ENABLING DIRECT COMPRESSION TABLET DEVELOPMENT OF ...

ENABLING DIRECT COMPRESSION TABLET DEVELOPMENT OF

CELECOXIB THROUGH SOLID STATE ENGINEERING

A DISSERTATION

SUBMITTED TO THE FACULTY OF THE

UNIVERSITY OF MINNESOTA

BY

Kunlin Wang

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

Changquan Calvin Sun, Advisor

August 2020

i

Acknowledgements

Life as a Ph.D. student is a journey filled with gratitude. For my 5 years of Ph.D.

study as well as 2 years of undergraduate research at the Sun Lab, the first person I would

like to thank is my advisor, Prof. Changquan Calvin Sun. During my undergraduate

research in my junior and senior years, he has been a good guide and scientist who triggered

my interest in pharmaceutical materials science and lead my way to pursuing a Ph.D.

degree in Pharmaceutics. During my graduate study, Dr. Sun has been a great mentor in

science as well as a supporter who inspire me for attitude and solutions towards life’s

problems.

My appreciation also goes to my committee members, Dr. Raj Suryanarayanan, Dr.

Timothy Wiedmann, and Dr. Andreas Stein for reviewing my dissertation, offering

insightful comments and suggestions, and serving on my Ph. D. exam committees. I have

learned the power of logical thinking and analysis from Dr. Sury; the significance of

practical thinking from Dr. Wiedmann and the power of getting the big picture from Dr.

Stein from a different perspective. I am extremely grateful not only for their time and

patience, but for their trust in me as they offer me support in my development as a scientist.

My special thanks go to Dr. Chris Kulczar, who has been an exceptional mentor

and friend during my internship at Merck & Co., Inc at Rahway, NJ in 2019. I am grateful

for the opportunities I had to participate in exciting research projects and cutting edge

research within Merck. I also appreciate the help from Dr. Chris Kulczar in giving me the

opportunities to meet with many other great scientists across Merck. My thanks also goes

to the directors at Merck Animal Health: Dr. Chad Brown, Dr. Jonathan Eichman, Dr.

ii

Izabela Galeska and Dr. Charles Moeder who provided feedback and encouragements

during monthly meetings. I’m grateful for all other colleagues I met during my internship

for company and a wonderful summer: Dr. Michelle Fung, Dr. Ximeng Dow, Katie

Roughneen, Dr. Anagha Bhakay, Dr. Laura Northrup, Dr. Kelly Cung, Sharon Martin, Dr.

Mandana Azimi, Dr. Brenda Colón, Dr. Chen-Chao Wang, and Dr. Patrick Boyer.

I would like to extend my sincere thanks to my current and former colleagues in

the Materials Science and Engineering Laboratory: Dr. Reddy Perumalla, Dr. Limin Shi,

Dr. Shubhajit Paul, Dr. Chenguang Wang, Dr. Manish Kumar Mishra, Dr. Wei-Jhe Sun,

Dr. Shao-Yu Chang, Dr. Jiangnan Dun, Shenye Hu, Dr. Hongbo Chen, Ling Zhu, Sibo Liu,

Zhongyang Shi, Zhengxuan Liang, Yiwang Guo, Gerrit Vreeman, Joan Cheng, Shan Wang,

Prof. Benyong Lou, Dr. Majeed Ullah, Prof. Xingchu Gong, Prof. Jiamei Chen, Prof. Xin

He, Prof. Shuyu Liu, Dr. Hiroyuki Yamashita, Prof. Mingjun Wang, Prof. Wenxi Wang,

Dr. Cosima Hirschberg, Ms. Hongliang Wen, Prof. Wei Li and Prof. Jun Wen. Their

company, discussion, and assistance have been invaluable.

I would like to acknowledge the team members in preparations for the 50th Annual

Pharmaceutics Graduate Student Research Meeting (PGSRM) at University of Minnesota.

As a co-chair for Logistics, the memory for organizing the PGSRM 2018 is unforgettable

experience in life as graduate student.

I’d like to recognize my funding sources, the David J.W. Grant & Marilyn J. Grant

Fellowship in Physical Pharmacy, and the Edward G. Rippie Fellowship in Pharmaceutics

for the financial support of part of my graduate study.

iii

Last but not least, I’d like to express my deepest appreciation to my husband

Chenxing Pei, who is, and has always been, an accompany, a comrade, a best friend and a

supporter of me during this joint journey of life. My Ph.D. journey would not have been

possible without the support, encouragement, guidance, inspiration and unconditional love

from him, who understands me the best as a Ph.D. himself. I would also like to extend my

most sincere gratitude to my parents and parents-in-law, who have given me trust and

support, both emotionally and financially for my 9 years of overseas study. The

encouragement and unreserved love they gave me shapes me to be who I am now. Thanks

to my dearest friends at the University of Minnesota and across the world for the warm

friendship in all these years. My thanks also go to my 5-year-old twin cats, Laobing and

Bantang, who have grown up as the best ‘coworkers’ during my Ph.D. study.

iv

To my parents and my love

v

Abstract

Tablets are the most desirable solid oral dosage form for patients. Direct

compression (DC) tablet formulation is the most economical, robust and efficient way of

tablet manufacture. Being sensitive to properties of the Active Pharmaceutical Ingredient

(API), direct compression tablet formulation is not available for the high dose non-steroidal

anti-inflammatory drug, celecoxib (CEL) due to the undesirable properties of the

commercial solid form of CEL, including low bulk density, poor flowability and tablet

lamination issues.

The solid form used in commercially available CEL capsules is a polymorph of

CEL, Form III. Form III CEL is a needle shaped crystal, which is exceptionally elastic.

This high elasticity, verified by nanoindentation and three-point bending tests, is

unfavorable for good tablet quality and performance during high speed tableting. Through

understanding the molecular interactions by analyzing the CEL crystal structure, a

structural model for high elasticity is built and validated by Raman spectroscopy.

Interlocked molecular packing without slip plane and the presence of isotropic hydrogen

bond network are major structural features responsible for both the exceptional elastic

flexibility and high stiffness of the CEL crystal.

CEL Form III exhibits unsatisfactory flowability and tablet lamination issues for

DC tablet manufacturing. Pharmaceutically acceptable solvates of CEL offer better flow,

compaction and dissolution properties than CEL Form III. Two stoichiometric solvates of

CEL and N-methyl-2-pyrrolidone (NMP) are extensively characterized and examined,

which establishes a clear crystal structure-property relationship essential for crystal

vi

engineering of CEL. Through crystal engineering, a DC tablet formulation of CEL is

successfully developed using the dimethyl sulfoxide (DMSO) solvate of CEL. This

pharmaceutically acceptable solvate is highly stable and also exhibited much improved

manufacturability compared to CEL Form III, including better flowability, lower elasticity

and bulk density (superior tablet quality) as well as better dissolution performance.

As a Class II drug in the biopharmaceutics classification system with low solubility

and high permeability, the high dose of CEL is partially attributed to its limited solubility.

Amorphous CEL, although providing solubility advantages as the thermodynamically high

energy state, is unstable and prone to crystallization. The study of crystal growth of

amorphous CEL reveals a fast glass-to-crystal growth mode at room temperature with a

surface-enhanced mechanism. This paves the way for future development of a stable

amorphous solid dispersion tablet product of CEL with improved dissolution performance

and tablet manufacturability.

In summary, by understanding the structural origin of undesired properties of CEL,

successful development of the most patient-compliant tablet dosage form by direct

compression can be achieved. This sets an excellent example of utilizing a solid state

engineering approach to effectively overcome challenges encountered in direct

compression tablet development.

vii

Table of Contents

Acknowledgements .............................................................................................................. i

Abstract ............................................................................................................................... v

Table of Contents .............................................................................................................. vii

List of Tables ..................................................................................................................... xi

List of Figures ................................................................................................................... xii

Table of Abbreviations .................................................................................................. xviii

Chapter 1. Introduction ................................................................................................... 1

1.1 General Introduction ....................................................................................... 2

1.2 Literature Review............................................................................................ 4

1.2.1 Tablet Dosage Form - Prevalence and Advantages .................................... 4

1.2.2 Critical Attributes of a Successful Tablet Formulation .............................. 5

1.2.3 Materials Engineering for Drug Product Development .............................. 7

1.2.3.1 Preformulation and Process Chemistry ................................................... 7

1.2.3.2 Formulation and Process Development ................................................ 11

1.2.4 Crystal Engineering to Optimize API Properties ...................................... 15

1.2.5 Celecoxib .................................................................................................. 18

1.2.5.1 Background of Celecoxib ..................................................................... 18

1.2.5.2 Challenges for Direct Compression Tablet Manufacture ..................... 19

1.3 Objectives and Hypotheses ........................................................................... 21

1.4 Research Plan and Thesis Organization ........................................................ 23

Chapter 2. Exceptionally Elastic Crystals of Celecoxib Form III .............................. 25

2.1 Synopsis ........................................................................................................ 26

2.2 Introduction ................................................................................................... 27

2.3 Materials and Methods .................................................................................. 29

2.3.1 Materials ................................................................................................... 29

2.3.1.1 Single Crystal Preparation of Elastic Crystals of CEL Form III .......... 29

2.3.2 Methods..................................................................................................... 30

2.3.2.1 Single Crystal X-ray Diffraction Experiment ....................................... 30

2.3.2.2 Face Index of CEL Form III ................................................................. 30

viii

2.3.2.3 Three-point Bending and Looping Tests of CEL Crystals ................... 31

2.3.2.4 Nanoindentation .................................................................................... 32

2.3.2.5 Raman Spectroscopy ............................................................................. 32

2.4 Results and Discussion ................................................................................. 33

2.5 Conclusion .................................................................................................... 44

Chapter 3. Structural Insights into the Distinct Solid-state Properties and

Interconversion of Celecoxib N-Methyl-2-Pyrrolidone Solvates ................................ 46

3.1 Synopsis ........................................................................................................ 47

3.2 Introduction ................................................................................................... 48


3.3.1 Materials ................................................................................................... 50

3.3.2 Methods..................................................................................................... 50

3.3.2.1 Single Crystal X-ray Crystallography ................................................... 50

3.3.2.2 Preparation of NMP Solvate Bulk Powders.......................................... 51

3.3.2.3 Powder X-ray Diffraction (PXRD) ....................................................... 51

3.3.2.4 Thermal Analyses ................................................................................. 52

3.3.2.5 Hot Stage Microscopy (HSM) .............................................................. 52

3.3.2.6 Fourier Transform Infrared Spectroscopy (FTIR) ................................ 53

3.3.2.7 Energy Framework Calculations........................................................... 53


3.4.1 Crystal Structures of CEL-NMP Solvates ................................................ 54

3.4.2 Solid-state Properties of the Two CEL-NMP Solvates ............................. 57

3.4.3 Phase Transformation of di-NMP upon Heating ...................................... 60

3.4.4 Thermodynamic Stability Relationship of Two Solvates ......................... 64

3.4.5 Physical Stability on Storage .................................................................... 65

3.4.6 Structural Characteristics of the Two NMP Solvates ............................... 66

3.4.6.1 Spectroscopic Evidence for the Structural Differences Between Two

CEL-NMP Solvates .............................................................................................. 72

3.5 Conclusions ................................................................................................... 73

Chapter 4. Direct Compression Tablet Formulation of Celecoxib Enabled by A

Pharmaceutical Solvate .................................................................................................. 75

4.1 Synopsis ........................................................................................................ 76

4.2 Introduction ................................................................................................... 77


ix

4.3.1 Materials ................................................................................................... 79

4.3.2 Methods..................................................................................................... 79

4.3.2.1 Single Crystal X-ray Crystallography ................................................... 79

4.3.2.2 Preparation of Bulk CEL-DMSO Powder ............................................ 80

4.3.2.3 Powder X-ray Diffraction (PXRD) ....................................................... 81

4.3.2.4 Thermal Analyses ................................................................................. 81

4.3.2.5 Hot Stage Microscopy (HSM) .............................................................. 82

4.3.2.6 Nanoindentation of CEL-DMSO Single Crystals ................................. 82

4.3.2.7 Nanocoating of Microcrystalline Cellulose .......................................... 83

4.3.2.8 Powder Flow Property Assessment ...................................................... 83

4.3.2.9 Preparation of Tablet Formulations ...................................................... 84

4.3.2.10 Tabletability Profiling ....................................................................... 84

4.3.2.11 True Density Determination ............................................................. 85

4.3.2.12 Powder Bulk Density Measurement ................................................. 86

4.3.2.13 Compressibility and Compactibility ................................................. 86

4.3.2.14 In-die Elastic Recovery ..................................................................... 86

4.3.2.15 Intrinsic Dissolution Rate ................................................................. 87

4.3.2.16 Tablet Disintegration Time ............................................................... 87

4.3.2.17 Expedited Friability Test................................................................... 87

4.3.2.18 Dissolution of CEL Tablets and Capsules ........................................ 88


4.4.1 Physicochemical Properties of CEL-DMSO Solvate ............................... 89

4.4.2 Mechanical Properties of CEL-DMSO ..................................................... 95

4.4.3 Flowability of CEL-DMSO ...................................................................... 97

4.4.4 Development of A Direct Compression Tablet Formulation Using CEL-

DMSO Solvate .......................................................................................................... 98

4.4.5 Expedited Friability of CEL-DMSO Formulation D Tablets ................. 105

4.4.6 Disintegration and Dissolution of CEL-DMSO Formulation D Tablets 106

4.5 Conclusion .................................................................................................. 108

Chapter 5. Crystal Growth of Celecoxib from Amorphous State - Polymorphism,

Growth Mechanism, and Kinetics ............................................................................... 109

5.1 Synopsis ...................................................................................................... 110

5.2 Introduction ................................................................................................. 112

5.3 Materials and Methods ................................................................................ 114

x

5.3.1 Materials ................................................................................................. 114

5.3.2 Methods................................................................................................... 114

5.3.2.1 Preparation of Amorphous CEL Thin Films ....................................... 114

5.3.2.2 Crystal Growth Rate Measurements ................................................... 116

5.3.2.3 Arrhenius Analysis of Crystal Growth Rate ....................................... 117

5.3.2.4 Preparation of Polymorphic Seeds of CEL ......................................... 117

5.3.2.5 Raman Spectroscopy ........................................................................... 118

5.3.2.6 Micro X-ray Diffraction ...................................................................... 118

5.3.2.7 Thermal Analysis ................................................................................ 119

5.3.2.8 Scanning Electron Microscopy (SEM) ............................................... 119

5.4 Results and Discussion ............................................................................... 119

5.4.1 Differential Surface and Bulk Crystal Growth Rates ............................. 119

5.4.2 Temperature-dependent Crystal Growth Mechanisms ........................... 122

5.4.3 Polymorph Cross-nucleation ................................................................... 125

5.4.4 Relationship between Growth Mechanism and Crystal Morphology ..... 129

5.4.5 Effect of Polymorphism on Growth Rate ............................................... 131

5.4.6 Arrhenius Growth Kinetics of CEL Crystals .......................................... 133

5.5 Conclusions ................................................................................................. 138

Chapter 6. Research Summary and Future Work .................................................... 140

Research Summary ......................................................................................................... 141

Future Work .................................................................................................................... 145

Bibliography .................................................................................................................. 149

Appendix ........................................................................................................................ 186

xi

List of Tables

Table 2–1. Crystallographic information for CEL Form III crystals ............................... 34

Table 3–1. Strong hydrogen bonds that stabilize the tetramer in mono-NMP and di-NMP

solvate crystal structures. .................................................................................................. 57

Table 3–2. Crystallographic information of CEL mono-NMP and di-NMP solvates ..... 71

Table 4–1. Hydrogen bonds that stabilize the tetramer in the CEL-DMSO structure ..... 92

Table 4–2. Summary of tablet formulation compositions and weights. The total amount

of CEL in each tablet is 200 mg........................................................................................ 99

Table 5–1. Physical properties of Form III, Form I and amorphous CEL ..................... 115

Table 5–2. Kinetic parameters of CEL bulk and surface processes compared to known

examples of GC crystal growth in bulk glasses .............................................................. 135

xii

List of Figures

Figure 1–1. The materials science tetrahedron (MST) ...................................................... 5

Figure 1–2. Materials engineering activities involved in development of tablets ............. 8

Figure 1–3. Illustration of common crystalline forms of APIs used in crystal engineering

........................................................................................................................................... 10

Figure 1–4. Molecular structure of celecoxib, MW=381.373 g/mol ............................... 18

Figure 2–1. Synopsis figure of elastic CEL crystal ......................................................... 26

Figure 2–2. Molecular packing in CEL viewed into (01-1), (001) and (100) faces. ....... 29

Figure 2–3. Face index of needle-shaped CEL form III crystal ....................................... 34

Figure 2–4. Molecular conformation of CEL represented by perpendicular planes ........ 36

Figure 2–5. Intermolecular interactions pattern from two different orientations (a) and

(b). Interaction within the dimer is highlighted in the gray color. .................................... 36

Figure 2–6. (a) 1-dimensional chain along a-axis. (b) Corrugated layered structure of

CEL molecule, with yellow color for C-H···π interactions connecting the dimers.......... 37

Figure 2–7. Screenshots of bending tests on the bendable face (001) / (00-1) of a CEL

single crystal using forceps and a needle, showing four repeated cycles of bending and

elastic recovery(a) ~ (i). .................................................................................................... 38

Figure 2–8. Screen shots of bending on CEL crystal minor and major faces – when bent

on major (001) face, crystal is elastic; when bent on minor (011) or (011) face, the CEL

crystal is brittle. ................................................................................................................ 38

Figure 2–9. (a) - (c). Looping of elastic CEL crystal and its complete recovery. ........... 39

Figure 2–10. Radius of curvature for maximally bent loop of CEL ................................ 39

Figure 2–11. Nanoindentation characterization of the CEL crystal. a) Force-depth curve

and b) 3D scan of indented area. ...................................................................................... 40

Figure 2–12. Schematic representation of the molecular rearrangement in the elastic

crystal during the bending state. ....................................................................................... 41

Figure 2–13. Calculated Raman spectra of CEL.............................................................. 42

Figure 2–14. Raman spectra corresponding to the bent and straight crystals of CEL, with

red and blue circles indicate the point of data collection (inner arc and outer arc

xiii

respectively) a) lattice vibration, and b) aromatic symmetric in-plane bending (C–H)

modes of inner and outer arc of a bent CEL single crystal, as well as in the non-strained

straight state. ..................................................................................................................... 43

Figure 3–1. Chemical structure of a) celecoxib (MW = 381.37) and b) NMP (MW =

99.13) ................................................................................................................................ 49

Figure 3–2. PXRD of CEL-NMP solvates prepared using different methods, as compared

to patterns of CEL-NMP mono- and di-solvates and Form III CEL calculated from

corresponding single crystal structures ............................................................................. 54

Figure 3–3. ORTEP drawings of a) mono-NMP and b) di-NMP solvate structures

collected at room temperature (298 (2) K). The CF3 in both solvates and the NMP

molecule in mono-NMP are disordered. ........................................................................... 55

Figure 3–4. Tetramer with strong hydrogen bonded CEL and NMP molecules in a) CEL

mono-NMP solvate and b) CEL di-NMP solvate. Light blue dashed lines are strong

hydrogen bonds labeled as 1, 2, 1’ and 2’ in both structures. ........................................... 56

Figure 3–5. Overlay of CEL mono-NMP solvate (red) and di-NMP solvate (light blue).

The conformation of CEL is the same in the two structures, although the NMP in the

tetramers adopted flipped orientations (black arrows). N-H⋯O hydrogen bonds between

NH2 of CEL and the carbonyl O atom of NMP molecules are retained. .......................... 57

Figure 3–6. Morphology of a) mono-NMP solvate and b) di-NMP solvate of CEL. Scale

bars are 500 µm. ................................................................................................................ 58

Figure 3–7. Thermal properties of CEL Form III (black), CEL mono-NMP (red), and

CEL di-NMP (blue). (a) DSC and (b) TGA (Dashed lines represent first derivative of

TGA traces of the three solid forms). ............................................................................... 59

Figure 3–8. CEL mono-NMP solvate heated using HSM without silicone oil at a rate of

10 ºC/min. ......................................................................................................................... 59

Figure 3–9. HSM images as CEL-NMP disolvate was heated at 10 ºC/min and undergo a

series of events: desolvation, dissolution, and melting. .................................................... 61

Figure 3–10. CEL-NMP disolvate heated using HSM at a rate of 10 ºC/min, without

silicone oil. a) ~ b) no change from room temperature to 50 ºC. c) desolvation at 70 ºC,

forming polycrystalline mono-NMP crystals (the foggy image is a result of the

condensation of escaped NMP on the glass window of HSM), d) The mono-NMP

remained stable up to at least 108 ºC (image was clear again because the condensed NMP

evaporated). e) and f) melting of mono-NMP crystals 125 – 133.3 ºC. ........................... 62

Figure 3–11. DSC thermogram of CEL di-NMP solvate upon heating at 10 ºC/min.

Corresponding thermal events at various stages are also shown with PLM images. ....... 64

xiv

Figure 3–12. PXRD patterns of equilibrium solids obtained from slurry experiments and

comparisons with CEL-NMP calculated patterns from room temperature single crystal

structures. .......................................................................................................................... 65

Figure 3–13. Weak interactions in CEL mono-NMP solvate structure: a) intra-tetramer

and inter-tetramer to form 1-D chain along b-axis; b) inter-tetramer along a-axis; c)

between tetramers along c-axis. d) viewed into a-axis. .................................................... 67

Figure 3–14. Weak interactions in CEL di-NMP solvate a) Intra-tetramer along c-axis; b)

along b-axis; c) along a-axis; and d) view down a-aixs ................................................... 68

Figure 3–15. Energy framework of (a) mono-NMP viewed down a axis and (b) di-NMP

viewed down b axis. The pink parallelograms highlight a tetramer in each of the

frameworks. ...................................................................................................................... 68

Figure 3–16. Mono-NMP solvate crystal structure viewed along (a) b axis and (b) a axis,

(c)~(e) zig-zag void channels housing the strongly-bound NMP molecules .................... 69

Figure 3–17. Packing of di-NMP solvate crystal a) down the b-axis and b) down the a-

axis. In the top row, NMP molecules as part of the framework are colored blue, blue and

red rectangles/circles highlight the channels along the b- and a-axis, respectively; loosely

bound NMP molecules are colored red. In the bottom row, the loosely bound NMP

molecules are deleted to show the channels. .................................................................... 70

Figure 3–18. Comparison of unit cell parameters and packing for a) CEL mono-NMP

solvate and b) di-NMP solvate (loosely bound NMP molecules removed) down the b-axis

........................................................................................................................................... 71

Figure 3–19. FTIR spectra of CEL Form III, CEL mono-NMP solvate and di-NMP

solvate crystals .................................................................................................................. 73

Figure 4–1. Molecular structures of celecoxib (MW 381.37) and DMSO (MW 78.13). 77

Figure 4–2. Particle size of a) CEL-DMSO powder and b) CEL Form III as-received

powder. The scale bar is 50 m in both pictures. Since the shape of CEL Form III and

CEL-DMSO crystals are different, while CEL Form III is in needle shape with high

aspect ratio, it was made sure that the longest dimension of CEL-DMSO and CEL Form

III crystals match............................................................................................................... 89

Figure 4–3. a) Block morphology of CEL-DMSO solvate crystals and ORTEP diagrams

of CEL-DMSO single crystal structure at b) 298 K and c) 100 K. The DMSO molecule

and the CF3 group on CEL are disordered only in 298 K structure. ................................. 90

Figure 4–4. a) DSC and b) TGA thermographs of CEL-DMSO solvate crystals. The

dashed lines in b) represents the first derivative of weight signal. ................................... 90

xv

Figure 4–5. Polarized light microscope images of CEL-DMSO single crystal as it is

heated using a hotstage ..................................................................................................... 91

Figure 4–6. A tetramer synthon 𝑅42(8) comprised of two CEL and two DMSO

molecules .......................................................................................................................... 92

Figure 4–7. Packing of CEL-DMSO with removed DMSO molecule to highlight the void

space. Hydrogen atoms were removed for simplicity and clarity. a) Viewed down the b-

axis, the spiral is barely look-through. b) viewed down the a-axis, the spiral is obvious

and the DMSO occupies staggered positions in the lattice ............................................... 93

Figure 4–8. Powder X-ray patterns of CEL-DMSO solids and tablets. The simulated

pattern was used as a reference, which was obtained from the room temperature crystal

structure............................................................................................................................. 94

Figure 4–9. IDR data showing CEL-DMSO tests and CEL Form III tests at 25 °C. Three

CEL-DMSO curves showed different extents of supersaturation followed by rapid

precipitation and phase conversion to Form III. ............................................................... 95

Figure 4–10. Surface scans of indentation imprint after nanoindentation tests for a) CEL

Form III on (001) face and b) CEL-DMSO on (002) face. The hairline in b) is a crack in

the crystal. ......................................................................................................................... 96

Figure 4–11. a) Tabletability profiles and b) flowability of CEL-DMSO solvate, CEL

Form III and their Formulations D.................................................................................... 97

Figure 4–12. Severe lamination of tablets made from CEL Form III Formulation A. .. 100

Figure 4–13. a) The tabletability and b) the ejection force profiles of CEL-DMSO and

CEL Form III formulations who were formulated as listed in Formulation A. .............. 100

Figure 4–14. Ejection force of Formulations D for both CEL Form III and CEL-DMSO

......................................................................................................................................... 103

Figure 4–15. a) Compressibility and b) compactibility of CEL-DMSO and CEL Form III

Formulations D. .............................................................................................................. 104

Figure 4–16. 500 mg of CEL Form III-containing Formulation D powder cannot be

packed into the die, even if it is at the maximum depth of filling. The powder had to be

pushed into the die by hand for compaction into tablets. ............................................... 104

Figure 4–17. a) Lamination upon ejection of CEL Form III Formulation D tablets made

at 7kN. The inset is a close-up of the middle piece of the broken tablet, which is further

laminated into layers. b) Another laminated tablet of CEL Form III Formulation D made

at 7 kN. c) Intact tablet of CEL-DMSO Formulation D made at 7 kN. .......................... 104

xvi

Figure 4–18. a) Friability profiles and b) in-die elastic recovery profiles for the CEL-

DMSO Formulation D at 20 ms (solid squares) and 30 ms (open triangles) dwell times.

Trend lines are manually added to guide the eye. ........................................................... 106

Figure 4–19. Tablet dissolution tests of CEL-DMSO Formulation D vs. marketed 200

mg CEL capsules. ........................................................................................................... 107

Figure 5–1. Synopsis figure. Crystallization of celecoxib from amorphous phase exhibits

the phenomena of fast surface growth, fast bulk glass-to-crystal growth, and cross-

nucleation between two polymorphs. The fast growth mechanisms are disrupted by the

onset of fluidity controlled growth at temperatures above Tg. ........................................ 111

Figure 5–2. Sample configuration of surface and bulk amorphous thin films (top view)

......................................................................................................................................... 115

Figure 5–3. a) Growth of bulk Form I crystals at 70 °C over 6 days b) Linear regression

on increasing crystal radii over time. .............................................................................. 117

Figure 5–4. Surface and bulk crystal growth rate - temperature profiles of CEL a) Form

III CEL. b) Form I. Open symbols indicate growth mechanisms distinct from the

diffusion controlled growth (solid symbols). .................................................................. 120

Figure 5–5. Crystallization of CEL from amorphous phase observed under PLM and

Raman Microscope, a) Form III crystals grew on surface at 40 °C, and b) Form I crystal

grew in the bulk at 80 °C, c) The two polymorphs under Raman microscope, d) Raman

area scan of the boundary between Form I (orange red) and Form III (black). .............. 121

Figure 5–6. Liquid-embedded crystal in the diffusion-controlled crystal growth region of

CEL Form III, taken from 110 ºC sample. Crystal growth direction is indicated with an

arrow. a) Top view; b) tilted view in SEM. .................................................................... 122

Figure 5–7. Surface and bulk crystal growth rate - temperature profiles of Form I and III

of CEL. Characteristic crystal morphologies in different temperature regions are shown.

Open symbols indicate growth mechanisms distinct from the diffusion controlled growth

(solid symbols). Tt was determined from linear fitting, and the black and purple dashed-

lines aid visualization of transition zones on surface and in bulk. .................................. 125

Figure 5–8. Characterization of CEL Form I and Form III by powder X-ray diffraction

and micro-diffractometer. ............................................................................................... 126

Figure 5–9. DSC and TGA curves of Forms I and III of CEL crystallization from melt

......................................................................................................................................... 127

Figure 5–10. Characterization of CEL Form I and Form III by Raman spectroscopy;

characteristic peak of Form III is at 1154cm-1, which split into two peaks for Form I. 127

xvii

Figure 5–11. a) Characteristic peak representing CEL Form I used for Raman mapping

and b) Raman 1D line scan result. .................................................................................. 128

Figure 5–12. PLM images of CEL crystals grown from amorphous films at different

temperatures. B is for bulk and S is for surface samples. Scale bars are all 100 µm. ... 130

Figure 5–13. Growth rate profile of both crystal forms for surface samples of amorphous

CEL; the inset represents the growth rates of two polymorphs on linear scale .............. 132

Figure 5–14. Growth rate profile of both crystal forms for bulk samples of amorphous

CEL; the inset in the growth rates of two polymorphs in linear scale ............................ 132

Figure 5–15. Arrhenius relationship of CEL growth rate a) on the surface and b) in bulk

......................................................................................................................................... 136

Figure 5–16. Growth rate u, at Tt as a function of Tt / Tg of several small organic

molecules studied for GC growth mode. The black star represents CEL. ...................... 138

xviii

Table of Abbreviations

API Active Pharmaceutical Ingredient

ASD Amorphous Solid Dispersion

BCS Biopharmaceutics Classification System

CCDC Cambridge Crystallographic Data Centre

CEL Celecoxib

COX Cyclo-oxygenase

DC Direct Compression

DFT Density Functional Theory

DMA Dimethyl Acetamide

DMF Dimethyl Formamide

DMPU N, N′-Dimethylpropyleneurea

DMSO Dimethyl Sulfoxide

DPCP 1,2- diphenylcyclopentene

DPCH 1,2-diphenylcyclohexene

DSC Differential Scanning Calorimetry

FBRM Focused Beam Reflectance Measurements

FDA Food and Drug Administration

FT-IR Fourier Transform Infrared Spectroscopy

GC Glass-to-Crystal

GSF Griseofulvin

xix

HPC Hydroxypropyl Cellulose

HPMC-AS Hydroxypropyl Methylcellulose – Acetate Succinate

HSM Hot Stage Microscopy

ICH International Council for Harmonisation of Technical

Requirements for Pharmaceuticals for Human Use

IDR Intrinsic Dissolution Rate

IMC Indomethacin

IPB Isopropylbenzene

MCC Microcrystalline Cellulose

MgSt Magnesium Stearate

MST Materials Science Tetrahedron

NIF Nifedipine

NMP N-Methyl-2-Pyrrolidone

NSAID Nonsteroidal Anti-inflammatory Drug

OTP O-terphenyl

PAT Process Analytical Technology

PEG Polyethylene Glycol

PG Propylene Glycol

PLM Polarized Light Microscope

PXRD Powder X-ray Diffraction

QbD Quality by Design

RH Relative Humidity

xx

ROY 5-methyl-2-[(2-nitrophenyl)amino]-3-thiophenecarbonitrile

SCXRD Single Crystal X-Ray Diffraction

SEM Scanning Electron Microscope

SLS Sodium Lauryl Sulphate

TGA Thermogravimetric Analysis

TMU Tetramethyl Urea

TNB Tris-naphthyl Benzene

TP Testosterone Propionate

TPGS D-α-Tocopherol Polyethylene Glycol 1000 Succinate

1

Chapter 1.

Introduction

2

1.1 General Introduction

Among all dosage forms, tablets are the most widely produced dosage form and are

more advantageous than other forms in many ways: 1) elegancy in terms of different

shapes/sizes and high patient compliance; 2) precisely controlled dosing for patients; 3)

good physical and chemical stability; 4) faster and easier production and transportation and

5) more straightforward and economical manufacturing processes. 1 Although tablet

formulation is most desirable for both patients and pharmaceutical companies, it is not

always easy to successfully develop a tablet formulation. A balance among aspects of

preformulation and formulation, including physical and chemical stability, solubility,

powder flow properties, tablet friability and tensile strength, must be achieved during tablet

formulation design.

To attain the desired characteristics of the optimal formulation, multiple approaches

can be employed. Solid-state properties of active pharmaceutical ingredients (APIs) can be

engineered by screening for different solid forms, such as amorphous solid,

polymorph/cocrystal/salt/solvate and hydrate. For a given API solid form, particle and

powder properties can be engineered by particle size reduction, dry and wet granulation.

Lastly, fine-tuning tablet performance can be achieved by altering the formulation

compositions. For drugs that are less potent or of low solubility, high doses are usually

required for the tablet to demonstrate the desired therapeutic effect. For these types of drugs,

the final property of the formulation is highly dependent on the API properties, and the

space for manipulating drug product performance through formulation design is limited. In

this case, although high drug loading can be achieved by wet granulation and dry

3

granulation to uniformly disperse the API and densify the powder blend, 2 the granulation

process is often costly and requires extra steps and more strict process controls. In

comparison, direct compression is the more desirable way for tablet manufacturing. Out of

all these possible solid-state engineering methods, API engineering to improve the

properties is the most effective.

In this work, a non-steroidal anti-inflammatory drug for treating pain and arthritis,

celecoxib (CEL), is used as an example of high dose drug that is challenging for direct

compression (DC) tablet formulation development. The marketed CEL product is in form

of capsules, Celebrex®, which are usually prescribed in high doses of 200 mg or 400 mg.

