Polymer-Assisted Grinding, a Versatile Method for Polymorph Control

Jan 28, 2016 - The approach proposed in this paper can be readily applied to each system, where polarity is the main issue for polymorph control witho...
1 downloads 0 Views 9MB Size
Article pubs.acs.org/crystal

Polymer-Assisted Grinding, a Versatile Method for Polymorph Control of Cocrystallization Dritan Hasa,†,‡ Elvio Carlino,§ and William Jones*,† †

Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge, CB2 1EW, United Kingdom Department of Chemical and Pharmaceutical Sciences, University of Trieste, Piazzale Europa 1, 34127, Trieste, Italy § TASC-IOM-CNR, AREA Science Park, S.S.14 Km 163 Basovizza, I-34149 Trieste, Italy ‡

S Supporting Information *

ABSTRACT: Despite the great interest that cocrystals are currently gaining for their application to the design of new supramolecular structures with desired functional properties, studies concerning new experimental strategies capable of controlling polymorphism phenomena of a given system are scarcely reported. We propose herein the use of polymer-assisted grinding (POLAG) as a new method for the selective control of the product polymorphic form in a mechanochemical cocrystallization reaction. Specifically, to the model system selected in this study formed by caffeine and glutaric acid, we demonstrate that the polymorphic outcome can be controlled by modifying the number of monomer units of the catalyst from the shortest dimer to a polymer with chains of approximately 1000 units. The characteristics of each polymorphic form were investigated by low-dose high-resolution TEM, and the mechanistic aspects of the cocrystal formation were studied through a series of ex situ and interconversion experiments. The results suggest that for this system the modification of the catalyst chain length and, consequently, modification of polarity drives cocrystal formation toward the more stable polymorph. The approach proposed in this paper can be readily applied to each system, where polarity is the main issue for polymorph control without the risk of solvate formation.



INTRODUCTION

method that guarantees the discovery of all of the energetically reasonable polymorphs. The need for the development of efficient preparation techniques is even more logical because the traditional solutionbased methods are often limited, in part because of differences in solubility of cocrystal components and/or solvent−solute interactions.12 In this context, solid-state techniques, such as mechanochemistry, have been proposed for the screening of new multicomponent crystal forms.13 Such one-step methods are environmentally friendly, cheaper as compared with multistage solution synthesis, and simple because they do not require knowledge of the solubilities of the cocrystallizing compounds.14,15 The neat cogrinding of two or more components represents the simplest case of a mechanochemical cocrystallization reaction. The polymorphic outcome during the

Cocrystals, solid forms having two or more different molecules in the crystal structure,1−3 continue to gain significant interest for their application to the design of new supramolecular structures with desired functional properties.4−8 For a given cocrystal system, more than one polymorphic structure can be found,9 each with possibly different properties.10 Polymorph control, hence, amounts to the control of physicochemical properties and other characteristics of the crystalline material, as illustrated in the pharmaceutical case of the anti-HIV drug ritonavir.11 Despite the great importance of controlling the crystal product outcome for applications from material science to drug delivery,10 however, studies concerning the phenomenon of cocrystal polymorphism are not frequently reported.9 An explanation can be related to the inappropriateness of the experimental strategies used to discover the highest number of polymorphic forms of a given cocrystal. There is no general © 2016 American Chemical Society

