Anhydrous Caffeine Hydrochloride and Its Hydration - Crystal Growth

Synopsis. A study of the first reported crystal structure of anhydrous caffeine hydrochloride reveals irreversibly hydration without loss of crystalli...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/crystal

Anhydrous Caffeine Hydrochloride and Its Hydration Gabriela Gurau, Steven P. Kelley, Kristin R. Di Bona, Marcin Smiglak,† and Robin D. Rogers* Center for Green Manufacturing and Department of Chemistry, The University of Alabama, Box 870336, Tuscaloosa, Alabama 35487-0336, United States S Supporting Information *

ABSTRACT: A study of the first reported crystal structure of anhydrous caffeine hydrochloride reveals irreversible hydration without loss of crystallinity and decomposition with the evolution of HCl gas upon heating.

E

Single crystal structures have widespread relevance in the pharmaceutical sciences, not only for confirmation of molecular structure but also as standards for qualitative and quantitative analysis using powder X-ray diffraction (PXRD) in the design and screening of new APIs. Thus, the lack of any structural information for such a widely used salt as caffeine hydrochloride surprised us. We thus set out to use our methods of making liquid salts of APIs, where water must often be rigorously excluded, for the synthesis and isolation of an anhydrous caffeine hydrochloride. We found that anhydrous caffeine hydrochloride was rather easily and quantitatively isolated by stirring caffeine with concentrated HCl overnight followed by evaporation of the volatiles and drying under high vacuum (see Supporting Information (SI)). The proton transfer, and thus the identity of the compound, was confirmed by nuclear magnetic resonance (NMR), infrared (IR), and Raman spectroscopy, as well as elemental analysis. The 1H NMR spectrum in DMSO-d6 shows a broad peak at δ 10.93 ppm, corresponding to the N9−H proton. The slightly downfield shift of the imidazole proton (C8−H) is also a characteristic of salt formation. The IR spectrum (Figure S1, SI) shows similarities to that reported in the literature,24 with the characteristic N−H stretching vibration (νNH) appearing at 2294 cm−1. Single crystals were grown via slow diffusion of diethyl ether into a saturated ethanol solution, and the molecular structure was determined by single crystal X-ray diffraction (Figure 1). The structural results as discussed below imply that this salt can be regarded as the dehydrated form of caffeine hydrochloride dihydrate. The salt crystallizes in the space group C2/c, and the asymmetric unit consists of an essentially planar protonated (N9) caffeine cation and a chloride anion. The bond lengths are

ven after half of a century of investigations, caffeine, a xanthine alkaloid compound used mainly as a formulation or food additive and as a pharmaceutical model compound, still has the ability to surprise, especially when it comes to its chemical and biological properties. The world’s demand for new and more efficient drugs has rapidly increased, and caffeine, among other drugs, has become a widely investigated model compound in pharmaceutical laboratories.1−8 Therefore, it has become increasingly important to improve its performance and shelf life through the improvement of the physicochemical properties such as solubility, dissolution rate, bioavailability, and stability. Today, there are several strategies utilized to screen and design different formulations of active pharmaceutical ingredients (APIs)9−11 with desired improved properties,12 such as the use of salts13,14 or cocrystals.15 Recently, ionic liquids (ILs, defined as salts that melt below 100 °C) have gained the interest of pharmaceutical scientists due to the ability to tune the physical, chemical, and biological properties, thus offering a new strategy designed to overcome problems such as polymorphism, bioavailability, and solubility.16 Our interest in pure liquid salts of APIs17−22 led us to consider caffeine as a potential precursor for ILs. The pKb for caffeine suggests that it could readily form salts with tetrafluoroboric acid, bis(trifluoromethane)sulfonamine, or nitric acid, yet reported attempts to prepare caffeine ILs have been unsuccessful.23 This was (and is) curious and we set out to learn more about the limited salt forms of caffeine. Despite the intense study of caffeine over the years, the number of caffeine salts of any kind reported in the literature is very small,24−29 and their identity as salts is often in question. A thorough search of SciFinder and the Cambridge Structural Database (CSD)30 revealed only one pharmaceutically acceptable24 salt of caffeine, caffeine hydrochloride dihydrate.29 No crystal structure of the anhydrous form of the hydrochloride salt has been reported, even though its preparation has.28 © 2012 American Chemical Society

