A Dissertation Entitled Acetaldehyde Scavengers for Poly(Ethylene Terephthalate)

Home , Acetaldehyde, Aldehyde

A Dissertation

entitled

Acetaldehyde Scavengers for Poly(ethylene terephthalate): Chemistry of Reactions,

Capacity, and Modeling of Interactions

Brent A. Mrozinski

Submitted to the Graduate Faculty as partial fulfillment of the

requirements for the Doctor of Philosophy Degree in Engineering

Dr. Saleh A. Jabarin, Committee Chair

Dr. Dong-Shik Kim, Committee Member

Dr. Yong-Wah Kim, Committee Member

Dr. Steven E. LeBlanc, Committee Member

Dr. Arunan Nadarajah, Committee Member

Dr. Patricia R. Komuniecki, Dean College of Graduate Studies

The University of Toledo December 2010

This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author.

An Abstract of

Acetaldehyde Scavengers for Poly(ethylene terephthalate): Chemistry of Reactions, Capacity, and Modeling of Interactions

Brent A. Mrozinski

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Doctor of Philosophy Degree in Engineering

The University of Toledo December 2010

During the melting and processing of poly(ethylene terephthalate) (PET), degradation of

the material may occur. One of the more common degradation products is acetaldehyde

(AA). Due to its low boiling point, 21oC, AA is able to diffuse out of PET and into either the atmosphere or the packaged contents of the PET container. The diffusion of AA into packaged contents is of concern, because many food products have a limited threshold for the sweet, fruity taste and odor of AA. One of the ways to limit the AA affects is through the addition of AA scavenging agents. While these additives do not limit the generation of AA; they are designed to interact with and reduce the amount of AA that can be release from PET articles.

The purpose of this study was not only to study these AA and AA scavenger interactions and quantify their abilities in reducing AA concentrations in PET; it was also to develop an initial model to predict effectiveness of adding AA scavengers to multi-cavity PET injection molding systems. Through this work, it was determined that anthranilamide and meta-xylenediamine (MXDA) reduce AA concentrations in PET by means of a reaction mechanism. Alpha-cyclodextrin, however, scavenges AA through a hydrogen

iii

bonding/size-enclosing scheme. Regardless of the mechanism, it was proven that these

three scavengers are capable of reducing detectable AA concentrations in PET. It was

generally found that the greater the AA scavenger concentration, the great the effect.

Additionally, the changes in the physical properties of PET due to AA scavenger addition

were studied. It was shown that melt-blending these additives into PET could adversely

affect the intrinsic viscosity (I.V.) and color of the PET blend resin and/or container. The thermal properties and oxygen permeation of PET were not affected by AA scavenger addition.

The modification of an existing multi-cavity injection molding program was applied to

account for the addition of AA scavengers, to PET resin, when predicting the

accumulation of AA within PET preforms. The approach to modify this original program

and methodologies to quantify the appropriate kinetic terms has been described in detail.

Finally, the modified simulation program was then used to predict the effectiveness of

various AA scavenger/PET blends in reducing detectable AA concentrations in PET preforms. While complete agreement between the modeling results and observed trends

from single-cavity injection molding was not achieved, the groundwork was laid to make further improvements and advance predictability for future modeling programs.

iv To my wife, Whitney, you are the most important part of my life and I hope that I show you the constant and unwavering love and support that you show me everyday. I look forward to our journey together and all the joy it may bring.

To my mother, Nancy, thank you for too many things to list. Your love, guidance, and friendship throughout these first 29 years of my life have been unimaginable. Thank you for the tremendous examples of how to treat and respect others, how to be a parent, and how to be a loving and selfless spouse.

To my late father, Richard, who was not only a parent to me; he was also my best friend.

His tireless love, support, and encouragement still lives with me to this day. We shared many great moments together: going up north to cut firewood, playing catch, watching

Detroit Tigers baseball games, and trying to teach him about golf; a sport for which he had no interest except for the fact that it was important to me. His sudden passing on

December 19, 2006, following a short bout with cancer, left a tremendous void in my life and heart. We were always very close, but those last 3 months were filled with moments that I will never forget: many shared laughs, tears, and short walks through the house.

To quote a song from Keith Urban, “I only hope when I have my own family that everyday I see a little more of my father in me.”

Acknowledgments

First and foremost, I would like to thank Dr. Saleh A. Jabarin for giving me this tremendous opportunity to learn from him and conduct this research project at the

Polymer Institute, under his guidance. His knowledge, encouragement, and incredible patience have been not only appreciated, but greatly needed as well.

I would like to thank Dr. Mike Cameron for his help with the computational modeling work and Mrs. Elizabeth Lofgren for her help with the various analytical experiments and for reviewing this work. The generosity of their time and effort has been immensely appreciated. Thank you to Mr. Mike Mumford for his help with the processing equipment and experiments and to Mrs. Jackie Zydorczyk for her support and help.

Thank you to my fellow students at the Polymer Institute for your encouragement and suggestions throughout my project; especially to Mr. Thomas R. Matthews, Dr. Sung-Gi

Kim, and Dr. Kamal Mahajan for their assistance in conducting experiments. Thank you, as well, to the PET Industrial Consortium for their financial support for this work.

I would like to thank Dr. Dong-Shik Kim, Dr. Yong-Wah Kim, Dr. Steven E. LeBlanc, and Dr. Arunan Nadarajah for serving on my dissertation committee.

A final thank you is extended to my wife, my mother, and my entire family for their support and encouragement throughout these years.

Contents

Abstract iii

Acknowledgments vi

Contents vii

List of Figures xiii

List of Tables xxi

1 Introduction 1

1.1 Poly(ethylene terephthalate) Overview ...... 1

1.2 Synthesis of PET ...... 2

1.2.1 Melt-Phase Polymerization...... 2

1.2.2 Solid-State Polymerization ...... 5

1.2.3 Direct Melt-Phase to Higher I.V...... 7

1.2.4 PET Copolymers...... 8

1.3 Degradation of PET...... 9

1.3.1 Hydrolytic Degradation of PET...... 9

1.3.2 Thermal Degradation of PET...... 10

1.3.3 Thermal-Oxidative Degradation of PET ...... 10

1.4 Acetaldehyde in PET...... 11

1.4.1 Overview...... 11

vii

1.4.2 Amount of Acetaldehyde in PET...... 12

1.4.3 PET Degradation Routes That Generate Acetaldehyde...... 13

1.4.3.1 Thermal Decomposition of Hydroxyl End-Groups...... 14

1.4.3.2 Breakdown of Diethylene Glycol Molecules ...... 14

1.4.3.3 Reactions with Vinyl Ester End-Groups ...... 15

1.4.3.4 Presence of Oxygen...... 16

1.4.3.5 Presence of DEG Linkages within PET Chains ...... 17

1.4.3.6 Presence of Free Radicals ...... 19

1.4.4 Ways to Reduce the Amount of AA within PET...... 20

1.5 Rationale and Objectives...... 22

2 Literature Review 25

2.1 Minimizing AA During PET Degradation Mechanisms ...... 25

2.1.1 Limiting Thermal Degradation...... 25

2.1.2 Limiting Thermal-Oxidative Degradation...... 27

2.2 Minimizing AA by Choice of Polymerization Catalysts...... 29

2.3 Minimizing AA by Choice of PET Resins...... 30

2.4 Minimizing AA by Means of Acetaldehyde (AA) Scavengers...... 31

2.4.1 Reactive AA Scavengers ...... 31

2.4.1.1 Polyamides ...... 31

2.4.1.2 Low Molecular Weight Amides ...... 34

2.4.1.3 Polyamines ...... 35

2.4.1.4 Polyimines ...... 36

2.4.1.5 Polyols...... 37

viii

2.4.2 Size-Enclosing AA Scavengers ...... 39

2.4.2.1 Cyclodextrins...... 39

2.4.2.2 Zeolites...... 43

2.4.3 AA Scavenging Catalysts ...... 45

3 Experimental Work 47

3.1 Materials...... 47

3.2 Spectroscopic Techniques to Study Chemical Interactions...... 48

3.2.1 Nuclear Magnetic Resonance (NMR) Spectroscopy...... 48

3.2.1.1 Proton (1H) NMR ...... 49

3.2.1.2 1H-1H COrrelation SpectroscopY (COSY) ...... 50

3.2.2 Mass Spectrometry ...... 51

3.3 Twin-Screw Extrusion...... 52

3.3.1 Preparation of Alpha-Cyclodextrin/PET Blend Samples ...... 53

3.3.2 Preparation of Anthranilamide/PET Blend Samples ...... 55

3.3.3 Preparation of MXDA/PET Blend Samples...... 56

3.3.4 Preparation of Control PET Samples...... 58

3.4 Manufacturing of PET Containers...... 59

3.4.1 Injection Molding...... 60

3.4.2 Stretch-Blow-Molding...... 61

3.5 Gas Chromatography...... 62

3.5.1 Acetaldehyde Generation Analysis...... 63

3.5.2 Residual Acetaldehyde Analysis...... 66

3.6 Rheological Methods...... 69

3.6.1 Plate and Plate Rheometer...... 69

3.6.2 Capillary Rheometer...... 70

3.7 Color Analysis...... 70

3.8 Differential Scanning Calorimetry (DSC) Analysis...... 72

3.9 Oxygen Film Permeation...... 72

4 Results and Discussion 76

4.1 Chemical Mechanisms of AA and AA Scavenger Interactions ...... 76

4.1.1 AA and Anthranilamide...... 76

4.1.2 AA and MXDA...... 93

4.1.3 AA and Alpha-Cyclodextrin...... 98

4.2 Effectiveness of AA Scavengers in Reducing the Amount of AA in PET ...... 106

4.2.1 AA Generation Rates...... 107

4.2.2 Residual AA...... 116

4.2.2.1 Pelletized Samples...... 117

4.2.2.2 Preform Samples...... 119

4.2.2.3 Comparison of Results for Pelletized Samples and Preform

Samples...... 120

4.3 Physical Properties of AA Scavenger/PET Blend Samples ...... 121

4.3.1 Intrinsic Viscosity (I.V.)...... 121

4.3.1.1 Pelletized Samples...... 122

4.3.1.2 Preform Samples...... 123

4.3.1.3 Comparison of Results for Pelletized Samples and Preform

Samples...... 125

4.3.2 Color...... 126

4.3.2.1 Color Analysis...... 127

4.3.2.2 Appearance of 2-Liter Bottles ...... 128

4.3.3 Thermal Properties...... 129

4.3.3.1 Glass Transition Temperature (Tg)...... 132

4.3.3.2 Crystallization Behavior When Heating from the Glassy State 132

4.3.3.3 Melting Behavior...... 134

4.3.3.4 Crystallization Behavior When Cooling form the Melt ...... 135

4.3.4 Oxygen Film Permeation...... 136

4.4 Optimal AA Scavenger/PET Blends ...... 137

4.4.1 Anthranilamide/PET Blends...... 137

4.4.2 MXDA/PET Blends...... 139

4.4.3 Alpha-Cyclodextrin/PET Blends...... 141

4.5 Modeling...... 144

4.5.1 Predictive AA Generation Program...... 145

4.5.2 Modified AA Generation Program ...... 157

4.5.2.1 Numerical Analysis...... 164

4.5.2.1.1 Determination of k2 and a...... 165

4.5.2.1.2 Determination of k1, bb1, bb2, and b ...... 174

4.5.3 Modeling Results...... 182

5 Conclusions and Recommendations 191

5.1 Conclusions...... 191

5.1.1 Chemical Mechanisms of AA and AA Scavenger Interaction ...... 192

5.1.2 Effectiveness of AA Scavengers’ Apparent Reduction in Generated

AA ...... 194

5.1.3 Physical Properties of AA Scavenger/PET Blend Samples ...... 197

5.1.4 Optimal AA Scavenger/PET Blends ...... 199

5.1.5 Modeling...... 201

5.2 Recommendations...... 203

References 206

A 1H NMR Spectra of AA and Alpha-Cyclodextrin Titration Experiment 220

B Raw Data from AA Generation Experiments 228

C AA Generation Plots 248

D Arrhenius Plots 259

E Raw Data from Residual AA Experiments 259

F Raw Data from Melt Viscosity Measurements to Determine I.V. 273

G Raw Data from Color Measurements 277

H Derivation of Thermal Energy Equation 280

I AA Generation 60 Minute Curve Study Data 283

J Data Used to Determine the k2 and a Values 287

K Data Used to Determine the k1, bb1, bb2, and b Values 289

L Raw Data from Modeling Program 300

xii

List of Figures

1-1 The Repeating Chemical Structure of PET ...... 1

1-2 Reaction Scheme for TPA and EG to Produce BHET ...... 3

1-3 Reaction Scheme for DMT and EG to Produce BHET...... 4

1-4 Reaction Scheme for the Polymerization of BHET to Produce PET and EG...... 4

1-5 Chemical Structure of Cyclohexane 1,4 Dimethanol (CHDM) ...... 9

1-6 Chemical Structure of Isophthalic Acid (IPA)...... 9

1-7 Chemical Structure of Acetaldehyde...... 11

1-8 Acetaldehyde Concentration During the Lifecycle of PET ...... 13

1-9 Thermal Decomposition of Hydroxyl Eng-Groups to Produce AA...... 14

1-10 AA Formation from Diethylene Glycol (DEG) Molecules...... 15

1-11 AA Formation from Vinyl Ester End-Groups Reacting with Hydroxyl End-

Groups and Carboxyl End-Groups...... 16

1-12 AA Generation Due to the Presence of Oxygen ...... 17

1-13 AA Generation from Diethylene Glycol (DEG) Linkages...... 18

1-14 AA Generation Due to the Presence of Free Radicals ...... 20

2-1 The Chemical Structure of MXD6 ...... 32

2-2 Acetaldehyde Scavenging Reaction Between MXD6 and AA ...... 33

2-3 Chemical Structure of Anthranilamide ...... 34

2-4 Aldehyde Scavenging Reaction Between D-Sorbitol and Valeraldehyde ...... 39

xiii

2-5 Various Forms of Cyclodextrin...... 40

2-6 Chemical Structure of Alpha-Cyclodextrin...... 40

2-7 Proposed Fitting of Caproaldehyde and Cyclodextrin ...... 41

2-8 1H NMR Titration Experiment of 2-TeCD and GSH...... 42

3-1 Relative Location of IR Heating Zones with Respect to a Preform...... 61

1 4-1 H NMR Spectrum of AA in CDCl3 ...... 77

1 4-2 H NMR Spectrum of Anthranilamide in CDCl3 ...... 78

4-3 1H NMR Spectrum of the Reaction Between Anthranilamide and AA, in

o CDCl3, After Heating for 2 Days at 60 C ...... 81

4-4 1H –1H COSY NMR Spectrum of the Reaction Between Anthranilamide and

o AA, in CDCl3, After Heating for 2 Days at 60 C ...... 83

4-5 Proposed Reaction Mechanism #1 for Anthranilamide and AA...... 84

4-6 Predicted 1H NMR Spectrum for the Product Formed from the Proposed

Reaction Mechanism #1 (Figure 4-5)...... 85

4-7 Proposed Reaction Mechanism #2 for Anthranilamide and AA...... 86

4-8 Predicted 1H NMR Spectrum for the Product Formed from the Proposed

Reaction Mechanism #2 (Figure 4-7)...... 87

4-9 ESI Mass Spectrum of Anthranilamide in CDCl3 and Methanol...... 88

4-10 ESI Mass Spectrum of the Product from the Reaction Between Anthranilamide

and AA in CDCl3 and Methanol ...... 90

4-11 Proposed Reaction Mechanism #3 for Anthranilamide and AA...... 91

4-12 Predicted 1H NMR Spectrum for the Product Formed from the Proposed Reaction

Mechanism #3 (Figure 4-11)...... 92

xiv

1 4-13 H NMR Spectrum of MXDA in CDCl3 ...... 94

4-14 Proposed Reaction Scheme for MXDA and AA...... 96

1 4-15 H NMR Spectrum of the Reaction Between MXDA and AA in CDCl3 ...... 97

1 4-16 H NMR Spectrum of the Alpha-Cyclodextrin in D2O...... 99

1 1 4-17 H – H COSY NMR Spectrum of Alpha-Cyclodextrin in D2O...... 100

1 4-18 H NMR Spectrum of AA in D2O...... 101

4-19 Equilibrium Reaction Between AA and D2O...... 102

1 4-20 Predicted H NMR Spectrum of AA in D2O...... 102

4-21 Interaction Mechanism for AA and Alpha-Cyclodextrin...... 104

4-22 Peak Shifting of the Protons for AA and its Equilibrium Product When Titrated

with Alpha-Cyclodextrin (Solvent is D2O) ...... 106

4-23 AA Generation Plots for the 1200 ppm Anthranilamide/PET Blend...... 108

4-24 AA Generation Rates as a Function of Anthranilamide Concentration ...... 110

4-25 AA Generation Rates as a Function of Alpha-Cyclodextrin Concentration ...... 110

4-26 AA Generation Rates as a Function of MXDA Concentration...... 111

4-27 Arrhenius Plot of 10,000 ppm MXDA/PET Blend Sample...... 114

4-28 2-liter Blow-Molded PET Bottles ...... 129

4-29 DSC Heating Curve of the 5 Weight % Alpha-Cyclodextrin/PET Blend...... 131

4-30 DSC Cooling Curve of the 2.5 Weight % Alpha-Cyclodextrin/PET Blend ...... 131

4-31 Viscosity versus Shear Rate Curves for the Voridian CB12 PET Resin ...... 149

4-32 Arrhenius Plot for the Voridian CB12 PET Resin ...... 151

4-33 Temperature Profile as a Function of Radial Distance from the Center of a Flow

Channel over a Two Second Period of Time ...... 153

4-34 Distribution of AA as a Function of Radial Distance from the Center of the Flow

Channel...... 154

4-35 Distribution of Material to Fill Four Cavities within an Eight-Cavity Mold...... 155

4-36 Temperatures for the Various Cavities as a Function of Filling Times ...... 156

4-37 60 Minute AA Generation Curve for the 10,000 ppm Anthranilamide/PET

Blend...... 167

4-38 60 Minute AA Generation Curve for the 10,000 ppm Alpha-Cyclodextrin/PET

Blend ...... 168

4-39 60 Minute AA Generation Curve for the 10,000 ppm MXDA/PET Blend ...... 168

d[AA] 4-40 Plot of ln(R − ) Versus ln([AA]) for the 10,000 ppm G dt

Anthranilamide/PET Blend at 290oC ...... 170

d[AA] 4-41 Plot of ln(R − ) Versus ln([AA]) for the 10,000 ppm Alpha- G dt

Cyclodextrin/PET Blend at 290oC ...... 171

d[AA] 4-42 Plot of ln(R − ) Versus ln([AA]) for the 10,000 ppm MXDA/PET G dt

Blend at 290oC ...... 171

4-43 60 Minute AA Generation Data for the 10,000 ppm Anthranilamide/PET

Blend Fitted with Equation 22, Using a and k2 Values...... 172

4-44 60 Minute AA Generation Data for the 10,000 ppm Alpha-Cyclodextrin/PET

Blend Fitted with Equation 22, Using a and k2 Values...... 173

4-45 60 Minute AA Generation Data for the 10,000 ppm MXDA/PET Blend Fitted

with Equation 22, Using a and k2 Values...... 173

xvi

4-46 60 Minute AA Generation Data for the 10,000 ppm Anthranilamide/PET

o Blend Fitted with Equation 17; Using the a, k1 at 290 C, and b Values ...... 177

4-47 60 Minute AA Generation Data for the 10,000 ppm Alpha-Cyclodextrin/PET

o Blend Fitted with Equation 17; Using the a, k1 at 290 C, and b Values ...... 177

4-48 60 Minute AA Generation Data for the 10,000 ppm MXDA/PET Blend Fitted

o with Equation 17; Using the a, k1 at 290 C, and b Values...... 178

4-49 60 Minute AA Generation Data for the 10,000 ppm Anthranilamide/PET

nd Blend Fitted with Equation 17; Using the a, b, and bb2 Values and the 2

Iteration bb1 Value ...... 180

4-50 60 Minute AA Generation Data for the 10,000 ppm Alpha-Cyclodextrin/PET

nd Blend Fitted with Equation 17; Using the a, b, and bb2 Values and the 2

Iteration bb1 Value ...... 181

4-51 60 Minute AA Generation Data for the 10,000 ppm MXDA/PET Blend Fitted

nd with Equation 17; Using the a, b, and bb2 Values and the 2 Iteration bb1

Value ...... 181

4-52 Predicted Injection Molding Results for Various Anthranilamide/PET Blends

and for Various Manifold Designs; Modeled at 280oC...... 183

4-53 Predicted Injection Molding Results for Various Alpha-Cyclodextrin/PET

Blends and for Various Manifold Designs; Modeled at 280oC...... 183

4-54 Predicted Injection Molding Results for Various MXDA/PET Blends and for

Various Manifold Designs; Modeled at 280oC ...... 184

4-55 One Minute Simulated AA Generation at 280oC...... 185

4-56 One Minute Simulated AA Generation at 290oC...... 186

xvii

4-57 One Minute Simulated AA Generation at 300oC...... 186

4-58 Predicted Injection Molding Results for Various Anthranilamide/PET Blends,

Studied as a Function of Temperature; Modeled for a 48 Cavity Process...... 187

4-59 Predicted Injection Molding Results for Various Alpha-Cyclodextrin/PET

Blends, Studied as a Function of Temperature; Modeled for a 48 Cavity

Process...... 188

4-60 Predicted Injection Molding Results for Various MXDA/PET Blends, Studied

as a Function of Temperature; Modeled for a 48 Cavity Process ...... 188

A-1 AA and Alpha-Cyclodextrin, in D2O, Mixed in a 0.2 to 1 Ratio...... 220

A-2 AA and Alpha-Cyclodextrin, in D2O, Mixed in a 0.4 to 1 Ratio...... 221

A-3 AA and Alpha-Cyclodextrin, in D2O, Mixed in a 0.6 to 1 Ratio...... 222

A-4 AA and Alpha-Cyclodextrin, in D2O, Mixed in a 0.8 to 1 Ratio...... 223

A-5 AA and Alpha-Cyclodextrin, in D2O, Mixed in a 1 to 1 Ratio...... 224

A-6 AA and Alpha-Cyclodextrin, in D2O, Mixed in a 2 to 1 Ratio...... 225

A-7 AA and Alpha-Cyclodextrin, in D2O, Mixed in a 3 to 1 Ratio...... 226

C-1 AA Generation Plots for the Voridian CB12 PET Resin...... 248

C-2 AA Generation Plots for the “One-Time” Processed PET Sample...... 249

C-3 AA Generation Plots for the “Two-Times” Processed PET Sample ...... 249

C-4 AA Generation Plots for the “Three-Times” Processed PET Sample ...... 250

C-5 AA Generation Plots for the 10,000 ppm Anthranilamide/PET Blend Sample.....250

