients (APIs).1–5 While the creation of such materials has been
repeatedly demonstrated, studies into the crystallisation processes
and transformations between such materials has only recently
received attention.6–17 However, such understanding is required to
fully develop suitable crystallisation processes for the industrial
synthesis of pharmaceutical co-crystals.
The key source of information for the design of a solution-
based co-crystallisation is the ternary phase diagram (TPD), which
indicates the thermodynamically stable phase for
a given
composition.6 While previous studies have, in general, focused
on the development of methodologies for the construction of
TPDs and their application in the growth of a stable phase, less
work has been carried out on the kinetics of the crystallisation and
influence of conditions on the growth of the differing phases. The
solution mediated transformation between polymorphic co-
crystals has been reported for the carbamazepine–isonicotinamide
aSolid State Pharmaceutical Cluster, Materials and Surface Science Institute and the
Department of Chemical and Environmental Sciences, University of Limerick,
Limerick, Ireland. E-mail: denise.croker@merck.com; colin.seaton@ul.ie;
Fax: +353 61 213529; Tel: +353 61 234166
bSchool of Chemical Engineering and Analytical Science, University of Manchester,
Oxford Road, Manchester, M13 9PL, UK
cMSD Ballydine, Kilsheelan, Clonmel, Co., Tipperary, Ireland
2044 | CrystEngComm, 2013, 15, 2044–2047
This journal is ß The Royal Society of Chemistry 2013

View Article Online
CrystEngComm
Communication
indicated significant differences between the two components for
the majority of solvents and construction of the ternary phase
diagram for malic acid–caffeine–acetone indicated that the 1 : 2
system was metastable for this system.20 Selection of a solvent that
reduced the differences in solubility lead to the identification of a
system where both co-crystal phase were stable. Construction of
the TPD for malic acid–caffeine–ethyl acetate allowed for the
controlled growth of the 1 : 2 co-crystal.22
For polymorphic systems, solution mediated transformations
are an important process for the conversion of metastable phases
into stable crystalline phases.23,24 While the general mechanism is
the dissolution of the metastable phase and regrowth of the stable
one, these conversions may also occur through heterogeneous
nucleation driven by an epitaxial interaction between the two
phases due to similarities in selected surfaces of the two forms.24
For example, the solution mediated transformation of a to b
glutamic acid occurs by nucleation of the b phase on the {111}
faces of the a phase.25–28 Recently, work in identifying the different
possible transformation scenarios for polymorph transformations
has highlighted the balance between dissolution of the metastable
phase and the nucleation and growth of the stable phase.29
Similar factors will control the transformation of metastable co-
crystal phases to stable co-crystal phases, however unlike
polymorphic systems the stability of a given co-crystal is
dependant on the composition of the solution. Thus the
transformation process can be easily studied by the addition of
crystals of one co-crystal into a solution with a composition where
the other co-crystal is stable.
The system comprising p-toulenesulfonamide (TSA) and
triphenylphosphine oxide (PH3PO) was initially reported to form
as a 3 : 2 co-crystal30 but subsequent study has identified an
additional 1 : 1 co-crystal phase.16 Construction of TPDs in
acetonitrile and dichloromethane has allowed for the identifica-
tion of distinct solution compositions where each individual co-
crystal is the thermodynamically stable phase. Knowledge of these
solution compositions allows for preparation of a solution
saturated with respect to one specific co-crystal form. This study
investigated the behaviour of the 3 : 2 co-crystal of the TSA–PH3PO
system when it was placed in a solution saturated with respect to
the 1 : 1 co-crystal, and vice versa. The subsequent solution
mediated co-crystal transformation was observed with optical
microscopy in real time, and confirmed with PXRD on comple-
tion.
Fig. 1 Solubility diagram for TSA–Ph3PO in acetonitrile. The experimental
compositions studied are indicated by the green and red crosses.
60 uC and cooling at 0.1 uC min21 to 20 uC. The resulting solutions
were covered with parafilm into which a number of small holes
were punched to allow evaporation, and left to stand at room
temperature until the crystals increased to a suitable size (3–5
mm). The seed crystals were placed in 5 mL of saturated solution
in a temperature controlled optical cell at 20 uC, and a microscope
image taken every 15 seconds. The 3 : 2 co-crystal was placed in a
solution containing 0.20 mol dm23 TSA and 0.15 mol dm23
Ph3PO (Fig. 1, green crosses), and the 1 : 1 co-crystal in a solution
containing 0.56 mol dm23 TSA and 0.08 mol dm23 Ph3PO (Fig. 1,
red crosses).
The 3 : 2 crystal undergoes dissolution initially, and additional
crystals appear independently of the seed crystal (Fig. 2). Growth of
the new crystals continued until all the seed phase had dissolved.
PXRD of the resulting phase confirmed that the crystals were the
1 : 1 co-crystal (Fig. 4).
