USF Libraries
USF Digital Collections

Crystal engineering of nutraceutical cocrystals

MISSING IMAGE

Material Information

Title:
Crystal engineering of nutraceutical cocrystals
Physical Description:
Book
Language:
English
Creator:
Aboarayes, Dalia A
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Supramolecular chemistry
Cambridge structural database
Hydrogen bond
Resveratrol
Polymorphism
Dissertations, Academic -- Chemistry -- Masters -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: The work presented herein focus upon crystal engineering of nutraceutical cocrystals. Cocrystals are considered unique solid dosage form which has many advantages over other traditionally known solid forms. Furthermore, cocrystals have proven to improve stability, solubility and bioavailability of Active Pharmaceutical Ingredient (API) as shown in the case of carbamazepine and other APIs in previous studies. Crystal engineering is commonly used to design new solid forms based on the bases of supramolecular chemistry. In this study, crystal engineering based on intensive Cambridge Structural Database (CSD) analysis used to predict and design new cocrystals of targeted nutraceuticals. Two nutraceuticals were selected for this study; resveratrol and citric acid. The rationale behind selecting resveratrol was to improve its solubility and, accordingly, bioavailability.On the other hand, citric acid is known as a highly soluble and safe nutraceutical, and thus it can be used as a coformer. Five new cocrystals were prepared and characterized using a variety of techniques that include single crystal X-ray diffraction (XRD), powder X-ray diffraction (PXRD), differential scanning calorimetry (DSC), FT-IR, and thermo-gravimetric analysis (TGA). Most of the reported cocrystals were obtained using different techniques; solvent slow evaporation, mechanichemical approach, slurry, and from melt. Moreover, dissolution test has been performed on resveratrol and two of its cocrystals, using UV-vis spectrophotometer, where the data demonstrate that through cocrystallization with different cocrystal formers, solubility of resveratrol could be greatly modified, and further controlled. The polymorphism phenomenon is encountered, and accordingly addressed, herein where four novel polymorphs were obtained during cocrystallization attempts.Polymorphism has a significant importance in industry, in general, and in pharmaceutical industry, in particular, due to the vast differences in physical properties of polymorphs. Furthermore, the study of polymorphism provides valuable information essential to understand how different crystal forms are attained.
Thesis:
Thesis (M.S.)--University of South Florida, 2009.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Dalia A. Aboarayes.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 123 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002317560
oclc - 660248816
usfldc doi - E14-SFE0003061
usfldc handle - e14.3061
System ID:
SFS0027377:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200373Ka 4500
controlfield tag 001 002317560
005 20100902183527.0
007 cr cnu|||uuuuu
008 100902s2009 flua ob 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003061
035
(OCoLC)660248816
040
FHM
c FHM
049
FHMM
090
QD31.2 (Online)
1 100
Aboarayes, Dalia A.
0 245
Crystal engineering of nutraceutical cocrystals
h [electronic resource] /
by Dalia A. Aboarayes.
260
[Tampa, Fla] :
b University of South Florida,
2009.
500
Title from PDF of title page.
Document formatted into pages; contains 123 pages.
502
Thesis (M.S.)--University of South Florida, 2009.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
520
ABSTRACT: The work presented herein focus upon crystal engineering of nutraceutical cocrystals. Cocrystals are considered unique solid dosage form which has many advantages over other traditionally known solid forms. Furthermore, cocrystals have proven to improve stability, solubility and bioavailability of Active Pharmaceutical Ingredient (API) as shown in the case of carbamazepine and other APIs in previous studies. Crystal engineering is commonly used to design new solid forms based on the bases of supramolecular chemistry. In this study, crystal engineering based on intensive Cambridge Structural Database (CSD) analysis used to predict and design new cocrystals of targeted nutraceuticals. Two nutraceuticals were selected for this study; resveratrol and citric acid. The rationale behind selecting resveratrol was to improve its solubility and, accordingly, bioavailability.On the other hand, citric acid is known as a highly soluble and safe nutraceutical, and thus it can be used as a coformer. Five new cocrystals were prepared and characterized using a variety of techniques that include single crystal X-ray diffraction (XRD), powder X-ray diffraction (PXRD), differential scanning calorimetry (DSC), FT-IR, and thermo-gravimetric analysis (TGA). Most of the reported cocrystals were obtained using different techniques; solvent slow evaporation, mechanichemical approach, slurry, and from melt. Moreover, dissolution test has been performed on resveratrol and two of its cocrystals, using UV-vis spectrophotometer, where the data demonstrate that through cocrystallization with different cocrystal formers, solubility of resveratrol could be greatly modified, and further controlled. The polymorphism phenomenon is encountered, and accordingly addressed, herein where four novel polymorphs were obtained during cocrystallization attempts.Polymorphism has a significant importance in industry, in general, and in pharmaceutical industry, in particular, due to the vast differences in physical properties of polymorphs. Furthermore, the study of polymorphism provides valuable information essential to understand how different crystal forms are attained.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
590
Advisor: Michael J. Zaworotko, Ph.D.
653
Supramolecular chemistry
Cambridge structural database
Hydrogen bond
Resveratrol
Polymorphism
690
Dissertations, Academic
z USF
x Chemistry
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.3061



PAGE 1

Crystal Engineering of Nutraceutical Cocrystals by Dalia A. Aboarayes A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Chemistry College of Arts and Sciences University of South Florida Major Professor: Michael J. Zaworotko, Ph.D. Mohamed Eddaoudi, Ph.D. Abdul Malik, Ph.D. Date of Approval: July 17, 2009 Keywords: Supramolecular Chemistry, Cambri dge Structural Data base, Hydrogen Bond, Resveratrol, Polymorphism Copyright 2009, Dalia A. Aboarayes

PAGE 2

Dedication For my husband, family and friends

PAGE 3

Acknowledgements I would like to thank my advisor, Pr ofessor Michael J. Zaworotko, for the opportunity to conduct research under his su pervision, and for his advice and guidance throughout the Graduate Program. I would also like to thank Dr. Mohame d Eddaoudi and Dr. Abdul Malik, my committee members, for their helpful comments and encouragements. In addition, I would like to acknowledge all members of my research group, as well as Faculty and Staff of the Chemistry Depa rtment of University of South Florida, for their friendly accommodation. At last, I would like to express my d eepest thanks to my husband, family, and friends who constantly supported me throughout the period of studies.

PAGE 4

i Table of Contents List of Tables v List of Figures vi Abstract xi Chapter 1. Introduction 1 1.1. Crystal Engineering 1 1.2. Cambridge Structural Database 4 1.3. Cocrystals 6 1.4. Crystal Forms of Drugs and Biopha rmaceutical Classification System 10 Chapter 2. Nutraceutical Cocrystals 13 2.1. Introduction 13 2.2. Resveratrol 14 2.2.1. Introduction 14 2.2.2. CSD Analysis 18 2.2.3. Experimental 22

PAGE 5

ii 2.2.3.1. Resveratrol-caprolactam cocrystals (DA182901 and DA182902) 22 2.2.3.2. Resveratrol-flavone cocrystals (RESVONE 01 and RESVONE02) 23 2.2.3.3. Resveratrol and 4,4'-di pyridyl cocrystal(DA4) 24 2.2.4. Results and discussion 25 2.2.4.1. Resveratrol-caprolactam cocrystals (DA182901 and DA182902) 25 2.2.4.2. Resveratrol-flavone cocrystals (RESVONE 01 and RESVONE 02) 29 2.2.4.3. Resveratrol and 4, 4’-dip yridyl cocrystal (DA4) 32 2.2.5. Dissolution test 35 2.2.6. Conclusion 39 2.3. Citric acid 41 2.3.1. Introduction 41 2.3.2. CSD analysis 42 2.3.2.1. CSD statistics for acids 43 2.3.2.2. CSD statistics for alcohols 48 2.3.3. Experimental 52 2.3.3.1. Citric acidiso -nicotinamide cocrystal (DA005) 52 2.3.3.2. Dihydrate of citric acidiso -nicotinic acid cocrystal (DA02) 53 2.3.4. Results and discussion 53 2.3.4.1. Citric acidiso -nicotinamide cocrystal (DA005) 53 2.3.4.2. Dihydrate of citric acidiso -nicotinic acid cocrystal (DA02) 55 2.3.5. Conclusion 59 2.4. Flavanone single crystal 59

PAGE 6

iii 2.4.1. Experimental 60 2.4.2. Result and discussion 60 2.4.3. Conclusion 61 Chapter 3. Polymorphism 62 3.1. Introduction 62 3.1.1. Significance of polymorphism to pharmaceutical industry 64 3.1.2. Polymorphism and patents in pharmaceutical industry 69 3.1.3. Conformational polymorphism 72 3.1.4. Disappearing polymorphs 73 3.2. Experimental 75 3.2.1. Salicylamide polymorph II (DA008) 75 3.2.2. Iso -nicotinamide polymorph III (DA16712) 75 3.2.3. Trans -resveratrol polymorph II (DA2116) 76 3.2.4. Citric acid monohydrate polymorph II (C-azod) 76 3.3. Results and Discussion 77 3.3.1. Salicylamide polymorph II (DA008) 77 3.3.2. Iso -nicotinamide polymorph III (DA16712) 84 3.3.3 Trans -resveratrol polymorph II (DA2116) 88 3.3.4. Citric acid monohydrate polymorph II (C-azod) 91 3.4. Conclusion 93 References 94

PAGE 7

iv Appendices 100 Appendix 1. Experimental data 101 Appendix 2. Crystallographic data 117 Appendix 3. Comparing crystal structures of new polymorphs to that of known polymorphs 120

PAGE 8

v List of Tables Table 2.1. Different sources of resveratrol 15 Table 2.2. CSD statistics for phenols 18 Table 2.3. CSD statistics for acids 43 Table 2.4. CSD statistics for alcohols 48 Table 2.5. Comparison of the C-N-C angle of the pyridine and the C-O distances of the carboxylic acid Moieties in DA02 molecules (left) and iso -nicotinic acid (right) 58

PAGE 9

vi List of Figures Figure 1.1. Supramolecular homosynthons (left) and supramolecular heterosynthons (right) 3 Figure 1.2. Growth of the CSD since 1972 5 Figure 1.3. Various possible cr ystalline forms for an API: (a) pure API; (b) polymorph of API; (c) clathrate hydrate/solvate of API; (d) hydrate/solvate of API; (e) salt of API; (f ) pharmaceutical cocrystal 7 Figure 1.4. Crystal structure of qui nhydrone (CSD refcode: QUIDON) 7 Figure 1.5. The Hoogsteen base pairing in the cocrystal of 9-methyladenine and 1-methylthymine (CSD refcode: MTHMAD) 8 Figure 1.6. Comparing average plasma concentrations versus time of carbamazepine in Tegretol and that in carbamazepine: saccharin (cocrystal 1) 9 Figure 1.7. Comparing average plasma con centrations versus time of 2-[4-(4chloro-2-fluorophenoxy) phenyl] pyrim idine-4-carboxamide and that in glutaric acid cocrystal 10 Figure 1.8. The four classes of APIs based on BCS 11 Figure 2.1. PubMed search for word ‘resver atrol’ shows exponential increase in the number of resveratrol rela ted publications over years 16 Figure 2.2. Histogram fo r OHOH homosynthon 19 Figure 2.3. Histogram for OHCO primar y amide (left) and that for OHNH2 primary amide (right) 20 Figure 2.4. Histogram for OHCO secondary amide (left) a nd that for OHNH secondary amide (right) 21

PAGE 10

vii Figure 2.5. Histogram for OHCO 21 Figure 2.6. Histogram for OHNarom 22 Figure 2.7. Single crystal of resveratro l-caprolactam cocrystal (DA1829) 23 Figure 2.8. Single crystals of resverat rol-flavone cocrys tal (RESVONE01) 24 Figure 2.9. Single crystals of resver atrol and 4,4'-dipyridyl (DA4) 25 Figure 2.10. The molecular structure of trans -resveratrol molecule shows labeled rings and phenolic OH groups 25 Figure 2.11. Comparing the asymmetric unit of two polymorphs of 1:2 cocrystal of trans -resveratrol and -caprolactam DA182901 (left) with that of DA182902 (right) 26 Figure 2.12. Comparison of hydroge n bonding in DA182901 (up) and DA182902 (down) 27 Figure 2.13. Molecular structures of trans -resveratrol molecule (left), and that of flavone (right) 29 Figure 2.14. Comparing the asymmetric unit of two polymorphs of 1:2 cocrystal of trans -resveratrol and flavone, RESVON E01 (left) with that of RESVONE02 (right) 29 Figure 2.15. Comparison of hydrogen bondi ng in RESVONE01 (left) and RESVONE02 (right) 30 Figure 2.16. Comparison of hydrogen bondi ng pattern in RESVONE01 (left) and RESVONE02 (right) 30 Figure 2.17. The asymmetric unit of DA4 32 Figure 2.18. Two sets of tetrameric units in DA4 33 Figure 2.19. Spiral tapes of DA4 33 Figure 2.20. [2+2] photodimerization of 2 (4,4-bpe) and 2(resorcinol) 34 Figure 2.21. Dissolution profiles of Fluoxe tine HCl and its cocrystals 36 Figure 2.22. Dissolution profiles of itraconazo le and its cocrysta ls (itraconazole presented by white) 36

PAGE 11

viii Figure 2.23. Dissolution profiles of trans -resveratrol and its cocrystals 38 Figure 2.24. Molecular struct ure of citric acid 41 Figure 2.25. Histogram for COOHNarom 44 Figure 2.26. Histogram for COOHCl– 44 Figure 2.27. Histogram for COOHP=O 45 Figure 2.28. Histogram for OHCOOH (left) and that for OHHOOC (right) 46 Figure 2.29. Histogram for COOHCONH2 (left) and CONH2COOH (right) 46 Figure 2.30. Histograms for carboxylic acid dimers 47 Figure 2.31. Histogram for COOHCO 48 Figure 2.32. Histogram for OHCl– 49 Figure 2.33. Histogram for OHCONH2 (left) and that of OHNH2CO (right) 50 Figure 2.34. Histogram for OHP=O 50 Figure 2.35. Histogram for OHNarom 51 Figure 2.36. Histogram for OHCO 51 Figure 2.37. Single crysta ls of citric acidiso -nicotinamide cocrystal (DA005) 52 Figure 2.38. Single crystals of citric acidisonicotinic acid dihydrate cocrystal (DA02) 53 Figure 2.39. The asymmetric unit of DA005 53 Figure 2.40. Crystal packing in DA005 54 Figure 2.41. The asymmetric unit of DA02 55 Figure 2.42. Crystal packing in DA02 56 Figure 2.43. Comparison of the C-N-C a ngle of the pyridine and the C-O distances of the carboxylic acid Moieties in DA02 molecules (left) and iso -nicotinic acid (right) 57

PAGE 12

ix Figure 2.44. The molecular structure of flavanone 59 Figure 2.45. The asymmetric unit of flavanone 60 Figure 2.46. Hydrogen bonding in flavanone 61 Figure 2.47. Herringbone pattern in flavanone 61 Figure 3.1. Allotropes of carbon diamond (lef t), graphite (middle), and fullerene (right). 62 Figure 3.2. The molecular structure of a red polymorph of ROY (CSD refcode: CAXMEH03) 63 Figure 3.3. Paracetamol polymorphs, (left) form I showing puckered hydrogenbonded sheets, and (right) form II showing flat sheets. (CSD refcodes: HXACAN01 and HXACAN08, respectively) 65 Figure 3.4. Chloramphenicol urinary excre tion rates after oral administer of chloramphenicol palmitate with varying percents of polymorph B: M, 0%; N, 25%; O, 50%; P, 75%; and L, 100% 66 Figure 3.5. Blood serum levels of chloram phenicol after oral administer of chloramphenicol palmitate with varying percents of polymorph B 67 Figure 3.6. Molecular structure of ranitidine 70 Figure 3.7. General scheme shows di fferent types of polymorphs 73 Figure 3.8. Molecular struct ure of salicylamide 75 Figure 3.9. Molecular structure of iso -nicotinamide 75 Figure 3.10. Molecular structure of trans -resveratrol 76 Figure 3.11. Comparison of the dihedral angle in DA008 molecules A and B (left) with that of SALMID (right) 78 Figure 3.12. Comparison of hydrogen bondi ng in DA008 (up) to that in SALMID (down) 79 Figure 3.13. Illustration of the sandwich herringbone pattern in DA008 (left) and the herringbone pattern in SALMID (right) 80 Figure 3.14. Illustration of face to face inte ractions in DA008 (left) and that in SALMID (right) 80

PAGE 13

x Figure 3.15. Histogram for the dihedral angle of polymorphs of primary aromatic amides 81 Figure 3.16. Histogram for the dihedral a ngle of primary aromatic amides 82 Figure 3.17. Histogram for primary aromatic amide dimers 83 Figure 3.18. Histogram for primary aromatic amide catemers 83 Figure 3.19. Comparison of the dihedral angle in DA16712 (left), EHOWIH01 (middle), and molecules A a nd B in EHOWIH02 (right) 85 Figure 3.20. Comparison of hydrogen bonding in DA16712 (up), EHOWIH01 (middle), and in assemblies composed of molecules A and B respectively in EHOWIH02 (down) 86 Figure 3.21. Illustration of the herringbone pattern in DA16712 (left), sheets of EHOWIH01 (middle), and the he rringbone pattern in EHOWIH02 (right) 87 Figure 3.22. Comparison of hydrogen bonding in DA2116 (left) and that of DALGON (right) 90 Figure 3.23. Illustration of the packing in DA2116 (left) and that in DALGON (right) 90 Figure 3.24. Comparison of hydrogen bonding in C-azod (left) and that of CITARC (right) 92

PAGE 14

xi Crystal Engineering of Nutraceutical Cocrystals Dalia A. Aboarayes ABSTRACT The work presented herein focus upon crystal engineering of nutraceutical cocrystals. Cocrystals are considered unique solid dosage form which has many advantages over other traditi onally known solid forms. Furthermore, cocrystals have proven to improve stability, solubility a nd bioavailability of Active Pharmaceutical Ingredient (API) as shown in the case of carbamazepine and other APIs in previous studies. Crystal engineering is commonly used to design new solid forms based on the bases of supramolecular chemistry. In this study, crystal engineering based on intensive Cambridge Structural Database (CSD) analysis used to predict and design new cocrystals of targeted nutraceuticals. Two nutraceuticals were selected for this study; resveratrol and citric acid. The rationale behi nd selecting resveratrol was to improve its solubility and, accordingly, bioavailability. On the other ha nd, citric acid is known as a highly soluble and safe nutraceutical, and thus it can be used as a coformer. Five new cocrystals were prepared and characterized using a variety of techniques that include single crystal X-ray

PAGE 15

xii diffraction (XRD), powder X-ray diffraction (P XRD), differential scanning calorimetry (DSC), FT-IR, and thermo-gravimetric analysis (TGA). Most of the reported cocrystals were obtained using different techniques; solvent slow evaporation, mechanichemical approach, slurry, and from melt. Moreover, dissolution test has been performed on resveratrol and two of its cocrystals, us ing UV-vis spectrophotometer, where the data demonstrate that through cocrystallization with different cocrystal formers, solubility of resveratrol could be greatly m odified, and further controlled. The polymorphism phenomenon is encount ered, and accordingly addressed, herein where four novel polymorphs were obt ained during cocrystallization attempts. Polymorphism has a significant importance in industry, in general, and in pharmaceutical industry, in particular, due to the vast differences in phys ical properties of polymorphs. Furthermore, the study of polymorphism provi des valuable information essential to understand how different crystal forms are attained.

