ARTICLE pubs.acs.org/crystal
Ionic Co-crystals of Organic Molecules with Metal Halides: A New Prospect in the Solid Formulation of Active Pharmaceutical Ingredients Dario Braga, Fabrizia Grepioni,* Giulio I. Lampronti, Lucia Maini, and Alessandro Turrina Dipartimento di Chimica “G. Ciamician”, Universita degli Studi di Bologna, via Selmi 2 - 40126 Bologna, Italy
bS Supporting Information ABSTRACT: Ionic co-crystals (ICCs) are constituted of inorganic salts and organic molecules. These crystals combine ionic interactions with classical supramolecular bonding such as hydrogen and dipole-bonding interactions. Such is the case of the hydrogen bonds between NH and CdO dipoles in primary and secondary amides when these cocrystallize with an ionic material such as CaCl2. Here, we report our results of the preparation and characterization of a series of ICCs in which the molecular component is an active pharmaceutical ingredient (API) or a precursor of APIs, namely, barbituric acid, diacetamide, malonamide, nicotinamide, and piracetam, while the inorganic salt coformer is CaCl2. CaCl2 has been chosen, inter alia, for its nontoxicity and potential applications in the pharmaceutical field. Preparative methods include conventional crystallizations from solution, as well as slurry and solid state techniques (grinding and kneading). All crystal structures reported herein were determined either from single crystal diffraction data or from powder diffraction data, using simulated annealing procedures. Crystalline products were analyzed by differential scanning calorimetry, thermogravimetric analysis, and variable temperature X-ray powder diffraction. Intrinsic dissolution rate measurements were also performed on nicotinamide and piracetam ICCs.
’ INTRODUCTION Crystal engineers aim to make crystals by design, that is, to assemble molecular or ionic components into a crystalline supramolecular network starting from a knowledge, hence a predictive capacity, of the way molecules—the “building blocks”— will cling to each other to construct a stable crystalline edifice. In the case of molecular crystals, ionic crystals containing molecular ions, and coordination networks, the interactions between building blocks are mainly of noncovalent nature, for example, coordination bonds between ligands and metal centers, Coulombic attractions and repulsions between ions, interactions of the van der Waals type, for example, π-stacking, interactions between halogen atoms as well as hydrogen bonds in the whole range of strength from weak interactions of the CH 3 3 3 O type to strong “charge assisted” bonds between ions, and, of course, combinations of these linkages.1 There is a strong current interest in the application of crystalengineering approaches to the preparation of molecular cocrystals, viz. of stable aggregates of two or more molecular components that form stable crystalline materials on their own at ambient conditions.2 It is, however, difficult to define a precise difference between co-crystals and other multicomponent aggregates such as solvates, molecular salts (e.g., when proton transfer along a hydrogen bond is involved), and molecular complexes.3 Cocrystals are important because of their possible applications in the pharmaceutical field and generally in any field where the final r 2011 American Chemical Society
product is used and commercialized as a solid phase.2 Moreover, cocrystals often show physical and chemical properties (solubility and instrinsic dissolution rate, melting point, color, etc.) that are usually different from those of their separate components.3 At the same time, the cocrystallization process does not affect the chemical integrity of the molecular compounds—this is particularly important in the case of active pharmaceutical ingredients (APIs). At the academic level, the research on molecular cocrystals yields new information on molecular recognition, assembly, and packing. Recently, we have reported on a new class of ionic cocrystals (ICCs) formed by an organic molecule and an inorganic salt,4a like an alkali or alkaline earth halide: In these compounds, the organic molecule, which is solid as a pure compound at ambient conditions, acts as a sort of solvating molecule toward the ions.4 As a matter of fact, ICCs are often hydrated, and water molecules compete with the organic component for ions coordination. ICCs can be obtained either by classic crystallization methods, i.e., by mixing components in solution in adequate stoichiometric ratios, or by solid-state reactions between solid components (grinding or kneading) with no or limited solvent involvement. It should be stressed that it is not yet possible to predict whether a certain organic molecule will form an ICC with a given Received: September 8, 2011 Revised: October 14, 2011 Published: October 18, 2011 5621
dx.doi.org/10.1021/cg201177p | Cryst. Growth Des. 2011, 11, 5621–5627
Crystal Growth & Design
ARTICLE
Table 1. ICCs Obtained as Single Crystals or Powders from Solution and Solid-State Processes ICC formula malonamide 3 CaCl2 3 2H2O piracetam2 3 CaCl2 3 2H2O nicotinamide 3 CaCl2 3 H2O
synthesis technique
structural solution
solution/kneading
single crystal
kneading kneading
powder powder
nicotinamide2 3 CaCl2 3 2H2O
solution
single crystal
diacetamide 3 CaCl2 3 5H2O
solution/kneading
single crystal
barbituric acid 3 CaCl2 3 5H2O piracetam2 3 SrCl2 3 2H2O
solution/kneading
single crystal
solution
single crystal
Figure 1. Organic components used in the preparation of ICCs with CaCl2.