C-6 AA Generation Plots for the 500 ppm Anthranilamide/PET Blend Sample...... 251

C-7 AA Generation Plots for the 200 ppm Anthranilamide/PET Blend

Sample...... 251

xviii

C-8 AA Generation Plots for the 100 ppm Anthranilamide/PET Blend Sample...... 252

C-9 AA Generation Plots for the 50,000 ppm Alpha-Cyclodextrin/PET Blend

Sample...... 252

C-10 AA Generation Plots for the 25,000 ppm Alpha-Cyclodextrin/PET Blend

Sample...... 253

C-11 AA Generation Plots for the 10,000 ppm Alpha-Cyclodextrin/PET Blend

Sample...... 253

C-12 AA Generation Plots for the 5000 ppm Alpha-Cyclodextrin/PET Blend Sample.254

C-13 AA Generation Plots for the 1200 ppm Alpha-Cyclodextrin/PET Blend Sample.254

C-14 AA Generation Plots for the 500 ppm Alpha-Cyclodextrin/PET Blend Sample...255

C-15 AA Generation Plots for the 10,000 ppm MXDA Blend Sample...... 255

C-16 AA Generation Plots for the 1200 ppm MXDA Blend Sample...... 256

C-17 AA Generation Plots for the 500 ppm MXDA Blend Sample...... 256

C-18 AA Generation Plots for the 200 ppm MXDA Blend Sample...... 257

C-19 AA Generation Plots for the 100 ppm MXDA Blend Sample...... 257

D-1 Arrhenius Plot for the “One-Time” Processed PET Sample...... 259

D-2 Arrhenius Plot for the “Two-Time” Processed PET Sample ...... 260

D-3 Arrhenius Plot for the “Three-Time” Processed PET Sample ...... 260

D-4 Arrhenius Plot for the 10,000 ppm Anthranilamide/PET Blend Sample...... 261

D-5 Arrhenius Plot for the 1200 ppm Anthranilamide/PET Blend Sample...... 261

D-6 Arrhenius Plot for the 500 ppm Anthranilamide/PET Blend Sample...... 262

D-7 Arrhenius Plot for the 200 ppm Anthranilamide/PET Blend Sample...... 262

D-8 Arrhenius Plot for the 100 ppm Anthranilamide/PET Blend Sample...... 263

xix

D-9 Arrhenius Plot for the 50,000 ppm Alpha-Cyclodextrin/PET Blend Sample ...... 263

D-10 Arrhenius Plot for the 25,000 ppm Alpha-Cyclodextrin/PET Blend Sample ...... 264

D-11 Arrhenius Plot for the 10,000 ppm Alpha-Cyclodextrin/PET Blend Sample ...... 264

D-12 Arrhenius Plot for the 5000 ppm Alpha-Cyclodextrin/PET Blend Sample...... 265

D-13 Arrhenius Plot for the 1200 ppm Alpha-Cyclodextrin/PET Blend Sample...... 265

D-14 Arrhenius Plot for the 500 ppm Alpha-Cyclodextrin/PET Blend Sample ...... 266

D-15 Arrhenius Plot for the 1200 ppm MXDA/PET Blend Sample...... 266

D-16 Arrhenius Plot for the 500 ppm MXDA/PET Blend Sample...... 267

D-17 Arrhenius Plot for the 200 ppm MXDA/PET Blend Sample...... 267

D-18 Arrhenius Plot for the 100 ppm MXDA/PET Blend Sample...... 268

List of Tables

1.1 Typical I.V. Ranges for Various PET Uses ...... 2

3.1 The Thermal History for Each Alpha-Cyclodextrin/PET Blend...... 55

3.2 The Thermal History for Each Anthranilamide/PET Blend...... 56

3.3 The Thermal History for Each MXDA/PET Blend ...... 58

3.4 Optimized Stretch-Blow-Molding Parameters...... 62

3.5 Explanation of the Variables from Equation 1...... 65

3.6 Explanation of the Variables from Equations 2 and 3...... 66

3.7 Melt Viscosity Testing Conditions...... 69

3.8 Instrument Parameters for the Capillary Rheometry Analysis...... 70

3.9 Explanation of L, a, b, and YI Values...... 71

3.10 Explanation of the Variables in Equation 8 ...... 74

3.11 Explanation of the Variables in Equation 9 ...... 74

3.12 Explanation of the Variables in Equation 10 ...... 75

1 4.1 Peak Assignment for the H NMR Spectrum of AA in CDCl3 ...... 77

1 4.2 Peak Assignment for the H NMR Spectrum of Anthranilamide in CDCl3...... 79

4.3 Peak Assignment for the 1H NMR Spectrum of the Reaction Between

o Anthranilamide and AA, in CDCl3, After Heating for 2 Days at 60 C...... 82

1 4.4 Peak Assignment for the H NMR Spectrum of MXDA in CDCl3 ...... 94

xxi

4.5 Peak Assignment for the 1H NMR Spectrum of the Reaction Between MXDA

and AA, in CDCl3...... 97

1 4.6 Peak Assignment for the H NMR Spectrum of Alpha-Cyclodextrin in D2O ...... 100

1 4.7 Peak Assignment for the H NMR Spectrum of AA in D2O ...... 103

4.8 AA Generation Rates ...... 109

4.9 AA Generation Rates of Control Samples ...... 113

4.10 Activation Energies...... 115

4.11 Residual AA Data for Pelletized Samples...... 118

4.12 Residual AA Data for Preform Samples ...... 119

4.13 Comparison of the Residual AA Data for Pelletized and Preform Samples...... 120

4.14 I.V. Data for Pelletized Samples ...... 123

4.15 I.V. Data for Preform Samples...... 124

4.16 Comparison of the I.V. Data for Pelletized and Preform Samples ...... 125

4.17 L, a, and b Values and Yellowness Index of Pelletized Samples...... 127

4.18 Glass Transition Temperature (Tg) Data ...... 132

4.19 Crystallization Behavior Data When Heating from the Glassy State ...... 133

4.20 Melting Behavior Data When Heating from the Glassy State ...... 134

4.21 Crystallization Behavior Data When Cooling form the Melt ...... 135

4.22 Oxygen Film Permeability ...... 136

4.23 Explanation of the Terms in Equation 11...... 146

4.24 Explanation of the Terms in Equation 12...... 147

4.25 Explanation of the Terms in Equations 13 and 14 ...... 148

4.26 Capillary Rheometry Results ...... 150

4.27 Rheology Constants for the Predictive AA Generation Program ...... 150

4.28 Explanation of the Terms in Equation 15...... 151

xxii

4.29 Explanation of the Terms in Equation 16...... 152

4.30 Variables Needed to Run the Predictive AA Generation Program ...... 157

4.31 Explanation of the Terms in Equation 17...... 159

4.32 Explanation of the Terms in Equation 18...... 159

4.33 Explanation of the Terms in Equations 19 and 20 ...... 161

4.34 Review of the Residual AA Data for Preform Samples...... 162

4.35 Explanation of the Terms in Equations 22 and 23 ...... 163

4.36 Variables Needed to Modify the Predictive AA Generation Program...... 164

o 4.37 Calculated k2, for 290 C, and a Value for Each Scavenging Agent...... 172

4.38 b, bb1, and bb2 Values for Each Scavenging Agent Determined Through

Multiple Linear Regression...... 176

4.39 Final a, b, bb1, bb2, and k1 Values for Each AA Scavenging Agent ...... 182

A.1 Location of the AA and Alpha-Cyclodextrin Protons for each of the AA and

Alpha-Cyclodextrin NMR Titration Experiments...... 227

A.2 Change in Location of the Protons Representing AA and its D2O Equilibrium

Product, Due to the Presence of Alpha-Cyclodextrin ...... 227

B.1 AA Generation Data for the Voridian CB12 PET Resin ...... 228

B.2 AA Generation Data for the “One-Time” Processed PET Sample ...... 229

B.3 AA Generation Data for the “Two-Times” Processed PET Sample...... 230

B.4 AA Generation Data for the “Three-Times” Processed PET Sample...... 231

B.5 AA Generation Data for the 10,000 ppm Anthranilamide/PET Blend Sample .....232

B.6 AA Generation Data for the 1200 ppm Anthranilamide/PET Blend Sample ...... 233

B.7 AA Generation Data for the 500 ppm Anthranilamide/PET Blend Sample ...... 234

B.8 AA Generation Data for the 200 ppm Anthranilamide/PET Blend Sample ...... 235

xxiii

B.9 AA Generation Data for the 100 ppm Anthranilamide/PET Blend Sample ...... 236

B.10 AA Generation Data for the 50,000 ppm Alpha-Cyclodextrin/PET Blend

Sample...... 237

B.11 AA Generation Data for the 25,000 ppm Alpha-Cyclodextrin/PET Blend

Sample...... 238

B.12 AA Generation Data for the 10,000 ppm Alpha-Cyclodextrin/PET Blend

Sample...... 239

B.13 AA Generation Data for the 5000 ppm Alpha-Cyclodextrin/PET Blend Sample .240

B.14 AA Generation Data for the 1200 ppm Alpha-Cyclodextrin/PET Blend Sample .241

B.15 AA Generation Data for the 500 ppm Alpha-Cyclodextrin/PET Blend Sample ...242

B.16 AA Generation Data for the 10,000 ppm MXDA/PET Blend Sample ...... 243

B.17 AA Generation Data for the 1200 ppm MXDA/PET Blend Sample ...... 244

B.18 AA Generation Data for the 500 ppm MXDA/PET Blend Sample ...... 245

B.19 AA Generation Data for the 200 ppm MXDA/PET Blend Sample ...... 246

B.20 AA Generation Data for the 100 ppm MXDA/PET Blend Sample ...... 247

E.1 Residual AA Data for the Control PET Pelletized Samples ...... 269

E.2 Residual AA Data for the Anthranilamide/PET Blend Pelletized Samples...... 269

E.3 Residual AA Data for the Alpha-Cyclodextrin/PET Blend Pelletized Samples....270

E.4 Residual AA Data for the MXDA/PET Blend Pelletized Samples...... 270

E.5 Residual AA Data for the PET Control Preform Samples ...... 270

E.6 Residual AA Data for the Anthranilamide/PET Blend Preform Samples ...... 271

E.7 Residual AA Data for the Alpha-Cyclodextrin/PET Blend Preform Samples...... 271

E.8 Residual AA Data for the MXDA/PET Blend Preform Samples ...... 272

xxiv

F.1 Melt Viscosity Data for the Control PET Pelletized Samples...... 273

F.2 Melt Viscosity Data for the Anthranilamide/PET Blend Pelletized Samples...... 273

F.3 Melt Viscosity Data for the Alpha-Cyclodextrin/PET Blend Pelletized Samples .274

F.4 Melt Viscosity Data for the MXDA/PET Blend Pelletized Samples...... 274

F.5 Melt Viscosity Data for the Control PET Preform Samples ...... 274

F.6 Melt Viscosity Data for the Anthranilamide/PET Blend Preform Samples...... 275

F.7 Melt Viscosity Data for the Alpha-Cyclodextrin/PET Blend Preform Samples....275

F.8 Melt Viscosity Data for the MXDA/PET Blend Preform Samples ...... 276

G.1 Color Data for the Voridian CB12 PET Control Samples ...... 277

G.2 Color Data for the Anthranilamide/PET Blend Samples ...... 278

G.3 Color Data for the MXDA/PET Blend Samples ...... 278

G.4 Color Data for the Alpha-Cyclodextrin/PET Blend Samples ...... 279

H.1 Definition of Terms Listed in Equations 12 and 29 to 36...... 282

I.1 AA Generation Data for the CB12 PET Resin...... 283

I.2 AA Generation Data for the 10,000 ppm Anthranilamide/PET Blend ...... 284

I.3 AA Generation Data for the 10,000 ppm Alpha-Cyclodextrin/PET Blend ...... 285

I.4 AA Generation Data for the 10,000 ppm MXDA/PET Blend ...... 286

J.1 Calculated Data Based on 60 Minute AA Generation Data, at 290oC, for the

10,000 ppm Anthranilamide/PET Blend...... 287

J.2 Calculated Data Based on 60 Minute AA Generation Data, at 290oC, for the

10,000 ppm Alpha-Cyclodextrin/PET Blend...... 288

J.3 Calculated Data Based on 60 Minute AA Generation Data, at 290oC, for the

10,000 ppm MXDA/PET Blend...... 288

xxv

o K.1 Anthranilamide/PET Blend Data, at 280 C, Calculated for the Original ln(k2)

Versus ln([S0]) Plot ...... 289

o K.2 Anthranilamide/PET Blend Data, at 290 C, Calculated for the Original ln(k2)

Versus ln([S0]) Plot ...... 290

o K.3 Anthranilamide/PET Blend Data, at 300 C, Calculated for the Original ln(k2)

Versus ln([S0]) Plot ...... 291

K.4 Alpha-Cyclodextrin/PET Blend Data, at 280oC, Calculated for the Original

ln(k2) Versus ln([S0]) Plot ...... 292

K.5 Alpha-Cyclodextrin/PET Blend Data, at 290oC, Calculated for the Original

ln(k2) Versus ln([S0]) Plot ...... 293

K.6 Alpha-Cyclodextrin/PET Blend Data, at 300oC, Calculated for the Original

ln(k2) Versus ln([S0]) Plot ...... 294

o K.7 MXDA/PET Blend Data, at 280 C, Calculated for the Original ln(k2) Versus

ln([S0]) Plot ...... 295

o K.8 MXDA/PET Blend Data, at 290 C, Calculated for the Original ln(k2) Versus

ln([S0]) Plot ...... 296

o K.9 MXDA/PET Blend Data, at 300 C, Calculated for the Original ln(k2) Versus

ln([S0]) Plot ...... 297

K.10 Multiple Linear Regression Data Used to Determine the b, bb1, and bb2 Values

for the Anthranilamide/PET Blends...... 298

K.11 Multiple Linear Regression Data Used to Determine the b, bb1, and bb2 Values

for the Alpha-Cyclodextrin/PET Blends ...... 298

xxvi

K.12 Multiple Linear Regression Data Used to Determine the b, bb1, and bb2

Values for the MXDA/PET Blends...... 299

L.1 Predicted AA Generation Results for a 16 Cavity Injection Molding Process,

Modeled at 280oC...... 300

L.2 Predicted AA Generation Results for a 24 Cavity Injection Molding Process,

Modeled at 280oC...... 301

L.3 Predicted AA Generation Results for a 32 Cavity Injection Molding Process,

Modeled at 280oC...... 302

L.4 Predicted AA Generation Results for a 48 Cavity Injection Molding Process,

Modeled at 280oC...... 303

L.5 Predicted AA Generation Results for a 48 Cavity Injection Molding Process,

Modeled at 270oC...... 304

L.6 Predicted AA Generation Results for a 48 Cavity Injection Molding Process,

Modeled at 290oC...... 305

L.7 Predicted AA Generation Results for a 48 Cavity Injection Molding Process,

Modeled at 300oC...... 306

xxvii

Chapter 1

Introduction

1.1 Poly(ethylene terephthalate) Overview

Poly(ethylene terephthalate) (PET), shown in Figure 1-1, is a common polyester known for its optical, mechanical, and thermal properties. The combination of these properties makes PET applicable to many industries for a variety of uses.1 Some of the more common applications include:

• synthetic fiber to manufacture both apparel and carpets for the textile industry

• tire cord for the transportation industry

• angioplasty balloons for the medical industry

• transparent films for the office supply industry

• containers for the packaging industry

Figure 1-1: The repeating chemical structure of PET

For many packaging applications PET is a preferred material. When PET is properly oriented it provides good optical transparency, high impact strength, and good gas barrier properties.1-3 PET is particularly known to be a good barrier against the permeation of carbon dioxide and water vapor.1, 2 Over the past twenty-five years PET has become the leading packaging material for carbonated soft drinks, sports drinks, and water.

Depending on the final use, the molecular weight of the PET resin may vary. It is a common practice within the industry to identify PET samples by their respective intrinsic viscosities (I.V.) rather than their molecular weights.4 Table 1.1 shows typical I.V. ranges for a few common applications of PET.1, 4

Table 1.1: Typical I.V. ranges for various PET uses Uses Intrinsic Viscosity (I.V.) (dL/g) Textiles 0.5 to 0.65 Film and Tape 0.65 to 0.75 Bottles 0.70 to 1.00 Tire Cord 1.00

1.2 Synthesis of PET

As shown in Table 1.1, the I.V., or molecular weight, of PET can vary depending on its application. Identification of the intended use is important because PET manufacturing techniques can vary depending on the I.V. that is desired. To date, there are three common processes used to manufacture PET: melt-phase polymerization, solid-state polymerization, and direct melt-phase polymerization to high I.V. material. It can be important to choose the right technique because each has its own benefits and drawbacks.

1.2.1 Melt-Phase Polymerization

PET is typically synthesized in a two-step polymerization process.1, 3, 4 The first step is known as melt-phase polymerization. During this process ethylene glycol (EG), classified as a diol, is reacted with either terephthalic acid (TPA) or dimethyl terephthalate (DMT), both can be classified as di-acids. When TPA is used, a self- catalyzed esterification reaction occurs to produce bis-hydroxyl terephthalate (BHET) and water; as shown in Figure 1-2.1, 3, 4

O O HO

HO O HO + 2 + 2 H2O OO OH

OOH OH TPA EG BHET Water

Figure 1-2: Reaction scheme for TPA and EG to produce BHET

When DMT is used, a catalyzed transesterification reaction produces BHET and acetaldehyde (AA), a byproduct.1, 3 Some of the common catalysts used for this reaction include acetates of lithium (Li), calcium (Ca), magnesium (Mg), zinc (Zn), or lead (Pb); and oxides of Pb. This reaction scheme is shown in Figure 1-3.

CH3 HO O O O HO

2 + + 2 CH3 O H OH OO OO

CH3 OH DMT EG BHET AA

Figure 1-3: Reaction scheme for DMT and EG to produce BHET

Once BHET is produced, either from EG and TPA or EG and DMT, it must be polymerized to form PET. PET is obtained through a catalyzed, high temperature transesterification reaction; shown in Figure 1-4.1, 3 This reaction is an equilibrium reaction and thus the byproduct, ethylene glycol (EG), must be removed to obtain a high yield of PET.4 The catalysts for this reaction include acetates of antimony (Sb), Zn, or

Pb; and oxides of Sb, germanium (Ge), or Pb.1, 4

O O HO

O HO HO n O O H + n OO O n O OH

BHET PET EG

Figure 1-4: Reaction scheme for the polymerization of BHET to produce PET and EG

Traditional melt-phase polymerization techniques have a limit to the molecular weight, or

I.V., that can be achieved.1, 3, 4 The constraint is due to the difficulty in removing the reaction by-products (particularly EG), BHET, and oligomers from the viscous PET melt.1 Removal of these un-wanted by-products is needed to continually drive the equilibrium reaction forward and thus continually increase the degree of polymerization of the PET. Traditional melt-phase polymerization can be used to produce PET to be used as textile or as film or tape.1, 4 The desired I.V. for these applications is achievable with melt-phase polymerization alone. When higher I.V. PET is needed to manufacture containers or tire cord, for example, the I.V. needs to be increased beyond what melt- phase polymerization techniques can yield. While recent advancements5-12 in melt-phase polymerization technologies have increased the achievable I.V. ranges, traditionally a second polymerization technique was required to produce higher I.V. PET. This second step is known as solid-state polymerization (SSP).1, 3, 4

1.2.2 Solid-State Polymerization

Similar to melt-phase polymerization, solid-state polymerization (SSP) increases molecular weight (or I.V.) by driving the various PET end-groups to react with one another and thus increasing the length of the polymer chains.1, 3, 4 As the end-groups react with one another, by-products are formed. Removal of these by-products (water,

EG, AA, etc.) is achieved by continual inert gas purging or by applying vacuum pressure.1 As with melt-phase polymerization, by-product purging is a necessity to progress the SSP equilibrium reactions forward and ultimately reach the desired I.V.

During the SSP process, solid PET pellets are heated well above the polymer’s glass transition temperature (Tg) but below its melting temperature. This temperature range is typically between 200 and 240oC.1 As the temperature is increased, mobility of the polymer chains also increases. This increases the likelihood/ability of the polymer chains’ end-groups to find and react with one another. If the temperature is excessively increased, however, thermal degradation can occur; causing random chain scission to occur.1, 3 Random chain scission leads to the formation of low molecular weight byproducts and the loss of molecular weight from the PET chains. The thermal degradation of PET will be discussed in greater detail in Section 1.3.2.

The fundamental difference between melt-phase and solid-state polymerization is the phase of the reactants during the respective polymerization. Melt-phase polymerization is typically performed between 270 and 285oC; which is above the melting temperature of PET, usually listed to be above 255oC. SSP, however, is carried out at a much milder temperature range; between 200 and 240oC. Therefore, at SSP conditions PET is in a solid, rubbery state and not the viscous liquid seen during melt-phase polymerization.

This makes it much easier for a purging gas (or vacuum) to remove the volatile degradation products and reaction by-products that form during polymerization.1 The greater ease of by-product removal allows the polymerization process to progress beyond the limitation observed during melt-phase polymerization. Additionally, the milder SSP reaction temperature causes fewer side reactions to occur.1, 3 The combination of less side reactions and easier by-product removal create a more efficient route for the end-

groups of PET to react with one another and ultimately for the molecular weigh to increase more rapidly.

1.2.3 Direct Melt-Phase to Higher I.V.

As previously mentioned, two-step polymerization (melt-phase polymerization followed by solid-state polymerization) has been the traditional method to achieve high molecular, or high I.V., PET resins. In recent years, however, new melt-phase polymerization techniques5-12 have been developed that are now able to produce high molecular weight

PET resins without the need for solid-state polymerization. This technology has especially become prevalent when producing PET resins to be used in the manufacturing of containers. These new processes use the same starting materials (EG and TPA or EG and DMT) and follow the same chemistry as the traditional two-step method (Figures 1-2 to 1-4).