Received: January 18, 2016 Published: January 28, 2016 1772

DOI: 10.1021/acs.cgd.6b00084 Cryst. Growth Des. 2016, 16, 1772−1779

Crystal Growth & Design

Article

HRTEM Analysis. The specimens have been prepared for transmission electron microscopy experiments by depositing the asprepared powders on a copper grid previously covered with an amorphous carbon film. The specimens were studied using a JEOL 2010F UHR microscope equipped with a low spherical aberration coefficient (Cs = 0.47 mm ±0.01 mm) objective pole piece. The microscope was operated at room temperature at an accelerating voltage of 200 kV. Electron-optical conditions used enabled a spatial resolution in HRTEM at an optimal defocus of 0.19 nm.22 As the material is highly sensitive to electron irradiation, the experimental conditions were chosen using a density of current as low as 1 pA/cm2 (as measured on the microscope phosphorus detector) to minimize the electron dose on the sample, thereby enabling lattice imaging at atomic resolution.23,24 The morphology and crystal structure of the specimens were studied by imaging in bright field and phase contrast high resolution TEM together with the relevant diffractograms. The diffractograms were derived from each region of interest by DigitalMicrograph.25 As the particles in the specimens are randomly oriented with respect to the direction of the primary electron beam, extensive experiments have been performed to detect those particles oriented along, or very close to, a specific zone axis, thus enabling the unequivocal identification of the relevant polymorph by comparing the experimental reciprocal vectors ratio, angle and length to those calculated for each zone axis of Form I and Form II. The experimental diffractograms have also been compared to the simulated patterns of pristine caffeine and glutaric acid to rule out their eventual presence. All of the considered crystal structures and relevant diffraction patterns were simulated by JEMS.26 To highlight the presence of the longer polymer chains in the imaging experiments, some bright field pictures were acquired with strong out of focus to make visible the Fresnel’s fringes that follow the polymer profiles.27

mechanochemical neat cocrystallization of reactants, however, will exclusively depend on the grinding conditions, such as the type of mechanical activator and mode of treatment, and generally, there is a lower flexibility on the process conditions for the control of the polymorphic outcome. Better control over the product outcome via mechanochemistry can be achieved by grinding with a minimal addition of a liquid (liquidassisted grinding, LAG) with the liquid having a catalytic role during the cocrystallization reaction.16,17 Varying the chemical nature of the liquid and its amount represent the two most significant variable parameters, which have a pronounced effect on the product crystal outcome. Conversely, although the amount of liquid used in LAG reactions is generally low, often it is sufficient to promote the formation of crystal solvates. This phenomenon is frequent during cocrystal screening by LAG, and several examples of cocrystal solvates are reported in the literature.18−20 Recently, we reported a new approach, namely, polymerassisted grinding (POLAG), which eliminates the risk of solvate formation during the mechanochemical production of cocrystals.21 In that study, the primary objective was to compare the use of a polymer during cocrystal screening through grinding with already existing methods, such as LAG and neat grinding. As such, we developed an experimental data set with three different cocrystal systems previously reported in the literature to demonstrate that POLAG can be an equivalent alternative to the classical grinding techniques for cocrystal discovery and which could also guarantee similar advantages in terms of mechanochemical reaction rate.21 In this study, we propose another important advantage of POLAG over the classical LAG reaction in controlling polymorphic outcome during the mechanochemical cocrystallization of organic molecules. Whereas in LAG reactions the polymorph generated can be controlled using different liquids with different chemical properties (for example, polar and nonpolar liquids), POLAG offers the possibility of controlling the microenvironment polarity through a polymer engineering strategy, without a change in the functional groups, by simply modifying the number of monomer units (i.e., polymer molecular weight).





RESULTS AND DISCUSSION The experimental data set was developed using a previously reported cocrystal of the model pharmaceutical compound caffeine (CAF, Figure 1).16 The system under study consists of two polymorphic forms of the molecular cocrystal containing CAF and glutaric acid (GLA, Figure 1) in a 1:1 stoichiometry. CAF-GLA cocrystal Form I and Form II have similar hydrogen-

EXPERIMENTAL SECTION

Materials. Anhydrous CAF, GLA, EG derivatives, and other chemicals used in this study were purchased from Sigma-Aldrich Company, Ltd. (Gillingham, UK) and used without further purification. Grinding Procedure. CAF-GLA cocrystals were prepared mechanochemically following the experimental conditions reported in ref 16: 200 mg of equimolar material (119 mg of CAF and 81 mg of GLA) and 30 mg of an EG derivative (the optimal amount of polymer when used as an additive for cocrystallization varies from 10 to 15 wt %21) were transferred in a 15 mL steel jar containing two grinding balls of 7 mm. The mixture was then ground for 60 min in a Retsch MM200 grinding mill at a frequency of 25 Hz. In the case of interconversion experiments, 200 mg of pure cocrystal polymorphic Form I or Form II and 30 mg of an EG derivative were ground for 60 min at 30 Hz. PXRD Analysis. The resulting product material of each grinding experiment was characterized by powder X-ray diffraction (PXRD). The diffraction analyses were performed at room temperature using a Panalytical X’Pert Pro Diffractometer with Ni-filtered Cu Kα radiation (wavelength 1.5418 Å) equipped with an RTMS X’celerator detector. A small amount of cocrystal powder (20−30 mg) was gently pressed on a glass slide to give a flat surface and subsequently analyzed, collecting the data in the 2 theta range 3−50°.