PAGE 16

1 Chapter 1. Introduction 1.1. Crystal Engineering The term “crystal engineering” was first introduced by Pepinsky1 in 1955, but it was not widely employed until the pioneering works of Schmidt2 in topochemical reactions of cinnamic acids. Later, Desira ju defined crystal engineering as “the understanding of intermolecular in teractions in the context of crystal packing and in the utilization of such understanding in the desi gn of new solids with desired physical and chemical properties”.3 Thus, crystal engineering is based on the concept of forming new solids with desired properties, using nonc ovalent interactions, such as hydrogen or ionic bonding. Understanding of these interactions can be employed to obtain novel molecular entities based on the elements of design in supramolecular chemistry. The early beginnings of supramolecular ch emistry started with works of Johannes Diderik van der Waals in 1873. However, it wa s not until 1987 when the field gained a lot of attention mainly due to the 1987 Nobe l Prize in chemistry awarded to Jean-Marie Lehn, Donald J. Cram, and Charles J. Pedersen as a result of their works in supramolecular chemistry. Supramolecular chemis try is unique in th at it relies only on relatively weak forms of inte ractions. Thus, no covalent in teractions are involved in supramolecular chemistry. Therefore it is commonly known as “chemistry beyond the

PAGE 17

2 molecules”. Of particular interest among non covalent interactions to supramolecular chemistry are hydrogen bonding, electrostatic in teractions, metal coordination, and van der Waals interactions. The large range of non-covale nt intermolecular interactio ns include electrostatic interactions (ion-ion, ion-di pole and dipole-dipole intera ctions), coordination bonds, hydrogen bonds, halogen bonds, stacking and Van der Waals interactions. The concept of molecular-recognition could be considered to have emerged in 1890 when Fischer suggested "lock and key" model for enzyme-substrate interactions. However, the definition of hydrogen bonding could be traced back to Latimer and Rodebush in 1920.4One important aspect of the weak in termolecular interactions involved in supramolecular chemistry is the reversibilit y of the relatively w eak bonding interactions involved. This allows molecules to self-assemb le and further to self-correct the overall structure. The term Self-assembly refers to processe s in which a disordered system of preexisting components forms an organized structur e or pattern as a consequence of specific, local interactions among the components th emselves, without ex ternal direction. The molecular recognition between tw o molecules is accomplished through complimentary interactions of specific func tional groups in the two molecules. Such specific complimentary interacting parts of th e two molecules are usually referred to as supramolecular synthons. The term “s ynthon” was introduced by Corey in 1967.5

PAGE 18

3 A broader definition of supramolecular synthons made by Desiraju et al. where supramolecular synthons could be visualized as “structural units within supermolecules which can be formed and/or assembled by known or conceivable synthetic operations involving intermolecu lar interactions”.6 Supramolecular synthons can then be classified to either supramolecular homosynthons or supramolecular heterosynthons7,8. Supramolecular homosynthons are those synthons formed between the same functionalities, while supramolecular homosynthons are those formed between the different functionalities. Examples of the two types are shown in Figure 1.1, where two functionalities are complementary to each other and can be a ssembled through two points of recognitions to form supramolecular homosynthons or heterosynthons. Figure 1.1. Supramolecular homosynt hons (left) and supramolecular heterosynthons (right)9 Many case studies of supramolecular ho mosynthons or heterosynthons were recently presented in research. Examples of supramolecular homosynthons can be seen in carboxylic acid dimers10 and amide dimers.11Supramolecular homosynthons can be seen in carboxylic acidamide12, carboxylic acidNarom13, and phenolNarom.14

PAGE 19

4 Years ago, Feynman15 asked his famous question “What would the properties of materials be if we could really arrange th e atoms the way we want them?”. However, advances in field of crysta l engineering might be the answer to that question. In this context, we have to note that cocrystals are amenable for design using crystal engineering principles, which are based on supramolecular chemistry. The research presented herein is ba sed on exploiting the concept of donoracceptor molecular recognition of complementary functional groups to form novel supermolecules, namely; nutraceutical cocrystals. 1.2. Cambridge Structural Database As we discussed earlier, unde rstanding of supramolecular interactions is the key to crystal engineering of novel crystalline entities which could be achieved through careful inspection of exis ting hydrogen bond interactions in previously reported supermolecules, hence facilitating the supramolecular retrosynthesis.16 18Cambridge Structural Database (CSD), the product of Cambridge Crystallographic Da ta Center (CCDC)19, is invaluable tool to study a wide collection of reported crystal structures that enables id entification of the o ccurrence frequency of distinct supramolecular synthons. CCDC began its activities in 1965, led by Dr. Olga Kennard at Cambridge University, as a depository of crystal stru ctures for organic and metal-organic small molecules studied by X-ray or neutron diffr action techniques. Si nce its beginning, the

PAGE 20

5 number of reported crystalline structures is exponentially increasing over the years. The last reported number of entries is 423752, according to August 2008 update. Figure 1.2. Growth of the CSD since 197220 Thus, CCDC is the center for archiving th ese crystal structures which could be easily visualized, classified and analyzed through an integral co llection of software application impeded into the CSD program This collection contains the following softwares: search and information retrie val (ConQuest), structure visualization (Mercury), numerical analysis (Vista ), and database creation (PreQuest). According to Kennard et al., about twenty years after the firs t release of CSD, “The systematic analysis of large numbers of related structures is a powerful research technique, capable of yielding results that could not be obtaine d by any other method”.21

PAGE 21

6 Thus, CSD is considered as an invaluable re search tool, which can be used for design and prediction of new crystalline entities. 1.3. Cocrystals The term cocrystal is used to desc ribe “a long known but little-studied”22 class of crystalline compounds. For years, many terms have been used to describe cocrystals. Some of these terms are: molecular compounds23, addition compounds24, organic molecular compounds25, molecular complexes26, solid-state complexes27, and heteromolecular crystals.28 Even recently, there is still a debate around the most appropriate name for this class of crystalline compounds.29,30 However, Dr. Zaworotko research group defines cocrystal as “a multiple component crystal in which all components are solid under ambient conditions when in their pure form. These components co-exist as a stoichiometric ratio of a target molecule or ion and a neutral molecular cocrystal former(s)”.22 This definition makes clear difference between cocrystals and salts; where in cocrystal at least one of the molecules is neutral, while in slats both molecules are ionized. The de finition also excludes hydrates and solvates.

PAGE 22

7 Figure 1.3. Various possible crystalline fo rms for an API: (a) pure API; (b) polymorph of API; (c) clathrate hydrate/solv ate of API; (d) hydrate/solvate of API; (e) salt of API; (f ) pharmaceutical cocrystal22 As we mentioned earlier, cocrystals are a long known class of crystalline compounds. This can be seen from works of Wohler31 in 1844, when he synthesized the first cocrystal between benzoquinone and hydroquinone. However, thecrystal structure was determined later by Sakurai32,33 in the 1960’s. Figure 1.4. Crystal structure of qui nhydrone (CSD refcode: QUIDON)

PAGE 23

8 The first scientific use of cocrysta ls was reported in work of Hoogsteen34,35, when he studied the base pairing in the cocrystal of 9-methyladenine and 1-methylthymine, in the contest of DNA base pairing. However, co crystals were popularized latter by work of Etter36. Figure 1.5. The Hoogsteen base pairing in the cocrystal of 9-methyladenine and 1methylthymine (CSD refcode: MTHMAD) The term “pharmaceutical cocrystals” is us ed to describe “A multiple component crystal in which at least one component is mo lecular and a solid at room temperature (the cocrystal former) and forms a supramolecula r synthon with a molecular or ionic API”8 Many pharmaceutical cocrystals are currently synthesized, where they represent a novel class of APIs, which has more advantages compared to traditionally known crystalline forms, such as; salts, solvates, hydrates, and polymorphs. However, pharmaceutical cocrystals have shown to improve physical properties of APIs, such as solubility and therefore bioavailability. Two APIs37 and two of their cocrystals were studied in animal models.

PAGE 24

9 Carbamazepine (Tegretol) is an anti-epileptic agent, which has many drawbacks38,39, such as; narrow therapeutic window, autoinduction of metabolism and dissolution-limited bioavailability, and high tendency to pol ymorphism. Tegretol and carbamazepine: saccharin cocrystals9 were tested in dog model to compare carbamazepine bioavailability in both. As shown in figure 1.6., it is found that the cocrystal shows higher plasma le vels than that of Tegretol. Figure 1.6. Comparing average plasma concentrations versus time of carbamazepine in Tegretol and that in carbamazepine: saccharin (cocrystal 1)37 Another case study40 was performed using 2-[4 -(4chloro-2-fluorophenoxy) phenyl] pyrimidine-4-carboxamide, a low sol ubility, sodium channel blocker. 2-[4-(4chloro-2-fluorophenoxy) phenyl] pyrimidine4-carboxamide and its glutaric acid

PAGE 25

10 cocrystal were tested in dog model. As show n in figure 1.7., it is found that the cocrystal shows higher plasma levels than that of pure API. Figure 1.7. Comparing average plasma concen trations versus time of 2-[4-(4chloro-2-fluorophenoxy) phenyl] pyrimidi ne-4-carboxamide and that in glutaric acid cocrystal 40 It can be concluded from the above discus sion that cocrystal is a promising class of crystalline drugs, which can afford some advantages over the other traditional forms. 1.4. Crystal Forms of Drugs and Biopharm aceutical Classification System Over the years, different routes were developed for drug administration, such as oral, topical, and parenteral. Orally administered drugs cons titute about 85% of the most sold drugs in the USA and Europe. Moreover, orally administered drugs are available in

PAGE 26

11 different solid or liquid dosage forms amongst which oral solid dosage forms are the most common among other available dosage forms. The active pharmaceutical ingredients (API s) in solid dosage forms could either be present as amorphous or crystalline forms. As shown in figure 1.3., there are different possible crystalline forms for an API such as : pure API, polymorph, hydrate, solvate, salt, or cocrystal. In all of the aforementioned cases, crystallinit y of API directly affects its physical properties, mainly sol ubility and dissolution rate. The Biopharmaceutical Classification Syst em (BCS) is a recently developed guidancewhich “correlates in vitro drug product dissolution and in vivo bioavailability”.41 The BCS predicts gastrointestinal dr ug absorption based on dissolution and permeability data. In this system, two parame ters, namely; solubility and permeability are used to classify oral drugs into four classes, as shown in figure 1.8. Figure 1.8. The four classes of APIs based on BCS

PAGE 27

12 Those two parameters are defined by FDA42 as following: a drug substance is considered HIGHLY SOLUBLE when the highest dose strength is soluble in < 250 ml water over a pH range of 1 to 7.5. Furthermor e, a drug substance is considered HIGHLY PERMEABLE when the extent of absorp tion in humans is determined to be > 90% of an administered dose based on mass-balance or in comparison to an intravenous reference dose. According to this classification system about 70% of commercially available drugs have inherently low aqueous solubilit y, which therefore limit their bioavailability. It becomes evident that drug bioavailability is mainly affected by its solubility, or more specifically, solubility of its solid form. Therefore, as cocrystal solid forms were demonstrated to enhance solubility of APIs cocrystal solid forms represent interesting targets for studies concerning enhanc ed bioavailability of solid APIs.

PAGE 28

13 Chapter 2. Nutraceutical Cocrystals 2.1. Introduction For many years, food has been used for more than its nutritional value. It is well known that ancient civilizations used herbal products as a remedy for many diseases. The term “pharmacognosy” was first used by Schmidt in 1811, which is derives from the Greek words pharmakon (drug), and gnosis or "knowledge". It is defined latter by The American Society of Pharmacognosy to desc ribe “the study of the physical, chemical, biochemical and biological properties of dr ugs, drug substances, or potential drugs or drug substances of natural origin as well as the search for new drugs from natural sources.43 However, formulation of a specific term to describe food or its active ingredients, which have medicina l values, was eagerly required. The term ‘Nutraceutical’ was coined by Dr. Stephen De Felice, founder and chairman of Foundation of Innova tion in Medicine, in 1976 to describe “food, or parts of food, that provide medical or health benef its, including the preven tion and treatment of disease”.44 It is believed that nutraceuticals can ha ve a rule in preventing and treating many diseases45 ; starting from common cold, weight pr oblems and ending with cardio vascular diseases and even cancer.46 Nutraceuticals include many food and food products, including vitamins, soy products, glucosamine, chondroitin, and many polyphenols and flavonoids (resveratrol,

PAGE 29

14 ellagic acid, and quercetin). The vast majority of nutraceuticals are extracted from plant origins; such as fruits, vegetables, roots, and rhizomes. However, many nutraceuticals are derived from animal origins; vitamins and amino acids. Nutraceuticals can be used as targets for cocrystal formation because many of them have major problems with solubility and bioavailability, and cocrystal can solve those problems, as discussed earlier in section 1.1.3. Furthermore, nutraceutical cocrystals are patentable as they meet the criteria required for patents.47 Nutraceuticals have a big market currently and annual growth rates of nutraceuticals are predicted to increase. This rate expected to be up to 25% for some products.44 However, nutraceutical cocrystals can offer an opportunity for exponential growth of nutraceuticals market. 2.2. Resveratrol 2.2.1. Introduction Resveratrol [ trans -3,5,4'-trihydroxystilbene] is a nutra ceutical, extracted mainly from skin of red grapes. It can also be obtaine d from other dietary sources as shown in table 2.1.

PAGE 30

15 Table 2.1. Different sources of resveratrol48 Source Trans -resveratrol concentration Red wines 0.1–14.3 mg/L White wines <0.1–2.1 mg/L Ports and sherries <0.1 mg/L Grapes 0.16–3.54 g/g Dry grape skins 24.06 g/g Red grape juices 0.50 mg/L White grape juices 0.05 mg/L Cranberry juice 0.2 mg/L Blueberries 32 ng/g Peanuts 0.02–1.92 g/g Resveratrol attracted a lot of attention when Jang et al .46 published their paper in 1997, reporting anticarcinogenic activity of resver atrol. Since that si gnificant discover, research on resveratrol biological activity ha s greatly increased, which is reflected as exponential increase in the number of scientif ic publications concer ning resveratrol, as seen in figure 2.1.

PAGE 31

16 Figure 2.1. PubMed search for word ‘resvera trol’ shows exponential increase in the number of resveratrol related publications over years48 However, medical evidences have shown that resveratrol can participate in preventing and treating many diseases48 such as; cancer, cardiovascular diseases, myocardial infarction, brain da mage…etc. It also plays an important role in supporting immunity and reducing stress. There are many mechanisms48 proposed to participate in resveratrol biological activity. Anticarcinogenic activity originates fr om its antioxidant pr operties, in addition to its ability to inhibit cyclooxygenase (COX1 and COX2), ornithine decarboxylase, and angiogenesis. Moreover, resver atrol activity against heart di seases originates from its antioxidant properties, in addi tion to its ability to inhibit platelet aggregation and its vasodilation effect. Furthermore, various sp ecies treated with resveratrol had shown

PAGE 32

17 lifespan extensions. Up to 70% increase in lifespan in some species was reported, but unfortunately this effect is unknown in mammals. However, administer of enough quantity of resveratrol is re quired to get those beneficial effects. This means that resveratro l has to reach to certai n blood level, i.e. it has to be bioavailable. Unfortunately, resver atrol bioavailability is low because of two factors. First, solubility of resveratrol is low as 0.019 mg/ml. Second, oral resveratrol undergoes first pass effect, thus, it is pr edisposed to metabolism by cytochrome P450. Hence, higher doses are required which has two main drawbacks. Toxic effects are reported upon administering 1 or more g per kg (body weight).49 Furthermore, estimated cost of administering enough doses is a bout $6,800 based on annual consumption of 2.7 kg. Interestingly, Baur et al .48 commented on this situation as following “Therefore, blocking the metabolism of resveratro l, developing analogues with improved bioavailability, or finding new, more potent compounds that mimic its effects will become increasingly important”. Hence, resveratrol is a valid target for cocrystallization and the following sections describe how cocrys tals of resveratrol were designed and the subsequent observed effect(s) on resveratrol solubility. However, CSD search for resveratro l reveals one crystal structure of transresveratrol, while no cocrystals were reported on CSD (A ugust 2008 update).