The marketed product contains the most stable polymorph, Form III of CEL, which has

undesired properties for DC tablets. 3, 4 Wet granulation was employed in the original

Celebrex® capsules to improve the poor flow properties of API for capsule filling. 4

However, this process is complex, costly and possess risks of overgranulation and chemical

and physical stability. 5 In addition, these capsules are of relatively large sizes and thus

lower patient compliance than tablets. In addition, the lag time before releasing the drug

from capsules is unfavorable for pain medications like CEL, which is preferred to have

faster action. To develop a DC tablet formulation of CEL, the unwanted properties of Form

III CEL needs to be overcome. This includes poor flow properties, low bulk density and

tablet lamination issue during high speed direct compression. 4

Here we aim to improve CEL properties through solid-state engineering to enable

the development of a suitable DC tablet. This is achieved using a combined crystal

engineering and Quality by Design (QbD) approach. The guiding principles behind this

4

work is the materials science tetrahedron (MST), which describes the relationship between

material structure, properties, processing and performance. 6

1.2 Literature Review

1.2.1 Tablet Dosage Form - Prevalence and Advantages

Solid oral dosage forms have been the most widely used for both patient compliance

and physical stability. Among solid oral dosage forms, tablets have been the most widely

produced and prescribed drug among all dosage forms. 7 For patients, solid oral dosages

forms are usually taken without complications, have accurate dosing, easy to store, and can

be taken without pain and extra dedication. Tablets stands out among all oral solid dosages

forms due to its elegancy, which can be made into different shapes, colors, and tastes. In

addition, a tablet can also be significantly smaller than capsules with the same amount of

drug content. To the pharmaceutical industry, the manufacturing of tablets is more robust,

time-saving and economical. 8

There are many types of tablets for different functions, routes of administration,

and release profiles. 9 Tablets categorized for special function include chewable tablets, 10

lozenges, 11 bilayer tablets, 12-14 orally disintegrating tablets 15, 16 and effervescent tablets.

Buccal tablets 17 are administered with slightly different routes of administration. Film

coated tablets are often made for ease of swallowing or for an extended release profile. 18

Depending on the purpose of development, tablet formulations can be altered to

attain improving specific properties. For drug product quality control, many important

aspects of tablet performance need to be assessed for all kinds of tablets, including

5

bioavailability, stability, manufacturability, safety and efficacy. 19 According to MST, the

performance of a drug product is closely related to materials structure, properties and

processing (Figure 1–1). 6 These key attributes will be reviewed in detail in the next section.

Figure 1–1. The materials science tetrahedron (MST)

1.2.2 Critical Attributes of a Successful Tablet Formulation

A successful tablet formulation should satisfy the criteria for content uniformity,

tablet weight, tablet hardness, tablet friability, disintegration time and dissolution

performance. Such performance of a tablet product stems from its properties, the process

for manufacturing during drug product, and the structure of materials.

Here, the material properties include the physicochemical properties (solubility of

API, thermal properties, hygroscopicity, etc.), micromeritics, powder flow properties and

mechanical properties (elasticity and plasticity, brittleness, tabletability, compressibility

and compactibility) of to both the API and the excipients. They are closely related to the

6

material ‘structure’, which include the crystal structure, granule structure, and tablet

structure (e.g., porosity and compositions). 6

A tablet formulation consists of both the API and other inactive ingredients. Once

the drug molecular structure is selected, the different solid forms of the API affect its

properties and thus performance of the drug product. The inactive ingredient, commonly

called excipients, can play different roles in a tablet formulation. A diluent or filler is used

to provide bulk volume, which can also aid in achieving content uniformity of mixtures or

modify tableting properties of tablets. The binder is used for increasing particle bonding

strength or as the granulation facilitator. A tablet disintegrant is important for immediate

release formulations for breaking up the tablets or granules when it is in contact with water,

achieving faster API release. A lubricant is used to reduce the friction between powder

mixture and the equipment surfaces and thus alleviate sticking and heat generation. A

glidant is sometimes used to reduce adhesion between particles and thus improving

flowability. Other components, including flavor or colorant for taste masking and

appearance, respectively, and surface coating are sometimes used for specific purposes,

like taste masking of bitter-tasting drugs. The selected excipients must also be chemically

compatible with the API to minimize any possible degradation of the API. 20 Knowing the

different properties of the API solid forms and functional excipients is useful for

developing a tablet product that best serves the patients’ need.

The ‘processing’ category of MST, including the key parameters of operations

during manufacturing, is another degree of freedom during product development that

connects the ‘structure’ of API and excipients with their ‘properties’ to achieve optimal

7

performance. Examples of processing are milling to change primary particle size; dry and

wet granulation process for altering particle properties; powder blending time and sequence;

and tableting speed and pressure.

A successful tablet product development requires a deep understanding of the

material and processing attributes as outlined by the MST. Only with these in mind, can a

formulation be truly designed using Quality by Design (QbD) approach. 2

1.2.3 Materials Engineering for Drug Product Development

1.2.3.1 Preformulation and Process Chemistry

The impact of pharmaceutical materials engineering spans a wide range from

structure, properties and processing to the final drug product performance. Materials

engineering includes two categories: preformulation (realized through process chemistry)

and formulation (realized through manufacturing process control). The hierarchical

organization and details of materials engineering approaches for each step is summarized

in Figure 1–2.

Solid form selection is of paramount importance during drug product development,

laying a foundation for attaining critical attributes of the drug product, including stability,

dissolution performance, manufacturability and tablet quality. 21-25 One of the most popular

solid-state engineering approaches is utilizing amorphous solids to increase solubility and

bioavailability of drugs. Amorphous solids, do not have long-range order of molecules and

tend to convert to their crystalline counterparts because they have higher thermodynamic

free energy. 26 To attain sufficient physical stability and high dissolution rates, APIs are

8

Fig

ure

1–2. M

ater

ials

en

gin

eeri

ng a

ctiv

itie

s in

volv

ed i

n d

evel

opm

ent

of

table

ts

9

often molecularly dispersed in polymeric matrices to form amorphous solid dispersions

(ASDs).27 ASDs are of increasing interest to the pharmaceutical industry due to the need

to improve the solubility of increasing numbers of emerging new drugs belonging to the

Biopharmaceutics Classification System (BCS) classes II and IV. 28 Some drug products

enabled by ASDs include Noxafil, Novir, Kaletra, Onmel, and Sporanox, etc. Among 24

marketed ASD products, 20 of them are made into tablets. 29

Developing an amorphous solid dispersion requires careful selection of polymers

and the method of generating ASD product. Hot-melt extrusion, spray drying,

coprecipitation are the main scalable manufacturing methods for ASDs, while film casting,

rotary evaporation and melt-quenching are often used in laboratory scale to screen ASDs.

29 The choice of polymer and manufacturing process for ASD is based on its performance,

i.e. stability, drug-polymer miscibility, in vitro dissolution performance and in vivo

bioavailability. 30, 31 Understanding the crystallization mechanisms and kinetics of pure

amorphous API is of fundamental significance for polymer screening and processing

conditions, 32, 33 since crystallization negates the solubility advantages of amorphous drug

products. 34

A crystalline form of drug substances has been traditionally utilized in developing

tablet products due to their better physical stability. For drugs in the Biopharmaceutics

Classification System (BCS) classes I and III, 35 the crystalline state of APIs is preferred

due to its robustness in stability and flexibility with processing and formulation. Different

crystalline forms of an API molecule include polymorphs, salts, hydrates, solvates and

cocrystals (Figure 1–3). Crystal engineering is a way of materials engineering to obtain

10

desired solid-state properties of API by manipulating crystal structure. 22 The drug solid

form also has to be chemically compatible with the excipients used and physically stable

during manufacturing and administration. For example, salt disproportionation is

sometimes encountered if formulated with specific excipients like dicalcium phosphate

dihydrate (DCPD). 20, 36

Figure 1–3. Illustration of common crystalline forms of APIs used in crystal engineering

After the lead solid form of API is chosen, particle engineering is sometimes

utilized to modify the crystal shape and size, which also have practical implications on

performance of the final tablet product. For different crystal shapes, the surface area of

specific crystal faces vary, and may exhibit distinct properties. Celecoxib plate shaped

crystals exhibit higher intrinsic dissolution rate and better wettability than needle shaped

crystals of similar size, attributing to the favorable exposure of hydrophilic facets of plate

crystals. 37 Crystal shape can also affect the flow properties, 38 tableting performance 39 and

punch sticking propensities. 40, 41

11

Particle size is also an essential aspect of particle engineering. Decreasing API

particle size by as micronization or nanosizing has been a ubiquitous method of improving

the dissolution performance of drug product and thus its bioavailability. 42-44 Smaller

particle size is also favorable for higher tablet tensile strength as it provides more surface

area available for bonding. 45, 46 However, smaller particles also exhibit deteriorated

powder flowability due to increased cohesion. 47 The balance between detrimental effect

of small particle size on poor flow and the beneficial effect on tablet tensile strength and

dissolution performance needs to be balanced during the tablet product development

process.

The different crystal shape and particle size distribution can be achieved by

controlling process parameters of crystallization and / or milling. Jet milling, ball milling,

wet milling, high pressure homogenization and cryogenic milling are common types of

milling methods used in industry. 48 Recently, extensive efforts have been made at the drug

substance/drug product interface for simultaneously controlling the API morphology and

size for better manufacturability with the aid of co-processed excipients for direct

compression in continuous manufacturing settings. 49, 50

1.2.3.2 Formulation and Process Development

Formulation and process development come after the lead API and particle

attributes have been determined. The main purpose of this stage is to select the best route

to formulate the API to attain properties suitable for high speed tableting, preferably using

DC process. There could be more or fewer steps required to optimize the powder blend for

satisfactory critical attributes required for DC. Depending on the dose, API and particulate

12

properties, project timeline, patient demographics and availability of technologies and

platforms, different formulation strategies can be employed to make the final blend for

tableting.

The tablet manufacturing of a direct compression formulation requires only

excipient selection and proper mixing. 8, 51 Aside from the tableting parameters, including

tableting speed, 52, 53 compaction pressure and selection of tooling, 54 which needs to be

considered for all tablet formulations, direct compression only requires the assessment of

excipient functionalities and blending intensities. Selection of functional excipients helps

to build a successful formulation by addressing issues, such as flowability and content

uniformity of the blend. 55 Dry nanocoating efficiently improves the flow. 56 Poloxamers

can serve as lubricants to reduce the ejection force without sacrificing dissolution. 57 This

development process is simple, economical and robust for most APIs. However, for potent

drugs with very low doses, content uniformity may be an issue if prepared only by simple

mixing. 58 For high dose drugs, on the other hand, the properties and performance are highly

dependent on the API itself if direct compression is employed. When the API solid form is

fixed, other formulation strategies may be pursued to solve problems in either

manufacturability or dissolution performance caused by undesirable properties.

Granulation is the most used technique other than DC. The most widely used

granulation techniques are wet granulation, dry granulation and melt granulation (Figure

1–2). 59 Melt-granulation is sometimes categorized as wet-granulation due to somewhat

similar processes, where the liquid binder in wet granulation is substituted with a low-

melting point solid. A variety of equipment are also available for granulation, which is

13

sometimes incorporated as one important component in the continuous manufacturing line

of pharmaceuticals, such as twin screw wet granulation. 60 Granulation is a complicated

process, which involves optimization of many critical parameters and characterization

techniques are needed to obtain the desired product performance.

Wet granulation has gained great prevalence in formulation and processing of APIs.

5, 61, 62 There are many aspects in wet granulation that needs to be evaluated for ideal granule

performance. First of all, high-shear wet granulations, fluid bed granulation, spray-drying

granulation and twin screw granulation are all wet granulation techniques with different

operating principles. 5 Take high shear wet granulation as an example, the operating

conditions of the high shear wet granulator critically affects the properties of granules and

thus the quality of tablets. The water level, impeller speed, viscosity of binder solution all

influence the granule properties and thus tablet quality. 63-69 Wet granulation also mandates

stable solid forms, since phase changes are more likely to take place during granulation

and drying processes. Many process analytical technologies (PAT) have been developed

for on-line or in-line monitoring of the granulation process for potential phase changes,

particle size distribution, etc. These include near infra-red spectroscopy (Near-IR, Raman

spectroscopy, focused beam reflectance measurements (FBRM) and powder X-ray

diffraction (PXRD). 62, 70-73 To sum up, wet granulation is complicated, time-consuming

and requires delicate control. Nevertheless, it is available for APIs with certain properties

and the when choice of solvent/binder is appropriate to avoid solid form changes during

granulation and drying.

14

Dry granulation is another route for improving particle properties for tablet

compression. 74-76 For APIs that are unstable for wet granulation processes, dry granulation

may be a viable method to improve particulate properties. Just like wet granulation, dry

granulation also has several process parameters that needs to be optimized for desired

granule properties. The type of roller compaction equipment (roll gap, roll diameter, roll

width, roll surface texture, feeding and de-aeration), processing parameters (roll speed,

feeding speed, compaction pressure) and formulation variables all impact the granule

properties. 75, 77-81 However, cautions need to be taken for the deterioration in tabletability

of granules after dry granulation.82

Melt-granulation is utilized in some pharmaceutical formulations to incorporate

low melting point excipients as granulating liquid by heating the granulation vessel to

above its melting temperature. 83-85 In addition to parameters similar to wet granulation, the

melt granulation involves processing conditions optimized to melt the solid binder. To sum

up, there are many factors to be considered in either wet granulation or dry granulation

processes, making granulation a time-consuming, complex and expansive formulation

process. The direct compression is the most preferred way of tablet manufacturing,

attributing to its robust, economical and straightforward characteristics. Since direct

compression tablet formulations is sensitive to API properties, especially for high dose

drugs, ways to alter the API properties in favor of direct compression are of great interest

to pharmaceutical industry, especially in early development stages. As an effective way to

alter drug properties, advancements in crystal engineering has been proven useful in drug

product development.

15

1.2.4 Crystal Engineering to Optimize API Properties

The field of crystal engineering has significantly advanced in terms of both the

understanding of the structure-property relationship between crystal structures and API

properties, and enabling QbD design for desired physicochemical properties and

manufacturability. 22, 86-88 Crystal engineering can be crucial to obtaining desired properties

in both the preformulation and formulation stages. Crystal engineering can be divided into

two parts - the first is understanding the root of the properties that needs to be changed –

which is the ‘structure-property relationship’ aspect. 22, 86 The second part is the

‘engineering’ aspect – to obtain the desired properties of API by altering its crystal

structure.22, 88-90

Many efforts has been made to understand the molecular origins of mechanical

properties of molecular crystals. Panda et al. described the molecular origin of the plastic

behavior of bendable hexachlorobenzene molecular crystals. It was argued that upon stress,

layers of molecules slide against each other with restorative Cl···Cl interactions, which act

cohesively to recover the adhesion between the layers. 91 These studies expand the

understanding of molecular origins of mechanical properties of solids, e.g. elasticity and

plasticity, and lay the foundation for the design of solid forms with desired properties.

Many researchers have also made efforts to design for wanted drug mechanical

properties with crystal engineering. 89, 92 93 The Reddy research group has successfully

induced plasticity of molecular single crystals by introducing plastically active slip planes

in the structure via carefully placing selected noninterfering van der Waals, π-stacking, and

hydrogen bonding groups, on different compounds. 94 While it is imperative to engineer

16

the mechanical properties of molecular crystals, more efforts have been made on using

crystal engineering to improve the solubility and dissolution rates of pharmaceutical APIs.

As detailed in Figure 1–3, polymorphs, solvates/hydrates, cocrystals and salts are all

commonly used to improve solubility and dissolution rates in pharmaceutical materials. 87,

95, 96 Cocrystallization of APIs has been shown to improve the solubility of many drugs -

for example, mebendazole 97 , ketoconazole 98 and posaconazole 99. Solvates of some APIs

are also found to increase the dissolution rate of many drugs, 100 e.g. rifampicin 101,

olanzapine 102 and spironolactone 103. Screening, selection or synthesis of polymorphs, 104

105 hydrates 106 and salts 98, 107, 108 are all successful crystal engineering methods for

solubility and dissolution enhancement during drug development. Some work has also

shown improvement in the stability of an API through hydration 109 or salt formation 110.

The manufacturability aspect of crystal engineering is less well explored compared

to the single crystal mechanical properties and solubility. Previous work done in our lab

surveyed crystal engineering approaches to overcome a variety of manufacturability

challenges. For example, poor tablet mechanical properties has been a challenge for many

APIs, preventing them to be made into tablet dosage form via direct compression. Caffeine

is such an example. To overcome this challenge, the caffeine – methyl gallate cocrystal

was developed and was shown to improve mechanical properties, with tabletabilities better

than caffeine and methyl gallate. 111 Cocrystallization of ibuprofen and flurbiprofen with

nicotinamide also exhibited improved tabletability, along with other simultaneously

improved properties like hygroscopicity, and dissolution performance. 112 A high dose

tablet formulation of 5-fluorocytosine was successfully developed into a tablet by salt

17

formation with oxalate acid. 113 Karki et al. were also able to make intact tablets of a drug

with high capping tendency, paracetamol, through cocrystallization. 114

The flow properties, an important parameter during high speed tablet

manufacturing, can also be improved by crystal engineering. The flow properties of a

cohesive powder, citric acid, was greatly improved by hydrate formation. 38 Punch sticking,

another practical problem encountered during tablet compression, 115 can also be addressed

via crystal engineering. The high sticking propensity of acesulfame free acid was

pronouncedly alleviated by forming a potassium salt, the mechanism of which was

explained by surface chemistry and reduction in plasticity. 116 Cocrystallization of

celecoxib and proline have also been proven to greatly reduce the sticking of celecoxib to

the punch surface by similar mechanisms. 117

As a summary, the crystal form of API is crucial developing a direct compression

tablet formulation. Crystal engineering is an effective way to improve API properties not

only at preformulation stage for properties like stability and solubility, but also for

manufacturability and processing during formulation development. According to the MST

(Figure 1–1), through understanding of the structural landscape of different crystal forms

and their relationships to properties and performance, one can use crystal engineering

methods to obtain the desired API properties to enable the tablet manufacture using the

direct compression process.

18

1.2.5 Celecoxib

1.2.5.1 Background of Celecoxib

Celecoxib (Figure 1–4) is a widely prescribed nonsteroidal anti-inflammatory drug

(NSAID) developed by Pfizer to treat arthritis, osteoarthritis, rheumatoid arthritis,

ankylosing spondylitis and acute pain, with also analgesic and antipyretic characteristics.

118-120 Celecoxib (CEL) is a cyclo-oxygenase (COX) inhibitor that selectively inhibits

COX-2 over COX-1 both in vitro and ex vivo. 118 CEL is commercially available as

capsules, with available doses of 100 mg, 200 mg and 400 mg. 119 CEL is a class II

compound in the Biopharmaceutics Classification System (BCS), having high permeability

and low solubility. 121 The absolute oral bioavailability of CEL in dogs was 64-88 % if

given as solution and 22-40 % if given as capsule. 121 The absolute bioavailability of CEL

in human was not studied due to its low water solubility. 122

Figure 1–4. Molecular structure of celecoxib, MW=381.373 g/mol

Celecoxib was developed as a first-in-class COX-2 inhibitor, approved for its role

in treating osteoarthritis and rheumatoid arthritis, with potential in treating acute pain. 118

It soon became a blockbuster drug and widely prescribed by doctors. All dosages (100 to

19

400 mg twice daily) has clinically shown similar efficacy to conventional NSAIDs (e.g.

500 mg naproxen) in pain relief and improvement in functionality. Meanwhile, CEL has a

much safer gastrointestinal profile and is well tolerated by patients. 118 The recommended

dose of CEL is 200 mg/day for osteoarthritis, and 100 to 200 mg twice daily for rheumatoid

arthritis in adults. No significant food effect was observed in healthy human subjects. 121

1.2.5.2 Challenges for Direct Compression Tablet Manufacture

The marketed Celebrex® capsules utilized the most thermodynamically stable -

Form III as the API solid form of CEL. 3, 4 Aside from the Celebrex® capsules which are

of relatively large sizes, only a fixed dose combination named CONSENSI® (amlodipine

and celecoxib) was approved by the FDA in 2018, but no tablet containing only CEL is

available. Although direct compression is the most simple and efficient way of tablet

manufacture, several problems need to be solved to successfully produce CEL tablets.

From the manufacturability perspective, CEL form III has very low bulk density 123

and thus very poor flow properties. 124 This is problematic for effective mixing, successful

die filling during high speed tableting. CEL Form III also exhibits severe punch sticking

problems during tableting. 115 Lamination is also a prominent issue related to CEL tablets

at high tableting speeds. In addition, the high dose (200 mg) required to elicit a therapeutic

benefit makes the tablet formulation highly sensitive to the CEL solid form properties in

terms of manufacturability. Since CEL absorption is limited by solubility, its dose could

be further decreased if its solubility is increased, leaving more room for functional

excipients to attain desired properties of the tablet formulation.

20

Much effort has been focused on increasing the solubility and thus the

bioavailability of CEL. Guzmán et al. have used high throughput screening methods to

screen for the excipients that can delay the precipitation of a high kinetic solubility sodium

salt form of CEL. 125 It was shown that the CEL sodium salt formulation with TPGS/HPC

and Pluronic F127/HPC exhibits complete absorption (100% oral bioavailability) in dogs

at 5mg/kg dose, whereas the Celebrex® only achieved 40% bioavailability at the same dose.

125 Other efforts have been made to develop amorphous solid dispersions to increase its

bioavailability. Much work has been focusing on identifying a polymer that could

effectively inhibit the nucleation of CEL during dissolution. 126 Others have explored the

effect of surfactants on the crystallization and dissolution performance of CEL amorphous

solid dispersions (ASDs) as spray-dried particles 127, 128 or hot-melt extrudates 129 . A

ternary solid dispersion of CEL has also been proposed for both quick drug release and

nucleation inhibition. 27 Other techniques explored for enhancing CEL solubility and

dissolution rate include a solid phospholipid nanoparticle dispersion via freeze-drying or

spray-drying; 130 nanoparticles generated by antisolvent precipitation and high pressure

homogenization techniques 131; CEL glass solutions produced by hot-melt extrusion and

supercritical carbon dioxide. 132

Some research has touched on solving problems related to the manufacturability of

CEL. For punch sticking of CEL Form III, Wang et al. utilized a cocrystallization method

to overcome this challenge 117; Chen et al. employed spherical crystallization with a

polymer as a way to alleviate sticking 133. The compaction of amorphous and crystalline

CEL, and CEL ternary ASD were compared 134, and the compression-induced

destabilization in melt-quenched CEL was examined by low-frequency Raman

21

Spectroscopy. 135 The mechanical properties of a CEL glass were studied by

nanoindentation for effect of drug loading and RH. 136

So far, little attention was focused on the processability and manufacturability of

CEL solid forms for tablet compression. No efforts were made to evaluate and understand

the solid forms of CEL for the purpose of developing the direct compression tablets. As a

high dose drug, CEL formulation is sensitive to the API properties. Therefore, overcoming

the poor tablet quality (lamination) problem and poor flow of CEL Form III by crystal

engineering is of practical interest. Per the crystal engineering principle, the design is

achieved by first understanding the molecular origin of the properties of CEL, which is

then followed by rational proposal and engineer for the desired crystal properties. By

employing amorphous solids as the API solid form, the potential dose needed to exhibit

adequate therapeutic effects would be lowered. Thus, the stability and crystallization

mechanisms of amorphous CEL are critical to such designs for direct compression tablet

manufacture. The stability of amorphous CEL and its crystal growth mechanism in its solid

state have not been studied before.

1.3 Objectives and Hypotheses

The main goal of this thesis is to use solid state engineering methods to solve the

issues associated with the existing commercial Form III of celecoxib that hinder the

development of a direct compression tablet formulation. The guiding hypothesis of this

thesis is that solid-state engineering is effective in overcoming deficient properties of CEL

and enables the successful development of a direct compression tablet formulation. In

testing this hypothesis, it is necessary to characterize the solid-state properties of various

22

novel solid forms of CEL, including the crystal growth mechanism and kinetics of

amorphous CEL and structure-property relationships of new crystal forms of CEL.

The objectives of this thesis are:

1. To understand the structural origin of the exceptional elasticity of CEL Form

III crystals from a molecular perspective. Its rarely seen high elasticity is

unfavorable for good tablet quality and performance during high speed

tableting.

2. To screen for a pharmaceutically acceptable solvate with better flow,

compaction and dissolution properties that can be employed to develop a DC

tablet formulation. The first step of this exercise is to gain a clear understanding

of the relationship between crystal structure and solid-state properties and

mechanical properties of these new crystal forms. New insights into the

structure-property relationship lay a foundation for future crystal engineering.

3. To develop and comprehensively characterize a direct compression tablet

formulation, a suitable solid form with improved manufacturability.

4. To prepare and understand the physical stability of amorphous CEL by studying

the crystal growth mechanism and role of polymorphism to lay a foundation for

future development of CEL ASDs for direct compression tablets. The enhanced

solubility of amorphous CEL is expected to reduce dose and hence, leaving

more room of using functional excipients for developing a tablet formulation of

high performance.

23

1.4 Research Plan and Thesis Organization

In Chapter 2, the structural origin of the exceptionally high elastic flexibility of

CEL Form III crystals is explained at the molecular level. The high elasticity is illustrated

by both qualitative three point bending and quantitative nanoindentation tests. A structural

model is developed to explain the highly restorative forces among the CEL molecules in

Form III needle crystals of CEL. This model is verified by observing changes in micro-

Raman spectra of different regions of the bent crystal. The understanding derived from this

work can be used to design crystals with desired elasticity and, thus, better compaction

properties during high speed tableting.

In Chapter 3, two newly discovered stoichiometric solvates of CEL with N-Methyl-

2-Pyrrolidone are characterized. A complex and intriguing conversion from CEL-NMP

disolvate to monosolvate is carefully studied. The drastically different stabilities of CEL-

NMP disolvate and monosolvate is evaluated and analyzed from their crystal structures.

The established structure-property relationship aids the future development of a stable

solvate that is suitable for direct tablet compression.

In Chapter 4, a pharmaceutically acceptable dimethyl sulfoxide (DMSO) solvate of

CEL, CEL-DMSO is characterized in terms of crystal structure, stability, physicochemical

properties, flow properties, compaction properties and dissolution. Then, CEL-DMSO is

used to enable the development of a a direct compression tablet formulation. The CEL-

DMSO formulation exhibits superior flow properties, easy die-filling and satisfactory

tablet quality. The CEL-DMSO tablets show faster disintegration and release than the

24

commercial capsules, which successfully demonstrates the potential of the rarely explored

pharmaceutical solvates in solving formulation problems.

In Chapter 5, the amorphous CEL is studied. The amorphous CEL is prepared by

melt-quenching, and its stability is assessed by monitoring its crystal growth rate at

different temperatures. The results show that below the glass transition temperature of

amorphous CEL, the crystal growth rate increase abruptly by activating a new mechanism

called Glass-to-Crystal growth mechanism. The surface-enhanced crystal growth indicates

that elimination of free surface can stabilize amorphous CEL. Different polymorphs of

CEL also exhibited different the crystal growth profile. The results from this study lay a

foundation for the future development of direct compression tablets based on ASD of CEL.

25

Chapter 2.

Exceptionally Elastic Crystals of Celecoxib Form III

This chapter has been published as an ACS Editors’ Choice research article in

Chemistry of Materials.

Wang, K.†, Mishra, M. K.†, Sun, C. C.*, Exceptionally Elastic Single-Component

Pharmaceutical Crystals. Chem. Mater. 2019, 31 (5), 1794-1799.

26

2.1 Synopsis

We report here the structural origins of the elastically bendable single component

pharmaceutical crystal, celecoxib Form III. Interlocked molecular packing without slip

plane and the presence of isotropic hydrogen bond network are major structural features

responsible for both the exceptional elastic flexibility and high stiffness of the celecoxib

crystal revealed by bending and nanomechanical studies. The molecular model of the

exceptional elasticity is rationalized by the inhomogeneous spatial separations of

molecules in the bent crystal, which is further confirmed by micro-Raman spectroscopy.

Celecoxib crystal, exhibiting both therapeutic effects and elastic mechanical behavior,

could be used to manufacture functional micro-devices with novel medical applications.

Figure 2–1. Synopsis figure of elastic CEL crystal

27

2.2 Introduction

Molecular crystals with exceptional mechanical properties have gained prominent

attention due to the wide range of potential applications, such as pharmaceuticals,

explosives, flexible optoelectronics, bioinspired natural fibers, food and fine chemicals.6,

89, 92, 137-147 The mechanical behavior of molecular crystals depends on the types of the

atoms, ions or molecules, molecular arrangement, and strength of intermolecular

interactions.6, 89, 92, 137-139, 148 Property engineering in molecular crystals is the main

objective of the third-generation crystal engineering, which is accomplished via the

modulation of the underlying structure.137, 138 Crystal engineering is also widely used in

designing active pharmaceutical ingredients (APIs) with optimum physicochemical

properties, such as solubility, dissolution rate, bioavailability, and mechanical properties.6,

137, 138, 149-18 Understanding the mechanical behavior of pharmaceutical crystals in the

context of crystal packing, slip planes, and crystal defects is of significant importance both

scientifically and practically.149-156 For instance, the success in the processes of milling and

tableting of an API depends, to a large extent, on its mechanical properties.6 A key to

effective design of pharmaceutical crystals with optimum mechanical properties is the

establishment of an appropriate structure-mechanical property relationship.

Mechanical plasticity in most molecular crystals can be rationalized based on

crystal packing anisotropy and the presence of slip systems.6, 89, 92, 137-139 Hand-twisted

helical crystals have also been recently designed by introducing two-dimensional plasticity

with a fair degree of isotropic crystal packing.137 In contrast, elasticity predominately

depends on the isotropy of both crystal packing and strength of intermolecular

28

interactions.6, 89, 92, 137-139 Examples of elastic behavior of molecular crystals and co-crystals

have been reported and rationalized.138, 139, 147, 148, 157-166 Elastic crystals of drugs with

therapeutic characteristics could be a class of materials with novel mechano-

pharmaceutical applications. 6, 89, 92, 137-139, 167-169 For example, a pharmaceutical crystal with

combined elasticity and optical fluorescent properties could be exploited for medical and

diagnostic imaging in biomedical applications.167, 168 Despite the appealing aspects of

elastically bending crystals with biopharmaceutical activities, their design is extremely

difficult due to a lack of clear understanding of the structure-elasticity relationship at the

molecular level.

So far, the known elastic pharmaceutical crystals are all multi-component, e.g.,

cocrystal, solvate, and salt. 22, 170, 36 The first reported elastic molecular crystal was a

caffeine cocrystal solvate.157 Subsequently, an elastic biocrystal of clofazimine

hydrochloride salt was discovered.170 In both cases, the solvent and chloride anion play an

integral role in the structural stability and elasticity during bending. Recently, a cocrystal

of 4,4′‐azopyridine and probenecid, a uricosuric agent, was shown to exhibit unique elastic

behavior under external stimuli of heat, light, or mechanical load.171 However, an elastic

pharmaceutical single component crystal remains to be discovered. Herein, we report the

first elastic crystal of a blockbuster drug, celecoxib Form III (CEL, Figure 2–2). CEL is a

COX-2 selective nonsteroidal anti-inflammatory drug for treating pain and inflammation

associated with osteoarthritis or rheumatoid arthritis, and other acute pain in adults.172, 173

Although no extraneous molecules are present in the crystal lattice, CEL still exhibits

surprisingly elastic bending behavior. We conducted bending and indentation experiments

29

to quantify the elastic properties of the CEL crystal and identified responsible features in

the crystal structure and key intermolecular interactions.

Figure 2–2. Molecular packing in CEL viewed into (01-1), (001) and (100) faces.

2.3 Materials and Methods

2.3.1 Materials

Celecoxib (CEL) was purchased from Aarti Drugs Pvt Ltd. (Mumbai, India), and

used as received. The solid-state polymorphic form of as-received CEL was confirmed by

powder X-ray diffraction to be form III.

2.3.1.1 Single Crystal Preparation of Elastic Crystals of CEL Form III

200 mg of Form III as-received Celecoxib was dissolved in 10 mL methanol. The

methanol solution of CEL was allowed to slowly evaporate in the chemical hood until

dryness. Very long needles were crystallized from the solution and used in looping,

bending and single crystal X-ray diffraction experiments.

30

2.3.2 Methods

2.3.2.1 Single Crystal X-ray Diffraction Experiment

Single crystal X-ray diffraction (SCXRD) of form III CEL was performed on a

Bruker D8 Venture diffractometer (Bruker AXS Inc., Madison, Wisconsin) equipped with

a Bruker PHOTON-II CMOS detector. The data collection was done at 100 (2) K with a

MoKα radiation source (IμS 3.0 microfocus tube). Data integration was performed with the

SAINT program, the SADABS program was used for scaling and absorption correction

purposes and XPREP was used for space group determination and data merging. The

crystal structure was solved and refined using the ShelXle program (a graphical user

interface for SHELXL174). The crystal structure was solved using SHELXT (Intrinsic

Phasing) methods. The hydrogen atoms were either placed geometrically from the

difference Fourier map or allowed to ride on their parent atoms in the refinement cycles.

All non-hydrogen atoms were refined with anisotropic displacement parameters. CCDC

1875184 contains the supplementary crystallographic data for CEL form III for this paper.

These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by

emailing [email protected], or by contacting The Cambridge Crystallographic

Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223 336033.

2.3.2.2 Face Index of CEL Form III

Face indexing of needle–shaped CEL form III crystals was done on a Bruker D8

Venture diffractometer (Bruker AXS Inc., Madison, Wisconsin) equipped with a Bruker

PHOTON-II CMOS detector. The major face is (001) along the c axis; the minor faces

along c axis are (011̅) or (01̅1) face. Another minor face is (100).

31

2.3.2.3 Three-point Bending and Looping Tests of CEL Crystals

The three-point bending, and looping tests were carried out using metallic needle

and forceps on the major face (001) of CEL needled crystals. The video was recorded on

screen using a Polarized Light Microscope (Leica M165C, Leica Microsystems Inc.,

Buffalo Grove, IL). Thickness and radius of the bent CEL crystal was measured from

video screenshots by Image J program. 175

For deflections of a beam in pure bending, i.e. shear component is not included, the

Euler-Bernoulli beam theory is used for strain calculation upon bending. 176

𝜖𝑥 =𝑦

𝑅× 100% Eqn. 2–1

Here 𝜖𝑥 is the normal strain, 𝑦 is the distance between the plane of interest to the

neutral plane (whose length and location does not change during bending), and 𝑅 is the

radius of curvature of the neutral plane. Since the side view cross-section of the needle-

shaped CEL crystal beam is a rectangle (011̅ or 01̅1 face), the neutral plane is therefore

the central plane, parallel to the bending plane (001). Thus, the maximum bending strains

are exerted to the innermost and outermost planes of this crystal, with the innermost plane

being compressed (shortened) and the outermost plane being stretched (elongated).