inorganic salt or crystallize out as separate materials. However, a knowledge of preferred hydrogen-bonding interactions of the organic moiety allows some insight. For example, in primary and secondary amides, it is known that NH and CdO groups will act, respectively, as donor and acceptor groups for hydrogen bonds but also as “solvating” groups toward negative and positive ions, respectively. Thus, it is reasonable to consider these molecules as suitable candidates for the design of new ICCs. It has in fact been shown that urea and barbituric acid form ICCs where the CdO and NH dipoles coordinate metal cations and halide anions, respectively.4,5 An a priori knowledge of the metal coordination patterns would also be of great advantage. 6 Unfortunately, the coordination number and geometry of alkaline and alkali earth metals are not easily predictable, and this fact adds problems when the structures need to be determined from powder diffraction data alone, as for some of the cases discussed herein. Crystal structures of the urea4 3 CaCl 2 and oxamide 3 CaCl 2 3 2H 2O ICCs have also been recently solved from powder diffraction data and reported in the context of a paper describing a new crystallographic approach to structure solution from powder diffraction data.4b In the present work, we explore the “solvating properties” toward the inorganic salt CaCl2 of a number of amides that are relevant to the pharmaceutical field, either as APIs or as precursors for the synthesis of drugs, namely, barbituric acid, urea, oxamide, diacetamide, malonamide, nicotinamide, and piracetam (Figure 1). CaCl2 was chosen as the inorganic counterpart also because of its nontoxicity and potential applications in the pharmaceutical field. The new ICCs discussed hereafter are reported in Table 1. Crystal structures were determined either from single crystal data or from powder diffraction data. Crystalline products were analyzed with differential scanning calorimetry (DSC),
thermogravimetric analysis (TGA), and variable-temperature X-ray powder diffraction (XRPD). Intrinsic dissolution rate (IDR) measurements were performed on ICCs containing the APIs nicotinamide and piracetam.
’ EXPERIMENTAL DETAILS AND STRUCTURE DETERMINATION FROM XRPD DATA All reagents and solvents were purchased from Sigma-Aldrich and used without further purification. Synthesis in Solution. CaCl2 (0.1 mmol) and 0.1 mmol of the organic precursor were dissolved in 20 mL of absolute ethanol; the solution was left to evaporate at room temperature, yielding crystalline barbituric acid 3 CaCl2 3 5H2O, diacetamide 3 CaCl2 3 5 H2O, malonamide 3 CaCl2 3 2H2O, and nicotinamide2 3 CaCl2 3 2 H2O. Solid piracetam2 3 SrCl2 3 2H2O was obtained by dissolving piracetam (0.2 mmol) and SrCl2 (0.1 mmol) in 20 mL of absolute ethanol; the solution was left to evaporate at room temperature. Solid State Synthesis. Barbituric acid 3 CaCl2 3 5H2O, diacetamide 3 CaCl2 3 5H2O and malonamide 3 CaCl2 3 2H2O were obtained by manually kneading barbituric acid (1 mmol), urea (4 mmol), oxamide (1 mmol), diacetamide (1 mmol), or malonamide (1 mmol) with CaCl2 (1 mmol) in an agate mortar for 20 min with a drop of ethanol; all reactions were quantitative. Nicotinamide 3 CaCl2 3 H2O was obtained by ball milling (using a Retsch MM 200 Mixer Mill) nicotinamide (1 mmol) and CaCl2 (1 mmol) for 120 min with a drop of ethanol; the reaction was quantitative. Piracetam2 3 CaCl2 3 2H2O was obtained by ball milling piracetam (2 mmol) and CaCl2 (1 mmol) for 120 min with a drop of water; the reaction