PAGE 33

18 2.2.2. CSD Analysis As we discussed earlier in section 1.2.2., CS D search can serve as a tool to predict and design new cocrystals. Therefore, deta iled CSD search conducted herein using Conquest v. 1.10, Aug 2008 update. Filters applied to the searches to limit the results according to the following criteria: 3D coordinates determined, R factor < 7.5%, no ions and only organics. The alcohols were exclude d from the search due to the difference between acidity of phenols and alcohols wh ich therefore affects their supramolecular interactions. The results are summarized in table 2.2. were raw data indicates results in presence of competitive hydrogen bonding groups wh ile refined data is that in absence of any competitive groups. Table 2.2. CSD statistics for phenols Functional group Raw data Refined data Bond distance range () mean () OHOH 1045/7069 (14.8%) 129/216 (59.7%) 2.504-3.069 2.817(3) Primary amide OHCONH2 OHNH2CO 26/81(32%) 29/81(35.8%) 6/6 (100%) 6/6 (100%) 2.607-3.497 2.765-3.162 2.944(57) 3.018(13) secondary amide OHCO OHNH 24/52 (46.1%) 23/52 (44.2%) 10/10 (100%) 7/10 (70%) 2.603-2.767 2.892-3.489 2.683(11) 3.271(47) OHCO 619/1653 (37.5%) 162/247 (65.6%) 2.525-3.098 2.776(4) OHNarom 338/595 (56.8%) 107/130 (82.3%) 2.515-3.114 2.750(3) CSD Conquest v. 1.10, Aug 2008 update, organics only, 3D coordinates, R<7.5%, no ions

PAGE 34

19 CSD search reveals 7069 entries with phe nolic OH. In presence of competitive groups, 1045 (14.8%) entries exhibit OH OH supramolecular homosynthon. The percentage increases to 59.7% in absen ce of competitive functional groups. The histogram for bond distance ra nge is shown in figure 2.2. and it reveals bond distance range of 2.504-3.069 (mean = 2.817(3) ). Figure 2.2. Histogram for OHOH homosynthon Furthermore, several CSD searches were car ried out to investigate supramolecular heterosynthons. CSD search for structures with both phenolic OH and primary amide reveals 81 entries. In presence of competitive groups, 26 (32%) entries exhibit OHCONH2 supramolecular heterosynthons while 29 (35.8%) entries exhibit OHNH2CO supramolecular heterosynthons. The pe rcentage increases to 100%, for both OHCONH2 and OHNH2CO, in absence of compe titive functional groups. The histogram for bond distance ra nge is shown in figure 2.3. and it reveals bond distance range of 2.607-3.497 (mean = 2.944(57) ) and 2.765-3.162 (mean = 3.018(13) ), for OHCONH2 and OHNH2CO, respectively.

PAGE 35

20 Figure 2.3. Histogram for OHCO primary amide (left) and that for OHNH2 primary amide (right) Another CSD search reveals 52 entries with both phenolic OH and secondary amide. In presence of competitive groups 24 (46.1%) entries exhibit OHCONH supramolecular heterosynthons while 23 (44.2%) entries exhibit OHNHCO supramolecular heterosynthons. The percen tages increase to 100% for OHCONH and 70% for OHNHCO, in absence of compe titive functional groups. The histogram for bond distance range is shown in figure 2.4. and it reveals bond distance range of 2.6032.767 (mean = 2.683(11) ) and 2.892-3.489 (mean = 3.271(47) ), for OHCONH and OHNH CO, respectively.

PAGE 36

21 Figure 2.4. Histogram for OHCO secondary amide (left) and that for OHNH secondary amide (right) To investigate supramolecular heteros ynthons between phenolic OH and aldehyde or ketone, carbonyl group is used as a repres entative for the two later groups. CSD search reveals 1653entries with both phenolic OH and carbonyl group. In presence of competitive groups, 619 (37.5%) entries exhibit OHCO supramolecular heterosynthons. The percentages increase to 65.6% in absence of competitive functional groups. The histogram for bond distance ra nge is shown in figure 2.5. and it reveals bond distance range of 2.525-3.098 (mean = 2.776(4) ). Figure 2.5. Histogram for OHCO

PAGE 37

22 Another CSD search reveals 595 entries with both phenolic OH and aromatic nitrogen, Narom. In presence of competitive groups, 338 (56.8%) entries exhibit OHNarom supramolecular heterosynthons. The per centages increase to 82.3% in absence of competitive functional groups. The hist ogram for bond distance range is shown in figure 2.6. and it reveals bond distance ra nge of 2.515-3.114 (mean = 2.750(3) ). Figure 2.6. Histogram for OHNarom 2.2.3. Experimental All the chemicals used during the preparation were obtained from commercial sources and used as received. All crystal lization experiments were conducted in an unmodified atmosphere and the solven ts were distilled prior to use. 2.2.3.1. Resveratrol-caprolactam co crystals (DA182901 and DA182902) 22.8mg (0.100 mmol) of resv eratrol and 22.6mg (0.200 mmol)of -caprolactam were dissolved in 5ml of hot acetone. The solution was allo wed to slowly evaporate at room temperature and pale yellow crystals of DA182901 were harvested after one day

PAGE 38

23 (yield = 36.2mg, 0.0796 mmol, 79.7%). DA182901 can also be obtained th rough solvent drop grinding of trans resveratrol (22.8mg, 0.100 mmol), -caprolactam (22.6mg, 0.200 mmol) and acetone (20l) for 5 min in an agat e pestle and mortar. DA182901 can also be obtained from slurry of 1:2 of trans -resveratrol and -caprolactam in acetone. (Melting point = 107oC). After few weeks of first report of DA182901, crystals of DA182901 were transformed to DA182902 crystals (Melting point = 113oC). However, DA182901 were not obtained after appear ance of DA182902 crystals. Figure 2.7. Single crystal of resveratr ol-caprolactam cocrystal (DA1829) 2.2.3.2. Resveratrol-flavone cocrystals (RESVONE 01 and RESVONE02) 11.4 mg (0.050 mmol) of trans -resveratrol and 22.2 mg (0.100 mmol) of flavone were dissolved in 1ml of acetone to result in a clear solution. The solution was allowed to slowly evaporate at room temperature result ing in yellow glassy material and a white precipitate. Re-dissolving in hot 1ml of ethyl acetate then al lowing the solvent to slowly evaporate at room temperature resulted in pa le yellow crystals of RESVONE01 that were harvested after one day (yield = 22.5mg, 0.0334 mmol, 67%).

PAGE 39

24 RESVONE01 can also be obtained from slurry of trans -resveratrol (114 mg, 0. 500 mmol), flavone (222mg, 1.00 mmol)) and 2 ml acetone, which was stirred for five hours then filtered. The filtrate was allowed to slowly evaporate at room temperature to obtain crystals of RESVONE01. RESVONE01can also be obtained by grinding of trans -resveratrol (11.4mg, 0.050 mmol) and flavone (22.2mg, 0.100 mmol) for 5 mi n in an agate pestle and mortar using either 20l of acetone, ethyl acetate or equal mixture of acetone and ethyl acetate. (Melting point = 160oC). Crystals of RESVONE02 were repor ted during preparation of RESVONE01 crystals. RESVONE02 crystals appeared once and then disappeared. Efforts to prepare RESVONE02 resulted always in obtaining RESVONE 01 crystals Figure 2.8. Single crystals of resverat rol-flavone cocrystal (RESVONE01) 2.2.3.3. Resveratrol and 4,4'-dipyridyl cocrystal(DA4) 22.8 mg (0.100 mmol)of resveratrol and 15.6 mg (0.100 mmol) of 4,4'-dipyridyl were dissolved in 6ml of acetone by h eating on a hotplate until a clear solution was

PAGE 40

25 obtained. The solution was allowed to slowly evaporate at room temperature and pale yellow crystals of DA4 were harvested after one day (yield = 34mg, 0.0368, 88.5%). DA4 can also be obtained through solvent drop grinding of 22.8 mg (0.100 mmol)of resveratrol, 15.6mg (0.100 mmol) of 4,4'-dipyridyl and acetone (10ul) for 10 min in an agate pestle and mortar. DA4 can al so be obtained from sl urry of resveratrol (228mg, 1 mmol), 4,4'-dipyridyl (156mg, 1 mmol) and 4 ml acetone (Melting point = 236 oC). Figure 2.9. Single crystals of resver atrol and 4,4'-dipyridyl (DA4) 2.2.4. Results and discussion 2.2.4.1. Resveratrol-caprolactam co crystals (DA182901 and DA182902) OHbOHcOHa A BFigure 2.10. The molecular structure of trans -resveratrol molecule shows labeled rings and phenolic OH groups

PAGE 41

26 The single crystal x-ray analysis of DA 1829 crystals reveals that it forms a 1:2 ratio cocrystal between trans -resveratrol and -caprolactam. Trans -resveratrol molecules form head-to-tail chains through phenolic supramolecular homos ynthons, OHOH (D = 2.7905(12) ). Each trans -resveratrol molecule is hydroge n bonded to three molecules of -caprolactam through the following synthons: OHaCO, OHbCO, and OHcNH (D = 2.5725(13), 2.6587(12), and 2.9299(14) re spectively). One of the -caprolactam molecules is further linked to another -caprolactam molecule to form a centrosymmetric dimer, NHCO (D = 2.8871(14) ). The whole structures shows extended sheets of trans -resveratrol and -caprolactam molecules. Those sh eets are further extended to form three dimensional structures. Figure 2.11. Comparing the asym metric unit of two polymorphs of 1:2 cocrystal of trans -resveratrol and -caprolactam, DA 182901 (left) with that of DA182902 (right) The single crystal x-ray anal ysis of DA182902 crystals reve als that it forms a 1:2 ratio cocrystal between transresveratrol and -caprolactam. Each trans -resveratrol molecule is hydrogen bonded to four molecules of -caprolactam through the following

PAGE 42

27 synthons: OHaCO, OHaNH, OHbCO, OHcNH and OHcCO (D = 2.690(12), 2.992(12), 2.705(14), 2.896(13), and 2.639(14) ). The whole structures shows extended sheets of trans -resveratrol and -caprolactam molecules. Those sheets are further extended to form three dimensional structures. Figure 2.12. Comparison of hydrogen bonding in DA182901 (up) and DA182902 (down) Comparing the two polymorphs, DA182901 shows both supramolecular homosynthons and supramolecular heterosynthons in the same crystal structure. On the other hand, DA182902 shows only supramolecular hete rosynthons in its crystal structure.

PAGE 43

28 Based on CSD search for phenols, OH OH supramolecular homosynthons were observed in 14.78% of the entries in presence of other competitive groups. On the other hand, in absence of competitive groups OHCONH supramolecu lar heterosynthons were reported in 100% of the entries, wh ile OHNHCO were reported in 70% of the entries. Thus, crystal structure of DA182902 was more expected than that of DA182902. All observed hydrogen bonding distances are wi thin the accepted range based on CSD search discussed in section 2.2.2. There is also a major difference between the dihedral angles of the two aromatic ring planes of resveratrol. That dihedral angle is 15.41 in DA182901 while it becomes smaller as 5.89 in DA182902. The latter value is close to that angle in DALGON which is 5.33. This case of polymorphism originates from flexibility of trans -resveratrol molecule, thus, it can be considered as c onformational polymorphism. At the same time the two polymorphs adopt different packing, s o, it can also be considered as packing polymorphism. Efforts to reproduce DA182901 crystals resulted in obtaining crystals of DA182902 which indicates that DA182902 is the most stable form. The difficulty to obtain DA182901 crystals may result from contamination by DA182902 seeds. This indicates that DA182902 is the thermodyna mically favored while DA182901 is the kinetically more favored form.

PAGE 44

29 2.2.4.2. Resveratrol-flavone cocrystals (RESVONE 01 and RESVONE 02) OHbOHcOHa A B O O Figure 2.13. Molecular structures of trans -resveratrol molecule (left), and that of flavone (right) The single crystal x-ray analysis of RE SVONE01 crystals reveals a 1:2 ratio cocrystal between trans -resveratrol and flavone, respect ively. The structure exhibit molecules of trans -resveratrol bridged th rough hydrogen bonding to flavone molecules, whereas flavone molecules aligned perpendicular to trans -resveratrol molecules. The hydrogen bond interactions between trans -resveratrol and flavone molecules occur through the phenolic OH and CO of resverat rol and flavone, respectively, with OHCO distances of 2.715(11), 2.721(12), and 2.759(12) . The structure also contains two flavone molecules involved in interaction with a head-t o-tail arrangement, centroidto-centroid distance of 3.680 . Weak hydrogen bonding between phenolic OH and CH of trans -resveratrol molecules are present, OHCH (D = 3.413 and 3.498 ). Figure 2.14. Comparing the asym metric unit of two polymorphs of 1:2 cocrystal of transresveratrol and flavone, RESVONE01 (le ft) with that of RESVONE02 (right)

PAGE 45

30 The single crystal x-ray analysis of RE SVONE02 crystals reveals a 1:2 ratio cocrystal between trans -resveratrol and flavone, respectively. The hydrogen bond interactions between trans -resveratrol and flavone molecu les occur through the phenolic OH and CO of resveratrol and flavone, resp ectively, with OHCO distances of 2.773(6), 2.737(4), and 2.736(4) . The struct ure also contains two fla vone molecules involved in stacking with a head-to-tail arrangement centroid-to-centroid distance of 3.675 . The overall structures show s alternating molecules of trans -resveratrol and flavone whereas flavone molecules aligned perpendicular to trans -resveratrol molecules. Figure 2.15. Comparison of hydrogen bonding in RESVONE01 (left) and RESVONE02 (right) Figure 2.16. Comparison of hydrogen bo nding pattern in RESVONE01 (left) and RESVONE02 (right)

PAGE 46

31 It is observed that in both structur es, oxygen ether did not involve in any hydrogen bonding interactions. Furthermore, phenolic supramolecular homosynthon did not observe in any of the two polymorphs. Thus, supramolecular heterosynthons were dominant over supramolecular homosynthons. Th is is consistent with CSD search for phenols which reveals that in absence of competitive groups, OHCO supramolecular heterosynthons are observed in 65.59% of entries. Furthermore, OHOH supramolecular homosynthons were not expected as their occurrenc e reported only in 14.78% entries, in presence of competitiv e groups. Moreover, all observed hydrogen bonding distances are within the accepted range based on CSD search discussed in section 2.2.2. Thus, the structures reported here in are considered to be expected based on CSD search. This case of polymorphism can be explained in terms of packing and variations in dihedral angles because transresveratrol and flavone are both flexible molecules. The dihedral angle of the two aromatic ring planes of transresveratrol is 7.46 in RESVONE01 while it becomes larger as 12.49 in RESVONE02. The earlier value is close to that angle in DALGON which is 5.33. On the other hand, the dihedral angles between chromone and phenyl rings, in flavone molecules, are also di fferent in the two polymorphs. The angles values are 6.65 and 3.79 in RESVONE01. On RESVONE 02, the values are 9.42 and 13.17. Since this case of polymorphism originates from flexibility of transresveratrol and flavone molecules, thus, it can be considered as c onformational polymorphism. At

PAGE 47

32 the same time the two polymorphs adopt differe nt packing, so, it can also be considered as packing polymorphism. Efforts to reproduce RESVONE02 crystals resulted in obtaining crystals of RESVONE01 which indicates that RESVONE01 is the most stable form. The difficulty to obtain RESVONE02 crystals may result from contamination by RESVONE01 seeds. This indicates that RESVONE01 is the th ermodynamically favored while RESVONE02 is the kinetically more favored form. 2.2.4.3. Resveratrol and 4, 4’-dipyridyl cocrystal (DA4) The single crystal x-ray analysis of DA4 cr ystals reveal that it forms a 2:3 ratio cocrystal between resverat rol and 4,4'-dipyridyl. Figure 2.17. The asymmetric unit of DA4 Two sets of tetrameric units are formed between transresveratrol and 4,4'dipyridyl molecu les forming OHNarom supramolecular heterosynthons through the two phenolic OH groups of resveratrol located at meta position to each other and aromatic nitrogen atom of 4,4'-dipyridyl. These te trameric units are linked through OHNarom hydrogen bond formed between the third pheno lic OH group of resveratrol and aromatic

PAGE 48

33 nitrogen atom of 4,4'-dipyridyl forming spiral tapes. The hydrogen bond distances between the phenolic OHar omatic nitrogen (OHNarom) are as following: 2.804(5), 2.707(4), 2.850(5), 2.762(5), 2.725(5), and 2.726(5) . Figure 2.18. Two sets of tetrameric units in DA4 Figure 2.19. Spiral tapes of DA4 It is observed that phenolic supram olecular homosynthons did not observe in DA4. Thus, supramolecular he terosynthons were dominant over supramolecular homosynthons. This is consistent with CSD search for phenols which reveals that in absence of competitive groups, OHNarom supramolecular heterosynthons are observed in 82.31% of entries. Furthermore, OH OH supramolecular homosynthons were not expected as their occurrence reported only in 14.78% entries, in presence of competitive

PAGE 49

34 groups. Moreover, all observed hydrogen bonding distances are within the accepted range based on CSD search discussed in sec tion 2.2.2. Thus, DA4 crystal structure is considered to be expected based on CSD search. Furthermore, the observed structure wa s expected based on Aoyama et al.50 and MacGillivray et al.51-53 work on [2+2] photodimerization in case of resorcinol and 4,4bpe derivatives, were OHNarom supramolecular heterosynthons were occasionally observed. Figure 2.20. [2+2] photodimerization of 2 (4,4-bpe) and 2(resorcinol)51 Based on Schmidt work, parallel olefins separated by < 4.2 should be used for photodimerization.54 However, 4,4'-dipyridyl molecules in DA4 are parallel and separated by 3.916 and 3.764. In fact, there are no olefins in the case pr esented herein. But the ability of this cocrystal to exhibit this topology and dist ances predict that co crystallization of transresveratrol with other derivatives of 4,4'-dipyridyl may re sult in cocrystals which become suitable for photodimerization.

PAGE 50

35 2.2.5. Dissolution test As discussed earlier in section 1.4., biopharmaceutical classification system (BCS) considers aqueous solubility as one of the factors that directly affect drug bioavailability. However, the vast majority of APIs tend to have low aqueous solubility, therefore, improving their solubility is neces sary to improve their bioavailabilities. Aqueous solubility could be enhanced by particle size reduction, lyophilization, additives, and forming new crystalline fo rms such as new polymorphs, salts, or cocrystals. Several studies of pharmaceutical cocrystals demonstrated that solubility of a particular API could be cont rolled via cocrystallization.55,56 Fluoxetine hydrochloride (Prozac), a commonly used antidepressant, has b een studied in this context. Aqueous dissolution rate of three cocr ystals of fluoxetine HCl were determined and compared to that of fluoxetine HCl. As shown in fi gure 2.24., alteration of dissolution rate of fluoxetine HCl has been observed. Fluoxetine HCl: su ccinic acid cocrystal exhibits 2-fold increase in dissolution rate, while the rate d ecreases to the half in case of fluoxetine HCl: benzoic acid cocrystal. In addi tion, a slight increase in disso lution rate has been reported in case of fluoxetine HCl: fumaric acid cocr ystal. Thus, it beco mes evident that dissolution rate of an API can be modifie d, and further controlle d, by cocrystallization. Furthermore, this case could be considered as an obvious illustrati on of how cocrystals can afford higher aqueous dissolutio n rates comparing to salts.

