Received 15 April 2013
Solvates of the antifungal drug griseofulvin: structural, thermochemical and conformational analysis
aCrystallization and Particle Science, Institute of Chemical and Engineering Sciences, 1 Pesek Road, Jurong Island, Singapore 627833, and bDepartment of Chemical and Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, Singapore 117576
Four solvates of an antifungal drug, griseofulvin (GF), were discovered. All the solvates were characterized by differential scanning calorimetry, thermogravimetric analysis, and their crystal structures were determined by single-crystal X-ray diffraction. The solvents that form the solvates are acetonitrile, nitromethane and nitroethane (2:1 and 1:1). It was found that all the solvates lose the solvent molecules from the crystal lattice between 343 and 383 K, and that the melting point of the desolvated materials matched the melting point of the solvent-free GF (493 K). The conformation of the GF molecule in solvent-free form was found to be significantly different from the conformations found in the solvates. Solution stability studies revealed that the GF-acetonitrile solvate transforms to GF and that GF-nitroethane (1:1) solvate transforms to GF-nitroethane (2:1) solvate. On the other hand, GF-nitromethane and GF-nitroethane (2:1) solvates were found to be stable in solution. Our results highlight the importance of the co-crystallization technique in the pharmaceutical drug development; it not only expands the solid form diversity but also creates new avenues for unraveling novel solvates.
Finding a suitable solid form for a non-ionic, poorly soluble active pharmaceutical ingredient (API) is an utmost challenge in drug development (Morissette et al., 2004; Shan & Zaworotko, 2008; Blagden et al., 2007). This is primarily because of the inability to apply the most common method of solubility improvement, i.e. salt formation. Therefore, alternative solid forms such as the amorphous form, metastable polymorphs, hydrates, solvates and co-crystals are considered for development (Morissette et al., 2004). Solvates are referred to as crystalline solids that incorporate one or more solvent molecules in the crystal lattice. As with other solid forms, solvates may show unique physical and pharmaceutical properties (Byrn et al., 1999). It has also been reported that desolvation of a solvate yields a particular polymorph, thus suggesting potential implications for polymorph control (Giron, 1995; Wirth & Stephenson, 1997). This has been demonstrated by a number of APIs, such as indomethacin, axitinib, sulfamerazine etc. (Aitipamula et al., 2011; Aitipamula, Chow & Tan, 2012; Desikan et al., 2005; Joshi et al., 1998; Braun et al., 2009). Therefore, from a pharmaceutical and supramolecular perspective, knowledge of solvate-formation ability of an API is important in an attempt to find an optimal solid form for drug development.
Griseofulvin (GF, Fig. 1) is an antifungal drug administered orally for the treatment of dermatomycoses including ringworm, athlete's foot, and infections of the scalp and nails (Chan & Friedlander, 2004; Kassem et al., 2006). According to the Biopharmaceutics Classification System, GF is a Class II drugs which means it has low solubility and high permeability (Kasim et al., 2004). We have recently reported a co-crystal hydrate of GF with an artificial sweetener, acesulfame (Aitipamula, Vangala et al., 2012). Performance of the co-crystal hydrate has been evaluated and compared with the known amorphous form of GF, which suggested a greater physical stability, and higher solubility and dissolution rate for the co-crystal hydrate. Notably, all our attempts to find co-crystals with various other pharmaceutically acceptable co-crystal formers have not been successful (Aitipamula, Vangala et al., 2012). However, some of these attempts helped to unravel novel solvates of GF. In this paper we report structural and thermochemical analysis, desolvation studies and microscopic analysis of four solvates of GF with the solvents acetonitrile, nitromethane and nitroethane. Interestingly, nitroethane forms 1:1 and 2:1 solvates concomitantly (Fig. 2). We recall here that GF is known to form several crystalline solvates. For example, Grant and Abougela reported solvates with a series of fatty acids (Grant & Abougela, 1981). GF also forms solvates with alkyl halides, alkyl dihalides, chloroform, benzene and 1,4-dioxane (Shirotani et al., 1988; Sekiguchi et al., 1964, 1968, 1976; Cheng et al., 1979).
| || Figure 1 |
Molecular structure of GF and the solvents that produced solvates reported in this paper.
| || Figure 2 |
A microscopic image showing concomitant crystallization of 1:1 (blocks) and 2:1 (needles) GF-nitroethane solvates.