The appeal of these new melt-phase polymerization methods are believed to reduce the overall cost of production. Not only does SSP require additional reactors and energy to run the process, it is also a time consuming operation. Elimination of the SSP step would, in theory, allow PET producing plants to have higher throughput of resin.

Elimination of the SSP step, however, may need to be carefully considered. The inherent advantages of SSP, as previously discussed, are something that could be critical to the final product. Traditional solid-state polymerized PET resins possess minimal unwanted

by-products and good thermal stability; criteria which one-step polymerization techniques may not be able to meet.

1.2.4 PET Copolymers

There are applications where it is desirable to slightly alter the physical properties of poly(ethylene terephthalate) in order to better meet the needs of its end-use. To do this,

PET resins can be synthesized as copolymers instead of homopolymers.1, 3, 4 A copolymer is a polymer composed of two or more repeating monomer units. The chemistry presented up to this point details the reactions to polymerize a PET homopolymer.

The synthesis of a PET copolymer is achieved by replacing a small amount of the original raw materials (EG, TPA, or DMT) with another reactant(s). The copolymer concentration in the final product is typically less than 10%. Cyclohexane dimethanol

(CHDM) and isophthalic acid (IPA) are two of the more common reactants used as substitutes to create PET copolymer resins.3, 4 When compared to EG, the increased size of CHDM alters the structured packing of the polymer chains. This phenomenon affects the resin’s crystallinity and therefore lowers its melting point.4 The carboxyl end-groups of TPA are in a 1, 4 (“para”) configuration, where as the carboxyl end-groups of IPA are in a 1, 3 (“meta”) configuration. The “meta” configuration gives a slight angle to the polymer chain. This again alters the resin’s crystallinity and thus its melting point.4 The chemical structures of CHDM and IPA are shown in Figure 1-5 and 1-6, respectively.

HO Figure 1-5: Chemical structure of cyclohexane 1, 4 dimethanol (CHDM)

OOH

OH Figure 1-6: Chemical structure of isophthalic acid (IPA)

1.3 Degradation of PET

The melting and processing of PET resin into manufactured articles frequently results in at least some degradation of the material.1, 3, 4 During the extrusion or injection molding process the polymer can be subjected to moisture, oxygen, and/or elevated temperatures.

Each of these can cause at least one route of degradation. There are three main degradation processes that can occur during PET processing: hydrolytic degradation, thermal degradation, and thermal-oxidative degradation.1, 3

1.3.1 Hydrolytic Degradation of PET

Hydrolytic degradation occurs when of water reacts with PET at elevated temperatures.

This reaction can result in the reduction of I.V. or molecular weight; resulting in the generation of hydroxyl and carboxyl end-groups.1, 3, 13 PET is known to be a hygroscopic

material and thus it will to absorb moisture from the atmosphere.1, 3 To limit the effects of hydrolytic degradation, PET must be properly dried prior to processing the material. It is generally observed that PET should contain less than 50 parts per million (ppm) of water to be considered “properly dried”.1

1.3.2 Thermal Degradation of PET

Thermal degradation occurs when PET is exposed to high temperatures. This results in random chain scission, forming carboxyl and vinyl ester end-groups.13, 14 The formation of these end-groups leads to further reduction in I.V. or molecular weight, discoloration, formation of oligomers, and formation of low molecular weight byproducts.1, 3

Minimizing the effects of thermal degradation can be achieved by limiting the temperature, the residence time, and the shear heating that occurs during extrusion or injection molding.15

1.3.3 Thermal-Oxidative Degradation of PET

Thermal-oxidative degradation occurs when oxygen reacts with PET at elevated temperatures. It results in the reduction of I.V. or molecular weight, formation of carboxyl end-groups, generation of low molecular weight byproducts, discoloration, and formation of branched chains.1, 3, 4, 16 To limit thermal-oxidative degradation, PET should be melted and/or processed under vacuum or in an inert environment. For example, the

oxygen within the headspace of an extruder can be flushed by means of a nitrogen purge or a vacuum.

1.4 Acetaldehyde in PET

1.4.1 Overview

One of the more common byproducts resulting from the degradation of PET is acetaldehyde (AA).17 AA is commonly found as a natural component within many foods; including citrus fruits, bread, wine, and milk. 2, 18 AA is known to have a sweet, fruity taste and odor.2, 18, 19 This small organic compound, shown in Figure 1-7, is also very volatile. The boiling point of acetaldehyde is 21oC.19

CH3

OH Figure 1-7: Chemical structure of acetaldehyde

The presence of acetaldehyde within PET packages has been known to result in adverse effects.2, 20 With a boiling point that is lower than room temperature, AA is able to diffuse out of PET and into either the atmosphere or into the packaged contents. The diffusion of AA into packaged contents is a concern because many food products have a limited threshold for the taste of acetaldehyde. This is especially true when bottling

water because the taste of pure water is so sensitive that even a small amount of AA is detectable by consumers.21

1.4.2 Amount of Acetaldehyde in PET

The amount of acetaldehyde that is present within PET varies greatly during the polymer’s lifecycle. As previously discussed in greater detail during Section 1.2, synthesis of the material begins with melt-phase polymerization. For a two-step polymerization process (Section 1.2.1), melt-phase polymerization yields an amorphous

PET resin of relatively low I.V., typically around 0.60 dL/g, which possesses a high amount of AA. It is not uncommon for this resin to contain more than 20 ppm of AA.

To reduce the amount of degradation byproducts and prepare the PET resin for the second polymerization step, the amorphous resin is subsequently dried and crystallized.

This can reduce the amount of AA to less than 15 ppm.

The second polymerization step, solid-state polymerization, is conducted to further polymerize the PET resin and increase its I.V. Inherent to the SSP process, degradation byproducts, such as AA, are removed from the polymer’s matrix. SSP can reduce the AA concentration from less than 15 ppm to less than 3 ppm. Additional drying of the solid- state polymerized PET resin can ultimately yield an AA concentration of less than 1 ppm.

The final step is to use this PET resin to manufacture articles for consumers to use. For instance, preforms can be injection molded. These preforms will ultimately be stretch-

blow-molded into food or beverage containers. The melting and processing that occurs during injection molding results in some PET degradation; increasing the AA content to around 10 ppm or less. It is this 10 ppm of AA that is of concern to the PET container manufacturers for foods and beverages. If an excessive amount of AA migrates from the

PET container to the packaged contents, the taste of the packaged product could be undesirably altered. Figure 1-8 shows a graphical depiction of this example, showing how AA content changes throughout the lifecycle of PET.

Yields Amorphous Crystallization Crystallized Melt-Phase PET Pellets Polymerization PET Pellets [>20 ppm] [<15 ppm]

SSP

Dried SSP Drying Injection Molding SSP PET PET Pellets Pellets [<1 ppm] [<3 ppm]

[≤ 10 ppm] Figure 1-8: Acetaldehyde concentration during the lifecycle of PET

1.4.3 PET Degradation Routes That Generate Acetaldehyde

There are several identified PET degradation routes that result in the generation of acetaldehyde. These AA producing chemical reactions result from two of the three core degradation mechanisms of PET: thermal degradation and thermal-oxidative

degradation.1, 3 The factors that drive these reactions are: temperature, hydroxyl end- groups, diethylene glycol (DEG) molecules, vinyl ester end-groups, oxygen, DEG linkages, and free radicals.

1.4.3.1 Thermal Decomposition of Hydroxyl End-Groups

The chemical reaction in Figure 1-9 shows the thermal decomposition of hydroxyl end-

1, 3, 22 groups into carboxyl end-groups and AA (CH3CHO). The precursors to this reaction are the presence of hydroxyl end-groups and elevated temperature. Therefore, this reaction can be characterized as thermal degradation.

Figure 1-9: Thermal decomposition of hydroxyl end-groups to produce AA

1.4.3.2 Breakdown of Diethylene Glycol Molecules

Figure 1-10 shows how acetaldehyde can be generated from diethylene glycol molecules.1, 3, 22 In this reaction, two PET chains terminated by hydroxyl end-groups react to form a larger PET chain that is connected by an anhydride linkage. Also produced in this reaction is a DEG molecule. This DEG molecule can then undergo a dehydration reaction to produce ethylene oxide vinyl ether and water. The ethylene glycol vinyl ether molecule can subsequently decompose to produce two molecules of acetaldehyde.

Figure 1-10: AA formation from diethylene glycol (DEG) molecules

1.4.3.3 Reactions with Vinyl Ester End-Groups

Of all the factors that lead toward the generation of AA, researchers3, 22, 23 have shown that the most prominent is the concentration of vinyl ester end-groups in PET. Vinyl ester end-groups typically form during a random chain scission reaction; as illustrated in

Figure 1-11. As previously mentioned in Section 1.3.2, this reaction is characteristic of thermal degradation.

Once the vinyl ester end-group is formed, it can generate AA by two different mechanisms. The first route is by reacting with another PET chain terminated by a hydroxyl end-group. This reaction creates a larger PET chain, connected by an ethylene linkage, and a molecule of acetaldehyde. The second route occurs when the vinyl ester end-group reacts with a carboxyl end-group. This reaction ultimately yields a larger PET chain, connected by an anhydride linkage, and a molecule of AA.

Figure 1-11: AA Formation from vinyl ester end-groups reacting with hydroxyl end- groups and carboxyl end-groups

1.4.3.4 Presence of Oxygen

Not only can vinyl ester end-groups form by random chain scission and thermal degradation, they can also be produced by thermal-oxidative degradation.1, 3, 16 Oxygen that is present during the melting and processing of PET can react with the ethylene linkage of a PET chain to produce a branched hydroperoxide group. This hydroperoxide group may then decompose to form free radicals. As illustrated in Figure 1-12, these free radicals will eventually yield two PET chains, one terminated by a vinyl ester end-group and one terminated by a hydroxyl end-group. As previously shown in Figure 1-11, these functional groups will react with one another to yield a larger PET chain and a molecule of acetaldehyde.

Figure 1-12: AA generation due to the presence of oxygen

1.4.3.5 Presence of DEG Linkages within PET Chains

The process for synthesizing PET, melt-phase polymerization, is usually carried out between 270 to 285oC. At these elevated temperatures, it is common for a small amount of ethylene glycol (EG) to react with itself to form diethylene glycol (DEG). Since EG and DEG are both diols, it is possible for DEG to replace EG during the synthesis of the

PET chains. When this occurs, a DEG linkage connects the terephthalate groups of PET

rather than an EG linkage. The disadvantage of this linkage, however, is that it is susceptible to being attacked by oxygen; as shown in Figure 1.13.3

Figure 1-13: AA generation from diethylene glycol (DEG) linkage

When oxygen attacks the DEG linkage it forms a branched hydroperoxide group; similar to the one formed in Figure 1-12. With elevated temperature, this hydroperoxide group decomposes to form free radicals. In time, these free radicals lead to the formation of

PET chains terminated by vinyl ester end-groups and hydroxyl end-groups. As shown in

Figures 1-11, 1-12, and 1-13; ultimately the reaction of vinyl ester end-groups and hydroxyl end-groups produces a larger PET chain and a molecule of acetaldehyde.

1.4.3.6 Presence of Free Radicals

Figures 1-12 and 1-13 have shown how the presence of oxygen can lead to the generation of free radicals and eventually AA.3, 4 Figure 1-14 also shows a degradation reaction scheme that results from the presence of a free radical. This time, however, the free radical is generic and could have been produced by another mechanism; such as the exposure of PET to UV light.

When the free radical reacts with PET, as shown in Figure 1-14, it shifts from the generic species to the ethylene portion of the PET chain. The instability of the free radical causes the chain to split in two, forming a PET chain terminated by a vinyl ester end-group and a

PET chain terminated by an unstable end-group. In either case, as illustrated by Figure

1-14, the presence of the unstable end-group and the presence of the vinyl ester end- group will ultimately lead to the formation of acetaldehyde.

Figure 1-14: AA generation due to the presence of free radicals

1.4.4 Ways to Reduce the Amount of AA within PET

As previously mentioned in Sections 1.4.1 and 1.4.2, the amount of AA in PET is of great concern for manufacturers of food and beverage packaging. Acetaldehyde that is generated during manufacturing can, with time, diffuse from the PET container and into the packaged contents. It is known that an excessive amount of acetaldehyde can affect the desired taste of many food products.2, 20 The most extreme scenario exists for bottled

water distributors. The taste of pure water is very subtle and is unable to mask the taste of even a few parts per million of AA.21

There are a few common techniques that manufacturers use to limit the generation of AA for PET packages. The first solution is to optimize the processing conditions. This includes minimizing the melt temperature, shear rate, and residence time that the polymer is exposed to during injection molding.13, 15 Optimizing these parameters limits the temperature and amount of time the polymer is exposed to these harsh conditions.

Of all the AA producing degradation mechanisms discussed in Sections 1.3 and 1.4.3, thermal degradation has the greatest impact on the generation of AA. Thermal degradation leads to random chain scission reactions, resulting in the formation of vinyl ester end-groups. As mentioned in Section 1.4.3.3, the vinyl ester end-group concentration has been shown to have the most direct influence on the amount of AA that will be generated.23

A second solution to limiting AA formation is to use of low AA generating PET resins, sometimes referred to as “water-grade” resins. These are resins which have been

“specifically tailored to the needs of water containers”.2 One of these particular requirements is limiting the amount of AA that is generated during container manufacturing.

There are instances, however, where acetaldehyde levels are required to be lower than any optimization of conditions or resin can achieve. In these situations, compounds called AA scavengers can be melt-blended into the PET matrix. These additives do not limit the degradation of PET or the generation of AA; acetaldehyde scavengers work by interacting with AA to reduce the levels at which it is released from PET.20, 24

1.5 Rationale and Objectives

A study by Suloff24 examined the effectiveness of three different aldehyde scavenging agents: poly(meta-xylene adipamide) (MXD6), D-sorbitol, and alpha-cyclodextrin. He prepared his samples by dry-blending each scavenging agent with PET pellets and then thermally pressing those blends into films. Each film was stored in its own aqueous mixture, comprised of various aldehydes, for up to 14 days. He quantified the aldehyde sorption ability of each scavenger by monitoring the change in concentration of each aldehyde, within each solution, over time. Suloff24 showed that aldehyde scavenging/PET films were more effective at reducing aldehyde concentrations than control films. He also showed that the scavenging agents preferred smaller molecular weight aldehydes to larger molecular weight aldehydes.

Suloff’s work24 examined the ability of his thermally pressed scavenging agent/PET blend films to remove aldehydes from an aqueous solution. The purpose of the current research project is to expand upon Suloff’s work, while focusing solely on acetaldehyde

(AA) scavenging in PET. The goal is to develop a comprehensive understanding of the

overall effects of melt-blending AA scavengers into poly(ethylene terephthalate).

Through this research an understanding of the influence that the AA scavengers have upon the physical properties of PET will also be developed. These physical properties include: thermal properties, thermal stability, intrinsic viscosity (I.V.), barrier properties, color, and physical appearance.

Through the knowledge obtained by studying the interactions that occurs between acetaldehyde scavengers and PET, a greater understanding will be achieved and the overall benefit of adding the scavengers can be specifically evaluated. In addition, greater understanding of AA scavengers should help in the development of better sequestering systems. It is also through this research that an initial model will be developed for a multi-cavity injection molding system. This model will be used to determine the amount of AA scavengers that will be needed to melt-blend with PET in order meet desired AA concentration requirements for various packaging systems.

To meet these goals, the following five objectives have been identified as the focal points of this work:

1. Determine the chemical interactions/reactions that occur between the various

scavenging agents and AA.

2. Investigate the effectiveness of the AA scavengers in reducing the amounts of

acetaldehyde that are present in PET.

3. Study any changes in the physical properties of PET due to the addition of the AA

scavengers.

4. Determine the optimal amounts of AA scavengers to melt-blend with PET resins

during injection molding.

5. Create a predictive model to quantify the overall effectiveness of the AA

scavengers.

Chapter 2

Literature Review

2.1 Minimizing AA During PET Melt Processing

There has been a significant amount of research focusing on reducing the presence of acetaldehyde (AA) in PET. The majority of this work has concentrated on the degradation routes that generate AA as a byproduct. As previously mentioned in Section

1.4.3, thermal degradation and thermal-oxidative degradation both lead to the generation of AA.1, 3 Thermal degradation and thermal-oxidative degradation are two of the three main degradation routes of PET. Hydrolytic degradation, the last of the three main PET degradation routes, does not directly lead to the formation of AA. In fact, research has shown that the presence of water during PET processing actually reduces the amount of

AA that is generated.1

2.1.1 Limiting Thermal Degradation

As previously stated in Section 1.3.2, thermal degradation of PET occurs when the polymer is exposed to excessive temperatures. Researchers have shown that exposure to these extreme conditions results in random chain scission reactions. Marshall and Todd25

believed these reactions occurred at the end of the polymer chains. Both Goodings17 and

Ritchie26 felt that thermal degradation reactions occur at the ester linkages. Although the true location of this degradation mechanism may be of debate, it is agreed that random chain scission reactions lead to the generation of additional carboxyl and vinyl ester end- groups.

Thermal degradation is known to have several other effects upon PET besides the generation of additional carboxyl and vinyl ester end-groups. With the addition of these generated end-groups comes a reduction in the polymer’s molecular weight and intrinsic viscosity (I.V.). Thermal degradation has also been known to cause discoloration within the polymer, as well as the formation of oligomers and low molecular weight byproducts.1, 3

One of the low molecular weight byproducts that are formed through the thermal degradation of PET is acetaldehyde. AA has been shown to be the most prominent of all the byproducts generated through thermal degradation.17 Some researchers have determined that AA comprises 80% of all the generated, gaseous byproducts. During thermal degradation, the formation of AA occurs predominantly by means of the vinyl ester end-groups.23 As illustrated in Figure 1-11, in Section 1.4.3.3, the vinyl ester end- group can generate AA by reacting with either a carboxyl end-group or a hydroxyl end- group.

Shukla, et al15 performed an exhaustive study examining the effects that various injection molding parameters can have on the degradation of PET and the generation of AA. Their work revealed that increasing the processing temperature by 10oC will cause the AA concentration within PET to double. In addition, Shukla, et al15 showed there exists strong relationships between an injection molder’s shear rate and the generation of AA, as well as between the polymer’s processing time and the amount of AA that is generated.

Intuitively, to minimize the effects of thermal degradation and the amount of generated

AA, a balance must be made between the processing temperature, the residence time, and the shear heating that occurs during extrusion or injection molding.

2.1.2 Limiting Thermal-Oxidative Degradation

While hydrolytic degradation is reported to be the most aggressive form of PET degradation, researchers14, 16 have shown thermal-oxidative degradation to be a more disruptive than thermal degradation. Thermal-oxidative degradation, as previously discussed in Section 1.3.3, occurs when oxygen reacts with PET at elevated temperatures.

By comparing various melting environments, Jabarin and Lofgren14 showed that thermal- oxidative degradation had higher reaction rates than thermal degradation; indicating that degradation occurs more rapidly. Yoda, et al16 found that thermal-oxidative degradation can lead to the formation of branched and cross-linked chains; and in some circumstances, thermal-oxidative degradation can even lead to gel formation. Branched/cross-linked chains and gel formation only result from thermal-oxidative degradation.

Both thermal degradation and thermal-oxidative degradation share many common traits.1

First, both mechanisms require excessive temperature to degrade the PET chains. Second, for the most part, both degradation systems result in similar effects upon PET: reduction of I.V. or molecular weight, formation of carboxyl end-groups, discoloration, and generation of low molecular weight byproducts. Third, both degradation systems produce AA as one of their more prominent byproducts.

Figure 1-12, located in Section 1.4.3.4, details the reaction scheme from which AA is produced by thermal-oxidative degradation. The exposure to elevated temperatures leaves the ethylene linkage of a PET chain susceptible to be attacked by oxygen. The formed branched hydroperoxide group will decompose and form free radicals. The formation of two free radicals causes the PET chain to split in two. One PET chain is terminated by a vinyl ester end-group and the other is terminated by a hydroxyl end- group. These functional end-groups then react with one another to re-form a PET chain and a molecule of AA.

Ideally, to minimize the chance of thermal-oxidative degradation from occurring, PET should be melted and/or processed under an inert environment. An example of this is the continual flushing of the headspace of an extruder with nitrogen gas to displace any oxygen. Vacuum pressure could also be applied in a similar manner; continually removing oxygen and any generated gaseous byproducts.

2.2 Minimizing AA by Choice of Polymerization Catalysts

The manufacturing of PET requires the use of various metal acetates and/or oxides to drive both the esterification and transesterification reactions.1, 3, 4, 27 Examples of these catalysts have been previously mentioned in Section 1.2.1.1. It is well documented that catalyst systems play a critical role in the degradation of PET, and eventual generation of

AA. In separate publications, Zimmermann27, Rieckmann and Völker4, and Jabarin3 all provide a summary of these investigations.

Zimmermann and co-workers27 have extensively studied both the thermal and thermal- oxidative degradation of PET. This work was performed by using a variety of catalyst systems. Zimmermann identifies cobalt (Co), cadmium (Cd), nickel (Ni), and zinc (Zn) as the most aggressive catalysts which lead to PET degradation. Kao, et al28 verified

Zimmermann’s work, reporting that the use of Co, copper (Cu), and Zn acetates increase the rate of degradation within PET. Derivatives of antimony (Sb), meanwhile, have been acknowledged by Zimmermann and Kim29 as catalysts that do no accelerate PET degradation. Separately, Granzow, et al30 and Zimmermann27 have shown that phosphorus (P) based additives are able to provide stabilization and reduce the rate of degradation.

2.3 Minimizing AA by Choice of PET Resins

As previously mentioned in Section 1.4.4, one avenue toward limiting the generation of

AA is through the use of specifically designed PET resins.2 These PET resins are sometimes referred to as “water-grade” resins, since they have been “specifically tailored to the needs of water containers”.2 Since water does not require carbonation, the strength requirements of the PET container can be lessened; compared to the strength requirements of carbonated soft drink (CSD) PET containers. The PET water bottles need only enough strength to house the packaged water and survive an impact.