PAGE 51

36 Figure 2.21. Dissolution profiles of Fluoxetine HCl and its cocrystals55 Another case study on disso lution rate was performed on itraconazole (Sporanox), an antifungal drug with low aqueous solubility, which is marketed in the amorphous form. Figure 2.25. shows the dissoluti on profiles of itraconazole and four of its cocrystals, which reveals up to 20-fold increas e of itraconazole dissolution rate, while cocrystallized with organic acids, compar ed to that of amor phous itraconazole. Figure 2.22. Dissolution profiles of itraco nazole and its cocrystals (itraconazole presented by white)56

PAGE 52

37 However, no reports exist in current liter ature concerning diss olution studies of nutraceutical cocrystals, in general, and transresveratrol, in particular. Therefore, we report herein a dissolution study of transresveratrol and two of its cocrystals. UV-vis spectrophotometer was employed in this study du e to presence of two aromatic rings in transresveratrol which represent the chromophores. Since dissolution rate can greatly be aff ected by particle size, all samples were sieved using ASTM sieve in order to keep particle sizes between 53m-75m. Because of the inherent low solubility of resveratrol in water, aqueous medium was not suitable to conduct the intended investigation. Moreover, 25% ethanol: water solvent system resulted in poor dissolution of the solids. In a 50% aqueous ethanol (v/v), satisfactory dissolution profiles were obtained and thus this solvent system was employed in this study. Since both trans -resveratrol and flavone have clos ely similar conjugated systems, i.e. chromophores, there was an interfere nce between their UV absorption spectra. However, that interference was avoided by measuring absorbance at 345 nm where absorbance of trans -resveratrol only could be recorded in this region. Calibration plot for trans -resveratrol was created using a range of known concentrations of trans -resveratrol and recording the corresponding absorbance at 345 nm. Furthermore, little shif t in absorbance profile of trans -resveratrol was detected in case of its cocrystals, which could be expl ained as a result of interactions between trans resveratrol and its coformers. Thus, additi onal calibration plots were created for each cocrystal where the concentration of resv eratrol in each case is known based on the

PAGE 53

38 composition in the single crystal structure. Trans -resveratrol constitutes 50.2 wt% and 34.0 wt% in the case of transresveratrol-caprolactam (DA182902) and transresveratrolflavone (RESVON01) cocr ystals, respectively. All the experiments were carried out at 296 K, and three sets of each sample were performed to keep the accuracy. Using stir bars, the samples were stirre d at 125 rpm for four hours. Fixed amounts of the aliquot were drawn at various time in tervals, and filtered using 0.25 mm syringe filters equipped with 0.45 m nylon membranes. Proper dilutions were performed to obtain acceptable absorption values, and th en UV absorbance was recorded. Finally, concentration of trans -resveratrol was calculated usi ng calibration plots and dilution factors. Finally, dissolution profiles were created using time versus concentration of trans -resveratrol, as shown in figure 2.26. Figure 2.23. Dissolution profiles of trans -resveratrol and its cocrystals 0 10 20 30 40 50 60 5050150250mg / mLTime, minResveratrol Dissolution 1:2 Caprolactam;Rel. Sol.=56.4; mp=112.1C Resveratrol; Rel. Sol.=1; mp=271.6C 1:2 Flavone;Rel. Sol.=0.9; mp=159.9C

PAGE 54

39 The dissolution profiles indi cate a 4-fold increase of trans -resveratrol dissolution rate in case of trans -resveratrol: caprolactam cocrys tal (DA182902). On the other hand, trans -resveratrol: flavone cocrystal (RESVONE01) has a low dissolution rate. This can be explained in terms of relative solubility57 and trends in melting points. Relative solubility could be calculated as a ratio of solubility of the cocrystal former to that of the drug in the same solvent system. It is believed that substance with high melting point has strong self interactions, and therefore, it will not dissoci ate easily, and in other word s, it will not dissolve easily. The opposite is true for low melting point substance. It is found that relative solubility of trans -resveratrol: caprolactam cocrystal is higher than that of trans -resveratrol: flavone cocrystal. Th is is in agreement with trends in melting points, trans -resveratrol: flavone cocrysta l is higher than that of trans resveratrol: caprolactam cocrystal. These results are in agreement with RodriguezHornedo et al.57 In summary, cocrystallization could be us ed to control solubility of nutraceuticals, either by increasing or decrea sing their solubilitis. 2.2.6. Conclusion The choice of transresveratrol as a target for cocrystallization attempts conducted in this study is justified due to the rapi dly growing interest in current literature establishing desirable pharmacological propert ies of resveratrol and, equally important, the known poor solubility and bioavailability of resveratrol. The principles and concepts of crystal engineering were implemented in developing a suitable strategy to produce

PAGE 55

40 cocrystals of resveratrol. Thr ough a statistical analys is of CSD, phenols, in general, were found to be able to participate in hydroge n bond interactions forming supramolecular heterosynthons with a variety of functional groups including secondary amide, aromatic nitrogen, and carbonyl. Therefore, molecules that contain these functionalities were employed as cocrystal formers in the design strategy for resveratrol cocrystallization. Three cocrystals of transresveratrol are presented herein and it is found that they can be obtained by different cocrys tallization techniques. Interestingly, two cases of polymorphism in cocrystals are reported herein. CSD search for polymorphic cocrystals shows only 40 cases, which represents 1.9% of the total number of cocrystals availabl e on CSD (August 2008 update). Moreover, dissolution test was performed on resveratrol and two of its cocrystals, using UV-vis spectrophotometer, where the data demonstrate that through cocrystallization with different cocrystal form ers, dissolution rate of resveratrol could be greatly modified, and further controlled. In conclusion, cocrystals are amenable to design and prediction using basics of crystal engineering and supramolecular ch emistry, where CSD is considered as invaluable tool in this context. Overall, nutraceutical cocrystals represent viable and readily accessible crystalline forms to enhance, and further control, the solubility of nutraceuticals.

PAGE 56

2 2 i n a b p r a c a n e x p r b f o s a U a c .3. Citric a c .3.1. Introd Citric n a variety o f b out 8% of t Beca u r eservative. c id cycle. H o n d in pharm x cipient.59 I t r eparations. icarbonates o r m salts of Furth e a fe substanc U nited State s c id is consi d c id uction acid [2-hy d f fruits. Ho w t heir dry we i Fi gu u se of its an t It also play s o wever, citr i aceutical in d t is widely u The charac t or carbonat e APIs. e rmore, U.S e; therefore s ).60 It also h a d ered to be h d roxy-1,2,3p w ever, it is m i ght.58 u re 2.24. M o t ioxidant pr o s an importa n i c aci d has m d ustry, in p a u sed as flavo t eristic effer v e s with citri c Food and D it is a mem b a s high LD 5 h ighly solub l 41 p ropanetric a m ainly extra c o lecular str u o perties, citr i n t role in bi o m any uses a n a rticula r sinc ring and sta b v escence of c acid. More D rug Admi n b er of EAF U 5 0 (3000 mg / l e in water ( 1 a rboxylic aci c ted from ci t u cture of ci t i c acid is wi d o chemistry a n d applicati o e it has a pa r b ilizing age n antacids is a over, citric a n istration (F D U S list (Eve r / kg in rat). I n 1 330mg/ml) d] is a nutra t rus fruits as t ric acid d ely used as a s an interm e o ns in indus t r ticular val u n t in many p a result of c o a cid is used i D A) conside r ything Add e n addition t o ceutical, ex i it constitut e a natural f o e diate in cit r t ry, in gener a u e as an p harmaceuti c o mbining i n many cas r citric acid e d to Food i n o its safety, c i sts e s o od r ic a l, c al es to as a n the c itric

PAGE 57

42 Thus, citric acid is safe, hi ghly soluble, and abundant in low cost, therefore, it could be considered as a valid cocrystal former. However, CSD search for citric acid re veals a two polymorphs of unhydrous citric acid and a crystal structure of citric acid monohydrate (CSD refcodes: CITRAC10, CITRAC11 and CITARC, respectiv ely). Four crystal struct ures were reported as cocrystals formed between citric acid and the following cocrystal formers: 2,5-bis(4Pyridyl)-1,3,4-oxadiazole, betaine, the ophylline monohydrate, and caffeine (CSD refcodes: MEBRUH, XOBHIF, KIGKAN, a nd KIGKER). (CSD August 2008 update). 2.3.2. CSD analysis As we discussed earlier in section 1.2.2., CS D search can serve as a tool to predict and design new cocrystals. Therefore, deta iled CSD search conducted herein using Conquest v. 1.10. Filters applied to the searches to limit th e results according to the following criteria: 3D coordinates determ ined, R factor < 7.5%, no ions and only organics. The search included acids and alcohol s because they are th e targeted functional groups in citric acid. Interestingly, both acids and alcohols can serve as hydrogen bond donors or acceptors. Phenols were excluded fr om the search for alcohols due to the difference between acidity of phenols and alcohols which therefore affects their supramolecular interactions. The results are summarized in tables 2.3. and 2.4. were raw data indicates results in presence of co mpetitive hydrogen bonding groups while refined data is that in absence of any competitive groups.

PAGE 58

43 2.3.2.1. CSD statistics for acids Citric acid molecule contains three carboxylic acid groups. Those groups can serve as hydrogen bond donors, OH, and acceptors, CO. The results of CSD search for carboxylic acids are summarized in table 2.3. and it is based on 5690 entries. Table 2.3. CSD statistics for acids Functional group Raw data Refined data Bond distance () mean () COOHNaro m 457/607 (75.29%) 119/126 (94.44%) 2.510-2.827 2.652(2) COOHCl171/267 (64.04%) 51/51 (100%) 2.772-3.238 3.001(4) COOHP=O 68/122 (56%) 17/17 (100%) 2.417-2.852 2.579(6) COOHOH OHCOOH 523/1176 (44.47%) OHHOOC 501/1176 (42.60%) 173/230 (75.22%) 186/230 (80.87%) 2.607-3.000 2.4142.989 2.790(3) 2.655(3) COOHCONH2 78/177 (44.07%) 10/19 (52.63%) COOHCONH2 (2.501-2.713) CONH2COOH (2.809-3.247) 2.583(5) 2.957(9) COOHCOOH Dimers 1785/5690(31.37%) Catemers 156/5690 (2.74%) Dimers 384 /474 (81%) Catemers 35/474 (7.4%) 2.536-2.983 2.649(1) COOHCO 147/597 (24.62%) 67/ 178 (37.64%) 2.4 51-2.863 2.689(5) CSD Conquest v. 1.10, Jan 2008 update, organics only, 3D coordinates, R<=7.5%, no ions CSD search reveals 607 entr ies with both COOH and Narom. In presence of competitive groups, 457 (75.29%) entries exhibit COOHNarom supramolecular heterosynthons. The percentage increases to 94.44% in absence of competitive functional groups. The histogram for bond distance range is shown in figure 2.28. and it reveals bond distance range of 2.510-2.827 (mean = 2.652(2) ).

PAGE 59

44 Figure 2.25. Histogram for COOHNarom Another CSD search for structur es contain both COOH and Cl–reveals 267 entries. In presence of competitive gr oups, 171 (64.04%) entries exhibit COOH Cl–supramolecular heterosynthons The percentage increases to 100% in absence of competitive functional groups. The histogram for bond distance range is shown in figure 2.29. and it reveals bond distance rang e of 2.772-3.238 (mean = 3.001(4) ). Figure 2.26. Histogram for COOHCl–

PAGE 60

45 Another CSD search for structures with both COOH and P=O reveals 122 entries. In presence of competitive groups, 68 (56 %) entries exhibit COOHP=O supramolecular heterosynthons. The percentage increases to 100% in absence of competitive functional groups. The histogram for bond distance range is shown in figure 2.30. and it reveals bond distance range of 2.417-2.852 (mean = 2.579(6) ). Figure 2.27. Histogram for COOHP=O Another CSD search for structures w ith both COOH and alcoholic OH reveals 1176 entries. In presence of competitive groups, 523 (44.47%) entries exhibit OHCOOH supramolecular heterosynthons while 501 (42.60%) entries exhibit OHHOOC supramolecular heterosynthons. In absence of co mpetitive functional groups, the percentage increases to 75.22% and 80.87%, fo r OHCOOH and OHHOOC, respectively. The histogram for bond distance range is shown in figure 2.31. and it reveals bond distance range of 2.607-3.000 (mean = 2.790(3) ) and 2.414 -2.989 (mean = 2.655(3) ), for OH COOH and OHHOOC, respectively.

PAGE 61

46 Figure 2.28. Histogram for OHCOOH (lef t) and that for OHHOOC (right) Another CSD search for structures w ith both COOH and primary amide reveals 177 entries. In presence of competiti ve groups, 78 (44.07%) entries exhibit COOHCONH2 supramolecular heterosynthons. The percentage increases to 52.63%, in absence of competitive functional groups. Th e histogram for bond distance range is shown in figure 2.32. and it reveals bond di stance range of 2.501-2.713 (mean = 2.583(5) ) and 2.809-3.247 (mean = 2.957(9) ), for COOHCONH2 and CONH2COOH respectively. Figure 2.29. Histogram for COOHCONH2 (left) and CONH2COOH (right)

PAGE 62

47 For search of carboxylic acid supramolecu lar homosynthons, CSD search reveals 5690 entries with carboxylic acid. In pr esence of competitive groups, 1784 (31.35%) entries exhibit acid dimers while156 (2.74%) en tries exhibit acid catem ers. In absence of competitive functional groups, the percentages increase to 81% and 7.4%, for acid dimers and catemers, respectively. The histogram for bond distance range is shown in figure 2.33. and it reveals bond distance rang e of 2.536-2.983 (mean = 2.649(1) ). Figure 2.30. Histograms for ca rboxylic acid dimers To investigate supramolecular hetero synthons between carboxylic acids and aldehydes or ketones, carbonyl group is used as a representative for the two later groups. CSD search reveals 597 entries with both COOH and CO. In presence of competitive groups, 147 (24.62%) entries e xhibit COOHCO supramol ecular heterosynthons. The percentage increases to 37.64% in absen ce of competitive functional groups. The histogram for bond distance ra nge is shown in figure 2.34. and it reveals bond distance range of 2.451-2.863 (mean = 2.689(5) ).

PAGE 63

48 Figure 2.31. Histogram for COOHCO 2.3.2.2. CSD statistics for alcohols Citric acid molecule contai ns one alcoholic OH. This group can serve as hydrogen bond donor and acceptor. The results of CSD sear ch for carboxylic acids are summarized in table 2.4. and it is based on 25035 entries. Phenols were excluded from this search because the difference between acidity of al cohols and phenols, as discussed earlier. Table 2.4. CSD statistics for alcohols Functional group Raw data Refined data Bond distance () mean () OHCl389/529 (75.53%) 0 2.853-3.496 3.116(4) OHCONH2 144/252 (57.14%) OHCONH2 141/252 (55.95%) OHNH2CO 50/54 (92.59%) OHCONH2 49/54 (90.74%) OHNH2CO 2.607-2.997 2.737-3.198 2.759(6) 2.990(6) OHP=O 202/368 (54.89%) 32/48 (66.67%) 2.429-2.899 2.714(4) OHNarom 457/930 (49.14%) 70/102 (68.63%) 2.590-3.099 2.821(4) OHCOOH COOHCOOH 523/1176 (44.47%) COOHHOOC 501/1176 (42.60%) 173/230 (75.22%) 186/230 (80.87%) 2.607-3.000 2.414-2.989 2.790(3) 2.655(3) OHCO 1694/3886 (4 3.59%) 552/851 (64.86% ) 2.413-3.098 2.825(2) OHOH 5184/18475 (28. 06%) 824/1055 (78.10%) 2.510-3.070 2.797(5) CSD Conquest v. 1.10, Jan 2008 update, organics only, 3D coordinates, R<=7.5%, no ions

PAGE 64

49 CSD search for structures contain both OH and Cl–reveals 529entries. In presence of competitive groups, 389 (75.53%) entries exhibit OHCl–supramolecular heterosynthons. Interestingly, in absence of competitive functiona l groups, no entries with both OH and Cl– were found. The histogram for bond distance range is shown in figure 2.35. and it reveals bond distance ra nge of 2.853-3.496 (mean = 3.116(4) ). Figure 2.32. Histogram for OHCl– Another CSD search for structures with both OH and primary amide reveals 252 entries. In presence of competitive groups 57.14% of the entries exhibit OHCONH2 while 55.95% exhibit OHNH2CO supramolecular hetero synthons. In absence of competitive functional groups, the percentages increase to 92.59% and 90.74%, for OHCONH2 and OH NH2CO, respectively. The histogra m for bond distance range is shown in figure 2.36. and it reveals bond di stance range of 2.607-2.997 (mean = 2.759(6) ) and 2.737 -3.198 (mean = 2.990(6) ), for OHCONH2 and OH NH2CO, respectively

PAGE 65

50 Figure 2.33. Histogram for OHCONH2 (left) and that of OHNH2CO (right) Another CSD search for structures with both OH and P=O reveals 368 entries. In presence of competitive groups, 202 (54.89%) entries exhibit OHP=O supramolecular heterosynthons. The percentage increases to 66.67% in absence of competitive functional groups. The histogram for bond distance range is shown in figure 2.37. and it reveals bond distance range of 2.429-2.899 (mean = 2.714(4) ). Figure 2.34. Histogram for OHP=O Another CSD search reveals 930 entries with both OH and Narom. In presence of competitive groups, 457 (49.14%) entries exhibit OHNarom supramolecular heterosynthons. The percentage increases to 68.63% in absence of competitive functional