Received: June 29, 2012 Revised: August 2, 2012 Published: August 3, 2012 4658

dx.doi.org/10.1021/cg300878j | Cryst. Growth Des. 2012, 12, 4658−4662

Crystal Growth & Design

Article

Figure 3. Comparison of anion environments in anhydrous caffeine hydrochloride (left) vs caffeine hydrochloride dihydrate (right). Blue lines indicate intermolecular contacts less than the sum of the van der Waals radii.

Figure 1. 50% probability ORTEP diagram of anhydrous caffeine hydrochloride.

not significantly different from those in other caffeine salts26,29 or cocrystals.1,2 The lack of changes in bond length and the positioning of anions primarily around the imidazolium portion of the molecule suggest the positive charge is concentrated largely on the five-membered ring. Each cation makes short (less than the sum of the van der Waals radii) contacts with five anions and two other cations (Figure 2). The chloride ions interact exclusively with the imidazole portion of the molecule (Figure 3). Four chloride ions accept hydrogen bonds with hydrogen atoms on the ring and methyl groups. One chloride ion makes a contact perpendicular to the ring plane to a carbon atom, which is known to be the most favorable cation−anion interaction in imidazolium cations.31 On the other hand, the pyrimidine portion of the molecule interacts with other cations. There is a cation-π interaction between pyrimidine portion of one cation and the imidazole portion of another cation. The closest approach between stacked rings is between O11 and C4 at 3.312(3) Å, which is greater than the sum of the van der Waals radii but close enough to indicate an inter-ring interaction. The carbonyl groups also make out-of-plane short contacts with carbon atoms on neighboring cations which appear to be dipole− dipole interactions. Because the cation−anion interactions are localized to one part of the molecule and cation−cation interactions are localized to the other, the crystal lattice consists of cation “bilayers” which are separated by layers of anions along the b axis (Figure 4).

Figure 4. Packing down the b axis of anhydrous caffeine HCl (left) and caffeine HCl dihydrate (right). Color coding on axes: red = a; blue = c.

The packing of anhydrous caffeine hydrochloride can be described through various supramolecular synthons. Two caffeine cations and two chloride anions form a centrosymmetric, 10-membered hydrogen bonded ring. This unit, which can be thought of as a dimer of two caffeine hydrochloride formula units, stacks with other dimers along the crystallographic b axis via cation−anion stacking and cation−cation π stacking between the imidazole and pyrimidine rings of the cation to form an infinite ribbon. The dimers also form infinite zigzag chains along the ac diagonal via the short contacts between the carbonyl groups and the neighboring pyrimidine rings. The cation−anion orientation in the dihydrate is similar to that in the anhydrous crystal: a chloride ion is located above the imidazole ring plane in the appropriate position for the major imidazolium cation−anion contact, and the cation−cation interactions also occur to a large extent via the pyrimidine

Figure 2. Comparison of cation environments in anhydrous caffeine hydrochloride (left) vs caffeine hydrochloride dihydrate (right). Blue lines indicate intermolecular contacts less than the sum of the van der Waals radii. 4659

dx.doi.org/10.1021/cg300878j | Cryst. Growth Des. 2012, 12, 4658−4662

Crystal Growth & Design

Article

The powder X-ray diffraction (PXRD) pattern of freshly prepared, anhydrous caffeine hydrochloride matches that simulated from the single crystal data suggesting a homogeneous phase (Figure 7).

