Polymorphism of a Simple Organic Amide - Crystal Growth & Design


Polymorphism of a Simple Organic Amide - Crystal Growth & Design...

0 downloads 66 Views 513KB Size

CRYSTAL GROWTH & DESIGN

Polymorphism of a Simple Organic Amide Melanie R. Hauser,† Lev Zhakarov,† Kenneth M. Doxsee,*,† and Tonglei Li‡ Department of Chemistry, UniVersity of Oregon, Eugene, Oregon 97403-1253, and Department of Pharmaceutical Sciences, UniVersity of Kentucky, Lexington, Kentucky 40536

2008 VOL. 8, NO. 12 4428–4431

ReceiVed February 15, 2008; ReVised Manuscript ReceiVed August 15, 2008

ABSTRACT: Glycolanilide (2-hydroxy-N-phenylacetamide), a simple amide, may be reproducibly crystallized in either of two polymorphic forms. The crystal and molecular structures of each polymorph have been determined, revealing that the two polymorphs differ in the extent and arrangement of intermolecular hydrogen bonding. The more stable polymorph displays one-dimensional chains linked through hydrogen bonds between the amide carbonyl and the R-hydroxyl group. In the metastable polymorph, analogous chain structures are overlain with cyclic hydrogen-bonded amide dimers, with an additional set of hydrogen bonding contacts joining the chain and cyclic motifs. Introduction The term “polymorphism” has been used in the context of crystallography since at least the early 1820s, when Mitscherlich reported the existence of several distinct structures for complex arsenate/phosphate salts (NaH2PO4 · H2O/NaH2AsO4 · H2O and Na2HPO4 · H2O/Na2HAsO4 · H2O).1 Although the concept of polymorphism appears to be caught in the throes of ambiguity, for example, with regard to dynamic isomers, tautomers, and solvates,2 the most commonly accepted definition was originally formulated by McCrone3 in 1965: “a solid crystalline phase of a given compound resulting from the possibility of at least two different arrangements of the molecules of that compound in the solid state.” According to Lehman,4 enantiotropic polymorphism pertains to two crystal forms that undergo reversible phase changes between one another, while monotropic polymorphism corresponds to two crystal forms where one s a kinetically trapped, metastable form s undergoes an irreversible phase change to the other, the thermodynamically more stable form. Lehman noted that crystallization of organic compounds from solution not infrequently affords the less stable form of a monotropic pair. With the impacts of polymorphism ranging from processing (drug compounding and stability) to biomedical (establishment of therapeutic dosages) to legal (potential loss of patent protection), polymorph screening is now an essential element in pharmaceutical research and development.5-7 Here, we report the reproducible crystallization of two polymorphs of a simple organic amide, glycolanilide (1), a discussion of the crystal structures of these two polymorphs, and the results of thermal and calculational studies establishing their relative energetics.

Experimental Section Preparation of Glycolanilide (1). To a solution of 7.240 g of N-thionylaniline (51.8 mmol, Aldrich) in 90 mL of CH3CN was added 3.97 g of glycolic acid (52.2 mmol, Aldrich). The reaction mixture was stirred overnight at ambient temperature under a nitrogen atmo* To whom correspondence should be addressed. E-mail: [email protected]. Phone: (541) 346-2846. Fax: (541) 346-2023. † University of Oregon. ‡ University of Kentucky.