Figure 1. Structures of CAF and GLA and the hydrogen bonding present in the two conformational polymorphs of the CAF-GLA cocrystal: Form I (top) viewed along the b axis and Form II (bottom) viewed along the a axis. 1773

DOI: 10.1021/acs.cgd.6b00084 Cryst. Growth Des. 2016, 16, 1772−1779

Crystal Growth & Design

Article

bond synthons and exhibit identical secondary architecture, such that a two-dimensional sheet results from an array of linear O−H···O hydrogen-bonded tapes established between the hydroxyl groups of one GLA molecule and the carbonyl of the next. The second carboxylic group of GLA interacts with the imidazole ring of CAF to form an O−H···N hydrogen bond (Figure 1). Both of the CAF-GLA polymorphic forms can be prepared via mechanochemistry and solution crystallization.16 Neat cogrinding of CAF and GLA gives cocrystal Form I. The same polymorphic form can also be prepared by LAG with nonpolar liquid, such as hexane and heptane, whereas Form II can be obtained by LAG with polar liquids, such as acetonitrile and dichloromethane.16 CAF-GLA cocrystal Form I is metastable and converts to Form II under high humidity conditions or when crystals are dispersed in acetonitrile.28,29 In the present study, we used a series of ethylene glycol (EG) derivatives as catalysts for the mechanochemical cocrystallization of CAF and GLA because this nontoxic class of molecules shows wide-ranging application in biological and pharmaceutical contexts.30 The resulting material of each grinding experiment was characterized by powder X-ray diffraction (PXRD) and, in particular, high resolution transmission electron microscopy (HRTEM) experiments. Co-grinding of CAF and GLA in the presence of 15 wt % of diethylene glycol (DEG), the smallest oligomer of EG, provided pure polymorphic Form II (Figures 2 and 3).

Figure 3. Representative TEM results of CAF-GLA prepared with DEG: (a) bright field image showing the morphology of the specimen consisting of small crystalline particles, (b) HRTEM image of a particles along with the relevant diffractogram shown in (c) and indexed as Form II oriented along the [2,−4,3] zone axis.

Figure 4. Representative TEM results of CAF-GLA prepared with TEG: (a) HRTEM image showing the morphology of small particles assembled to create chain-like structures; the arrows marked by 1 and 2 point to particles whose diffractograms indicate Form II oriented along the [3,5,1] zone axis. The diffractogram of the particle marked as 1 is shown in (c). b) HRTEM image of another specimen region: the arrows point to the particles 1−3. Particles 1 and 2 are particles of Form I oriented along the [1,1,2] zone axis, and particle 3 is made of Form II oriented along the [1,−3,2] zone axis. (d) Diffractogram of particle 1 in (b); (d) diffractogram of particle 3 in (b).

observed (Figure 2). The presence of Form I has also been confirmed by high resolution TEM (HRTEM) along with the relevant diffractograms, as shown in Figure 4b and d. HRTEM experiments enable a study of the shape and crystal properties of individual particles, but for the quantification, the PXRD has a higher statistical relevance as it explores a much bigger specimen volume. The presence of polymorphic Form I would appear to be more evident in the PXRD pattern of CAF-GLA prepared using hexaethylene glycol (HEG) (Figures 2 and 5). From Figure 2, it is also worth noting the presence of a significant preferential orientation phenomenon, which is present in the PXRD patterns of the cocrystal products prepared with DEG, TEG, HEG, and PEG 200, because these molecules are viscous liquids at room temperature. Instead, in the case of CAF-GLA prepared in the presence of a semisolid product, such as PEG 1000, the preferential orientation phenomenon appears to be significantly lower. Interestingly, the scenario appears to be reversed when the number of EG units was increased, such as in the case of

Figure 2. PXRD patterns of the products prepared via POLAG in the presence of different EG derivatives. The calculated PXRD patterns16 of CAF-GLA polymorphs differ from each other, especially in the low range of the 2 theta angle.