GF was purchased from Junda Pharmaceuticals Ltd, People's Republic of China, and used as received. Analytical grade solvents were used for the crystallization experiments.
Crystallization of GF from common organic solvents resulted in block-shaped crystals that belonged to the pure GF (Malmros et al., 1977; Puttaraja & Sakegowda, 1982). Block-shaped crystals of acetonitrile solvate (GF-acetonitrile) were obtained when a 1:1 molar ratio of GF (100 mg, 0.283 mmol) and gentisic acid (43.7 mg, 0.283 mmol) was co-crystallized from acetonitrile. Nitromethane solvate was obtained from an attempted co-crystallization of GF (100 mg, 0.283 mmol) and saccharin (51.8 mg, 0.283 mmol) in a 1:1 molar ratio from nitromethane. Slow evaporation of a saturated solution of GF in nitroethane resulted in block- and needle-shaped crystals (Fig. 2). 1H NMR spectroscopy1 and thermogravimetric analysis (TGA; to be discussed later) confirmed both the crystals to be nitroethane solvates, but with different stoichiometries. Whereas block-shaped crystals contained the GF and nitroethane in a 1:1 molar ratio, the needle-shaped crystals belong to a 2:1 solvate. All the solvates were reproduced consistently by the experimental methods described.
X-ray reflections were collected on a Rigaku Saturn CCD area detector with graphite-monochromated Mo K radiation ( = 0.71073 Å). Data were collected and processed using CrystalClear (Rigaku, 2007) software. Structures were solved by direct methods and SHELXTL (Sheldrick, 2008) was used for structure solution and least-squares refinement. Non-H atoms were refined anisotropically. All H atoms were fixed at idealized positions. All C-H distances were normalized to the neutron diffraction value of 1.083 Å. In the crystal structure of GF-nitroethane (1:1), electron densities that correspond to the solvent, nitroethane, molecules could not be modelled. Hence, these were squeezed out using PLATON/SQUEEZE (Spek, 2009). The data thus obtained without the nitroethane molecules were used for further crystallographic refinement. 1H NMR spectroscopy and TGA analysis confirmed the 1:1 stoichiometry of GF and nitroethane. Crystallographic data for the structures described in this paper are listed in Table 1. Details of hydrogen-bond parameters in the crystal structures are given in Table 2.
Powder samples obtained in desolvation and slurry experiments were identified using a D8 Advance X-ray powder diffractometer (Bruker AXS GmbH, Germany) with Cu K radiation ( = 1.54056 Å). The voltage and current applied were 35 kV and 40 mA, respectively. Samples were placed on the sample holder which has 1 mm thickness and 1.5 cm diameter. The sample was scanned within the scan range 2 = 5-50° continuous scan, with a scan rate of 2° min-1. The PXRD patterns were plotted using OriginPro7.5.
DSC was performed with a Perkin-Elmer Diamond DSC with an autosampler. Crystals taken from the mother liquor were blotted dry on a filter paper and placed in crimped but vented aluminium sample pans. The sample size was 2-5 mg and the temperature range was typically 303-523 K at a heating rate of 5 K min-1. The samples were purged with a stream of flowing nitrogen (20 ml min-1). The instrument was calibrated using indium as the reference material.
TGA was performed on a TA instruments TGA Q500 thermogravimetric analyzer. Approximately 15 mg of the sample was added to an alumina crucible. The samples were heated over the temperature range 298-573 K at a constant heating rate of 5 K min-1. The samples were purged with a stream of flowing nitrogen at 40 ml min-1 throughout the experiment.