was quantitative. TGA. TGA measurements were performed using a PerkinElmer TGA7 in the temperature range 30400 °C under an N2 gas flow, at a heating rate of 5 °C min1. DSC. DSC measurements were performed with a PerkinElmer Diamond. Samples (35 mg) were placed in open aluminum pans. Heating was carried out at 5 °C min1 for all cocrystals, in the temperature range 30160 °C. IDR. Measurements were carried out with a Varian Cary 50 Spectrophotometer equipped with a fiber optic dip probe. Five standard solutions in physiological solution (0.1 M NaCl) at concentrations of 6.25, 12.50, 25.00, 50.00, and 100.00 mg L1 were used to calculate a calibration curve for both piracetam and nicotinamide (correlation coefficients were 0.99476 and 0.99839, respectively). We analyzed the dissolution rate in water at room temperature of piracetam (commercial, form III),8 nicotinamide (commercial reagent, form I),9 piracetam2 3 CaCl2 3 2H2O, and nicotinamide 3 CaCl2 3 H2O. We measured the absorbance and used the linear part of the spectrum between 0.1 and 0.5 min, its slope corresponding to the dissolution rate in that interval of time, expressed in Abs min1. The Abs min1 values were then interpolated in the calibration curve to find the dissolution rate of the analytes expressed as mg L1 min1. X-Ray Single-Crystal Diffraction Experiments. Singlecrystal X-ray diffraction data were collected at room temperature with an Oxford Diffraction Xcalibur diffractometer equipped with a CCD detector. Mo Kα radiation (λ = 0.71073 Å) was used. SHELX977a was used for structure solution and refinement. Nonhydrogen atoms were refined anisotropically. Hydrogen atoms bound to carbon atoms were added in calculated positions. Hydrogen atoms bound to nitrogen and oxygen atoms were located from a Fourier map, and their positions were 5622
dx.doi.org/10.1021/cg201177p |Cryst. Growth Des. 2011, 11, 5621–5627
5623
a
7.02
7.21
2
502.027(8)
107.858(2)
98.564(1)
103.036(2)
9.2781(1) 6.1967(1)
Powder data.
7.02 1.031
0.0715
0.0318
1.05
0.022
5365/2364
0.931 2.8827.33
1.697
4
1288.3(2)
90
130.34(1)
90
21.060(2) 8.9583(7)
triclinic 9.6884(2)
RF2 χ2
0.1497
0.0551
1.059
0.019
3488/1881
0.816 3.2329.3
1.388
2
723.12(8)
90
102.815(7)
90
7.2632(5) 12.9379(8)
monoclinic 8.9583(7)
7.21
0.0682
wR2 (all)
monoclinic 7.8917(5)
P1
243.06
C6CaCl2N2O2
nicotinamide 3 CaCl2 3 H2Oa
Rwp
0.845
0.0251
R1 (obsd)
Rint
GoF
2625/1715
0.011
reflns collected/unique
1.628
2
Z
1.124 3.5027.50
508.32(5)
V (Å3)
μ (mm1) θ range (°)
85.297(5)
γ (deg)
dcalcd (mg cm3)
71.208(5)
80.134(5)
b (Å) c (Å)
β (deg)
8.0258(5) 11.5035(6)
a (Å)
α (deg)
triclinic
5.9055(4)
space group
P 21/c
P 21
P1
crystal system
C4H14CaCl2N2O8 329.15
C4H17CaCl2NO7 302.17
C3H10CaCl2N2O4
294.11
barbituric acid 3 CaCl2 3 5H2O
Mr
diacetamide 3 CaCl2 3 5H2O
chemical formula
malonamide 3 CaCl2 3 2H2O
0.088
0.0384
1.093
0.026
3759/1561
0.709 3.1529.21
1.534
4
1694.1(2)
90
117.775(7)
90
6.872(3) 12.8167(7)
21.740(1)
monoclinic
C2/c
391.27
C12H16CaCl2N4O4
nicotinamide2 3 CaCl2 3 2H2O
0.0750
0.0377
0.933
0.028
3524/1506
2.985 3.1428.79
1.584
2
1003.84(9)
90
106.520(7)
90
9.8519(4) 11.8823(7)
8.9444(5)
monoclinic
P 21/n
478.87
C12H24SrCl2N4O6
piracetam2 3 SrCl2 3 2H2O
9.54 2.202
9.11
2
969.31(2)
90
108.0589(6)
90
9.6092(1) 11.9274(1)
8.89548(8)
monoclinic
P 21/n
407.14
C12CaCl2N4O6
piracetam2 3 CaCl2 3 H2Oa
Table 2. Crystallographic Details for Malonamide 3 CaCl2 3 2H2O, Diacetamide 3 CaCl2 3 5H2O, Barbituric Acid 3 CaCl2 3 5H2O, Nicotinamide 3 CaCl2 3 H2O, Nicotinamide2 3 CaCl2 3 2H2O, Piracetam2 3 SrCl2 3 2H2O, and Piracetam2 3 CaCl2 3 2H2O