Similarly, the mechanochemical cocrystallization of CAF and GLA in the presence of a slightly longer oligomer, such as triethylene glycol (TEG), also provided CAF-GLA Form II (Figures 2 and 4). However, in the PXRD pattern of CAF-GLA cocrystal Form II prepared with TEG, low traces of polymorphic Form I particularly distinguishable from the appearance of the typical peak at 7 degrees 2 theta was 1774

DOI: 10.1021/acs.cgd.6b00084 Cryst. Growth Des. 2016, 16, 1772−1779

Crystal Growth & Design

Article

Figure 5. Representative TEM results of CAF-GLA prepared with HEG: (a) bright field image showing the morphology of the specimen made of isolated particles and chains (both crystalline). (b) HRTEM image with the areas marked as 1 and 2 corresponding to the diffractograms shown in (c) and (d), respectively, and revealing the presence of both Form I and Form II.

Figure 7. Representative TEM results of CAF-GLA Form I prepared with PEG 1000: (a) bright field image showing the morphology of the specimen, which is made of clusters of tapes of crystalline material; the region marked in red is seen at higher magnification in (b), which is a HRTEM image of some portions of the tapes with different orientations with respect to the primary electron beam; in particular, the arrow points to an area oriented close to a zone axis that enables recognition of the crystal structure as shown in (c) and (d). c) Phase contrast high resolution image; the arrow points to the region where the diffractogram was acquired. (d) Diffractogram belonging to [−2,0,1] zone axis of Form I. (e) Arrangement of molecules in Form I crystal oriented along the [−2,0,1] zone axis.

polyethylene glycol 200 (PEG 200, Figures 2 and 6). In fact, the PXRD pattern of the cocrystal product prepared via

depends on the liquid polarity: cocrystal Form I is generated by LAG with nonpolar liquids, whereas Form II can be obtained with liquids of a higher polarity. The reason for such a strong dependence on the liquid polarity can be found in the crystal packing differences between the two polymorphs.16 In fact, although the two cocrystals show a common secondary level of architecture (Figure 1), their tertiary structure is quite different (Figure 8). Specifically, the orientation of Form I along the c

Figure 6. Representative TEM results of CAF-GLA prepared with PEG 200: (a) bright field image showing the morphology of the sample made of small crystalline particles; the inset was acquired with a strong out of focus and reveals the presence of strands of material attached to the particles. (b) HRTEM image along with the relevant diffractogram (e) of a particle of Form II; (c) HRTEM image along with the relevant diffractogram (d) of a particle of Form I.

POLAG using PEG 200 corresponds to the calculated pattern of the polymorphic Form I, and only a trace amount of Form II (distinguishable from the presence of the characteristic peak at 8.5 degrees of 2 theta) can be detected (Figure 2). A higher molecular weight polymer, such as PEG 1000, provided the pure polymorphic Form I (Figures 2 and 7). It is also noteworthy that the morphology of the material is different in the case of PEG1000 (Figure 7a). As in most of the cases, the crystalline Form I is found in the shape of a bundle of tapes (Figure 7a−c) and more seldomly as isolated particles. As mentioned previously, the polymorphic outcome during the CAF and GLA mechanochemical cocrystallization strongly

Figure 8. Crystal packing diagrams of CAF-GLA cocrystal polymorphic Form I (left) and Form II (right).

axis shows how the sheets stack along the b axis, forming nonpolar cleavage planes (Figure 8). Instead, in the crystal packing of CAF-GLA Form II, the sheets stack in a staggered fashion; thus, nonpolar cleavage planes are not observed.16 The nonpolar planes observed in Form I, consequently, would possibly be more stable in a nonpolar environment. 1775