Regardless of sign conventions for shortening (negative) and elongation (positive), the

numerical values of 𝜖𝑥 for the innermost and outermost plane are the same. If the measured

thickness of CEL crystal is represented by 𝑡, and the measured diameter of curvature is 𝑑,

the absolute value for 𝜖𝑥 is thus:

32

𝜖𝑥 =

𝑡 2⁄

𝑅 2⁄× 100% Eqn. 2–2

2.3.2.4 Nanoindentation

The TI 980 Triboindenter (Hysitron Inc., Minneapolis, MN, USA) system was used

to perform nanoindentation experiments. A standard Berkovich diamond indenter tip was

employed. Indentations were performed on the (001) face (bending face) of CEL form III

crystal under force control mode, with 5 mN/s loading and unloading rate and a peak load

of 5 mN held for 20 s. A total of 15 valid indentations were used to calculate Elastic

Modulus (𝐸) and Hardness (𝐻) using Oliver-Pharr method.177 The 20𝜇𝑚 × 20𝜇𝑚 area

containing the indentation mark was scanned for visualization of surface topology.

2.3.2.5 Raman Spectroscopy

Bent crystals for Raman spectroscopy were prepared by gluing two ends of an

elastic CEL crystal onto a glass slide, preventing it from recovering to the normal straight

configuration. Raman spectra of straight and bent CEL crystals were collected using a

confocal Raman microscope (Witec alpha 300R, WITec, Ulm, Germany), equipped with a

CCD detector. CEL straight and bent crystal samples were excited using a 532 nm

Omnichrome Argon ion laser. For straight CEL crystal, the 10× lens was used, and a 5-

second integration time with 20 accumulations was utilized for spectra collection. For the

bent CEL crystal, both the inner part and outer part of the curve were tested using a 100×

lens, which provides a laser beam size of 2 μm in dimeter. For the inner and outer points

on the bent crystal, the spectral acquisition integration time was 5 seconds and the average

of 10 spectrum accumulations was obtained. The vibrational computations of CEL was

calculated by using Becke-3-Lee Yang Parr (B3LYP) density functional theory (DFT)

33

method with 6-31G* basis set in ground state using Gaussian-09 program to assign the

relevant bands in the experimental spectra. 178 The positive value of the calculated

vibrational wavenumbers show that the optimized molecular structure is stable.

2.4 Results and Discussion

Acicular shaped CEL crystals, 3 to 5 mm in length and 0.05-0.1 mm in width, were

obtained by slow evaporation of a methanol solution at room temperature. The previously

reported crystal structure of CEL was of insufficient quality (R factor = 8.8%).172 For

accurate structural analysis, we solved CEL crystal structure with a higher quality (R factor

= 4%) (Table 2–1). The two dominant faces of CEL crystals were (001)/(00-1) and (01-

1)/(0-11), and the end faces of the crystals were (100)/(-100), identified from a face

indexing experiment using Single Crystal X-ray Diffraction (SCXRD) (Figure 2–2 &

Figure 2–3).

34

Table 2–1. Crystallographic information for CEL Form III crystals

Formula C17H14F3N3O2S

Molecular weight 381.37

Crystal system Triclinic

Space group P-1

a (Å) 5.0627(5)

b (Å) 10.0139(11)

c (Å) 16.3981(17)

α (°) 89.874(4)

β (°) 85.867(3)

γ (°) 80.530(4)

Volume (Å3) 817.84(15)

Z/Z′ 2/1

ρcalc (g/cm3) 1.549

F(000) 392

Temperature (K) 100(2)

R1 0.0409

wR2 0.0834

Goodness-of fit 1.10

CCDC No. 1875184

Figure 2–3. Face index of needle-shaped CEL form III crystal

A crystal structure analysis of CEL shows that the 4-methylphenyl and

trifluoromethyl-1H-pyrazol rings are approximately in plane (18°) and are each

approximately perpendicular (89.56° and 83.78°, respectively) to the benzenesulfonamide

(𝟎𝟏𝟏)

(𝟎𝟏𝟏) (𝟎𝟎𝟏)

(𝟎𝟎𝟏)

35

ring (Figure 2–4). The -NH2 group of benzenesulfonamide of each CEL molecule is

bonded with N of the pyrazol ring to another CEL molecule (N–H···N, D, d, θ: 3.038 Å,

2.19 Å, 159°) to form a hydrogen bonded dimer (Figure 2–5). The dimer extends along

the a-axis to form a 1-dimensional (1D) chain fortified by N–H···O (2.922 Å, 2.09 Å,

166°), C–H···N (3.517 Å, 2.740 Å, 139.79°), C–H···π (3.869 Å, 2.966 Å, 159.98°), and

C–H···π (3.805 Å, 2.976 Å, 143.97°) hydrogen bonds (Figure 2–5b and Figure 2–6a).

Neighboring 1D chains are connected by various weak intermolecular interactions, i.e., C–

H···F dimer (3.516 Å, 2.59 Å, 165.2°), C–H···F linear (3.578 Å, 2.696 Å, 149.95°), and

C–H···π dimer (3.936 Å, 2.971 Å, 168.4°). CEL dimers from neighboring 1D chains are

offset by 4.896 Å along the direction of chain. The connection of dimers via C–H···π

dimer (3.936 Å, 2.971 Å, 168.4°) hydrogen bond leads to corrugated layers parallel to the

(100) plane (Figure 2–6 b). Each corrugated layer is connected by bifurcated C–H···O

dimer (3.255 Å, 2.56 Å, 129°; 3.358 Å, 2.819 Å, 116.82°), C–H···F dimer (3.516 Å, 2.59

Å, 165.2°) and, C–H···F linear (3.578 Å, 2.696 Å, 149.95°) to form interlocked packing

along (001) plane (Figure 2–5 and Figure 2–6). Overall, the structural features of CEL

satisfy the essential criteria for elastic crystals with a specific bendable face: 1) interlocked

isotropic packing without any slip planes, which prevents long-range molecular movement

to avoid plastic deformation during flexing; and 2) presence of multiple dispersive

interactions in orthogonal directions, which leads to “structural buffering” to allow

reversibly stretched intermolecular interactions during elastic bending.157, 158, 166

36

Figure 2–4. Molecular conformation of CEL represented by perpendicular planes

Figure 2–5. Intermolecular interactions pattern from two different orientations (a) and

(b). Interaction within the dimer is highlighted in the gray color.

37

Figure 2–6. (a) 1-dimensional chain along a-axis. (b) Corrugated layered structure of

CEL molecule, with yellow color for C-H···π interactions connecting the dimers.

An acicular CEL crystal could be easily and repeatedly bent (Figure 2–7a-i), when

a load was applied to the major crystal face, (001)/(00-1), by a metallic needle. An arc was

formed without crystal breakage (Figure 2–7b), which was rapidly reversed upon

withdrawal of the applied load (Figure 2–7c). This elastic bending phenomenon was highly

reversible, as it could be performed multiple times without any sign of fatigue. However,

when the load was applied to the minor face (0-11) / (01-1) of the same crystal, the crystal

broke before appreciable strain could be developed (See Figure 2–8). Therefore, the CEL

crystal is a 1D elastic crystal, since elastic deformation occurs at only one face. In addition,

CEL is a rare example of single component elastic crystal without halogen bonds. Thus,

unlike what was thought before, the presence of halogen bonds is not a prerequisite for

elastic organic crystals. 31

38

-

Figure 2–7. Screenshots of bending tests on the bendable face (001) / (00-1) of a CEL

single crystal using forceps and a needle, showing four repeated cycles of bending and

elastic recovery(a) ~ (i).

Figure 2–8. Screen shots of bending on CEL crystal minor and major faces – when bent on

major (001) face, crystal is elastic; when bent on minor (011̅) or (01̅1) face, the CEL crystal is

brittle.

39

The CEL crystal could be elastically bent to make a closed loop (Figure 2–9). The

complete shape recovery from the loop reflects remarkable elastic flexibility.179 The elastic

strain of the looped crystal, calculated using the Euler–Bernoulli beam-bending theory,180

is about 3.56% (Figure 2–10). This is significantly higher than the maximum elastic strain

(< 0.5%) of most crystalline materials.181 The highest reported value of maximum elastic

strain for known elastic polyhalogenated based molecular crystals is 2% for 2,3‐

Dichlorobenzylidine‐4‐chloroaniline.158, 166 Thus, the elastic strain of CEL crystal is at least

78% higher than the previously known most elastically flexible organic crystal.

Figure 2–9. (a) - (c). Looping of elastic CEL crystal and its complete recovery.

Figure 2–10. Radius of curvature for maximally bent loop of CEL

40

The mechanical properties of the bendable major face, (001)/(00-1), of CEL

crystals were characterized using nanoindentation (Figure 2–11), which provided a highly

accurate measurement of the mechanical properties of molecular crystals by examining an

essentially defect-free area of a sample.89, 138, 153-156, 182-185 The elastic modulus (E) and

hardness (H) were 16.27 ± 0.43 GPa and 0.45 ± 0.02 GPa, respectively. This E value is

significantly higher (28%) than the highest E reported for elastically bendable organic

crystal (N-2,5-dichlorobenzylidine-4-iodo aniline).166 This is attributed to the presence of

a large number of relatively weak hydrogen bonds (C–H···π, C–H···F, C–H···O, and C–

H···N) between CEL dimers, which are fortified by strong N–H···N and N–H···O

hydrogen bonds, in the CEL crystal compared to those in N-2,5-dichlorobenzylidine-4-

iodo aniline, where only weak hydrogen bonds (C–H···Cl) and halogen bond (type 1 Cl···I)

are present. Thus, incorporating both strong and weak intermolecular interactions in the

crystal structure may be an effective strategy for designing highly elastic crystals with high

stiffness.

Figure 2–11. Nanoindentation characterization of the CEL crystal. a) Force-depth curve

and b) 3D scan of indented area.

41

All proposed qualitative mechanistic models predict an inhomogeneous molecular

distribution in the lattice of bent elastic crystals.157, 158 The splaying of molecules and

sliding of the molecular layers in the elastically deformed crystal affect the intermolecular

interactions in the three domains, i.e., outer, inner, and central arc (Figure 2–12).186 In the

outer arc, the pronounced weakening of the intermolecular interactions is expected due to

the larger intermolecular separations due to tensile strain. However, intermolecular

interactions are stronger in the inner arc due to the closer proximity among molecules

resulting from the compressive strain, which may even introduce new intermolecular

interactions. The structural arrangement in the neutral axis (middle part) of the bent crystal

is the same as the straight crystal. Such inhomogeneous interactions in a bent crystal can

be experimentally characterized by in-situ micro-Raman spectroscopy, which can measure

the structural inhomogeneity in the elastically bent CEL crystal (Figure 2–13).186

Figure 2–12. Schematic representation of the molecular rearrangement in the elastic

crystal during the bending state.

42

Figure 2–13. Calculated Raman spectra of CEL

Careful inspection of the Raman spectra in the three domains reveals broadening

and shifting in bending modes of lattice vibration and aromatic C-H bonds.187 Thus, the

crucial role of aromatic C-H bonds in the elasticity of CEL crystal (Figure 2–5) is

confirmed. The 102.84 cm−1 band (Figure 2–14a) of a straight crystal corresponds to the

strong lattice vibrations of CEL. Therefore, the changes in the peak indicate the prominence

of the strong lattice vibration in strained domains. Strengthening the intermolecular

interactions and the formation of multiple, new hydrogen bonds due to closer proximity of

molecules in the compressed inner arc result in a significant broadening of the 102.84 cm−1

band. However, the same 102.84 cm-1 band in the outer arc shows a blue shift with

broadening up to 113 cm-1 due to the fewer and weaker intermolecular interactions as the

43

molecules are farther apart. The aromatic C–H group of 4-methylphenyl and pyrazol ring

of CEL are involved in the formation of several weak interactions, such as C–H···F, C–

H···O, and C–H···N, which act as a structural buffer during elastic bending.

Figure 2–14. Raman spectra corresponding to the bent and straight crystals of CEL, with

red and blue circles indicate the point of data collection (inner arc and outer arc

respectively) a) lattice vibration, and b) aromatic symmetric in-plane bending (C–H)

modes of inner and outer arc of a bent CEL single crystal, as well as in the non-strained

straight state.

The bands at 1596 cm-1 and 1613 cm-1 of unbent crystal correspond to the

symmetric in-plane bending (C–H) modes of aromatic C–H group (Figure 2–14b). The

1596 cm-1 band shows broadening with increased intensity in the compressed inner arc due

to the formation of multiple hydrogen bonds involving the aromatic C–H groups. However,

the same band in the outer arc became sharper and blue-shifted due to weakening of the

intermolecular interactions as well as strengthening of the C–H bonds. The band at 1613

44

cm-1 shows significant broadening in the compressed inner arc, which is expected due to

strengthening of the interactions involving aromatic C-H group. The intensity of the 1613

cm-1 band is diminished with broadening in the outer arc. This effect is expected for

structural modifications due to induced maximum strain and, hence, longer hydrogen bonds

present in the outer arc. These observations validate the mechanistic bending model of

elastic CEL crystal shown in Figure 2–12 and confirm the inhomogeneous spatial

separations of molecules during elastic bending.

2.5 Conclusion

In summary, we report the first elastically bendable single component

pharmaceutical crystal. Two major structural features responsible for the superior elasticity

in CEL are: (1) Corrugated structure with criss-cross interlocked packing without slip

planes, and (2) numerous weak and a few strong dispersive interactions in orthogonal

directions. An interlocked packing without slip planes restricts long range molecular

movement away from the equilibrium positions during elastic bending. Furthermore, the

bending and nanoindentation experiments revealed a new strategy for designing highly

elastic crystals by incorporating both strong and weak intermolecular interactions. The

structural perturbations model was validated by a Raman spectroscopic analysis, which

confirmed the inhomogeneous spatial separation of molecules corresponding to the

expected changes in distances among molecules in different domains of bent crystal. A

structural understanding of the mechanical elasticity in this single component

pharmaceutical crystal can guide future designs of molecules for mechano-pharmaceutical

45

applications, including low-cost flexible smart functional microdevices for biomedical

applications with advantageous mechanical properties.188

46

Chapter 3.

Structural Insights into the Distinct Solid-state Properties and

Interconversion of Celecoxib N-Methyl-2-Pyrrolidone Solvates

This chapter has been submitted as research article to Crystal Growth and Design.

Wang, K., Wang, C., Sun, C. C.*, Structural insights into the distinct solid-state

properties and inter-conversion of Celecoxib N-Methyl-2-Pyrrolidone solvates. Cryst.

Growth Des. 2020, reivision submitted.

47

3.1 Synopsis

In an effort to develop a tablet product of celecoxib by overcoming its poor

physicochemical properties using a pharmaceutically acceptable solvate, we isolated two

stoichiometric N-Methyl-2-Pyrrolidone (NMP) solvates of CEL, mono-NMP and di-NMP

solvates. Here, we report the preparation, characterization, and structural origin of different

properties of the two solvates, including a rare and complex heating induced phase

transformation from di-NMP to mono-NMP solvate.

48

3.2 Introduction

Celecoxib (CEL, Figure 3–1, C17H14F3N3O2S) is a non-steroidal anti-inflammatory

drug that selectively inhibits COX-2 for the treatment of pain and inflammation associated

with a number diseases, such as arthritis and osteoarthritis.119 As a BCS class II drug, the

poor solubility of CEL in aqueous media leads to slow dissolution and, thus, delayed onset

of action for pain relief. CEL is marketed as capsules, available in 100, 200 and 400 mg

strengths, under the brand of Celebrex®. An issue with capsules is the larger size when

compared to tablets of the same content, which leads to lower patient compliance because

of the difficulty with swallowing. An improved drug product of celecoxib would be one

that can release CEL more rapidly and in smaller size.

Among available options, crystal engineering is a promising approach to

simultaneously improve solubility, dissolution, and tablet manufacturability. 189, 190 The

most commonly employed crystal engineering techniques are cocrystallization, 96, 111 salt

formation, 107 and hydrate formation. 38, 191 A less well explored but still effective approach

is the use of a pharmaceutically acceptable solvate. 192 Organic solvents that can potentially

be used for this approach include ethanol, dimethyl sulfoxide (DMSO), N-methyl-2-

pyrrolidone (NMP), propylene glycol (PG), ethanol and polyethylene glycol (PEG), as long

as the consumed amount of the solvent accompanying the therapeutic dose of the drug

remains safe. 193 An example of this approach is Crixivan, which is marketed as indinavir

sulphate - ethanolate.194 In addition to being a potential API solid form candidate, solvates

are frequently used as intermediate crystalline forms during API synthesis for impurity

rejection via crystallization . 192 195

49

Known solid forms of CEL include polymorphs I, II, III, and IV, 3, 196 an amorphous

form, 32 sodium salt hydrates, 125, 197 some solvates,3, 198 and various cocrystals with

common cocrystal formers 199-204 or other drugs such as venlafaxine 205 and tramadol

hydrochloride. 206 The room temperature thermodynamically most stable Form III is used

to manufacture commercial CEL capsules. The choice of capsule dosage form was partially

attributed to its very poor flowability 4 and high punch sticking propensity,115, 133, 207, 208

which make the commercial manufacturing of tablet products challenging. In an effort to

simultaneously solve these problems using an alternative solid form to enable the

development of a CEL tablet formulation, we explored possible pharmaceutical solvates of

CEL. The focus of this work was to characterize CEL mono-NMP and di-NMP solvates

and provide molecular level understanding of their distinct properties based on an analysis

of their crystal structures. An intriguing series of phase changes of the di-NMP solvate

upon heating was captured and explained.

Figure 3–1. Chemical structure of a) celecoxib (MW = 381.37) and b) NMP (MW =

99.13)

N

N

CF3

CH3

SH2N

O

O

C17H14F3N3O2S MW 381.37

N

CH3

C5H9NO MW 99.13

O

Celecoxib NMP

a) b)

50


3.3.1 Materials

Celecoxib (Figure 3–1) was purchased from Aarti Drugs Pvt Ltd. (Mumbai, India).

N-methyl-2-pyrrolidone (NMP, Figure 3–1, PharmasolveTM) was obtained from Ashland

Global Specialty Chemicals Inc (Covington, Kentucky, USA). All materials were used as

received.

3.3.2 Methods

3.3.2.1 Single Crystal X-ray Crystallography

Single crystals of CEL mono-NMP solvate were prepared from a saturated solution

of CEL in NMP via slow evaporation at room temperature. Single crystals for structural

determination of CEL di-NMP solvate were prepared by cooling a solution of 2.8 g of CEL

in 5 mL of NMP from 60 ºC to 3 ºC.

Single crystal X-ray diffraction (SCXRD) of both CEL-NMP solvates was

performed on a Bruker D8 Venture diffractometer (Bruker AXS Inc., Madison, Wisconsin)

equipped with a Bruker PHOTON-II CMOS detector. Data collections were done at both

298 K and 100 K for di-NMP solvate with a MoKα radiation source (IμS 3.0 micro focus

tube). Data integration was performed with the SAINT program, the SADABS program

was used for scaling and absorption correction, and XPREP was used for space group

determination and data merging. The crystal structure was solved using SHELXT-16

(Intrinsic Phasing) and refined using SHELXL-16, which were executed from the ShelXle

graphical user interface. 209 Hydrogen atoms were either placed geometrically from the

51


The disordered CF3 group of CEL in the CEL-NMP 1:2 structure at 298 K was modeled,

where two sets of three F atoms’ site-occupancy factors were constrained to add up to unity.

198 To ensure a reasonable geometry, the C-F distances and the F-F distances were

restrained to a common refined value of 1.33 Å and 2.14 Å, respectively. All F atoms were

refined with anisotropic displacement parameters. The anisotropic displacement

parameters in both disordered parts are modeled to be equivalent.

3.3.2.2 Preparation of NMP Solvate Bulk Powders

A slurry was prepared by suspending an excess amount of CEL solid (~ 2.5 g) in

NMP (~ 5 mL) in a 20 mL scintillation vial for at least 10 min by a magnetic stirrer at a

temperature of interest ranging from 3 ºC to 45 ºC. A 1:1 (w : w) seed mixture of the two

solvates was then added. The desired temperature was maintained by means of a water-

jacketed beaker connected to a temperature-controlled water bath. After at least 3 days, the

solid was vacuum filtered in a Buchner funnel for 10 min, and then analyzed for solid forms

by powder X-ray diffraction. Preliminary work showed that both solvates remained phase

stable after vacuum filtration for 10 min. Such information allowed the determination of

the thermodynamic stability of the two solvates in NMP at different temperatures, which

was used to bracket the thermodynamic transition temperature between the two solvates.

3.3.2.3 Powder X-ray Diffraction (PXRD)

Powders were scanned over a 2θ range of 5°−35° on a wide-angle X-ray

diffractometer (X’Pert PRO; PANalytical Inc., West Borough, MA) using Cu Kα radiation

(45 kV and 40 mA) at a step size of 0.0167° and a dwell time of 1.15 s. The simulated

52

PXRD patterns of CEL Form III, mono-NMP, and di-NMP solvates were calculated from

respective 298 K single crystal structures using the Mercury software (v4.3.1 Cambridge

Crystallographic Database Centre, Cambridge, UK).

3.3.2.4 Thermal Analyses

Thermal behaviors of samples were characterized using a Differential Scanning

Calorimeter (DSC, Q1000; TA Instruments, New Castle, DE), heated from room

temperature to 180 °C at a rate of 10 °C/min. A sample was placed in an aluminum pan

which was then crimped with an aluminum lid. The cell constant for heat flow was

calibrated using indium and the cell temperature was calibrated with indium and

cyclohexane. The DSC cell was purged with helium gas at 25 mL/min. The maximum

temperature used was below the degradation onset temperature of CEL.32 The DSC sample

of each of the two CEL-NMP solvates was a single crystal with surface mother liquor

removed using an absorbent tissue (Kimwipe). The weight loss upon heating of different

solids was determined on a thermogravimetric analyzer (TGA, Q50, TA Instruments, New

Castle, DE). Samples were placed in an open aluminum pan and heated to 300 ºC at a rate

of 10 ºC /min under 60 mL/min nitrogen purge.

3.3.2.5 Hot Stage Microscopy (HSM)

Both NMP solvates were observed under a polarized light microscope (PLM)

(Nikon Eclipse E200, Nikon, Tokyo, Japan) equipped with a DS-Fi1 microscope digital

camera for capturing images and videos. Most crystal samples were dispersed in silicone

oil between a glass slide and a cover glass; oil and cover glass were not used in few tests

of di-NMP crystals. The samples were heated on a temperature stage (Linkam LTS 420,

53

Linkam Scientific Instruments, Ltd., Waterfield, U.K.) from room temperature to 180 ºC at

a rate of 10 ºC/min and observed under PLM for on-line monitoring of thermo events.

3.3.2.6 Fourier Transform Infrared Spectroscopy (FTIR)

IR spectra of the powder samples were collected using a high resolution FT-IR

spectrometer, equipped with a built-in iS50 diamond attenuated total reflection (ATR)

component (Nicolet iS50; Thermo Scientific, Waltham, MA) and a DLaTGS detector. A

total of 32 scans were collected and averaged for each spectrum. The range of IR spectra

collected was 400-4000 cm−1 at a resolution of 2 cm−1. IR data were processed using

OMNIC software (v9.2, Thermo Scientific, Waltham, MA).

3.3.2.7 Energy Framework Calculations

The pairwise intermolecular interaction energy was estimated using

CrystalExplorer 210 and Gaussian09 178 with experimental geometry of CEL mono-NMP

and di-NMP solvates. 211 The hydrogen positions were normalized to standard neutron

diffraction values before energy calculation using the CE-B3LYP electron densities model.

The total intermolecular interaction energy between a given molecular pair was the sum of

the electrostatic, polarization, dispersion, and exchange-repulsion components with scale

factors of 1.057, 0.740, 0.871, and 0.618, respectively. Intermolecular interactions between

two molecules with the closest inter-atomic distance of more than 3.8 Å were ignored. The

interaction energies were graphically presented as a framework by connecting centers of

mass of molecules with cylinders, with the radii proportional to the total inter-molecular

interaction energies. The crystal packing topology was analyzed using a Python program.

212

54


3.4.1 Crystal Structures of CEL-NMP Solvates

Three methods were used to prepare solvate crystals during initial solvate screening

for CEL solvates: slurry at room temperature, cooling from 60 °C to 3 °C and slow

evaporation of saturated CEL solutions in NMP at room temperature. The powder X-ray

diffraction (PXRD) patterns of the solids obtained from cooling and slurry methods were

identical, which were different from the pattern of the solids prepared by slow evaporation

(Figure 3–2). Since both PXRD patterns were different from the known CEL polymorphs,

the formation of two new solid forms of CEL was indicated.

Figure 3–2. PXRD of CEL-NMP solvates prepared using different methods, as compared

to patterns of CEL-NMP mono- and di-solvates and Form III CEL calculated from

corresponding single crystal structures

55

The solved single crystal structures (Figure 3–3) revealed the two forms are NMP

solvates with CEL:NMP ratios of 1:1 (mono-NMP) and 1:2 (di-NMP). PXRD patterns of

the powders from slow evaporation matched well with the calculated PXRD pattern of the

mono-NMP solvate, while the crystals from cooling and slurry methods matched well with

the calculated PXRD pattern of the di-NMP solvate (Figure 3–2).

Figure 3–3. ORTEP drawings of a) mono-NMP and b) di-NMP solvate structures

collected at room temperature (298 (2) K). The CF3 in both solvates and the NMP

molecule in mono-NMP are disordered.

In the mono-NMP solvate crystal, two NMP molecules and two CEL molecules

form a tetramer, which is stabilized by strong hydrogen bonds (𝑅42(8), Figure 3–4) and

serves as a building block for the CEL mono-NMP crystal. All NMP molecules are bonded

with CEL via strong hydrogen bonds (Table 3–1), marked a1, a2, a1’, and a2’, where a1’

and a2’ are symmetrically equivalent to a1 & a2 in the tetramer (Figure 3–4a). In the di-

NMP crystal, however, only half of the NMP molecules are strongly bonded with CEL to

also form the tetramer through strong hydrogen bonds (𝑅42(8), Table 3–1), marked b1, b2,

b1’ and b2’ (Figure 3–4b). The other NMP molecules fill in space between the tetramers

and do not form any strong hydrogen bonds with either CEL or other NMP molecules. It

is noteworthy that this tetramer synthon is preserved in both mono-NMP and the di-NMP

56

solvate structures (Figure 3–4). The tetramers in both mono-NMP and di-NMP crystals

are similar in terms of both molecular positions and strengths of hydrogen bonds (Figure

3–4 and Table 3–1). There is no significant difference in the conformation of CEL

molecules. However, although the orientations of NMP molecules in the two tetramers

flipped, the carbonyl oxygen on NMP is pinned to the same location in both structures

(Figure 3–5) to act as the acceptor of the strong N-H⋯O hydrogen bond within the

tetramer. Thus, the change in NMP orientation only influences the weak intermolecular

interactions but does not affect the hydrogen bonding patterns of the tetramer. The

hydrogen bond connectivity of the tetramer remains unchanged (Table 3–1).

Figure 3–4. Tetramer with strong hydrogen bonded CEL and NMP molecules in a) CEL

mono-NMP solvate and b) CEL di-NMP solvate. Light blue dashed lines are strong

hydrogen bonds labeled as 1, 2, 1’ and 2’ in both structures.

57

Table 3–1. Strong hydrogen bonds that stabilize the tetramer in mono-NMP and

di-NMP solvate crystal structures.

D-H⋯A

(N-H⋯O)

d (H⋯A)

(Å)

d (D⋯A)

(Å)

∠(DHA)

(º)

mono-NMP a1, a1’ 2.026 2.889 167.98

a2, a2’ 2.043 2.883 161.48

di-NMP b1, b1’ 2.129 2.963 162.79

b2, b2’ 1.973 2.889 165.18

Figure 3–5. Overlay of CEL mono-NMP solvate (red) and di-NMP solvate (light blue).

The conformation of CEL is the same in the two structures, although the NMP in the

tetramers adopted flipped orientations (black arrows). N-H⋯O hydrogen bonds between

NH2 of CEL and the carbonyl O atom of NMP molecules are retained.

3.4.2 Solid-state Properties of the Two CEL-NMP Solvates

The CEL-NMP monosolvate crystals obtained from slow evaporation were blocks,

while that of the di-NMP solvate crystals from cooling were blocks with higher aspect

ratios (Figure 3–6).

58

Figure 3–6. Morphology of a) mono-NMP solvate and b) di-NMP solvate of CEL. Scale

bars are 500 µm.

The mono-NMP solvate had a single endothermic event in DSC at a peak

temperature of 141 ºC, which is much lower than the peak melting temperature of Form III

CEL (163 ºC) (Figure 3–7a).32 Hot stage microscopy (HSM) data confirmed this thermal

event as melting (Figure 3–8). No weight loss was recorded by TGA until the temperature

reached ~120 ºC (Figure 3–7b), which corresponds to the onset of mono-NMP solvate

melting (Figure 3–7a). Therefore, the NMP molecules are tightly bound and could not

escape the mono-NMP crystal before the melting occurs. The total weight loss did not

reach the theoretical NMP content in mono-NMP crystals (20.64 wt %) until approximately

250 ºC, which is more than 30 ºC higher than the boiling point of NMP (202 ºC). Since

pure CEL liquid starts to evaporate at temperatures above 200 ºC (Figure 3–7b), a part of

the weight loss of mono-NMP solvate at 250 ºC may be attributed to CEL. This means that

NMP was not completely removed from the CEL mono-NMP melt at 250 ºC. This is

possible if NMP and CEL form an azeotrope in liquid phase because of the four strong N-

H∙∙∙O hydrogen bonds that stabilize the tetramers (Figure 3–4, Table 3–1).

59

Figure 3–7. Thermal properties of CEL Form III (black), CEL mono-NMP (red), and

CEL di-NMP (blue). (a) DSC and (b) TGA (Dashed lines represent first derivative of

TGA traces of the three solid forms).

Figure 3–8. CEL mono-NMP solvate heated using HSM without silicone oil at a rate of

10 ºC/min.

The DSC thermogram of the di-NMP solvate shows two separate endothermic

events. The first sharper endotherm corresponds to a step weight loss of about 10.33 %,

corresponding to about 0.6 mol of NMP, up to ~110 ºC. We attribute this to the removal of

the more loosely bound NMP molecules from the di-NMP solvate lattice (Figure 3–4b) to

form mono-NMP crystals (Figure 3–4a). The loss of approximately 0.6, instead of 1,

60

stoichiometric NMP gravimetrically (Figure 3–7b) happened because part of the NMP

removed from the di-NMP crystals remained as liquid at temperatures below the boiling

point of NMP (202 ºC). The liquid NMP dissolved some amount of CEL di-NMP and

mono-NMP crystals. The dissolution process of mono-NMP is reflected by the curved

baseline immediately following the first endothermic peak (Figure 3–7a), which would

have been flat had dissolution not occurred. 213 The second endotherm, which corresponds

to the melting of mono-NMP, is about 10 ºC below the melting point of neat mono-NMP

crystals. This observation is consistent with the depression of the melting point of mono-

NMP by the presence of liquid NMP in the hermetically sealed DSC pan. Meanwhile, the

shape of TGA curve above 120 ºC is similar to that of mono-NMP (Figure 3–7b), which

is attributed to the evaporation of NMP followed by simultaneous evaporation of CEL and

NMP above 200 º C.

3.4.3 Phase Transformation of di-NMP upon Heating

To better understand and simulate the crimped DSC pan environment, the crystals

used in the HSM study were covered with silicone oil when observed between the glass

slip and the cover glass. Corresponding to the first endothermic peak for desolvation (50.9

- 70.2 ºC) in the DSC thermogram of the CEL di-NMP solvate, simultaneous crystal

dissolution was seen because di-NMP crystals shrunk in size upon heating (Figure 3–9 b-

g,). This is possible because, after leaving the di-NMP crystals at a temperature below its

boiling point (202 ºC), NMP exists as a liquid and is in intimate contact with the remaining

di-NMP crystals and sealed by silicone oil. As the temperature rises, the desolvation and

dissolution processes occur simultaneously (Figure 3–9 b-g). The dissolution of di-NMP

61

crystals cannot be seen from the DSC curve because it coincided with desolvation of di-

NMP (Figure 3–7a). To decouple the desolvation from the dissolution processes, we

heated up the di-NMP crystals in HSM without submerging them in silicone oil, allowing

evaporation of NMP. Under this condition, di-NMP single crystals converted to

polycrystalline mono-NMP crystals in the temperature range of 50 - 70 ºC, followed by

melting of mono-NMP solvate (Figure 3–10). No dissolution was observed during the

entire temperature range. This confirms that the desolvation of di-NMP in this temperature

range indeed leads to the mono-NMP solvate, and that the crystal shrinkage of di-NMP in

Figure 3–9b-g was due to dissolution in NMP in a sealed pan condition.

Figure 3–9. HSM images as CEL-NMP disolvate was heated at 10 ºC/min and undergo a

series of events: desolvation, dissolution, and melting.

62

Figure 3–10. CEL-NMP disolvate heated using HSM at a rate of 10 ºC/min, without

silicone oil. a) ~ b) no change from room temperature to 50 ºC. c) desolvation at 70 ºC,

forming polycrystalline mono-NMP crystals (the foggy image is a result of the

condensation of escaped NMP on the glass window of HSM), d) The mono-NMP

remained stable up to at least 108 ºC (image was clear again because the condensed NMP

evaporated). e) and f) melting of mono-NMP crystals 125 – 133.3 ºC.

It is important to point out that, during heating in the 50 - 70 ºC temperature range,

new mono-NMP crystals grew while the original di-NMP crystals shrunk (dissolved) until

completely disappeared (Figure 3–9b-g). This suggests that the mono-NMP solvate has a

lower solubility in NMP than di-NMP solvate. Hence, the mono-NMP solvate is

thermodynamically more stable than the di-NMP solvate at this temperatures range. In

summary, three events are happening within the first endothermic peak in the DSC

thermogram of di-NMP solvate (Figure 3–9a) 1) desolvation of CEL di-NMP solvate to

form mono-NMP solvate, 2) dissolution of disolvate into NMP, and 3) growth of CEL

mono-NMP solvate crystals. With increasing temperatures above 70 ºC, (Figure 3–9g, h),

63

where the CEL concentration in NMP is slightly higher than the saturation solubility CEL-

NMP monosolvate, only subtle growth of few CEL-NMP crystals are observed, while the

majority of CEL-NMP 1:1 crystals remains unchanged without undergoing melting,

dissolution or desolvation. After the temperature reaches 79.5 ºC (Figure 3–9h), the

solubility of CEL-NMP monosolvate has ramped to the point where the CEL concentration

in NMP is below its solubility. Thus, dissolution of the mono-NMP solvate started, Figure

3–9 i-j) where CEL-NMP monosolvate crystals become smaller in size. This dissolution

process continues all the way to the onset temperature of CEL-NMP monosolvate melting,

which has an onset of and a peak melting temperature of 120.3 ºC, and 141.0 ºC,

respectively, for clean monosolvate (Figure 3–7a, red DSC trace). For disolvate crystals

that had undergone desolvation to monosolvate, however, a depressed melting

phenomenon prelude. Figure 3–9k) shows the beginning of melting for CEL-NMP

monosolvate, which completely melted at 133.4 ºC (Figure 3–9l). This depressed peak

melting temperature was the result of the presence of NMP liquid surrounding the CEL-

NMP monosolvate crystals. To sum up, two events happened from Figure 3–9g) to l): 1)

dissolution and 2) melting of CEL mono-NMP solvate crystals. The combination of these

two events explained the declining disolvate DSC baseline after the first endotherm

(Figure 3–7a), as well as the depressed melting peak following the declining baseline. This

is an excellent example that demonstrates the benefit of HSM in aiding the clear

interpretation of complex thermal events registered by DSC. The DSC trace marked with

thermal events captured by HSM is shown in Figure 3–11.