Crystal Growth & Design ARTICLE
dx.doi.org/10.1021/cg201177p |Cryst. Growth Des. 2011, 11, 5621–5627
Crystal Growth & Design
ARTICLE
Figure 2. Experimental (red crosses), calculated (green curve), and difference (purple curve) powder patterns for piracetam2 3 CaCl2 3 2 H2O; the x-axis is in degrees of 2θ. Peak positions are marked in black.
Figure 3. Experimental (red crosses), calculated (green curve), and difference (purple curve) powder patterns for nicotinamide 3 CaCl2 3 H2O; the x-axis is in degrees of 2θ. Peak positions are marked in black.
refined. PLATON7b and SCHAKAL997c were used for hydrogenbonding analysis and molecular graphics, respectively. Relevant crystallographic details are listed in Table 2. XRPD Experiments. For structure solution and refinement purposes, X-ray powder diffractograms in the 2θ range 570° (step size, 0.01°; time/step, 50 s; 0.02 rad soller; VxA 40 40) were collected on a Panalytical X'Pert PRO automated diffractometer equipped with an X'Celerator detector. For phase identification purposes, X-ray powder diffractograms in the 2θ range 540° (step size, 0.02°; time/step, 20 s; 0.04 rad soller; VxA 40 40) were collected on a Panalytical X’Pert PRO automated diffractometer equipped with an X'Celerator detector. All data were collected in BraggBrentano geometry, using Cu Kα radiation without a monochromator. Variable Temperature X-Ray Diffraction. X-ray powder diffractograms in the 2θ range 550° were collected on a Panalytical X'Pert PRO automated diffractometer equipped with an X'Celerator detector and an Anton Paar TTK 450 system for measurements at controlled temperature. The data were collected in open air in BraggBrentano geometry using Cu Kα radiation without a monochromator. Structure Determination of Piracetam2 3 CaCl2 3 2H2O. The crystal structure of piracetam2 3 CaCl2 3 2H2O could not be solved directly from powder diffraction data by simulated annealing, due to strong preferential orientation effects. The problem was circumvented by preparing and growing single crystals of piracetam2 3 SrCl2 3 2H2O, which turned out to be isostructural with piracetam2 3 CaCl2 3 2H2O. The single-crystal structure determination of the strontium ICC allowed both solution by analogy and refinement by Rietveld analysis of the calcium compound. Rietveld refinement was performed with the software GSAS.10 A shifted Chebyshev with 8 coefficients and a Pseudo-Voigt function (type 4) were used to fit background and peak shape, respectively. A spherical harmonics model was used to describe preferred orientation. Soft constraints were applied on bond
Figure 4. Packing patterns for (a) malonamide 3 CaCl2 3 2H2O, (b) piracetam2 3 CaCl2 3 2H2O, (c) nicotinamide 3 CaCl2 3 H2O, (d) nicotinamide2 3 CaCl2 3 2H2O, (e) diacetamide 3 CaCl2 3 5H2O, and (f) barbituric acid 3 CaCl2 3 5H2O. Color codes: orange, calcium cations and coordination polyhedra; green, chloride ions; gray, carbon; blue, nitrogen; and red, oxygen. Hydrogen atoms not shown for clarity. The software TOPOS was used for all graphical representations.13