The morphology of the desolvated products was captured using field emission scanning electron microscopy (SEM) at 10.0 kV (Jeol JSM 6700, Japan). Samples were dispersed on adhesive copper tapes and coated with a layer of gold for 70 s using sputter coater (Cressington 208HR, Watford, England).
An analysis of the Cambridge Structural Database (CSD Conquest Version 1.15, November 2012 update) revealed that the crystal structures of the GF solvates with benzene, chloroform, 1,2-dichloromethane, 1,2-dichloroethane, bromochloromethane, bromoethane and dibromomethane have been reported. However, only the GF-1,2-dichloroethane (Refcode: VADGOW; Shirotani et al., 1988) and GF-chloroform (Refcode: MATZEO; Cheng et al., 1979) solvates contain the three-dimensional coordinates. An analysis of the crystal structure of the 1,2-dichloroethane revealed that the solvent molecules are located in the channels formed by the GF molecules along the crystallographic a axis (Shirotani et al., 1988). The crystal structure of the chloroform solvate has also been described (Cheng et al., 1979). It was found that the chloroform molecules in the solvate are arranged in layers perpendicular to the c axis, in which direction the solvent molecules can escape upon desolvation.
A good quality block-shaped crystal of GF-acetonitrile that was obtained from the co-crystallization experiment using GF and gentisic acid was subjected to single-crystal X-ray diffraction. The pertinent crystallographic data are listed in Table 1. The asymmetric unit consists of two molecules each of GF and acetonitrile. As shown in Fig. 3, the symmetry-independent molecules of GF form columns along the crystallographic a axis involving C-HO (HA distance, 2.28-2.52 Å, D-HA angle, 109-173°) interactions from the methyl of the methoxy substituent and carbonyl of the cyclohexenone ring, and a short ClO (oxygen of the methoxy substituent, 3.07 and 3.09 Å) interaction. These columns are interconnected to each other via C-H (2.92 Å, 122°; 2.93 Å, 128°) interactions involving the methyl group on the cyclohexenone ring and the centroid of the phenyl ring of the GF. The overall crystal structure consists of channels along the crystallographic a axis. These channels are filled with the guest acetonitrile molecules that are stabilized via C-HO (2.39-2.44 Å, 140-150°) interactions involving the methyl group of the acetonitrile and carbonyl groups of both the symmetry-independent molecules of GF.
| || Figure 3 |
Crystal structure of the GF-acetonitrile solvate showing channels that are occupied by acetonitrile molecules. Symmetry-independent molecules of GF and acetonitrile are colored differently.
Co-crystallization of GF and saccharin in a 1:1 molar ratio from nitromethane produced block-shaped crystals of nitromethane solvate (GF-nitromethane). The crystal structure was solved in the orthorhombic space group P212121. The crystallographic asymmetric unit consists of one molecule each of GF and nitromethane. The GF molecules interact with each other via several C-HO (2.33-2.48 Å, 116-160°) interactions and generate channels along the crystallographic b axis (Fig. 4). Similar to the crystal structure of GF-acetonitrile solvate, the crystal structure of GF-nitromethane solvate also features a short ClO (3.06 Å) interaction. The nitromethane molecules fill the channels and are stabilized within via C-HO (2.43 Å, 159°; 2.51 Å, 124°) interactions.
| || Figure 4 |
Crystal structure of the GF-nitromethane (1:1) solvate showing the nitromethane molecules located in the channels formed by GF molecules.