64

Figure 3–11. DSC thermogram of CEL di-NMP solvate upon heating at 10 ºC/min.

Corresponding thermal events at various stages are also shown with PLM images.

3.4.4 Thermodynamic Stability Relationship of Two Solvates

The mono-NMP solvate was suggested to be more stable at higher temperatures by

the HSM data under a non-equilibrium condition (heated at 10 ºC /min). To better

understand the thermodynamic stability relationship between the two solvates, we

performed competitive slurry experiments, where a mixture of the two crystalline phases

was allowed to equilibrate at different temperatures for a period of 72 hours, and the

equilibrium solid phases were then harvested and analyzed by powder X-ray diffraction

(Figure 3–12. Such data bracketed the transition temperature (Tt) between the two solvates

65

to be 35.5 ºC < Tt < 36 ºC. The di-NMP solvate is more stable below Tt, while the mono-

NMP solvate is more stable above Tt (Figure 3–12).

Figure 3–12. PXRD patterns of equilibrium solids obtained from slurry experiments and

comparisons with CEL-NMP calculated patterns from room temperature single crystal

structures.

3.4.5 Physical Stability on Storage

Although the di-NMP solvate is more stable in NMP solution at room temperature

(Figure 3–12), CEL di-NMP solvate undergoes partial desolvation to form mono-NMP

solvate at room temperature when exposed to open air for 7 days or house vacuum

overnight. In contrast, the mono-NMP solvate is stable in open air for up to 12 months. A

66

clear explanation of the drastically different stability of the two solvates requires an

analysis of their crystal structures.

3.4.6 Structural Characteristics of the Two NMP Solvates

In addition to the strong hydrogen bonding interactions within each tetramer

building block in both solvates (Table 3–1), there are many weak interactions that help to

fortify the tetramer (Figure 3–136a and Figure 3–13a). In the mono-NMP solvate, the

tetramers interact with each other only via weak hydrogen bonds in all directions (Figure

3–13). In the di-NMP solvate, a half of the NMP molecules strongly interact with CEL to

enable the formation of tetramer synthon; while other NMP molecules interact with the

tetramers through weak hydrogen bonds (Figure 3–14). A noticeable common structural

feature of both solvates is that the tetramers form a 1-D chain running along the

crystallographic b-axis. The stability of NMP molecules in the two solvate crystals is

quantitatively assessed using the pair-wise interaction energies graphically presented as the

energy frameworks, in which the tetramers are clearly visible as parallelograms in both

structures (Figure 3–15). The inter-tetramer interaction strength is weaker than intra-

tetramer interactions for both solvate crystals. The energy frameworks of mono-NMP

(Figure 3–15a) and di-NMP (Figure 3–15b) solvates match well with the hydrogen bond

networks present in the two crystals (Figure 3–13 and Figure 3–14, respectively). There

are also some weak interactions along b-axis to stabilize the 1-D chain in both solvates

(Figure 3–13a & Figure 3–14b, Figure 3–15). For the mono-NMP solvate, all NMP

molecules are part of the rigid tetramers and they are arranged in staggered positions to

form an irregular zig-zag solvent spiral running along the b axis (Figure 3–16). The high

67

energy barrier to removing NMP molecules explains the stability of CEL mono-NMP

solvate.

Figure 3–13. Weak interactions in CEL mono-NMP solvate structure: a) intra-tetramer

and inter-tetramer to form 1-D chain along b-axis; b) inter-tetramer along a-axis; c)

between tetramers along c-axis. d) viewed into a-axis.

a

c

a

b

b

c

a) b)

c) d)

a c

b

68

Figure 3–14. Weak interactions in CEL di-NMP solvate a) Intra-tetramer along c-axis; b)

along b-axis; c) along a-axis; and d) view down a-aixs

Figure 3–15. Energy framework of (a) mono-NMP viewed down a axis and (b) di-NMP

viewed down b axis. The pink parallelograms highlight a tetramer in each of the

frameworks.

a)

b

c a

c)

b c

a b)

b c

a

d)

a

c

69

Figure 3–16. Mono-NMP solvate crystal structure viewed along (a) b axis and (b) a axis,

(c)~(e) zig-zag void channels housing the strongly-bound NMP molecules

In the di-NMP solvate crystals, consistent with the weak hydrogen bonding network

of di-NMP solvates (Figure 3–14d), the NMP molecules outside of the tetramers fill

channels along the b-axis with weak interactions among each other (Figure 3–15b) and the

1-D chain of tetramers. These weak interactions correspond to weak overall interaction

energy (Figure 3–15b). The loosely bound NMP molecules in di-NMP solvate structure

form channels along both the crystallographic a and b axes (Figure 3–17), which allows

the weakly interacting NMPs to escape easily. This accounts for the instability of di-NMP

solvate crystals in air, vacuum, and under heat. This distinct interaction strengths of the

tightly bound and loosely bound NMP also explain the observation that the di-NMP solvate

a) b)

c) d)

e)

70

always desolvate to form the mono-NMP solvate instead of CEL under different

conditions, unlike in some other disolvates and dihydrates. 214, 215

Figure 3–17. Packing of di-NMP solvate crystal a) down the b-axis and b) down the a-

axis. In the top row, NMP molecules as part of the framework are colored blue, blue and

red rectangles/circles highlight the channels along the b- and a-axis, respectively; loosely

bound NMP molecules are colored red. In the bottom row, the loosely bound NMP

molecules are deleted to show the channels.

The structure and energy framework of the di-NMP solvate also explains the ease

of transformation from di-NMP to crystalline mono-NMP solvates without sacrificing

crystallinity (Figure 3–9, Figure 3–10), since the escape of the loosely bound NMP from

the channels does not affect the strongly bound 1-D chains of tetramers along the b-axis

(Figure 3–17 and Figure 3–18). Compared to the mono-NMP solvate, the unit cell of the

di-NMP solvate has a significantly decreased 𝛽 angle, an expanded c-axis, and a larger unit

cell volume (Figure 3–18, Table 3–2), as a result of inserting loosely-bound NMP

molecules in between the rigid tetramers to fill the channels.

71

Figure 3–18. Comparison of unit cell parameters and packing for a) CEL mono-NMP

solvate and b) di-NMP solvate (loosely bound NMP molecules removed) down the b-axis

Table 3–2. Crystallographic information of CEL mono-NMP and di-NMP solvates

Mono-NMP

solvate

Di-NMP

solvate*

Di-NMP

solvate*

Temperature 298K 298K 100K

Crystal system Monoclinic Monoclinic Monoclinic

Space group P21/c P21/c P21/c

Unit cell dimensions

a=11.9978(4)Å

b=9.0896(3)Å

c=21.9732(8)Å

α= 90°

β=102.55(5)°

γ = 90°

a=11.8667(7)Å

b=8.5932(7)Å

c=28.8601(22)Å

α=88.049(2)°

β=90°

γ=90°

a=11.8677(5)Å

b=8.3366(3)Å

c=28.1410(13)Å

α=88.937(2)°

β=90°

γ=90°

Volume 2349.36 (14) Å3 2941.24 (4) Å3 2783.69 (2) Å3

Z 4 4 4

Density (calculated) 1.36 g/cm3 1.31 g/cm3 1.38 g/cm3

R 0.0475 0.0738 0.0527

*Note that the unit cell parameters for both 298K and 100K di-NMP solvate structures

were transformed to the non-standard setting of monoclinic P21/c space group for the

ease of crystal structure comparison with the mono-NMP solvate. However, their CIFs

are reported in the standard unit cell settings.

72

3.4.6.1 Spectroscopic Evidence for the Structural Differences Between Two CEL-

NMP Solvates

The different hydrogen bonding scenarios of the two groups of NMPs in CEL di-

NMP solvate are also revealed by the infrared (IR) spectra (Figure 3–19). The IR spectra

of Form III and the two NMP solvates have peaks at 1154, 1234 and 1344 cm-1 in common,

which correspond to stretching of functional groups of characteristic of CEL molecule, i.e.,

O=S=O asymmetric stretch, the CF3 group stretch, and the O=S=O symmetric stretch,

respectively. Compared to the IR spectra of two solvates, the main difference of CEL Form

III is the peak broadening of O=S=O asymmetric stretch at 1154 cm-1. This is attributed to

the existence of both strong and weak hydrogen bonds with the O=S=O oxygen in Form

III as a proton acceptor, 216 while the same oxygen atoms only involve weak hydrogen

bonds in both NMP solvates. The signature peak for NMP molecule in both mono-NMP

and di-NMP solvates is at 1650 cm-1, representing the associated carbonyl (C=O) group of

NMP, which is absent in the CEL form III spectrum (Figure 3–19, box in gray). The O

atom on this C=O group is the acceptor for the strong N-H⋯O hydrogen bond which is

required to form the tetramer. Since this tetramer is present in both mono-NMP and di-

NMP structures but not in Form III, the peak at 1650 cm-1 is present in spectra of the two

solvates but not of Form III. The peak unique to the di-NMP solvate resides at 1679 cm-1

(Figure 3–19). This corresponds to the stretching motion of unbound carbonyl (C=O)

group, in contrast to the 1650 cm-1 peak for the bound carbonyl group. This is consistent

with the fact that, in addition to the strongly hydrogen bonded NMP molecules in the mono-

NMP solvate, weakly hydrogen bonded, channel-filling NMP molecules also exist as

another type of NMP in di-NMP solvate crystals (Figure 3–4, Figure 3–15). Thus, the

73

FTIR spectrum analysis corroborates the analysis of intermolecular interactions based on

crystal structure visualization and energy framework.

Figure 3–19. FTIR spectra of CEL Form III, CEL mono-NMP solvate and di-NMP

solvate crystals

3.5 Conclusions

In this study, two stoichiometric (1:1 and 1:2) NMP solvates of CEL were prepared

and fully characterized. The di-NMP solvate undergoes a complex phase transformation

process when heated, including desolvation, dissolution of the di-NMP solvate, and

growth, dissolution and melting of the mono-NMP solvate. This is explained clearly based

on complementary thermal analytical techniques. The different stability and

interconversion of mono- and di-NMP solvates of CEL were understood based on their

crystal structures. The understanding of very different roles of solvent molecules in the two

NMP solvates of CEL serves as a good example of characterizing pharmaceutical solvates

74

and hydrates using complementary techniques to fully understand their structure-property

relationship, which is important for successful drug development where such solid forms

are pertinent.

75

Chapter 4.

Direct Compression Tablet Formulation of Celecoxib Enabled

by A Pharmaceutical Solvate

This chapter is in preparation as research article.

Wang, K., Sun, C. C.*, Direct Compression Tablet Formulation of Celecoxib Enabled by

a Pharmaceutical Solvate. 2020.

76

4.1 Synopsis

Celecoxib, an anti-inflammatory drug for pain and arthritis, is currently only

available in capsule form. To reduce the onset time for a faster action and to lower the

manufacturing cost, the tablet dosage form is more preferred. However, the commercial

celecoxib (Form III) is not suitable for direct compression (DC) tablet manufacture due to

its poor flow, low bulk density, and tablet lamination. In this work, we overcome these

challenges using a pharmaceutically acceptable dimethyl sulfoxide (DMSO) solvate of

celecoxib. Aided with the DMSO solvate, an acceptable DC tablet formulation was

successfully developed to manufacture tablets containing 200 mg celecoxib, with

satisfactory manufacturability, disintegration, and in vitro dissolution performance.

77

4.2 Introduction

Celecoxib (CEL, Figure 4–1) is a blockbuster NSAID drug for treating arthritis,

available as capsules in the strengths of 100, 200 and 400 mg. The current capsule products

on the market contain the most thermodynamically stable form at room temperature, Form

III of CEL, which is micronized and wet-granulated with excipients to enhance flow for

satisfactory capsule-filling. 4 Compared to capsules, tablets are preferred due to the lower

manufacturing cost and short onset time, which is especially important for pain

medications, such as CEL. However, although a fixed dose combination product of CEL,

Consensi ® (2.5/5/10 mg amlodipine and 200 mg celecoxib) tablets, was approved by the

FDA in 2018, no tablet product of only CEL is available. A tablet formulation containing

wet-granulated CEL granules was patented, but never marketed. 4 For tablet

manufacturing, the wet granulation process is expensive and inefficient compared to the

direct compression (DC) process, which only requires mixing and compression. 8, 189

However, Form III CEL is unfit for the DC process because of its poor powder flowability,

124 low bulk density, 123 poor tablet quality (lamination tendency), and high punch sticking

propensity. 117, 133, 207

Figure 4–1. Molecular structures of celecoxib (MW 381.37) and DMSO (MW 78.13).

N

N

CF3

CH3

SH2N

O

O

Celecoxib

H3CS

CH3

O

DMSO

78

Crystal engineering is an effective technique for overcoming problematic

physicochemical and mechanical properties of active pharmaceutical ingredients (APIs).

87, 93 Common crystal engineering approaches include cocrystallization, 96, 111, 116, 217-220 salt

formation 107 and hydrate formation. 38, 191, 221, 222 A less well explored crystal engineering

approach is the formation of pharmaceutically acceptable solvates. 192 Crixivan (indinavir

sulphate - ethanolate) is one of few examples of marketed products using an API solvate

crystal form. 194

CEL has four known polymorphs (Forms I to IV), 3, 196 an amorphous form, 32

sodium salt hydrates, 125, 197 several solvates, 3, 198 and cocrystals with both common

cocrystal formers 199-204 and drugs 205, 206. In this work, we aimed to develop a DC tablet

formulation of CEL by simultaneously improving the flow, bulk density, and mechanical

properties of CEL using a pharmaceutically accepted dimethyl sulfoxide (DMSO, Figure

4–1) solvate of CEL (CEL-DMSO) for the following reasons: 1) DMSO is a class 3 solvent

categorized by International Council for Harmonisation of Technical Requirements for

Pharmaceuticals for Human Use (ICH) with an exposure limit of 50 mg/day, 223 which is

higher than the expected DMSO intake of 34 mg/day along with 200 mg of CEL; 2) CEL-

DMSO crystalizes in block shape, which is expected to exhibit better flowability, higher

bulk powder density, and easier die filling than the needle shaped CEL; 3) DMSO, being a

polar solvent commonly used to dissolve hydrophobic drugs, may improve the dissolution

rate of CEL; and 4) good physical stability of CEL-DMSO.

79


4.3.1 Materials

CEL (Form III) was purchased from Aarti Drugs Pvt Ltd. (Mumbai, India) and used

as received. DMSO was obtained from Alfa Aesar (Tewksbury, MA, USA). Fused silica

(M-5P, Cab-o-sil; CABOT Corporation, Tuscola, IL, USA), microcrystalline cellulose

(MCC, Avicel PH102, FMC Biopolymer, Philadelphia, PA), mannitol (Pearlitol 100SD,

Roquette, Lestrem, France), crospovidone (Kollidon CL-F, BASF, Ludwigshafen,

Germany), sodium lauryl sulphate (Kollidon SLS, BASF, Ludwigshafen, Germany) and

magnesium stearate (Mg St, Mallinckrodt, St Louis, MO) were obtained from respective

suppliers. Cab-o-sil was de-agglomerated by passing through a #30 mesh sieve before use.

All other materials were used as received. Celebrex® capsules (200 mg CEL strength) were

purchased from Pfizer Pharmaceuticals LLC and used as a reference in dissolution testing.

4.3.2 Methods

4.3.2.1 Single Crystal X-ray Crystallography

Single crystals of CEL-DMSO were obtained from slow evaporation of a saturated

solution of CEL in DMSO at room temperature. Single crystal X-ray diffraction (SCXRD)

was performed on a Bruker D8 Venture diffractometer (Bruker AXS Inc., Madison,

Wisconsin) equipped with a MoKα radiation source (IμS 3.0 micro focus tube) and a

Bruker PHOTON-II CMOS detector. Full sets of crystal structure data were collected and

solved at both 298 K and 100 K. Data integration was performed with the SAINT program.

Scaling and absorption correction was done via the SADABS program, and space group

determination and data merging was done with XPREP. The crystal structure was solved

80

and refined using SHELXT-16 (Intrinsic Phasing), which was executed through ShelXle

graphical user interface. 209 Hydrogen atoms were placed either geometrically from the


The CF3 group of CEL and the DMSO molecule in the 298 (K) CEL-DMSO structure were

disordered. The CF3 group disorder was modeled as two sets of three F atoms with site-

occupancy factors constrained to sum to unity. To ensure a reasonable geometry, the C-F

distances were restrained to a common refined value of 1.33 Å. All F atoms were refined

with anisotropic displacement parameters, where the anisotropic displacement parameters

are modeled to be equivalent in both disordered parts. The disorder of DMSO molecule

was modeled as two sets of DMSO molecules, with site-occupancy factors constrained to

sum to unity. The anisotropic displacement parameters of the bonds in two sets of DMSO

molecules were restrained to be equal.

4.3.2.2 Preparation of Bulk CEL-DMSO Powder

For tablet formulation development, a batch of CEL-DMSO was prepared by a

slurry method, i.e., suspending an excess amount of CEL solid (~ 80 g) in DMSO (~ 30

mL) in a 100 mL beaker by a magnetic stirrer at room temperature. After 4 days, the solids

were collected via vacuum filtration for 10 min in a Buchner funnel. The surface residual

solvent was washed away by briefly spraying deionized water on to the powder in Buchner

funnel during vacuum filtration. The powder was then dried in a vacuum oven with air

flowing on the top of the powder for 3 days. The washing and drying did not cause

detectable phase impurity of CEL-DMSO by powder X-ray diffraction, hot stage polarized

light microscope and differential scanning calorimetry.

81

4.3.2.3 Powder X-ray Diffraction (PXRD)

Powders were scanned over the 2θ range of 5°−35° on a wide-angle X-ray

diffractometer (X’Pert PRO; PANalytical Inc., West Borough, MA) equipped with Cu Kα

radiation (45 kV and 40 mA). A 1/16 º divergence slit and a 1/8 º anti-scattering slit were

used on the incident beam side. The anti-scattering slit on the diffracted beam side was 5.5

mm (X’ Celerator). Data collection was done at a step size of 0.0167° and a dwell time of

1.15 s. The simulated PXRD patterns of CEL-DMSO and CEL (Form III) were calculated

from corresponding single crystal structures solved at 298 K. 216 using the Mercury

software (v4.3.1 Cambridge Crystallographic Database Centre, Cambridge, UK).

4.3.2.4 Thermal Analyses

Thermal behaviors of samples were characterized using Differential Scanning

Calorimetry (DSC, Q1000; TA Instruments, New Castle, DE). Samples were placed in

aluminum pans and then crimped with aluminum lids. Samples were heated from room

temperature to 180 °C, which is below the degradation onset temperature of CEL, 32 at a

rate of 10 °C/min. The cell constant for heat flow was calibrated using indium and the cell

temperature was calibrated with indium and cyclohexane. The DSC cell was purged with

helium gas at 25 mL/min. For CEL-DMSO, the sample for DSC analysis was a small piece

cut from larger single crystals obtained from slow evaporation with surface mother liquor

removed using an absorbent tissue (Kimwipe®). The weight loss upon heating of CEL-

DMSO was obtained using a thermogravimetric analyzer (TGA, Q50, TA Instruments,

New Castle, DE). Samples were placed in an open aluminum pan and heated to 300 ºC at

a rate of 10 ºC /min under 60 mL/min nitrogen purge.

82

4.3.2.5 Hot Stage Microscopy (HSM)

CEL-DMSO crystals were examined using a polarized light microscope (PLM)

(Nikon Eclipse E200, Nikon, Tokyo, Japan) equipped with a temperature-controlled

sample stage (Linkam LTS 420, Linkam Scientific Instruments, Ltd., Waterfield, U.K.) and

a DS-Fi1 microscope digital camera for image-capturing and The crystals were observed

while being heated from room temperature to 180 ºC at a rate of 10 ºC/min without silicone

oil.

4.3.2.6 Nanoindentation of CEL-DMSO Single Crystals

Nanoindentation experiments were performed on a TI-980 Triboindenter (Hysitron

Inc., Minneapolis, MN, USA) system using a standard Berkovich diamond indenter tip. A

fused silica sample with known elastic modulus (E) of 69.9 GPa and hardness (H) of and

9.2 GPa was used to calibrate the area function of the Berkovich tip. 224 The crystal was

fixed on a glass slide using Super glue (Scotch, 3M, Maplewood, MN), which was then

mounted on the sample puck by pulling vacuum. Before each measurement, the crystal

surface was scanned to generate the topographical images of 20 × 20 𝜇𝑚 size. Only

surfaces with an average roughness less than 20 nm were used for indentation. The major

crystal face (002) of two CEL-DMSO crystals was indented under force control mode, with

5 mN/s loading and unloading rate. The indenter was held at a peak load of 5 mN for 20 s.

A total of 15 indents were made to calculate 𝐸 and 𝐻 using the Oliver-Pharr method,

assuming a Poisson’s ratio of 0.3 177. The 20 𝜇𝑚 × 20 𝜇𝑚 area containing the indentation

mark was scanned again after the indentation to gain more information.

83

4.3.2.7 Nanocoating of Microcrystalline Cellulose

The flow properties of MCC can be profoundly improved by silica nanocoating 56

using a conical mill (Model U3, Quadro Engineering Corp., Waterloo, OT, Canada).

Before nanocoating, 0.5 wt % fumed silica was first blended with 50 g MCC with a spatula

before the mixture was passed through a #100 mesh sieve (150 µm opening, US Standard

Sieve). The mixture was then passed through the comill screen with round 0.995 mm

openings (7B039R03125) and a type I impeller (7B16121004, sharp edge forward,

corresponding to more scraping motion during milling) at 2200 rpm. A total of 20 comilling

cycles were done to ensure uniform coating.

4.3.2.8 Powder Flow Property Assessment

The flow properties of powder with comparable particle sizes were characterized

using a shear cell (RST-XS, Dietmar Schulze, Wolfenbüttel, German), with a 10 mL cell

at 3kPa preshear normal stress. The normal stresses for shear testing were 230, 1000, 1500,

2000, and 2500 Pa (230 method). Shear failure stress at each normal stress was used to

construct a yield locus, from which the unconfined yield strength (𝑓𝑐) and major principal

stress ( 𝜎𝑛 ) were obtained by drawing Mohr’s circles. Flowability index ( 𝑓𝑓𝑐 ) was

calculated using Eqn. 4–1. All flow tests were all triplicated, with the relative humidity

during measurements ranged 39 % ~ 42 %.

𝑓𝑓𝑐 =𝜎𝑛

𝑓𝑐 Eqn. 4–1

84

4.3.2.9 Preparation of Tablet Formulations

All formulation components, except MgSt, were first mixed roughly in an amber

glass bottle with spatula, and then using a Turbula mixer (Glen Mills Inc., Clifton, NJ) at

49 rpm for 10 minutes. MgSt, the lubricant, was added only after this for another 5-minute

mixing by Turbula at 49 rpm to avoid over-lubrication.

4.3.2.10 Tabletability Profiling

Tablets of each formulation were made on a compaction simulator (Styl’ One

Evolution, MedelPharm, Beynost, France), simulating a Korsch XL100 press. To attain the

same dose of 200 mg per tablet, tablet weights were 600 mg for CEL-DMSO formulations,

and 500 mg for CEL formulations, made using 11.28 mm and 9.53 mm flat face tooling,

respectively. Different tooling sizes were used to maintain a suitable diameter to thickness

ratio of 2 - 2.32. A dwell time of 30 ms (16,320 tablets/hour) was used to make tablets for

expedited friability, disintegration time, and dissolution tests for formulations containing

CEL-DMSO. For comparing tabletability profiling, flat face round tooling (8 mm diameter)

was used to compress 200 mg tablets using a dwell time of 100 ms, since tablets from the

CEL Form III formulation severely laminated at shorter dwell times.

The tabletabilities of the pure CEL Form III and CEL-DMSO were characterized

using 8 mm flat face round tooling with external lubrication on a Materials Testing

Machine (ZwickRoell 1485, Ulm, Germany). The compaction pressures for tabletability

profiles ranged 12 - 370 MPa.

85

Tablet tensile strength was characterized by breaking the tablets diametrically using

a texture analyzer (TA-XT2i, Texture Technology Corps, NY, equipped with a 30 kg load

cell, trigger force of 5 g, The test speed was 0.01 mm/s. Tablet tensile strength (σ) was

calculated from breaking force, F, tablet diameter, D, and thickness, h using Eqn. 4–2. 225

𝜎 =2𝐹

𝜋𝐷ℎ Eqn. 4–2

4.3.2.11 True Density Determination

Accurate powder true density, ρt, is crucial for accurate analysis of powder

compaction properties. Helium pycnometry is generally used to measure powder ρt for

anhydrous or non-hygroscopic materials. For formulations containing MCC, ρt by helium

pycnometry is overestimated because of the release of water in MCC during measurement.

226 Hence, we obtained the true density using the Sun method by non-linear fitting of Eqn.

4–3 to tablet density (ρ) versus compaction pressure (P) data, Using Origin 2020

(OriginLab Corp, Northampton, MA).

𝑃 =1

𝐶[(1 − 𝜀𝑐) −

𝜌

𝜌𝑡− 𝜀𝑐 ln (

1 −𝜌𝜌𝑡

𝜀𝑐)] Eqn. 4–3

Here, 1/C is a parameter relating to material plasticity, where the lower 1/C value

corresponds to higher material plasticity, 227 and εc is the critical porosity at which the

powder starts to gain rigidity and mechanical strength. 228, 229

86

4.3.2.12 Powder Bulk Density Measurement

Each powder (∼5 mL) was poured in a 10mL graduated cylinder and its weight was

recorded. The bulk density is recorded as the weight divided by volume (g/cm3).

4.3.2.13 Compressibility and Compactibility

Compressibility describes the relationship between tablet porosity (ε) and pressure.

For individual tablets, ε was calculated from ρ and ρt according to Eqn. 4–4. The ρ was

calculated from tablet weight, thickness, and diameter of tablets.

ε = 1 − 𝜌𝑡𝑎𝑏𝑙𝑒𝑡

𝜌𝑡𝑟𝑢𝑒 Eqn. 4–4

Compactibility describes the relationship between 𝜎 and 𝜀, which was fitted with

equation Eqn. 4–5, when possible 230). Here, 𝜎0 is the tensile strength at zero porosity, and

𝑏 is an empirical constant. 𝜎0 can be used to describe the bonding strength of materials 231.

σ = 𝜎0𝑒−𝑏𝜀 Eqn. 4–5

4.3.2.14 In-die Elastic Recovery

In-die elastic recovery of tablets made with compaction simulator is calculated

using Eqn. 4–6, where ℎ and ℎ0 are tablet thicknesses at maximum and zero compaction

pressures, respectively.

ER = 100(ℎ − ℎ0)/ℎ0 Eqn. 4–6

87

4.3.2.15 Intrinsic Dissolution Rate

The intrinsic dissolution rate (IDR) was determined using the rotating disc method.

Powder of CEL or CEL-DMSO were compressed with a 2000-lb force on a Carver Press

(Wabash, IN, USA), in a custom-made stainless-steel die and punch for 2 min to obtain a

pellet (6.39 mm diameter) for dissolution experiments. This eliminates possible particle

size effects that affect powder dissolution. While rotating at 200 rpm, the die and pellet

assembly was submerged into a 200 mL of pH 12 dissolution medium of tribasic sodium

phosphate buffer containing 1% sodium lauryl sulfate 232 at 23.5 ºC in a water-jacketed

beaker. A UV–vis fiber optic probe (Ocean Optics, Dunedin, FL) was used to continuously

monitor the UV absorbance of the solution at λ = 252 nm. The concentration-time profiles

of Celecoxib were obtained from the absorbance data and a previously established

calibration curve. All IDR experiments were triplicated.

4.3.2.16 Tablet Disintegration Time

Three tablets of CEL-DMSO formulation (600 mg) were made at ~7 kN using the

Styl’ One compaction simulator at 30 ms dwell time. Tablet disintegration time (DT) was

determined using a disintegration tester (DJ-1, Tianjin Guoming Medical Equipment CO.,

LTD). Tablets were immersed into a beaker containing 800 mL deionized water, with

temperature maintained at 37°C by a circulating water bath.

4.3.2.17 Expedited Friability Test

Tablet friability profiles were determined using an expedited friability method. 54

Coded tablets prepared under different compaction forces were weighed and loaded into a

friabilator (Model F2, Pharma Alliance Group Inc., Santa Clarita, CA) at 25 rpm for 4 min

88

(corresponding to 100 drops). The final weight of each tablet was noted, and percentage

weight loss of each tablet was plotted against compaction force. The range of compaction

force corresponding to less than 1% tablet weight loss was identified from the friability

plot.

4.3.2.18 Dissolution of CEL Tablets and Capsules

Dissolution performance of formulated CEL-DMSO tablets and commercial CEL

capsules (Celebrex®, 200 mg) was tested in 500 mL of the FDA recommended pH 12

medium of tribasic sodium phosphate buffer with 1% SLS 232 at 37°C in a water-jacketed

beaker controlled by water bath with circulating water. An overhead paddle stirrer rotating

at 50 rpm was used. The concentration of dissolved CEL was continuously monitored by a

UV–vis fiber optic probe (Ocean Optics, Dunedin, FL) at absorbance of λ = 252 nm. All

dissolution experiments were triplicated.


To develop a tablet formulation of CEL with improved manufacturability, the poor

flow, low bulk density and high lamination propensity of CEL (Form III) need to be tackled

effectively. It was achieved in this work through forming a DMSO solvate of CEL. A batch

of CEL-DMSO powder with particle size similar to that of as-received CEL was prepared

for use in formulation (Figure 4–2).

89

Figure 4–2. Particle size of a) CEL-DMSO powder and b) CEL Form III as-received

powder. The scale bar is 50 m in both pictures. Since the shape of CEL Form III and

CEL-DMSO crystals are different, while CEL Form III is in needle shape with high

aspect ratio, it was made sure that the longest dimension of CEL-DMSO and CEL Form

III crystals match.

4.4.1 Physicochemical Properties of CEL-DMSO Solvate

CEL-DMSO is a 1:1 solvate (space group of P21/c) that crystallizes into hexagonal

block shaped crystals (Figure 4–3). When heated, CEL-DMSO solvate is very stable,

which does not undergo desolvation below its melting point. The DSC thermograph

(Figure 4–4a) shows a single sharp endothermic peak (onset temperature of 105.4 ±

0.02°C, heat of fusion of 77.06 ±1.01 J/g, n=3). The endothermic event in DSC were

confirmed by hot stage microscopy studies to be melting of crystals ranged 103 °C - 111°C

(Figure 4–5), matching well with the range of the endothermic peak in DSC thermogram.

b) a)

90

Figure 4–3. a) Block morphology of CEL-DMSO solvate crystals and ORTEP diagrams

of CEL-DMSO single crystal structure at b) 298 K and c) 100 K. The DMSO molecule

and the CF3 group on CEL are disordered only in 298 K structure.

Figure 4–4. a) DSC and b) TGA thermographs of CEL-DMSO solvate crystals. The

dashed lines in b) represents the first derivative of weight signal.

a) b) c)

b) a)

91

Figure 4–5. Polarized light microscope images of CEL-DMSO single crystal as it is

heated using a hotstage

The absence of desolvation before melting suggests a higher stability of CEL-

DMSO than many other solvates and hydrates in the literature, which routinely lose solvent

molecule before melting. 233-236 Thus, CEL-DMSO displays adequate physical stability

important for pharmaceutical tablet formulation development. The high stability of CEL-

DMSO is attributed to the fact that two CEL molecules and two DMSO molecules form a

tetramer (Figure 4–6) via very strong hydrogen bonds (𝑅42(8) & Table 4–1). This tetramer

is the basic building block of the CEL-DMSO crystal structure. Each tetramer is stabilized

by four strong hydrogen bonds, labeled 1, 2, 1’ and 2’, where 1’ and 2’ are symmetrically

equivalent to 1 and 2 in the tetramer (Figure 4–6, Table 4–1). The DMSO molecules

occupy staggered positions in zig-zag channels of the CEL-DMSO crystal lattice,

meanwhile stabilized by strong interactions with CEL molecules (Figure 4–7). These

structural features all lay the foundation of the exceptional stability of this CEL-DMSO

29.9℃ 100℃ 103℃

107℃ 110℃ 111.8℃

92

solvate. The TGA curve shows that the total weight loss did not reach the theoretical

DMSO content in CEL-DMSO crystals (17 wt %) until approximately 230 ºC (Figure 4–

4b), which is about 40 ºC higher than the boiling point of DMSO (189 ºC). Since pure CEL

liquid starts to evaporate at temperatures above 200 ºC (Figure 4–4b), a part of the weight

loss of CEL-DMSO solvate around 230 ºC may be attributed to CEL. This means that

DMSO was not completely removed from the CEL-DMSO melt at 230 ºC. This is possible

if DMSO and CEL form an azeotrope in liquid phase, because of the stability of tetramers

(Figure 4–6). A similar phenomenon was observed in CEL mono-NMP solvate crystals

(Chapter 3).