Fundamentally, a reduction in the strength requirements for PET packaging means that the material’s molecular weight or I.V. can be reduced. Generally, “water-grade” PET bottles can be manufactured with resins that range in I.V. between 0.70 and 0.76 dL/g; whereas, bottles manufactured for CSD packaging require an I.V. between 0.80 and 0.84 dL/g. Some20 believe that since these “water-grade” resins have lower intrinsic viscosities, they are inherently exposed to less shear heating than are CSD grade resins, during processing. Shukla, et al15 showed that minimizing shear heating can significantly reduce the amount of AA that is generated in PET. Other researchers,1, 31 however, feel that AA generation is unrelated to molecular weight or I.V. It is their belief that chemical composition (catalysts, monomers, etc.) is the dominant factor in whether a PET resin produces more or less AA than another resin.

2.4 Minimizing AA by Means of Acetaldehyde (AA) Scavengers

Acetaldehyde (AA) scavengers are additives, which when melt-blended into PET resin are designed to interact with any acetaldehyde that is present in the PET matrix.20, 24

There are at least three different mechanisms through which this interaction occurs. One type of scavenger is designed to react with any generated AA and the second is designed to lock AA into its structure. Both of these systems do not minimize the amount of AA that is generated during the processing of PET; they simply limit its release by interacting with generated AA and thus its effects upon the packaged contents. A third type of scavenger is a catalyst system that converts AA into another compound which possesses different migration and/or flavor threshold properties.

2.4.1 Reactive AA Scavengers

2.4.1.1 Polyamides

Polyamides have been identified by several researchers32-46 as possible AA scavengers for PET. Several Eastman Chemical patents describe the use of polyester/polyamide blends to improve the flavor retention for polyester packaging applications. U.S. Patent

6,042,90832 identifies an exhaustive list of possible low molecular weight partially aromatic polyamides and low molecular weight aliphatic polyamides that can be used as

AA scavengers. Of the two general classifications listed, Long, et al32 states that aromatic polyamides are more preferable than aliphatic polyamides because the aromatic

polyamides tend to be more easily dispersed and produce less haze. The patent describes blends having a polyamide composition between 0.05% and 2%, by weight; the remaining balance of the blend is polyester.

Long, et al32 suggests two particular polyamides to be the most effective at AA scavenging: poly(hexamethylene adipamide) and poly(meta-xylene adipamide) (MXD6).

It is recommended that the number average molecular weight for poly(hexamethylene adipamide) should be between 3,000 to 6,000 g/mol and an inherent viscosity of 0.4 to

0.9 dL/g. Long claims that MXD6 should possess a number average molecular weight between 4,000 to 7,000 g/mol and an inherent viscosity of 0.3 to 0.6 dL/g. The chemical structure for MXD6 is shown in Figure 2-1.

H2N H2C CH2 NH OOC (CH2)4 COOH

Figure 2-1: The chemical structure of MXD6

Polyamides, particularly MXD6, have been added to PET to help improve its barrier properties.47 When polyamides are melt-blended with PET a yellowing phenomenon of the polymer blend typically results. Bandi, Mehta, and Schiraldi48 studied the mechanism from which chromophores are generated when PET is melt-blended with polyamides. Of particular interest were PET/MXD6 blends. Bandi, et al48 linked the color generation to the formation of an imine group, which is the result of a reaction between the amine end- group of MXD6 and the aldehyde group of AA. It is through this reaction, shown in

Figure 2-2, that MXD6 exhibits its use as an AA scavenger when melt-blended with PET.

It should be noted that the reaction shown in Figure 2-2 does generate a byproduct, water

(H2O).

H2N H2C CH2 NH OOC (CH2)4 COOH + H3C CH

(MXD6) (AA)

H3C HC N H2C CH2 NH OOC (CH2)4 COOH

Figure 2-2: Acetaldehyde scavenging reaction between MXD6 and AA24

Other polyamides, besides MXD6, have also been evaluated as AA scavengers for PET.

In WO Patent 9728218, Turner and Nicely44 proclaim the scavenging ability of polyester/polyesteramide blends. The addition amount of the polyesteramide ranges from

0.05% to 2% by weight. Ciba Specialty Chemicals36, 42, 43, 45 patented polyacrylamide, polymethacrylamide, and copolymers of polyacrylamide and polymethacrylamide as AA scavengers for PET. Both polyacrylamides and polymethacrylamides possess a branched amine groups that allow the polymers to scavenge AA in a similar manner as MXD6.

U.S. Patent 7,022,390 teaches that optimal polymethacrylamide concentration can vary between 0.03% to 1%, by weight, when melt-blended with PET.

2.4.1.2 Low Molecular Weight Amides

U.S. Patent 7,550,20349 lists several low molecular weight amides that can potentially be used to scavenge generated AA within PET. This list includes: anthranilamide, malonamide, salicyclamide, o-mercaptobenzamide, N-acetylglycinamide, and 2- aminobenzenesulfonamide. Of particular interest is anthranilamide, which was patented by Rule, Shi, and Huang50 for its use as an AA scavenger.

As shown in Figure 2-3, the chemical structure of anthranilamide is comprised of two functional groups, an amide group and an amine group, attached to a substituted benzene ring. According to Rule, et al50, the reaction between anthranilamide and acetaldehyde produces “water and a resulting organic compound comprising an unbridged five or six member ring including the at least two heteroatoms”. The patent goes on to described two more details: one, “the unbridged 5 or 6 member ring of the resulting organic compound is bonded to the aromatic ring” and, two, “the two heteroatoms are both nitrogen”.

NH2

Figure 2-3: Chemical structure of anthranilamide

According to Rule, et al50, it is the formation of this second ring structure that makes anthranilamide an appealing AA scavenger. Patent 6,274,212 claims, “Unlike the prior art methods that depend on the formation of inherently colored imines, the formation of unbridged 5 or 6 member ring structures do not inherently result in color formation”. It was the generation of imine functional groups that Bandi48 suggested was the reason for color generation in PET/polyamide blend systems. In addition, Rule, et al50 also claims that “thermodynamics often favor ring formation more than imine formation; thus, significantly lower amounts of the organic additive compound of this invention can effectively decrease the AA content of melt-processed polyesters”. Addition levels for anthranilamide are recommended to be between 10 and 1000 ppm.

2.4.1.3 Polyamines

Patent literature shows that polyamines can be added to oxygen scavenging systems to react with and stabilize the byproducts that result from those reactions. In U.S. Patent

5,942,297 Speer, et al51 identifies a list of polyamines that can react with the various aldehydes and alcohols that are formed during oxygen scavenging reactions. This list includes: polymers and copolymers of allylamine, polymers and copolymers of diallylamine, polymers and copolymers of vinyl amine, poly(D-glucosamine), silica- supported polymeric amines, and amine functionalized silicas. The scavenging reaction between terminal amines and aldehydes follow a chemistry similar to that in the reaction scheme shown in Figure 2-2. This condensation reaction generates an imine compound and water, as a byproduct.

Similar to the application depicted by Speer, et al51, Ching, et al52 claims a similar use for polyamines. U.S. Patent 6,057,013 describes a multi-layer oxygen scavenging system in which one of the layers is comprised of an oxygen scavenging material, a second layer is comprised of byproduct neutralizing materials, a third layer is an oxygen barrier, and a fourth layer is a polymeric selective barrier. The intent of adding these byproduct neutralizing materials is to prevent byproducts of the oxygen scavenging reactions from diffusing through the multi-layer structure and into the packaged contents. The list of neutralizing materials listed in Patent 6,057,01352 includes low molecular weight amines, amine-containing polymers, and polyamines. The identified low molecular weight amines are: dipropylenetriamine, tris(3-aminopropylene)amine, N,N,N’N’-tetrakis(3- aminopropyl)ethylenediamine and 1,12 dodecanediamine. The identified amine- containing polymers are: glycols containing amine groups such as polyethylene glycol with two amines and polypropylene glycol with two amines, and dimethylaminoethanol grafted ethylene-methyl acrylate copolymers. The identified polyamines are: pentaethylene hexamine (PEHA), triethylene tetraamine, polyvinyl oxazoline, and comparable higher molecular weight products.

2.4.1.4 Polyimines

In U.S. Patent 5,362,784, Brodie, et al53 teaches that a specific class of polyimines can be used to scavenge aldehydes in polyesters. The class of polyimines mentioned by Brodie, et al is known as polyalkylene imines (PAI); more specifically the authors talk of polyethylene imine (PEI). It is recommended that PEI has an average molecular weight

of at least 2500 g/mol, is not heavily branched, and has a low amount of tertiary amines.

The optimal blending of PEI and PET is claimed by the authors to be between 0.01 and

10:100, by weight. It is mentioned within Patent 5,362,78453 that the use of binding agents can be used to “lock in” the polyalkylene imines into the polymer matrix.

It should be noted that Brodie, et al discuss the use of PAI and PEI as aldehyde scavengers in two more U.S. Patents; 5,284,89254 and 5,413,82755. The application of these two patents, however, concern blending the polyimines with polyolefin materials.

2.4.1.5 Polyols

Polyols are a class of alcohols that contain multiple hydroxyl groups. The AA scavenging ability of these compounds has been claimed by several researchers in patent literature. McNeely and Woodward56 describe a PET modified by an alkoxylated polyol; intended to enhance the material’s melt-strength. The specific polyols mentioned in U.S.

Patent 5,250,333 are alkoxylated trimethylolethane, alkoxylated trimethylolpropane, alkoxylated pentaerythritol, and the alkoxylated dimmer of pentaerythritol. McNeely and

Woodward teach that their optimal concentration is between 100 and 50,000 ppm.

While McNeely and Woodward56 did not claim the AA scavenging ability of polyols in their patent, Al-Malaika’s patents57, 58 do state that ability. In WO Patent 006665957 and

U.S. Patent 6,936,20458, Al-Malaika describes the use of polyol/PET blend systems as a way to reduce AA. These multiple hydroxylic compounds should ideally possess

between 3 to 8 hydroxyl groups and be present between 0.0001 and 2%, by weight. The specific polyols mentioned in these patents are triglycerin, trimethylolpropane, dipentaerythritol, tripentaerythritol, D-mannitol, xylitol, and D-sorbitol.

In WO Patent 010072, Eckert, et al59 also describes the use of polyols as AA scavengers in PET systems. The authors of this patent claim that desired polyols have at least one primary hydroxyl group and another primary, secondary, or tertiary hydroxyl in the 2 and/or 3 position. The identified list of polyols that fit the criteria includes xylitol, mannitol, and sorbitol. To reduce AA, the concentration of these additives can range between 50 and 5000 ppm; although polyol content can be as high as 25% by weight.

An example of the aldehyde scavenging ability of a polyol is shown in Figure 2-4. In this reversible reaction scheme, two molecules of D-sorbitol react with one molecule of valeraldehyde to produce a higher molecular weight acetal and a byproduct, water. A similar reaction occurs when AA is the sought after compound. This also yields water as the byproduct.

Odorisio and Andrews60 also identified a group of polyols that can be used as AA scavengers in PET. This group consists of homo- and copolymers of polyhydric alcohol- containing polymers which are derived from 2-propenoic acid ester monomers. Within this group, the most preferable is poly(glyceryl methacrylate) homopolymer. U.S. Patent

7,138,45760 teaches that poly(glyceryl methacrylate) homopolymer’s ideal concentration can vary between 0.01 and 5%, by weight.

Figure 2-4: Aldehyde scavenging reaction between D-sorbitol and valeraldehyde24

2.4.2 Size-Enclosing AA Scavengers

2.4.2.1 Cyclodextrins

Cyclodextrins are cyclic oligosaccharides consisting of six, seven, or eight repeat glucopyranose units.24 Figure 2-5 shows the how the classification of cyclodextrin changes with the number of glucopyranose repeat units; it also shows how the internal and external diameters change with the number of repeat units.61 Figure 2-6 shows the chemical structure of alpha-cyclodextrin.

Figure 2-5: Various forms of cyclodextrin62

Figure 2-6: Chemical structure of alpha-cyclodextrin24

Wood, et al have several patents63-70 claiming that cyclodextrin can encapsulate AA, as wells as other permeates, into its molecular structure. An example of this interaction mechanism is shown in Figure 2-7, with caproaldehyde as the guest molecule.

Cyclodextrin does not inhibit the generation of AA during the melt processing of PET.

As with other AA scavengers, it merely limits the diffusion of AA out of the polymer’s

matrix.62 According to Wood, et al, the optimal amount of cyclodextrin melt-blended into PET ranges between 0.1 to 5 %, by weight.

Figure 2-7: Proposed fitting of caproaldehyde and cyclodextrin24

Several authors71-74 have shown that 1H NMR (proton nuclear magnetic resonance spectroscopy) can be used to validate cyclodextrin’s size-enclosing mechanism through

NMR titration studies. NMR titration studies involve varying the concentration of the inclusion species (guest) to the concentration of cyclodextrin (host). 1H NMR spectra are recorded for several samples of varying host to guest ratios. A plot is then made of the change in the chemical peak shift (y-axis), in ppm or ∆δ units, versus the host to guest ratio (x-axis), in units of concentration.

One such NMR titration study was performed by Hao, et al71 who studied the complex formation of 2, 2’-ditelluro-bis(β-cyclodextrin) (2-TeCD) with glutathione (GSH). This study was performed at room temperature, using deuterium oxide (D2O) as the solvent.

Figure 2-8 shows the change in the chemical shift (∆δ) for a proton of the guest molecule

(GSH), labeled H5 proton, as the concentration of the host molecule (2-TeCD) is increased.

Figure 2-8: 1H NMR titration experiment of 2-TeCD and GSH71

NMR measures the overall chemical environment which a particular hydrogen or carbon atom experiences; depending on the analysis method: proton (1H) NMR or carbon-13

(13C) NMR. It does this by averaging the chemical shift seen for each particular proton or carbon-13 atom. When the guest molecules are dissolved in a solvent, the 1H NMR yields a standard chemical shift for each of its protons. This particular chemical shift is based on the chemical environment which those protons experience in that particular solvent.

As host molecules are added, guest molecules move from the solvent to the internal structure of the host molecules. While the guest molecules are inside the internal structure of the host molecule, hydrogen bonding occurs between the protons of the guest

molecules and the protons which line the internal structure of the host molecules. The

NMR instrument interprets this hydrogen bonding as a change in the chemical environment of the guest molecule’s protons. The instrument responds with a different chemical shift for the guest molecules’ protons than previously seen when they were in solvent alone. As the concentration of the host molecules increases, the number of guest molecules which occupy the internal structure of the host molecules also increases. Since the NMR instrument averages the chemical shift for each particular proton, the change in the chemical shift changes correspondingly; as shown in Figure 2-8.

The change in the chemical shift, of the guest molecules’ protons, occurs until a saturation point is reached. This saturation point is achieved when all of the guest molecules are occupying the internal structure of host molecules; thus changes in the chemical shift of the guest molecules protons can no longer occur. The saturation point is indicated in Figure 2-8 as the position where the slope of the graph flattens and it equals zero. Figure 2-8 shows that the saturation point for Hao’s experiment71 occurs when a

2:1 host-to-guest ratio is achieved.

2.4.2.2 Zeolites

Zeolites75 are a class of microporous, crystalline solids that occur both naturally and synthetically. They are generally composed of aluminum (Al), silicon (Si), and oxygen

(O). The polarity of their internal pore allows cations such as sodium (Na), calcium (Ca), and potassium (K) and small molecules such as water to fill their internal cavities.

Massey and Callander76 noted that the propensity for zeolites to absorb water, into their internal cavities, is so strong that if water is absent from the system zeolites will allow any material into their cavities as long as it is small enough to fit through the pores and enter the internal structure. This filter mechanism allows zeolites to act as “molecular sieves” and separate smaller molecules from larger ones.

In U.S. Patent 4,391,971, Massey and Callander76 teach of a process for reducing AA content by passing molten PET through a filter containing a molecular sieve. The described molecular sieve is a zeolite composition and is claimed to be located “between an extruder outlet and a receiving mold of forming die”. Massey and Callander76 also state that their process will improve the brightness and color of the resulting PET resin.

While Massey and Callander76 pass molten PET through a zeolite based filter to reduce

AA content, Mills, et al77 teach that zeolites can also be melt-blended into the polymer’s matrix to achieve the same result. The authors claim that the addition of small- or medium-pore zeolites can aid in minimizing AA concentration without reducing the clarity of the final PET article. Small-pore zeolites have an eight tetrahedral structure with an internal diameter of 4.1 angstroms (Å); while medium-pore zeolites have a ten- ring system with an ellipsoidal tubular diameter of 5.5 Å by 5.6 Å. WO Patent

Application 942937877 teaches that the optimal addition amount of these zeolites is between 100 and 1000 ppm.

2.4.3 AA Scavenging Catalysts

Researchers have shown that certain catalyst systems can also act as AA scavengers. As previously discussed in Section 1.2.1.1, the catalysts used for PET polymerization include: acetates of antimony (Sb), lithium (Li), calcium (Ca), magnesium (Mg), zinc

(Zn), and lead (Pb); as well as oxides of Sb, Pb, and germanium (Ge). AA scavenging catalysts are not added to assist in the polymerization of PET; there sole intent is to convert the molecular structure of AA into another product which has different migration and/or flavor threshold properties.

Go and Burzynski78 describe a method for manufacturing PET resins with enhanced thermal stability against the generation of AA. Their process incorporates an alkali metal salt of ethylenediaminetetraacetic acid (EDTA) into melt-phase polymerization. U.S.

Patent 4,357,46178 teaches that the alkali metal should be either sodium (Na) or potassium (K). The authors claim that the amount of EDTA can range between 0.001 and 0.2 mol percent. They indicate that amounts above 0.5 percent can lead to discoloration.

In U.S. Patent 6,569,47979 and W.O. Patent 013090080, Rule teaches that an oxidation catalyst can be used to convert AA to acetic acid. Rule claims that the conversion of AA to acetic acid occurs as oxygen gas permeates the PET packaging. Therefore, this reaction is also capable of decreasing the migration of oxygen through PET containers.

Converting AA to acetic acid is appealing because acetic acid has a taste threshold that is

more than 1000 times greater than acetaldehyde. The oxidation catalysts described in these patents are simple cobalt (Co) or manganese (Mn) salts; or Co or Mn salts comprised of an amine, phosphine, or alcohol complex. Rule79, 80 notes that simple salts can include: cobalt acetate, cobalt octoate, cobalt naphilhenate, manganese acetate, manganese octoate, and manganese naphthenate. Examples of amine complexes include

EDTA and glycine. The concentration of these catalysts should be between 1 and 500 ppm, but most preferably between 5 and 50 ppm.

Rule and Shi81 describe a PET resin comprised of a hydrogenation catalyst and at least one source of reactive hydrogen. In this application, the hydrogenation catalyst can be a

Group VIII metal (zero valent nickel, palladium, or platinum) or a metal hydride (tin or titanium based complex). According to U.S. Patent 7,041,350, the catalyst’s amount can range between 0.1 and 100 ppm; optimally between 5 and 50 ppm. The identified reactive hydrogen source is poly(methylhydro)siloxane (PMHSO) and its concentration should be 1 to 50 time greater than the amount of AA.

Chapter 3

Experimental Work

3.1 Materials

The acetaldehyde (AA) scavenging capabilities of three materials were evaluated.

Anthranilamide was purchased from Sigma-Aldrich, meta-xylenediamine (MXDA) was donated by Mitsubishi Gas Chemical America, Inc, and alpha-cyclodextrin was supplied by the Wacker Chemical Corporation. Acetaldehyde was also purchased from Sigma-

Aldrich and used to spectroscopically study the chemical interactions that occur between

AA and the various scavenging materials. The spectroscopic analysis, particularly NMR, was conducted in the presence of deuterated solvents. Deuterium oxide (D2O) and deuterated chloroform (CDCl3) were both purchased from Cambridge Isotope

Laboratories.

The various AA scavengers were also melt-blended with a commercially available poly(ethylene terephthalate) (PET) resin. Voridian CB12, produced by Eastman

Chemical, is a PET copolymer resin with an initial I.V. of 0.84 dL/g. It is designed to be used in the manufacturing of carbonated soft drink (CSD) containers.

3.2 Spectroscopic Techniques to Study Chemical Interactions

AA scavengers work by interacting with the acetaldehyde that is present within PET.

These chemical interactions prevent AA from diffusing from the matrix of the polymer and into either the packaged contents of the container or into the atmosphere. Studying and determining these chemicals interactions was achieved through two spectroscopic techniques: nuclear magnetic resonance (NMR) spectroscopy and mass spectroscopy.

NMR provided the primary information, while mass spectroscopy was used to supplement the NMR data.

3.2.1 Nuclear Magnetic Resonance (NMR) Spectroscopy

Nuclear magnetic resonance (NMR) spectroscopy is a powerful analytical tool used to perform structure elucidation and determine the identity of unknown compounds.82 NMR works on the premises that certain nuclei (1H, 13C, etc.) have magnetic properties when in the presence of a magnetic field.83 When an electromagnetic pulse is applied, energy is absorbed by the nuclei and then sent back out. The energy is remitted at a specific resonance, explicit to the studied nuclei’s chemical environment.82, 83

The most commonly studied nuclei in NMR are 1H and 13C. The appeal of studying these two nuclei is due to their relative abundance within organic compounds. Although 1H

NMR is about 5700 times more sensitive than 13C NMR83, both techniques are very powerful and capable of providing significant data. These NMR methods are sometimes

referred to as “one-dimensional” techniques. The reason for this designation is because the resulting spectrum from 1H or 13C NMR appears relative to only one axis or scale.

In addition to “one-dimensional” NMR, there also exists a family of “two-dimensional”

NMR techniques. “Two-dimensional” NMR yields spectra that are shown relative to two scales; one scale is on the x-axis and the other scale is on the y-axis. The data can be correlated between these two axes to show which nuclei are coupled or connected to one- another83; information that “one-dimensional” techniques may not be able provide.

“Two-dimensional” NMR techniques enhance the information that “one-dimensional” methods provide, making it possible to decipher the chemical structure of compounds too difficult with only “one-dimensional” NMR spectra.

3.2.1.1 Proton (1H) NMR

In order to properly identify the interaction mechanisms between the various AA scavengers and acetaldehyde, proton nuclear magnetic resonance (1H NMR) was used.

1H NMR was chosen over 13C NMR for two reasons: one, 1H NMR is approximately

5700 times more sensitive than 13C NMR and two, 1H NMR is able to provide quantitative data; something 13C NMR cannot always provide.82 For this project, the

NMR Varian Inova 600 MHz spectrometer, located in the University of Toledo’s

Instrumentation Center, was used.

