ENABLING DIRECT COMPRESSION TABLET DEVELOPMENT OF ...
Transcript of 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
© Kunlin Wang 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 Materials and Methods .................................................................................. 50
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 Results and Discussion ................................................................................. 54
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
4.3 Materials and Methods .................................................................................. 79
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 Results and Discussion ................................................................................. 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 Materials and Methods
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
difference Fourier map or allowed to ride on their parent atoms in the refinement cycles.
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 Results and Discussion
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 Materials and Methods
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
difference Fourier map or allowed to ride on their parent atoms in the refinement cycles.
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.
4.4 Results and Discussion
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 Materials and Methods
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 Results and Discussion
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).
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.; Fábián, 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 (38�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
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.
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.
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.
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.
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.