The Role of Cocrystals in Solid-State Synthesis: Cocrystal-Controlled Solid-State Synthesis of Imides Miranda L. Cheney, Gregory J. McManus, Jason A. Perman, Zhenqiang Wang, and Michael J. Zaworotko*
CRYSTAL GROWTH & DESIGN 2007 VOL. 7, NO. 4 616-617
Department of Chemistry, UniVersity of South Florida, CHE205, 4202 East Fowler AVenue, Tampa, Florida 33620-5250 ReceiVed February 19, 2007
ABSTRACT: Cocrystal-controlled solid-state synthesis (C3S3) of imides occurs via heating of cocrystals formed between anhydride and aromatic amine cocrystal formers. Cocrystals that comprise two or more molecules (cocrystal formers1) that are solids under ambient conditions represent a longknown2 class of compound. However, they remain relatively unexplored; a Cambridge Structural Database (CSD)3 survey reveals that they represent less than 0.5% of published crystal structures. Nevertheless, their potential impact upon pharmaceutical formulation4 and green chemistry5 is of topical and growing interest. In particular, that all components are solids under ambient conditions has important practical considerations, because synthesis of cocrystals can be achieved via solid-state techniques (mechanochemistry)6 and chemists can execute a degree of control over the composition of a cocrystal, by invoking molecular recognition, especially hydrogen bonding, during the selection of cocrystal formers. These features distinguish cocrystals from another broad and well-known group of multiple component compounds, solvates. Solvates are much more widely characterized than cocrystals,7 although this could change because most molecular compounds are solids under ambient conditions. Whereas solid-state organic synthesis represents a wellestablished area of research,8 cocrystal-controlled solid-state synthesis (C3S3) is presently limited to photodimerizations or photopolymerizations9 and nucleophilic substitution.10 In the case of the former, one cocrystal former typically serves to align or “template” the reactant, which is the other cocrystal former. In the case of the latter, both cocrystal formers are reactants, although there are examples in which the reactive moieties are in the same molecule and therefore generate polymeric structures.11 C3S3 offers broad potential in the context of green chemistry, and herein, we address the issue of whether or not C3S3 can effect the formation of imides, a class of compound that is generally prepared via condensation of acid anhydrides and primary amines (Scheme 1).12 Cocrystals are accessible via solvent-drop grinding, i.e., two or more solid cocrystal formers milled in the presence of a small amount of solvent.6,13 A selected group of anhydrides and primary amines were investigated to determine the following: if they form cocrystals via solvent-drop grinding under ambient conditions14 and if the ground mixtures so obtained can be converted to imides simply by applying heat. The majority of reactants studied were observed to form imides after heating, but it was not always possible to isolate a cocrystal. Indeed, our CSD survey reveals no previous examples of cocrystals between amines and anhydrides.7 However, two combinations of cocrystal formers were isolated as cocrystals that facilitate high-yield, low-waste formation of imides. Scheme 1. Condensation Reaction between Acid Anhydrides and Primary Amines, Facilitated by Cocrystal Formation
Figure 1. Condensation reaction of NTCDA and MNA, which proceeds via cocrystal 1 to produce the dimide 2. 1 can be isolated as a powder from solvent-drop grinding involving DMF but can also be isolated as single crystals from 1:1 dioxane:MeCN.
As revealed by Figure 1, 1,4,5,8-naphthalenetetracarboxylic dianhydride (NTCDA) and 2-methyl-4-nitroaniline (MNA) form a 1:2 cocrystal, 1,15 which converts cleanly to diimide, 2, when heated at 180 °C for 3 h (75% yield attributed to volatility of MNA). 2 was recrystallized from DMF or DMSO, affording solvated single crystals of 2.16 1 can be prepared from solution, by solvent-drop grinding or solvent-drop grinding followed by heat and is sustained by charge transfer interactions between the aromatic rings of NTCDA and MNA, which are separated by centroid-plane distances of ca. 3.32 Å. The amino moieties form infinite chains along the b-axis via amine-nitro hydrogen bonds (NH‚‚‚O, 2.946 Å). The purple color exhibited by 1 contrasts with the starting materials (pale yellow) and product (orange) and is indicative of charge transfer17 (the solid-state UV-vis spectrum of 1 exhibits a broad band at ca. 600 nm). The carbonyl carbon atom of the NTCDA and amino nitrogen atom of the MNA are separated by only 3.42 Å, i.e. well within the 4.2 Å limit of the topochemical postulate.18 Interestingly, solvent-drop grinding with other solvents affords mixtures of NTCDA and MNA. However, heating of these mixtures at 130 °C (i.e., above the melting point of MNA) results in immediate formation of 1 and additional heating at 180 °C for 3 h affords 2. DSC analysis of 1 suggests that the condensation occurs at ca. 160 °C. These observations suggest that the formation of 1 is a key step for facilitating or even controlling the condensation process. NTCDA and 3-aminobenzoic acid (ABA) also form a purple cocrystal via solvent-drop grinding with DMF. However, the cocrystal undergoes condensation to the corresponding diimide under ambient conditions and attempts at crystallization resulted in the formation of the less-reactive 1,4-dioxane solvate of the cocrystal,19 3 (Figure 2). The ABA molecules in 3 form centrosymmetric dimers via NH‚‚‚OdC(3.066 Å) hydrogen bonds, which link to additional dimers through 1,4-dioxane molecules (OH‚‚‚O, 2.643 Å) to generate chains that stack along the a-axis. NTCDA molecules stack in between the ABA dimers with a plane-centroid distance
10.1021/cg0701729 CCC: $37.00 © 2007 American Chemical Society Published on Web 03/15/2007
Communications
Crystal Growth & Design, Vol. 7, No. 4, 2007 617
Figure 2. Condensation reaction of ABA and NTCDA via the cocrystal solvate 3, which dehydrates to generate 4.