The 1H NMR work began by analyzing the various AA scavengers and acetaldehyde independently. Each material was dissolved in an NMR tube using an appropriate deuterated solvent. Its spectrum was then obtained and analyzed to characterize each 1H

NMR peak for the given compound.

After obtaining the spectrum for each respective component, the next step was to mix acetaldehyde with each scavenger and study any changes in the resulting 1H NMR spectrum. Both acetaldehyde and an AA scavenger were separately dissolved using the same deuterated solvent. The weights of each component were recorded so that the proper stoichiometric ratio was achieved. These solutions were then mixed together in an

NMR tube and analyzed.

3.2.1.2 1H-1H COrrelation SpectroscopY (COSY)

Proton–Proton (1H-1H) COrrelation SpectroscopY (COSY) is a common “two- dimensional” NMR technique. This experiment reveals the coupling/connection pattern of the protons that compose the studied chemical compound83; helping to map-out the chemical structure. The output of this experiment is the sample’s 1H NMR spectrum on both the y- and x-axes. Since these spectra are obtained by the same magnetization, there is a symmetry that yields “diagonal peaks”. The “off diagonal peaks” or “cross peaks” are what indicate the coupled protons of the analyzed sample.

1H-1H COSY NMR was used to help decipher those 1H NMR spectra that were, alone, too hard to interpret. The sample preparation for this experiment was the same as for 1H

NMR. Separately, acetaldehyde and an AA scavenger were dissolved using the same deuterated solvent. These two solutions were stoichiometrically mixed together and then analyzed. Analysis was conducted via the NMR Varian Inova 600 MHz spectrometer, located in the University of Toledo’s Instrumentation Center.

3.2.2 Mass Spectrometry

Mass spectrometry is a versatile, analytical tool that can be used to both qualitatively and quantitatively determine an unknown analyte based upon its mass.82, 83 Upon entering a mass spectrometer, a chemical compound is ionized by some ionization method incumbent within the machine. These generated ions are then separated, by the instrument’s separations method, based upon their mass to charge ratio.82 The spectrometer’s detector subsequently quantifies the amount of ions at each mass to charge ratio, yielding a chart illustrating the abundance of each ion at the various ratios.83

There are various types of mass spectrometers, each possessing a different ionization method and a different mass analyzer.82, 83 For this project, the Esquire liquid chromatograph/mass spectrometer (LC-MS), located in the University of Toledo’s

Instrumentation Center, was used. This mass spectrometer possesses electrospray ionization, as its ionization source, and an ion trap, as its mass analyzer. The intent of the

mass spectrometry experiments was to compliment and validate the results seen from the

NMR work.

The preparation of each sample was as follows. A small amount of the sample, typically between 10 to 15 milligrams, was dissolved in a solvent; usually chloroform. Next, to aid with the ionization process, a very small aliquot of methanol was added to the dissolved sample. Prior to analysis, the mass spectrometer was purged with pure solvent, in order to clean the instrument of any residual impurities. Experimental samples were then directly injected into the mass spectrometer. The resulting spectrum was recorded and analyzed.

3.3 Twin-Screw Extrusion

The melt-blending of AA scavengers into PET resin was performed by means of a

Werner and Pfleiderer ZSK-30 twin-screw extruder; operating at 300 revolutions per minute (rpms) and 280oC. Attached to this extruder are a nitrogen tank and a vacuum pump. Nitrogen gas is fed into the throat of the extruder to displace any oxygen, creating an inert environment for PET melting and processing. The vacuum pump is attached to the barrel of the extruder and is used to remove any volatile chemicals that are generated during PET processing.

Experimental samples were prepared by melt-blending each of the three scavenging agents with the CSD PET resin. For each AA scavenger, a “master-batch” AA

scavenger/PET blend sample was initially extruded. Further concentrations of each AA scavenger/PET blend system were made by diluting the appropriate “master-batch” blend.

A determined amount of the “master-batch” sample was re-extruded with a known amount of virgin PET. For comparative purposes, control samples were also extruded under similar conditions. These samples were the same CSD PET resin, however, without the addition of any AA scavenging agents.

Prior to extrusion, all PET samples were dried overnight at 150oC and under vacuum; in either a Conair Franklin desiccant hopper/dryer or a small vacuum oven. This step was taken to limit the presence of moisture and thus minimize the effect of hydrolytic degradation during extrusion.

3.3.1 Preparation of Alpha-Cyclodextrin/PET Blend Samples

To prepare the alpha-cyclodextrin/PET master-batch sample, determined amounts of pure resin and the AA scavenger were weighed and then placed in separate vacuum ovens.

The virgin PET sample was dried overnight at 150oC, while alpha-cyclodextrin was dried at 80oC. After drying, the alpha-cyclodextrin and PET samples were removed from their respective ovens and dry-mixed in a small metal bucket. The mixing was done as quickly as possible to minimize the resin’s absorption of moisture from the atmosphere. This

PET blend was then extruded, pelletized, and collected. The concentration of alpha- cyclodextrin within this master-batch sample was 5 % by weight.

Through the dilution of this 5 weight % alpha-cyclodextrin/PET master-batch sample, further blend concentrations were obtained. A 2.5 weight % blend was produced by melt-blending equal parts, by weight, of the 5 weight % master-batch sample with the virgin, CSD PET resin. Each of these components was weighed and placed in a metal bucket that was dried overnight in a small vacuum oven, at 150oC. The next morning this sample was extruded, pelletized, and collected. A similar process was followed to yield 1 weight % (or 10,000 ppm) and 0.5 weight % (or 5000 ppm) alpha-cyclodextrin/PET blend samples.

To obtain the least concentrated alpha-cyclodextrin/PET samples, the 1 weight % sample was used as the concentrate instead of the 5 weight % master-batch sample. The thought was that by using a more diluted blend the AA scavenger would have a better chance of being uniformly dispersed, within the PET resin, during the extrusion process. A better distribution of the AA scavenger should provide a more homogeneous blend, limiting variability within the pellets.

This methodology may have increased the possibility of acquiring more uniform AA scavenger dispersions, at the lowest concentrations; however, it also increased the thermal histories of these PET blends. As an example, portions of the 500 ppm alpha- cyclodextrin/PET blend sample have been processed up to three times:

• Extruded once: the virgin PET resin added to dilute the 1 weight % blend to the

desired 500 ppm concentration

• Extruded twice: the virgin resin melt-blended with the 5 weight % master-batch

sample to yield an overall alpha-cyclodextrin concentration of 1 weight %

• Extruded three times: the 5 weight % master-batch sample that was extruded with

virgin resin to dilute the blend’s alpha-cyclodextrin concentration to 1 weight %

blend

Table 3.1 breaks down the thermal histories of the various alpha-cyclodextrin/PET blend samples. This table shows how much of each sample was extruded once, twice, and even three times to achieve the final, desired AA scavenger concentration.

Table 3.1: The thermal history for each alpha-cyclodextrin/PET blend Scavenger Concentration Percentage of Processed Material weight % ppm 1x 2x 3x 5.0 50,000 100.00 - - 2.5 25,000 50.00 50.00 - 1.0 10,000 80.00 20.00 - 0.5 5000 90.00 10.00 - 0.12 1200 88.00 9.60 2.40 0.05 500 95.00 4.00 1.00

3.3.2 Preparation of Anthranilamide/PET Blend Samples

The preparation of the anthranilamide/PET blend samples was very similar to the procedure used to prepare the alpha-cyclodextrin/PET blend samples. The anthranilamide/PET master-batch sample was prepared by weighing determined amounts of the AA scavenger and pure resin. Each sample was place in its own vacuum oven and dried overnight; virgin PET at 150oC and anthranilamide at 80oC. The next morning, these two components were dry-mixed in a small metal bucket and then extruded and

pelletized. The anthranilamide concentration of this master-batch sample was 1 weight %.

Similar to the process described in Section 3.3.1, the anthranilamide/PET master-batch sample was further diluted to make additional blend samples. The 1 weight % (or 10,000 ppm) anthranilamide/PET master-batch blend was reprocessed and diluted to produce a

1200 ppm blend and a 500 ppm blend. The lowest anthranilamide/PET blend samples, the 200 ppm and the 100 ppm blends, were prepared from the 1200 ppm concentrate. As discussed in Section 3.3.1, a more diluted blend was used with the hope that it would produce less variability and more homogeneity among these samples. Table 3.2 has been prepared to show the amount of each anthranilamide/PET blend that has been extruded once, twice, and even three times.

Table 3.2: The thermal history for each anthranilamide/PET blend Scavenger Concentration Percentage of Processed Material weight % ppm 1x 2x 3x 1.0 10,000 100.00 - - 0.12 1200 88.00 12.00 - 0.05 500 95.00 5.00 - 0.02 200 83.33 14.67 2.00 0.01 100 91.67 7.33 1.00

3.3.3 Preparation of MXDA/PET Blend Samples

The process to produce the MXDA/PET blends varied slightly in comparison to the methods discussed in Sections 3.3.1 (alpha-cyclodextrin/PET blends) and 3.3.2

(anthranilamide/PET blends). The change in preparation methods was because at room

temperature MXDA is a liquid84; while alpha-cyclodextrin85 and anthranilamide86 are crystalline materials. To blend this scavenger with PET, MXDA was pumped into the throat of the extruder at a desired rate; through the use of a pump. Prior to extrusion, the virgin PET resin was dried overnight at 150oC in a Conair Franklin hopper/dryer to eliminate moisture.

The next morning, the process began by extruding PET resin until a steady-state was established. Once steady-state was achieved, the pump was turned on, blending MXDA with the molten polymer. Collection of the pelletized MXDA/PET blend sample began three minutes after the pump was initially turned on. As with the previous processes, a vacuum pump and nitrogen gas were used to limit the degradation of the material during extrusion. Since MXDA is known to be very reactive, a plastic tarp was used to create a tent that surrounded the twin-screw extruder. This precaution was used to direct any volatiles toward the lab hood that was located above the extruder. The resulting

MXDA/PET master-batch sample had an AA scavenger concentration of 1 weight %. It should be noted that this sample had a slight greenish tint to its appearance.

The various MXDA/PET blend samples were produced by the same manner as described

(Section 3.3.2) for the range of anthranilamide/PET blends. The 1 weight % (or 10,000 ppm) master-batch blend was used to make a 1200 ppm blend and a 500 ppm blend. The

200 ppm and 100 ppm blend were created from the 1200 ppm MXDA/PET concentrate.

Table 3.3 shows the amount of each MXDA/PET blend that has been extruded once, twice, and even three times. Since the MXDA/PET blend samples were prepared in the

same manner as the anthranilamide/PET blend samples, the thermal histories shown in

Tables 3.2 and 3.3 are identical to one another.

Table 3.3: The thermal history for each MXDA/PET blend Scavenger Concentration Percentage of Processed Material weight % ppm 1x 2x 3x 1.0 10,000 100.00 - - 0.12 1200 88.00 12.00 - 0.05 500 95.00 5.00 - 0.02 200 83.33 14.67 2.00 0.01 100 91.67 7.33 1.00

3.3.4 Preparation of Control PET Samples

A pure PET resin, without the addition of any AA scavenging agent, was extruded under conditions equivalent to those experienced during the melt-blending process of the various AA scavenger/PET blends. The resin used to establish this control sample was the same CSD PET resin which was melt-blended to make each of the various AA scavenger/PET blend samples. To simulate the sample blending methods discussed in

Sections 3.3.1, 3.3.2, and 3.3.3, a total of three control samples were prepared. Prior to processing, all of these samples were dried overnight at 150oC in a Conair Franklin hopper/dryer to eliminate moisture.

The process to prepare these three control samples began by initially extruding, pelletizing, and collecting a large amount of the CSD PET resin. About one-half of this

“one-time processed” control sample was set aside for analysis. The other half of this sample was re-extruded to establish the “two-times processed” control sample. Again,

about one-half of this “two-times processed” sample was set aside for analysis and the other half was re-extruded to create the “three-times processed” control sample.

While this approach does not replicate the exact blend ratios established in Sections 3.3.1,

3.3.2, and 3.3.3; it does allow for comparisons to be made. For any analytical method, the evaluation of each control sample produces data which can be proportioned with the data from the other two control samples. This method creates theoretical values that can be compared to any AA scavenger/PET blend sample as long as the thermal histories of these two samples match. Matching the thermal histories is achieved by proportioning the correct amount of the “one-time processed”, “two-times processed”, and “three-times processed” control samples to equal the ratio for any particular AA scavenger/PET blend sample shown in Tables 3.1, 3.2, and 3.3.

3.4 Manufacturing PET Containers

Producing PET containers is generally performed by means of a two stage process. The first step is known as injection molding, a process that melts PET resin and pushes the viscous polymer to fill a mold. For container manufacturing, the article that is produced is known as a preform. Transforming these performs into containers is achieved by a separate process known as stretch-blow-molding.

This second and final step of the container manufacturing progression heats the PET preforms to temperatures that are slightly above the glass transition temperature (Tg) of

the material, but below its crystallization temperature. When a polymer is heated above its Tg, it transitions from a glassy state into a rubbery state. These rubbery preforms are then mechanically stretched, by a rod, and blown to fill another mold. The result of the stretch-blow-molding process is a final PET container of a desired volume and design.

3.4.1 Injection Molding

The Arburg 320S machine was used to injection mold preforms. This is a single cavity injection molder that has a 55-ton capacity and a reciprocating screw. The processing temperature was controlled to be 280oC, with a nozzle temperature of 290oC. The injection pressure was set at 1500 bar and the cooling time was 10 seconds. The mold used this work produces preforms specifically designed for 2-liter bottle manufacturing.

In preparation for injection molding, two concentrations of each AA scavenger were prepared. A determined amount of an already extruded and melt-blended AA scavenger/PET blend was dry-blended with a determined amount of virgin PET. Each dry-blended sample was then dried overnight, at 150oC, in a Conair Franklin hopper/drier.

After drying, a transfer pipe from the Conair Franklin drier was attached to the throat of the injection molder for automated resin loading. A pure PET resin sample was also prepared by a similar process. This control sample provided a benchmark, to be compared against, for analytical testing.

For each sample set, the first ten preforms were discarded; allowing the machine to reach a steady-state until sample collecting began. Once the machine reached steady-state, twenty samples were collected and immediately placed in a freezer. This was to prevent any acetaldehyde from diffusing from the PET preforms and into the atmosphere; a vital step toward assuring accurate results when analyzing the residual AA content of those samples. The remaining preforms were set aside to be stretch-blow-molded into 2-liter bottles.

3.4.2 Stretch-Blow-Molding

Transforming the prepared preforms into 2-liter bottles was achieved through a stretch- blow-molding process. This process begins by heating a PET preform above its Tg, as a result of exposure to infrared radiation (IR). The IR heating system consists of twelve quartz lamps, each rated at 1600 watts; with peak filament temperatures of 2200K at 240 volts. These twelve zones can be adjusted to alter the temperature profile. Figure 3-1 shows a drawing of a preform and approximately where the twelve zones are located.

1 2 3 4 5 6 7 8 9 10 11 12 Figure 3-1: Relative location of IR heating zones with respect to a perform

After loading a preform onto a rotating mandrel, the IR heater box passes the PET preform twice; returning to its starting position. The speed at which the heater box passes the preform and the voltages of the twelve IR heaters were optimized yield the best possible bottles. Additionally, the stretch rod pressure and the blow pressures were optimized to yield the best bottle appearance. Table 3.4 shows the optimized stretch- blow-molding conditions for Voridian CB12 PET based samples.

Table 3.4: Optimized stretch-blow-molding parameters Stretch Rod Pressure 100 psi Low Blow Pressure 60 psi High Blow Pressure 200 psi Heater Box Speed 230 Heating Settings -- Zone 1 235 volts Zone 2 235 volts Zone 3 235 volts Zone 4 200 volts Zone 5 125 volts Zone 6 100 volts Zone 7 135 volts Zone 8 135 volts Zone 9 165 volts Zone 10 165 volts Zone 11 180 volts Zone 12 180 volts

3.5 Gas Chromatography (GC)

Chromatography is a common family of analytical techniques used as separation methods and used to quantify analytes. Chromatography methods are defined by their respective stationary and mobile phases.82 Gas chromatography (GC), for example, uses a carrier gas as the mobile phase and a solid column as the stationary phase.

During GC, an inert carrier gas moves the gaseous sample through a long column, which is stored within an oven, and eventually to the instrument’s detector where it will be quantified. As the sample moves through the column, it separates into its individual chemical species based on the affinity that each individual chemical species has toward interacting with the column (stationary phase).82 If the strength of non-covalent interactions is strong between the chemical compound and the stationary phase, then it will have a longer retention time within the column than a chemical that does not interact as well with the material composing the stationary phase.82

Gas chromatography is commonly used to analyze the presence of acetaldehyde in PET by two different techniques. The first method determines the rate at which AA is generated during the processing of PET resin. The second technique, known as headspace analysis, is used to determine the amount of AA that remains trapped within

PET following processing.

3.5.1 Acetaldehyde Generation Analysis

To quantify how much acetaldehyde is generated during processing, the AA generation rates were studied by a method described by Kim and Jabarin23. This technique simulates the heating conditions that are needed to process PET. Measurements are made through the use of a Perkin-Elmer Automatic Thermal Desorption System (ATD 400) coupled to a Perkin-Elmer AutoSystem XL Gas Chromatograph. The gas chromatograph contains a

Stabilwax®-DA column, measuring 30 meters in length and has an internal diameter of

0.32 mm. The column temperature was 60oC and used a helium gas purge. The GC uses a flame ionization detector (FID) to sense the amount of analtyes that have passed through the system. Perkin-Elmer’s TurboChrom software then interprets the FID’s response into the quantified peak areas of each analyte.

Samples for this measurement are PET pellets, virgin or with AA scavengers, which have been dried and crystallized overnight, at 120oC, in a vacuum oven. For each sample, two

PET pellets are weighed using a five decimal place analytical balance; the combined target weight for these pellets is 0.03 ± 0.01 grams. These two pellets are placed in a cylindrical Teflon sample chamber, separated by quartz wool. This packed sample chamber is placed into a metal sample tube. Each sample tube that is placed in the ATD

400 is melted for a specified residence time and at a specified temperature; establishing only one data point along the AA generation curve. Multiple samples, measured at varying residence times, are therefore required to create an AA generation curve. The slope of this curve is the AA generation rate for the studied PET sample, at the evaluated temperature.

The instrument is calibrated by injecting ten microliters (µL) of a 1.002 µg/µL concentrated standard into a metal sample tube, containing a Teflon sample chamber packed with Tenax®; an analyte absorbing agent. This standard is purged into this sample tube for five minutes at a pressure of 0.5 pounds per square inch (psi). This process was repeated at least three times in order to obtain an averaged peak area. These prepared standards are placed in the ATD 400 and heated for 10 minutes at 250oC. The

instrument’s response is determined by dividing the averaged peak area, for the AA standards, by the amount of sample injected. A typical value for the GC’s instrument

µVolts × seconds response for this set-up is in the vicinity of 50,000 ; µ means “micro”. µgramsAA

Equation 1 shows the equation used to calculate the amount of AA for each sample.

Table 3.5 provides an explanation of the variables in Equation 1.

E 1 AA = PA × (Equation 1) ESW SIR

Table 3.5: Explanation of the variables from Equation 1 Variable Meaning Units

AA Amount of acetaldehyde µgrams AA or ppmAA gramsPET

EPA Peak area of the experimental sample µVolts × seconds

ESW Weight of the experimental PET sample gramsPET SIR Instrument response obtained from the µVolts × seconds standardized sample µgramsAA

AA generation rates were established at three different temperatures (280, 290, and

o 300 C) so that the activation energy (EA) could be determined by means of the Arrhenius equation. Equation 1 shows the Arrhenius equation in its most common form. Table 3.6 provides an explanation of the variables displayed in Equations 2 and 3. Equation 3 shows the Arrhenius equation in a derived form. Using this equation, the activation energy can be determined graphically by plotting the natural log of the rate versus the inverse of the temperature, in degrees Kelvin. The slope of this plot is the activation

energy divided by the gas constant, and the y-intercept is the natural log of the pre- exponential factor.

 E   A   RT  k = A× e (Equation 2)

Table 3.6: Explanation of the variables in Equations 2 and 3 Variable Meaning Units k Rate ppm

min. A Pre-exponential factor ppm

min. E Activation energy Joules A mole R Gas constant Joules

mole× Kelvin T Temperature Kelvin

 E A 1  ln()k = ln (A )−  ×  (Equation 3)  R T 

3.5.2 Residual Acetaldehyde Analysis

As AA is generated through the degradation of PET it has the ability to diffuse out of the polymer’s matrix; due to AA’s low boiling point of 21oC.19 A portion of the generated

AA, however, usually remains trapped within the manufactured PET article as residual

AA. Quantifying the concentration of residual AA in PET packages is a critical aspect for many packaging applications because many foods and beverages have a limited threshold for AA.2, 20

For this work, the quantification of residual AA was performed by a headspace analysis technique. This technique coupled a Perkin-Elmer TurboMatrix 40 (TM 40) Headspace

Sampler with a Perkin-Elmer AutoSystem XL Gas Chromatograph. The GC’s settings for headspace analysis are identical to the parameters that were previously mentioned in

Section 3.5.1.

For this measurement, processed samples (either melt-blended by twin screw extrusion or injection molded into preforms) were immediately collected and placed in a freezer to prevent the AA from volatilizing out of the PET matrix. Each sample, in its amorphous state, was ground by means of a small grinder made by the Tekmar Company. The amorphous PET sample was placed in the grinder and then saturated with liquid nitrogen to keep the polymer cold. The resulting ground powder was then separated using a sieve combination of 20 mesh, 40 mesh, and solid bottom. Only the 20 mesh sample was collected and used for this analysis. This is done in order to maximize surface area and increase amount of AA that can diffuse from PET. The ground PET samples were, again, immediately placed back into the freezer to prevent the diffusion of AA.

Prior to analysis, the PET powder samples were removed from the freezer and then weighed on a five decimal place analytical balance. The samples are weighed within a glass sample vial which is capped and sealed immediately after weighing. During analysis, each sample is heated for 60 minutes at 150oC and at 18 psi. This temperature does not melt the PET, it simply volatizes the residual amount of AA that is trapped within the PET into the headspace of the glass vial. Once the sixty minutes has lapsed,

the TM 40 injects a needle and extracts a sample of the gaseous headspace. This sample is sent through the GC column to be quantified.