The 2:1 GF-nitroethane solvate crystallizes in the orthorhombic space group P212121. Two molecules of GF and one molecule of nitroethane constitute the asymmetric unit. As shown in Fig. 5, GF molecules generate layers in the crystallographic bc plane via several C-HO (2.34-2.45 Å, 127-157°) interactions and a relatively longer ClO (3.19 Å and 3.21 Å) interaction compared with the ClO interactions observed in the other GF solvates discussed in this paper. In contrast to the previous solvates, the overall crystal structure does not contain channels for the guest inclusion. The nitroethane molecules are located between the layers and stabilized via C-HO (2.27 Å, 160°; 2.49 Å, 155°) interactions.
| || Figure 5 |
Crystal structure of the GF-nitroethane (2:1) solvate showing the nitroethane molecules located between the layers formed by GF molecules. Symmetry-independent molecules of GF are colored differently.
The 1:1 GF-nitroethane solvate crystallizes in the monoclinic space group C2. The nitroethane molecule was found to be highly disordered, hence it was incorporated in the model using PLATON/SQUEEZE (Spek, 2009). GF molecules interact with each other via C-HO (2.26-2.45 Å, 115-178°) and ClO (3.14 Å) interactions and generate channels along the crystallographic c axis (Fig. 6). The disordered nitroethane molecules fill the channels (see the supporting information ).
Fig. 7 shows an overlay of different coformers of GF found in its solvates and its parent crystal structure. Selected torsion angles for various conformers are given in Table 3. GF conformers found in the reported crystal structures of chloroform and 1,2-dichloroethane solvates were also compared. A careful analysis of the conformational differences revealed that there is a significant conformational change due to the formation of solvates, and all the conformers found in the solvates show only minor conformational differences arising mainly because of free rotation of the cyclohexenone ring with respect to the benzofuran ring moiety. Interestingly, the conformer found in the parent crystal structure of GF (red) was not observed in any of the solvates. Both the symmetry-independent molecules of GF in GF-acetonitrile adopt a similar conformation which is nearly similar to the conformations found in the crystal structures of GF-nitromethane and GF-nitroethane (1:1). The symmetry-independent molecules of GF in GF-nitroethane (2:1) adopt a similar conformation. Computed conformer energies (Materials Studio, http://www.accelrys.com ), which are tabulated in Table 3, revealed that all the conformers found in the solvates are higher in energy than the conformer in the parent GF crystal structure. Notably, small conformational differences in solvates are very well correlated with minor differences in their conformer energies with a maximum difference between the conformation energy of only 8.4 kJ mol-1 (Table 3).
+The crystal structure did not contain the H atoms, and hence the conformer energy was not calculated.
| || Figure 7 |
An overlay diagram comparing different conformers of GF found in its solvates with the conformer in the parent crystal structure. Color codes: blue and green - GF-acetonitrile; magenta - GF-nitromethane; brown and cyan - GF-nitroethane (2:1); yellow - GF-nitroethane (1:1); orange - GF-chloroform (from Cheng et al., 1979); grey - GF-1,2-dichloroethane (from Shirotani et al., 1988), and red - GF (from Malmros et al., 1977).
Mahieu et al. (2013) recently found two novel polymorphs of GF from melt crystallization, which were mainly characterized by PXRD. Whether or not these polymorphs feature any of the conformers observed in the solvates can only be proved by full structural analysis by single-crystal X-ray diffraction.
Crystals obtained from the crystallization batches were air dried before they were subjected to differential scanning calorimetry (DSC) or thermogravimetric analysis (TGA). As shown in Fig. 8, DSC thermograms of all the solvates show two distinct peaks. Whereas the first endotherm was ascribed to solvent release from the crystal lattice, the second endotherm corresponds to melting of the desolvated material. The onset temperature for the solvent release is in accordance with the crystal packing, such that the solvents which are incorporated in the crystal lattice by strong interactions have higher Tonset values (Table 4).
| || Figure 8 |
DSC profiles of the GF solvates.
The GF:solvent ratio obtained from weight loss measurements in the TGA analysis are in agreement with the crystal structure analysis (Table 4). In the case of 1:1 GF·nitroethane solvate, the TGA analysis supports the 1H NMR analysis (see the supporting information ) and confirms the molar ratio. All the solvates show a clear weight loss step for the solvent release (Fig. 9).
| || Figure 9 |
TGA profiles of the GF solvates.