PAGE 66

51 groups. The histogram for bond distance range is shown in figure 2.38. and it reveals bond distance range of 2.590-3.099 (mean = 2.821(4) ). Figure 2.35. Histogram for OHNarom To investigate supramolecular hete rosynthons between alcoholic OH and aldehydes or ketones, carbonyl group is used as a representative for the two later groups. CSD search reveals 3886 entries with both OH and CO. In presence of competitive groups, 1694 (43.59%) entries exhibit OH CO supramolecular heterosynthons. The percentage increases to 64.86% in absen ce of competitive functional groups. The histogram for bond distance ra nge is shown in figure 2.39. and it reveals bond distance range of 2.413-3.098 (mean = 2.825(2) ). Figure 2.36. Histogram for OHCO

PAGE 67

52 CSD search reveals 18475 entries with al coholic OH. In presence of competitive groups, 5184 (28.06%) entries exhibit OH OH supramolecular homosynthons. The percentage increases to 78.10% in absen ce of competitive functional groups. Bond distance range is found to be 2.510-3.070 (mean = 2.780(5) ). 2.3.3. Experimental All the chemicals used during the preparation were obtained from commercial sources and used as received. All crystal lization experiments were conducted in an unmodified atmosphere and the solven ts were distilled prior to use. 2.3.3.1. Citric acidiso -nicotinamide cocrystal (DA005) 38.4 mg (0.200 mmol) of citric ac id and 73.2 mg (0.600 mmol) of iso nicotinamide were dissolved in 5ml of hot methanol until a clear solution was obtained. The solution was allowed to slowly evaporate at room temperature and colorless crystals of DA005 were harvested afte r one day (yield = 106.8mg, 0.257 mmol, 95.7%). (Melting point = 148oC). Figure 2.37. Single crystals of citric acidiso -nicotinamide cocrystal (DA005)

PAGE 68

53 2.3.3.2. Dihydrate of citric acidiso -nicotinic acid cocrystal (DA02) 76.8mg (0.400mmol) of citric acid and 49.2mg (0.400mmol) of iso -nicotinic acid were dissolved in 5ml of methanol by hea ting on a hotplate, filtered and allowed to slowly evaporate at room temperature. Colo rless crystals of DA02 were harvested after one day (yield = 114.5mg, 0.241 mmol, 90.9%). (Melting point = 187oC). Figure 2.38. Single crystals of citric acidiso -nicotinic acid dihydrate cocrystal (DA02) 2.3.4. Results and discussion 2.3.4.1. Citric acidiso -nicotinamide cocrystal (DA005) The single crystal x-ray crystal structur e of DA005 indicates that molecules of iso -nicotinamide form hydrogen bonds with molecu les of citric acid with a stoichiometry of 2:1. The asymmetric unit of DA005 cont ains two crystallogra phically independent iso nicotinamide molecules and a citric acid molecule. Figure 2.39. The asymmetric unit of DA005

PAGE 69

54 Each molecule of citric acid is linked to three molecules of iso -nicotinamide, OHacidNarom (D = 2.521(3) ) NHamideCOacid (D = 2.832(5) and 2.905(5) ). There is further an intramolecular hydrogen bondi ng in citric acid molecules, OHacidOHalcohol (D = 2.902(2) and 2.937(4) ). Two iso -nicotinamide molecules form noncentrosymmetric amide dimer through, NHamideCOamide (D = 2.894(2) and 2.905(5) ). Each of the aromatic nitrogen atoms of iso -nicotinamide are further included in a hydrogen-bond interaction with terminal carboxylic acid group of a citric acid molecule OHacidNarom (D = 2.521(3) and 2.625(2) ) to form zigzag tapes. These tapes ar e connected through the middle carboxylic acid moieties of citric acid molecules. Figure 2.40. Crystal packing in DA005 Based on CSD search discussed earlier, raw se arch is applicable in this case rather than refined search because we have four groups which can serve as hydrogen bonding donors or acceptors. Certainly, there are three functional groups which can act as hydrogen bonding donors or acceptors; COOH, alcoholic OH, and primary amide. On the other hand, Narom can act as hydrogen bonding acceptor only.

PAGE 70

55 However, all interactions reported in DA005 were expected based on CSD search (raw data), and this can be e xplained as following: COOHNarom supramolecular heterosynthons were observed in 75.29% of entries, COOHCONH2 were observed in 44.07% of entries, and COOHOHalc were observed in 42.60% of entries. All observed hydrogen bonding distances are within th e accepted range based on CSD search discussed in sections 2.3.2.1 and 2.3.2.2. On the other hand, two supramolecular he terosynthons were not observed while they have chance to appear base d on CSD search. T hose are HOOCOHalc (44.47%), and OHalcNarom (49.14%). In summary, COOH supramolecular hetero synthons were more expected than COOHCOOH supramolecular homosynthons, a nd this is the case reported herein. 2.3.4.2. Dihydrate of citric acidiso -nicotinic acid cocrystal (DA02) The single crystal x-ray structure analysis of DA02 reveals that the asymmetric unit contains a dihydrate of 1:2 cocrystal of citric acid and iso -nicotinic acid, where iso nicotinic acid molecules are present in zwitterionic form. Figure 2.41. The asymmetric unit of DA02

PAGE 71

56 Each molecule of citric acid is linked to two molecules of iso -nicotinic acid, OHacidCO and OHalcohol CO (D = 2.641(2) and 2.645(2) respectively). The indicated molecule of citric acid is fu rther linked to three water molecules, OHacidOHwater (D = 2.556(2) and 2.621(2) ), OHalcoholOHwater (D = 2.676(2) ), and OHwaterCOacid (D = 2.991(2) ). The two zwitterions of iso -nicotinic acid molecules constitute molecular tapes through th e formation of charge assisted NH+CO supramolecular heterosynthons (D = 2.668(2) and 2.676(2) ). Su ch tapes are running in anti-parallel fashion and are conne cted by two water molecules, OHwaterCO (D = 2.683(2) and 2.740(2) ) and OHwaterOHwater (D = 2.707(3) ). The citric acid molecules connect such anti-p arallel tapes through hydroxyl and carboxylic acid moieties forming sheets. Figure 2.42. Crystal packing in DA02 This zwitterionic structure is based on charge-assisted hydrogen bonding rather than neutral hydrogen bonding interactions. Th is makes it a unique structure and thus, CSD search discussed earlier is not applicab le to this case. It is known that charge-

PAGE 72

57 assisted hydrogen bonding inte ractions are stronger than neutral hydrogen bonding interactions, which make distan ces shorter in the earlier case than the latter one. Thus, COis stronger than neutral CO and this expl ains dominance of ch arge-assisted hydrogen bonding over neutral hydrogen bonding in this crystal structure. Furthermore, water molecules are able to act as hydrogen bonding donors or acceptors. Hence, they played important rule in sustaining the crystal structure as they involved in four out of eight of the hydrogen bonding intera ctions exhibited herein. To confirm that iso -nicotinic acid molecules in DA02 are zwitterions, further study of the crystal structure has to be perf ormed. In addition to IR spectroscopy, two measurements can be used for this pur pose, C-O distances and C-N-C angles of pyridines.61 -66 Figure 2.43. Comparison of the C-N-C angle of the pyridine and the C-O distances of the carboxylic acid Moieti es in DA02 molecules (left) and iso -nicotinic acid (right) It is known that in neutral carboxylic acid, C-O distance is longer than that of C=O. In case of carboxylate, no C=O is pres ent, and thus the distance of both C-O moieties are in between those of C-O and C= O. Moreover, C-N-C angle is larger in protonated pyridine than those of neutral ones.

PAGE 73

58 Table 2.5. Comparison of the C-N-C angle of the pyridine and the C-O distances of the carboxylic acid Moieties in DA02 molecules (left) and iso -nicotinic acid (right) Analyzed crystallographic parameter Zwitterionic molecules Iso -nicotinic acid molecule C-O distances 1.260 , 1.270 , 1.244 , 1.236 1.215 , 1.294 C-N-C angles 121.91, 121.81 118.91 Furthermore, p K a [p K a(base)-p K a (acid)] is traditionally used in pharmaceutical industry as an indicator of sa lt formation. It is known that p K a >3 is usually associated with salt formation67, If the value of p K a is between 2 and 3, formation of either a salt or a cocrystal cannot be predicted. However, p K a for citric acid and pyridyl moiety of iso -nicotinic acid = 4.88-2.93= 1.95 which does not meet that criterion for salt formation. However, it is clear that iso -nicotinic acid molecule has both acidic and basic moieties. Thus, it might tend to present as zwitterion. Mo reover, we can also c onsider presence of COOH and Narom in para position as a factor which mi ght stabilize the zwitterions in this structure. However, CSD search for iso -nicotinic acid reveals one neutral crystal structure and there are no cocrystals of iso -nicotinic acid. Thus, no more information about iso nicotinic acid cocrysta ls can be obtained.

PAGE 74

2 s u p r a c s u c o v r e p c r i n 2 a n .3.5. Concl u Citric u ch as safet y r eparations. c id; carbox y u ggested th a o nsidered a s Two n arious anal y e ported, and arameters. In co n r ystal engin e n valuable to o .4. Flavano n Singl e n attempt to u sion acid was c h y low cost, h Two functi o y lic acid and a t functional i s viable can d n ovel cocry s y tical techni q this remark n clusion, co c e ering and s u o l in this co n n e sin g le cr y e crystal of f cocrystalli z Fi g ur e h osen for co c h igh solubil i o nalities we r alcoholic O i ties like ar o d idates. s tals of citri c q ues. Interes able situati o c rystals are a u pramolecu l n text. y stal f lavanone [2 z e trans -resv e 2.44. The m 59 c rystallizati o i ty, and its a b r e used as a t O H. Therefo r o matic nitro g c acid were o tingly, a hy d o n was confi r a menable to l ar chemistr y ,3-Dihydro f eratrol and f m olecular s t o n based on m b undance i n t arget for cr y r e, CSD ana l g en and pri m o btained an d d rate of zwi t r med by so m design and p y where CS f lavone] is r e f lavanone. t ructure of m any of its c n various ph a y stal engine e l ysis was co n m ary amide c d fully chara c t terionic coc r m e crystallo g p rediction u D is consid e e ported for fi flavanone c haracterist i a rmaceutica l e ring of citr i n ducte d and c ould be c terized usi n r ystal is g raphic sing basics o e red as fi rst time du r i cs l i c it n g o f r ing

PAGE 75

60 2.4.1. Experimental 112 mg (0.499 mmol) of flava none were dissolved in 5m l of acetone and heated on a hotplate. The solution allowed to slow ly evaporating at room temperature. 2.4.2. Result and discussion The crystal structure of flavanone is m onoclinic with one molecule of flavanone in the asymmetric unit. Figure 2.45. The asymmetric unit of flavanone The crystal structure does not show strong hydrogen bonding since flavanone does not have strong hydrogen bond donors, although some forms of weak hydrogen bonding were observed, CHCO (D = 3.437 and 3.476 ) and CHO (D = 3.475 ). The structure sustained by CHinteractions, CHcentroid (D = 3.604 and 3.834 ) which lead to herringbone patt ern as shown in figure 2.49.

PAGE 76

61 Figure 2.46. Hydrogen bonding in flavanone Figure 2.47. Herringbone pattern in flavanone 2.4.3. Conclusion Single crystal of flavanone is reported for first time during an attempt to cocrystallize trans -resveratrol and flavanone. In th e crystal structure, weak hydrogen bond interactions, CHCO, CHO and electrostatic CHinteractions are observed. Flavanone could be considered as rela tively safe; LD50 is 75 mg/kg (bird-wild bird species), therefore, it can be used as a cocrystal former.

PAGE 77

62 Chapter 3. Polymorphism 3.1. Introduction The term Polymorphism (Greek: poly = many, morph = form) was first introduced around 1822 by Mitscherlich68, while the earliest report of the phenomenon is most probably found in the works of Klaproth (1788 ). Klaproth reported two different phases of calcium carbonate, calcite and aragon ite. The phenomenon, for molecules and complexes, is closely related to allotropism, in elements, where different structural forms of the same element can exhibit quite different physical properties. Allotropism is best exemplified by diamond, graphite, and fullerene s, three allotropes of carbon, Figure 3.1. Figure 3.1. Allotropes of carbon diamond (left), graphite (middle), and fullerene (right).

PAGE 78

63 The stunning case of highest number of polymorphs is the case of ROY69,70 [5Methyl-2-[(2-nitrophenyl) amino]-3-thiophen ecarbonitrile]. It named ROY because the crystal colors of its polymorphs are red, orange, and yellow. ROY which has eight polymorphs and it is an ideal example of c onformational polymorphism as discussed in section 3.1.3. Figure 3.2. The molecular structure of a red polymorph of ROY (CSD refcode: CAXMEH03) A polymorph is defined by McCrone (1965) as “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”.71 While this definition might imply that structures where molecules exhibit different conformations are not considered polymorphs, McCrone pointed out that polymorphs will be essentially the same in liquid and vapor states but different in crystal structure.72 However, this later “safe” defi nition of polymorphs can accommodate conformational polymorphs.

PAGE 79

64 As the number of published crystal struct ures has grown exponentially since the original works of Klaproth, the number of pol ymorphs has also increased. According to McCrone, all compounds are polymorphic, a nd the number of known polymorphs of a single compound is proportional to time a nd effort spent researching this compound.71 Other terms related to polymorphism are commonly used in literature. One of these terms is “pseudopolymorphism”, which is used to describe solvates and hydrates. The use of this term was a matter of a recent scientif ic debate. Rogers70, Seddon73, and Bernstein74 opposed its use on the basis of bein g a misleading and an ambiguous term where they recommended journa l editors to eliminate it from current usage. However, Desiraju69,75 argued that despite its ambiguity it has to be retained because it is too common and it has a scientific value. Nangia76 also agrees on retain ing its usage as it is clear enough to all scientists and he pointed out its scient ific and legal importance. 3.1.1. Significance of polymorphism to pharmaceutical industry Polymorphism has significant applications in pharmaceutical industry. The choice of the right polymorph is critical during formulation of pharmaceutical preparations.77 As we pointed out earlier, polymorphs exhibit si gnificantly different physic al properties, it is widely recognized that different polymorphi c forms of the same active pharmaceutical ingredient (API) exhibit differe nt stabilities, solubilitis, me lting points, processabilities, bioavailabilities, particle flow, etc.77 Therefore, it becomes evident that physical properties of APIs can be optimized thr ough control, and further design, of suitable polymorphs. A striking example for this phe nomenon is demonstrated in the case of

PAGE 80

65 paracetamol. Two polymorphs of paracetamol form I50 and II (Haisa 1974), reported by Haisa et al., figure 3.3., exhib it widely different compressi bility. This difference in compressibility greatly affects processi ng of the drug, and thus represents a good example of the effect of polymorphism on targeted physical properties in API formulation. The difference in compressi bility between the two polymorphs could directly be linked to the difference in the underlying p acking patterns in the two polymorphs. In polymorph I, hydrogen-bonded paracetamol molecules crystallize in puckered hydrogen-bonded sheets, resisting compressibility as the sheets do not slip easily over each other. On the other hand, pa racetamol molecules in form II exhibit flat sheets that can slip easily over each other, and thus demonstrate better compressibility. Figure 3.3. Paracetamol polymorphs, (lef t) form I showing puckered hydrogenbonded sheets, and (right) form II showin g flat sheets. (CSD refcodes: HXACAN01 and HXACAN08, respectively) Effect of polymorphism on bioavailability can be demonstrated by the case of chloramphenicol palmitate. Chloramphenicol is an antimicrobial agent which has a strong bitter taste and therefore it cannot be administered orally. Edgerton78 has synthesized a tasteless chloramphenicol palmitate which is a p oorly soluble ester. It does not dissolve in

PAGE 81

66 mouth while administered orally and it under goes hydrolysis in the small intestine to form active chloramphenicol. The rate of este r hydrolysis, which is the rate of absorption determinant factor, depends on the selection of the right polymorph79 as chloramphenicol palmitate exhibits polymorphism80-82 Aguiar et al.96 carried out an experiment to study the absorption of two polymorphs of chloramphenicol palmitate A a nd B. Varied concentrations of A and B were administered orally by human volunt eers and urine and blood samples were collected over a 24 hr period. Urinary excr etion data are shown in Figure 3.4. which demonstrate that concentration of chloramphe nicol equivalent increases as concentration of form B increases. Blood concentrations of free chloramphenicol were also plotted versus percent concentration of form B. This shows that chloramphenicol blood serum levels proportionally increases by increasi ng concentration of polymorph B and this linear relationship is shown in figure 3.5. This study reveals that API absorption is directly related to the type a nd concentration of the polymorph. Figure 3.4. Chloramphenicol urinary excreti on rates after oral administer of chloramphenicol palmitate with varying percents of polymorph B: M, 0%; N, 25%; O, 50%; P, 75%; and L, 100%96

PAGE 82

67 Figure 3.5. Blood serum levels of chlora mphenicol after oral administer of chloramphenicol palmitate with varying percents of polymorph B96 One important aspect of polymorphism in pharmaceuticals is susceptibility of polymorphs to interconvert77, where many consequences can result from conversion of the kinetically favored form to the therm odynamically stable form either during the manufacturing processes or upon st orage (shelf life). This may result in precipitation of the low solubility stable form which ma y lead to pharmaceutically unacceptable preparations. Consequently, interconversion into the low solubility, thermodynamically stable, form may limit pharmacological utility of the API77. Furthermore, as the solubility of API is greatly affected by particle size, precipitation and subseque nt particle growth may lead to variations in particle size distribution in a given formulation, which subsequently affect the solubi lity of an active ingredient. In their review, Haleblian and McCrone72 provided many examples for the consequences of phase interconversion in pharmaceutical industry. Considering parenteral preparations, preci pitation may result in significan t particle growth which may affect the syringibility of th e product. Precipitatio n in suspensions might lead to caking,

PAGE 83

68 which could be significant enough to affect the API uniformity and furthermore might prevent resuspension by shaking. Other formulations as creams and suppositori es can also significa ntly be affected by phase conversion. In those cases it is essential to determine the solubility of the API in its vehicle as phase interconversion may result in nucleation which could lead to pharmaceutically and cosmetically unacceptable preparations. In case of creams, use of an unsuitable polymorph might lead to phase conversion and precipitation of a more st able phase. Significant partic le growth could arise in the vehicle yielding grainy particles, which might result in cosmetically unacceptable creams in which the active ingredient is unevenly distributed. Some suppository bases are polymorphic a nd the selection of an appropriate polymorph could affect the melting characteri stics of the preparation. The low melting phase may become softer or even liquefy during the shelf life. Being very soft may preclude the ability to administer that dos age form. On the other hand, formulation of suppositories with a high melting phase coul d result in a hard product which may not melt upon administration.72 This position could be demonstrated by Theobroma oil84 as a suppository base exhibiting polymorphism. The metastable -form has a melting point of 30C, while the thermodynamically stable -form has a higher melting point. Using the appropriate method of manufacture is essential to permit the crystall ization of the more stable, higher melting point -form. From the above discussion, it can be seen that investigation of polymorphism and selection of the suitable polymorph is cri tical in pharmaceutical industry. Therefore,

PAGE 84

69 polymorphism can be seen as an opportunity to tailor tune preparations. Since the thermodynamically stable form has lower solub ility than the metastable form, retaining the metastable form can improve the effi cacy of some pharmaceutical preparations.77 3.1.2. Polymorphism and patents in pharmaceutical industry As we mentioned previously, polymorphs ar e different in their crystal structures and thereby they can vary in their physic al properties. Hence, discovery of a new polymorph of an API can be considered as an opportunity to improve its physical characteristics. Furthermore, the new form may be considered as a new invention and thus it might be awarded a patent. There are three criteria to patent a new entity. It has to be novel, useful and nonobvious. New polymorphic form is novel as it has a distinct solid-state structure. It might be useful if it provides improved physical pr operties of a certain API. It is also nonobvious since many efforts to predict new polymorphic crystal st ructures are not successful so far. 70 According to the Oxford English Dictionar y, patent is defined as “a license to manufacture, sell, or deal in an article or co mmodity, to the exclusion of other persons; in modern times, a grant from the government to a person or persons conferring for a certain definite time the exclusive pr ivilege of making, using, or se lling some new invention”. This definition implies that the purpose of th e patent is to prevent other producers from making, using or selling the invention cove red by the patent, and thereby many legal consequences can result upon patent infringement.