sphere, and then the solvent was removed on a rotary evaporator. To the resulting yellow oil was added 100 mL of diethyl ether, leading to precipitation of crude 1 as a white powder, 4.436 g (56.6%). 1H NMR (300 MHz, CD3CN): δ 8.679 (s, 1H, NH), 7.653-7.624 (m, 2H, ArH), 7.384-7.332 (m, 2H, ArH), 7.162-7.109 (m, 1H, ArH), 4.075 (d, 2H, CH2), 3.755 (t, 1H, OH). 13C NMR (75.435 MHz, CD3OD): δ 128.68, 124.40, 120.37, 61.78. IR (KBr): 3389 (O-H), 3290 (N-H), 1664 (CdO), 1550 (amide II) cm-1. MS (ESI): m/z 173.9 ([M + Na]+), 79. Crystallization of Glycolanilide (1). Compound 1 was crystallized from acetonitrile/xylenes (mixed isomers) by slow evaporation of homogeneous solutions at varying concentrations and three distinct temperatures. In a representative crystallization, to 0.67 mL of a 0.15 M solution of 1 in acetonitrile was added 2 mL of xylenes. The resulting homogeneous solution was placed in an uncapped 12 × 75 mm test tube and allowed to stand undisturbed at 20 °C, leading to the formation of thin white needles. Analogous crystallizations were carried out with initial concentrations of 1 ranging from 0.15 M to 1.1 M and at temperatures of +20 °C, 4 °C, and -25 °C, affording either needles (1-II) or blocks (1-I, see Discussion). Structure Determination. X-ray diffraction data for 1-I and 1-II were collected on a Bruker Smart Apex diffractometer at 173(2) K using Mo KR radiation (λ ) 0.71070 Å). Absorption corrections were applied by SADABS.8 Structures were solved using direct methods completed by subsequent difference Fourier syntheses, and refined by full matrix least-squares procedures on F2. All non-H atoms were refined with anisotropic thermal parameters. The H atoms were found on the residual density maps and refined with isotropic thermal parameters. The crystallographic data and some details of data collection, solution, and refinement of the crystal structure are given in Table 1. All calculations were performed using SHELXTL 6.10 package.9 Differential Scanning Calorimetry. Samples of 1-I (11.719 mg) and 1-II (5.063 mg) were weighed into aluminum pans, onto which were crimped aluminum lids. The samples were analyzed on a 2920 MDSC V2.6A, first increasing the temperature from -20 °C to +150 at 1 °C min-1, then reducing the temperature at the same rate. Energy Calculations. Crystal 03 was used for the optimization and energy calculations of crystals and single molecules.10 Lattice constants were kept constant during the optimization of the crystal structures. The basis set superposition error (BSSE) was corrected using the counterpoise method.11 The basis set used was B3LYP/6-21G**. No diffusion function was considered due to the periodicity of Bloch functions that were used to construct the local functions for the DFT calculations. The energy convergence of the optimizations and energy calculations was set to 10-7 Hartree. Root-mean-squares (rms) were set to 0.0003 and 0.0012 atomic units for energy gradient and atomic displacement, respectively. All calculations were conducted on a Linux cluster. For the purposes of these calculations, polymorph 1-I, which

10.1021/cg800171h CCC: $40.75  2008 American Chemical Society Published on Web 10/02/2008

Polymorphism of a Simple Organic Amide

Crystal Growth & Design, Vol. 8, No. 12, 2008 4429

Table 1. Crystallographic Data and Summary of the Data Collection and Structure Refinement 1-I formula C8H9NO2 fw 151.16 cryst syst orthorhombic space group Pbca T, K 153(2) a, Å 14.167(3) b, Å 5.6343(13) c, Å 18.694(5) R, deg 90 β, deg 90 γ, deg 90 V, Å3 1492.1(6) Z, Z′ 8, 1 cryst dimens, mm 0.24 × 0.18 × 0.10 dcalcd, g · cm-3 1.346 µ, mm-1 0.098 56.38 2θ max, deg Tmin/Tmax 0.604 9127 [0.0280] N measd [Rint] N ind 1780 N obs (I g 2(I)) 1538 no. of params 136 R (I g 2σ(I)); Rw (I g 2σ(I)) 0.0389; 0.1035 R (all); Rw (all) 0.0452; 0.1086 GOF 1.028

1-II C8H9NO2 151.16 monoclinic P21/n 153(2) 5.7611(3) 12.2633(6) 10.3091(5) 90 94.4750(10) 90 726.12(6) 4, 1 0.30 × 0.19 × 0.16 1.383 0.100 56.54 0.857 6150 [0.0218] 1716 1598 136 0.0377; 0.1007 0.0401; 0.1039 1.019

crystallized in the space group P21/n, was converted to P21/c, since Crystal 03 supports the latter alternative space group.