Figure 4–6. A tetramer synthon 𝑅42(8) comprised of two CEL and two DMSO molecules

Table 4–1. Hydrogen bonds that stabilize the tetramer in the CEL-DMSO structure

N-H⋯O bond name 1 & 1’ 2 & 2’

d (H⋯A) (Å) 2.071 2.043

d (D⋯A) (Å) 2.931 2.883

∠(DHA) (º) 163.34 162.36

1

1’

2

2’

93

Figure 4–7. Packing of CEL-DMSO with removed DMSO molecule to highlight the void

space. Hydrogen atoms were removed for simplicity and clarity. a) Viewed down the b-

axis, the spiral is barely look-through. b) viewed down the a-axis, the spiral is obvious

and the DMSO occupies staggered positions in the lattice

The high solid-state stability of CEL-DMSO is important for removing the surface

residual DMSO without causing desolvation. Phase pure CEL-DMSO powder was

obtained by spraying deionized water onto the CEL-DMSO powder during vacuum

filtration and then placing the powder in a vacuum oven at room temperature and dried for

3 days with air flowing over the powder. The CEL-DMSO powder was also stable upon

storage at ambient conditions for at least 9 months in a closed container. Additionally, the

CEL-DMSO was stable against high compaction pressure since the XRD pattern of a tablet

(200 mg) compressed at 325 MPa showed no sign of phase conversion when compared to

the simulated pattern from room temperature single crystal structure (Figure 4–8).

Therefore, no phase stability related issues are expected during API preparation and tablet

manufacturing. It is also useful to point out that the DMSO molecules and -CF3 group of

CEL in the crystal are disordered over two sites at 298 K, but disorders are absent at 100K

(Figure 4–3). Thus, the two disordered sites correspond to slightly different interaction

energies, with the energy difference comparable in magnitude to the kinetic energy

difference of the molecules between the two temperatures. At 298 K, the disordered DMSO

b) a)

94

molecules occupy two positions with 0.84 and 0.16 occupancies with only subtle changes

to the position of O atom that forms the tetramer with N-H on sulfonamide of CEL. The

disorder at 298 K did not affect the crystal packing much.

Figure 4–8. Powder X-ray patterns of CEL-DMSO solids and tablets. The simulated

pattern was used as a reference, which was obtained from the room temperature crystal

structure.

CEL-DMSO exhibited a higher intrinsic dissolution rate than CEL in deionized

water at 25 °C (Figure 4–9). However, it is important to point out that conversion of CEL-

DMSO to CEL Form III occurred quickly (in 10 – 30 s) during IDR measurements.

Therefore, washing CEL-DMSO with water. To remove residual DMSO requires careful

monitoring and control during large scale manufacturing.

95

Figure 4–9. IDR data showing CEL-DMSO tests and CEL Form III tests at 25 °C. Three

CEL-DMSO curves showed different extents of supersaturation followed by rapid

precipitation and phase conversion to Form III.

4.4.2 Mechanical Properties of CEL-DMSO

The elastic modulus (E) and hardness (H) are important parameters that could

potentially be used to predict tableting performance of APIs. 90, 237, 238 CEL Form III crystal

is exceptionally elastic, 216 with an E of 16.27 ± 0.43 GPa and an H of 0.45 ± 0.02 MPa

when the (001) crystal face was indented. Therefore, CEL falls in the category of stiff

crystals based on E and intermediate-stiffness based on H. 239 This is consistent with the

poor tabletability of CEL since highly elastic materials are generally poorly compressible

and tends to form laminated tablets due to high elastic recovery in die. 238 Compared to

CEL, the elasticity of CEL-DMSO is greatly reduced, with an E of to be 5.574 ± 0.31 GPa

when indented on the major (002) crystal face. This places CEL-DMSO in the category of

compliant molecular crystals. 239 The H of the (002) face of CEL-DMSO (0.29 ± 0.02

96

MPa) is also much lower than CEL. However, the CEL-DMSO crystal fractured during the

nanoindentation test (Figure 4–10), indicating higher brittleness of CEL-DMSO than CEL.

This also means that actual H of CEL-DMSO is likely higher had the fracture not taken

place.

Figure 4–10. Surface scans of indentation imprint after nanoindentation tests for a) CEL

Form III on (001) face and b) CEL-DMSO on (002) face. The hairline in b) is a crack in

the crystal.

The tabletability plots of CEL-DMSO powder is significantly lower than CEL

(Figure 4–11), which may partially be attributed to the brittleness of CEL-DMSO crystals.

For example, more brittle materials tend to develop defects in the tablets during

decompression, which generally leads to poorer tabletability. Hence, the inclusion of a

plastic excipient in the formulation is expected to be critical for tablet formulation of CEL-

DMSO to avoid compaction related problems. In this regard, MCC is a good candidate

because of its high plasticity and excellent tabletability. 240

b) a)

97

Figure 4–11. a) Tabletability profiles and b) flowability of CEL-DMSO solvate, CEL

Form III and their Formulations D.

4.4.3 Flowability of CEL-DMSO

Although the longest dimensions were similar (Figure 4–2), the flowability index

(ffc) of CEL-DMSO powder (7.1 ± 0.2) was almost two times that of the ffc of CEL powder

(4.0 ± 0.3) (Figure 4–11b). The bulk density was also higher than that of CEL Form III

b)

a)

98

(0.64 vs. 0.11 g/cm3). The improved flowability of CEL-DMSO as attributed to its more

equi-dimensional shape, which is more favorable to powder flow than the needle like CEL

crystals (Figure 4–2, Figure 4–3). The CEL-DMSO crystals with a lower aspect ratio

alleviates the detrimental effect of particle entanglement among needle-shaped crystals on

powder flowability. Despite the improvement, the flowability of CEL-DMSO is still poorer

than that of Avicel PH 102 (ffc = 12.2 ± 0.8), which exhibits flowability minimally required

for high speed tableting. 241 However, the improvement in flowability by CEL-DMSO is

sufficient to enable the development of a DC tablet formulation with the aid of functional

excipients.

4.4.4 Development of A Direct Compression Tablet Formulation Using CEL-

DMSO Solvate

Based on the flowability and tabletability of CEL-DMSO, we designed the first DC

tablet formulation (Formulation A) containing 50% CEL-DMSO (Table 4–2). This

formulation corresponds to 200 mg CEL when the total tablet weight is 500 mg.

Formulation A also has 23% MCC as the plastic binder to enhance tablet strength; mannitol

Pearlitol 100SD was used to enhance the flow of the formulation; 55 crospovidone was used

as a disintegrant; 242 SLS was added as a dual functionality wetting agent and lubricant 243

and Mg St was used as the internal lubricant. 244 Formulation A (ffc = 14.1 ± 1.6) exhibited

significantly better flowability than both CEL-DMSO alone and Avicel PH102. When

CEL-DMSO in Formulation A was replaced by CEL, the flowability profoundly

deteriorated (ffc = 3.7 ± 0.1), to be similar to CEL alone. Thus, at 50% CEL loading, the

flowability of the formulation is dominated by CEL. At this drug loading, the use of a better

99

flowing crystal form, CEL-DMSO in this case, leads to profound improvement in the

flowability of the formulation. Additionally, CEL containing Formulation A could not be

made into intact tablets at low compaction pressures at a 20 ms dwell time due to severe

lamination (Figure 4–12). When CEL-DMSO was used in Formulation A, all tablets were

intact below 300 MPa but very mild lamination occurred at higher pressures where hairline

cracks on the side of tablets were observed. Since faster compression aggravates tablet

capping and lamination, 245, 246 a slower speed (corresponding to a dwell time of 100 ms)

was used to make intact tablets for delineating their tabletabilities. At this speed, the

tabletabilities of the two formulations were comparable, sufficiently strong tablets ( >

2MPa tensile strength) could be made for both formulations (Figure 4–13a). However, the

ejection force of CEL-DMSO Formulation A ~ 550 N at 150 MPa compaction pressure,

which is higher than the preferred 400 N. 247 Thus, Formulation A requires optimization to

be more suitable for commercial manufacturing.

Table 4–2. Summary of tablet formulation compositions and weights. The total amount

of CEL in each tablet is 200 mg.

Components (%) Formulation A Formulation B Formulation C Formulation D

API (CEL or CEL-DMSO) 50 40 40 40

Avicel PH102 23 41.5 - -

Mannitol Pearlitol 100SD 20 11 - -

Nanocoated Avicel PH102 - - 52.5 52.25

Crospovidone 4.75 4.75 4.75 4.75

SLS 2 2 2 2

MgSt 0.25 0.75 0.75 1

Tablet weight (CEL) 400 mg 500 mg 500 mg 500 mg

Tablet weight (CEL-DMSO) 500 mg 600 mg 600 mg 600 mg

100

Figure 4–12. Severe lamination of tablets made from CEL Form III Formulation A.

Figure 4–13. a) The tabletability and b) the ejection force profiles of CEL-DMSO and

CEL Form III formulations who were formulated as listed in Formulation A.

To minimize possible damages of high ejection force to tablets and dies during

large scale commercial manufacturing, two changes were made to arrive at Formulation B

(Table 4–2), i.e. 1) increasing the amount of MgSt from 0.25 wt % to 0.75 wt % to reduce

the ejection force seen in Formulation A; and 2) using more of the plastic MCC to replace

some brittle mannitol to mitigate the lamination issue at high compaction pressures. Since

CEL-DMSO is also brittle (Figure 4–10), the loading of active crystals was also reduced

b) a)

101

from 50 wt % to 40 wt %. This necessitates a larger tablet (600 mg) to deliver 200 mg of

CEL equivalent dose (Table 4–2).

The ffc of CEL-DMSO containing Formulation B is 14.0 ± 0.01, indicating its

satisfactory flowability for high speed tableting. However, very light hairline cracks on the

side of tablets were still observed only at high compaction pressures. Therefore, the

formulation composition changes did alleviate the observed lamination problem, but

further formulation optimization is needed to fully eliminate tablet defects. In Formulation

C (Table 4–2), all mannitol was removed to further reduce brittleness of the formulation.

However, the elimination of freely flowing mannitol also likely deteriorates the flowability

of the formulation. Hence, silica nanocoated MCC was used to ensure sufficient flowability

for high speed tablet manufacturing. 56 In fact, the ffc of CEL-DMSO containing

Formulation C (22.0 ± 2.7) indicates excellent flowability. Importantly, defect-free tablets

were made in the entire pressure range at 20 ms dwell time. However, the ejection force,

ranging from 556 to 725 N, is still slightly too high. To address this problem, we increased

the amount of MgSt to 1 wt % in Formulation D (Table 4–2).

As expected, the ejection force profile of Formulation D is much improved

compared to Formulation C, with ejection force lower than 400 N in the entire pressure

range (Figure 4–14). Thus, the increased level of MgSt is effective in addressing the high

ejection force. We then assess Formulation D against other important aspects of tablet

manufacturing. As expected from the minor change in formulation, the ffc of CEL-DMSO

containing Formulation D (19.0 ± 1.70), is comparable to that of Formulation C. Its

flowability is excellent. The flowability is significantly higher than that of CEL-containing

102

Formulation D (4.5 ± 0.1). Clearly, the better flowability of CEL-DMSO is translated into

better flowability of the DC tablet formulation (Figure 4–11b). The CEL-DMSO

containing Formulation D also exhibited adequate tabletability, although it is slightly lower

than that of the CEL-containing Formulation D (Figure 4–11a). Similar to flowability, the

different tabletability of the two solid forms is translated into different tabletability of

Formulation D. This is attributed to the smaller bonding strength between CEL-DMSO

formulation than the CEL Form III formulation (Figure 4–15). With the same excipient

matrix composition, only the solid form of CEL is different in the two formulations - CEL

Form III or CEL-DMSO. Thus, the difference in compaction properties of these

formulations are attributed to the properties of different solid forms. The same

compressibility of both formulations (Figure 4–15a) indicates the bonding area of both

formulations are similar. Whereas the higher 0 of CEL-DMSO formulation D (3.68 MPa)

than the 0 of CEL Form III Formulation D (2.99 MPa) indicates that CEL Form III

formulations has higher bonding strength (Figure 4–15b). The higher bonding strength

overrides the tableting performance of CEL Form III formulation through the bonding area-

bonding strength interplay, thus causing the CEL Form III Formulation D to have slightly

higher tabletability than CEL-DMSO Formulation D.

So far, Formulation D exhibits excellent flowability, adequate tabletability (with

tablets free from lamination), and acceptable ejection force. The much lower bulk density

of the CEL-based Formulation D (0.20 g/cm3), compared to the CEL-DMSO formulation

(0.48 g/cm3), led to the difficulty of packing 500 mg of powder into the die even at the

maximum filling depth (Figure 4–16). The low bulk density along with the poor

103

flowability (ffc = 3.7 ± 0.1) made it impossible to prepare tablets with the desired tablet

weight using the CEL-containing Formulation D. When 500 mg powder was force-filled

into the die, aided by pressing the powder bed with hand to consolidate the powder, no

intact tablets could be made at 20 ms dwell time due to severe lamination upon ejection

(Figure 4–17 a & b, where pictures of a tablet made at 7 kN using CEL formulation D has

exhibited severe lamination). In contrast, intact tablets could be made from the CEL-

DMSO based Formulation D under identical compression conditions (Figure 4–17c).

Overall, the CEL-DMSO based Formulation D exhibits adequate manufacturability.

Hence, it is further examined for friability and dissolution to confirm its suitability for

developing a DC tablet formulation of CEL.

Figure 4–14. Ejection force of Formulations D for both CEL Form III and CEL-DMSO

104

Figure 4–15. a) Compressibility and b) compactibility of CEL-DMSO and CEL Form III

Formulations D.

Figure 4–16. 500 mg of CEL Form III-containing Formulation D powder cannot be

packed into the die, even if it is at the maximum depth of filling. The powder had to be

pushed into the die by hand for compaction into tablets.

Figure 4–17. a) Lamination upon ejection of CEL Form III Formulation D tablets made

at 7kN. The inset is a close-up of the middle piece of the broken tablet, which is further

b) a)

b) a) c)

105

laminated into layers. b) Another laminated tablet of CEL Form III Formulation D made

at 7 kN. c) Intact tablet of CEL-DMSO Formulation D made at 7 kN.

4.4.5 Expedited Friability of CEL-DMSO Formulation D Tablets

Friability is critical for tablet quality in terms of general handling and shipping, and

it determines suitable packaging of the final product. For example, individual blister

packaging may be needed for more friable tablets to protect their integrity, while bulk

tablets can be put into bottles if they are not friable. Typically, tablet friability should be

<1%. 248 However, tablet friability of a given formulation depends on the mechanical

strength of the tablets, which is in turn determined by tableting speed, compression force,

tooling design, and tablet size. A suitable set of compression conditions can be determined

from a friability profile, which can be determined using a material-sparing and expedited

method. 54 In this work, the expedited friability test of CEL-DMSO Formulation D tablets

compressed at 20 ms dwell time using flat-faced round tooling (Figure 4–18a) suggests

that tablets with < 1% friability can be obtained in the compression force range of 4.6 –

21.7 kN. Below 4.6 kN, the tablets are friable because of their low mechanical strength.

Above 21.7 kN, tablets are friable likely because of defects introduced into tablets due to

higher extent of elastic recovery. This is supported by the coincidence of both rising tablet

friability (Figure 4–18a) and the deflection of the in-die elastic recovery profile at 10 kN

compression force (Figure 4–18b). The link between elastic recovery and friability is

further supported by the effects of tableting speed on friability and elastic recovery. At a

slower speed, corresponding to a dwell time of 30 ms, both friability and elastic recovery

are overall lower than those at higher speed. (Figure 4–18, open triangles). This is sensible

106

because longer dwell time allows particles to undergo larger extent of plastic deformation,

which both dissipates more elastic energy and increases tablet strength. 53

Figure 4–18. a) Friability profiles and b) in-die elastic recovery profiles for the CEL-

DMSO Formulation D at 20 ms (solid squares) and 30 ms (open triangles) dwell times.

Trend lines are manually added to guide the eye.

With the 30 ms dwell time, the friability and in-die elastic recovery at 7.5 kN is

predicted from corresponding profiles to be ~ 0.5% and ~4.6%, respectively (Figure 4–

18a). Three tablets made at 7.4 ± 0.1 kN, exhibited a friability of 0.47 % ± 0.09 % and

an elastic recovery of 4.5 % ± 0.2 %, supporting the validity of these profiles.

4.4.6 Disintegration and Dissolution of CEL-DMSO Formulation D Tablets

Three tablets of the CEL-DMSO Formulation D compressed at 7.70 ± 0.05 kN all

disintegrated in less than 15 seconds in deionized water at 37 °C. The fast disintegration of

these tablets is advantageous for CEL to exhibit faster onset of action.

Compared to the commercial 200 mg Celebrex capsules, the release of CEL from

the CEL-DMSO Formulation D tablets is much faster, largely because of an absence of the

b) a)

107

lag time, corresponding to the very fast tablet disintegration process (Figure 4–19).

Although the Celebrex capsules have a ~2 min lag time before dissolution initiated, it

exhibited a relatively fast dissolution rate, likely because of the smaller particle size of the

micronized CEL Form III used in the capsules 4. If needed, the dissolution performance of

CEL-DMSO tablet can likely be further improved by using a suitable precipitation inhibitor

to slow down or even prevent the crystallization of CEL during the course of dissolution

27, 126, 249.

Figure 4–19. Tablet dissolution tests of CEL-DMSO Formulation D vs. marketed 200

mg CEL capsules.

Aided with CEL-DMSO, we have successfully developed a DC tablet formulation

of CEL that can be manufactured on a high speed press with improved dissolution

performance compared to marketed capsules. The efficient development of this tablet

formulation was facilitated by predictive tools and material-sparing techniques for

108

characterizing formulations, as repeatedly shown before. 247, 249-251 The formulation

optimization employed a materials science based decision making process, which

invariably starts with clear understanding of deficiencies of a prototype formulation, and

is followed with changes in formulation to specifically address such deficiencies. This

successful example highlights the largely neglected applications of pharmaceutically

acceptable solvates in tablet product development. With pharmaceutical solvates added to

the toolbox, one has more flexibility in solving formulation problems.

4.5 Conclusion

We have developed a direct compression tablet formulation of CEL using DMSO

solvate of CEL, which exhibits improved flowability, manufacturability, tablet quality, and

faster dissolution over those of CEL. The formulation development process was efficient

because of the employment of predictive techniques for assessing powder flowability and

tableting performance. This study is another example that shows the benefits of combining

crystal engineering and materials science to enable the development of high quality tablet

products efficiently.

109

Chapter 5.

Crystal Growth of Celecoxib from Amorphous State -

Polymorphism, Growth Mechanism, and Kinetics

This chapter has been published as a research article in Crystal Growth and Design.

Wang, K.; Sun, C. C.*, Crystal Growth of Celecoxib from Amorphous State:

Polymorphism, Growth Mechanism, and Kinetics. Cryst. Growth Des. 2019, 19 (6),

3592-3600.

110

5.1 Synopsis

The crystal growth kinetics of amorphous celecoxib (CEL), an anti-inflammatory

BCS class II drug, was systematically investigated for developing effective stabilization

strategies to enhance solubility of CEL. Compared to bulk, the surface crystal growth was

much faster, e.g., ~80 fold higher at Tg + 3 °C. Both surface and bulk growth near and

below Tg was accelerated over that predicted from diffusion-limited process. Similar to

some other organic small molecule glasses, the faster surface and bulk growth near and

below Tg are attributed to solid-state crystal growth and Glass-to-Crystal growth

mechanisms, respectively. These two solid-state growth modes were disrupted by fluidity

differently upon heating, with a much wider transition zone of CEL on the surface

compared to the bulk. Interestingly, the surface transition zone is the widest among all

systems studied so far. The phenomenon of cross-nucleation was also observed between

CEL polymorphs, forms I and III, which also exhibited different crystal growth rates in the

diffusion-controlled region both on surface and in bulk.

111

Figure 5–1. Synopsis figure. Crystallization of celecoxib from amorphous phase exhibits

the phenomena of fast surface growth, fast bulk glass-to-crystal growth, and cross-

nucleation between two polymorphs. The fast growth mechanisms are disrupted by the

onset of fluidity controlled growth at temperatures above Tg.

20 40 60 80 100 120 140 160

1E-12

1E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3

Surface Form III

Bulk Form III

Surface Form I

Bulk Form I

Gro

wth

rat

e (m

/s)

Temperature (°C)

Tt=41.4°C for

surface (III)

Tt=44.8°C for

bulk (III)

Tg=51.8 °C

Tm=160.8 °C

112

5.2 Introduction

There is an increasing percentage of poorly soluble compounds, i.e.,

biopharmaceutical classification system (BCS) classes II and IV compounds, under

development in the pharmaceutical industry. 28, 252 For example, approximately 39 % of

698 oral immediate release drugs on the market and 60 % of 28,912 new chemical entities

under development fall in the BCS classes II and IV. 28 Consequently, there has been a

rising demand for enabling techniques to improve solubility and/or dissolution of drugs.

253 Common ways to improve solubility and/or dissolution include 1) adoption of a more

soluble solid form of a compound, such as salts, 107 co-crystals 254 and amorphous solid;

150 2) particle size reduction to increase surface area; 255 3) use of cosolvents, 256 surfactants,

lipids, or complexing agents. 257 Among these methods, the use of amorphous solids has

gained much interest as a universal solubility enhancing technique, owing to their

inherently higher thermodynamic activity than corresponding crystalline counterparts. 258,

259 However, an amorphous drug tends to convert to its thermodynamically more stable

crystalline form during storage, which negates its potential solubility advantages.260 Thus,

mechanistic understanding and control of the crystallization behavior of amorphous drugs

are of practical importance. 261-264

It has been shown that crystal growth from the amorphous state can still be

surprisingly fast at the storage condition, which is well below its glass transition

temperature, Tg, due to two mechanisms: 1) a solid-state glass-to-crystal (GC) growth in

bulk that exceeds the rate of bulk diffusion of molecules; 265-267 2) fast surface crystal

growth due to the considerably higher molecular mobility on the surface of an organic glass

113

than in the bulk. 268-271 Among the 15 organic small molecules exhibiting bulk GC growth

so far, 268, 270-274 there are three common features : 1) an abrupt increase in the crystal

growth rate near Tg, compared to that extrapolated from growth rate data above Tg

(attributed to diffusion-controlled growth mechanism); 272 2) temperature dependent

growth morphology of crystals; and 3), lower activation energy (Ea) of GC growth than

that for diffusion-controlled growth. 275 Several mechanisms have been suggested to

explain GC growth, including the homogeneous nucleation-based crystallization model ;

276 the tension-induced interfacial mobility; 277 and solid-state growth by local fracture and

surface mobility. 278 However, none of these theories could completely explain all features

of GC growth. 272 The fast surface molecular diffusion was established by following the

surface grating decay of organic glasses. 279, 280 It explains the phenomenon of crystals

growing hundreds of nanometers above the surface.281 This mechanism was further

indirectly verified by showing arrested surface crystal growth upon nanocoating the glass

surface. 282, 283

In this study, we used Celecoxib (CEL, Figure 1–4) as a model compound to

further probe the prevailing crystallization mechanisms of a pharmaceutical organic glass,

both on surface and in bulk. CEL is a widely prescribed nonsteroidal anti-inflammatory

drug (NSAID) for treating pain and inflammation. 172 It is a BCS class II compound,

exhibiting low solubility and high permeability. 284 The bioavailability of CEL is expected

to be improved through solubility enhancement by the use of its amorphous form. 285 A

highly stable CEL glass has been prepared before by vapor deposition and its stability was

studied at a single temperature. 286 We found that CEL also exhibits fast GC growth in the

bulk and fast crystal growth on the surface, near and below its Tg. We also observed that

114

both mechanisms are disrupted by the onset of fluidity as previously observed in other

model systems. 270 The intriguing phenomenon of abrupt increase in growth rate near Tg of

amorphous compounds as temperature decreases has been attributed to the activation of

diffusionless solid-state growth mechanisms. A unique feature observed in CEL is that it

exhibits presently the widest transition zone between solid-state growth and diffusion-

controlled growth both on surface and in bulk. This makes CEL a good model system for

future studies to further elucidate the molecular mechanism underlying this phenomenon.

In addition, we have also observed different growth rates and activation energies of two

CEL polymorphs in the diffusion-controlled region, which indicates that crystal structures

play a role in crystal growth kinetics. A clear understanding of the kinetics of various

crystal growth mechanism is useful to developing effective stabilization strategies of

amorphous drugs, such as surface elimination by nanocoating.282, 283


5.3.1 Materials

Form III CEL was purchased from Aarti Drugs Pvt Ltd. (Mumbai, India) and used

as received.

5.3.2 Methods

5.3.2.1 Preparation of Amorphous CEL Thin Films

Amorphous CEL thin films were prepared by melt-quenching. Approximately 3~5

mg CEL crystals were sandwiched between a glass slide and a rectangular cover glass

orthogonally placed, keeping the overlapping area constant (24mm × 24mm). The

115

sandwiched sample was then melted on a hot stage (Linkam LTS 420, Linkam Scientific

Instruments, Ltd., Waterfield, UK) at 180 °C and held for 5 min until the molten CEL

spread and cover the entire overlapped area. When bulk crystal growth rate was measured,

the cover glass was left in place. When surface crystal growth rate was measured, the cover

glass was peeled off to expose the free surface after the sample had been quenched to room

temperature (Figure 5–2). The thickness of films was estimated to be 3.7 ~ 6.2 μm, based

on the total surface area, sample weight, and true density of amorphous CEL (Table 5–1)

measured using a helium pycnometer (Ultrapyc 1200e, Quantachrome Instruments,

Boynton Beach, Florida). This film thickness was used to avoid the known phenomenon

of subsurface nucleation in supercooled liquid, which is about 48.3 ± 7.4 𝜇 m for

acetaminophen. 287

Figure 5–2. Sample configuration of surface and bulk amorphous thin films (top view)

Table 5–1. Physical properties of Form III, Form I and amorphous CEL

CEL solid form Melting point or Tg (°C) Density (g/mL)

Form III 160.80 1.549 288

Form I 163.45 -

Amorphous 51.80 1.392

Bulk Free Surface

Melt

Glass slide

Cover glass

Melt

116

5.3.2.2 Crystal Growth Rate Measurements

Amorphous CEL samples were kept in Drierite® containing desiccators, which

were placed in ovens at temperatures of interest. They were periodically withdrawn and

observed under a polarized light microscope (PLM) (Nikon Eclipse E200, Nikon, Tokyo,

Japan) equipped with a DS-Fi1 microscope digital camera for capturing digital images to

enable crystallite size measurements. Crystal growth in the 100 – 160 ℃ temperature range

was measured from video clips captured during hot stage microscopy experiments because

it was too fast to reliably measure by taking still images. The size of a crystal that grew as

compact spherulites/polycrystalline circles was evaluated by measuring the orthogonal

diameters of the crystal circle and taking the average. The average radii of the crystal

spherulites were then plotted as a function of time (Figure 5–3). For crystals that grew as

single crystals, the growth rate was taken as the linearly advancing speed of the crystals’

growing front. Each crystal growth rate was calculated by tracking the growth of 3 to 39

different crystals. More crystals were tracked at lower temperatures when parallel samples

could be set up.

117

Figure 5–3. a) Growth of bulk Form I crystals at 70 °C over 6 days b) Linear regression

on increasing crystal radii over time.

5.3.2.3 Arrhenius Analysis of Crystal Growth Rate

Growth rate – temperature data was analyzed using the Arrhenius Eqn. 5–1, where

A is a pre-exponential factor, 𝐸𝑎 is the activation energy of the crystal growth (𝑘𝐽/𝑚𝑜𝑙), 𝑇

is the absolute temperature (K), and 𝑅 is the universal gas constant ( 8.314 ×

10−3 𝑘𝐽 𝑚𝑜𝑙−1𝐾−1). The slope of a line on a ln(k) vs. 1/T plot yields −𝐸𝑎 𝑅⁄ .

5.3.2.4 Preparation of Polymorphic Seeds of CEL

Form III CEL seeds were prepared by storing melt-quenched amorphous samples

in a 40 ºC oven for 1~2 days. Form I CEL seeds were prepared by placing the melt-

quenched samples in an 80 ºC oven overnight. Samples containing appropriate seeds were

then stored at temperatures of interest to determine crystal growth rates.

𝑘 = 𝐴 ∗ 𝑒−𝐸𝑎 (𝑅𝑇)⁄ Eqn. 5–1

118

5.3.2.5 Raman Spectroscopy

Raman spectra of the amorphous and crystallized CEL were collected using a

confocal Raman microscope (Witec alpha 300R, WITec, Ulm, Germany), equipped with a

CCD detector for identifying solid forms of CEL samples. Surface and bulk thin film

samples were excited using a 532 nm Omnichrome Argon ion laser. A 100× lens was used,

providing a laser beam size of 2 μm in diameter. For each point tested, the spectral

acquisition integration time was 10 s and the average of two spectrum accumulations was

obtained. Both line scan and area scan were performed at the interface of two CEL

polymorphs, using 5 s and 3 s integration time per point, respectively. For line scan, the

scan length was 100 μm with 1 μm resolution (100 points total). For the area image scan,

a 100μm × 20 μm strip was scanned with 1 μm resolution. The intensities of the non-

overlapping peak ranging from 1113 to 1143 cm-1 representing Form I was measured and

is plotted against the relative positions on the sample. In the resulting image of the 2D scan,

higher brightness represents higher intensities of this peak.

5.3.2.6 Micro X-ray Diffraction

A wide-angle Bruker-AXS Micro diffractometer (Bruker-AXS, Madison,

Wisconsin) with a 2.2 kW sealed Cu X-ray source was utilized to collect X-ray diffraction

patterns of crystals grown from melts. A 0.8 mm size collimator was selected for a better

resolution. Data were collected using an area detector at room temperature over 5 - 35° 2θ

range with a total exposure time of 300 s. The glass slide containing crystals of interest

grown from melt was glued to the vertical sample holder. The holder can be moved to allow

the X-ray beam to focus on the crystal spherulites of interest.

119

5.3.2.7 Thermal Analysis

Melting point was determined using a Differential Scanning Calorimeter (DSC,

Q1000; TA Instruments, New Castle, DE) at a heating rate of 10 °C/min. Prior to DSC

study, the degradation temperature of CEL was determined by thermogravimetric analysis

(TGA) (Q50, TA Instruments, New Castle, DE). Maximum temperature used in DSC

experiments was kept below the corresponding degradation temperature. The samples for

DSC were prepared by scratching off the crystals of interest, which are grown from

amorphous CEL.

5.3.2.8 Scanning Electron Microscopy (SEM)

The surface crystals grown from melt at 110 ºC on a glass slide were cut by a

tungsten carbide point scriber/etching pen to obtain a small piece of glass containing a

complete crystallite to be analyzed. The cut piece of glass sample with surface crystal on

top was glued to a carbon tape on the bottom, and then adhered to a SEM stub. The sample

surface was coated with platinum coating of approximately 75Å thickness using an ion-

beams sputter (IBS/TM200S; VCR Group Inc., San Clemente, CA). The surface textures

were analyzed using a JEOL 6500F scanning electron microscope (JEOL Ltd., Tokyo,

Japan), operated at SEI mode with an accelerated voltage of 5kV.


5.4.1 Differential Surface and Bulk Crystal Growth Rates

The growth rate profile of CEL is summarized in Figure 5–4, where Tg and Tm

represent glass transition temperature of amorphous CEL (51.8 °C) and melting

120

temperature of form III CEL (160.8 °C), respectively. Under PLM, the crystals are

birefringent and the areas surrounding the crystals are amorphous (Figure 5–5a-b). Both

CEL Forms I and III were observed during the course of this study. Form III bulk growth

rate data between 60 °C and 100 °C could not be collected because only Form I grew in

bulk samples in this temperature range, even when seeded with Form III. The missing data

points of surface growth rate for either Form III or Form I crystals at ≥ 120 °C was due to

the difficulty in capturing fast crystal growth into a flowing liquid.

Figure 5–4. Surface and bulk crystal growth rate - temperature profiles of CEL a) Form

III CEL. b) Form I. Open symbols indicate growth mechanisms distinct from the

diffusion controlled growth (solid symbols).

a) b)

121

Figure 5–5. Crystallization of CEL from amorphous phase observed under PLM and

Raman Microscope, a) Form III crystals grew on surface at 40 °C, and b) Form I crystal

grew in the bulk at 80 °C, c) The two polymorphs under Raman microscope, d) Raman

area scan of the boundary between Form I (orange red) and Form III (black).

From the data obtained, the surface growth rates of Form III were 1~2 orders of

magnitude higher than those in the bulk below and near Tg of CEL (Tg =51.8 °C) (Figure

5–4a) and the enhancement effect appears to increase with decreasing temperature. This is

similar to that observed in several other small organic molecules, where the faster surface

crystal growth rate is sustained by the orders of magnitude faster molecular diffusion on

the surface than in the bulk, 261-263, 279 as suggested by the strong positive correlation

between the surface crystal growth rate and surface diffusion coefficients of o-terphenyl

(OTP), nifedipine (NIF) and indomethacin (IMC).279, 281 This phenomenon is also

responsible for the faster growth along crevices and voids in an amorphous sample.268 Also

Crystalline

Amorphous Crystalline

Amorphous a) b)

c)

Form I Form III

d)

20 m

122

similar to other molecules, 262, 268, 289 the relative difference between the surface and bulk

growth rates of CEL decreases with increasing temperature, with a convergence

temperature of about 100 °C (Figure 5–4a and b). At and above the convergence

temperature, the differences between the surface and bulk growth rates cease to exist

because wetting and embedding of surface crystals by surrounding liquid. 289 This was

confirmed for CEL Form III surface crystals at 110 °C (Figure 5–6).

Figure 5–6. Liquid-embedded crystal in the diffusion-controlled crystal growth region of

CEL Form III, taken from 110 ºC sample. Crystal growth direction is indicated with an

arrow. a) Top view; b) tilted view in SEM.

5.4.2 Temperature-dependent Crystal Growth Mechanisms

With decreasing temperature, CEL Form III crystal growth rate abruptly increased

in the 25 – 60 °C range (Tg is 51.8 C for CEL) for both surface and bulk growth (Figure

5–4a). Similar observations were made in several other organic materials. 261, 268, 271, 272, 290

The abrupt increase in bulk growth rate was attributed to the GC mechanism, 272, 275 while

the abrupt increase in the surface growth rate was related to the activation of a solid-state

crystal growth mechanism that differs from fluidity-limited growth at higher temperatures.

a) b)

123

279, 280 Both mechanisms are disrupted by the onset of fluidity when the temperature is

sufficiently high for the diffusion-controlled process to play a more significant role in

crystal growth. The vertex of the transition region, termination temperature (Tt), marks the

temperature above which the GC growth mode or the solid-state growth mode is disrupted

by onset of fluidity-based crystal growth. 270 The temperature range between Tt and onset

of the diffusion-controlled growth may be defined as the transition zone. 270 291, 292 In the

transition zone, both the solid-state growth mechanisms and fluidity based mechanism are

in play.