DOI: 10.1021/acs.cgd.6b00084 Cryst. Growth Des. 2016, 16, 1772−1779

Crystal Growth & Design

Article

and is calculated on the basis of the cohesive energy density (the energy necessary to separate the molecules of a solid or a liquid to a sufficient distance so as to eliminate all molecular interactions) divided by the molar volume.34 Subsequently, Hansen extended the SP concept through a subdivision of the total SP (δtot) into three partial parameters: dispersion parameter (δd), polar parameter (δp), and a parameter associated with hydrogen bonds (δh).35 Several studies have proposed different approaches to calculate Hansen’s SP. In this study, we used an improved group contribution parameter recently proposed by Just et al.36 The Hansen’s SP of the EG derivatives were calculated by including the average number of the repeating units based on the declared molecular weight. The results are reported in Figure 9. It is important to emphasize that, despite the term, the definition of SP does not cover the solubility of a solid because the entropy effects of the solubilization are not considered. Instead, the SP can be used to give an estimation of the enthalpic contribution to the mixing energy, relying on the chemical rule of similarity to define the interaction strength between different species.36 In mechanochemical cocrystallization, the reactants interact with the third component in which they are dispersed, which can be a liquid (in the case of LAG) or a polymer in the case of POLAG. Hence, the chemical nature of the microenvironment where the reactants are dispersed is likely to influence the early stages of cocrystal nucleation and growth.14,15,17 We therefore speculate that the SP, specifically the polarity parameter δp, can be used as an indicator of the nature (in terms of polarity) of the microenvironment where CAF and GLA are dispersed and cocrystallize. From Figure 9, it can be observed that in the case of DEG the polar parameter (yellow histogram) represents the major component of the total SP (gray histogram). Hence, the mechanochemical cocrystallization of CAF and GLA in the presence of DEG would occur in a polar environment, and CAF-GLA Form II would be the favored polymorphic form. The polar parameter remains the major component in the case of TEG, although its value is significantly lower. In the case of HEG, δp is still present, but another parameter, namely δd, is the major component. In longer chain EG derivatives such as PEG 200, the δp influence drastically decreases and the total SP

This hypothesis would also suggest that the polarity of the different EG derivatives used in this study depends on the number of monomer units (Figure 9).

Figure 9. (top) Schematic representation of how the polar zones are less influential in long chain oligomers. (bottom) Calculated solubility parameters of the EG derivatives according to Just et al.36

We used the solubility parameter (SP) approach to better understand the characteristics of the EG derivatives. SP, also referred to as cohesion parameters, is frequently used for estimating the interaction capacities of various organic liquids, polymers, and solids.31−33 SP was introduced by Hildebrand

Figure 10. PXRD patterns of the CAF-GLA cocrystals after different milling time intervals prepared via POLAG with DEG (left) and PEG 1000 (right). 1776

DOI: 10.1021/acs.cgd.6b00084 Cryst. Growth Des. 2016, 16, 1772−1779

Crystal Growth & Design

Article

Figure 11. Study of the mechanochemical interconversion of CAF-GLA cocrystal Form I (left) and Form II (right) in the presence of 15 wt % of different EG derivatives.

(δtot, gray histogram) is represented mainly by the dispersion component δd (green histogram). However, because the dispersion parameter remains almost constant in all of the EG derivatives, it can be speculated that δd does not have a significant role in the polymorph control of the CAF-GLA cocrystal. It is also worth noting that the hydrogen bonding component (blue histograms) decreases passing from HEG to PEG 1000. In fact, it is reported that the proportion of hydroxyl functions in relation to the number of monomers decreases to 9% in PEG 1000.36 Finally, we investigated the mechanochemical formation of CAF-GLA as a function of the grinding time through a series of ex situ experiments. Pure CAF and GLA were ground in the presence of each EG derivative (DEG, TEG, HEG, PEG 200, and PEG 1000) for different periods of time, and the product was immediately analyzed by PXRD. As illustrative examples, in Figure 10 is reported the mechanochemical CAF-GLA cocrystal formation in the presence of DEG and PEG 1000 as a function of time. In the case of POLAG experiments with PEG 1000, the reflections of CAF-GLA polymorphic Form I can be observed after 5 min of grinding together with some reflections of the starting materials. The diffraction peaks of the crystalline starting materials progressively disappear with longer grinding time. The POLAG of CAF-GLA Form I in the presence of PEG 1000 appears to proceed without any intermediate (Figure 10). Conversely, the formation pathway of CAF-GLA in the presence of DEG would appear to be more intriguing. The PXRD pattern of the cocrystal product ground for 5 min shows the reflections of CAF-GLA Form I together with the reflections of the starting materials, whereas the first reflections of the cocrystal polymorphic Form II appear after 15 min of grinding. With a longer grinding time, the reflections of Form I become less evident, and the PXRD pattern resembles the calculated pattern of polymorphic Form II. After 45 min of grinding, the reflections typical of Form I totally disappear (Figure 10). The cocrystal formation pathways in the presence of the other EG derivatives (TEG, HEG, and PEG 200) are similar to the two cases represented in Figure 10 (Supporting Information). The data suggest that the mechanochemical formation of CAF-GLA Form II includes the formation of kinetic Form I as a transitional stage, which subsequently