Results and Discussion Glycolanilide (1, 2-hydroxy-N-phenylacetamide) was originally prepared through the direct reaction between glycolic acid and aniline by Iwamoto and Farson,12 who noted that crystallization from water gave concentration-dependent changes in crystal form. Each crystal form was reported to melt sharply at 92 °C, both alone and in mixtures with other crystal forms, leading Iwamoto and Farson to conclude that these crystal forms represented external morphological differences rather than polymorphism. In the course of our studies of the chelating properties of R-hydroxy carbonyl compounds,13 we prepared glycolanilide using the method reported by Shin and Kim,14 via the rather unusual reaction between N-thionylaniline (PhsNdSdO) and glycolic acid. Slow evaporation of acetonitrile/xylene solutions of glycolanilide reproducibly afforded either blocky prismatic or needle-like crystals (Figure 1), or mixtures of the two forms, with the particular form obtained appearing to be dependent on both the initial concentration of the solution and on the temperature at which crystallization was carried out. At the lowest concentrations (0.15 M), needles formed regardless of temperature (ranging from -25 to +20 °C), and at the highest concentrations (1.1 M), exclusively blocks were obtained. At intermediate concentrations (0.45 - 0.6 M), mixtures of both needles and blocks were obtained, with proportionately greater quantities of blocks at lower temperatures. Given these observations, it appeared plausible that polymorphism, not simple morphological variation, was at the root of the changes in crystal form, both in our hands and as noted by Iwamoto and Farson. In order to establish polymorphism, we carried out single crystal X-ray diffraction analysis of both crystal forms, as well as a variety of thermal analyses. Single crystal X-ray diffraction analysis confirmed the existence of polymorphism. Table 1 presents a summary of the crystallographic data for the two crystal forms, and Figure 2 illustrates their structures. In the polymorph forming the blocky

crystals (1-I), all the atoms within the glycolanilide molecule are essentially coplanar with the exception of the two hydrogens of the methylene group and the hydrogen of the hydroxyl group, which is twisted out of the plane by ca. 76° (103.9°). In the needle-shaped polymorph (1-II), the structure is very similar, with a relatively small but notable change in the C-C-O-H dihedral angle (114.1°) and a slight twist of the amide bond (O-C-N-H dihedral angle of 173.61°) relative to that seen in 1-I (177.23°). Although these differences in molecular structural parameters are really rather subtle, they allow for, or result from, a dramatic change in the hydrogen bonding networks for the two polymorphs. Polymorph 1-I presents a simple O-H · · · OdC hydrogen bonded chain structure (Figure 3a), representing a unitary C(5) chain in the graph set nomenclature pioneered by Etter for the description of crystalline networks.15,16 Interestingly, the amide N-H bonds are not involved in hydrogen bonding in this polymorph, despite the frequency with which such hydrogen bonding interactions are observed in other amide-containing systems.17 Polymorph 1-II, in marked contrast, does incorporate N-H as well as O-H hydrogen binding interactions in its packing, resulting in a significantly more complicated secondary bonding motif. This hydrogen bonding pattern may be “parsed” into two simpler unitary motifs, with C(5) chains completely analogous to those seen in 1-I overlain with hydrogen-bonded R22(10) dimers (formed through N-H · · · OH interactions, Figure 3b). These two unitary motifs converge on one another through the formation of additional R44(12) cyclic hydrogen bonding interactions (Figure 4). In order to establish the relative stabilities of the two polymorphs experimentally, physical changes in crystal morphology were observed during slow heating from room temperature to the melting point (92.5-94 °C). Samples of each polymorph were heated side-by-side in melting point capillaries at ca. 2-3 °C min-1 in a heated block melting point apparatus. At 50 °C, thickening of the needles of 1-II was evident, while 1-I appeared unchanged. After reaching 60 °C, both tubes were removed from the heat source and examined under an optical microscope, revealing the unchanged morphology of 1-I and the presence of small blocks as well as the thickened needles of 1-II. The samples were returned to the melting point apparatus, and heating was continued. At 70 °C, the samples were again removed from the heat source and examined under an optical microscope. Distinct blocks, including one large blocky crystal, as well as some remaining thickened needles, were observed for 1-II, while 1-I, again, remained unchanged. The samples were again returned to the melting point apparatus, and heating was continued until both samples melted rather sharply at 92.5-94 °C (reported mp 92-94 °C). This simple set of observations argues strongly for the metastable nature of polymorph 1-II. Interestingly, the transformation from 1-II to the more stable polymorph, 1-I, is kinetically very facile, suggesting that the morphological variations noted in 1946 by Iwamoto and Farson12 may in fact have been due to polymorphism rather than simple habit modification despite their observation of identical melting points for the various crystal forms. More quantitative evidence for the metastability of 1-II was obtained by differential scanning calorimetry (DSC). Both 1-I and 1-II display a melting endotherm at 93 °C. Whereas the DSC trace for 1-I is otherwise featureless, 1-II displays an exotherm with an onset of ca. 55 °C and a peak at 63 °C, consistent with a phase transition to the more stable polymorph. Interestingly, on completion of the DSC analysis, incomplete