of 3.13 Å. The solid-state UV-vis spectrum of 3 exhibits a broad band at 550 nm, consistent with charge transfer. The shortest distance between the amine nitrogen atoms and the carbon atoms of the carbonyl moieties is 3.14 Å. Therefore, 3 also obeys the topochemical postulate and converts to diimide 4 after heating for 24 h at 200 °C (99% yield).20 The different reactivities of 1, 3, and the solvate of 3 are presumably an artifact of crystal packing. Whereas C3S3 can be confirmed only in the two reactions for which isolation of an anhydride-amine cocrystal occurred, conversion of solvent-ground anhydride/amine mixtures to imides was a more general occurrence. Solvent-grinding followed by heating therefore appears to represent a feasible and possibly general methodology for preparation of imides. We plan to further explore methodologies for isolation of cocrystals and address other condensation reactions in order to determine if C3S3 can play a more general role in solid-state organic synthesis. Acknowledgment. We gratefully acknowledge Transform Pharma for financial support of this research. Supporting Information Available: X-ray crystallographic information in CIF format; FTIR, XPD, DSC, and UV-vis in PDF format. This material is available free of charge via the Internet at http://pubs.acs.org.Cocrystalcontrolled solid-state synthesis (C3S3) of imides occurs via heating of cocrystals formed between anhydride and aromatic amine cocrystal formers.
References Almarsson, O ¨ .; Zaworotko, M. J. Chem. Commun. 2004, 17, 1889. Wo¨hler, F. Annalen 1844, 51, 153. Allen, F. H. Acta Crystallogr., Sect. B 2002, 58, 380. (a) Vishweshwar, P.; McMahon, J. A.; Bis, J. A.; Zaworotko, M. J. J. Pharm. Sci. 2006, 95, 499. (b) Li, Z. J.; Abramov, Y.; Bordner, J.; Leonard, J.; Medek, A.; Trask, A. V. J. Am. Chem. Soc. 2006, 128, 8199-8210. (c) Remenar, J. F.; Morissette, S. L.; Peterson, M. L.; Moulton, B.; MacPhee, J. M.; Guzma´n, H. R.; Almarsson, O ¨ . J. Am. Chem. Soc. 2003, 25, 8456. (d) Childs, S. L.; Chyall, L. J.; Dunlap, J. T.; Smolenskaya, V. N.; Stahly, B. C.; Stahly, G. P. J. Am. Chem. Soc. 2004, 126, 13335. (5) Anastas, P. T.; Warner, J. C. Green Chemistry: Theory and Practice; Oxford University Press: New York, 1998. (6) Shan, N.; Toda, F.; Jones, W. Chem. Commun. 2002, 20, 2372. (7) 1642 co-crystals are reported in the Cambridge Structural Database versus 10 575 solvates; version 5.27 (May 2006) 3D coordinates, R < 0.075, no ions, organics only. (1) (2) (3) (4)
(8) (a) Tanaka, K. SolVent-Free Organic Synthesis; Wiley: New York, 2003. (b) Kaupp, G. Top. Curr. Chem. 2005, 254, 95. (c) Tanaka, K.; Toda, F. Chem. ReV. 2000, 100, 1025. (9) (a) MacGillivray, L. R.; Reid, J. L.; Ripmeester, J. A. J. Am. Chem. Soc. 2000, 122, 7817. (b) Fowler, F. W.; Lauher, J. W. J. Phys. Org. Chem. 2000, 13, 850. (10) Etter, M. C.; Frankenbach, G. M.; Bernstein, J. Tetrahedron Lett. 1989, 30, 3617. (11) Foxman, B. M.; Sandor, R. B. Tetrahedron 2000, 56, 6805. (12) Solomons, T. W. G.; Fryhle, C. B. Organic