Figure 5. (Left) Basic hydrogen-bonded caffeine hydrochloride dimeric unit and (right) dimers forming sheets consisting of stacks along the b axis and zigzag chains along the ac diagonal. Color coding on axes: red = a, green = b, blue = c.

portion of the molecule. The most striking difference between the two structures is that there are no direct contacts between the cation and anion in the dihydrate. The cations donate hydrogen bonds to water molecules, which in turn donate hydrogen bonds to chloride ions, resulting in a two-dimensional hydrogen bonded sheet consisting of a layer of cations, then water molecules and anions, then cations. The orientation of the cations relative to each other is not the same as in the dihydrate, and the cation−cation contacts occur between different atoms in anhydrous caffeine hydrochloride vs the dihydrate. Nonetheless, the nature of the cation−cation contacts appears to be the same  dipole− dipole interactions between the negatively charged carbonyl oxygen atoms and positively charged carbon atoms. The cation−cation bilayer type packing is similar in both compounds. Ultimately it appears that water molecules alter the packing only slightly but greatly increase the number of favorable interactions, especially anion−water−anion hydrogen bonding.

Figure 7. Comparison of PXRD patterns of caffeine HCl dihydrate, anhydrous caffeine HCl, and the simulated PXRD pattern from the single crystal structure of anhydrous caffeine HCl.

Analysis of powders of the title compound exposed to different relative humidities (RH) provided insight into the ready hydration of this compound. Anhydrous caffeine hydrochloride was stored at different RH (43, 75, and 98%), and after less than 24 h, the material showed full conversion to the dihydrate. This rapid hydration, which is similar to the hydration of anhydrous caffeine, may provide a reason for the difficulty in isolating the caffeine hydrochloride salt as a crystalline anhydrous salt. A series of time-resolved PXRD scans were taken in which the PXRD pattern was measured continuously and repeatedly for two days with each spectrum requiring approximately 10 min of collection time (Figure 8). The sample studied had been stored for two years, and some evidence of hydration was already evident from the first PXRD scan. Subsequent PXRD scans revealed that anhydrous caffeine hydrochloride continues to rehydrate rapidly and without loss of crystallinity when exposed to ambient air. The diffraction peaks of the anhydrous form steadily decreased, while those of the hydrate steadily increased with no peaks which do not match one of the two structures. The data suggest a crystal-to-crystal transition from anhydrate to hydrate without any intermediate polymorphs. An analysis of the two crystal structures suggests that the hydration occurs via diffusion of water molecules through the anion layer of the crystal lattice. During hydration, the cations are slightly reoriented to accommodate the change in size. Only the weaker cation−cation interactions are perturbed; the cation−anion (or cation−water for the hydrate) interactions are similar in both species. The cation reorientation results in a change in the length of the c axis, as indicated by a shift in the (0,0,2) PXRD peak. Because there are no pores or channels in the crystal structure, it is most likely that the salt cannot be dehydrated without loss of crystallinity.33 The hydration to the crystalline hydrate and

Figure 6. (Left) Comparison of packing of anhydrous caffeine hydrochloride (green) and the dihydrate (red) showing the difference in cation orientation relative to the anions; (right) relative orientations of the cations perpendicular to the b axis.

Thermogravimetric analysis (TGA, with isocratic heating at 5 °C min−1 under nitrogen atmosphere; Figure S2, SI) shows mass loss occurring in two steps with total decomposition occurring by 263 °C. The decomposition temperatures were determined from both the onset to 5 wt % mass loss and from the onset to complete decomposition. The T5% onset temperature observed at 114 °C was attributed to HCl loss. After the first decomposition step, the TGA curve follows the pattern of that of anhydrous caffeine32 (Figure S1, SI). To confirm the dissociation into the individual components, a sample was heated at 100 °C for 1 h under vacuum. The pH of the trapped gas was highly acidic and had the characteristic smell of HCl gas. 4660

dx.doi.org/10.1021/cg300878j | Cryst. Growth Des. 2012, 12, 4658−4662

Crystal Growth & Design

Article

Figure 8. Time-resolved PXRD scans showing the (0,0,2) peaks for anhydrous caffeine hydrochloride and the dihydrate. Time progression is along the z axis, red to blue.