distances and angles of both molecules. One overall thermal parameter was adopted for piracetam heterocycle and another one for piracetam amide group. Refinement converged with χ2 = 2.202, Rwp = 9.11, and RF2 = 9.54. Figure 2 shows experimental, calculated, and difference curves. Structure Determination of Nicotinamide 3 CaCl2 3 H2O. Powder diffraction data were analyzed with the software EXPO2010, the updated version of EXPO2009,11 which is designed to analyze either monochromatic and nonmonochromatic data. Peaks were automatically chosen in the 2θ range 540°, and a triclinic cell was found, using the algorithm N-TREOR,12 with a volume of 502 Å3. The volume is consistent with the presence of two nicotinamide 3 CaCl2 units plus two or four water molecules; therefore, the structure was likely to be centrosymmetric (space group nr. 2, P1), with Z = 2 and Z0 = 1. The structure was then solved by simulated annealing using (i) one nicotinamide molecule, (ii) one Ca2+ 3 3 3 Owater or Ca2+ 3 3 3 (Owater)2 fragment 5624
dx.doi.org/10.1021/cg201177p |Cryst. Growth Des. 2011, 11, 5621–5627
Crystal Growth & Design
Figure 5. Coordination polyhedra around the calcium cation in (a) malonamide 3 CaCl2 3 2H2O, (b) piracetam2 3 CaCl2 3 2H2O, (c) nicotinamide 3 CaCl2 3 H2O, (d) nicotinamide2 3 CaCl2 3 2H2O, (e) diacetamide 3 CaCl2 3 5H2O, and (f) barbituric acid 3 CaCl2 3 5H2O. Color codes: orange, calcium cations; green, chloride ions; blue, nitrogen; red, amido oxygen; and light blue, water oxygen.
taken from known structures of similar compounds, (iii) two independent chloride ions, and (iv) one or two independent water molecules. Ten runs per simulated annealing trial were set, and various cooling rates (defined as the ratio Tn/Tn1), from 0.90 to 0.98, were used. Several trials with different fragments were performed assuming the formula to be either nicotinamide 3 CaCl2 3 H2O or nicotinamide 3 CaCl2 3 2H2O: Most solutions with two water molecules per formula unit showed an overlapping of the water molecules, thus indicating the presence of only one water molecule per formula unit. The best solution consistent with the formula nicotinamide 3 CaCl2 3 H2O was chosen for a Rietveld refinement, which was performed with the software GSAS. A shifted Chebyshev function with eight parameters and a Pseudo-Voigt function (type 4) were used to fit background and peak shape, respectively. Soft costraints were applied on bond distances and angles of the nicotinamide molecule, and a planar group restraint was applied to the aromatic ring. An overall thermal parameter for atoms of the same element in the nicotinamide molecule was adopted. Refinement converged with χ2 = 1.031, Rwp = 7.21, and RF2= 7.02. Figure 3 shows the experimental, calculated, and difference diffraction patterns.
’ RESULTS AND DISCUSSION Figure 4 shows a schematic representation of the ICCs structures with CaCl2 discussed herein (see also Table 1). It is easy to appreciate how in all of these ICCs, the organic molecules act as kind of special “solvent molecules” toward calcium chloride. The CdO dipoles establish electrostatic interactions
ARTICLE
Figure 6. Relevant hydrogen-bonding interactions involving the organic moieties in (a) barbituric acid 3 CaCl2 3 5H2O [OCO 3 3 3 (H)Owater in the range 2.809(3)2.982(3); N(H) 3 3 3 Cl, 3.263(4) and 3.275(3) Å], (b) diacetamide 3 CaCl2 3 5H2O [OCO 3 3 3 (H)Owater, 2.777(8) and 2.993(7); N(H) 3 3 3 Cl, 3.227(5) Å], (c) malonamide 3 CaCl2 3 2H2O [N(H) 3 3 3 Cl in the range 3.310(2)3.481(2); N(H) 3 3 3 Owater, 3.144(3) Å], (d) nicotinamide 3 CaCl2 3 H2O [N(H) 3 3 3 Cl, 3.29(1) and 3.31(1)], (e) nicotinamide2 3 CaCl2 3 2H2O [N(H) 3 3 3 Cl, 3.380(3); N(H) 3 3 3 N, 2.953(2) Å], and (f) piracetam2 3 CaCl2 3 2H2O [N(H) 3 3 3 Cl, 3.42(1) and 3.63(1) Å; H atoms not located, but N(H) 3 3 3 Cl hydrogen bonds are comparable to those observed in the Sr analogue]. Color codes are as in Figure 5; gray, carbon; light gray, hydrogen. Hwater atoms are not shown for clarity.