Since bulk GC growth mode in the glassy state is a solid-state process due to local

molecular mobility rather than the global process, it has been suggested that the much

higher mobility near fracture-created surfaces is responsible for the GC growth. 278 In

contrast, the α process is more closely related to the diffusion-controlled mode of crystal

growth.272 It was suggested that the crystallization of amorphous CEL above Tg is

diffusion-controlled (associated with α relaxation process), 292 whereas the crystallization

below Tg is not.291 To probe the prominence of the α relaxation of CEL in GC growth mode,

the number of molecules (𝑛𝛼 ) being added to the crystalline phase in one structural

relaxation time 𝜏𝛼 was calculated by 𝑛𝛼 = 𝑢𝜏𝛼 𝑎⁄ . Taking the estimated molecular

diameter, a, of 10.29 Å and 𝜏𝛼 of 100 s at Tg 275, the 𝑛𝛼 for CEL is calculated to be 67~5443

in GC growth mode over the range of measured GC growth rate, u. The diameter of CEL

was estimated by treating CEL molecules as spheres having a packing coefficient of 70%,

and solve for a, where the sphere volume is the unit cell volume of CEL Form III. This is

much greater than the expected 𝑛𝛼 (< 1) if the dominating mechanism were bulk-diffusion-

124

controlled growth.265 This calculation further supports that α relaxation is not responsible

for GC growth and is consistent with the view that local molecular motions like surface

diffusion 278 and relaxation 41 play a role. 291

The transition from bulk GC growth and surface solid-state growth processes to the

fluidity based mechanism also differentiated CEL in two ways. Firstly, the transition zone

of CEL Form III crystal from GC or solid-state based growth to diffusion-controlled growth

is wider on surface (22 °C) than in bulk (12 °C) (Figure 5–4a and Figure 5–7). Both are

substantially wider than all previous reported surface and bulk transition zones of other

compounds, e.g., OTP (11°C for surface and 5 °C for bulk) and NIF (13 °C for surface and

3 °C for bulk). 272 For some compounds, e.g., griseofulvin 268 and ROY (YN, R05 and ON

polymorphs), 272 the transition zones for the surface crystal growth were not obvious,

showing smooth growth rate profiles. Secondly, the surface Tt of Form III (41.4 °C) is

lower than that in bulk (44.8 °C). The higher Tt in the bulk suggests that the solid-state

growth in the bulk can persist until a higher temperature, indicating possibly different

mechanisms by which fluidity affects surface and bulk crystal growth. Tt for both surface

and bulk crystal growth of CEL Form I could not be determined due to the lack of data in

GC growth region (Figure 5–4b).

As the bulk growth rates for Forms I and III gradually converge with increasing

temperature in the diffusion-controlled region (Figure 5–7), a maximum growth rate near

the melting point of CEL is observed. The decrease in growth rate when approaching the

melting point, despite the higher mobility of CEL molecules, is attributed to the reduction

125

in thermodynamic driving force for crystallization near the melting point instead of a new

crystal growth mechanism. 26

Figure 5–7. Surface and bulk crystal growth rate - temperature profiles of Form I and III

of CEL. Characteristic crystal morphologies in different temperature regions are shown.

Open symbols indicate growth mechanisms distinct from the diffusion controlled growth

(solid symbols). Tt was determined from linear fitting, and the black and purple dashed-

lines aid visualization of transition zones on surface and in bulk.

5.4.3 Polymorph Cross-nucleation

The Form I and Form III CEL crystallites were easily distinguishable by PLM since

they differed in color (when viewed between crossed polarizers) and in texture. Form III

20 40 60 80 100 120 140 160

1E-12

1E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3

Surface Form III

Bulk Form III

Surface Form I

Bulk Form I

Gro

wth

rat

e (m

/s)

Temperature (°C)

Tt=41.4°C for

surface (III)

Tt=44.8°C for

bulk (III)

Tg=51.8 °C

Tm=160.8 °C

126

crystallites were composed of fibrous crystals and appeared dark and dense (Figure 5–5a,

c). These crystals exhibited micro X-ray diffraction peaks (10.69 º, 12.97 º, 14.80 º, 16.07

º, 17.96 º, 19.6 º, 21.52 º, 22.42 º, 24.60 º, 27.02 º, 27.72 º, 29.56 º, 32.2 º) characteristic of

Form III (Figure 5–8).293 Form I crystallites were composed of plate-like crystals with

smoother surfaces (Figure 5–5b, c) and exhibited an X-ray diffraction pattern having peaks

(16.4 º, 18.24 º, 19.08 º, 21.96 º, 23.08 º, 25.28 º, 28.72 º) characteristic of the Form I

(Figure 5–8).293 The melting points, determined by DSC of crystallites isolated from the

samples also confirmed their Form I and Form III polymorphic identity (Figure 5–9).293

When characterized by Confocal Raman Spectroscopy, the main difference is that the

single peak at 1154 cm-1 for Form III is split into two peaks (1139.83 and 1160 cm-1) for

Form I (Figure 5–10). This spectral difference was subsequently used to identify

polymorphic nature of crystallites or to map samples by Raman microscopy (Figure 5–5d

and Figure 5–11).

Figure 5–8. Characterization of CEL Form I and Form III by powder X-ray diffraction

and micro-diffractometer.

127

Figure 5–9. DSC and TGA curves of Forms I and III of CEL crystallization from melt

Figure 5–10. Characterization of CEL Form I and Form III by Raman spectroscopy;

characteristic peak of Form III is at 1154cm-1, which split into two peaks for Form I.

100 200 300

-20

-10

0

Form I

Form III

Weight %

Temperature (°C)

He

at F

low

(m

W)

45

50

55

60

65

70

75

80

85

90

95

100

We

igh

t %

128

Figure 5–11. a) Characteristic peak representing CEL Form I used for Raman mapping

and b) Raman 1D line scan result.

Figure 5–10 shows the Raman spectra for amorphous, Form I, and Form III CEL.

Forms I and III have many differences in their Raman spectra. The 1154 cm-1 peak for

Form III is split into two peaks at 1139 cm-1 and 1160 cm-1. The 1139 cm-1 peak was used

for mapping (details in Figure 5–11). Other characteristic peaks are 326 cm-1, 447 cm-1,

760 cm-1, 3118 cm-1, 3230 cm-1,3338 cm-1, and 3349 cm-1 for Form III; and 1295 cm-1,

1311 cm-1, 1408 cm-1, 3053 cm-1, and 3258 cm-1 For Form I. There are also some minor

peak shifts and splitting, e.g., The 409 cm-1 peak in Form I is shifted to 395 cm-1 in Form

b)

a) 1113 to 1143 cm

-1

Relative positions scanned (µm)

129

III spectrum; and the 1596 cm-1 peak in Form I is shifted to 1574 cm-1 in Form III. The 624

cm-1 peak in Form I is split into 626 cm-1 and 644 cm-1 peaks in Form III.

CEL also exhibited cross-nucleation behavior between Form I and III polymorphs,

where heterogeneous nucleation of a crystalline phase is induced by crystals of the same

chemical composition but a different phase. 294 During surface growth, Form III nucleated

from Form I seeds at temperatures from 25 to 60 ºC because only Form III grew from either

Form I or Form III seeds. At ≥65 ºC, Form I could nucleate on Form III because both

polymorphs grew from Form III seeds, but Form III never nucleated on Form I. During the

bulk growth of CEL, Form III could nucleate on Form I below 60 ºC because only Form

III grew regardless of polymorphic nature of seed crystals. Above 60 ºC, Form I seeds

always led to growth of Form I. Meanwhile, above 60 ºC, cross-nucleation on Form III

seeds was complex. In the temperature range of 60 ºC - 90 ºC, Form I always nucleated on

Form III so that Form III never grew. However, Form III seeds could either cross-nucleate

Form I or grow into larger polycrystalline Form III in temperature range of 100 ºC - 140

ºC. Cross-nucleation was confirmed by both 1D Raman line scan (Figure 5–11) and 2D

Raman mapping (Figure 5–5c & d). It is useful to mention that when cross-nucleation

occurred, the fast growing polymorph always nucleated on the slow growing polymorph in

the respective temperature ranges.294

5.4.4 Relationship between Growth Mechanism and Crystal Morphology

The morphology of crystals grown from melt is highly dependent on the

temperature (Figure 5–12), eluding to the different underlying growth mechanisms. The

GC growth (30 and 40 °C) of Form III surface and bulk CEL led to compact spherulites,

130

while Form III crystals grew as fibers at 55 - 60 °C (part of the transition zone). At the

beginning of the diffusion-controlled region (70 °C), Form III grew as fibers while Form I

grew as polycrystalline crystals with well-defined growth front. As the temperature further

entered the diffusion-controlled region (80 - 150 °C), both Form III and Form I grew with

polycrystalline texture. At 160 °C (1-2 °C below melting temperature), the CEL crystals

grew into single crystals of Form III. Similar trend was also observed in ROY polymorphs.

275 However, some ROY polymorphs grew in the bulk as both fast-growing fibers and slow-

growing crystalline spherulites between Tt and 1.15 Tg.272 Nevertheless, fibrous bulk

growth of CEL within the transition zone (44.8 – 56.6 °C) only grew as compact spherulites

(Figure 5–12).

Figure 5–12. PLM images of CEL crystals grown from amorphous films at different

temperatures. B is for bulk and S is for surface samples. Scale bars are all 100 µm.

131

5.4.5 Effect of Polymorphism on Growth Rate

In the diffusion-controlled region, molecules arriving at the growth front must also

take on the correct orientation and conformation before it is accepted into the crystalline

phase. Therefore, crystal packing pattern may influence crystal growth rate from the same

amorphous material. Comparing crystal growth rates of different polymorphs offers an

opportunity to assess the relative contributions of this step relative to diffusion. For CEL,

the surface growth rate of Form I was consistently higher (18 % - 60 %) than that of Form

III over the temperature range of 70 - 150 °C (Figure 5–13). Therefore, the step of adjusting

molecular conformation and orientation required for Form III crystal noticeably slowed

down CEL crystal growth on the surface. Within the temperature range where bulk growth

rates of Forms I and III could both be measured (100 - 150 °C), Form I bulk growth rate

was about 8 – 43 % higher than that of Form III (Figure 5–14). Thus, the process of

molecules adjusting conformation and orientation also played a significant role in crystal

growth in the bulk.

132

Figure 5–13. Growth rate profile of both crystal forms for surface samples of amorphous

CEL; the inset represents the growth rates of two polymorphs on linear scale

Figure 5–14. Growth rate profile of both crystal forms for bulk samples of amorphous

CEL; the inset in the growth rates of two polymorphs in linear scale

0 20 40 60 80 100 120 140 160 180

1E-12

1E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3Transition

Zone (III) Diffusion-controlled Growth

Surface Form III Solid-state

Surface Form III Diffusion

Surface Form I Diffusion

Gro

wth

rat

e u (

m/s

)

Temperature (°C)

Tt=41.4 °C for

surface (III)

GC Growth

Tg=51.8 °C T

m=160.8 °C

0 20 40 60 80 100 120 140 160 180

1E-12

1E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3

Gro

wth

Rat

e u

(m

/s)

Temperature (°C)

Bulk Form III Diffusion

Bulk Form III GC

Bulk Form I Diffusion

GC Growth Diffusion-controlled Growth

Tg=51.8 °C

Tt=44.8 °C for

bulk (III)

Tm=160.8 °C

Transition

Zone (III)

133

The slightly higher percent increase in growth rate of Form I over Form III on the

surface (18 % - 60 %) than that in the bulk (8 % - 43 %) (Figure 5–13 and Figure 5–14)

may be simply because of the different temperature regions over which both polymorphs

can grow, i.e., 70 - 150 C for surface growth and 100 - 160 C for bulk growth. The

relative difference became smaller at higher temperatures because molecules can rotate, re-

orientate, or undergo conformational changes more rapidly. Thus, its effect on the time

taken for the different polymorphs to grow is reduced with temperature. This is supported

by the fact that, below and near corresponding Tg, differences in growth rates between α

and γ polymorphs of IMC (70 - 570 %) and Forms I and IV polymorphs of carbamazepine

(440 – 600 %) were significantly larger than that between CEL Forms I and III at

temperatures significantly above Tg. 261, 295

5.4.6 Arrhenius Growth Kinetics of CEL Crystals

To gain more insight into the kinetics of the crystal growth process, the surface and

bulk growth rate data for Forms I and III were fitted to the Arrhenius equation (Eqn. 5–1)

to obtain the corresponding activation energies, Ea (Table 5–2). Both diffusion-controlled

and GC growth kinetics follow the Arrhenius relationship, as shown by the linear

dependence of ln(k) on 1/T (Figure 5–15). However, the points started to digress from

linearity above 120 °C and was therefore not included for fitting. This digression could be

attributed to the more prominent role of the reduced free energy difference between crystal

and the melt when approaching melting point (zero at the melting point). In both surface

and bulk crystallization, the Ea is significantly lower for solid-state growth/GC growth than

for diffusion-controlled growth. This is aligned with the fact that GC growth requires only

134

minor local molecular mobility (e.g., β process), which is easier than global mobility of

molecules (e.g., α process) in the diffusion-controlled growth region. 296

In the solid-state growth and GC growth region, the Ea for CEL Form III surface

growth (68 kJ/mol) is lower than that for bulk growth (91 kJ/mol). Both activation energies

for surface and bulk lie in the middle of the range of Ea in systems studied for GC growth

so far (30 – 185 kJ/mol) (Table 5–2). Since the surface crystal growth has been correlated

with surface diffusion coefficient, 297 and both surface and bulk diffusion coefficients

follow Arrhenius relationship, 279, 298 Ea based on diffusion coefficients is expected to be

comparable to that based on growth rate. In fact, the Ea of surface diffusion coefficient

(146 kJ/mol) is lower than that of the bulk diffusion coefficient (368 kJ/mol) for tris-

naphthyl benzene (TNB).298 This is consistent with the observed lower Ea for the surface

growth than the bulk growth of CEL.

Another observation is that Ea values of both CEL polymorphs in the diffusion-

controlled regions, both on surface and in the bulk, are higher than those in the solid-state

growth/GC growth region. This is consistent with the fact that below Tt, the solid-state

growth mechanisms dominate over diffusion mechanism. In addition, most of previous

work on growth kinetics only included the bulk crystal growth process. As a result, the

different Ea values between surface and bulk crystallization processes have not been

established. The lower surface Ea than that of bulk Ea (~25% for Form III CEL) in solid-

state growth region, and ~42% for Form I CEL in diffusion-controlled region) is reported

here for the first time.

135

Table 5–2. Kinetic parameters of CEL bulk and surface processes compared to known

examples of GC crystal growth in bulk glasses

Systems Growth mode Tg, K Tt, K 𝒍𝒐𝒈 𝒖 at Tt,

m/s

𝑬𝒂,

kJ/mol Ref. #

ROY-YT04 bulk GC 260 269 -7.2 83 272

ROY-Y bulk GC 260 265 -7.4 76 272

ROY-OP bulk GC 260 265 -7.5 88 272

ROY-R bulk GC 260 261 -8.2 71 272

TP (TPm) bulk GC 272 280 -6.1 127 273

NIF (β) bulk GC 315 316 -8.5 126 271

IMC (γ) bulk GC 315 306 -10.4 131 270

OTP bulk GC 246 249 -7.8 65 276

DPCP bulk GC 222 225 -7.8 30 274

DPCH bulk GC 230 233 -8.1 45 274

salol bulk GC 222 226 -7.1 57 272

IPB bulk GC 129 127 -9.5 - 299

toluene bulk GC 118 116 -8.3 50 272

GSF bulk GC 361 360 -8.5 185 268

CEL (III) bulk GC 325 317 -9.5 91 This work

CEL (III) surface diffusion 325 315 -9.1 68 This work

136

Figure 5–15. Arrhenius relationship of CEL growth rate a) on the surface and b) in bulk

a)

b)

137

In the diffusion-controlled regions (≥ 63.3 C on the surface and ≥ 56.6 C for

bulk), Forms I and III have similar Ea on the surface, being 186 kJ/mol and 185 kJ/mol,

respectively. The bulk growth of Form I has a significantly higher Ea (265 kJ/mol) than

that on the surface. The bulk Ea of CEL Form I in the diffusion-controlled regions is the

highest in all systems reported so far (Griseofulvin had Ea of 236 kJ/mol in diffusion-

controlled region).268 The Form III doesn’t have enough data for obtaining bulk Ea in this

temperature range, but similar observation to the Form I growth is expected.

The crystal growth rates at Tt normalized by Tg, i.e., Tt/Tg, for small organic

molecules studied so far follow a common trend (Figure 5–16). This trend is upheld when

bulk crystal growth of CEL is added, which suggests a universal physical mechanism for

the GC crystal growth model among these investigated compounds.270 In more than half of

these systems, faster GC growth can persist up to a temperature above Tg (Figure 5–16).

270 The Tt/Tg for CEL GC growth in bulk is among the lowest (comparable to that of γ-

IMC), compared to other compounds studied. Therefore, the Tt is lower than the Tg of bulk

CEL, indicating an earlier disruption by the onset of diffusion-controlled growth

mechanism than other systems.

138

Figure 5–16. Growth rate u, at Tt as a function of Tt / Tg of several small organic

molecules studied for GC growth mode. The black star represents CEL.

5.5 Conclusions

We have systematically studied the crystal growth kinetics of amorphous CEL both

on the surface and in bulk. Upon cooling, the crystal growth rates near and below Tg

(51.8°C) for both surface and bulk Form III CEL abruptly increased from those predicted

from diffusion-controlled crystal growth mechanism, agreeing with the known solid-state

and GC growth mechanism. The nearly 80-fold increase in the surface growth rate of CEL

Form III crystal over bulk growth in the glassy region is attributed to the fast molecular

surface diffusion. These two fast crystal growth mechanisms are influenced by the onset of

fluidity differently, showing higher Tt and a wider transition zone on the surface. The

transitions zones of both surface and bulk growth of CEL are the widest among all known

0.96 0.97 0.98 0.99 1.00 1.01 1.02 1.03 1.04

-11

-10

-9

-8

-7

-6

ROY-YT04 ROY-Y

ROY-OP ROY-R

TP (TPm) NIF ()

IMC () OTP

DPCP DPCH

salol IPB

toluene GSF

CEL (bulk)

log u

at

Tt (

m/s

)

Tt /T

g

139

small organic molecules exhibiting GC growth. CEL also exhibited cross-nucleation

phenomenon, which has been characterized by micro-Raman mapping. Form I exhibited

faster growth rate than Form III in all crystal growth processes, where both followed the

Arrhenius relationship. The activation energies for diffusion controlled mechanisms are

higher than surface solid-state growth/bulk GC growth mechanisms. In summary, CEL

exhibits several intriguing phenomena in crystal growth from the amorphous state,

including fast surface growth, GC-growth, cross-nucleation, and growth mechanism

dependent crystallite morphologies. Compared to other molecules investigated for growth

mechanisms, two unique features with CEL include: 1) the widest transition zone between

surface solid-state growth/bulk GC growth and diffusion-controlled growth, and 2) the

highest activation energy in the diffusion controlled growth in the bulk. Data presented

here can also help the development and stabilization of amorphous based tablet products

of CEL with improvement in solubility and dissolution properties.

140

Chapter 6.

Research Summary and Future Work

141

Research Summary

Celecoxib (CEL) is a non-steroidal anti-inflammatory drug (NSAID) that is widely

prescribed for treating arthritis, osteoarthritis and rheumatoid arthritis associated pain and

inflammation. It is a high dose drug only available as capsules, and not in the most patient-

compliant tablet dosage form. The low bulk density, high elasticity and poor flowability of

commercial CEL Form III makes it difficult to be formulated into tablets suitable for high

speed direct compression tablet manufacture. In this thesis, the shortcomings of the current

Form III of CEL are overcome by solid state engineering, and a direct compression tablet

formulation is successfully developed via crystal engineering.

Understanding the molecular origin of the properties of CEL Form III that

challenge direct compression tablet manufacturing is the first step of developing strategies

to mitigate these problems via Quality by Design (QbD) approach. In Chapter 2, the

exceptionally high elasticity of CEL is examined using analytical techniques – micro-

Raman spectroscopy and nanoindentation, and understood on a molecular level. A large

needle-shaped single crystal of CEL Form III is grown, which is quantitatively bent for a

high elastic strain of 3.65%; after which the crystal recovered to straight conformation.

Nanoindentation suggested a high Elastic Modulus (E) of 16.27 ± 0.43 GPa. A model is

proposed based on analysis of key intermolecular interactions within CEL Form III crystals,

which is then confirmed by micro-Raman spectroscopy. The results suggest numerous

weak and a few strong intermolecular interactions establish the high elasticity on the major

(001) face of CEL crystals. This implicates that future crystal engineering approach to

modify elasticity could aim at the intermolecular interactions within the crystal structure.

142

In search of alternative solid forms of CEL, pharmaceutically acceptable solvates

of CEL are screened as an alternative way to improve the tablet manufacturability of CEL

through crystal engineering. During the screening of CEL solvates, two stoichiometric

solvates of CEL and N-methyl-2-pyrrolidone (NMP) are discovered with stoichiometric

ratios of 1:1 and 1:2 CEL:NMP. The two solvates have distinct crystal structures and

physicochemical properties. An intriguing conversion from CEL di-NMP solvate to mono-

NMP solvate as CEL di-NMP solvate is heated up is captured using Differential Scanning

Calorimetry (DSC). A hot stage combined with polarized microscopy (HSM) reveals a

series of events including desolvation, dissolution of di-NMP solvate, crystal growth,

dissolution and melting of mono-NMP solvate, which match with each step in the DSC

curve. The mono-NMP solvate of CEL is stable at room temperature while di-NMP solvate

easily converts to mono-NMP solvate at ambient conditions. Crystal structure analysis and

the energy framework reveal the tightly bound NMP in mono-NMP crystal structure.

Channels of loosely bound NMP molecules corresponding to 1 stoichiometric NMP in di-

NMP structure are found, in addition to the tightly bound NMP molecules that have very

similar packing as mono-NMP solvates. These structural features contribute to the ultra-

stability of mono-NMP solvate and the easy conversion form di-NMP to mono-NMP

solvate without sacrificing crystallinity. Infrared spectroscopy further confirmed the two

interaction states of NMP molecules in the di-NMP solvate structure. This study details the

structure-property relationship of the two CEL-NMP solvates and their interconversions,

and illustrates the importance of using orthogonal analytical techniques to characterize

solid state properties of drugs (Chapter 3).

143

To develop a direct compression (DC) tablet formulation for CEL, a

pharmaceutically acceptable solvate - celecoxib dimethyl sulfoxide (DMSO) solvate is

characterized and selected as suitable for DC tablet of better manufacturability and

dissolution performance. The challenges facing commercial Form III powder for

developing a DC formulation include low bulk density, poor flowability and tablet

lamination issues. Through crystal engineering, we utilize the DMSO solvate of CEL to

greatly improve the bulk density by 6 fold, and the flow factor (ffc) by 2 fold. By reducing

the elastic modulus (E) of CEL Form III by 1/3, CEL-DMSO tablets exhibit less in-die

elastic recovery, and thus much alleviated lamination problem. The lamination of CEL-

DMSO tablets are completely eliminated by developing a DC table formulation, which is

a result of step-by-step data-oriented decision making process. Through optimization of

formulations by evaluating the tablet quality, friability, ejection force, tabletability and

dissolution performance, a DC tablet formulation is successfully developed. This is an

example of combining a crystal engineering approach and guiding principles of the

material science tetrahedron (MST) to develop a successful tablet formulation of a high

dose drug (Chapter 4).

As a biopharmaceutics classification system (BCS) class II drug, CEL dissolution

is limited by solubility. Thus, high doses of CEL are required to elicit a therapeutic benefit.

To increase the solubility and thus the bioavailability of CEL, its amorphous form is studied

for potential development as a candidate for tablet formulation. With the increased

bioavailability, the dose can be potentially lowered, leaving more room for excipients to

fine-tune the formulation performance as a DC tablet formulation. One problem about

amorphous solids is its instability and tendency to convert to it crystalline counterpart. In

144

Chapter 5, the stability of amorphous CEL is evaluated through monitoring the crystal

growth rate of melt-quenched amorphous CEL at a range of temperatures. The mechanism

of crystal growth is found to be different on the free surface and in bulk of melt. Faster

surface crystal growth than bulk is found across all temperature ranges. Below the glass

transition temperature Tg of CEL, the glass-to-crystal (GC) growth mode is activated,

resulting in fast crystal growth rates at pharmaceutically relevant temperatures. Cross-

nucleation between CEL Forms I and III is also observed, which play a role in the crystal

growth profiles. The crystal growth rate follows Arrhenius kinetics, with the activation

energy of surface crystal growth lower than in the bulk. This study systematically studied

the crystal growth kinetics of amorphous CEL. It lays the foundation for future

development of amorphous solid dispersion (ASD) based tablet products with

improvement in solubility and dissolution properties.

145

Future Work

This thesis has taken celecoxib (CEL) as a model drug and delved into utilizing

solid state engineering strategies to obtain the desired properties suitable for direct

compression (DC) tablet manufacturing. Guided by Quality by Design (QbD) principles

and Material Science Tetrahedron (MST), the physicochemical properties and tablet

manufacturability of celecoxib have been greatly improved, with a viable DC tablet

formulation successfully developed.

Solid-state engineering stems from understanding of molecular origins of

properties. From the verified model of highly elastic CEL Form III crystals (Chapter 2),

the high elasticity of (001) face of CEL is well understood. However, the minor

(01̅1)/(011̅) faces of CEL needle crystals exhibits brittle fracture when subjected to 3-

point bending experiments. When subjected to an external force, the changes in molecular

interactions of minor (01̅1)/(011̅) faces are different from the major (001) face, which may

lead to drastically different properties. The size of crystals faces also leads to different

mechanical behavior. This exploitation can be achieved by growing larger single crystals

of needle-shaped CEL Form III crystals, for which the nanoindentation could be performed

on minor faces of the crystal. The different mechanical behavior on different faces of CEL

contribute to the mechanical performance during tableting to different extents. If a crystal

is highly anisotropic, the effect of major crystal faces may dominate the tablet behavior of

the crystal; conversely, the tableting properties can be equally influenced by every face of

an isotropic crystal. The details of correlating anisotropic mechanical properties of crystals

with the tableting performance requires further studies with selective model compounds.

146

The insights obtained from this research will enrich the landscape of ‘structure-property-

performance’ within MST, especially in connection to tableting behaviors of

pharmaceutical materials.

During solvate screening of CEL, 6 solvates were discovered and characterized.

The 6 solvent molecules involved are dimethyl acetamide (DMA), dimethyl formamide

(DMF), N-methyl-2-pyrrolidone (NMP), dimethyl sulfoxide (DMSO), tetramethyl urea

(TMU) and N, N′-Dimethylpropyleneurea (DMPU). The mechanical properties (E and H)

of the 6 solvates are characterized by nanoindentation. All six solvates are highly stable

solvates having high melting points below which no desolvation occurs. One commonality

of the mechanical properties among the 6 solvates are their fracture tendency after

nanoindentation tests (e.g. CEL-DMSO solvate fracture, Figure 4–10). Another similarity

is that all 6 solvates had lower E and H on their major face (002) than CEL Form III on

(001) face. The tendency to fracture of CEL solvate crystals indicate that they are more

brittle than CEL Form III. The fracture mechanism, fracture plane, and fracture toughness

remained unexplored. Further studies of fracture behavior will gain structural insights into

the mechanisms of fracture. The fracture toughness has also been correlated with milling

behavior in molecular crystals. 300 The detailed study of fracture toughness of the 6 CEL

solvates will also have indications on their susceptibility to milling and size reduction, as

well as the milling induced disorder and instability, which are important process parameters

during drug product development.

A direct compression tablet formulation is successfully developed in Chapter 4

through crystal engineering methods by utilizing a stable and pharmaceutically stable

147

solvate – CEL-DMSO solvate. The CEL-DMSO tablet exhibits better dissolution profile

with faster onset than Celebrex® capsules. However, the intrinsic dissolution rate (IDR)

study of pure CEL-DMSO solvate crystals reveals the fast conversion and precipitation to

thermodynamically stable Form III upon contact with dissolution medium. This suggests a

much higher thermodynamic solubility of CEL-DMSO than CEL Form III, leading to fast

supersaturation generation followed by rapid precipitation due to instability in solution.

Maintaining the supersaturation after CEL-DMSO is in contact with water and thus

maintaining its solubility advantage, will effectively increase the bioavailability of CEL.

Selection of polymer inhibitors to be added into the tablet formulation will be an effective

measure to solve this problem. The dose of CEL can also be lowered potentially through

this measure. Polymer screening can be done through monitoring the nucleation induction

times of CEL crystals under presence of different types of polymers in saturated CEL

solution. The longer nucleation induction time corresponds to better ability of the

polymeric material to stabilize CEL in dissolution media, forming the ‘spring and

parachute’ curve to sustain high CEL concentrations.

In Chapter 5, the stability of amorphous CEL is evaluated by understanding the

crystal growth mechanisms of amorphous CEL at a range of temperatures. The fast glass-

to-crystal growth below its glass transition temperature Tg, especially on the surface,

indicates its instability as amorphous state. To stabilize the amorphous CEL, an amorphous

solid dispersion (ASD) of CEL needs to be developed. First, the polymer type and grade

needs selected to be miscible with CEL without phase separation; it also needs to stabilize

amorphous CEL and prevents crystallization from happening both in solid state. Lastly,

this polymer should also be a good stabilizer when the ASD is being dissolve in dissolution

148

media. In addition to the enhancement in solubility and dissolution, the manufacturability

of ASD also needs to be evaluated. For a direct compression tablet formulation, the flow

properties, punch sticking propensities and tabletability should be considered. Preliminary

studies have shown that the CEL and HPMC-AS ASD prepared by fast solvent evaporation

exhibited much less punch sticking propensity compared to CEL Form III. This is another

example of using solid-state engineering method to overcome manufacturability challenges.

The flow and tabletability of this ASD remain to be evaluated. However, the different

processing conditions and particle engineering techniques can be employed in the DC

tablet formulation design of CEL ASDs under the MST landscape.

Finally, a successful direct compression tablet formulation of CEL with improved

stability, dissolution performance and excellent manufacturability can be successfully

developed via solid-state engineering strategies.

149

Bibliography

1. Gibson, M., Pharmaceutical Preformulation and Formulation: A Practical Guide

from Candidate Drug Selection to Commercial Dosage Form. 2nd ed.; CRC Press: 2016.

2. Sun, C. C.; Hou, H.; Gao, P.; Ma, C.; Medina, C.; Alvarez, F. J., Development

of a high drug load tablet formulation based on assessment of powder manufacturability:

moving towards quality by design. J. Pharm. Sci. 2009, 98 (1), 239-47.

3. Ferro, L. J.; Miyake, P. S. Polymorphic crystalline forms of celecoxib. US 7476744

B2, Jan. 13, 2009.

4. Gao, D.; Hlinak, A. J.; Mazhary, A. M.; Truelove, J. E.; Vaughn, M. B. Celecoxib

Compositions. Jan.8, 2015.

5. Suresh, P.; Sreedhar, I.; Vaidhiswaran, R.; Venugopal, A., A comprehensive

review on process and engineering aspects of pharmaceutical wet granulation. Chemical

Engineering Journal 2017, 328, 785-815.

6. Sun, C. C., Materials science tetrahedron--a useful tool for pharmaceutical research

and development. J. Pharm. Sci. 2009, 98 (5), 1671-87.

7. Kottke, M. J.; Rudnic, E. M., Modern Pharmaceutics. In Modern Pharmaceutics,

Banker, G. S.; Siepmann, J.; Rhodes, C., Eds. CRC Press: Boca Raton, 2002.

8. Li, Z.; Zhao, L.; Lin, X.; Shen, L.; Feng, Y., Direct compaction: An update of

materials, trouble-shooting, and application. Int. J. Pharm. 2017, 529 (1-2), 543-556.

9. Dosage forms. U.S. Food & Drug Administration. 2018. Available at:

https://www.fda.gov/industry/structured-product-labeling-resources/dosage-forms.

(accessed August 11, 2020).

https://www.fda.gov/industry/structured-product-labeling-resources/dosage-forms

150

10. Suzuki, H.; Onishi, H.; Takahashi, Y.; Iwata, M.; Machida, Y., Development of

oral acetaminophen chewable tablets with inhibited bitter taste. Int. J. Pharm. 2003, 251

(1-2), 123-132.

11. Codd, J. E.; Deasy, P. B., Formulation development and in vivo evaluation of a

novel bioadhesive lozenge containing a synergistic combination of antifungal agents. Int.

J. Pharm. 1998, 173 (1-2), 13-24.

12. Chang, S. Y.; Sun, C. C., Interfacial bonding in formulated bilayer tablets. Eur. J.

Pharm. Biopharm. 2020, 147, 69-75.

13. Chang, S. Y.; Li, J. X.; Sun, C. C., Tensile and shear methods for measuring

strength of bilayer tablets. Int. J. Pharm. 2017, 523 (1), 121-126.

14. Chang, S. Y.; Sun, C. C., Minimum Interfacial Bonding Strength for Bilayer

Tablets Determined Using a Survival Test. Pharm. Res. 2019, 36 (10), 139.

15. Wang, C.; Hu, S.; Sun, C. C., Expedited Development of Diphenhydramine Orally

Disintegrating Tablet through Integrated Crystal and Particle Engineering. Mol. Pharm.

2017, 14 (10), 3399-3408.

16. Al-Khattawi, A.; Mohammed, A. R., Challenges and emerging solutions in the

development of compressed orally disintegrating tablets. Expert Opin Drug Discov 2014,

9 (10), 1109-20.

17. Choi, H.-G.; Kim, C.-K., Development of omeprazole buccal adhesive tablets with

stability enhancement in human saliva. J. Control. Release 2000, 68 (3), 397-404.

18. Munday, D.; Fassihi, A., Controlled release delivery: Effect of coating composition

on release characteristics of mini-tablets. Int. J. Pharm. 1989, 52 (2), 109-114.

151

19. Yu, L. X., Pharmaceutical quality by design: product and process development,

understanding, and control. Pharm. Res. 2008, 25 (4), 781-91.

20. Thakral, N. K.; Kelly, R. C., Salt disproportionation: A material science

perspective. Int. J. Pharm. 2017, 520 (1-2), 228-240.

21. Vippagunta, S. R.; Brittain, H. G.; Grant, D. J. W., Crystalline solids. Advanced

Drug Delivery Reviews 2001, 48 (1), 3-26.