converts to polymorphic Form II. Similar behavior was also observed during the solution crystallization of CAF-GLA Form II.37 We also investigated the possibility that the interconversion of Form I is related to the catalyst polarity. Additional experiments confirmed that CAF-GLA Form I converts to the stable Form II after 60 min of grinding in the presence of DEG and TEG, whereas it is more stable in the presence of HEG and PEG 200 and no conversion can be observed in the presence of PEG 1000 (Figure 11). Form II, in contrast, is stable and does not convert in the EG derivatives except in PEG 1000, where a partial conversion is observed (Figure 11).



CONCLUSIONS In summary, we have successfully demonstrated that POLAG is an effective method to selectively control the polymorphic outcome of a mechanochemical cocrystallization reaction. The polarity factor was selected because it appears to be one of the main reasons for polymorph diversity during screening. Specific to the CAF-GLA system considered in this study, the product outcome was controlled by simply changing the number of monomer units of the EG derivatives during POLAG experiments. The characterization part of the study clearly demonstrated the reasons how and why the polymer chain length influenced the polymorphic outcome. Indeed, the calculations of the solubility parameters, together with a series of ex situ and interconversion experiments, suggested that alteration of chain length modified the catalyst polarity, and thus directs the CAF-GLA cocrystal formation toward the favored polymorphic form. Additionally, HRTEM analysis also suggested that the characteristics of the products obtained in the presence of different EG derivatives change. TEM experiments show two extremes in the morphology of the materials between DEG and PEG 1000 cocrystals: PEG 1000 sample evidenced the formation of relatively large ribbons of layered crystalline material, whereas in the DEG specimen, the material is made of isolated small particles. In other specimens, an intermediate situation appears to exist with the presence of isolated particles and particles merged into elongated rods. It could be argued that Form I, crystallized in the presence of PEG1000, finds a favorable low polarity environment in which to grow as long ribbons, whereas the presence of higher polarity EG polymers promotes the formation of Form II and the 1777