4430 Crystal Growth & Design, Vol. 8, No. 12, 2008

Hauser et al.

Figure 1. Photomicrographs of (a) blocks of polymorph 1-I and (b) needles of polymorph 1-II.

Figure 2. Structure of glycolanilide polymorphs (a) 1-I and (b) 1-II.

to 80 °C. Mounting the sample in a high-temperature epoxy minimizes these shifts, which appear to result from minor changes in position of the relatively large crystallites used for the powder determination. In contrast, the powder diffraction pattern of polymorph 1-II changes dramatically between 50 and 60 °C, nicely corresponding with the exotherm observed in the DSC trace for this polymorph, with only minor additional changes in peak intensity observed upon continued heating to 80 °C.

Figure 3. Hydrogen bonding motifs: (a) C(5) chains in 1-I and 1-II; (b) R22(10) dimers in 1-II.

Figure 4. Overall hydrogen bonding in polymorph 1-II.

sealing of the sample pans allowed the slightly volatile glycolanilide to escape. Its deposition on cooler edges of the pans in the form of small needles (1-II) provides one additional suggestive demonstration of the relative stabilities of the two polymorphs, with kinetic trapping of the metastable phase under the conditions of sublimation.18 Polymorph 1-I displays only minor shifts in X-ray powder diffraction peak intensity as the temperature is raised from 40

Given the hydrogen bonding motifs of the two polymorphs, it is tempting to assign at least a component of their relative stabilities to entropic terms, with the more rigidly defined cyclic R22(10) dimers and additional R44(12) cyclic hydrogen bonding interactions present in 1-II being less favorable entropically than the looser and more extended C(5) chain structures present in both polymorphs. Unfortunately, attempted Kissinger analysis19 was unsuccessful, with the exotherm seen at 63 °C shifting inseparably into the melting endotherm under higher scan rates. The lattice energy of the crystals was evaluated by application of Crystal 03, using periodic density functional theory (DFT) methods augmented by long-range van der Waals energies that were computed empirically by three damped, analytical energy models based on interatomic distances and predefined parameters.20-23 Our previous analyses of dozens of organic crystals indicated that this approach is capable of producing lattice energies that are in satisfactory agreement with experimental values.24,25 For crystals involving intermolecular hydrogen bonding, the lattice energy produced by using the medium damping function replicates the experimental value more closely than the other two damping functions. Using this approach, polymorphs 1-I and 1-II were calculated to have energies of -106.43 kJ mol-1 and -101.11 kJ mol-1, respectively,

Polymorphism of a Simple Organic Amide

consistent with the greater stability of 1-I suggested by experimental analyses (vide supra).