Calibration for this technique is similar to the procedure that is followed to perform experimental measurements. 2, 4, 6, 8, and 10 µL aliquots of standardized AA solution are respectively injected, by means of a syringe, into a glass sample vial. The standardized sample has a typical concentration of around 1.002 µg/µL. Each of these five samples is heated for 15 minutes at 150oC and at 18 psi. The instrument’s response factor is determined by plotting the resulting peak areas against the respective AA concentrations; the origin is also used as a data point. The instrument’s response factor is the slope of this linear line. Typical instrument response values for headspace

µVolts × seconds determination vary between 600 and 900 . µgramsAA

Determining AA content within PET samples is calculated by means of Equation 1; previously shown in Section 3.5.1. Equation 1 uses the experimental sample’s weight

(ESW), the experimental sample’s resulting peak area (EPA), and the instrument’s response factor (SIR) from the calibration standards to tabulate the AA content within the PET sample. Table 3.5, also located within Section 3.5.1, provides further explanation of the variables in Equation 1.

3.6 Rheological Methods

3.6.1 Plate and Plate Rheometer

Determination of a PET sample’s intrinsic viscosity (I.V.) was performed by measuring its melt viscosity. Conversion from melt viscosity to intrinsic viscosity was made by evaluating the melt viscosities of standardized samples that possess precisely known I.V.s.

The solvent used to measure the intrinsic viscosities of standard PET samples was composed of 60% phenol and 40% tetrachloroethane.

Melt viscosity measurements were made by a RDA III viscoelastic tester from

Rheometric Scientific. Measurements were made at using parallel plate and plate geometry. Before making any measurements, the PET samples were crystallized and dried, at 140oC, overnight. Table 3.7 summarizes the test conditions for the melt viscosity measurements.

Table 3.7: Melt viscosity testing conditions Temperature 280oC Motor Dynamic Test Delay 120 seconds Gap 1 mm Strain 15% Environment Nitrogen Shear Rate 10 radians/second

3.6.2 Capillary Rheometer

An Instron Capillary Rheometer was used to study the viscosity of PET resin as a function of varying shear rates. Pelletized, virgin PET resin was dried overnight, at

140oC, prior to analysis. Measurements were made at three different temperatures (260,

270, and 280oC) so that the data could be extrapolated to predict the material behavior for an even wider range of temperatures. The resin samples were melted under an inert, nitrogen environment. Once the samples were completely melted within the capillary column, a desired crosshead speed was set and the force was measured; this was varied over a range of crosshead speeds. Using this data, along with instrumentation parameters, the apparent shear rates, shear stresses, and apparent viscosities were tabulated. Plots of viscosity versus shear rate yielded the desired rheology curves. Mathematical constants were determined from these plots and were used to predict the resin’s rheology behavior within the multi-cavity injection molding modeling program. Table 3.8 lists the instrumentation parameters for this analysis.

Table 3.8: Instrument parameters for the capillary rheometry analysis Capillary Diameter 0.0304 inches Capillary Length 1.0080 inches Barrel Diameter 0.3750 inches

3.7 Color Analysis

Changes in color, due to the processing and/or addition of additives melt-blended into

PET, were analyzed by means of a Hunter Lab Color/Difference Meter D25-2. The

instrument was initially calibrated using the standard colored plates supplied the company. Experimental measurements were made on crystallized PET pellets, at room temperature. Response from the instrument yields L, a, and b values for each sample.

These values were then converted to Y, X%, and Z% values; according to Equations 5, 6, and 7, respectively. Ultimately, the Y, X%, and Z% values were used to calculate a yellowness index (YI). The yellowness index was calculated according to ASTM D

192587, which is described by Equation 4.

(125× (X % − Z%)) YI = (Equation 4) Y

Where: Y = 0.01× L2 (Equation 5)

 a × L  X % = ()0.01× L2 +   (Equation 6)  175 

 b × L  Z% = ()0.01× L2 −   (Equation 7)  70 

Table 3.9 gives an explanation of the L, a, b, and YI values that are obtained from the instrument or calculated using above equations.

Table 3.9: Explanation of L, a, b, and YI values Variable Meaning L Measures lightness and varies from 100, for perfect white, to 0, for black; approximately as the eye would evaluate it a Measures redness when plus, gray when zero, and greenness when minus b Measures yellowness when plus, gray when zero, and blueness when minus YI Yellowness index of nearly white plastic samples; positive values indicate increase yellowness, while negative values indicate decreased yellowness or increased blueness

3.8 Differential Scanning Calorimetry (DSC) Analysis

Differential scanning calorimetry (DSC) is a common technique used to analyze the thermal properties of polymeric materials. DSC analysis was performed to see if any changes in the thermal properties of the PET resin resulted from the melt blending of AA scavenging agents into PET. Measurements were made by means of a Perkin-Elmer DSC

7; with an attached nitrogen purge to prevent oxidative degradation from occurring during melting of the samples. Prior to analysis, samples were vacuum dried overnight at

120oC. Each sample was heated to 300oC, held for 5 minutes to remove all of its inherent crystallinity, and then rapidly quenched to 40oC creating a completely amorphous sample.

The sample was then reheated at 10oC per minute to give the melting behavior, the crystallization behavior, as well as a value for the glass transition temperature (Tg).

Cooling the sample at a rate of 10oC per minute also indicated its crystallization behavior when cooled from the melt.

3.9 Oxygen Film Permeation

Understanding the oxygen permeation rate for a given polymeric material is vital information when packaging oxygen-sensitive food or beverages. Oxygen from the atmosphere permeates through the plastic wall of the package, interacting with the packaged contents. Depending on the sensitivity of the contents, this interaction can alter the shelf-life of the packaged material. Studying the oxygen film permeation was

conducted to determine if the addition of AA scavengers would have any affect upon the oxygen barrier properties of PET.

In order to understand the oxygen permeation rate a coulometric method was used, similar to the one described in ASTM Procedure D 3985.88 This procedure involves the use of the MoCon OxTran 1050 Oxygen Permeation Analyzer. This instrument uses a single coulometric detector that is switched by a valve to any of the 10 samples cells contained on the apparatus.

Samples from sidewalls of stretch blow molded bottles were cut into four inch squares; two samples of each material were analyzed. For each sample, the average thickness was calculated after measuring the sample’s thickness at nine evenly distributed points using a

Magna-Mike® 8500. With the sensor turned off, the samples were placed in the respective chambers of the analyzer and sealed in place. Initially, baseline measurements were made by purging the instrument of any oxygen with nitrogen gas. Once a baseline for each chamber was established, the purging gas was changed to oxygen. Over time the oxygen permeates from one side of the chamber, through the polymeric samples, and to the other side of the chamber where the oxygen purge is detected. The instrument output is an electrical current that corresponds to the amount of oxygen present. This current increases with time, until steady state is reached.

Equation 8 shows the calculation used to determine the oxygen gas transmission rate

(GTR). Table 3.10 gives an explanation of the variables used in Equation 8.

GTR = (AV − BV )× IF (Equation 8)

Table 3.10: Explanation of the variables in Equation 8 Variable Meaning Units GTR gas transmission rate cc × STP

100in 2 × day AV voltage for oxygen permeation mV BV base voltage mV IF instrument factor, which accounts for the cell area cc × STP

and the conversion factor for the detector mV ×100in 2 × day

It should be noted that the baseline voltage is an averaged value, based upon two measurements. The voltage for oxygen permeation is the actual daily value; therefore, the oxygen gas transmission rate is calculated each day and not an averaged value. This value is not averaged is due daily fluctuations in barometric pressure. As shown in

Equation 9, the oxygen permeance is inversely proportional to the barometric pressure.

Table 3.11 gives an explanation of the variables used in Equation 9.

GTR OP = (Equation 9) ∆P

Table 3.11: Explanation of the variables in Equation 9 Variable Meaning Units OP oxygen permeance cc × STP

100in 2 × day × atmospheres GTR gas transmission rate cc × STP

100in 2 × day ∆P change in pressure atmospheres

Since the fluctuations in barometric pressure are accounted for, the oxygen permeance may be averaged over the time of analysis. This averaged value for oxygen permeance can now be used, along with the average thickness of the sample, to determine the oxygen permeability, shown below in Equation 10. Table 3.12 gives an explanation of the variables used in Equation 10.

P = OP × AT (Equation 10)

Table 3.12: Explanation of the variables in Equation 10 Variable Meaning Units P oxygen permeability cc × STP × mil

100in 2 × day × atmospheres OP oxygen permeance cc × STP

100in 2 × day × atmospheres AT average sample thickness mils

Chapter 4

Results and Discussion

4.1 Chemical Mechanisms of AA and AA Scavenger Interactions

The acetaldehyde (AA) scavenging interactions of anthranilamide, meta-xylenediamine

(MXDA), and alpha-cyclodextrin were studied by various nuclear magnetic resonance

(NMR) and mass spectrometry experiments. Proper determination of these chemical interactions/reactions involved the identification by which these three scavenging agents sequester AA and the exact stoichiometry for each of these mechanisms. For each experiment, the AA scavenging mechanism was studied under the most ideal circumstances. Most notably, the presence of PET was omitted from each system; eliminating diffusion as a controlling step. Each reactant was dissolved in an appropriate solvent and mixed only with the other, respective reactant and solvent solution.

4.1.1 AA and Anthranilamide

The investigation into the AA scavenging reaction between anthranilamide and acetaldehyde began by obtaining the individual 1H NMR spectra of these two components. Figure 4-1 shows the 1H NMR spectrum of AA, dissolved in deuterated

chloroform (CDCl3). Under each prominent signal there is an integration factor, indicating the number of protons which are represented by that particular peak. From left to right, the 1 to 3 ratio shown in Figure 4-1 correlates the one hydrogen atom in AA’s aldehyde (O=CH) group to the three hydrogen atoms located in its methyl (CH3) group.

The scale on the x-axis is in ppm or δ units. Table 4.1 provides a list of the assigned peaks for Figure 4-1.

2.07 9.66 9.65 9.65

0.93 3.00

10 9 8 7 6 5 4 3 2 1 1 Figure 4-1: H NMR spectrum of AA in CDCl3

1 Table 4.1: Peak assignment for the H NMR spectrum of AA in CDCl3 Peak Location Peak Integration Peak Assignment (ppm) Type Factor Chemical Functional Compound Group 2.07 Doublet 3 AA CH3 7.24 Singlet - Chloroform-d 9.65 Quartet 1 AA O=CH

The 1H NMR spectrum of anthranilamide, also dissolved in deuterated chloroform, is shown in Figure 4-2. A prominent feature of this spectrum is the broad singlet that appears at 5.67 ppm. The peak’s broadness is due to the fact that it represents exchangeable protons, which are rapidly traded when in an appropriate solution.83

Generally, these protons are present on heteroatoms such as oxygen (O), sulfur (S), and nitrogen (N). The peak at 5.67 ppm represents the four protons which comprise both the amide (O=CNH2) and the amine (NH2) groups of anthranilamide. Also shown in Figure

4-2, a number is adjacent to each of anthranilamide’s four methine (CH) groups; all are located within the ring formation. This numbering system corresponds to Table 4.2 and is used to ease the identification of these CH groups which are represented by the peaks centered at 6.63, 6.665, 7.20, and 7.34 ppm.

1 Figure 4-2: H NMR spectrum of anthranilamide in CDCl3

1 Table 4.2: Peak assignment for the H NMR spectrum of anthranilamide in CDCl3 Peak Peak Integration Peak Assignment Location Type Factor Chemical Functional Ring (ppm) Compound Group Position 5.67 Singlet 4.5 Anthranilamide NH2 and - O=CNH2 6.63 Triplet 1 Anthranilamide CH 3 6.665 Doublet 1 Anthranilamide CH 4 7.20 Triplet 1 Anthranilamide CH 2 7.24 Singlet - Chloroform-d - 7.34 Doublet 1 Anthranilamide CH 1

Anthranilamide and AA were each dissolved in deuterated chloroform, in separate NMR tubes. These two solutions were combined into one tube, at room temperature, and then analyzed by 1H NMR. The resulting spectrum, however, showed no evidence that a reaction between anthranilamide and AA took place. This spectrum was simply a combination of the 1H NMR spectra of AA (Figure 4-1) and anthranilamide (Figure 4-2).

The conclusion from this experiment was that energy must be added to this system to initiate a reaction.

A second attempt to analyze the AA scavenging reaction between anthranilamide and acetaldehyde was made by again dissolving each component in deuterated chloroform.

These two solutions, in separate NMR tubes, were then combined into one NMR tube that was subsequently was sealed. To seal this glass NMR tube, the contents (AA, anthranilamide, and deuterated chloroform solution) and the bottom of the tube were frozen in liquid nitrogen. Throughout this process, vacuum pressure was applied to continually remove air from the system. The neck of the NMR tube was then heated with a gas flame to melt the glass. When the glass reached a sufficient temperature, the tube

was twisted and a seal was created. The creation of this seal ensured that this solution could be heated without the risk of volatilizing and losing any of the components.

Figure 4-3 shows the 1H NMR spectrum that resulted after heating the anthranilamide and AA solution, in the sealed tube, for two days at 60oC. This sample was kept just under the boiling point deuterated chloroform, which is 62oC. Interpretation of Figure

4-3 indicates that two compounds still remain in this solution, and one of these components is still acetaldehyde. The presence of AA is confirmed by the doublet at

2.21 ppm and the quartet at 9.80 ppm. These peaks closely mirror the location of AA’s peaks in Figure 4-1; the integration factors are close matches as well. There are two pieces of evidence within Figure 4-3 that indication the second component is a product generated by a reaction between anthranilamide and AA.

The first bit of evidence is the appearance of a new peak (doublet) at 1.50 ppm, and has an integration factor of three. The location and integration factor indicate this peak represents a methyl group (CH3). The fact that this peak is a doublet means that it is coupled to another functional group comprised of only one proton.

The second piece of evidence confirming a reaction between anthranilamide and AA is the formation of a quartet peak at 5.06 ppm. This peak’s integration factor equals one, indicating it corresponds to one proton. Additionally, the peak is a quartet; meaning it is coupled to three other protons. Previously stated, the methyl group at 1.50 ppm was

coupled to an unknown peak representing one proton. The combination of these two pieces of information points toward the formation of a HC-CH3 linkage.

Figure 4-3: 1H NMR spectrum of the reaction between anthranilamide and AA, in o CDCl3, after heating for 2 days at 60 C

A peak assignment list for the 1H NMR spectrum shown in Figure 4-3 is provided in

Table 4.3. Within this list, there is a broad singlet at 6.24 ppm that has yet to be determined. The broadness of the peak and integration factor indicates that it correlates to an exchangeable proton bonded to a heteroatom (oxygen, nitrogen, sulfur, etc.). While in solution, however, these protons exchange so rapidly that it is not always possible to correlate the integration factor with the number of protons for which peak truly represents.

Table 4.3: Peak assignment for the 1H NMR spectrum of the reaction between o anthranilamide and AA, in CDCl3, after heating for 2 days at 60 C Peak Peak Integration Peak Assignment Location Type Factor Chemical Functional (ppm) Compound Group 1.50 Doublet 3 Reaction Product CH3 2.21 Doublet 15.5 AA CH3 5.06 Quartet 1 Reaction Product Undetermined 6.24 Singlet 1 Reaction Product Undetermined 6.68 Doublet 1 Reaction Product CH 6.87 Triplet 1 Reaction Product CH 7.27 Singlet - Chloroform-d 7.31 Triplet 1 Reaction Product CH 7.89 Doublet 1 Reaction Product CH 9.80 Quartet 4 AA O=CH

To provide further clarity and proof of a reaction between anthranilamide and AA, a 1H-

1H COSY (COrrelation SpectroscopY) NMR experiment was conducted. The scale of the peak intensities, as seen in Figure 4-4, is very difficult to read because of the dominant size of AA’s methyl group peak at 2.21 ppm. Figure 4-4 validates the earlier hypothesis that the quartet peak at 5.06 ppm is coupled with the doublet at 1.50 ppm.

This spectrum, however, shows no other evidence of this peak at 5.06 ppm being coupled to any other protons. This means one of two things: either this group is not bonded to any other group, besides the methyl group, or it is bonded to a heteroatom(s).

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

8.0

8.5

9.0

9.5

9 8 7 6 5 4 3 2 Figure 4-4: 1H –1H COSY NMR spectrum of the reaction between anthranilamide and o AA, in CDCl3, after heating for 2 days at 60 C

Analysis of the previous two NMR spectra can lead to the prediction of at least two different mechanisms which are able to describe the reaction that occurs between anthranilamide and AA. The first proposed reaction mechanism is shown in Figure 4-5.

In this reaction, anthranilamide and AA react to produce a larger organic compound.

This compound has a terminal methyl group (CH3) and a CH group which is attached to two heteroatoms; oxygen and nitrogen.

O O O OH

+ NH2 NH2

H3C H NH CH3 NH2

(Anthranilamide) (AA)

Figure 4-5: Proposed reaction mechanism #1 for anthranilamide and AA

As a way to help confirm reaction mechanism #1, shown in Figure 4-5, the ChemSketch software package was used to predict the 1H NMR spectrum for the reaction product.

The predicted 1H NMR spectrum, shown in Figure 4-6, shows a similar pattern to the actual spectrum, shown in Figure 4-3; with one exception. In Figure 4-3 the quartet appears at 5.06 ppm, while in Figure 4-6 that peaks appears at about 5.53 ppm. Further analysis was needed to prove if this 0.5 ppm difference is significant or if this value lies within the error of the predictive software program.

14 7.32[13] 13

12 O 11 11 CH3 3 10 12 4 2 1.30[12] 10 NH2 1.32[12] 13 5 1 8 9 6 NH OH 7 9 8

7 Molecular Weight = 180.2 g/mol 6

5.51[8] 3 5.53[8] 7.24[5] 2 8.05[3] 7.22[5] 5.54[8] 8.07[3] 7.12[4] 5.49[8] 0.20[7,9] 1

0 10 9 8 7 6 5 4 3 2 1 0 Figure 4-6: Predicted 1H NMR spectrum for the product formed from the proposed reaction mechanism #1 (Figure 4.5)

A second proposed reaction mechanism for anthranilamide and AA is shown in Figure

4-7. This reaction results in the formation of water and a two ring structured, organic compound. Similar to the reaction product formed in proposed reaction mechanism #1, shown in Figure 4-5, this compound has a terminal methyl group (CH3) and a CH group which is attached to two heteroatoms. This time, however, the CH group bonds to two nitrogen atoms instead; of a nitrogen atom and an oxygen atom.

O O O + NH NH 2 H H2O + H C H N 3 H CH3 NH2

(Anthranilamide) (AA) Figure 4-7: Proposed reaction mechanism #2 for anthranilamide and AA

ChemSketch was again used to help evaluate the reaction products from another proposed reaction mechanism between anthranilamide and AA. Previously it was shown that the only significant difference between the predicted 1H NMR spectrum from the first proposed reaction mechanism (Figure 4-6) and the actual spectrum (Figure 4-3) was the location of the CH quartet peak; 5.53 and 5.06 ppm respectively. This time, however, the predicted 1H NMR spectrum, shown in Figure 4-8, of the reaction product shown in

Figure 4-7 closely matches the actual spectrum shown in Figure 4-3. The location of the predicted quartet peak, representing the CH group, is 4.98 ppm; very close to the actual location of 5.06 ppm.

10 1.49[11] 1.47[11] 9

8 O 12 10 5 9 6 7 NH 1

8 4 2 N CH3 6 7 H 11 3

5 Molecular Weight = 162.2 g/mol

4 4.08[1]

7.16[8] 2 7.14[8] 7.64[10] 7.06[9] 4.98[2] 7.66[10] 6.40[7] 1 4.96[2]

0 10 9 8 7 6 5 4 3 2 1 0 Figure 4-8: Predicted 1H NMR spectrum for the product formed from the proposed reaction mechanism #2 (Figure 4-7)

Based on the data presented up to this point, it is difficult to truly distinguish between these two proposed reaction mechanisms. Even though the 1H NMR spectra for proposed reaction mechanism #1 (Figure 4-6) and proposed reaction mechanism #2 (Figure 4-8) are very similar, there is one drastic difference. The reaction product for proposed reaction mechanism #1 has a molecular weight of 180.2 grams/mol; while, the reaction product for proposed reaction mechanism #2 has a molecular weight of 162.2 grams/mol.

One way to distinguish between the two mechanisms would be through the use of mass spectrometry.

Electrospray ionization (ESI) mass spectrometry was used to analyze both a sample of anthranilamide dissolved in deuterated chloroform and a solution of anthranilamide and

AA dissolved in deuterated chloroform. A small aliquot of methanol was added to both samples to aid with the ionization process. The mass spectrum of the anthranilamide solution is shown in Figure 4-9. Within this figure, it can be seen that the molecular ion peak is located at 159.1 m/z (mass to charge ratio). The molecular weight of anthranilamide, however, is known to be 136.1 grams/mol. The difference between these two masses is 23 grams/mol, the molecular weight of sodium (Na). Sodium is a major component of glass containers, which happens to be what the sample was stored in prior to analysis.

Intens. All, 0.0-0.1min (#1-#10) x104

159.1 (molecular ion peak) – 136.1 (molecular weight of

5 anthranilamide) = 23 (molecular weight of sodium ion)

159.1

120.1

213.1

0 100 120 140 160 180 200 220 m/z Figure 4-9: ESI mass spectrum of anthranilamide in CDCl3 and methanol

Figure 4-10 shows the mass spectrum of the product from the reaction between anthranilamide and AA. The molecular ion peak for this product is 185.5 m/z. As previously mentioned, this sample was also stored in glass. Therefore, it is assumed that

Na is the ion which is attached to the product. Subtracting the mass of sodium from the molecular ion peak yields a mass of 162.5 grams/mol. This molecular weight (162.5 grams/mol) corresponds very well with the product which is formed from the proposed reaction mechanism #2 (162.2 grams/mol); shown in Figure 4-7. The molecular weight of the product formed in proposed reaction mechanism #1, shown in Figure 4-5, is 180.2 grams/mol. The use of both the mass spectrum and the 1H NMR spectral data confirm that proposed reaction mechanism #2 is the reaction scheme by which anthranilamide acts as an AA scavenger.