It has been recently found that the two novel polymorphs of GF [forms (II) and (III)] can only be prepared by crystallization from the melt (Mahieu et al., 2013). It has also been found that any traces of form (I) in the melt lead to the preferential crystallization of form (I). Therefore, isolation of the metastable forms (II) and (III) remains a challenge. All the solvates reported in this paper were desolvated at ambient conditions to find whether the desolvation would help to isolate forms (II) and (III). The resulting solids were analyzed by PXRD. All the PXRD patterns matched well with the PXRD pattern of form (I) (see the supporting information ). Therefore, this confirms the previous observation that form (I) is the stable form at ambient conditions.
Scanning electron microscope (SEM) analysis of the desolvated products revealed that all the crystals retain their morphology even after removal of the solvent (see supporting information ). A closer observation of the crystals revealed several pores on the surface of the desolvated crystals which could be due to the release of the solvent molecules from the crystals.
In general, the stability of solvates is dependent on factors such as temperature, solvent, pressure etc. (Giron, 1995; Wirth & Stephenson, 1997). A solvate may be the most stable form in its own solvent but upon exposure to ambient conditions, the solvate becomes metastable (Samas et al., 2011). Therefore, evaluation of the stability of solvates is important. In the case of GF solvates, thermal analysis (Figs. 8 and 9) suggested that the solvates desolvate in the temperature range 343-383 K. The stability of the solvates when in contact with their respective solvents was evaluated by slurry experiments. Excess amounts of solvate samples were slurried for a day in saturated solutions of the respective solvents. Filtered and air-dried powder samples were characterized by PXRD. PXRD patterns of the powder samples obtained from the slurry experiments are compared with simulated PXRD patterns of the solvates in Fig. 10. This suggests that acetonitrile solvate is unstable in solution and completely converted to GF, and GF-nitroethane (1:1) solvate converted to GF-nitroethane (2:1) solvate (Fig. 10). On the other hand, both the GF-nitromethane and GF-nitroethane (2:1) solvates remained unchanged after 1 d of slurrying at room temperature (Fig. 10).
| || Figure 10 |
Comparison of the simulated PXRD patterns of the solvates with the PXRD pattern of the powders obtained from the slurry experiments: (a) GF form (I), (b) GF-acetonitrile - simulated, (c) GF-nitromethane - simulated, (d) GF-nitroethane (2:1) - simulated, (e) GF-nitroethane (1:1) - simulated, (f) powder from slurry experiment on GF-acetonitrile, (g) powder from slurry experiment on GF-nitromethane, (h) powder from slurry experiment on GF-nitroethane (2:1), (i) powder from slurry experiment on GF-nitroethane (1:1). Notice that only the PXRD patterns of the samples obtained from the slurry experiments on GF-nitromethane and GF-nitroethane (2:1) match their respective simulated PXRD patterns.
Phase transformation of GF-nitroethane (1:1) to GF-nitroethane (2:1) was also observed in crystallization experiments. When the crystallization experiments were conducted from a saturated solution of GF in nitroethane, both the crystals of 2:1 and 1:1 solvates were obtained (Fig. 2). On the contrary, when GF was crystallized from excess nitroethane (diluted solutions) both the crystals appeared in the beginning but all the crystals of 1:1 solvate subsequently transformed to the 2:1 solvate crystals. This suggests that the 2:1 solvate is thermodynamically more stable than the 1:1 solvate at room temperature. Interestingly, while solvent mediated polymorphic phase transformations are well known in the literature (Thirunahari et al., 2009; Maher et al., 2012; Prohens et al., 2012; Zimmermann et al., 2012; Gomez et al., 2012), examples of phase transformation of a solvate to another solvate (of the same solvent) with different stoichiometry are seldom reported. A recent precedent of this kind has been reported by Tanaka et al. (2005): an organic host, 2,3-bis-fluoren-9-ylidene succinic acid, forms a 1:1 solvate (needles) with ethanol which completely transforms to 1:2 solvate (prisms) in the solvent. It is to be noted that this transformation results in a solvate with a higher proportion of guest molecules. However, in the case of GF-nitroethane solvates, the phase transformation results in complete structural reorganization such that part of the solvent is expelled from the crystal lattice to give a solvate with fewer guest molecules than its initial solvate.