PAGE 85

c o R i n c o p r c o i n i n m h y u a v a l a c t o One o o ncerning r a R anitidine h y n its initial p Poly m o ntaining h y r ocess had s o mpressed g n methylate d n dustrial pr o m ethylated s p Cons e y drochlorid e sing concen t v oided the c l lowed usin g c etate, whic h o be less hy g o f the well-d a nitidine hy d y drochloride olymorph, l a Fi gu m orph I was y drogen chl o s everal disa d g as which is d spirit and e o duction req u p irit, handli n e quently, Gl a e preparatio n t rated hydr o c onsequence s g a single so h can be eas g roscopic th a ocumented l d rochloride ( was patent e a ter designa t u re 3.6. Mo l prepared b y o ride follow e d vantages. H not easy to h e thyl acetate u ires prepar a n g the hydro a xo worked n In 1980, G o chloric acid s of using h y lvent rather ily recovere a n the earlie r 70 l egal cases o ( Zantac) w h e d directly a f t ed polymo r l ecular str u y dissolving r e d by addin g H ydrogen chl h andle. Mo r has to be fr e a tion of larg e gen chlorid e on develop i G laxo succe e instead of h y drogen chl o than the mi x d after prep a r one. o f patented p h ich is used f ter discove r r ph I. u cture of ra n r anitidine b a g ethyl acet a oride is cor r r eove r the s o e shly prepa r e quantities e caused ser i i ng alternati v ed ed to prep h ydrogen ch l o ride gas. T h x ture of met h a ration. Fur t p olymorphic for treatme n r y in 1977 b y n itidine a se in methy l a te to the sol u r osive and s u o lution of h y r e d Further m of the hydr o i ous proble m v e metho d s f are ranitidi n l oride gas a n h e new proc e h ylated spir i t hermore, th e drugs is the n t of peptic u y Glaxo Gr o l ated spirit u tion. This u pplied as y drogen chl o m ore, since t o gen chlorid e m s.85 f or ranitidin e n e hydrochl o n d thus they e dure also i t and ethyl e product fo u case u lce r o up o ride t he e in e o ride u nd

PAGE 86

71 During characterization of the later produc t, it found to have different IR and XRPD than the earlier product and thus Glaxo concluded that it is a different polymorph of ranitidine hydrochloride, designated form II. Given the advantages of form II over form I, Glaxo decided to use form II as the active ingredient of Zantac. Appropriate commercial proce dure for form II preparation has been developed86 and new patent application has been filled by Glaxo. As Zantac sales reached around $3.5 billion annually by 1991; other pharmaceutical firms looked forward to take a share of these profits by marketing form I of Zantac in 1995, the expiration date of the first patent. Novopharm Ltd. chemists worked on preparation of form I but they ended with form II. Novopharm claimed that since form II produced by following the pr oduction procedure in the first patent, therefore, the product was always form II and the second patent is invalid. Thus, Novopharm sought to market form II and fille d an abbreviated new drug application at the FDA. Novopharm notified Glaxo that the se cond patent is invalid and therefore Glaxo sued Novopharm for infringement of the s econd patent. Glaxo proved by evidence that following the procedure described in the first pa tent leading to form I rather than form II and thus, the court decided that the second patent is valid. Thereafter, Novopharm worked faithfully and succeeded in producing form I of ranitidine hydrochloride, and filled an a bbreviated new drug application. Glaxo sued Novopharm claiming that the product is a mixtur e of forms I and II, rather than a pure form I. Novopharm proved that the produc t is 99% form I with about 1% impurities

PAGE 87

72 which may include form II and therefore the court allowed Novopharm to market the product. From the above discussion it is clear th at the consequences of obtaining a new form of polymorphic drug can go beyond the dire ct effects on physical properties to be a source of litigation. 3.1.3. Conformational polymorphism The term conformational polymorphism wa s first introduced by Corradini (1973) to describe the phenomenon in which flexib le molecules crystallize in different conformations. From conformati onal perspective, organic mol ecules can be seen as rigid or flexible. Rigid molecules can adopt pack ing polymorphism while flexible molecules might be able to adopt conformationa l polymorphism. This phenomenon can be explained in the light of the intramolecula r rotation energy about single bonds. Since this energy is relatively low, 1-3 kcal mol-1 87, molecules with torsiona l degrees of freedom can adopt several conformations in solutions88 and thus they might be able to crystallize in different conformations. It is also important to notice that the ac tivity of the biological molecules and APIs depends mainly on their conformations, thus, investigation of different conformational polymorphs is essential in drug development.88

PAGE 88

73 Figure 3.7. General scheme shows different types of polymorphs83 3.1.4. Disappearing polymorphs Bernstein et al .89 used the term “disappearing polymorphs” to describe the situations when a particular polymorph is no longer obtained. This phenomenon is widely recorded in literature where it is difficult to reproduce a metastable form after nucleation of a thermodynamically more stable form.

PAGE 89

74 The crystallization process composed of formation of nuclei followed by their growth. The nucleation step is the determin ant step in crystallization. One of the commonly used techniques to induce growth of a certain polymorph is by intentional seeding. In some cases seeding might happen unintentionally and lead to formation of undesirable polymorph. This unint entional seedin g may result from contamination with seeds of the more stable form. The area aff ected by contamination might be small as a one lab but it can expand to a town, country or in some cases it might lead to “universal seeding”.89 Since different forms of a polymorphic compound can be produced by variations in some experimental conditions as temp erature and pressure, Jacewicz and Nayler90 argued that disappearing of a polymorph is temporary and that a given form can reappear by finding the appropriate experimental condit ions. They concluded that “Any authentic crystal form should be capabl e of being re-prepared, although selection of the right conditions may require some time and trouble”.90 The phenomenon of disappearing polymor phism can also be explained by applying Ostwald rule of stages91 which explains how interconvert happens in case of polymorphism. The rule states that duri ng crystallization of a polymorphic compound, the least stable form crystallizes first follo wed by more stable phases. The conversion of the metastable form to the more stable form occurs step by step to result finally in crystallization of the most stable form.

PAGE 90

3 s o u n 3 w s l d 3 .2. Experi m All t h o urces and u n modified a .2.1. Salic yl 13.7 m w ere dissolv e l owly evapo ays. .2.2. Iso -ni c m ental h e chemicals u sed as recei a tmosphere a l amide pol y Fi g u r m g (0.0999 m e d in 5ml of rate at roo m c otinamide p Fi g ure used durin g ved. All cry a nd the solv e morph II ( D r e 3.8. Mol e m mol) of sa l methanol b y m temperatur e p ol y morph 3.9. Molec u 75 g the prepar a stallization e e nts were di s D A008) e cular stru c l icylamide a y heating on e The cryst a III (DA167 u lar struct u a tion were o b e xperiments s tilled prior t c ture of sali c a nd 112.8 m g a hotplate, f a ls of DA00 12) u re of iso -ni c b tained fro m were cond u t o use. cy lamide g (0.300 m m f iltered and a 8 were obta i c otinamide m commerci a u cted in an m ol) of ribof l a llowed to i ned after t w a l l avin w o

PAGE 91

3 r e h C 3 1 a c t e 3 a z a n a f 6.6 mg (0.3 0 e sveratrol w e ours follow e C rystals of D .2.3. Trans 1.4 mg (0.0 4 c etate and h e e mperature. .2.4. Citric 76.8 m z odicarbon a n d allowed t f ter two day 0 0 mmol) o f e re dissolve d e d by gradu a A16712 we r resveratrol Fi g ure 3 4 99 mmol) o e ated on a h o Crystals of D a cid mono h m g (0.400 m a mide were d t o slowly ev a s. f isonicotin a d in 5ml of e a l cooling a n r e obtained a pol y morp h 3 .10. Molec u o f trans -res v o tplate. The D A2116 we r hy drate pol y m mol) of citr i d issolved in a porate at r o 76 a mide and 1 e thyl acetat e n d allowed t o a fter one da y h II (DA211 6 u lar struct u v eratrol wer e solution all o r e obtained a y morph II ( i c acid and 2 5ml of met h o om temper a 1.4 mg (0.0 4 e by heating o slowly ev a y 6 ) u re of trans e dissolved i n o wed to slo w a fter one da y C-azod) 2 3.2 mg (0.2 h anol by hea t a ture. Cryst a 4 99 mmol) o on a hotplat a porate at ro o resveratro l n 5ml of ac e w ly evapora t y 00 mmol) o f t ing on a ho t a ls of C-azo d o f trans e (60C) fo r o m tempera t l e tonitrile or e t ing at roo m f t plate, filter e d were obtai n r 12 t ure. e thyl m e d n ed

PAGE 92

77 3.3. Results and Discussion 3.3.1. Salicylamide polymorph II (DA008) Salicylamide [2-hydroxybenzamide] is an active ingredient in BC Powder, which is over-the-counte r pain relief product.92 The CSD search shows one crystal structure of salicylamide (SALMID) w ith no polymorphic structures. However the presence of the dihedral angle of the amide-benzene ring planes suggests that polymorphism of salicylamide may occur. The crystals of DA008 were obtained during an attempt to prepare a cocrystal of riboflavin and salicylamide using methanol as a solvent. The crystal structure of DA008 consists of two crystallographically-indepe ndent molecules of salicylamide in the asymmetric unit. The two molecules differ ma inly in the dihedral angle of the amidebenzene ring planes, 15.90o and 2.51o, in molecules A and B, respectively. The two molecules form noncentrosymmetric am ide dimers with bond distances (NH2CO) of 2.867(6) and 2.910(6) . The amide molecules do not form catemers because the amide carbonyl is involved in intramolecula r hydrogen bonding with the phenolic OH, OHaCOa (D = 2.511(5) ) and OHbCOb (D = 2.497(5) ), while the amide amine is involved in intermolecular hydrogen bonding with the neighbor phenolic OH, OHaNH2a (D = 2.891(6) ).Chains of salicylamide mole cules result from phenolic OH and amide NH2 (D = 2.891(6) ) intermolecular H-bond interactions of molecules A.

PAGE 93

78 Figure 3.11. Comparison of the dihedral angle in DA008 molecules A and B (left) with that of SALMID (right) The structure shows sheets formed by amide dimers of molecules A and B. Molecules of A are further linked where stacking occurs between the aromatic rings (face-to-face interacti ons) with interplanar distance of 3.480 . Further interactions towards neighboring sheets occur through CHinteractions between molecules A and B, CHacentroidb (D= 3.781 ). Molecules B are furthe r connected to molecules B in the other sheets through CHbCOb with a distance of 3.267 . The overall hydrogen bonding results in sandwich herringbone patter n as represented in Figure 3.13.

PAGE 94

79 Figure 3.12. Comparison of hydrogen bonding in DA008 (up) to that in SALMID (down) The crystal structure of SALMID is monoclinic with one molecule of salicylamide in the asymmetric unit. The dihe dral angle of the amide-benzene ring planes is 4.62. The salicylamide molecules form centrosymmetric amide dimers with bond distances (NH2CO) of 2.940 . The amide molecules do not form catemers because the amide carbonyl is involved in intramolecu lar hydrogen bonding with the phenolic OH, OHCO (D = 2.492 ), while the amide amin e is involved in intermolecular hydrogen bonding with the neighbor phenolic OH, OHNH2 (D = 2.812 ). Chains of salicylamide molecules result from phenolic OH and amide NH2 (D = 2.812 ) intermolecular H-bond interactio ns of salicylamide molecules. The sheets are linked

PAGE 95

80 through CHstacking, CHcentroid (D = 3.721 ) and through stacking with interplanar distances of 3.261 and 3.346 . The overall hydrogen bonding results in herringbone pattern as shown in figure 3.16. Figure 3.13. Illustration of the sandwich herringbone pattern in DA008 (left) and the herringbone pattern in SALMID (right) Figure 3.14. Illustration of face to face in teractions in DA008 (left) and that in SALMID (right) The CSD search shows that 16.32% (3 9/239) of primary aromatic amide structures are polymorphic. This polymorphism originates mainly from their flexibility due to the dihedral angle of the amide-arom atic ring planes. Since polymorphism in most

PAGE 96

81 of these cases result from variation in c onformation rather than the packing, those structures exhibit conformational polymor phism. The histogram shown in figure 3.15 reveals the dihedral angle of polymorphs of primary aromatic amides and it shows 32 structures with minimum of 0.681, ma ximum of 75.833, and the mean is 24.944. Furthermore, The CSD search for the dihedral angle of primary aromatic amides shows 228 structures with minimum of 0.034 and maximum of 86.334, while the mean is 21.536. The histogram of the dihedral angle is shown in figure 3.16. The dihedral angles of molecule B in DA008 and that of SALMID are within the range of majority of the CSD entries, while that of molecule A in DA008 does not locate in the average value. However, the dihedral angle of molecule A of DA008 is still has accepted value. Figure 3.15. Histogram for the dihedral angl e of polymorphs of primary aromatic amides

PAGE 97

82 Figure 3.16. Histogram for the dihedral angle of primary aromatic amides As mentioned earlier, the CSD search reve als 239 crystal structures with primary aromatic amides. 144 entries (60.25%) exhi bit amide dimers with bond range of 2.7053.189 (mean 2.927 ). Amide catemers were observed in 177 en tries (74.06%) with bond range of 2.596-3.242 (mean 2.940 ). The hi stograms are shown in figures 3.17. and 3.18., respectively. This indicates that amid e catemers are more dominant than amide dimers in primary aromatic amides. Howeve r, no amide catemers were observed in any of the two polymorphs because the amide carbonyl is involved in intramolecular hydrogen bonding with the phenolic OH. Hence, there are no available amide moieties to form catemers. Further CSD analysis shows 26 structures with both primary aromatic amide and phenol. Only five structures do not cont ain any competitive hydrogen bonding donors or acceptors, namely HXBNZM, HXBNZM01, SALMID, SALMID01, and VIDMAX. The phenolic OH in all of those structures is bifurcated in hydrogen bonding with amide moieties forming OHNH2 and OHCO supramolecular he terosynthons which is also

PAGE 98

83 observed in DA008 and SALMID. All hydrog en bonding distances are within the distance ranges discussed above and that discussed in Chapter 2. concerning phenol hydrogen bonding interactions. Figure 3.17. Histogram for prim ary aromatic amide dimers Figure 3.18. Histogram for primar y aromatic amide catemers Crystals of DA008 were obtained concomita ntly with SALMID crystals, and they transformed later to SALMID crystals. Effo rts to reproduce DA008 cr ystals resulted in obtaining crystals of SALMID which indicate s that SALMID is the more stable form.

PAGE 99

84 The difficulty to obtain DA008 crystals may result from contamination by SALMID seeds. This indicates that SALMID is the thermodynamically favored while DA008 is the kinetically more favored form according to Ostwald rule of stages.91 This hypothesis can be supported by the fact that DA008 has the highest Z value among the known structures of salicylamide.93 This case of polymorphism originates from flexibility of salicylamide molecule, thus, it can be considered as conforma tional polymorphism. At the same time both polymorphs adopt different packing, so, it can also be considered as packing polymorphism. 3.3.2. Iso -nicotinamide polymorph III (DA16712) Isonicotinamide [pyridine-4-carboxamide] is a nicotinamide analogue has shown that it can enhance the ac tivity of antitumor drug, 2-amino-l,3,4-thiadiazole.94 The CSD search shows two polymorphs of isonicotinamide (CSD refcodes: EHOWIH01 and EHOWIH02). Both structures are monoclinic and they differ in the dihedral angle of the amide-benzene ring planes, the number of molecules in the asymmetric unit and in thei r hydrogen bonding patterns. The crystals of DA16712 were obtained during attempts to prepare cocrystal of trans -resveratrol and isonicotinamide using ethyl acetate as a solvent. The crystal structure of DA16712 is orthorhom bic with one molecule of isonicotinamide in the asymmetric unit. The dihedral angle of the amide-benzene ring planes is 31.75. The isonicotinamide molecules are extended through amide catemers with bond distances (NH2CO) of 3.031(3) . The isonicotinamide molecules are further forming head to

PAGE 100

85 tail chains through amide NH2Narom (D = 2.995(3) ). The overall structure shows herringbone pattern sustained by CHinteractions, CHcentroid (D = 3.664 ). Figure 3.19. Comparison of the dihedr al angle in DA16712 (left), EHOWIH01 (middle), and molecules A and B in EHOWIH02 (right) Crystals of EHOWIH01 were pr epared by recrystallization of isonicotinamide from methanol, nitrobenzene or nitromethane.95 The crystal structure of EHOWIH01 is monoclinic with one molecule of isonicotinamide in the asymmetric unit. The dihedral angle of the amide-benzene ring planes is 32.41. The isonicotinamide molecules form centrosymmetric amide dimers with bond distances (NH2CO) of 2.9366(16) . The amide molecules are further extended th rough amide catemers with bond distances (NH2CO) of 2.9354(14) . The isonicotinamide molecules form extended sheets sustained by weak hydrogen bonding interacti ons between aromatic CH and adjacent aromatic nitrogen, CHNarom (D = 3.4275(19) and 3.4790(19) ), and by stacking with interplanar distances of 3.477.