Further examination of the patent by Rule, et al50 also indicates that proposed mechanism

#2 is the correct reaction scheme. U.S. Patent 7,550,203 describes the interaction of anthranilamide and AA, in the presence of PET, as: “combining with polyester an organic additive compound comprising at least two hydrogen-substituted heteroatoms bonded to carbons of the organic additive compounds, the organic additive compound being reactive with acetaldehyde in the polyester to form water and a resulting organic compound comprising an unbridged 5- or 6-member ring including at least two heteroatoms”. Further on in the patent50, the authors state that “the two heteroatoms are both nitrogen”.

The mass spectrum shown in Figure 4-10 produces one more peak of interest; the peak at

161.5 m/z (mass to charge ratio). ESI mass spectrometry ionizes a compound by adding an ion such as hydrogen (H), sodium (Na), etc. While ESI does not typically knock off a hydrogen atom, if it did the peak at 161.5 m/z would correlate very well with the mass of the product from reaction mechanism #2 (162.2 grams/mol). As just mentioned, this phenomenon does not typically occur. In essence, a compound which has a molecular weight of about 160 grams/mol, ionized with a hydrogen atom, would correspond with the secondary molecular ion peak shown in Figure 4-10.

Intens. All, 0.0-0.3min (#1-#26) x104

185.5

1.0

185.5 (molecular ion peak) – 23 (molecular weight of sodium ion) = 162.5 (molecular weight of product)

0.8

161.5

0.6

0.4

213.5 0.2 146.5

199.4 168.1 205.6 149.4 219.6 120.4 0.0 100 120 140 160 180 200 220 m/z Figure 4-10: ESI mass spectrum of the product from the reaction between anthranilamide and AA in CDCl3 and methanol

Figure 4-11 shows a proposed reaction mechanism for anthranilamide and AA in which the final product possesses a molecular weight of 160.2 grams/mol. Proposed reaction mechanism #3, Figure 4-11, starts with proposed reaction mechanism #2 (Figure 4-7) but adds one more step to the reaction. In this de-saturation reaction, the organic product formed in reaction mechanism #2 gives off a di-hydrogen molecule that yields a similar two-ring structured organic compound, now with a double bond in its second ring. The formation of this final product, shown in Figure 4-11, has been observed by several authors.89-92 Abdel-Jalil, et al89 showed that this product can be produced through the addition of time, energy, and a catalyst.

O O O + NH NH 2 H H2O + H C H N 3 H CH3 NH2

(Anthranilamide) (AA) O

NH H2 + N CH3

Figure 4-11: Proposed reaction mechanism #3 for anthranilamide and AA

Beyond what was previously stated, Rule, et al50 give no further indication of the exact composition for the resulting organic compound that is formed. Further experiments were required to determine if proposed mechanism #3 (Figure 4-11) is more correct than proposed mechanism #2 (Figure 4-7). This entailed periodic 1H NMR analysis to study

the possible double bond formation in the second ring of the organic compound produced by the anthranilamide and AA reaction.

To begin this work, ChemSketch was used to predict a 1H NMR spectrum of the final reaction product that is shown in Figure 4-11. Figure 4-12 shows this predicted spectrum.

Next, the sealed NMR tube which contained the dissolved mixture of anthranilamide and

AA was heated over a four week period, at 60oC. Throughout this time, periodic 1H

NMR spectra were obtained and analyzed.

2.36[12] 20

O 11 7 2 8 3 NH 15 1 4 6 9 N CH3 10 5 12

10 Molecular Weight = 160.2 g/mol

5 7.67[8]

7.71[10] 7.64[9] 7.73[10]

0 10 9 8 7 6 5 4 3 2 1 0

Figure 4-12: Predicted 1H NMR spectrum for the product formed from the proposed reaction mechanism #3 (Figure 4-11)

While each 1H NMR spectra confirmed the final product in proposed mechanism #2

(Figure 4-7), there proved to be no tangible evidence of the double bond formation on the second ring. Never was there a spectrum that resembled the appearance of Figure 4-12.

It is important to note, however, that this experiment was conducted without the presence of a catalyst; which was used in the experiments performed by Abdel-Jalil, et al89. It is therefore possible that the formation of the final product in proposed reaction mechanism

#3 could be obtained when anthranilamide is added to PET. All PET resins contain a small amount of residual catalyst within their matrix. When anthranilamide is added, to act as an AA scavenger, this residual amount of catalyst could drive the reaction to form the final product shown in proposed reaction scheme #3 (Figure 4-11).

4.1.2 AA and MXDA

The investigation into the AA scavenging reaction between meta-xylenediamine

(MXDA) and acetaldehyde began in the same manner as previously discussed in Section

4.1.1. The initial step was to obtain the individual 1H NMR spectra for AA and MXDA.

Previously shown and discussed, Figure 4-1 shows the 1H NMR spectrum of AA dissolved in deuterated chloroform (CDCl3). Similarly, MXDA was also dissolved in

1 CDCl3 and analyzed by means of H NMR.

Figure 4-13 shows the 1H NMR spectrum for MXDA in the presence of deuterated chloroform. The peaks appearing between 7.0 and 7.3 ppm are enhanced for detailed viewing; making it easier to see the multiplicity and integration factors. It can be seen

that the ratio of integration factors, from left to right, is 1:1:2:4:4. Similar to anthranilamide, MXDA’s 1H NMR spectrum contains a very broad peak. As previously mentioned, this broad peak indicates the presence of exchangeable protons. These protons are from MXDA’s two primary amine groups (NH2) and are represented by the peak located at 1.35 ppm. Table 4.4 provides a list of the assigned peaks for Figure 4-13.

3.81

1 7.14 7.13 7.22 7.25 7.23 2 4

3 1.06 0.96 1.98

7.30 7.25 7.20 7.15 7.10 7.14 7.13 7.22 1.35 7.25

2.01 4.11 4.00

7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 1 Figure 4-13: H NMR spectrum of MXDA in CDCl3

1 Table 4.4: Peak assignment for the H NMR spectrum of MXDA in CDCl3 Peak Peak Integration Peak Assignment Location Type Factor Chemical Functional Ring (ppm) Compound Group Position 1.35 Singlet 4 MXDA Two NH2 - 3.81 Singlet 4 MXDA Two CH2 - 7.135 Doublet 2 MXDA Two CH 2 7.22 Singlet 1 MXDA CH 1 7.25 Triplet 1 MXDA CH 3

Similar to the mixing procedure for anthranilamide and AA, MXDA and AA were each dissolved in separate NMR tubes using deuterated chloroform. These two individual solutions were then combined into one tube. The mixing of the two solutions resulted in an instantaneous reaction that occurred at room temperature; forming a solid, orange product. Further dilution of the resulting product altered its color from orange to a dark yellow, slight greenish appearance.

The appearance of this sample was similar to that of the 1 weight % MXDA/PET blend sample; following twin-screw extrusion. The reason for this color formation was documented by Bandi, et al.48 Through the study of polyamide/PET blends, the authors proved that this color generation was the result of a reaction between the amine group, from MXD6, and generated AA, from PET. This reaction results in the formation of imine (N=CH) groups; which Bandi, et al48 proved to be the chromophores. MXDA is the monomer from which MXD6 is manufactured. MXDA has two primary amines, while MXD6 has only one.

A reaction scheme between MXD6 and AA has been previously presented in Figure 2-2.

Using this as a guide, a proposed reaction mechanism for MXDA and AA is shown in

Figure 4-14. Similar to the reaction for MXD6 and AA, the aldehyde group (O=CH) of

AA reacts with a primary amine group from MXDA. Since MXDA has two primary amine groups, one molecule of MXDA can react with up to two molecules of AA. The result of this reaction can therefore generate up to two imine groups and up to two molecules of water, as a byproduct.

H2N H2C O CH2 NH2

+ 2 H3C CH

(MXDA) (AA)

H3C HC N H2C CH2 N CH CH3 + 2 H2O

Figure 4-14: Proposed reaction scheme for MXDA and AA

The 1H NMR spectrum representing the reaction between MXDA and AA is shown in

Figure 4-15. This spectrum contains two pieces of evidence that verify the reaction scheme shown in Figure 4-14. First, it can be seen that the reaction forms water, evident by the singlet at 1.90 ppm. Second, a new peak is formed at 7.80 ppm. This peak is a quartet and represents the two imine (HC=N) groups that were the result of the reactions between the two amine groups from MXDA and the aldehyde group from AA. Table 4.5 provides a complete list of the identified peaks that are shown in Figure 4-15.

The 1H NMR spectrum of the reaction between MXDA and AA, Figure 4-15, also shows another interesting fact. Analysis of this spectrum shows that only two components are present: AA and the resulting product. This indicates that any MXDA that was originally present has since completely reacted. This is validated by the fact that the singlet at 1.35 ppm, present in Figure 4-13, is absent in the 1H NMR spectrum shown in

Figure 4-15. This singlet represented the two primary amines of MXDA, and the absence of this peak indicates that no MXDA is present in this system.

1 2.18 2.18

2 4 3 7.13 7.12 7.15 4.51 7.26 7.24 1.99 7.25 1.98

0.90 0.80 1.65

7.30 7.25 7.20 7.15 7.10 7.13 9.76 9.77 1.90 7.12 7.15 7.80 7.80 7.26

1.43 1.72 1.65 4.00 6.26

9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1 Figure 4-15: H NMR spectrum of the reaction between MXDA and AA in CDCl3

Table 4.5: Peak assignment for the 1H NMR spectrum of the reaction between MXDA and AA in CDCl3 Peak Peak Integration Peak Assignment Location Type Factor Chemical Functional Ring (ppm) Compound Group Position 1.90 Singlet - H20 - 1.985 Doublet 4.5 AA CH3 - 2.18 Doublet 6 Reaction Product Two CH3 - 4.51 Singlet 4 Reaction Product Two CH2 - 7.135 Doublet 2 Reaction Product Two CH 2 7.15 Singlet 1 Reaction Product CH 1 7.24 Singlet - Chloroform-d - 7.26 Triplet 1 Reaction Product CH 3 7.80 Quartet 2 Reaction Product Two HC=N - 9.765 Quartet 1.5 AA O=CH -

4.1.3 AA and Alpha-Cyclodextrin

The previously presented 1H NMR spectra for the anthranilamide and the MXDA scavenging reactions have all used deuterated chloroform (CDCl3) as the solvent.

Deuterated chloroform is a very common solvent because it has a low boiling point, is relatively inexpensive, and dissolves a plethora of organic compounds. The solubility of cyclodextrins in deuterated chloroform, however, is very low. Alpha-cyclodextrin contains hydroxyl groups that make it a polar molecule; while CDCl3 is non-polar. The general rule for solubility states that “like dissolves like”. In other words, alpha- cyclodextrin will be more soluble in a polar solvent like deuterium oxide (D2O).

Studying the AA scavenging mechanism between alpha-cyclodextrin and acetaldehyde was initially similar to the approach discussed in Sections 4.1.1 and 4.1.2. This process began by obtaining the individual 1H NMR spectra for alpha-cyclodextrin and AA.

Figure 4-16 shows the 1H NMR spectrum of alpha-cyclodextrin in deuterium oxide; also shown is alpha-cyclodextrin’s repeat unit. The peaks shown in this spectrum correspond only to the hydrogen atoms which are bonded to carbon atoms (CH groups). Not seen in this spectrum are the hydrogen atoms bonded to oxygen atoms; hydroxyl groups. This phenomenon is a result of the chosen solvent. The hydrogen atoms of the hydroxyl groups are rapidly exchanged with the deuterium atoms in deuterium oxide. This exchange causes these hydrogen atoms to be grouped into the water peak at 4.8 ppm.

4.80 4.80

O OH 2 OH H H 3 1 H H 4 O 5H H H 6 6 O OH n 5.07 3.87 5.07 3.92 4.00 3.60 3.89 3.66 3.98 3.66 3.64 3.64 4.02 3.86 3.59

1.00 1.121.431.59 1.04

5.2 5.1 5.0 4.9 4.8 4.7 4.6 4.5 4.4 4.3 4.2 4.1 4.0 3.9 3.8 3.7 3.6 3.5 1 Figure 4-16: H NMR spectrum of alpha-cyclodextrin in D2O

In order to aid in the peak assignment for Figure 4-16, a 1H-1H NMR COSY experiment was conducted. The resulting 1H-1H COSY spectrum of alpha-cyclodextrin, dissolved in deuterium oxide, is shown in Figure 4-17. Due to the dominant size of the water peak at

4.80 ppm, the scale of the peak intensities is very difficult to read. The coupling pattern of this spectrum was used to assign the various peaks to their proper CH protons, within alpha-cyclodextrin’s repeat unit. The complete peak assignment list is shown in Table

4.6.

3.5

3.6

3.7

3.8

3.9

4.0

4.1

4.2

4.3

4.4

4.5

4.6

4.7

4.8

4.9

5.0

5.0 4.5 4.0 1 1 Figure 4-17: H – H COSY NMR spectrum of alpha-cyclodextrin in D2O

1 Table 4.6: Peak assignment for the H NMR spectrum of alpha-cyclodextrin in D2O Peak Peak Integration Peak Assignment Location Type Factor Chemical Functional Ring (ppm) Compound Group Position 3.60 Triplet 1 α-Cyclodextrin CH 4 3.61 Doublet of 1 α-Cyclodextrin CH 2 Doublets 3.83 Triplet 1 α-Cyclodextrin CH 5 3.88 Triplet 2 α-Cyclodextrin CH2 6 3.96 Triplet 1 α-Cyclodextrin CH 3 4.80 Singlet - D2O - 5.03 Doublet 1 α-Cyclodextrin CH 1

100

The 1H NMR spectrum of AA that is dissolved in deuterated chloroform has been previously discussed and shown, in Figure 4-1. For this part of the study, however, alpha-cyclodextrin was dissolved in deuterium oxide; not deuterated chloroform. The 1H

NMR spectrum of AA was therefore re-obtained using deuterium oxide as the solvent.

This spectrum is shown in Figure 4-18.

4.80 1.34 1.33 2.25 2.25 5.25 5.26 9.68 4.83

1.00 1.72 3.40 5.46

9 8 7 6 5 4 3 2 1 1 Figure 4-18: H NMR spectrum of AA in D2O

Comparing Figures 4-1 and 4-18, it can be seen that there is a dramatic difference between the 1H NMR spectra of AA. The reason for this disparity is the result of the solvent that is used to dissolve AA. In the presence of deuterium oxide, which is slightly acidic, AA reacts with D2O to form an equilibrium product; as illustrated in Figure 4-19.

ChemSketch was used to predict the 1H NMR spectrum of a solution containing AA and

101

the acetal-based equilibrium product shown in Figure 4-20. This proposed equilibrium reaction is confirmed through the comparison of the predicted 1H NMR spectrum, Figure

4-20, and the actual spectrum, Figure 4-18. The spectral patterns of these two spectra are very similar; within the error of the software program.

2 O O D D D 3 O 4 1 H3C H3C O D Figure 4-19: Equilibrium reaction between AA and D2O

2.00[9] 1.31[6] 10

8 O 7 7 H C O 8 3 CH3 D2O 6 3 1 D 9 2 H O 6 8a 4 D 5

3 9.33[8a] 5.78[3] 5.79[3]

2 5.81[3] 5.77[3] 1

0 10 9 8 7 6 5 4 3 2 1 0 1 Figure 4-20: Predicted H NMR spectrum of AA in D2O

102

The peak assignment list, shown in Table 4.7, corresponds to the 1H NMR spectrum of

AA in deuterium oxide, shown in 4-18. Figure 4-18 shows the approximate ratio of integration factors, from left to right, is 1:2:3:6. Looking back at Figure 4-1, the 1H NMR spectrum of AA in deuterated chloroform, the ratio of integration factors is 1:3. Looking at the peak assignment in Table 4.7, the ratio of AA’s protons, within its two functional groups, is still 1:3. The ratio for the acetal-based equilibrium product is 2:6; simplified to be 1:3. The doubling phenomenon indicates that the equilibrium product is twice as prominent as AA, within this solution.

1 Table 4.7: Peak assignment for the H NMR spectrum of AA in D2O Peak Peak Integration Peak Assignment Location Type Factor Chemical Functional Group (ppm) Compound Group Number 1.12 Doublet 6 Equilibrium Product CH3 3 2.04 Doublet 3 AA CH3 1 4.60 Singlet - D2O - 5.04 Quartet 2 Equilibrium Product CH 4 9.47 Quartet 1 AA O=CH 2

According to the literature24, 63-70 cyclodextrins act as AA scavengers by a size-enclosing mechanism. Cyclodextrins have hydrophilic exterior structures and lipophilic internal structures.24 This hydrophilic exterior makes water, or deuterium oxide, the solvent of choice to dissolve cyclodextrins. The lipophilic interior makes it favorable for aldehydes and other organics to enter its internal cavity.24, 62 As depicted in Figure 4-21, alpha- cyclodextrin encapsulates AA into its cyclical structure without the need for a chemical reaction. The force by which AA is held inside of alpha-cyclodextrin is hydrogen bonding.

103

Figure 4-21: Interaction mechanism for AA and alpha-cyclodextrin

Several authors71-74 have shown that 1H NMR can be used to validate cyclodextrin’s size- enclosing mechanism through NMR titration studies; as discussed in Section 2.4.2.1. A similar experimental procedure was followed in an attempt to reproduce the results, seen in Figure 2-8. In this experiment, however, alpha-cyclodextrin is the host molecule and

AA is the guest molecule. Samples were prepared by separately weighing and then dissolving each component in D2O. These two solutions, in separate vessels, were then combined into one NMR tube to achieve the desired molar ratios. Nine samples were prepared in all:

• Pure alpha-cyclodextrin

• Pure AA

• 0.2 to 1 (AA to alpha-cyclodextrin)

• 0.4 to 1 (AA to alpha-cyclodextrin)

• 0.6 to 1 (AA to alpha-cyclodextrin)

• 0.8 to 1 (AA to alpha-cyclodextrin)

104

• 1 to 1 (AA to alpha-cyclodextrin)

• 2 to 1 (AA to alpha-cyclodextrin)

• 3 to 1 (AA to alpha-cyclodextrin)

The protons that were monitored during this NMR titration study were previously identified in Figure 4-19 and Table 4.7. For AA there are two sets of protons which can be monitored: the aldehyde proton (group 2) and the methyl protons (group 1). For the acetal-based equilibrium product, the methyl protons (group 3) and the CH proton (group

4) were also monitored.

Figure 4-22 shows the comprehensive results from the NMR titration experiment that was conducted to study alpha-cyclodextrin’s AA scavenging mechanism; Appendix A contains the individual spectra from this study. This plot shows that as the concentration of AA increases, relative to that of alpha-cyclodextrin, the chemical shift of protons for

AA and its equilibrium product also increase until a saturation point is reached. The saturation point for the AA and alpha-cyclodextrin complex occurs at a one to one ratio.

This implies that every molecule of alpha-cyclodextrin can sequester only one molecule of AA. Figure 4-22 provides the experimental proof to confirm the interaction mechanism between AA and alpha-cyclodextrin, which has been reported for other host/guest complexes.61, 71-74, 93

105

0.070

0.060

Methyl Protons (AA) 0.050 Aldehyde Proton (AA)

Methyl Protons (AA's Equilibrium Product) 0.040 CH Proton (AA's Equilibrium Product)

0.030 Peak Shifting (ppm) 0.020

0.010

0.000 00.511.522.533.5 Guest to Host Ratio

Figure 4-22: Peak shifting of the protons for AA and its equilibrium product when titrated with alpha-cyclodextrin (solvent is D2O)

4.2 Effectiveness of AA Scavengers in Reducing the Amount of AA in

PET

The second objective of this work was to investigate the efficiency of these three scavenging agents (anthranilamide, MXDA, and alpha-cyclodextrin) in reducing the amount of AA that is present in PET. The intent of these scavenging agents is not to limit PET degradation and reduce the amount of generated AA. The purpose of adding

AA scavengers to PET is to interact with generated AA, reducing the amount that is able to migrate and affect the packaged food or beverage.

106

Acetaldehyde concentrations in PET have been studied through two gas chromatography methods. The first method quantifies the rate at which AA is generated during the processing of PET. The second technique, known as headspace analysis, is used to determine the amount of AA that remains residually trapped within PET.

4.2.1 AA Generation Rates

To quantify how much AA is created during processing, the AA generation rates of each sample were established. These measurements were made using a Perkin-Elmer

Automatic Thermal Desorption System (ATD-400) and AutoSystem XL Gas

Chromatograph. Details of the gas chromatograph column and testing conditions for these experiments were previously discussed in Section 3.5.1.

Under isothermal conditions, the establishment of AA generation rates were achieved by varying the sample’s heating time between 9 and 17 minutes. For each sample, rates were established at three different temperatures: 280, 290, and 300oC. This allowed for the determination of a sample’s activation energy (EA) through the derived Arrhenius equation; Equation 3 (Section 3.5.1). Compilations of raw data from these AA generation experiments are shown in Appendix B.

As an example, Figure 4-23 shows the AA generations rates that were determined for the

1200 ppm anthranilamide/PET blend sample. It can be seen from this figure that the AA generation of this sample at 280oC is around 1.0 ppm/minute. Increasing the temperature

107

by 10oC raises the rate to about 1.9 ppm/minute. At 300oC, the AA generation rate for the 1200 ppm anthranilamide/PET blend sample is 3.0 ppm/minute. This trend exemplifies the fact that increasing the melting/processing temperature of PET also increases the rate of degradation within the polymer.

45.0 280 C 40.0 290 C 35.0 300 C 30.0

25.0 y = 2.9939x - 10.456 R2 = 0.991 20.0 y = 1.8965x - 5.3717 R2 = 0.9924 15.0 Acetaldehyde (ppm) y = 1.0138x - 1.2626 R2 = 0.9278 10.0

5.0

0.0 0 5 10 15 20 Time (minutes)

Figure 4-23: AA generation plots for the 1200 ppm anthranilamide/PET blend

Plots like Figure 4-23 were prepared for each sample analyzed throughout this work: the virgin PET sample, the extruded PET control sample, and the various AA scavenger/PET blend samples; they are located within Appendix C. Table 4.8 lists the AA generation rates at 280oC, 290oC, and 300oC for each sample. The general trend is that as AA scavenger concentration increases, the AA generation rate decreases. This is better illustrated by Figures 4-24, 4-25, and 4-26. Respectively, these plots show the AA

108

generation rate as a function of scavenger concentration and temperature for the anthranilamide, alpha-cyclodextrin, and MXDA blend samples.