with the calcium cations, while the NH dipoles form hydrogen bonds with the chloride anions. When the cocrystal is a hydrate, the organic molecules are in competition with water in the coordination effort. The calcium coordination polyhedra in the various crystals expectedly show large differences in shape and in the number of coordinating atoms. In malonamide 3 CaCl2 3 2H2O, for example, two malonamide molecules, two water molecules, and two chloride anions are coordinated to the calcium cation, which lies on a crystallographic inversion center (Figure 5a). In the two isomorphous compounds piracetam2 3 CaCl2 3 2H2O and piracetam2 3 SrCl2 3 2H2O, on the other hand, four piracetam molecules and two water molecules are involved in the coordination of the calcium cation (Figure 5b). In both the monohydrate nicotinamide 3 CaCl2 3 H2O and the dihydrate nicotinamide2 3 CaCl2 3 2H2O, the calcium cation interacts with two nicotinamide molecules, the difference between the two systems being in the relative number of water molecules and chloride ions in the coordination polyhedra, which is two and two, and one water molecule and three in the two cocrystals, respectively (Figure 5c,d). Furthermore, the polyhedra share a ClCl edge two by two in nicotinamide2 3 CaCl2 3 2H2O. The coordination polyhedron is a pentagonal bipyramid (coordination seven) in diacetamide 3 CaCl2 3 5H2O (Figure 5e): The calcium cations interact with both CdO terminations of one diacetamide molecule and five water molecules each. The coordination polyhedron is instead a 5625
dx.doi.org/10.1021/cg201177p |Cryst. Growth Des. 2011, 11, 5621–5627
Crystal Growth & Design
ARTICLE
Figure 8. XRPD patterns for piracetam 3 CaCl2 3 2H2O at room temperature (bottom curve) and at 180 °C (top curve). Figure 7. XRPD patterns for nicotinamide 3 CaCl2 3 H2O at room temperature (bottom curve) and at 240 °C (top curve).
bicapped trigonal prism (coordination eight) in barbituric acid 3 CaCl2 3 5H2O (Figure 5f) with the calcium cations interacting with three barbituric acid molecules and five water molecules and the polyhedra sharing one waterwater edge two by two. Structural properties and organization make ICCs analogous to solvates and hydrates of inorganic salts. Therefore, the presence of hydrogen-bonding acceptor sites on the organic components brings about additional supramolecular requirements that have to be fulfilled on construting the new crystalline edifices, which also accounts for the stability of the cocrystal with respect to the organic and inorganic components separately. Relevant hydrogen-bonding interactions7b,c involving the organic moieties are visualized in Figure 6. A further point worth noting is that crystalline nicotinamide 3 CaCl2 3 H2O differs from all other ICCs we have studied until now in that calcium coordination involves also the nicotinamide heterocyclic nitrogen. This is not unexpected, because the heterocyclic nitrogen acts as an acceptor for hydrogen bonding. Thus, the possibility of forming stable ICCs is likely to be extended to other kinds of acceptors and donors such as aromatic heterocylic nitrogens. In terms of preparation methods, all ICCs described herein were synthesized by mechanochemical solid-state techniques (grinding and kneading) with quantitative reactions, except for nicotinamide2 3 CaCl2 3 2H2O, which can only be obtained as a crystalline material directly from solution. Dry grinding of nicotinamide2 3 CaCl2 3 2H2O produces nicotinamide 3 CaCl2 3 H2O. Physical Properties of ICCs with Piracetam and Nicotinamide. As stated in the introduction, the effect of ICCs formation on physical properties has been tested, because of their possible applications in the pharmaceutical field. IDR measurements (see the Experimental Details) show that ICCs of piracetam and nicotinamide have a lower IDR in physiological solution than the corresponding pure APIs: 136.4 vs 262.8 mg L1 min1 and 187.7 vs 379.4 mg L1 min1, respectively. These differences are significant, as ICC formation almost halves the IDR value. Beside the difference in IDR, the thermal stability of both APIs is enhanced in ICCs with respect to the pure components. While
pure nicotinamide melts at 132 °C, nicotinamide 3 CaCl2 3 H2O is stable up to 142 °C, as shown by variable temperature powder XRPD analysis (see Figure 7).14 At this temperature, nicotinamide 3 CaCl2 3 H2O converts into another crystalline form, presumably the corresponding anhydrous nicotinamide 3 CaCl2 ICC, which is still stable at 240 °C. Analogously, while pure piracetam melts at 127 °C,15 the ICC piracetam2 3 CaCl2 3 2H2O transforms at around 140 °C (see Figure 8) into another crystalline compound (possibly the anhydrous form), which is still stable at 180 °C. So far, we have not managed to determine the structures of these high temperature, stable phases, but it is our interest to explore these issues in a future work.