Chemistry; John Wiley and Sons: Hoboken, NJ, 2003. (13) (a) Trask, A. V.; van de Streek, J.; Motherwell, W. D. S.; Jones, W. Cryst. Growth Des. 2005, 5, 2233. (b) Bis, J. A.; Vishweshwar, P.; Middleton, R. A.; Zaworotko, M. J. Cryst. Growth Des. 2006, 6, 1048. (14) Anhydrides: NTCDA, pyromellitic anhydride, maleic anhydride, phthalic anhydride, 3,3′,4,4′-biphenyltetracarboxylic dianhydride, 1,8naphthalic anhydride, 3,4,9,10-perylenetetracarboxylic anhydride. Amines: MNA, ABA, melamine, 1,4-phenylenediamine, 1,5-naphthalenediamine, 1-adamantylamine, triphenylmethylamine. In a typical reaction, a stoichiometric amount of anhydride and amine (total weight of 100 mg) and 20 µL of solvent (chloroform, cyclohexane, DMSO, DMF, ethyl acetate, methanol, toluene, water) were handground using an agate pestle-and-mortar for 4 min. Products were analyzed by powder X-ray diffraction and solid-state FTIR spectroscopy. (15) Crystal data for 1: monoclinic, space group P1h, a ) 8.307(2) Å, b ) 8.894(2) Å, c ) 9.179(2) Å, R ) 110.230(4)°, β ) 103.873(5)°, γ ) 97.109(5)°, V ) 601.6(3) Å3, Z ) 1, Fcalcd ) 1.58 mg/m3, T ) 100 K, µ ) 0.123 mm-1, 2412 reflections measured, 1923 independent reflections, [I > 2σ(I)], R1 ) 0.0788, wR2 ) 0.2227, crystal size: 0.10 × 0.08 × 0.04 mm3. (16) Crystal data for 2‚DMF: triclinic, space group P1h, a ) 11.405(2) Å, b ) 12.457(2) Å, c ) 15.885(3) Å, R ) 90.911(3)°, β ) 100.914(3)°, γ ) 115.482(2)°, V ) 1988.2(6)Å3, Z ) 1, Fcalcd ) 1.47 mg/ m3, T ) 100 K, µ ) 0.110 mm-1, 7493 reflections measured, 6138 independent reflections, [I > 2σ(I)], R1 ) 0.1060, wR2 ) 0.2653, crystal size: 0.20 × 0.15 × 0.10 mm3. Crystal data for 2‚DMSO: triclinic, space group P1h, a ) 11.309(3) Å, b ) 12.204(4) Å, c ) 16.403(5)Å, R ) 89.651(6)°, β ) 77.343(5)°, γ ) 64.623(6)°, V ) 1985.8(10) Å3, Z ) 1, Fcalcd ) 1.48 mg/m3, T ) 100 K, µ ) 0.161 mm-1, 8259 reflections measured, 6472 independent reflections, [I > 2σ(I)], R1 ) 0.0915, wR2 ) 0.2108, crystal size: 0.10 × 0.08 × 0.05 mm3. (17) Kuroda, R.; Higashiguchi, K.; Hasebe, S.; Imai, Y. Cryst. Eng. Comm. 2004, 6, 463. (18) Schmidt, G. M. J. Pure Appl. Chem. 1971, 27, 647. (19) Crystal data for 3‚1,4-dioxane: triclinic, space group P1h, a ) 7.006(2) Å, b ) 9.990(3) Å, c ) 10.190(4) Å, R ) 83.081(7)°, β ) 81.603(7)°, γ ) 76.866(8)°, V ) 684.2(4) Å3, Z ) 1, Fcalcd ) 1.530 mg/m3, T ) 100 K, µ ) 0.119 mm-1, 2810 reflections measured, 2188 independent reflections, [I > 2σ(I)], R1 ) 0.0677, wR2 ) 0.1546, crystal size: 0.30 × 0.09 × 0.07 mm3. (20) 4 was recrystallized from pyridine as the disordered 0.5 pyridine solvate of the 1:2 complex of 4 with pyridine. Crystal data for 4‚ 2.5pyridine: triclinic, space group P1h, a ) 77.1569(14) Å, b ) 8.1962(17) Å, c ) 15.448(3) Å, R ) 98.287(4)°, β ) 102.202(4)°, γ ) 96.564(4)°, V ) 866.6(3) Å3, Z ) 1, Fcalcd ) 1.39 mg/m3, T ) 100 K, µ ) 0.099 mm-1, 3645 reflections measured, 2852 independent reflections, [I > 2σ(I)], R1 ) 0.0869, wR2 ) 0.2019, crystal size 0.30 × 0.20 × 0.10 mm3.
CG0701729