DOI: 10.1021/acs.cgd.6b00084 Cryst. Growth Des. 2016, 16, 1772−1779

Crystal Growth & Design

Article

(9) Aitipamula, S.; Chow, P. S.; Tan, R. B. H. Polymorphism in cocrystals: a review and assessment of its significance. CrystEngComm 2014, 16, 3451−3465. (10) Bernstein, J. Polymorphism in Molecular Crystals; Oxford University Press: Oxford, UK, 2002. (11) Bauer, J.; Spanton, S.; Henry, R.; Quick, J.; Dziki, W.; Porter, W.; Morris, J. Ritonavir: an extraordinary example of conformational polymorphism. Pharm. Res. 2001, 18, 859−866. (12) Rodriguez-Hornedo, N.; Nehm, S. J.; Seefeldt, K. F.; PaganTorres, Y.; Falkiewicz, C. J. Reaction Crystallization of Pharmaceutical Molecular Complexes. Mol. Pharmaceutics 2006, 3, 362−367. (13) Braga, D.; Maini, L.; Grepioni, F. Mechanochemical preparation of co-crystals. Chem. Soc. Rev. 2013, 42, 7638−7648. (14) James, S. L.; Adams, C. J.; Bolm, C.; Braga, D.; Collier, P.; Friscic, T.; Grepioni, F.; Harris, K. D. M.; Hyett, G.; Jones, W.; et al. Mechanochemistry: opportunities for new and cleaner synthesis. Chem. Soc. Rev. 2012, 41, 413−447. (15) Boldyreva, E. Mechanochemistry of inorganic and organic systems: what is similar, what is different? Chem. Soc. Rev. 2013, 42, 7719−7738. (16) Trask, A. V.; Motherwell, S. W. D.; Jones, W. Solvent-drop grinding: green polymorph control of cocrystallisation. Chem. Commun. 2004, 890−891. (17) Friscic, T.; Jones, W. Recent advances in understanding the mechanism of cocrystal formation via grinding. Cryst. Growth Des. 2009, 9, 1621−1637. (18) Britton, D.; Chantooni, M. K.; Kolthoff, I. M. The Structures of Polymorph I at 181 K and Polymorph II at 297 K of the Bis(waterdichloropicric acid)-1,4,7,10,13,16-Hexaoxacyclooctadecane Complex, 2[H20.C6(NO2)3Cl2(OH)].(CH2CH20)6. Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 1988, 44, 303−306. (19) Bhattacharya, S.; Saha, B. K. Guest-induced isomerization of net and polymorphism in Trimesic Acid−Arylamine complexes. Cryst. Growth Des. 2011, 11, 2194−2204. (20) Madusanka, N.; Eddleston, M. D.; Arhangelskis, M.; Jones, W. Polymorphs, hydrates and solvates of a co-crystal of caffeine with anthranilic acid. Acta Crystallogr., Sect. B: Struct. Sci., Cryst. Eng. Mater. 2014, 70, 72−80. (21) Hasa, D.; Schneider Rauber, G.; Voinovich, D.; Jones, W. Cocrystal formation through mechanochemistry: from neat and liquidassisted grinding to polymer-assisted grinding. Angew. Chem., Int. Ed. 2015, 54, 7371−7375. (22) Spence, J. C. H. Experimental High-Resolution Electron Microscopy, 2nd ed.; Oxford University Press: Oxford, UK, 1988. (23) De Caro, L.; Carlino, E.; Caputo, G.; Cozzoli, P. D.; Giannini, G. Electron diffractive imaging of oxygen atoms in nanocrystals at subångström resolution. Nat. Nanotechnol. 2010, 5, 360−365. (24) De Caro, L.; Carlino, E.; Vittoria, F. A.; Siliqi, D.; Giannini, C. Keyhole electron diffractive imaging (KEDI). Acta Crystallogr., Sect. A: Found. Crystallogr. 2012, 68, 687−702. (25) http://www.gatan.com/products/tem-analysis/gatanmicroscopy-suite-software. (26) http://cimewww.epfl.ch/people/stadelmann/jemsWebSite/ jems.html. (27) Reimer, L. Transmission Electron Microscopy: Physics of image formation and microanalysis; Springer-Verlag: Berlin, Heidelberg, New York, Tokyo, 1984. (28) Thakuria, R.; Eddleston, M. D.; Chow, E. H. H.; Lloyd, G. O.; Aldous, B. J.; Krzyzaniak, J. F.; Bond, A. D.; Jones, W. Use of in situ Atomic Force Microscopy to follow phase changes at crystal surfaces in real time. Angew. Chem., Int. Ed. 2013, 52, 10541−10544. (29) Yu, Z. Q.; Chow, P. S.; Tan, R. B. H.; Ang, W. H. Supersaturation control in cooling polymorphic co-crystallization of caffeine and glutaric acid. Cryst. Growth Des. 2011, 11, 4525−4532. (30) Henning, T. Polyethylene glycols (PEGs) and the pharmaceutical industry. PharmaChem 2002, 2, 57−59. (31) Karger, B. L.; Snyder, L. L. R.; Eon, C. Expanded solubility parameter treatment for classification and use of chromatographic solvents and adsorbents. Anal. Chem. 1978, 50, 2126−2136.

transformation of Form I to Form II. The polar interaction between particle and polymer could eventually saturate the functional groups at the particle surface, reducing particle ripening and hence size. Modifying the microenvironment polarity in LAG cocrystallization would require a change of solvent. This can be risky because the new solvent molecules may be included in the cocrystal structure, leading to the formation of a solvate and therefore requiring the use of further solvent screens. As a result, the number of experiments performed would possibly increase. The new approach proposed in this paper can be readily applied to each system where polarity is the main issue for polymorph control without risk of solvate formation. POLAG represents an unconventional combination between three important topics: crystal engineering, mechanochemistry, and polymer chemistry. We believe that such a high versatility would give the possibility in future studies to extend the use of POLAG toward the (one-step) mechanochemical synthesis of advanced functional materials.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.cgd.6b00084. Additional ex/in situ grinding experiments using TEG, HEG, and PEG 200 (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS D.H. thanks the University of Trieste for partially supporting this study (FRA 2013). Project PRIN 2012 NOXSS partially supported experiments by E.C.