GF is a versatile host with the ability to form solvates with aliphatic fatty acids, alkyl halides, nitroalkanes and some other organic small molecules. This ability can be traced to its conformational rigidity and awkward shape. The GF molecule has a benzofuran moiety together with a cyclohexenane ring. In addition, the molecule also has three methoxy substituents, two on the benzofuran ring and one on the cyclohexenone ring. All these geometrical features confer an awkward shape to the GF molecule which prevents it from packing efficiently. In the present study, solid form screening of GF resulted in four novel solvates with acetonitrile, nitromethane and nitroethane (1:1 and 2:1). Crystal structures of all the solids were determined by single-crystal X-ray diffraction. Thermal analysis confirmed that the solvates lose the solvent molecules in the temperature range 343-383 K. Desolvation of the solvates provided form (I) of GF. Solution stability studies revealed that nitromethane and nitroethane (2:1) solvates are stable in their respective solvents, but acetonitrile solvate transforms to GF and GF-nitroethane (1:1) solvate converts into GF-nitroethane (2:1) solvate.
Serendipitous discovery of novel solid forms (polymorphs, solvates, hydrates etc.) from attempted co-crystallization experiments is not uncommon, and recent reports revealed several such occurrences (Nath & Nangia, 2011; Li et al., 2011; Aitipamula, Chow & Tan, 2012; Wenger & Bernstein, 2007). Two of the solvates reported herein were first identified from attempted co-crystallization experiments, suggesting a greater significance of the co-crystallization technique in an attempt to discover not only co-crystals but also novel solvates/hydrates of pharmaceutical relevance.
This work was supported by the Science and Engineering Research Council of A*STAR (Agency for Science, Technology and Research), Singapore. We thank Dr Martin K. Schreyer for helpful discussions.
Aitipamula, S., Chow, P. S. & Tan, R. B. H. (2011). CrystEngComm, 13, 1037-1045.
Aitipamula, S., Chow, P. S. & Tan, R. B. H. (2012). CrystEngComm, 14, 691-699.
Aitipamula, S., Vangala, V. R., Chow, P. S. & Tan, R. B. H. (2012). Cryst. Growth Des. 12, 5858-5863.
Blagden, N., de Matas, M., Gavan, P. T. & York, P. (2007). Adv. Drug Deliv. Rev. 59, 617-630.
Blessing, R. H. (1995). Acta Cryst. A51, 33-38.
Braun, D. E., Gelbrich, T., Kahlenberg, V., Tessadri, R., Wieser, J. & Griesser, U. J. (2009). Cryst. Growth Des. 9, 1054-1065.
Byrn, S. R., Pfeiffer, R. R. & Stowell, J. G. (1999). Solid-State Chemistry of Drugs. West Lafayette, IN: SSCI.
Chan, Y. C. & Friedlander, S. F. (2004). Curr. Opin. Infect. Dis. 17, 97-103.
Cheng, K. C., Shefter, E. & Srikrishnan, T. (1979). Int. J. Pharm. 2, 81-89.
Desikan, S., Parsons, R. L., Davis, W. P., Ward, J. E., Marshall, W. J. & Toma, P. H. (2005). Org. Process Res. Dev. 9, 933-942.
Flack, H. D. (1983). Acta Cryst. A39, 876-881.