During the conversion of the 1 : 1 crystal, the newly formed
crystals cluster closely around the seed crystal and may be
associated with the dissolving surface (Fig. 3). As each experiment
is undertaken without stirring, a local increase in the super-
saturation level surrounding the dissolving crystal may occur. This
would increase the probability of nucleation and results in the
growth of the expected phase in the vicinity of the seed crystal. The
The saturation concentrations of TSA and Ph3PO within the
region in acetonitrile solution where the 1 : 1 and 3 : 2 co-crystal
are respectively stable, were initially measured using a Crystal
16TM system from Avantium Technologies (Fig. 1). For this
3
system, both composition lines for the two co-crystals cross the
1 : 1 co-crystal solubility, confirming that the 1 : 1 co-crystal
displays congruent dissolution, while the 3 : 2 displays incon-
gruent dissolution as shown by the previously determined TPD.16
The behaviour of each co-crystal in a solution saturated with
respect to the other co-crystal respectively was investigated with
optical microscopy.{ Seed crystals of the 1 : 1 and 3 : 2 co-crystal
were produced by heating a 10 : 20 : 70 (TSA : Ph3PO : MeCN)
and a 20 : 10 : 70 (TSA : Ph3PO : MeCN) mass fraction solution to
Fig. 2 Solution mediated transformation of the 3 : 2 co-crystal into the 1 : 1 co-
crystal in MeCN. Each image represents a time lapse of 10 minutes. The starting
crystal was approximately 4 mm in its longest dimension.
This journal is ß The Royal Society of Chemistry 2013
CrystEngComm, 2013, 15, 2044–2047 | 2045

View Article Online
Communication
CrystEngComm
Fig. 3 Solution mediated transformation of the 1 : 1 co-crystal into the 3 : 2 co-
crystal in MeCN. Each image represents a time lapse of 20 minutes. The starting
crystal was approximately 4 mm in its longest dimension.
Fig. 6 BFDH predicted morphology for the 1 : 1 co-crystal showing the
molecular packing and indexing of selected crystal faces.
final product in this case was confirmed as the 3 : 2 co-crystal
using PXRD (Fig. 5).
The recorded images do not clearly define if heterogeneous
nucleation occurred on the surface of the co-crystal. As the seed
crystals are poorly formed, the habits of the crystals have not been
characterised and so identification of the molecular features on
each surface preventing such interactions is limited. However,
prediction of the morphology by BFDH gives qualitatively similar
results to the experimental shapes (Fig. 6 and 7). From the
predicted morphologies, the 1 : 1 co-crystal is bounded by
predominately aromatic functionalities (Fig. 6), while the 3 : 2
co-crystal has an even mixture of aromatic and sulfonamide
functional group present on the surfaces (Fig. 7). Thus the
dominant faces of the two phases lack complementary surfaces
and the potential for strong directional interactions to direct a
surface to surface interaction as observed in the epitaxially driven
polymorphic transformations.23 Lattice matching calculations
using epicalc31 also indicate that there is no lattice registry
between the predicted dominant faces of either phase.
Fig. 4 PXRD of 3 : 2 single crystal which was equilibrated in MeCN saturated
with the 1 : 1 co-crystal and the product of this equilibration. Reference pattern
for the two co-crystals are also included.
Fig. 5 PXRD of 1 : 1 single crystal which was equilibrated in MeCN saturated
with the 3 : 2 co-crystal and the product of this equilibration. Reference pattern
for the two co-crystals are also included.
Fig. 7 BFDH predicted morphologies for the 3 : 2 co-crystal showing the
molecular packing and indexing of selected crystal faces.
2046 | CrystEngComm, 2013, 15, 2044–2047
This journal is ß The Royal Society of Chemistry 2013

View Article Online
CrystEngComm
Communication
11 E. Gagniere, D. Mangin, F. Puel, C. Bebon, J.-P. Klein,
O. Monnier and E. Garcia, Cryst. Growth Des., 2009, 9,
3376–3383.
12 M. Habgood, M. A. Deij, J. Mazurek, S. L. Price and J. H. ter
Horst, Cryst. Growth Des., 2010, 10, 903–912.
13 S. Boyd, K. Back, K. Chadwick, R. J. Davey and C. C. Seaton, J.
Pharm. Sci., 2010, 99, 3779–3786.
14 E. Gagniere, D. Mangin, F. Puel, J.-P. Valour, J.-P. Klein and
O. Monnier, J. Cryst. Growth, 2011, 316, 118–125.
15 M. A. Oliveira, M. L. Peterson and R. J. Davey, Cryst. Growth
Des., 2011, 11, 449–457.
16 D. M. Croker, M. E. Foreman, B. N. Hogan, N. M. Maguire, C.
J. Elcoate, B. K. Hodnett, A. R. Maguire, Å. C. Rasmuson and S.
E. Lawrence, Cryst. Growth Des., 2012, 12, 869–875.
17 Y. Feng, L. Dang and H. Wei, Cryst. Growth Des., 2012, 12,
2068–2078.