PAGE 101

86 Figure 3.20. Comparison of hydrogen bonding in DA16712 (up), EHOWIH01 (middle), and in assemblies composed of molecules A and B respectively in EHOWIH02 (down) Crystals of EHOWIH02 were pr epared by recrystallization of isonicotinamide from different solvents; etha nol, water, THF, dioxane, etc.95 The crystal structure of

PAGE 102

87 EHOWIH02 is monoclinic with two molecules of isonicotinamide in the asymmetric unit. The two molecules differ mainly in th e dihedral angle of the amide-benzene ring planes, 25.44 and 24.82, in molecules A and B, respectively. Each of the two molecules give rise to an independent assembly, each contain the same hydrogen bonding pattern with subtle variations in bond distances. The isonicotinamide molecules are extended through amide catemers with bond distances NH2aCOa (D = 2.934(2) ) and NH2bCOb (D = 2.947(3) ). The isonicotinamide molecules are further forming head to tail chains through amide amine and aromatic nitrogen, NH2aNaroma (D = 2.982(3) ) and NH2bNaromb (D = 2.974(3) ). The overall stru cture shows herringbone pattern sustained by CHinteractions, CHacentroidb (D = 3.902 ) and by stacking between B molecules with inte rplanar distances of 3.539. Figure 3.21. Illustration of the herringb one pattern in DA16712 (left), sheets of EHOWIH01 (middle), and the herringbo ne pattern in EHOWIH02 (right) In contrast to salicylamide pol ymorphs, all three polymorphs of isonicotinamide exhibit amide catemers because there is no interference with such hydrogen bonding pattern. This can be predicted from CSD analysis discussed earlier in section 3.3.1. However, amide dimers were observed in EHOWIH01 in addition to amide catemers.

PAGE 103

88 This is also anticipated sin ce occurrence of amide dimers is reported in 60.25% of crystal structures of primary aromatic amides in CSD. Further CSD analysis reveals 99 structures with both primary aromatic amide and aromatic nitrogen. 22 of those structures do not contain any competitive hydrogen bonding donors or acceptors, and 72.73% (16/22) exhibit amide NH2Narom supramolecular heterosynthons. This is in accordance with reported structures of DA16712 and EHOWIH02. However, this was not the case in EHOWIH01 crystal structure. Overall, all hydrogen bonding distan ces and dihedral angles are located within the expected range based on the CSD sear ch discussed earlier in section 3.3.1. Efforts to reproduce DA16712 crystals re sulted in obtaining crystals of EHOWIH01 which indicates that EHOWIH01 is more stable form. The difficulty to obtain DA16712 crystals may result from contamination by EHOWIH01 seeds. This indicates that EHOWIH01 is the ther modynamically favored while DA16712 is the kinetically more favored form. This case of polymorphism originates from flexibility of isonicotinamide molecule, thus, it can be considered as c onformational polymorphism. At the same time all polymorphs adopt different packing, s o, it can also be considered as packing polymorphism. 3.3.3 Trans -resveratrol polymorph II (DA2116) Trans -resveratrol [ trans -3,5,4'-trihydroxystilbene] is a nutraceutical with antioxidant properties and many promising appl ications in medicine as discussed in

PAGE 104

89 section 2.2.1. The CSD search s hows one crystal structure of trans -resveratrol (DALGON) with no polymorphic structures. Ho wever the presence of the dihedral angle of the two aromatic ring planes suggests that polymorphism of trans -resveratrol may occur. The crystal structure of DA2116 is monoclinic with two molecules of trans resveratrol in the asymmetric unit. The dihedr al angles of the two aromatic ring planes are 8.51 and 2.17. Each A molecule of trans -resveratrol is hydrogen bonded to other six molecules of molecules B of trans -resveratrol through supramolecular homosynthons, OHAOHB. Two of these molecules are almost on the same plane with the indicated trans -resveratrol molecule while th e other four are aligned perp endicular to that plane. The bond distances are: OHAaOHBb (D = 2.668(2) ), OHAaOHBc (D = 2.706(2) ), OHAbOHBa (D = 2.710(2) ), OHAbOHBb (D = 2.677(2) ), OHAcOHBa (D = 2.715(2) ), and OHAcOHBc (D = 2.709(2) ). In the crystal structure, sheets of molecules A and B, alternatively stacke d, with interplanar distance of 3.210 are observed. The crystal structure of DALGON75 is monoclinic with one molecule of trans resveratrol in the asymmetric unit. The dihedr al angle of the two aromatic ring planes in DALGON is 5.33. Each molecule of trans -resveratrol is hydroge n bonded to other six molecules of trans -resveratrol through supramolecular homosynthons, OHOH. Two of these molecules are almost on the same plane with the indicated trans -resveratrol molecule while the other four are aligned pe rpendicular to that pl ane. The bond distances are: OHaOHb (D = 2.685 ), OHaOHc (D = 2.687 ), OHbOHb (D = 2.727 ), and

PAGE 105

90 OHcOHc (D = 2.754 ). The molecules are further connected through stacking with interplanar distance of 3.305 . Figure 3.22. Comparison of hydrogen bo nding in DA2116 (left) and that of DALGON (right) Figure 3.23. Illustration of the packin g in DA2116 (left) and that in DALGON (right) From the above discussion, it is clear that both DA2116 and DALGON adopt the same hydrogen bonding pattern with subtle variations in hydrogen bond distances. All hydrogen bonding distances are within the expected range based on CSD analysis discussed in section 2.2.2. However, the presence of OHOH supramolecular homosynthon was expected due to the abse nce of any other moieties which can participate in hydrogen bonding. Furthermore, CSD search reveals that in absence of any other competing hydrogen bonding moieties, 59.72% out of 216 entries shows OHOH supramolecular homosynthons.

PAGE 106

91 Since trans -resveratrol is a flexible molecule, both structures differ mainly in their conformations, and this is an obvious ex ample of conformational polymorphism. 3.3.4. Citric acid monohydrate polymorph II (C-azod) Citric acid is a nutraceutical with antioxidant properties and many pharmaceutical and industrial applications as discussed in section 2.3.1. Citric acid can present as anhydrous [2-hydroxy-1,2,3-propanetricar boxylic acid] or monohydrate [2-hydroxy1,2,3-propanetricarboxylic acid monohydrate]. However, CSD search reveals two polymorphs of citric acid anhydrate and only one structure of citr ic acid monohydrate. We record herein a new polymorph of citric acid monohydrate (C-azod). The crystal structure of C-azod is triclini c with one molecule of citric acid and one molecule of water in the asymmetric unit. Each molecule of citric acid forms two distinct centrosymmetric acid dimers with two citric acid molecules, OHCO (D = 2.625(5) and 2.645(5) ). The indicated mo lecule is hydrogen bonded to two water molecules, acidic OHacidOHwater (D = 2.675(5) and 3.035(8) ), and OHalcoholOHwater (D = 3.024(7) ). There is intram olecular hydrogen bonding between OHalcoholCOacid (D = 2.903(8) ). The overall structure shows sheets of citric acid and water molecules linked to other sheets through the same hydrogen bonding patterns in addition to OHalcoholCOacid (D = 2.875 ). The crystal structure if CITA RC is orthorhombic with one molecule of citric acid and one molecule of water in the asymmetric unit. Trimers formed of two citric acid molecules with one water molecule were observed, OHacidCOacid (D = 2.755 ),

PAGE 107

92 OHacidOHwater (D = 2.623 ), and OHwaterCOacid (D = 2.761 ). That water molecule is further linked to the citric acid molecule in the next layer, OHacidOHwater (D = 2.695 ).the alcohol moieties also share in connecting the layers, OHalcoholCOacid (D = 2.765 ) and OHalcohol OHwater (D = 2.768 ). There is in tramolecular hydrogen bonding between OHacidCOacid (D = 3.020 ). The overall structur e shows sheets of citric acid and water molecules connected to othe r sheets through the same hydrogen bonding patterns. Figure 3.24. Comparison of hydrogen bonding in C-azod (left) and that of CITARC (right) Comparing the two structures, both struct ures form two dimensional sheets which extend to form three dimensional structur es. But it is obvious that C-azod crystal structure is mainly controlled by acid dimers while water molecules are present as guest molecules. On the other hand, water molecule s in CITARC share with two molecules of citric acid to form trimers. However, ba sed on CSD analysis discussed in section 2.3.2., acid dimmer occurs in 31.35% of CSD entrie s in presence of other competing groups. Furthermore, OHalcoholCOacid supramolecular heterosynthons is reported in 44.47% of CSD entries in presence of other competing groups. This supramolecular heterosynthons

PAGE 108

93 observed in the two polymorphs of citric acid monohydrate. Moreover, CSD search reveals 5050 structures of organic hydrat es, 43 (0.85%) of those structures exhibit polymorphism. On the other hand, polymorphism in single component structures is more common, 1.4% (1882/134,086). Since the main difference between the two pol ymorphs is in their packing, this is considered as a case of packing polymorphism. 3.4. Conclusion The four cases of polymorphism described in this chapter dem onstrate two widely encountered types of polymorphism, namely conformational and packing polymorphism. Molecular flexibility in salicylamide, isonicotinamide, and transresveratrol resulted in isolation of conformational polymorphs while citric acid monohydrate exhibits packing polymorphism. Generally, new polymorphs were obtained by serendipity, but systematic study of polymorphism can also be performed by vary ing some factors such as; temperature, pressure, solvent, and additives. Polymorphism has a significant importan ce in industry, in general, and in pharmaceutical industry, in particular, due to th e vast differences in physical properties of polymorphs. Furthermore, the study of polym orphism provides valuable information essential to understand how different crystal forms are attained. Moreover, Aakeroy et al.95 argued that polymorphic com pounds are considered as good cocrystal formers since they have synthons flexibility.

PAGE 109

94 References (1) R. Pepinsky Phys. Rev. 1955 100 971. (2) G. M. Schmidt, Pure Appl. Chem. 1971 27 647. (3) G. R. Desiraju, Crystal Engineering—The Design of Organic Solids Elsevier, Amsterdam, 1989 (4) Latimer, W. M.; Rodebush, W. H. J. Am. Chem. Soc. 1920 42 1419-1433. (5) E. J. Corcy. Pure. Appl. Chem. 1967 14, 19. (6) Gautam R. Desiraju, Angew. Chem. Int. Ed. 1995 34 2311-2327. (7) Walsh, R. D. B.; Bradner, M. W.; Fleischman, S.; Morales, L. A.; Moulton, B.; Rodriguez-Hornedo, N.; Zaworotko, M. J. Chem. Commun. 2003 186-187. (8) Almarsson, O.; Zaworotko, M. J. Chem. Commun 2004 1889-1896. (9) Fleischman, S. G.; Kuduva, S. S.; McMahon, J. A.; Moulton, B.; Bailey Walsh, R. D.; Rodriguez-Hornedo, N.; Zaworotko, M. J. Cryst Growth Des 2003 3 909-919. (10) Leiserowitz, L., 1976 32 775-802. (11) Leiserowitz, L.; Schmidt, G. M. J. J. Chem. Soc. 1969 2372. (12) Leiserowitz, L.; Nader, F. 1977 33 2719-2733. (13) Vishweshwar, P.; Nangia, A.; Lynch, V. M. J. Org. Chem. 2002 67 556-565. (14) Vishweshwar, P.; Nangia, A.; Lynch, V. M. CrystEngComm 2003 5 164-168. (15) Feynmaan, R. Eng. Sci. 1960 22-36. (16) Etter, M. C. Acc. Chem. Res. 1990 23 120. (17) Etter, M. C. J. Phys. Chem. 1991 95 4601.

PAGE 110

95 (18) Desiraju, G. R. ed. The Crystal as a Supramolecula r Entity: Perspectives in Supramolecular Chemistry 2, Wiley and Sons, Chichester, 1996 (19) www.ccdc.cam.ac.uk/products/csd/ (20) http://www.ccdc.cam.ac.uk/ products/csd/statistics/ (21) Allen, F., H.; Kennard, O.; Taylor, R. Acc. Chem. Res. 1983 16, 146-153. (22) Shan, N.; Zaworotko, M. J. Drug Discov. Today 2008 13 440-446. (23) Ling, A. R.; Baker, J. L. J. Chem. Soc 1893 ; 63 1314. (24) Buck, J. S.; Ide, W. S. J. Am. Chem. Soc 1931 53 2784. (25) Anderson, J. S. Nature. 1937 140 583. (26) van Niekerk, J. N.; Saunder, D. H. 1948 1 44. (27) Hall, B.; Devlin, J. P. J. Phys. Chem 1967 71 465. (28) Pekker, S.; Kovats, E.; Oszlanyi, G.; Benyei, G.; Klupp, G.; Bortel, G.; Jalsovszky, I.; Jakab, E.; Borondics, F.; Kamaras, K.; Bokor, M.; Kriza, G.; Tompa, K.; Faigel, G. Nat Mater. 2005 4 764-767. (29) Desiraju, G. R. CrystEngComm. 2003 5 466-467. (30) Dunitz, J. D. CrystEngComm. 2003 5 506-506. (31)Whler F. Annalen 1844 51 153. (32) Sakurai, T., 1965 19 320-330. (33) Sakurai, T. Acta Crystallogr. 1968 B 24 403. (34) Hoogsteen, K., 1959 12 822-823. (35) Hoogsteen, K. Acta Crystallogr 1963 16, 907–916. (36) Etter, M. C. J. Phys. Chem 1991; 95 4601-4610. (37) Hickey, M. B.; Peterson, M. L.; Scoppettu olo, L. A.; Morrisette, S. L.; Vetter, A.; Guzmn, H.; Remenar, J. F.; Zhang, Z.; Tawa, M. D.; Haley, S.; Zaworotko, M. J.; Almarsson, . Eur. J. Pharm. Biopharm. 2007 67 112-119.

PAGE 111

96 (38) L. Bertilson, T. Tomson, Clin. Pharmacokinet 1986 11, 177–198. (39) M.C. Meyer, A.B. Straughn, E.J. Jarv i, G.C. Wood, F.R. Pelsor, V.P. Shah. Pharm. Res 1992 9, 1612–1616. (40) Eur. J. Pharm. Biopharm 2007 67 112-119. (41) Amidon GL, Lennernas H, Shah VP, Crison JR. Pharm Res 1995 12, 413-420. (42) http://www.fda.gov/ (43) http://www.phcog.org /definition.html (44) Lockwood, B. Nutraceuticals ; Pharmaceutical Press: London, UK, 2007 (45) Espn, J. C.; Garca-Conesa, M. T.; Toms-Barbern, F. A. Phytochemistry 2007 68 2986-3008. (46) Jang, M.; Cai, L.; Udeani, G. O.; Slowing, K. V.; Thomas, C. F.; Beecher, C. W. W.; Fong, H. H. S.; Farnsworth, N. R.; Kinghorn, A. D.; Mehta, R. G.; Moon, R. C.; Pezzuto, J. M. Science 1997 275 218-220. (47) Trask, A. V. Molecular Pharmaceutics 2007 4 301-309. (and references within). (48) J.A. Baur; D.A. Sinclair Nat. Rev. Drug Discov. 2006 493-506. (49) Crowell, J. A.; Korytko, P. J.; Morri ssey, R. L.; Booth, T. D.; Levine, B. S. Toxicol. Sci. 2004 82 614-619. (50) Aoyama, Y.; Endo, K.; Anzai, T.; Yamaguchi, Y.; Sawaki, T.; Kobayashi, K.; Kanehisa, N.; Hashimoto, H.; Kai, Y.; Masuda, H. J. Am. Chem. Soc. 1996 118 5562-5571. (51) MacGillivray, L. R.; Reid, J. L.; Ripmeester, J. A. J. Am. Chem. Soc. 2000 122 7817-7818. (52) Papaefstathiou, G. S.; MacGillivray, L. R. Org. Lett. 2001 3 3835-3838. (53) MacGillivray, L. R.; Atwood, J. L. J Sol. State Chem. 2000 152 199-210. (54) Schmidt, G. M. J. Pure Appl. Chem. 1971 27 647. (55) Childs, S. L.; Chyall, L. J.; Dunlap, J. T.; Smolenskaya, V. N.; Stahly, B. C.; Stahly, G. P. J. Am. Chem. Soc. 2004 126 13335-13342.

PAGE 112

97 (56) Remenar, J. F.; Morissette, S. L.; Pe terson, M. L.; Moulton, B.; MacPhee, J. M.; Guzman, H. R.; Almarsson, O. J. Am. Chem. Soc. 2003 125 8456-8457. (57) Good, D. J.; Rodrguez-Hornedo, N. Cryst. Growth Des. 2009 9 2252-2264. (58) Kristina L. Penniston, Stephen Y. Naka da, Ross P. Holmes, Dean G. Assimos. J. Endourol 2008 22 567-570. (59) Nyknen, P.; Lemp, S.; Aaltonen, M.; Jrjenson, H.; Veski, P.; Marvola, M. Int. J. Pharm. 2001 229 155-162. (60) http://vm.cfsan.fda .gov/~dms/eafus.html (61) Weyna, D. R.; Shattock, T.; Vishweshwar, P.; Zaworotko, M. J. Cryst. Growth. Des. 2009 9 1106-1123. (62) Mohamed, S.; Tocher, D. A.; Vickers, M.; Karamertzanis, P. G.; Price, S. L. Cryst. Growth. Des. 2009 9 2881-2889. (63) Aakeroy, C. B.; Fasulo, M. E.; Desper, J. Mol. Pharma. 2007 4 317-322. (64) Majerz, I.; Malarski, Z.; Sobczyk, L. Chem. Phys. Lett. 1997 274 361-364. (65) Cowan, J. A.; Howard, J. A. K.; McIntyre, G. J.; Lo, S. M.; Williams, I. D. 2003 59 794-801. (66) Jerzykiewicz, L. B.; Malarski, Z.; Sobczyk, L.; Lis, T.; Grech, E. J. Mol. Struct. 1998 440 175-185. (67) Stahl, P. H.; Wermuth, C. G., Ed. Handbook of Pharmaceutical Salts:Properties, Selection, and Use; International Un ion of Pure and Applied Chemistry ; VHCA, Wiley-VCH:Weinheim, NY, 2002. (68) E.Mitscherlich, Ann.Chim.Phys 1822 19 350. (69) Yu, L.; Stephenson, G. A.; Mitchell, C. A. ; Bunnell, C. A.; Snorek, S. V.; Bowyer, J. J.; Borchardt, T. B.; Stowell, J. G.; Byrn, S. R. J. Am. Chem. Soc. 2000 122 585591. (70) Chen, S.; Guzei, I. A.; Yu, L. J. Am. Chem. Soc. 2005 127 9881-9885. (71) W. C. McCrone, Polymorphism in ‘Physics and chem istry of the organic solid state, vol. 2 ed. D. Fox, M. M. Labes and A. Wei ssberger, Wiley Interscience, New York, 1965 pp. 725–767.