Giron, D. (1995). Thermochim. Acta, 248, 1-59.
Gomez, A., Antonio, S. G., de Araujo, G. L. B., Ferreira, F. F. & Paiva-Santos, C. O. (2012). CrystEngComm, 14, 2826-2830.
Grant, D. J. W. & Abougela, I. K. (1981). J. Pharm. Pharmacol. 33, 619-620.
Joshi, V., Stowell, J. G. & Byrn, S. R. (1998). Mol. Cryst. Liq. Cryst. 313, 265-270.
Kasim, N. A., Whitehouse, M., Ramachandran, C., Bermejo, M., Lennernäs, H., Hussain, A. S., Junginger, H. E., Stavchansky, S. A., Midha, K. K., Shah, V. P. & Amidon, G. L. (2004). Mol. Pharm. 1, 85-96.
Kassem, M. A., Esmat, S., Bendas, E. R. & El-Komy, M. H. (2006). Mycoses, 49, 232-235.
Li, J., Bourne, S. A. & Caira, M. R. (2011). Chem. Commun. 47, 1530-1532.
Maher, A., Croker, D. M., Rasmuson, Å. C. & Hodnett, B. K. (2012). Cryst. Growth Des. 12, 6151-6157.
Mahieu, A., Willart, J. F., Dudognon, E., Eddleston, M. D., Jones, W., Danède, F. & Descamps, M. (2013). J. Pharm. Sci. 102, 462-468.
Malmros, G., Wagner, A. & Maron, L. (1977). Cryst. Struct. Commun. 6, 463-470.
Morissette, S. L., Almarsson, Ö., Peterson, M. L., Remenar, J. F., Read, M. J., Lemmo, A. V., Ellis, S., Cima, M. J. & Gardner, C. R. (2004). Adv. Drug Deliv. Rev. 56, 275-300.
Nath, N. K. & Nangia, A. (2011). CrystEngComm, 13, 47-51.
Prohens, R., Portell, A. & Alcobé, X. (2012). Cryst. Growth Des. 12, 4548-4553.
Puttaraja, K. A. N. & Sakegowda, D. S. (1982). J. Crystallogr. Spectrosc. Res. 12, 415-423.
Rigaku (2007). CrystalClear. Rigaku Corporation, Tokyo, Japan.
Samas, B., Seadeek, C., Campeta, A. M. & Chekal, B. P. (2011). J. Pharm. Sci. 100, 186-194.
Sekiguchi, K., Horikoshi, I. & Himuro, I. (1968). Chem. Pharm. Bull. 16, 2495-2502.
Sekiguchi, K., Ito, K., Owada, E. & Ueno, K. (1964). Chem. Pharm. Bull. 12, 1192-1197.
Sekiguchi, K., Shirotani, K., Kanke, M., Furukawa, H. & Iwatsuru, M. (1976). Chem. Pharm. Bull. 24, 1621-1630.
Shan, N. & Zaworotko, M. J. (2008). Drug Discovery Today, 13, 440-446.
Sheldrick, G. M. (2008). Acta Cryst. A64, 112-122.
Shirotani, K., Suzuki, E., Morita, Y. & Sekiguchi, K. (1988). Chem. Pharm. Bull. 36, 4045-4054.
Spek, A. L. (2009). Acta Cryst. D65, 148-155.
Tanaka, K., Wada, S. & Caira, M. R. (2005). CrystEngComm, 7, 592-594.
Thirunahari, S., Aitipamula, S., Chow, P. S. & Tan, R. B. H. (2009). J. Pharm. Sci. 99, 2975-2990.
Wenger, M. & Bernstein, J. (2007). Mol. Pharm. 4, 355-359.
Wirth, D. D. & Stephenson, G. A. (1997). Org. Process Res. Dev. 1, 55-60.
Zimmermann, A., Frøstrup, B. & Bond, A. D. (2012). Cryst. Growth Des. 12, 2961-2968.