Conclusions
The solution mediated transformation between two co-crystal
phases has been demonstrated for the p-toulenesulfonamide–
triphenylphosphine oxide system in acetonitrile. In this system no
surface interaction appears to occur between the dissolving
metastable phase and the growing stable phase, unlike many
transformations observed for polymorphic systems. This may be
due to the differences in surface chemistry of the two phases and
suitable epitaxial interactions cannot be formed in this case. This
result indicates the importance of understanding co-crystal
stability when attempting to isolate a desired co-crystal form.
Notes and references
Experiments were performed on a 1 mL scale. Mass fractions, on a total
3
18 J. H. T. Horst and P. W. Cains, Cryst. Growth Des., 2008, 8,
2537–2542.
19 A. Jayasankar, L. S. Reddy, S. J. Bethune and N. Rodriguez-
Hornedo, Cryst. Growth Des., 2009, 9, 889–897.
20 K. Guo, G. Sadiq, C. C. Seaton, R. J. Davey and Q. Yin, Cryst.
Growth Des., 2010, 10, 268–273.
mass basis, were selected to represent regions in the TPD where the 3 : 2
co-crystal or the 1 : 1 co-crystal are independently stable as indicated in the
previously determined ternary phase diagram. The solutions were held at
60 uC for 1 hour to ensure complete dissolution and subsequently cooled at
1 uC min21 to 20 uC, and aged for 1 hour. The resulting solids were filtered
and analysed with powder X-ray diffraction for phase composition.
Reference PXRD patterns for the 3 : 2 and 1 : 1 co-crystal were generated
from their respective crystallographic information files (CIF) using Mercury
2.4.
21 C. C. Seaton, A. Parkin, C. C. Wilson and N. Blagden, Cryst.
Growth Des., 2009, 9, 47–56.
{ Optical microscopy was performed using a Zeiss Axioplan 2 polarizing
microscope and the Linksys image capture software.
22 T. Leyssens, G. Springuel, R. Montis, N. Candoni and S. Veesler,
Cryst. Growth Des., 2012, 12, 1520–1530.
¨
23 D. M. Croker and B. K. Hodnett, Cryst. Growth Des., 2010, 10,
2806–2816.
1 O. Almarsson and M. J. Zaworotko, Chem. Commun., 2004,
1889–1896.
2 N. Schultheiss and A. W. Newman, Cryst. Growth Des., 2009, 9,
2950–2967.
24 P. T. Cardew and R. J. Davey, Proc. R. Soc. London,Ser. A, 1985,
398, 415–428.
3 I. Miroshnyk, S. Mirza and N. Sandlert, Expert Opin. Drug
Delivery, 2009, 6, 333–341.
25 C. Cashell, D. Sutton, D. Corcoran and B. K. Hodnett, Cryst.
Growth Des., 2003, 3, 869–872.
4 N. Shan and M. J. Zaworotko, Drug Discovery Today, 2008, 13,
440–446.
5 A. V. Trask, W. D. S. Motherwell and W. Jones, Cryst. Growth
Des., 2005, 5, 1013–1021.
6 R. A. Chiarella, R. J. Davey and M. L. Peterson, Cryst. Growth
Des., 2007, 7, 1223–1226.
7 J. H. ter Horst, M. A. Deij and P. W. Cains, Cryst. Growth Des.,
2009, 9, 1531–1537.
8 R. R. Schartman, Int. J. Pharm., 2009, 365, 77–80.
9 K. Chadwick, R. J. Davey, G. Dent, R. G. Pritchard, C. A. Hunter
and D. Musumeci, Cryst. Growth Des., 2009, 9, 1990–1999.
10 A. Ainouz, J.-R. Authelin, P. Billot and H. Lieberman, Int. J.
Pharm., 2009, 374, 82–89.
26 C. Cashell, D. Corcoran and B. K. Hodnett, Chem. Commun.,
2003, 374–375.
27 E. Ferrari and R. J. Davey, Cryst. Growth Des., 2004, 4,
1061–1068.
¨
28 J. Scholl, D. Bonalumi, L. Vicum, M. Mazzotti and M. Mu¨ller,
Cryst. Growth Des., 2006, 6, 881–891.
29 M. A. O’Mahony, A. Maher, D. M. Croker, Å. C. Rasmuson and
B. K. Hodnett, Cryst. Growth Des., 2012, 12, 1925–1932.
30 G. Ferguson and C. Glidewell, J. Chem. Soc., Perkin Trans. 2,
1988, 2129–2132.
31 A. Hillier and M. D. Ward, Phys. Rev. B: Condens. Matter, 1996,
54, 14037–14051.
This journal is ß The Royal Society of Chemistry 2013
CrystEngComm, 2013, 15, 2044–2047 | 2047