Crystal Growth & Design, Vol. 8, No. 12, 2008 4431

(3)

Summary Glycolanilide, a simple amide, may be reproducibly crystallized in either of two polymorphic forms, differing in the extent and arrangement of intermolecular hydrogen bonding. The thermodynamically more stable polymorph (as determined by DSC and calculational analysis) displays onedimensional chains linked through hydrogen bonds between the amide carbonyl and the R-hydroxyl group; the amide NH group does not participate in hydrogen bonding. In the metastable polymorph, analogous chain structures are overlain with more conventional cyclic hydrogen-bonded amide dimers, with an additional set of hydrogen bonding contacts joining the chain and cyclic motifs. Given their molecular and structural simplicity, both calculational analyses and chemical intuition appear to lead to plausible explanations for the relative stabilities of the two polymorphs. Acknowledgment. We are grateful to Clay Mortensen for his assistance with X-ray powder diffraction analyses. We express our appreciation to the National Science Foundation and the U.S. Department of Education (Graduate Assistance in Areas of National Need Program) for their support of this work. Supporting Information Available: Crystallographic data files (.cif format) and differential scanning calorimetry (DSC) traces for polymorphs 1-I and 1-II. This information is available free of charge via the Internet at http://pubs.acs.org.

References (1) Mitscherlich, E. Abhl. Akad. Berlin 1822-23, 43–48. (2) Animated discussions at the 2007 annual meeting of the American Crystallographic Association, Salt Lake City, Utah, pointed rather

(4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17)

(18) (19) (20) (21) (22) (23) (24) (25)

clearly to the lack of consensus regarding a universally accepted functional description of polymorphism. McCrone, W. C. In Physics and Chemistry of the Organic Solid State; Fox, D.; Labes, M. M.; Weissberger, A. Eds.; Wiley Interscience: New York, 1965; Vol. 2, pp 725-767. Lehmann, O. Die Krystallanalyse; Wilhelm Engelmann: Leipzig, 1891. For a representative discussion, see Knapman, K. Modern Drug DiscoVery 2000, 3, 5354, 57. See, for example, Hilfiker, R. Ed. Polymorphism in the Pharmaceutical Industry; Wiley-VCH: New York, 2006. Snider, D. A.; Addicks, W.; Owens, W. AdV. Drug DeliVery ReV. 2004, 56, 391–395. Sheldrick, G. M. SADABS (2.01), Bruker/Siemens Area Detector Absorption Correction Program; Bruker AXS: Madison, Wisconsin, USA, 1998. Sheldrick, G. Bruker XRD: Madison, Wisconsin, USA. Dovesi, R.; Orlando, R.; Civalleri, B.; Roetti, C.; Saunders, V. R.; Zicovich-Wilson, C. M. Z. Kristallogr. 2005, 220, 571–573. Boys, S. F.; Bernardi, F. Mol. Phys. 1970, 19, 553–566. Iwamoto, H. K.; Farson, D. J. Am. Pharm. Assoc. 1946, 35, 50–52. Doxsee, K. M.; Ferguson, C. M.; Wash, P. L.; Saulsbery, R. L.; Hope, H. J. Org. Chem. 1993, 58, 7557–7561. Shin, J. M.; Kim, Y. H. Tetrahedron Lett. 1986, 27, 1921–1924. Etter, M. C.; MacDonald, J. C.; Bernstein, J. Acta Crystallogr. 1990, B46, 256–262. Etter, M. C. Acc. Chem. Res. 1990, 23, 120–126. See, for example, Palmore, G. T. R.; MacDonald, J. C. In The Amide Linkage: Structural Significance in Chemistry, Biochemistry, and Materials Science; Greenberg, A.; Breneman, C. M.; Liebman, J. F. Eds.; Wiley-Interscience: Hoboken, NJ, 2003. See, e.g. Sarma, B.; Roy, S.; Nangia, A. Chem. Commun. 2006, 491, 8–4920. Kissinger, H. E. Anal. Chem. 1957, 29, 1702–1706. Ahlrichs, R.; Penco, R.; Scoles, G. Chem. Phys. 1977, 19, 119–130. Aziz, R. A.; Chen, H. H. J. Chem. Phys. 1977, 67, 5719–5726. Hepburn, J.; Scoles, G.; Penco, R. Chem. Phys. Lett. 1975, 36, 451– 456. Wu, Q.; Yang, W. T. J. Chem. Phys. 2002, 116, 515–524. Li, T.; Feng, S. Pharm. Res. 2006, 23, 2326–2332. Feng, S.; Li, T. J. Chem. Theory Comput. 2006, 2, 149–156.

CG800171H