’ CONCLUSIONS AND OUTLOOK In this paper, we have reported the preparation and full structural characterization of a series of novel ICCs obtained quantitatively by mechanochemical treament of the organic molecules barbituric acid, diacetamide, malonamide, and piracetam, while the inorganic salt coformer is CaCl2. Nicotinamide2 3 CaCl 2 3 2H 2O was only obtained by direct crystallization from solution. All of these ICCs possess structures whose stability can be accounted for in terms of the “solvation properties” of the primary and secondary amides toward the anions and cations of the inorganic salt. These interactions undoubtedly provide the extra thermodynamic drive toward formation of cocrystals with respect to crystallization as separate entities. Water is often brought into the crystal, and this is also crucial in the formation of intermolecular and interionic hydrogen bonds. Removal of water still leads to the formation of stable compounds, but we have thus far been unable to fully characterize the high-temperature anhydrous phases. In the cases of piracetam2 3 CaCl2 3 2H2O and nicotinamide 3 CaCl2 3 H2O, we have also shown that some physical properties of ICC may differ significantly from those of pure organic components. For instance, the thermal stability of piracetam2 3 CaCl2 3 2H2O and nicotinamide 3 CaCl2 3 H2O is greatly enhanced with respect to that of the pure API themselves. On the other hand, IDR values for both compounds are lower for the ICCs than for pure piracetam and nicotinamide, respectively. The high-temperature behavior of ICCs can be very important for industrial processes as well. We are currently working on the 5626
dx.doi.org/10.1021/cg201177p |Cryst. Growth Des. 2011, 11, 5621–5627
Crystal Growth & Design crystal structure determination of the nicotinamide 3 CaCl2 3 H2O and piracetam2 3 CaCl2 3 2H2O high-temperature phases. The structure of nicotinamide 3 CaCl2 3 H2O showed also the possibility of exploiting N 3 3 3 Ca interactions as an alternative to CdO 3 3 3 Ca interactions: The heterocyclic nitrogen can play an analogous role in the metal atom coordination. We aim to further explore solvation properties of molecules characterized by different kinds of dipoles such as aromatic heterocyclic nitrogens. Clearly, the development of an adequate strategy for structure determination directly from powder data has been quintessential to the develeopment of this solid-state chemistry, since the products are easily obtained by mechanical mixing of the reactants. Furthermore, the possibility of using common laboratory diffraction instruments widens up the perspective of direct structural determination of complex structures of ICCs.
’ ASSOCIATED CONTENT Supporting Information. Crystallografic information files (cif) for all strucures described herein. This material is available free of charge via the Internet at http://pubs.acs.org.