PAGE 113

98 (72) Davey, R. J.; Blagden, N.; Potts, G. D.; Docherty, R. J. Am. Chem. Soc. 1997 119 1767-1772. (73) Henck, J.; Bernstein, J.; Ellern, A.; Boese, R. J. Am. Chem. Soc. 2001 123 18341841. (74) Vishweshwar, P.; McMahon, J. A.; Peterson M. L.; Hickey, M. B.; Shattock, T. R.; Zaworotko, M. J. Chem. Commun. 2005 4601-4603. (75) Nangia, A. Acc. Chem. Res. 2008 41 595-604. (76) Jaroslav Nvlt, Crystal Research and Technology 1995 30 443-449. (77) Bernstein, J., Polymorphism in Molecular Crystals ; Oxford University Press: Oxford, 2002 (78) Edgerton, W. H., U.S. pat. 2,662,006 (December 15,1953). (79) Menachemoff, E.; Harokeach, H. 1964 10 300 (80) Milosovich, G., Parke, Davis & C o., unpublished data. Tamura. G., and Kuwano, H. J. Pharm. Soc. Japan 1961 81 755. (81) Maruyama, M., Hayashi, N., and Kishi, M., TakamuteKendyusho Nempo. 1962 13 176. (82) Almirante, L.; DeCarneri, I.; Coppi, G. Farmaco (Pasia), Ed. Pract. 1960 15 471. (83) Nangia, A. Acc. Chem. Res. 2008 41 595-604. (84) A. E. Baily, “Melting and Solidification of Fats,” Inter-science, New York, N. Y., 1950 (85) Declaration of D T Collin 9 August 1982, pp 2-4, Annexe A in an affidavit filed by C J Robertson, Managing Director Apotex, 22 July 1994, in the matter of the Patents Act 1953 and in the matter of an applicat ion for extension of Patent No 184759 in the name of Allen & Hanburys Ltd and in the matter of opposition proceedings by Apotex NZ Ltd. Obtained under th e Official Information Act, 1982 (86) Ibid pp 3-4; Decision of the Commissioner of Patents 11/17, 17.5.96, p 19. (87) (a) Buttar, D.; Charlton, M. H.; Docherty, R.; Starbuck, J. J. Chem. Soc., Perkin Trans.1998 2 1998, 763–772. (b) Starbuck, J.; Docherty, R.; Charlton, M. H.; Buttar, D. A. J. Chem. Soc., Perkin Trans. 1999 2 677–691. (c) Bashkirava, A.; Andrews, P. C.; Junk, P. C.; Robertson, E. G.; Spiccia, L.; Vanderhoek, N. Chem. Asian J. 2007 2 530–538.

PAGE 114

99 (88) Desiraju, G. R. Cryst. Growth Des. 2008 8 3-5. (89) Dunitz, J. D.; Bernstein, J. Acc. Chem. Res. 1995 28 193-200. (90) Jacewicz, V, W.; Nayler, J. H. C. J. Appl. Crystallogr. 1979 12 396. (91) Ostwald, W. Zeits. Phys Chem. 1897 22, 289-330. (92) http://www.bcpowder.com/products_original.aspx (93) Desiraju, G. R. Cryst. Eng. Comm. 2007 9 91-92. (94)Oettgen, H., F.; Purple, J., R.; Coley, V., C.; Krakoff, I., H.; Burchenal, J. H. Cancer. Res. 1964 24 689-692. (95) Aakeroy, C. B.; Beatty, A. M. ; Helfrich, B. A.; Nieuwenhuyzen, M. Cryst. Growth. Des. 2003 3 159-165. (96) Armando J.; Aguiar, J.; Arlyn, W., K. ; Samyn, J., C. J. Pharm. Sci. 1967 56 847853

PAGE 115

100 Appendices

PAGE 116

101 Appendix 1. Experimental data The experiments were performed using DS C (TA instrument 2920), FT-IR (Nicolet Avatar 320 FTIR, solid state), X-ray powder di ffraction (Bruker AXS D8, Cu radiation), and TGA (STM6000). 1.1. Experimental Data for DA182901 DSC thermogram, FT-IR spectrum, and X-ray powder diffraction patterns of bulk sample (down) and calculated from the single crystal structure (up). 106.54C 262.83C -5 -4 -3 -2 -1 0Heat Flow (W/g) 050100150200250300Temperature (C) Sample: DA1829 Size: 4.2000 mg Method: RampDSCFile: C:...\Dalia\DA1829.001 Operator: Dalia Run Date: 13-Jul-08 23:57Exo Up Universal V2.6D TA Instruments 685.05 798.62 819.32 835.56 891.55 968.20 1087.62 1166.32 1198.92 1248.43 1352.11 1442.10 1487.37 1582.19 1623.74 1650.69 2923.57 3351.71 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 %T 1000 2000 3000 4000 cm-1

PAGE 117

102 1.2. Experimental Data for DA182902 DSC thermogram, FT-IR spectrum, X-ray pow der diffraction patterns of bulk sample (down) and calculated from the single crystal structure (up), and TGA. 051015202530354045 0.0 0.2 0.4 0.6 0.8 1.0 Experimental Calculated (1Relative intensity2

PAGE 118

103 686.72 846.58 960.43 1011.03 1167.22 1171.94 1201.77 1333.80 1439.14 1509.19 1578.52 2921.95 3400.65 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 100 102 %T 1000 2000 3000 4000 Wavenumbers (cm-1)

PAGE 119

104 1.3. Experimental Data for RESVONE 01 DSC thermogram, FT-IR spectrum, X-ray pow der diffraction patterns of bulk sample (down) and calculated from the single crystal structure (up), and TGA. 606.35 672.29 720.47 749.09 767.87 847.05 908.31 964.75 1165.87 1257.78 1386.82 1463.40 1515.14 1557.76 1584.45 1615.59 3155.46 76 78 80 82 84 86 88 90 92 94 96 98 %Transmittance 1000 2000 3000 4000 Wavenumbers (cm-1)

PAGE 120

105 051015202530354045 0.0 0.2 0.4 0.6 0.8 1.0 Calculated (29 Experimental (Relative intensity2

PAGE 121

106 1.4. Experimental Data for RESVONE 02 X-ray powder diffraction pattern s of bulk sample (down) a nd calculated from the single crystal structure (up). 1.5. Experimental Data for DA4 DSC thermogram, FT-IR spectrum, X-ray pow der diffraction patterns of bulk sample (down) and calculated from the single crystal structure (up), and TGA.

PAGE 122

107 618.61 799.02 836.11 966.66 998.78 1169.87 1209.18 1272.93 1302.52 1406.05 1512.33 1582.83 2357.62 2587.24 3003.68 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 %T 1000 2000 3000 4000 cm-1 051015202530354045 0.0 0.2 0.4 0.6 0.8 1.0 Calculated ( 2 ExperimentalRelative intensity2

PAGE 123

1 D ( d .6. Experi m D SC thermo g d own) and c m ental Data g ram, FT-I R alculated fr o for DA005 R spectru m X o m the singl e 108 X -ray powd e e crystal str u e r diffractio n u cture (up), a n patterns o f a nd TGA. f bulk sampl e e

PAGE 124

109 051015202530354045 0.0 0.2 0.4 0.6 0.8 1.0 Calculated (1 ExperimentaRelative intensity2 theta degree

PAGE 125

1 D ( d .7. Experi m D SC thermo g d own) and c m ental Data g ram, FT-I R alculated fr o for DA02 R spectru m X o m the singl e 110 X -ray powd e e crystal str u e r diffractio n u cture (up), a n patterns o f a nd TGA. f bulk sampl e e

PAGE 126

111

PAGE 127

112 1.8. Experimental Data for flavanone DSC thermogram, FT-IR spectrum, X-ray pow der diffraction patterns of bulk sample (down) and calculated from the single crystal structure (up), and TGA. 625.73 699.69 764.19 906.63 1066.32 1114.62 1148.20 1225.96 1301.83 1461.18 1604.73 1687.09 74 76 78 80 82 84 86 88 90 92 94 96%Transmittance 1000 2000 3000 4000 Wavenumbers (cm-1)

PAGE 128

113

PAGE 129

114 1.9. Experimental Data for DA16712 DSC thermogram and X-ray powder diffracti on patterns of bulk sample (down) and calculated from the single crystal structure (up).

PAGE 130

115 1.10. Experimental Data for DA2116 DSC thermogram, FT-IR spectrum, X-ray pow der diffraction patterns of bulk sample (down) and calculated from the single crystal structure (up), and TGA. 606.18 623.34 674.00 716.97 804.47 829.27 964.53 986.68 1009.46 1105.92 1145.26 1206.27 1323.82 1381.97 1442.38 1511.02 1584.21 2356.52 3203.73 91 92 93 94 95 96 97 98 99 100 101 %Transmittance 1000 2000 3000 4000 Wavenumbers (cm-1)

PAGE 131

116

PAGE 132

117 Appendix 2. Crystallographic data 2.1. Crystallographic data for transresveratrol-caprolactam polymorphic cocrystal (DA182901 and DA182902) Identification code DA182902 DA182901 Empirical formula C26 H34 N2 O5 C26 H34 N2 O5 Formula weight 454.55 454.55 Temperature 293(2) K 100(2) K Crystal system Orthorhombic Triclinic Space group P n a 21 P -1 Unit cell dimensions a = 12.6665(10) , = 90 a = 8.1803(2) , = 114.627(2) b = 8.2167(5) , = 90 b = 13.0557(3) , = 106.532(2) c = 22.7749(19) , = 90 c = 13.4915(5) , = 96.5990(10) Volume 2370.3(3) 3 1209.76(6) 3 Z, Z` 4, 1 2, 1 Final R indices [I>2sigma(I)] R1 = 0.0638 wR2 = 0.1471 R1 = 0.0403, wR2 = 0.1239 R indices (all data) R1 = 0.0954, wR2 = 0.1729 R1 = 0.0427, wR2 = 0.1280

PAGE 133

118 2.2. Crystallographic data for transresveratrol-flavone polymorphic cocrystal (RESVONE 01 and RESVONE 02) Identification code RESVONE01 RESVONE02 Empirical formula C44 H32 O7 C44 H32 O7 Formula weight 672.70 672.70 Temperature 293(2) K 293(2) K Wavelength 1.54178 1.54178 Crystal system, space group Monoclinic, P 21 Orthorhombic, P c a 21 Unit cell dimensions a = 7.8768(7) , = 90 a = 12.3620(5) , = 90 b = 23.796(3) , = 101.380(6) b = 19.0460(5) , = 90 c = 9.2317(8) , = 90 c = 14.4410(8) , = 90 Volume 1696.4(3) 3 3400.1(2) 3 Z, Z` 2, 1 4, 1 Goodness-of-fit on F2 1.004 0.946 Final R indices [I>2sigma(I)] R1 = 0.0549, wR2 = 0.1224 R1 = 0.0445, wR2 = 0.0985 2.3. Crystallographic data for transresveratrol and 4,4'-dipyridyl cocrystal(DA4) Identification code da4_0m Empirical formula C58 H48 N6 O6 Formula weight 925.02 Temperature 298(2) K Wavelength 1.54178 Crystal system Triclinic Space group P -1 Unit cell dimensions a = 9.4975(11) , = 91.319(7) b = 10.3208(13) , = 95.689(6) c = 26.494(3) , = 108.619(6) Volume 2444.8(5) 3 Z, Z` 2, 1 Data / restraints / parameters 7642 / 0 / 826 Goodness-of-fit on F2 1.000 Final R indices [I>2sigma(I)] R1 = 0.0691, wR2 = 0.1817

PAGE 134

119 2.4. Crystallographic data for flavanone single crystalIdentification code da2720_0m Empirical formula C15 H12 O2 Formula weight 224.25 Temperature 293(2) K Wavelength 1.54178 Crystal system Monoclinic Space group P 21/n Unit cell dimensions a = 10.598(5) , = 90 b = 5.543(3) , = 92.479(10) c = 19.105(9) , = 90 Volume 1121.3(9) 3 Z, Z` 4, 1 Goodness-of-fit on F2 0.795 Final R indices [I>2sigma(I)] R1 = 0.0470, wR2 = 0.1072 2.5. Crystallographic data for citric acidiso -nicotinamide cocrystal, DA005 (left) and that of citric acidiso -nicotinic acid cocrystal dihydrate, DA02 (right) Identification code DA005 da02m Empirical formula C18 H20 N4 O9 C18 H22 N2 O13 Formula weight 416.22 474.38 Temperature 178(2) K 100(2) K Crystal system Monoclinic Triclinic Space group C c P -1 Unit cell dimensions a = 31.39(2) , = 90 a = 8.243(3) , = 87.722(7) b = 5.319(4) , = 101.243(14) b = 10.427(4) , = 87.076(6) c = 11.762(8) , = 90 c = 12.761(5) , = 70.171(7) Volume 1926(2) 3 1030.1(7) 3 Z, Z` 4, 1 2, 1 Goodness-of-fit on F2 1.352 0.984 Final R indices [I>2sigma(I)] R1 = 0.0830, wR2 = 0.1931 R1 = 0.0438, wR2 = 0.1057

PAGE 135

120 Appendix 3. Comparing crystal structures of new polymorphs to that of known polymorphs 3.1. Comparing crystal structures of salicylamide polymorphs Salicylamide form I (SALMID) Salicylamide form II (DA008) Dihedral angles 4.62 15.90o and 2.51o Hydrogen bonding Amide dimers (2.940 ), Intramolecular hydrogen bonding (2.492 ), OHNH2 (2.812 ), CH stacking (3.721 ), stacking (3.261 and 3.346 ) Amide dimers (2.867 and 2.911 ), Intramolecular hydrogen bonding (2.512 and 2.497 ), OHNH2 (2.891 ), CHstacking (3.781 ), CHbCOb (3.267 ), stacking (3.480 ) Space group I 2/a, monoclinic P 21/n, monoclinic Z 8 8 Z 1 2 a () 12.920 6.483(5) b () 4.980 15.735(11) c () 21.040 12.632(9) () 90.00 90.00 () 91.80 100.335 () 90.00 90.00 Volume (A3) 1353.08 1267.68

PAGE 136

121 3.2. Comparing crystal structures of isonicotinamide polymorphs Iso -nicotinamide form I(EHOWIH01) Iso -nicotinamide form II (EHOWIH02) Iso -nicotinamide form III (DA16712) Dihedral angles 32.41 25.44 and 24.82 31.75 Hydrogen bonding Amide catemers (2.935 ), Amide dimers (2.937 ), CHNarom (3.427 and 3.479 ), – stacking (3.477) Amide catemers (2.935 and 2.947 ), NH2Narom (2.982 and 2.974 ), CHinteraction (3.902 ), – stacking (3.539) Amide catemers (3.031 ), NH2Narom (2.995 ), CHstacking (3.664 ) Space group P 21/c, monoclinic P 21/c, monoclinic P bca, orthorhombic Z 4 8 8 Z 1 2 1 a () 10.1756(11) 15.735(3) 10.1725(6) b () 5.7319(6) 7.9976(18) 7.4507(6) c () 10.0340(10) 9.885(3) 15.9013(9) () 90.00 90.00 90.00 () 98.042(7) 105.586(17) 90.00 () 90.00 90.00 90.00 Volume (A3) 579.483 1198.21 1205.2

PAGE 137

122 3.3. Comparing crystal structures of trans -resveratrol polymorphs Trans -resveratrol form I (DALGON) Trans -resveratrol form II (DA2116) Dihedral angles 5.33 8.51 and 2.17 Hydrogen bonding OHOH (2.685, 2.687, 2.727, and 2.754 ), – stacking (3.305 ) OHOH (2.668, 2.706, 2.710, 2.677, 2.715, and 2.709 ), – stacking (3.210 ) Space group P 21/c, monoclinic P 21/n, monoclinic Z 1 2 Z 4 8 a () 4.3791(5) 8.6551(3) b () 9.2158(11) 9.1847(3) c () 26.681(3) 26.6412(9) () 90.00 90.00 () 92.748(2) 93.090(2) () 90.00 90.00 Volume (A3) 1075.52 2114.75(12)

PAGE 138

123 3.4. Comparing crystal structures of ci tric acid monohydrate polymorphs Identification code Citric acid monohydrate form I (CITARC) Citric acid monohydrate form II (C-azod) Hydrogen bonding Acid dimers (2.625 and 2.645 ), OHacidOHwater (2.675 and 3.035 ), OHalcoholOHwater (3.024 ), OHalcoholCOacid (2.903 and 2.875) OHacidCOacid (2.755 and 3.020 ), OHacidOHwater (2.623 and 2.695 ), OHwaterCOacid (2.761 ), OHalcoholCOacid (2.765 ), OHalcohol OHwater (2.768 ) Space group P 21 21 21, orthorhombic P -1, triclinic Z 4 2 Z 1 1 a () 6.297(3) 6.781(3) b () 9.319(3) 7.630(4) c () 15.398(3) 8.788(4) () 90 93.041(10) () 90 106.201(8) () 90 100.819(10) Volume (A3) 903.581 426.1(4)