the ready loss of HCl if conditions are used to dehydrate this material could explain the difficulty in isolating crystalline, anhydrous caffeine hydrochloride and in the possible detection of “extra peaks” arising from any anhydrate present before full hydration. Defects in the crystals induced by grinding might also aid in the absorption of water. The reversible hydration and dehydration of caffeine, for instance, was shown to be accelerated by abrading the single crystal.34 We have synthesized and characterized the anhydrous form of caffeine hydrochloride and found its crystal structure to be very similar to that of the dihydrate analogue. The anhydrous salt undergoes a crystal-to-crystal transition to the dihydrate under ambient conditions but decomposes to evolve HCl gas on heating. The ability of the salt to absorb water appears to be due to the layered structure, which may permit the diffusion of water molecules into the crystal lattice requiring only minor reorientation of the cation to maintain crystallinity. Therefore, detection of crystalline, anhydrous caffeine hydrochloride requires careful preparation and handling. This observation is especially important in light of the prevalence of highthroughput screening methods for crystalline APIs. A new polymorph or pseudopolymorph can be easily overlooked if there is no loss of crystallinity in such a transition.



(1) Bučar, D.-K.; Henry, R. F.; Lou, X.; Borchardt, T. B.; Zhang, G. G. Z. Chem. Commun. 2007, 525−527. (2) Bučar, D.-K.; Henry, R. F.; Lou, X.; Duerst, R. W.; Borchardt, T. B.; MacGillivray, L. R; Zhang, G. G. Z. Mol. Pharmaceutics 2007, 4, 339−346. (3) Trask, A. V.; Streek, J. v. d.; Motherwell, W. D. S; Jones, W. Cryst. Growth Des. 2005, 5, 2233−2241. (4) Karki, S.; Frišcǐ ć, T.; Jones, W.; Motherwell, W. D. S. Mol. Pharmaceutics 2007, 4, 347−354. (5) Trask, A. V.; Motherwell, W. D. S; Jones, W. Cryst. Growth. Des. 2005, 5, 1013−1021. (6) Trask, A. V.; Motherwell, W. D. S.; Jones, W. Chem. Commun. 2004, 890−891. (7) Aitipamula., S.; Chow, P. S.; Tan, R. B. H. CrystEngComm 2012, 14, 2381−2385. (8) Brittain, H. G. Cryst. Growth Des. 2012, 12, 1520−1530. (9) Delori, A.; Frišcǐ ć, T.; Jones, W. CrystEngComm 2012, 14, 2350− 2362. (10) Shan, N.; Toda, F.; Jones, W. Chem. Commun. 2002, 2372− 2373. (11) 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 Adv. Drug. Delivery Rev. 2004, 56, 275−300. (12) Palucki, M.; Higgins, J. D.; Kwong, E.; Templeton, A. C. J. Med. Chem. 2010, 53, 5897−5905. (13) Stahl, P. H.; Wermuth, C. G. in Handbook of Pharmaceutical Salts; Stahl, P. H., Wermuth, C. G., Eds.; Verlay Helvetica Chimica Acta & Wiley-VCH: Zurich, 2008. (14) Serajuddin, A. T. M. Adv. Drug Delivery Rev. 2007, 59, 603−616. (15) Vishweshwar, P.; McMahon, J. A.; Bis, J. A.; Zaworotko, M. J. J. Pharm. Sci. 2006, 95, 499−516. (16) Hough, W. L.; Smiglak, M.; Rodriguez, H.; Swatloski, R. P.; Daly, D. T.; Pernak, J.; Grisel, J. E.; Carliss, R. D.; Soutullo, M. D.; Davis, J. H., Jr.; Rogers, R. D. New J. Chem. 2007, 31, 1429−1436. (17) Stoimenovski, J.; MacFarlane, D. R; Bica, K.; Rogers, R. D. Pharm. Res. 2009, 27, 521−526. (18) Bica, K.; Rogers, R. D. Chem. Commun. 2010, 1215−1217. (19) Bica, K.; Rijksen, C.; Nieuwenhuyzen, M.; Rogers, R. D. Phys. Chem. Chem. Phys. 2010, 12, 2011−2017. (20) Bica, K.; Shamshina, J.; Hough, W. L.; MacFarlane, D. R.; Rogers, R. D. Chem. Commun. 2011, 2267−2269. (21) Rogers, R. D.; Daly, D. T.; Gurau, G.; MacFarlane, D. R.; Turanjanin, J.; Dean, P. M.; Scott, J. L.; Bica, K.; Seddon, K. R. PCT Int. Appl. WO 2010078300 A1 20100708 2010. (22) Rogers, R. D.; Daly, D. T.; Swatloski, R. P.; Hough, W. L.; Davis, J. H., Jr. Smiglak, M.; Pernak, J.; Spear, S. K. PCT/US2006/ 039454; WO2007/044693 A2.