bS
ARTICLE
1990, 46, C34. (c) Keller, E. SCHAKAL99, Graphical Representation of Molecular Models; University of Freiburg: Freiburg, Germany, 1999. (8) (a) Fabbiani, F. P. A.; Allan, D. R.; Parsons, S.; Pulham, C. R. CrystEngComm 2005, 7, 179–186. (b) Admiraal, G.; Eikelenboom, J. C.; Vos, A. Acta Crystallogr., Sect. B 1982, B38, 2600–2606. (9) (a) Hino, T.; Ford, J. L.; Powell, M. W. Thermochim. Acta 2001, 374 (1), 85–92. (b) Wright, W. B.; King, G. S. D. Acta Crystallogr. 1953, 7, 283–288. (10) Larson, A. C.; Von Dreele, R. B. General Structure Analysis System (GSAS); Los Alamos National Laboratory LAUR 86-748: Los Alamos, NM, 2000. (11) Altomare, A.; Camalli, M.; Cuocci, C.; C., G.; A., M.; Rizzi, R. J. Appl. Crystallogr. 2009, 42, 1197–1202. (12) Altomare, A.; Giacovazzo, C.; Moliterni, A. G. G.; Rizzi, R. J. Appl. Crystallogr. 2001, 34, 704–709. (13) Blatov, V. A.; Carlucci, L.; Ciani, G.; Proserpio, D. M. CrystEngComm 2004, 6, 377–395. (14) Shen, J.; Zheng, J.; Che, Y.; Xi, B. J. Cryst. Growth 2003, 257, 136–140. (15) Maher, A.; Croker, D.; Rasmuson, A. C.; Hodnett, B. K. J. Chem. Eng. Data 2010, 55, 5314–5318.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT The University of Bologna and the MIUR (PRIN 2007) are acknowledged. We thank PolyCrystalLine srl for the use of their spectrophotometer for IDR measurements. ’ REFERENCES (1) (a) Braga, D. Chem. Commun. 2003, 2751–2754. (b) Braga, D.; Grepioni, F.; Maini, L. Chem. Commun. 2010, 46, 6232–6242. (2) (a) Schultheiss, N.; Newman, A. Cryst. Growth Des. 2009, 9 (6), 2950–2967. (b) Vishweshwar, P.; McMahon, J. A.; Peterson, M. L.; Hickey, M. B.; Shattock, T. R.; Zaworotko, M. J. Chem. Commun. 2005, 36, 4601–4603. (3) (a) Aakeroy, C. B.; Salmon, D. J. CrystEngComm 2005, 7 (72), 439–448. (b) Bond, A. D. CrystEngComm 2007, 9 (9), 833–834. (c) Childs, S. L.; Stahly, G. P.; Park, A. Mol. Pharmaceutics 2007, 4 (3), 323–338. (d) Desiraju, G. R. CrystEngComm 2003, 5, 466–467. (e) Dunitz, J. D. CrystEngComm 2003, 5, 506. (f) Mohamed, S.; Tocher, D. A.; Vickers, M.; Karamertzanis, P. G.; Price, S. L. Cryst. Growth Des. 2009, 9 (6), 2881–2889. (g) Vishweshwar, P.; McMahon, J. A.; Bis, J. A.; Zaworotko, M. J. J. Pharm. Sci. 2006, 95 (3), 499–516. (h) Good, D. J.; Rodriguez-Hornedo, N. Cryst. Growth Des. 2009, 9 (5), 2252–2264. (i) Friscic, T.; Jones, W. J. Pharm. Pharmacol. 2010, 62 (11), 1547–1559. (j) Sanphui, P.; Goud, N. R.; Khandavilli, U. B. R.; Nangia, A. Cryst. Growth Des. 2011, 11, 4135–4145. (4) (a) Braga, D.; Grepioni, F.; Maini, L.; Prosperi, S.; Gobetto, R.; Chierotti, M. R. Chem.Commun. 2010, 46, 7715–7717.(b) Lampronti, G. I.; Maini, L.; Braga, D.; Grepioni, F.; Capucci, D.; Cuocci, C., J. Appl. Crystallogr., submitted for publication. (5) (a) Lebioda, L.; Lewinski, K. Acta Crystallogr., Sect. B 1980, B36 (3), 693–695. (b) Lebioda, L.; Stadnicka, K.; Sliwinski, J. Acta Crystallogr., Sect. B 1979, B35 (1), 157–158. (6) Favre-Nicolin, V.; Cerny, R. J. Appl. Crystallogr. 2002, 35, 734–743. (7) (a) Sheldrick, G. M. SHELXL97; University of G€ottingen: Germany, 1997. (b) Speck, A. L. PLATON. Acta Crystallogr., Sect. A 5627
dx.doi.org/10.1021/cg201177p |Cryst. Growth Des. 2011, 11, 5621–5627