22. Datta, S.; Grant, D. J., Crystal structures of drugs: advances in determination,

prediction and engineering. Nat. Rev. Drug Discov. 2004, 3 (1), 42-57.

23. Newman, A.; Knipp, G.; Zografi, G., Assessing the performance of amorphous

solid dispersions. J. Pharm. Sci. 2012, 101 (4), 1355-77.

24. Baghel, S.; Cathcart, H.; O'Reilly, N. J., Polymeric Amorphous Solid Dispersions:

A Review of Amorphization, Crystallization, Stabilization, Solid-State Characterization,

and Aqueous Solubilization of Biopharmaceutical Classification System Class II Drugs. J.

Pharm. Sci. 2016, 105 (9), 2527-2544.

25. Patel, S.; Kou, X.; Hou, H. H.; Huang, Y. B.; Strong, J. C.; Zhang, G. G. Z.;

Sun, C. C., Mechanical Properties and Tableting Behavior of Amorphous Solid

Dispersions. J. Pharm. Sci. 2017, 106 (1), 217-223.

26. Yu, L., Amorphous pharmaceutical solids: preparation, characterization and

stabilization. Adv. Drug Deliv. Rev. 2001, 48 (1), 27-42.

27. Xie, T.; Taylor, L. S., Dissolution Performance of High Drug Loading Celecoxib

Amorphous Solid Dispersions Formulated with Polymer Combinations. Pharm. Res. 2016,

33 (3), 739-50.

152

28. Benet, L. Z., The role of BCS (biopharmaceutics classification system) and BDDCS

(biopharmaceutics drug disposition classification system) in drug development. J. Pharm.

Sci. 2013, 102 (1), 34-42.

29. Pandi, P.; Bulusu, R.; Kommineni, N.; Khan, W.; Singh, M., Amorphous solid

dispersions: An update for preparation, characterization, mechanism on bioavailability,

stability, regulatory considerations and marketed products. Int. J. Pharm. 2020, 586,

119560.

30. Sotthivirat, S.; McKelvey, C.; Moser, J.; Rege, B.; Xu, W.; Zhang, D.,

Development of amorphous solid dispersion formulations of a poorly water-soluble drug,

MK-0364. Int. J. Pharm. 2013, 452 (1-2), 73-81.

31. Liu, X.; Feng, X.; Williams, R. O.; Zhang, F., Characterization of amorphous solid

dispersions. J. Pharm. Investig. 2017, 48 (1), 19-41.

32. Wang, K.; Sun, C. C., Crystal Growth of Celecoxib from Amorphous State:

Polymorphism, Growth Mechanism, and Kinetics. Cryst. Growth Des. 2019, 19 (6), 3592-

3600.

33. Sun, Y.; Zhu, L.; Wu, T.; Cai, T.; Gunn, E. M.; Yu, L., Stability of amorphous

pharmaceutical solids: crystal growth mechanisms and effect of polymer additives. AAPS

J 2012, 14 (3), 380-8.

34. Law, D.; Schmitt, E. A.; Marsh, K. C.; Everitt, E. A.; Wang, W.; Fort, J. J.;

Krill, S. L.; Qiu, Y., Ritonavir-PEG 8000 amorphous solid dispersions: in vitro and in vivo

evaluations. J. Pharm. Sci. 2004, 93 (3), 563-70.

153

35. Amidon, G. L.; Lennernas, H.; Shah, V. P.; Crison, J. R., A theoretical basis for a

biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and

in vivo bioavailability. Pharm. Res. 1995, 12 (3), 413-20.

36. Patel, M. A.; Luthra, S.; Shamblin, S. L.; Arora, K. K.; Krzyzaniak, J. F.; Taylor,

L. S., Assessing the Risk of Salt Disproportionation Using Crystal Structure and Surface

Topography Analysis. Cryst. Growth Des. 2018, 18 (11), 7027-7040.

37. Modi, S. R.; Dantuluri, A. K. R.; Perumalla, S. R.; Sun, C. Q. C.; Bansal, A. K.,

Effect of Crystal Habit on Intrinsic Dissolution Behavior of Celecoxib Due to Differential

Wettability. Cryst. Growth Des. 2014, 14 (10), 5283-5292.

38. Sun, C., Improving powder flow properties of citric acid by crystal hydration. J.

Pharm. Sci. 2009, 98 (5), 1744-9.

39. Sun, C.; Grant, D. J., Influence of crystal shape on the tableting performance of L-

lysine monohydrochloride dihydrate. J. Pharm. Sci. 2001, 90 (5), 569-79.

40. Paul, S.; Taylor, L. J.; Murphy, B.; Krzyzaniak, J. F.; Dawson, N.; Mullarney,

M. P.; Meenan, P.; Sun, C. C., Toward a Molecular Understanding of the Impact of Crystal

Size and Shape on Punch Sticking. Mol. Pharm. 2020, 17 (4), 1148-1158.

41. Hooper, D.; Clarke, F. C.; Docherty, R.; Mitchell, J. C.; Snowden, M. J., Effects

of crystal habit on the sticking propensity of ibuprofen-A case study. Int. J. Pharm. 2017,

531 (1), 266-275.

42. Hecq, J.; Deleers, M.; Fanara, D.; Vranckx, H.; Amighi, K., Preparation and

characterization of nanocrystals for solubility and dissolution rate enhancement of

nifedipine. Int. J. Pharm. 2005, 299 (1-2), 167-77.

154

43. Shegokar, R.; Muller, R. H., Nanocrystals: industrially feasible multifunctional

formulation technology for poorly soluble actives. Int. J. Pharm. 2010, 399 (1-2), 129-39.

44. Kesisoglou, F.; Panmai, S.; Wu, Y., Nanosizing--oral formulation development

and biopharmaceutical evaluation. Adv. Drug Deliv. Rev. 2007, 59 (7), 631-44.

45. Khomane, K. S.; Bansal, A. K., Effect of particle size on in-die and out-of-die

compaction behavior of ranitidine hydrochloride polymorphs. AAPS PharmSciTech 2013,

14 (3), 1169-77.

46. Sun, C. C.; Himmelspach, M. W., Reduced tabletability of roller compacted

granules as a result of granule size enlargement. J. Pharm. Sci. 2006, 95 (1), 200-6.

47. Hou, H.; Sun, C. C., Quantifying effects of particulate properties on powder flow

properties using a ring shear tester. J. Pharm. Sci. 2008, 97 (9), 4030-9.

48. Loh, Z. H.; Samanta, A. K.; Sia Heng, P. W., Overview of milling techniques for

improving the solubility of poorly water-soluble drugs. Asian J. Pharm. Sci. 2015, 10 (4),

255-274.

49. Schenck, L.; Koynov, A.; Cote, A., Particle engineering at the drug substance, drug

product interface: a comprehensive platform approach to enabling continuous drug

substance to drug product processing with differentiated material properties. Drug Dev.

Ind. Pharm. 2019, 45 (4), 521-531.

50. Schenck, L.; Erdemir, D.; Saunders Gorka, L.; Merritt, J. M.; Marziano, I.; Ho,

R.; Lee, M.; Bullard, J.; Boukerche, M.; Ferguson, S.; Florence, A. J.; Khan, S. A.;

Sun, C. C., Recent Advances in Co-processed APIs and Proposals for Enabling

Commercialization of These Transformative Technologies. Mol. Pharm. 2020, 17 (7),

2232-2244.

155

51. Richman, M. D.; Fox, C. D.; Shangraw, R. F., Preparation and Stability of Glyceryl

Trinitrate Sublingual Tablets Prepared by Direct Compression. J. Pharm. Sci. 1965, 54,

447-51.

52. Sun, C. C., Dependence of ejection force on tableting speed-A compaction

simulation study. Powder Technol. 2015, 279, 123-126.

53. Tye, C. K.; Sun, C. C.; Amidon, G. E., Evaluation of the effects of tableting speed

on the relationships between compaction pressure, tablet tensile strength, and tablet solid

fraction. J. Pharm. Sci. 2005, 94 (3), 465-72.

54. Osei-Yeboah, F.; Sun, C. C., Validation and applications of an expedited tablet

friability method. Int. J. Pharm. 2015, 484 (1-2), 146-55.

55. Paul, S.; Chang, S. Y.; Dun, J.; Sun, W. J.; Wang, K.; Tajarobi, P.; Boissier, C.;

Sun, C. C., Comparative analyses of flow and compaction properties of diverse mannitol

and lactose grades. Int. J. Pharm. 2018, 546 (1-2), 39-49.

56. Zhou, Q.; Shi, L.; Chattoraj, S.; Sun, C. C., Preparation and characterization of

surface-engineered coarse microcrystalline cellulose through dry coating with silica

nanoparticles. J. Pharm. Sci. 2012, 101 (11), 4258-66.

57. Dun, J.; Osei-Yeboah, F.; Boulas, P.; Lin, Y.; Sun, C. C., A systematic evaluation

of poloxamers as tablet lubricants. Int. J. Pharm. 2020, 576, 118994.

58. Sun, W. J.; Aburub, A.; Sun, C. C., A mesoporous silica based platform to enable

tablet formulations of low dose drugs by direct compression. Int. J. Pharm. 2018, 539 (1-

2), 184-189.

59. Shanmugam, S., Granulation techniques and technologies: recent progresses.

Bioimpacts 2015, 5 (1), 55-63.

156

60. Keleb, E. I.; Vermeire, A.; Vervaet, C.; Remon, J. P., Twin screw granulation as

a simple and efficient tool for continuous wet granulation. Int. J. Pharm. 2004, 273 (1-2),

183-94.

61. Faure, A.; York, P.; Rowe, R. C., Process control and scale-up of pharmaceutical

wet granulation processes: a review. Eur. J. Pharm. Biopharm. 2001, 52 (3), 269-277.

62. Kristensen, H. G.; Schaefer, T., Granulation: A Review on Pharmaceutical Wet-

Granulation. Drug Dev. Ind. Pharm. 2008, 13 (4-5), 803-872.

63. Benali, M.; Gerbaud, V.; Hemati, M., Effect of operating conditions and physico–

chemical properties on the wet granulation kinetics in high shear mixer. Powder Technol.

2009, 190 (1-2), 160-169.

64. Shi, L.; Feng, Y.; Sun, C. C., Roles of granule size in over-granulation during high

shear wet granulation. J. Pharm. Sci. 2010, 99 (8), 3322-5.

65. Osei-Yeboah, F.; Feng, Y.; Sun, C. C., Evolution of structure and properties of

granules containing microcrystalline cellulose and polyvinylpyrrolidone during high-shear

wet granulation. J. Pharm. Sci. 2014, 103 (1), 207-15.

66. Shi, L.; Feng, Y.; Sun, C. C., Initial moisture content in raw material can

profoundly influence high shear wet granulation process. Int. J. Pharm. 2011, 416 (1), 43-

8.

67. Rajniak, P.; Mancinelli, C.; Chern, R. T.; Stepanek, F.; Farber, L.; Hill, B. T.,

Experimental study of wet granulation in fluidized bed: impact of the binder properties on

the granule morphology. Int. J. Pharm. 2007, 334 (1-2), 92-102.

157

68. Mackaplow, M. B.; Rosen, L. A.; Michaels, J. N., Effect of primary particle size

on granule growth and endpoint determination in high-shear wet granulation. Powder

Technol. 2000, 108 (1), 32-45.

69. Iveson, S. M.; Litster, J. D.; Hapgood, K.; Ennis, B. J., Nucleation, growth and

breakage phenomena in agitated wet granulation processes: a review. Powder Technol.

2001, 117 (1-2), 3-39.

70. Räsänen, E.; Rantanen, J.; Jørgensen, A.; Karjalainen, M.; Paakkari, T.; Yliruusi,

J., Novel Identification of Pseudopolymorphic Changes of Theophylline During Wet

Granulation Using Near Infrared Spectroscopy. J. Pharm. Sci. 2001, 90 (3), 389-396.

71. Wikstrom, H.; Marsac, P. J.; Taylor, L. S., In-line monitoring of hydrate formation

during wet granulation using Raman spectroscopy. J. Pharm. Sci. 2005, 94 (1), 209-19.

72. Huang, J.; Kaul, G.; Utz, J.; Hernandez, P.; Wong, V.; Bradley, D.; Nagi, A.;

O'Grady, D., A PAT approach to improve process understanding of high shear wet

granulation through in-line particle measurement using FBRM C35. J. Pharm. Sci. 2010,

99 (7), 3205-12.

73. Davis, T. D.; Morris, K. R.; Huang, H.; Peck, G. E.; Stowell, J. G.; Eisenhauer,

B. J.; Hilden, J. L.; Gibson, D.; Byrn, S. R., In situ monitoring of wet granulation using

online X-ray powder diffraction. Pharm. Res. 2003, 20 (11), 1851-7.

74. Arndt, O. R.; Baggio, R.; Adam, A. K.; Harting, J.; Franceschinis, E.;

Kleinebudde, P., Impact of Different Dry and Wet Granulation Techniques on Granule and

Tablet Properties: A Comparative Study. J. Pharm. Sci. 2018, 107 (12), 3143-3152.

75. Kleinebudde, P., Roll compaction/dry granulation: pharmaceutical applications.

Eur. J. Pharm. Biopharm. 2004, 58 (2), 317-26.

158

76. Nishii, K.; Horio, M., Chapter 6 Dry granulation. In Handbook of Powder

Technology, Salman, A. D.; Hounslow, M. J.; Seville, J. P. K., Eds. Elsevier Science B.V.:

2007; Vol. 11, pp 289-322.

77. Herting, M. G.; Kleinebudde, P., Roll compaction/dry granulation: effect of raw

material particle size on granule and tablet properties. Int. J. Pharm. 2007, 338 (1-2), 110-

8.

78. Mangal, H.; Kirsolak, M.; Kleinebudde, P., Roll compaction/dry granulation:

Suitability of different binders. Int. J. Pharm. 2016, 503 (1-2), 213-9.

79. Schiano, S.; Chen, L.; Wu, C.-Y., The effect of dry granulation on flow behaviour

of pharmaceutical powders during die filling. Powder Technol. 2018, 337, 78-83.

80. Sun, W. J.; Sun, C. C., Ribbon thickness influences fine generation during dry

granulation. Int. J. Pharm. 2017, 529 (1-2), 87-88.

81. Sun, W.-J.; Rantanen, J.; Sun, C. C., Ribbon density and milling parameters that

determine fines fraction in a dry granulation. Powder Technol. 2018, 338, 162-167.

82. Sun, C. C.; Kleinebudde, P., Mini review: Mechanisms to the loss of tabletability

by dry granulation. Eur. J. Pharm. Biopharm. 2016, 106, 9-14.

83. Vasanthavada, M.; Wang, Y.; Haefele, T.; Lakshman, J. P.; Mone, M.; Tong,

W.; Joshi, Y. M.; AT, M. S., Application of melt granulation technology using twin-screw

extruder in development of high-dose modified-release tablet formulation. J. Pharm. Sci.

2011, 100 (5), 1923-34.

84. McTaggart, C. M.; Ganley, J. A.; Sickmueller, A.; Walker, S. E., The evaluation

of formulation and processing conditions of a melt granulation process. Int. J. Pharm.

1984, 19 (2), 139-148.

159

85. Perissutti, B.; Rubessa, F.; Moneghini, M.; Voinovich, D., Formulation design of

carbamazepine fast-release tablets prepared by melt granulation technique. Int. J. Pharm.

2003, 256 (1-2), 53-63.

86. Desiraju, G. R., Crystal engineering: from molecule to crystal. J. Am. Chem. Soc.

2013, 135 (27), 9952-67.

87. Blagden, N.; de Matas, M.; Gavan, P. T.; York, P., Crystal engineering of active

pharmaceutical ingredients to improve solubility and dissolution rates. Adv. Drug Deliv.

Rev. 2007, 59 (7), 617-30.

88. York, P., Crystal Engineering and Particle Design for the Powder Compaction

Process. Drug Dev. Ind. Pharm. 2008, 18 (6-7), 677-721.

89. Mishra, M. K.; Ramamurty, U.; Desiraju, G. R., Mechanical property design of

molecular solids. Curr. Opin. Solid State Mater. Sci. 2016, 20 (6), 361-370.

90. Varughese, S.; Kiran, M. S.; Ramamurty, U.; Desiraju, G. R., Nanoindentation in

crystal engineering: quantifying mechanical properties of molecular crystals. Angew.

Chem. Int. Ed. 2013, 52 (10), 2701-12.

91. Panda, M. K.; Ghosh, S.; Yasuda, N.; Moriwaki, T.; Mukherjee, G. D.; Reddy,

C. M.; Naumov, P., Spatially resolved analysis of short-range structure perturbations in a

plastically bent molecular crystal. Nat. Chem. 2015, 7 (1), 65-72.

92. Reddy, C. M.; Krishna, G. R.; Ghosh, S., Mechanical properties of molecular

crystals-applications to crystal engineering. CrystEngComm 2010, 12 (8), 2296-2314.

93. Nangia, A. K.; Desiraju, G. R., Crystal Engineering: An Outlook for the Future.

Angew. Chem. Int. Ed. 2019, 58 (13), 4100-4107.

160

94. Krishna, G. R.; Devarapalli, R.; Lal, G.; Reddy, C. M., Mechanically Flexible

Organic Crystals Achieved by Introducing Weak Interactions in Structure: Supramolecular

Shape Synthons. J. Am. Chem. Soc. 2016, 138 (41), 13561-13567.

95. Steed, J. W., The role of co-crystals in pharmaceutical design. Trends Pharmacol

Sci 2013, 34 (3), 185-93.

96. Sun, C. C., Cocrystallization for successful drug delivery. Expert Opin. Drug Deliv.

2013, 10 (2), 201-13.

97. Chen, J.-M.; Wang, Z.-Z.; Wu, C.-B.; Li, S.; Lu, T.-B., Crystal engineering

approach to improve the solubility of mebendazole. CrystEngComm 2012, 14 (19).

98. Martin, F. A.; Pop, M. M.; Borodi, G.; Filip, X.; Kacso, I., Ketoconazole Salt and

Co-crystals with Enhanced Aqueous Solubility. Cryst. Growth Des. 2013, 13 (10), 4295-

4304.

99. Kuminek, G.; Cavanagh, K. L.; da Piedade, M. F. M.; Rodríguez-Hornedo, N.,

Posaconazole Cocrystal with Superior Solubility and Dissolution Behavior. Cryst. Growth

Des. 2019, 19 (11), 6592-6602.

100. Shefter, E.; Higuchi, T., Dissolution Behavior of Crystalline Solvated and

Nonsolvated Forms of Some Pharmaceuticals. J. Pharm. Sci. 1963, 52 (8), 781-91.

101. Henwood, S. Q.; Liebenberg, W.; Tiedt, L. R.; Lotter, A. P.; de Villiers, M. M.,

Characterization of the solubility and dissolution properties of several new rifampicin

polymorphs, solvates, and hydrates. Drug Dev. Ind. Pharm. 2001, 27 (10), 1017-30.

102. Thakuria, R.; Nangia, A., Olanzapinium Salts, Isostructural Solvates, and Their

Physicochemical Properties. Cryst. Growth Des. 2013, 13 (8), 3672-3680.

161

103. Salole, E. G.; Al-Sarraj, F. A., Spironolactone Crystal Forms. Drug Dev. Ind.

Pharm. 2008, 11 (4), 855-864.

104. Zanolla, D.; Perissutti, B.; Passerini, N.; Chierotti, M. R.; Hasa, D.; Voinovich,

D.; Gigli, L.; Demitri, N.; Geremia, S.; Keiser, J.; Cerreia Vioglio, P.; Albertini, B., A

new soluble and bioactive polymorph of praziquantel. Eur. J. Pharm. Biopharm. 2018,

127, 19-28.

105. Pudipeddi, M.; Serajuddin, A. T., Trends in solubility of polymorphs. J. Pharm.

Sci. 2005, 94 (5), 929-39.

106. Gopi, S. P.; Ganguly, S.; Desiraju, G. R., A Drug-Drug Salt Hydrate of Norfloxacin

and Sulfathiazole: Enhancement of in Vitro Biological Properties via Improved

Physicochemical Properties. Mol. Pharm. 2016, 13 (10), 3590-3594.

107. Serajuddin, A. T., Salt formation to improve drug solubility. Adv. Drug Deliv. Rev.

2007, 59 (7), 603-16.

108. Suresh, K.; Nangia, A., Lornoxicam Salts: Crystal Structures, Conformations, and

Solubility. Cryst. Growth Des. 2014, 14 (6), 2945-2953.

109. Trask, A. V.; Motherwell, W. D. S.; Jones, W., Pharmaceutical Cocrystallization:

Engineering a Remedy for Caffeine Hydration. Cryst. Growth Des. 2005, 5 (3), 1013-1021.

110. Perumalla, S. R.; Sun, C. C., Improved solid-state stability of salts by

cocrystallization between conjugate acid-base pairs. CrystEngComm 2013, 15 (29), 5756-

5759.

111. Chattoraj, S.; Shi, L. M.; Sun, C. C., Understanding the relationship between

crystal structure, plasticity and compaction behaviour of theophylline, methyl gallate, and

their 1: 1 co-crystal. CrystEngComm 2010, 12 (8), 2466-2472.

162

112. Chow, S. F.; Chen, M.; Shi, L.; Chow, A. H.; Sun, C. C., Simultaneously

improving the mechanical properties, dissolution performance, and hygroscopicity of

ibuprofen and flurbiprofen by cocrystallization with nicotinamide. Pharm. Res. 2012, 29

(7), 1854-65.

113. Perumalla, S. R.; Sun, C. C., Enabling tablet product development of 5-

fluorocytosine through integrated crystal and particle engineering. J. Pharm. Sci. 2014, 103

(4), 1126-32.

114. Karki, S.; FrisÌcÌicÌ, T.; FÃ¡biÃ¡n, L. s.; Laity, P. R.; Day, G. M.; Jones, W.,

Improving Mechanical Properties of Crystalline Solids by Cocrystal Formation: New

Compressible Forms of Paracetamol. Advanced Materials 2009, 21 (38â�“39), 3905-

3909.

115. Paul, S.; Taylor, L. J.; Murphy, B.; Krzyzaniak, J.; Dawson, N.; Mullarney, M.

P.; Meenan, P.; Sun, C. C., Mechanism and Kinetics of Punch Sticking of Pharmaceuticals.

J. Pharm. Sci. 2017, 106 (1), 151-158.

116. Paul, S.; Wang, C.; Wang, K.; Sun, C. C., Reduced Punch Sticking Propensity of

Acesulfame by Salt Formation: Role of Crystal Mechanical Property and Surface

Chemistry. Mol. Pharm. 2019, 16 (6), 2700-2707.

117. Wang, C.; Paul, S.; Sun, D. J.; Nilsson Lill, S. O.; Sun, C. C., Mitigating Punch

Sticking Propensity of Celecoxib by Cocrystallization: An Integrated Computational and

Experimental Approach. Cryst. Growth Des. 2020, 20 (7), 4217-4223.

118. Clemett, D.; Goa, K. L., Celecoxib: a review of its use in osteoarthritis, rheumatoid

arthritis and acute pain. Drugs 2000, 59 (4), 957-80.

163

119. Pfizer Inc. Celebrex (celecoxib) [packge insert]. U. S. Food and Drug

Administration. Available at:

https://www.accessdata.fda.gov/drugsatfda_docs/label/1999/21156lbl.pdf. (accessed

August 10, 2020).

120. Fort, J., Celecoxib, a COX-2--specific inhibitor: the clinical data. Am J Orthop

(Belle Mead NJ) 1999, 28 (3 Suppl), 13-8.

121. Paulson, S. K.; Vaughn, M. B.; Jessen, S. M.; Lawal, Y.; Gresk, C. J.; Yan, B.;

Maziasz, T. J.; Cook, C. S.; Karim, A., Pharmacokinetics of celecoxib after oral

administration in dogs and humans: effect of food and site of absorption. J Pharmacol Exp

Ther 2001, 297 (2), 638-45.

122. Davies, N. M.; McLachlan, A. J.; Day, R. O.; Williams, K. M., Clinical

pharmacokinetics and pharmacodynamics of celecoxib: a selective cyclo-oxygenase-2

inhibitor. Clin Pharmacokinet 2000, 38 (3), 225-42.

123. Gupta, V. R.; Mutalik, S.; Patel, M. M.; Jani, G. K., Spherical crystals of celecoxib

to improve solubility, dissolution rate and micromeritic properties. Acta Pharm. 2007, 57

(2), 173-84.

124. Kwon, H. J.; Heo, E. J.; Kim, Y. H.; Kim, S.; Hwang, Y. H.; Byun, J. M.; Cheon,

S. H.; Park, S. Y.; Kim, D. Y.; Cho, K. H.; Maeng, H. J.; Jang, D. J., Development and

Evaluation of Poorly Water-Soluble Celecoxib as Solid Dispersions Containing Nonionic

Surfactants Using Fluidized-Bed Granulation. Pharmaceutics 2019, 11 (3).

125. Guzman, H. R.; Tawa, M.; Zhang, Z.; Ratanabanangkoon, P.; Shaw, P.; Gardner,

C. R.; Chen, H.; Moreau, J. P.; Almarsson, O.; Remenar, J. F., Combined use of

https://www.accessdata.fda.gov/drugsatfda_docs/label/1999/21156lbl.pdf

164

crystalline salt forms and precipitation inhibitors to improve oral absorption of celecoxib

from solid oral formulations. J. Pharm. Sci. 2007, 96 (10), 2686-702.

126. Ilevbare, G. A.; Liu, H.; Edgar, K. J.; Taylor, L. S., Maintaining Supersaturation

in Aqueous Drug Solutions: Impact of Different Polymers on Induction Times. Cryst.

Growth Des. 2012, 13 (2), 740-751.

127. Chen, J.; Ormes, J. D.; Higgins, J. D.; Taylor, L. S., Impact of surfactants on the

crystallization of aqueous suspensions of celecoxib amorphous solid dispersion spray dried

particles. Mol. Pharm. 2015, 12 (2), 533-41.

128. Lee, J. H.; Kim, M. J.; Yoon, H.; Shim, C. R.; Ko, H. A.; Cho, S. A.; Lee, D.;

Khang, G., Enhanced dissolution rate of celecoxib using PVP and/or HPMC-based solid

dispersions prepared by spray drying method. J. Pharm. Investig. 2013, 43 (3), 205-213.

129. Meng, F.; Ferreira, R.; Zhang, F., Effect of surfactant level on properties of

celecoxib amorphous solid dispersions. J. Drug. Deliv. Sci. Tec. 2019, 49, 301-307.

130. Fong, S. Y.; Ibisogly, A.; Bauer-Brandl, A., Solubility enhancement of BCS Class

II drug by solid phospholipid dispersions: Spray drying versus freeze-drying. Int. J. Pharm.

2015, 496 (2), 382-91.

131. Homayouni, A.; Sadeghi, F.; Varshosaz, J.; Afrasiabi Garekani, H.; Nokhodchi,

A., Promising dissolution enhancement effect of soluplus on crystallized celecoxib

obtained through antisolvent precipitation and high pressure homogenization techniques.

Colloids Surf B Biointerfaces 2014, 122, 591-600.

132. Andrews, G. P.; Abu-Diak, O.; Kusmanto, F.; Hornsby, P.; Hui, Z.; Jones, D. S.,

Physicochemical characterization and drug-release properties of celecoxib hot-melt

extruded glass solutions. J. Pharm. Pharmacol. 2010, 62 (11), 1580-90.

165

133. Chen, H.; Paul, S.; Xu, H.; Wang, K.; Mahanthappa, M. K.; Sun, C. C., Reduction

of Punch-Sticking Propensity of Celecoxib by Spherical Crystallization via Polymer

Assisted Quasi-Emulsion Solvent Diffusion. Mol. Pharm. 2020, 17 (4), 1387-1396.

134. Joshi, A. B.; Patel, S.; Kaushal, A. M.; Bansal, A. K., Compaction studies of

alternate solid forms of celecoxib. Adv. Powder Technol. 2010, 21 (4), 452-460.

135. Be Rzins, K. R.; Fraser-Miller, S. J.; Rades, T.; Gordon, K. C., Low-Frequency

Raman Spectroscopic Study on Compression-Induced Destabilization in Melt-Quenched

Amorphous Celecoxib. Mol. Pharm. 2019, 16 (8), 3678-3686.

136. Osei-Yeboah, F. Improving powder tableting performance through materials

engineering. University of Minnesota, 2015.

137. Saha, S.; Desiraju, G. R., Crystal Engineering of Hand-Twisted Helical Crystals. J.

Am. Chem. Soc. 2017, 139 (5), 1975-1983.

138. Saha, S.; Mishra, M. K.; Reddy, C. M.; Desiraju, G. R., From Molecules to

Interactions to Crystal Engineering: Mechanical Properties of Organic Solids. Acc. Chem.

Res. 2018, 51 (11), 2957-2967.

139. Naumov, P.; Chizhik, S.; Panda, M. K.; Nath, N. K.; Boldyreva, E., Mechanically

Responsive Molecular Crystals. Chem. Rev. 2015, 115 (22), 12440-90.

140. Garcia-Garibay, M. A., Molecular crystals on the move: from single-crystal-to-

single-crystal photoreactions to molecular machinery. Angew. Chem. Int. Ed. 2007, 46

(47), 8945-7.

141. Fratzl, P.; Barth, F. G., Biomaterial systems for mechanosensing and actuation.

Nature 2009, 462 (7272), 442-8.

166

142. Cao, Y.; Li, H., Engineered elastomeric proteins with dual elasticity can be

controlled by a molecular regulator. Nat. Nanotechnol. 2008, 3 (8), 512-6.

143. Keten, S.; Xu, Z.; Ihle, B.; Buehler, M. J., Nanoconfinement controls stiffness,

strength and mechanical toughness of beta-sheet crystals in silk. Nat. Mater. 2010, 9 (4),

359-67.

144. Bronstein, H.; Chen, Z.; Ashraf, R. S.; Zhang, W.; Du, J.; Durrant, J. R.;

Tuladhar, P. S.; Song, K.; Watkins, S. E.; Geerts, Y.; Wienk, M. M.; Janssen, R. A.;

Anthopoulos, T.; Sirringhaus, H.; Heeney, M.; McCulloch, I., Thieno[3,2-b]thiophene-

diketopyrrolopyrrole-containing polymers for high-performance organic field-effect

transistors and organic photovoltaic devices. J. Am. Chem. Soc. 2011, 133 (10), 3272-5.

145. Schmidt-Mende, L.; Fechtenkotter, A.; Mullen, K.; Moons, E.; Friend, R. H.;

MacKenzie, J. D., Self-organized discotic liquid crystals for high-efficiency organic

photovoltaics. Science 2001, 293 (5532), 1119-22.

146. Terao, F.; Morimoto, M.; Irie, M., Light-driven molecular-crystal actuators: rapid

and reversible bending of rodlike mixed crystals of diarylethene derivatives. Angew. Chem.

Int. Ed. 2012, 51 (4), 901-4.

147. Worthy, A.; Grosjean, A.; Pfrunder, M. C.; Xu, Y.; Yan, C.; Edwards, G.; Clegg,

J. K.; McMurtrie, J. C., Atomic resolution of structural changes in elastic crystals of

copper(II) acetylacetonate. Nat. Chem. 2018, 10 (1), 65-69.

148. Brock, A. J.; Whittaker, J. J.; Powell, J. A.; Pfrunder, M. C.; Grosjean, A.;

Parsons, S.; McMurtrie, J. C.; Clegg, J. K., Elastically Flexible Crystals have Disparate

Mechanisms of Molecular Movement Induced by Strain and Heat. Angew. Chem. Int. Ed.

2018, 57 (35), 11325-11328.

167

149. Desiraju, G. R., Crystal Engineering: From Molecule to Crystal. J. Am. Chem. Soc.

2013, 135 (27), 9952-9967.

150. Babu, N. J.; Nangia, A., Solubility Advantage of Amorphous Drugs and

Pharmaceutical Cocrystals. Cryst. Growth Des. 2011, 11 (7), 2662-2679.

151. Sun, C. C., Decoding Powder Tabletability: Roles of Particle Adhesion and

Plasticity. Journal of Adhesion Science and Technology 2011, 25 (4-5), 483-499.

152. Smith, A. J.; Kavuru, P.; Wojtas, L.; Zaworotko, M. J.; Shytle, R. D., Cocrystals

of quercetin with improved solubility and oral bioavailability. Mol. Pharm. 2011, 8 (5),

1867-76.

153. Wang, C. G.; Paul, S.; Wang, K. L.; Hu, S. Y.; Sun, C. Q. C., Relationships among

Crystal Structures, Mechanical Properties, and Tableting Performance Probed Using Four

Salts of Diphenhydramine. Cryst. Growth Des. 2017, 17 (11), 6030-6040.

154. Mishra, M. K.; Ramamurty, U.; Desiraju, G. R., Solid solution hardening of

molecular crystals: tautomeric polymorphs of omeprazole. J. Am. Chem. Soc. 2015, 137

(5), 1794-7.

155. Mishra, M. K.; Desiraju, G. R.; Ramamurty, U.; Bond, A. D., Studying

microstructure in molecular crystals with nanoindentation: intergrowth polymorphism in

felodipine. Angew. Chem. Int. Ed. 2014, 53 (48), 13102-5.

156. Sanphui, P.; Mishra, M. K.; Ramamurty, U.; Desiraju, G. R., Tuning mechanical

properties of pharmaceutical crystals with multicomponent crystals: voriconazole as a case

study. Mol. Pharm. 2015, 12 (3), 889-97.

157. Ghosh, S.; Reddy, C. M., Elastic and bendable caffeine cocrystals: implications for

the design of flexible organic materials. Angew. Chem. Int. Ed. 2012, 51 (41), 10319-23.

168

158. Ghosh, S.; Mishra, M. K.; Kadambi, S. B.; Ramamurty, U.; Desiraju, G. R.,

Designing elastic organic crystals: highly flexible polyhalogenated N-benzylideneanilines.

Angew. Chem. Int. Ed. 2015, 54 (9), 2674-8.

159. Saha, S.; Desiraju, G. R., Using structural modularity in cocrystals to engineer

properties: elasticity. Chem. Commun. 2016, 52 (49), 7676-9.

160. Mukherjee, A.; Desiraju, G. R., Halogen bonds in some dihalogenated phenols:

applications to crystal engineering. IUCrJ 2014, 1 (Pt 1), 49-60.

161. Takamizawa, S.; Miyamoto, Y., Superelastic organic crystals. Angew. Chem. Int.

Ed. 2014, 53 (27), 6970-3.