REFERENCES

(1) Etter, M. C. Hydrogen bonds as design elements in organic chemistry. J. Phys. Chem. 1991, 95, 4601−4610. (2) Caira, M. R. Molecular complexes of sulfonamides. 2.1:1 complexes between drug molecules: sulfadimidine-acetylsalicylic acid and sulfadimidine-4-aminosalicylic acid. J. Crystallogr. Spectrosc. Res. 1992, 22, 193−200. (3) Fleischman, S. G.; Kuduva, S. S.; McMahon, J. A.; Moulton, B.; Bailey Walsh, R. D.; Rodriguez-Hornedo, N.; Zaworotko, M. J. Crystal Engineering of the composition of pharmaceutical phases: Multiplecomponent crystalline solids involving Carbamazepine. Cryst. Growth Des. 2003, 3, 909−919. (4) Jones, W. Organic molecular solids: properties and applications; CRC Press: New York, 1997. (5) Atwood, J. L.; Davies, J. E. D.; MacNicol, D. D.; Vogtle, F. W. Comprehensive supramolecular chemistry; Pergamon: Oxford, UK, 1996. (6) Cheney, M. L.; McManus, G. J.; Perman, J. A.; Wang, Z.; Zaworotko, M. J. The Role of Cocrystals in Solid-State Synthesis: Cocrystal-Controlled Solid-State Synthesis of Imides. Cryst. Growth Des. 2007, 7, 616−617. (7) Rodriguez-Hornedo, N. Cocrystals: Molecular Design of Pharmaceutical Materials. Mol. Pharmaceutics 2007, 4, 299−300. (8) Friscic, T.; Jones, W. Benefits of cocrystallisation in pharmaceutical materials science: an update. J. Pharm. Pharmacol. 2010, 62, 1547−1559. 1778

DOI: 10.1021/acs.cgd.6b00084 Cryst. Growth Des. 2016, 16, 1772−1779

Crystal Growth & Design

Article

(32) Ramos, A. C. D. S.; Rolemberg, M. P.; Moura, L. G. M. D.; Zilio, E. L.; Santos, M.D.F.P.D.; Gonzales, G. Determination of solubility parameters of oils and prediction of oil compatibility. J. Pet. Sci. Eng. 2013, 102, 36−40. (33) Hasa, D.; Perissutti, B.; Grassi, M.; Chierotti, M. R.; Gobetto, R.; Ferrario, V.; Lenaz, D.; Voinovich, D. Mechanochemical activation of vincamine mediated by linear polymers: Assessment of some ‘‘critical’’ steps. Eur. J. Pharm. Sci. 2013, 50, 56−58. (34) Hancock, B. C.; York, P.; Rowe, R. C. The use of solubility parameters in pharmaceutical dosage form design. Int. J. Pharm. 1997, 148, 1−21. (35) Hansen, C. M. The universality of the Solubility Parameter. Ind. Eng. Chem. Prod. Res. Dev. 1969, 8, 2−11. (36) Just, S.; Sievert, F.; Thommes, M.; Breitkreutz, J. Improved group contribution parameter set for the application of solubility parameters to melt extrusion. Eur. J. Pharm. Biopharm. 2013, 85, 1191−1199. (37) Yu, Z. Q.; Chow, P. S.; Tan, R. B. H. Design space for polymorphic co-crystallization: Incorporating process model uncertainty and operational variability. Cryst. Growth Des. 2014, 14, 3949− 3957.

1779

DOI: 10.1021/acs.cgd.6b00084 Cryst. Growth Des. 2016, 16, 1772−1779