ASSOCIATED CONTENT

S Supporting Information *

Materials and Experimental; IR spectra of anhydrous caffeine hydrochloride and anhydrous caffeine; TGA of anhydrous caffeine hydrochloride and anhydrous caffeine. This material is available free of charge via the Internet at http://pubs.acs.org.



REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +1-205/348-4323. Fax: +1-205/348-0823. Present Address †

IoLiTec, Ionic Liquids Technologies GmbH, Satzstrasse 184, D-74076 Heilbronn, Germany. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was initiated at the Queen’s University at Belfast. 4661

dx.doi.org/10.1021/cg300878j | Cryst. Growth Des. 2012, 12, 4658−4662

Crystal Growth & Design

Article

(23) Kerr, K.; Thomas, M.; Szreder, T.; Engel, R.; Wishart, J. F.; LallRamnarine, S. I. Abstract of Papers, 230th ACS National Meeting, Washington, DC, August 28−September 1, 2005; American Chemical Society: Washington, DC, 2005; CHED-092. (24) (a) Herbstein, F. H.; Kaftory, M.; Kapon, M.; Saenger, W. Z. Kristallogr., Kristallgeom., Kristallphys., Kristallchem. 1981, 154, 11−30. (25) Anton, K.; Beauchamp, A. L. Acta Crystallogr., Sect. A 1984, 40, C301. (26) Wang, W.-D.; Wang, H.; Feng, W.-J.; Jin, Z.-M.; Hu, M.-L. Z. Kristallogr.-New Cryst. Struct. 2006, 221, 375−376. (27) Chandramohan, A.; Gayathri, D.; Velmurugan, D.; Ravikumar, K.; Kanthaswamy, M. A. Acta Crystallogr., Sect. E: Struct. Rep. Online 2007, 63, 2495−2496. (28) Cook, D.; Regnier, Z. R. Can. J. Chem. 1967, 45, 2895−2897. (29) Mercer, A.; Trotter, J. Acta Crystallogr., Sect. B. 1978, 34, 450− 453. (30) Bruno, I. J.; Cole, J. C.; Edgington, P. R.; Kessler, M; Macrae, C. F.; McCabe, P.; Pearson, J.; Taylor, R. Acta Crystallogr., Sect. B 2002, 58, 389−397. (31) Hardacre, C.; Holbrey, J. D.; McMath, S. E. J.; Bowron, D. T.; Soper., A. K. J. Chem. Phys. 2003, 118, 273−278. (32) Bothe, H.; Cammenga., H. K J. Therm. Anal. Calorim. 1979, 16, 267−275. (33) Clarke, H. D.; Arora, K. K.; Bass, H.; Kavuru, P.; Ong, T. T.; Pujari, T.; Wojtas, L.; Zaworotko., M. J. Cryst. Growth Des. 2010, 10, 2152−2167. (34) Byrn, S. R.; Lin, C. T. J. Am. Chem. Soc. 1976, 98, 4004−4005.

4662

dx.doi.org/10.1021/cg300878j | Cryst. Growth Des. 2012, 12, 4658−4662