162. Ghosh, S.; Mishra, M. K.; Ganguly, S.; Desiraju, G. R., Dual Stress and Thermally

Driven Mechanical Properties of the Same Organic Crystal: 2,6-Dichlorobenzylidene-4-

fluoro-3-nitroaniline. J. Am. Chem. Soc. 2015, 137 (31), 9912-21.

163. Ahmed, E.; Karothu, D. P.; Naumov, P., Crystal Adaptronics: Mechanically

Reconfigurable Elastic and Superelastic Molecular Crystals. Angew. Chem. Int. Ed. 2018,

57 (29), 8837-8846.

164. Saini, A. K.; Natarajan, K.; Mobin, S. M., A new multitalented azine ligand: elastic

bending, single-crystal-to-single-crystal transformation and a fluorescence turn-on Al(iii)

sensor. Chem. Commun. 2017, 53 (71), 9870-9873.

165. Mir, S. H.; Takasaki, Y.; Engel, E. R.; Takamizawa, S., Ferroelasticity in an

Organic Crystal: A Macroscopic and Molecular Level Study. Angew. Chem. Int. Ed. 2017,

56 (50), 15882-15885.

169

166. Mishra, M. K.; Kadambi, S. B.; Ramamurty, U.; Ghosh, S., Elastic flexibility

tuning via interaction factor modulation in molecular crystals. Chem. Commun. 2018, 54

(65), 9047-9050.

167. Keswani, R. K.; Yoon, G. S.; Sud, S.; Stringer, K. A.; Rosania, G. R., A far-red

fluorescent probe for flow cytometry and image-based functional studies of xenobiotic

sequestering macrophages. Cytometry A 2015, 87 (9), 855-67.

168. Hayashi, S.; Asano, A.; Kamiya, N.; Yokomori, Y.; Maeda, T.; Koizumi, T.,

Fluorescent organic single crystals with elastic bending flexibility: 1,4-bis(thien-2-yl)-

2,3,5,6-tetrafluorobenzene derivatives. Sci Rep 2017, 7 (1), 9453.

169. Sitti, M.; Ceylan, H.; Hu, W.; Giltinan, J.; Turan, M.; Yim, S.; Diller, E.,

Biomedical Applications of Untethered Mobile Milli/Microrobots. Proc IEEE Inst Electr

Electron Eng 2015, 103 (2), 205-224.

170. Horstman, E. M.; Keswani, R. K.; Frey, B. A.; Rzeczycki, P. M.; LaLone, V.;

Bertke, J. A.; Kenis, P. J.; Rosania, G. R., Elasticity in Macrophage-Synthesized

Biocrystals. Angew. Chem. Int. Ed. 2017, 56 (7), 1815-1819.

171. Gupta, P.; Karothu, D. P.; Ahmed, E.; Naumov, P.; Nath, N. K., Thermally

Twistable, Photobendable, Elastically Deformable, and Self-Healable Soft Crystals.

Angew. Chem. Int. Ed. 2018, 57 (28), 8498-8502.

172. Vasu Dev, R.; Shashi Rekha, K.; Vyas, K.; Mohanti, S. B.; Rajender Kumar, P.;

Om Reddy, G., Celecoxib, a COX-II inhibitor. Acta Crystallogr. C: Cryst. Struct. Commun.

1999, 55 (12), IUC9900161.

173. Modi, S. R.; Dantuluri, A. K. R.; Puri, V.; Pawar, Y. B.; Nandekar, P.;

Sangamwar, A. T.; Perumalla, S. R.; Sun, C. C.; Bansal, A. K., Impact of Crystal Habit

170

on Biopharmaceutical Performance of Celecoxib. Cryst. Growth Des. 2013, 13 (7), 2824-

2832.

174. Hübschle, C. B.; Sheldrick, G. M.; Dittrich, B., ShelXle: a Qt graphical user

interface for SHELXL. J. Appl. Crystallogr. 2011, 44 (Pt 6), 1281-1284.

175. Rasband, W. S. Image J, U. S. National Institutes of Health, Bethesda, Maryland,

USA, 1997-2018.

176. Gere, J. M., Goodno, B. J., Mechanics of Materials. 8th ed.; Cengage Learning:

Stamford, CT, 2013.

177. Oliver, W. C.; Pharr, G. M., An improved technique for determining hardness and

elastic modulus using load and displacement sensing indentation experiments. J. Mater.

Res. 2011, 7 (6), 1564-1583.

178. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.;

Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji,

H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.;

Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida,

M.; Nakajima, T.; Honda , Y.; Kitao, O.; Nakai, H.; Vreven, T.; Jr. Montgomery, J. A.;

Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.;

Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant,

J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox,

J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R.

E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;

Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.;

171

Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.;

Fox, D. J. Gaussian 09, Revision E.01; Gaussian, Inc.: Wallingford, CT, 2009.

179. O. A. Bauchau, J. I. C., Structural Analysis. Springer: Dordrecht, 2009; p 173-221.

180. Timoshenko, S. P., History of Strength of Materials. McGrawHill: New York,

1953.

181. Telford, M., The case for bulk metallic glass. Mater. Today 2004, 7 (3), 36-43.

182. Varughese, S.; Kiran, M.; Ramamurty, U.; Desiraju, G., Nanoindentation in

Crystal Engineering: Quantifying Mechanical Properties of Molecular Crystals.

Angewandte Chemie-International Edition 2013, 52 (10), 2701-2712.

183. Mishra, M. K.; Varughese, S.; Ramamurty, U.; Desiraju, G. R., Odd-even effect

in the elastic modulii of alpha,omega-alkanedicarboxylic acids. J. Am. Chem. Soc. 2013,

135 (22), 8121-4.

184. Krishna, G. R.; Shi, L. M.; Bag, P. P.; Sun, C. C.; Reddy, C. M., Correlation

Among Crystal Structure, Mechanical Behavior, and Tabletability in the Co-Crystals of

Vanillin Isomers. Cryst. Growth Des. 2015, 15 (4), 1827-1832.

185. Liu, F.; Hooks, D. E.; Li, N.; Mara, N. A.; Swift, J. A., Mechanical Properties of

Anhydrous and Hydrated Uric Acid Crystals. Chem. Mater. 2018, 30 (11), 3798-3805.

186. Mishra, M. K.; Mishra, K.; Syed Asif, S. A.; Manimunda, P., Structural analysis

of elastically bent organic crystals using in situ indentation and micro-Raman

spectroscopy. Chem. Commun. 2017, 53 (97), 13035-13038.

187. Ewen Smith, G. D., Modern Raman Spectroscopy – A Practical Approach. John

Wiley & Sons, Ltd: West Sussex, England, 2005.

172

188. Cianchetti, M.; Laschi, C.; Menciassi, A.; Dario, P., Biomedical applications of

soft robotics. 2018; Vol. 3.

189. Chattoraj, S.; Sun, C. C., Crystal and Particle Engineering Strategies for Improving

Powder Compression and Flow Properties to Enable Continuous Tablet Manufacturing by

Direct Compression. J. Pharm. Sci. 2018, 107 (4), 968-974.

190. Williams, H. D.; Trevaskis, N. L.; Charman, S. A.; Shanker, R. M.; Charman,

W. N.; Pouton, C. W.; Porter, C. J., Strategies to address low drug solubility in discovery

and development. Pharmacol Rev 2013, 65 (1), 315-499.

191. Sun, C.; Grant, D. J., Improved tableting properties of p-hydroxybenzoic acid by

water of crystallization: a molecular insight. Pharm. Res. 2004, 21 (2), 382-6.

192. Hilfiker, R.; Raumer, M. v.; Griesser, U. J., The Importance of Solvates. In

Polymorphism in the Pharmaceutical Industry, 2nd ed.; Hilfiker, R.; Raumer, M. v., Eds.

Wiley-VCH: Weinheim, 2018.

193. Q3C Impurities: Residual Solvents: Maintenance Procedures for the Guidance for

Industry Q3C. International Council for Harmonisation-Quality.2018

https://www.fda.gov/drugs/guidances-drugs/international-council-harmonisation-quality

Accessed June 11, 2020

194. Zhang, C.; Kersten, K. M.; Kampf, J. W.; Matzger, A. J., Solid-State Insight Into

the Action of a Pharmaceutical Solvate: Structural, Thermal, and Dissolution Analysis of

Indinavir Sulfate Ethanolate. J. Pharm. Sci. 2018, 107 (10), 2731-2734.

195. Kerschgens, I.; Lefranc, J., Early Drug Development — Bringing a Preclinical

Candidate to the Clinic. Edited by Fabrizio Giordanetto. ChemMedChem 2019, 14 (12),

1224-1225.

https://www.fda.gov/drugs/guidances-drugs/international-council-harmonisation-quality

173

196. Lu, G. W.; Hawley, M.; Smith, M.; Geiger, B. M.; Pfund, W., Characterization

of a novel polymorphic form of celecoxib. J. Pharm. Sci. 2006, 95 (2), 305-17.

197. Remenar, J. F.; Tawa, M. D.; Peterson, M. L.; Almarsson, O.; Hickey, M. B.;

Foxman, B. M., Celecoxib sodium salt: engineering crystal forms for performance.

CrystEngComm 2011, 13 (4), 1081-1089.

198. Bond, A. D.; Sun, C. C., Intermolecular interactions and disorder in six isostructural

celecoxib solvates. Acta Crystallogr. C 2020, 76 (7), 632-638.

199. Bolla, G.; Mittapalli, S.; Nangia, A., Celecoxib cocrystal polymorphs with cyclic

amides: synthons of a sulfonamide drug with carboxamide coformers. CrystEngComm

2014, 16 (1), 24-27.

200. Remenar, J. F.; Peterson, M. L.; Stephens, P. W.; Zhang, Z.; Zimenkov, Y.;

Hickey, M. B., Celecoxib:nicotinamide dissociation: using excipients to capture the

cocrystal's potential. Mol. Pharm. 2007, 4 (3), 386-400.

201. Zhang, S.-W.; Brunskill, A. P. J.; Schwartz, E.; Sun, S., Celecoxib–Nicotinamide

Cocrystal Revisited: Can Entropy Control Cocrystal Formation? Cryst. Growth Des. 2017,

17 (5), 2836-2843.

202. Wang, X. J.; Zhang, Q.; Jiang, L. L.; Xu, Y.; Mei, X. F., Isostructurality in six

celecoxib co-crystals introduced by solvent inclusion. CrystEngComm 2014, 16 (48),

10959-10968.

203. Salaman, P.; Ramon, C.; Tesson, N., Co-crystals of celecoxib and L-proline. EP

2325172 A1 2011.

204. Almarsson, Ö.; Hickey, M. B.; Peterson, M.; J, Z. M.; Moulton, B.; Rodriguez-

Hornedo, N., Pharmaceutical co-crystal compositions of drugs such as carbamazepine,

174

celecoxib, olanzapine, itraconazole, topiramate, modafinil, 5-fluorouracil,

hydrochlorothiazide, acetaminophen, aspirin, flurbiprofen, phenytoin and ibuprofen. US

2007/0059356 A1 2007.

205. Salaman, P.; Ramon, C.; Sebastia, V. C.; Tesson, N.; Trilla Castano, M. Co-

crystals of Venlafaxine and Celecoxib. 2011.

206. Almansa, C.; Merce, R.; Tesson, N.; Farran, J.; Tomas, J.; Plata-Salaman, C. R.,

Co-crystal of Tramadol Hydrochloride-Celecoxib (ctc): A Novel API-API Co-crystal for

the Treatment of Pain. Cryst. Growth Des. 2017, 17 (4), 1884-1892.

207. Paul, S.; Wang, K.; Taylor, L. J.; Murphy, B.; Krzyzaniak, J.; Dawson, N.;

Mullarney, M. P.; Meenan, P.; Sun, C. C., Dependence of Punch Sticking on Compaction

Pressure-Roles of Particle Deformability and Tablet Tensile Strength. J. Pharm. Sci. 2017,

106 (8), 2060-2067.

208. Paul, S.; Sun, C. C., Modulating Sticking Propensity of Pharmaceuticals Through

Excipient Selection in a Direct Compression Tablet Formulation. Pharm. Res. 2018, 35

(6), 113.

209. Hubschle, C. B.; Sheldrick, G. M.; Dittrich, B., ShelXle: a Qt graphical user

interface for SHELXL. Journal of Applied Crystallography 2011, 44 (6), 1281-1284.

210. Turner, M. J.; Wolff, S. K.; Grimwood, D. J.; Spackman, P. R.; Jayatilaka, D.;

Spackman, M. A. CrystalExplorer17, University of Western Australia: 2017.

211. Turner, M. J.; Thomas, S. P.; Shi, M. W.; Jayatilaka, D.; Spackman, M. A., Energy

frameworks: insights into interaction anisotropy and the mechanical properties of

molecular crystals. Chem. Commun. 2015, 51 (18), 3735-8.

175

212. Bryant, M. J.; Maloney, A. G. P.; Sykes, R. A., Predicting mechanical properties

of crystalline materials through topological analysis. CrystEngComm 2018, 20 (19), 2698-

2704.

213. Mohan, R.; Lorenz, H.; Myerson, A. S., Solubility Measurement Using Differential

Scanning Calorimetry. Ind. Eng. Chem. Res. 2002, 41 (19), 4854-4862.

214. Timoumi, S.; Mangin, D.; Peczalski, R.; Zagrouba, F.; Andrieu, J., Stability and

thermophysical properties of azithromycin dihydrate. Arab. J. Chem. 2014, 7 (2), 189-195.

215. Gillon, A. L.; Davey, R. J.; Storey, R.; Feeder, N.; Nichols, G.; Dent, G.;

Apperley, D. C., Solid state dehydration processes: mechanism of water loss from

crystalline inosine dihydrate. J. Phys. Chem. B 2005, 109 (11), 5341-7.

216. Wang, K. L.; Mishra, M. K.; Sun, C. C., Exceptionally Elastic Single-Component

Pharmaceutical Crystals. Chem. Mater. 2019, 31 (5), 1794-1799.

217. Duggirala, N. K.; Perry, M. L.; Almarsson, O.; Zaworotko, M. J., Pharmaceutical

cocrystals: along the path to improved medicines. Chem. Commun. 2016, 52 (4), 640-55.

218. Almarsson, O.; Zaworotko, M. J., Crystal engineering of the composition of

pharmaceutical phases. Do pharmaceutical co-crystals represent a new path to improved

medicines? Chem. Commun. 2004, (17), 1889-96.

219. Cherukuvada, S.; Kaur, R.; Guru Row, T. N., Co-crystallization and small

molecule crystal form diversity: from pharmaceutical to materials applications.

CrystEngComm 2016, 18 (44), 8528-8555.

220. Dai, X.-L.; Chen, J.-M.; Lu, T.-B., Pharmaceutical cocrystallization: an effective

approach to modulate the physicochemical properties of solid-state drugs. CrystEngComm

2018, 20 (36), 5292-5316.

176

221. Hilfiker, R.; Raumer, M. v., Polymorphism in the Pharmaceutical Industry. 2nd

ed.; Wiley-VCH: Weinheim, 2018.

222. Healy, A. M.; Worku, Z. A.; Kumar, D.; Madi, A. M., Pharmaceutical solvates,

hydrates and amorphous forms: A special emphasis on cocrystals. Adv. Drug Deliv. Rev.

2017, 117, 25-46.

223. ICH, ICH harmonized guideline. Impurities: Guideline for Residual Solvents Q3C

(R6). Use, I. C. F. H. o. T. R. f. P. f. H., Ed. International Council For Harmonisation of

Technical Requirements for Pharmaceuticals for Human Use: 2016.

224. Carvajal-Nunez, U.; Elbakhshwan, M. S.; Mara, N. A.; White, J. T.; Nelson, A.

T., Mechanical Properties of Uranium Silicides by Nanoindentation and Finite Elements

Modeling. JOM 2017, 70 (2), 203-208.

225. Fell, J. T.; Newton, J. M., Determination of tablet strength by the diametral-

compression test. J. Pharm. Sci. 1970, 59 (5), 688-91.

226. Sun, C. C., True density of microcrystalline cellulose. J. Pharm. Sci. 2005, 94 (10),

2132-4.

227. Paul, S.; Sun, C. C., The suitability of common compressibility equations for

characterizing plasticity of diverse powders. Int. J. Pharm. 2017, 532 (1), 124-130.

228. Sun, C. C., A novel method for deriving true density of pharmaceutical solids

including hydrates and water-containing powders. J. Pharm. Sci. 2004, 93 (3), 646-53.

229. Kuentz, M.; Leuenberger, H., Pressure susceptibility of polymer tablets as a critical

property: a modified Heckel equation. J. Pharm. Sci. 1999, 88 (2), 174-9.

230. Ryshkewitch, E., Compression Strength of Porous Sintered Alumina and Zirconia.

J. Am. Ceram. Soc. 1953, 36 (2), 65-68.

177

231. Chang, S. Y.; Sun, C. C., Superior Plasticity and Tabletability of Theophylline

Monohydrate. Mol. Pharm. 2017, 14 (6), 2047-2055.

232. Dissolution Methods Database. Division of Bioequivalence, Office of Generic

Drugs, Center for Drug Evaluation and Research, U.S. Food and Drug Administration.

2019. Available at: https://www.accessdata.fda.gov/scripts/CDER/dissolution/index.cfm.

(accessed May 5, 2020).

233. Han, J.; Gupte, S.; Suryanarayanan, R., Applications of pressure differential

scanning calorimetry in the study of pharmaceutical hydrates. II. Ampicillin trihydrate. Int.

J. Pharm. 1998, 170 (1), 63-72.

234. Machiste, E., Characterization of carbamazepine in systems containing a

dissolution rate enhancer. Int. J. Pharm. 1995, 126 (1-2), 65-72.

235. Aitipamula, S.; Chow, P. S.; Tan, R. B., Solvates of the antifungal drug

griseofulvin: structural, thermochemical and conformational analysis. Acta Crystallogr. B

Struct. Sci. Cryst. Eng. Mater. 2014, 70 (Pt 1), 54-62.

236. Zhu, B.; Fang, X.; Zhang, Q.; Mei, X.; Ren, G., Study of Crystal Structures,

Properties, and Form Transformations among a Polymorph, Hydrates, and Solvates of

Apatinib. Cryst. Growth Des. 2019, 19 (5), 3060-3069.

237. Taylor, L. J.; Papadopoulos, D. G.; Dunn, P. J.; Bentham, A. C.; Dawson, N. J.;

Mitchell, J. C.; Snowden, M. J., Predictive milling of pharmaceutical materials using

nanoindentation of single crystals. Org. Process Res. Dev. 2004, 8 (4), 674-679.

238. Shi, L.; Sun, C. C., Overcoming poor tabletability of pharmaceutical crystals by

surface modification. Pharm. Res. 2011, 28 (12), 3248-55.

https://www.accessdata.fda.gov/scripts/CDER/dissolution/index.cfm

178

239. Wang, C.; Sun, C. C., The landscape of mechanical properties of molecular crystals.

CrystEngComm 2020, 22 (7), 1149-1153.

240. Thoorens, G.; Krier, F.; Leclercq, B.; Carlin, B.; Evrard, B., Microcrystalline

cellulose, a direct compression binder in a quality by design environment--a review. Int. J.

Pharm. 2014, 473 (1-2), 64-72.

241. Sun, C. C., Setting the bar for powder flow properties in successful high speed

tableting. Powder Technol. 2010, 201 (1), 106-108.

242. Gordon, M. S.; Rudraraju, V. S.; Dani, K.; Chowhan, Z. T., Effect of the mode of

super disintegrant incorporation on dissolution in wet granulated tablets. J. Pharm. Sci.

1993, 82 (2), 220-6.

243. Dun, J.; Osei-Yeboah, F.; Boulas, P.; Lin, Y.; Sun, C. C., A systematic evaluation

of dual functionality of sodium lauryl sulfate as a tablet lubricant and wetting enhancer.

Int. J. Pharm. 2018, 552 (1-2), 139-147.

244. Ertel, K. D.; Carstensen, J. T., Chemical, physical, and lubricant properties of

magnesium stearate. J. Pharm. Sci. 1988, 77 (7), 625-9.

245. Paul, S.; Sun, C. C., Gaining insight into tablet capping tendency from compaction

simulation. Int. J. Pharm. 2017, 524 (1-2), 111-120.

246. Garr, J. S. M.; Rubinstein, M. H., An investigation into the capping of paracetamol

at increasing speeds of compression. Int. J. Pharm. 1991, 72 (2), 117-122.

247. Chen, H.; Aburub, A.; Sun, C. C., Direct Compression Tablet Containing 99%

Active Ingredient-A Tale of Spherical Crystallization. J. Pharm. Sci. 2019, 108 (4), 1396-

1400.

179

248. United States Pharmacopeia and National Formulary. Tablet Friability. United

States Pharmacopeial Convention.Rockville, MD 2016

https://www.usp.org/sites/default/files/usp/document/harmonization/gen-

chapter/g06_pf_ira_32_2_2006.pdf Accessed August 19, 2020

249. Yamashita, H.; Sun, C. C., Expedited Tablet Formulation Development of a Highly

Soluble Carbamazepine Cocrystal Enabled by Precipitation Inhibition in Diffusion Layer.

Pharm. Res. 2019, 36 (6), 90.

250. Wang, C.; Hu, S.; Sun, C. C., Expedited development of a high dose orally

disintegrating metformin tablet enabled by sweet salt formation with acesulfame. Int. J.

Pharm. 2017, 532 (1), 435-443.

251. Yamashita, H.; Sun, C. C., Material-Sparing and Expedited Development of a

Tablet Formulation of Carbamazepine Glutaric Acid Cocrystal- a QbD Approach. Pharm.

Res. 2020, 37 (8), 153.

252. Amidon, G. L.; Lennernäs, H.; Shah, V. P.; Crison, J. R., A Theoretical Basis for

a Biopharmaceutic Drug Classification: The Correlation of in Vitro Drug Product

Dissolution and in Vivo Bioavailability. Pharm. Res. 1995, 12 (3), 413-420.

253. Williams, H. D.; Trevaskis, N. L.; Charman, S. A.; Shanker, R. M.; Charman,

W. N.; Pouton, C. W.; Porter, C. J., Strategies to address low drug solubility in discovery

and development. Pharmacol. Rev. 2013, 65 (1), 315-499.

254. Sugandha, K.; Kaity, S.; Mukherjee, S.; Isaac, J.; Ghosh, A., Solubility

Enhancement of Ezetimibe by a Cocrystal Engineering Technique. Cryst. Growth. Des.

2014, 14 (9), 4475-4486.

https://www.usp.org/sites/default/files/usp/document/harmonization/gen-chapter/g06_pf_ira_32_2_2006.pdf

https://www.usp.org/sites/default/files/usp/document/harmonization/gen-chapter/g06_pf_ira_32_2_2006.pdf

180

255. Hammond, R. B.; Pencheva, K.; Roberts, K. J.; Auffret, T., Quantifying solubility

enhancement due to particle size reduction and crystal habit modification: case study of

acetyl salicylic acid. J. Pharm. Sci. 2007, 96 (8), 1967-73.

256. Kamath, A. V.; Chang, M.; Lee, F. Y.; Zhang, Y.; Marathe, P. H., Preclinical

pharmacokinetics and oral bioavailability of BMS-310705, a novel epothilone B analog.

Cancer Chemother. Pharmacol. 2005, 56 (2), 145-53.

257. Rawat, S.; Jain, S. K., Solubility enhancement of celecoxib using beta-cyclodextrin

inclusion complexes. Eur. J. Pharm. Biopharm. 2004, 57 (2), 263-7.

258. Hancock, B. C.; Parks, M., What is the True Solubility Advantage for Amorphous

Pharmaceuticals? Pharm. Res. 2000, 17 (4), 397-404.

259. Law, D.; Schmitt, E. A.; Marsh, K. C.; Everitt, E. A.; Wang, W.; Fort, J. J.;

Krill, S. L.; Qiu, Y., Ritonavir–PEG 8000Amorphous Solid Dispersions: In vitro and In

vivo Evaluations. J. Pharm. Sci. 2004, 93 (3), 563-570.

260. Hancock, B. C.; Zografi, G., Characteristics and significance of the amorphous state

in pharmaceutical systems. J. Pharm. Sci. 1997, 86 (1), 1-12.

261. Wu, T.; Yu, L., Surface crystallization of indomethacin below Tg. Pharm. Res.

2006, 23 (10), 2350-5.

262. Zhu, L.; Wong, L.; Yu, L., Surface-enhanced crystallization of amorphous

nifedipine. Mol. Pharm. 2008, 5 (6), 921-6.

263. Zhu, L.; Jona, J.; Nagapudi, K.; Wu, T., Fast surface crystallization of amorphous

griseofulvin below T g. Pharm. Res. 2010, 27 (8), 1558-67.

181

264. Kothari, K.; Ragoonanan, V.; Suryanarayanan, R., Influence of molecular mobility

on the physical stability of amorphous pharmaceuticals in the supercooled and glassy

States. Mol. Pharm. 2014, 11 (9), 3048-55.

265. Ediger, M. D.; Harrowell, P.; Yu, L., Crystal growth kinetics exhibit a fragility-

dependent decoupling from viscosity. J. Chem. Phys. 2008, 128 (3), 034709.

266. Debenedetti, P. G.; Stillinger, F. H., Supercooled liquids and the glass transition.

Nature 2001, 410, 259.

267. Greet, R. J.; Turnbull, D., Glass Transition in o‐Terphenyl. J. Chem. Phys. 1967,

46 (4), 1243-1251.

268. Shi, Q.; Cai, T., Fast Crystal Growth of Amorphous Griseofulvin: Relations

between Bulk and Surface Growth Modes. Cryst. Growth. Des. 2016, 16 (6), 3279-3286.

269. Sun, Y.; Xi, H.; Ediger, M. D.; Richert, R.; Yu, L., Diffusion-controlled and

"diffusionless" crystal growth near the glass transition temperature: relation between liquid

dynamics and growth kinetics of seven ROY polymorphs. J. Chem. Phys. 2009, 131 (7),

074506.

270. Musumeci, D.; Powell, C. T.; Ediger, M. D.; Yu, L., Termination of Solid-State

Crystal Growth in Molecular Glasses by Fluidity. J. Phys. Chem. Lett. 2014, 5 (10), 1705-

10.

271. Ishida, H.; Wu, T.; Yu, L., Sudden rise of crystal growth rate of nifedipine near

T(g) without and with polyvinylpyrrolidone. J. Pharm. Sci. 2007, 96 (5), 1131-8.

272. Sun, Y.; Xi, H.; Chen, S.; Ediger, M. D.; Yu, L., Crystallization near glass

transition: transition from diffusion-controlled to diffusionless crystal growth studied with

seven polymorphs. J. Phys. Chem. B 2008, 112 (18), 5594-601.

182

273. Shtukenberg, A.; Freundenthal, J.; Gunn, E.; Yu, L.; Kahr, B., Glass-Crystal

Growth Mode for Testosterone Propionate. Cryst. Growth. Des. 2011, 11 (10), 4458-4462.

274. Powell, C. T.; Paeng, K.; Chen, Z.; Richert, R.; Yu, L.; Ediger, M. D., Fast crystal

growth from organic glasses: comparison of o-terphenyl with its structural analogs. J. Phys.

Chem. B 2014, 118 (28), 8203-9.

275. Sun, Y.; Xi, H.; Ediger, M. D.; Yu, L., Diffusionless Crystal Growth from Glass

Has Precursor in Equilibrium Liquid. The Journal of Physical Chemistry B 2008, 112 (3),

661-664.

276. Hikima, T.; Adachi, Y.; Hanaya, M.; Oguni, M., Determination of potentially

homogeneous-nucleation-based crystallization ino-terphenyl and an interpretation of the

nucleation-enhancement mechanism. Phys. Rev. B 1995, 52 (6), 3900-3908.

277. Tanaka, H., Possible resolution of the Kauzmann paradox in supercooled liquids.

Physical Review E 2003, 68 (1), 011505.

278. Powell, C. T.; Xi, H.; Sun, Y.; Gunn, E.; Chen, Y.; Ediger, M. D.; Yu, L., Fast

Crystal Growth in o-Terphenyl Glasses: A Possible Role for Fracture and Surface Mobility.

J. Phys. Chem. B 2015, 119 (31), 10124-30.

279. Zhang, W.; Brian, C. W.; Yu, L., Fast surface diffusion of amorphous o-terphenyl

and its competition with viscous flow in surface evolution. J. Phys. Chem. B 2015, 119

(15), 5071-8.

280. Zhu, L.; Brian, C. W.; Swallen, S. F.; Straus, P. T.; Ediger, M. D.; Yu, L., Surface

self-diffusion of an organic glass. Phys. Rev. Lett. 2011, 106 (25), 256103.

183

281. Sun, Y.; Zhu, L.; Kearns, K. L.; Ediger, M. D.; Yu, L., Glasses crystallize rapidly

at free surfaces by growing crystals upward. Proc. Natl. Acad. Sci. U. S. A. 2011, 108 (15),

5990-5.

282. Wu, T.; Sun, Y.; Li, N.; de Villiers, M. M.; Yu, L., Inhibiting surface

crystallization of amorphous indomethacin by nanocoating. Langmuir 2007, 23 (9), 5148-

53.

283. Li, Y.; Yu, J.; Hu, S.; Chen, Z.; Sacchetti, M.; Sun, C. C.; Yu, L., Polymer

nanocoating of amorphous drugs for improving stability, dissolution, powder flow, and

tabletability: The case of chitosan-coated indomethacin. Mol. Pharm. 2019.

284. Stephen R. Byrn, G. Z., Xiaoming (Sean) Chen, Solid-State Properties of

Pharmaceutical Materials. John Wiley & Sons, inc.: 2017; p 9.

285. Chawla, G.; Gupta, P.; Thilagavathi, R.; Chakraborti, A. K.; Bansal, A. K.,

Characterization of solid-state forms of celecoxib. Eur. J. Pharm. Sci. 2003, 20 (3), 305-

17.

286. Rodríguez-Tinoco, C.; Gonzalez-Silveira, M.; Ràfols-Ribé, J.; Garcia, G.;

Rodríguez-Viejo, J., Highly stable glasses of celecoxib: Influence on thermo-kinetic

properties, microstructure and response towards crystal growth. J. Non-Cryst. Solids 2015,

407, 256-261.

287. Shi, L.; Sun, C. C., Subsurface nucleation of supercooled acetaminophen.

CrystEngComm 2018, 20 (43), 6867-6870.

288. Wang, K.; Mishra, M. K.; Sun, C. C., An Exceptionally Elastic Single Component

Pharmaceutical Crystal. Chemistry of Materials 2019.

184

289. Hasebe, M.; Musumeci, D.; Powell, C. T.; Cai, T.; Gunn, E.; Zhu, L.; Yu, L.,

Fast Surface Crystal Growth on Molecular Glasses and Its Termination by the Onset of

Fluidity. J. Phys. Chem. B 2014, 118 (27), 7638-7646.

290. Xi, H.; Sun, Y.; Yu, L., Diffusion-controlled and diffusionless crystal growth in

liquid o-terphenyl near its glass transition temperature. J. Chem. Phys. 2009, 130 (9),

094508.

291. Mehta, M.; Ragoonanan, V.; McKenna, G. B.; Suryanarayanan, R., Correlation

between Molecular Mobility and Physical Stability in Pharmaceutical Glasses. Mol.

Pharm. 2016, 13 (4), 1267-77.

292. Dantuluri, A. K.; Amin, A.; Puri, V.; Bansal, A. K., Role of alpha-relaxation on

crystallization of amorphous celecoxib above T(g) probed by dielectric spectroscopy. Mol.

Pharm. 2011, 8 (3), 814-22.

293. Ferro LJ, M. P., Polymorphic crystalline forms of celecoxib. US 7476744 B2 2004.

294. Tao, J.; Jones, K. J.; Yu, L., Cross-Nucleation Between d-Mannitol Polymorphs in

Seeded Crystallization. Cryst. Growth. Des. 2007, 7 (12), 2410-2414.

295. M. Gunn, E.; A. Guzei, I.; Yu, L., Does Crystal Density Control Fast Surface

Crystal Growth in Glasses? A Study with Polymorphs. Cryst. Growth Des. 2011, 11 (9),

3979-3984.

296. Yu, L., Surface mobility of molecular glasses and its importance in physical

stability. Adv. Drug Deliv. Rev. 2016, 100, 3-9.

297. Huang, C.; Ruan, S.; Cai, T.; Yu, L., Fast Surface Diffusion and Crystallization

of Amorphous Griseofulvin. J. Phys. Chem. B 2017, 121 (40), 9463-9468.

185

298. Ruan, S.; Zhang, W.; Sun, Y.; Ediger, M. D.; Yu, L., Surface diffusion and surface

crystal growth of tris-naphthyl benzene glasses. J. Chem. Phys. 2016, 145 (6), 064503.

299. Hatase, M.; Hanaya, M.; Oguni, M., Studies of homogeneous-nucleation-based

crystal growth: significant role of phenyl ring in the structure formation. J. Non-Cryst.

Solids 2004, 333 (2), 129-136.

300. Taylor, L. J.; Papadopoulos, D. G.; Dunn, P. J.; Bentham, A. C.; Mitchell, J. C.;

Snowden, M. J., Mechanical characterisation of powders using nanoindentation. Powder

Technol. 2004, 143, 179-185.

186

Appendix

This appendix contains copies of copyright permissions granted by the journal

publishers where the work is published.

187

Chapter 2 - Exceptionally Elastic Crystals of Celecoxib Form III of this thesis is

adapted from the published original research article by Kunlin Wang at Chemistry of

Materials : https://pubs.acs.org/doi/10.1021/acs.chemmater.9b00040. Further permissions

related to the material excerpted should be directed to the ACS.

https://pubs.acs.org/doi/10.1021/acs.chemmater.9b00040

188

Chapter 5 - Crystal Growth of Celecoxib from Amorphous State - Polymorphism,

Growth Mechanism, and Kinetics is an original research article by Kunlin Wang published

on Crystal Growth and Design. It is adapted with permission from Wang, K.; Sun, C. C.,

Crystal Growth of Celecoxib from Amorphous State: Polymorphism, Growth Mechanism,

and Kinetics. Cryst. Growth Des. 2019, 19 (6), 3592-3600. Copyright (2020) American

Chemical Society.

ENABLING DIRECT COMPRESSION TABLET DEVELOPMENT OF ...

Documents

Transcript of ENABLING DIRECT COMPRESSION TABLET DEVELOPMENT OF ...