Table 4.8: AA generation rates Scavenger Type / Sample Concentration AA Generation Rate (ppm/min) (ppm) 280oC 290oC 300oC Virgin PET resin - 0.5 2.4 6.4 Extruded PET (Control) 0 1.6 2.9 8.1 100 1.4 3.8 7.5 200 1.9 4.8 7.1 Anthranilamide 500 1.4 2.6 6.0 1200 1.0 1.9 3.0 10,000 0.3 0.9 1.9 500 1.4 4.0 7.2 1200 0.8 3.1 6.2 5000 0.5 1.8 4.2 Alpha-Cyclodextrin 10,000 0.4 1.1 4.4 25,000 0.3 0.7 3.7 50,000 0.3 0.8 3.1 100 1.8 4.6 6.8 200 1.5 2.6 6.1 MXDA 500 1.5 2.3 5.4 1200 1.4 2.4 5.0 10,000 0.9 2.2 5.2

109

9.0

8.0 280 C 7.0 290 C 300 C 6.0

5.0

4.0

3.0

2.0 Generation Rate of AA (ppm/min.) AA of Rate Generation

1.0

0.0 0 2000 4000 6000 8000 10000 12000 AA Scavenger Concentration (ppm)

Figure 4-24: AA generation rate as a function of anthranilamide concentration

9.0

8.0 280 C 7.0 290 C 300 C 6.0

5.0

4.0

3.0

2.0 Generation Rate of AA (ppm/min.)

1.0

0.0 0 10000 20000 30000 40000 50000 60000 AA Scavenger Concentration (ppm)

Figure 4-25: AA generation rate as a function of alpha-cyclodextrin concentration

110

9.0

8.0 280 C 7.0 290 C 300 C 6.0

5.0

4.0

3.0

2.0 Generation Rate of AA (ppm/min.)

1.0

0.0 0 2000 4000 6000 8000 10000 12000 AA Scavenger Concentration (ppm)

Figure 4-26: AA generation rate as a function of MXDA concentration

There are two common behaviors that are observed in each of these plots (Figures 4-24,

4-25, and 4-26). The first familiar feature among these figures is that eventually the AA generation rate becomes independent of concentration; the slope of the plot nears zero.

For MXDA this appears to happen around 1200 ppm, for anthranilamide this phenomenon seems to occur between 1200 ppm and 10,000 ppm, and for alpha- cyclodextrin the slope looks to flatten in the region of 10,000 ppm (or 1 weight %). The difference between these values has to do with the molecular structure and interaction mechanism of these scavenging agents. As shown in Section 4.1, a molecule of both anthranilamide and alpha-cyclodextrin can only interact with one molecule of AA.

MXDA, however, can scavenge up to two molecules of AA.

111

The second common feature among Figures 4-24, 4-25, and 4-26, is that at 280 and

290oC there is an initial increase in the AA generation rate for each scavenging agent.

This deviation from the previously mentioned general trend is observed for the anthranilamide concentrations between 0 and 500 ppm, the alpha-cyclodextrin concentrations between 0 and 1200 ppm, and the MXDA concentrations between 0 and

200 ppm. Initially, this phenomenon was attributed to experimental error in the preparation and/or AA generation measurement of the processed PET control sample. To investigate this idea, a second control sample was produced and subsequently analyzed.

The AA generation results from this second sample upheld those from the first control sample.

Discussed in Section 3.3 are the twin-screw extrusion experiments that melt-blended the various scavenging agents with PET resin. Respectively, Sections 3.3.1, 3.3.2, and 3.3.3, describe in detail how each of the various AA scavenger/PET blend samples was prepared. Within these sections, Tables 3.1, 3.2, and 3.3 show the proportions of each sample that was extruded once, twice, and up to three times; this general concept is labeled as a sample’s thermal history.

It is known that increasing a PET sample’s thermal history will also increase its amount of degradation and consequently the AA concentration. To simulate this effect, a portion of the “one-time processed” PET control sample was re-extruded to establish a “two-time processed” control sample. A portion of this “two-time processed” PET sample was then

112

extruded again to produce the “three-time processed” control sample. A more detailed explanation is provided in Section 3.3.4.

The AA generation rates of the “one-time processed”, “two-times processed”, and “three- times processed” control samples were analyzed and are reported in Table 4.9. It can be seen from this data that as the thermal history of the PET sample increases, so does its

AA generation rate. This information shows that the initial increase in the AA generation rates of the lowest concentrated samples was the result of the sample preparation methodology.

Table 4.9: AA generation rates of control samples Sample Number of AA Generation Rate (ppm/min) Processing Times 280oC 290oC 300oC 1 1.6 2.9 8.1 Extruded PET 2 1.8 3.5 8.2 3 2.5 3.9 8.9

As previously mentioned, the intent of establishing AA generation rates at three different temperatures was to allow for the determination of each sample’s activation energy (EA).

Activation energy is obtained by means of the Arrhenius plot; which is derived from the

Arrhenius equation (Equation 2). The slope of this graph is the activation energy divided by the gas constant; the y-intercept is the natural log of pre-exponential factor. Figure

4-27 shows the Arrhenius plot for the 10,000 ppm (or 1 weight %) MXDA/PET blend sample. Similar plots were prepared for each sample and are located within Appendix D.

The comprehensive results from these individual graphs are shown in Table 4.10.

113

1.80

1.60

1.40

1.20 y = -28068x + 50.624 1.00 R2 = 0.9996 0.80

0.60 ln Rate

0.40

0.20

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 -0.20

-0.40 1/Temperature (1/K)

Figure 4-27: Arrhenius plot of 10,000 ppm MXDA/PET blend sample

According to the data in Table 4.10, the respective anthranilamide and MXDA blend samples possess similar activation energies; the lone exception being the 100 ppm samples. The similarity in values is attributed to the fact that anthranilamide and MXDA scavenge AA by similar mechanisms. In comparison, Table 4.10 shows that similarly concentrated alpha-cyclodextrin samples have higher activation energies. As previously shown, alpha-cyclodextrin sequesters AA by a mechanism that is completely different to those of anthranilamide and MXDA. This indicates that the activation energy, which corresponds to the generation of AA, is a function of the interaction method by which AA is scavenged.

114

Table 4.10 shows that four of the six evaluated alpha-cyclodextrin/PET blend samples have higher activation energies than any other sample. This demonstrates that it takes more energy to remove AA from the internal molecular structure of alpha-cyclodextrin than it is does to break the bonds that were formed by the reaction between AA and anthranilamide, or between AA and MXDA. Since these samples have higher activation energies than the virgin, unprocessed PET resin the data indicates that it requires more energy to remove AA from alpha-cyclodextrin’s structure than it does to generate AA during PET degradation.

Table 4.10: Activation energies Scavenger Type / Number of Concentration Activation Sample Processing Times (ppm) Energy (kJ/mol) Virgin PET resin 0 - 277 1 - 216 Extruded PET 2 - 196 (Control) 3 - 165 - 100 221 - 200 172 Anthranilamide - 500 189 - 1200 155 - 10,000 244 - 500 220 - 1200 267 Alpha- - 5000 281 Cyclodextrin - 10,000 324 - 25,000 347 - 50,000 327 - 100 179 - 200 188 MXDA - 500 166 - 1200 173 - 10,000 233

115

Examination of the data in Table 4.10 indicates that generally activation energy increases with increasing scavenger concentration. In other words, the greater the concentration of the scavenging agent, in a PET blend system, the more likely generated AA will be sequestered and not allowed to diffuse from the polymer. Traditionally, activation energies for PET resins have been shown to decrease with I.V. A comparison between

Tables 4.10 and 4.14, which will be discussed in Section 4.3.1.1, show that this is still true for the virgin PET and the three extruded PET control samples. The opposite, however, appears to be true for the various AA scavenger/PET blend samples. In this case, activation energy appears to increase with decreasing I.V. This phenomenon is assumed to be the combined result of both the sample blending method and the addition of AA scavenging agents, rather than a deviation from what has been previously observed.

4.2.2 Residual AA

As AA is generated, through the degradation of PET, it has the ability to diffuse out of the polymer. A portion of the generated AA, however, usually remains residually trapped within the matrix of PET. Since AA is able to diffuse out of PET, even at temperatures as low as 21oC,19 quantifying its residual concentration is important because many foods and beverages have a limited threshold for AA.2, 20

Residual AA was quantified by means of a procedure known as headspace analysis. This technique uses a Perkin-Elmer TurboMatrix 40 Headspace Sampler (TM-40) coupled to a

Perkin-Elmer AutoSystem XL Gas Chromatograph. Amorphous PET pellets or preform

116

samples are ground, in a liquid nitrogen environment, and then sieved to obtain a powder that increases the diffusion ability of AA. Samples are heated at an elevated temperature, not to melt the polymer but to volatize the residual AA that is trapped within the PET matrix. Details of the gas chromatograph column and testing conditions for these experiments were previously discussed in Section 3.5.2. Appendix E provides the comprehensive results for the residual AA experiments, for both the pelletized samples and preform samples.

4.2.2.1 Pelletized Samples

During each AA scavenger/PET melt-blending process, a portion of the extruded, pelletized sample was collected for headspace analysis. Immediately upon collection, this sample was place in a freezer to prevent the AA, within this PET sample, from volatilizing. At a later point in time this sample was further prepared to be analyzed.

Table 4.11 shows the residual AA results for the various pelletized AA scavenger/PET blend and control samples. Within this table, the effectiveness of each AA scavenger/PET blend sample’s ability to reduce the amount of residual AA has been quantified relative to the “one-time processed” control sample. It should also be noted that no data is reported for the 10,000 ppm (1 weight %) anthranilamide/PET blend sample because the entire allotment of sample was dried and crystallized prior to headspace analysis. It was assumed that the drying and crystallization processes removed much of the residual AA from this sample.

117

Table 4.11: Residual AA data for pelletized samples Scavenger / Number of Concentration Residual Amount of Sample Processing (ppm) AA (ppm) Reduction (%) Times PET resin 0 - 0.8 - 1 - 8.9 - Extruded PET 2 - 13.5 - (control) 3 - 14.5 - - 100 5.2 41.6 - 200 3.5 60.7 Anthranilamide - 500 3.0 66.3 - 1200 1.0 88.8 - 10,000 - - - 500 5.0 43.8 - 1200 3.3 62.9 Alpha- - 5000 2.8 68.5 Cyclodextrin - 10,000 2.6 70.8 - 25,000 2.8 68.5 - 50,000 2.6 70.8 - 100 4.1 53.9 - 200 3.3 62.9 MXDA - 500 3.6 59.6 - 1200 3.4 61.8 - 10,000 2.8 68.5

It can be seen that the results shown in Table 4.11 corroborate with the general trend that was observed for the previously discussed AA generation rate results. Generally, as the concentration of AA scavenging agent increases, the percent reduction of AA also increases. These results also indicate, as did Table 4.9, that the amount of AA (generated or residual) increases with increasing thermal histories. The virgin PET resin, that was not processed, has an initial residual AA concentration that is less than 1 ppm.

Processing this resin one time increases its residual AA content up to 8.9 ppm.

Processing this resin a second and then third time, increases the residual AA concentration to 13.5 and 14.5, respectively.

118

4.2.2.2 Preform Samples

Six of the AA scavenger/PET blend samples, along with a control sample, were chosen to be injection molded into preform samples for two reasons. The first reason was to blow- mold a portion of these preforms in to 2-liter bottles for further analysis. The second reason was to analyze the residual AA content, of these samples, that resulted from an injection molding process. Table 4.12 shows the results of this work.

Table 4.12: Residual AA data for preform samples Scavenger / Concentration Residual Amount of Sample (ppm) AA (ppm) Reduction (%) PET - 8.3 - 100 4.5 45.8 Anthranilamide 200 3.7 55.4 Alpha- 500 4.8 42.2 Cyclodextrin 1200 4.7 43.4 100 5.3 36.1 MXDA 200 4.6 44.6

The results shown in Table 4.12 clearly reveal two points. The first is that the addition of scavenging agents is successful in reducing the amount of detectable residual AA within

PET preforms. The second point confirms the previously observed results; the greater the amount of scavenging agent, the greater the reduction in AA. As with Table 4.11, the effectiveness of each AA scavenger/PET blend sample has been quantified relative to a

PET control sample.

119

4.2.2.3 Comparison of Results for Pelletized Samples and Preform Samples

Table 4.13 was prepared to compare the residual AA results for similar AA scavenger/PET blend pellet and preform samples. The sample blending process for the preform samples varied in comparison to the blending of the pelletized samples. The preform samples were produced by melt-blending virgin PET resin with a master-batch sample that was passed through the twin-screw extruder only once. In the end, these preform samples were composed of a portion that was processed once and the remaining amount was processed twice; twin-screw extruded once and injection molded once. As discussed in greater detail in Section 3.3, similarly concentrated pelletized samples contained portions that were extruded once, twice and three times.

Table 4.13: Comparison of the residual AA data for pelletized and preform samples Scavenger / Concentration Residual AA (ppm) Sample (ppm) Pellets Preforms PET - 8.9 8.3 100 5.2 4.5 Anthranilamide 200 3.5 3.7 500 5.0 4.8 Cyclodextrin 1200 3.3 4.7 100 4.1 5.3 MXDA 200 3.3 4.6

Since the scavenger concentrations for each of these sets of samples are assumed to be the same, the comparison of their residual AA content can be made; even though their overall thermal histories may differ. Table 4.13 shows that in most every case the residual AA concentration is fairly similar for each respective pelletized and preform

120

sample. The one common trend in both circumstances is that the more scavenging agent, the lower the residual AA content.

4.3 Physical Properties of AA Scavenger/PET Blend Samples

For sensitive PET packaging applications, reducing the amount of detectable AA can be an immense concern. This enhancement, however, is not acceptable if it comes at the cost of sacrificing the overall appearance and physical properties of PET. Studying changes in the physical properties of PET was a vital step in determining the overall benefit of adding AA scavenging agents. The properties that were analyzed to complete this objective are: intrinsic viscosity (I.V.), color, thermal properties, and oxygen permeability. For each type of analysis, the results for the various AA scavenger/PET blend samples were compared to those of an appropriate PET control sample.

4.3.1 Intrinsic Viscosity (I.V.)

The intrinsic viscosities (I.V.) of the various AA scavenger/PET blend and PET control samples were obtained by measuring their respective melt viscosities. Measurements were made by means of a RDA III viscoelastic tester, using parallel plate and plate geometry. For each sample, a conversion of melt viscosity to I.V. was made through the use of a calibration curve; established from the measurements of standardized samples possessing precisely known I.V.s. Further details of this experimental setup and testing conditions were discussed in greater detail in Section 3.6.1. All of the data from the melt

121

viscosity measurements, for both the pelletized samples and preform samples, are shown in Appendix F.

4.3.1.1 Pelletized Samples

Each AA scavenger/PET blend sample was prepared via twin-screw extrusion. During these melt-blending procedures, a portion of each blend was set aside for melt viscosity analysis. Prior to their evaluation, these samples were dried and crystallized overnight in a vacuum oven, at a temperature of 140oC.

Table 4.14 shows the I.V. for each sample and its change in comparison to that of virgin

PET resin. The general trend that is shown in this table indicates that increasing scavenger concentration results in decreasing I.V. It is believed, however, that increasing scavenger concentration is not the only reason for reductions in viscosity.

As previously mentioned, these AA scavenger/PET blend samples are composed of portions that have been extruded once, twice, and up to three times. It can be seen that one pass through the twin-screw extruded degrades the PET resin enough to reduce the

I.V. by 2.5%. A second and then third time through the extruder reduces the I.V. by

13.8% and 18.8%, respectively.

This sample blending method, however, does not completely explain the loss in I.V. for each sample. If that were the case, the “three times processed” sample should have the

122

lowest I.V. The I.V. of that sample is 0.65 dL/g; yet, the lowest I.V. (0.34 dL/g) belongs to the 1 weight % (or 10,000 ppm) MXDA/PET blend sample. This implies that the loss in each sample’s I.V. is due to a combination of both the addition of scavenging agents and the sample blending method used in this work.

Table 4.14: I.V. data for pelletized samples Scavenger / Number of Concentration I.V. Amount of Sample Processing Times (ppm) (dL/g) Reduction (%) PET resin 0 - 0.80 - 1 - 0.78 2.5 Extruded PET 2 - 0.69 13.8 3 - 0.65 18.8 - 100 0.67 16.3 - 200 0.68 15.0 Anthranilamide - 500 0.68 15.0 - 1200 0.66 17.5 - 10,000 0.51 36.3 - 500 0.73 8.8 - 1200 0.71 11.3 Alpha- - 5000 0.58 27.5 Cyclodextrin - 10,000 0.54 32.5 - 25,000 0.46 42.5 - 50,000 0.40 50.0 - 100 0.68 15.0 - 200 0.67 16.3 MXDA - 500 0.70 12.5 - 1200 0.68 15.0 - 10,000 0.34 57.5

4.3.1.2 Preform Samples

As previously mentioned in Section 4.2.2.2, seven PET blends were injection molded into preforms: six AA scavenger/PET blends and one control sample. Preforms of each sample type were put aside for melt viscosity measurements. The purpose of this work

123

was to isolate the effect of AA scavenger addition on I.V.; by keeping the thermal histories for the various AA scavenger/PET blend samples as much the same as possible.

Table 4.15 shows the results from these measurements.

Table 4.15: I.V. data for preform samples Scavenger / Sample Concentration (ppm) I.V. (dL/g) PET - 0.76 100 0.78 Anthranilamide 200 0.78 500 0.73 Alpha-Cyclodextrin 1200 0.70 100 0.77 MXDA 200 0.76

The data in Table 4.15 shows that the anthranilamide and MXDA samples show no reduction in I.V.; relative to the control sample (pure PET preforms). The alpha- cyclodextrin samples, however, did show a reduction in the viscosity. The 500 ppm alpha-cyclodextrin samples showed a 4% reduction and the 1200 ppm samples showed an

8% reduction; compared to the pure PET samples. This implies that it may be possible to add a small amount of AA scavenging agents, less than 500 ppm, and not affect the final

I.V. of the preform. The reason the alpha-cyclodextrin/PET blend samples have a lower

I.V. than the other samples is due to the addition levels of this scavenger. As an example, as shown in Table 4.14, when 500 ppm and 1200 ppm of anthranilamide or MXDA are melt-blended into PET, the resulting viscosities of these AA scavenger/PET blend samples are lower than those of the 500 ppm and 1200 ppm alpha-cyclodextrin/PET blend samples.

124

4.3.1.3 Comparison of Results for Pelletized Samples and Preform Samples

As previously discussed, the blending process for the preform samples varied in comparison to the blending for the pelletized samples; in terms of each respective sample’s overall thermal history. Since the concentrations are assumed to be the same, a comparison of their intrinsic viscosity values can be made among these sets of samples.

Table 4.16 shows the comparison between the I.V. of the pelletized samples and preform samples.

Table 4.16: Comparison of the I.V. data for pelletized and preform samples Scavenger / Concentration I.V. (dL/g) Sample (ppm) Pellets Preforms PET - 0.78 0.76 100 0.68 0.78 Anthranilamide 200 0.68 0.78 Alpha- 500 0.73 0.73 Cyclodextrin 1200 0.71 0.70 100 0.68 0.77 MXDA 200 0.67 0.76

While, the results for the alpha-cyclodextrin samples shows no change between the pelletized samples and the preform samples; changes in I.V. become apparent for the anthranilamide and MXDA samples. For both of the anthranilamide samples, the preform I.V. is 0.10 dL/g greater than the I.V. of the pellets. The same trend is true for the two MXDA samples; this time, however, the difference in I.V. is 0.09 dL/g.

There are two reasons for the disparities seen between pelletized and preform samples for both anthranilamide and the MXDA. The first reason, as previously mentioned, is that

125

the preform samples have less thermal history than the pelletized samples. The second reason is due to the mechanisms by which anthranilamide and MXDA each scavenger

AA. In each of these reactions, between anthranilamide and AA (Figure 4-7 or 4-11) or between MXDA and AA (Figure 4-14), water forms as a byproduct. It is well known that the presence of water decreases the I.V. Since the pelletized samples have a greater thermal history than the preform samples, there is a greater chance for more reactions with AA and a greater chance that residual water still exists in these samples.

4.3.2 Color

Part of PET’s appeal to the food and beverage industry is the combination of its excellent clarity and lack of color. This makes studying color generation in PET a vital step. As previously mentioned, the goal of adding AA scavengers to PET is to reduce the AA concentration without disturbing any of its desirable properties. If the addition of AA scavenging agents were to generate an undesirable color within PET, it could negatively affect the final appearance of the container and its attractiveness to the customer.

The color of PET samples can be studied by both the human eye and by analytical techniques. ASTM D 192587 describes a process by which the yellowness of white plastics can be quantified. In this method a meter is utilized, similar to the one used in this work, to measure the L, a, and b values of a crystallized polymer at room temperature.

The meaning of each of these three variables (L, a, and b) is described in Table 3.8.

Through a series of calculations these three values are then converted into Y, X%, and

126

Z%; as respectively shown in Equations 5, 6, and 7. Ultimately, the results of these conversions are used to calculate a yellowness index (YI) for each PET sample; as shown in Equation 4. Further information concerning this analytical procedure, the calibration of the instrument, and the equations used to tabulate the results can be found in Section

3.7.

4.3.2.1 Color Analysis

The determination of any color changes due to processing and/or AA scavenger additive addition, were made through the use of the Hunter Lab Color/Difference Meter D25-2.

The instrument was used to analyze the crystallized PET pellets of each blend; ultimately determining a yellowness index (YI) for each sample. Table 4.17 shows the L, a, and b values and the calculated yellowness index for each sample.

Table 4.17: L, a, and b values and yellowness index of pelletized samples Scavenger / Number Concentration Averaged Values Yellowness Sample of (ppm) Index Processing L a b Times PET resin 0 - 74.8 -1.8 -2.3 -7.2 1 - 68.5 -0.4 0.7 1.4 Extruded PET 2 - 71.5 -0.4 4.0 9.5 3 - 71.2 -0.4 5.6 13.7 - 100 66.2 -10.0 2.4 -4.3 - 200 69.9 -10.3 3.0 -2.9 Anthranilamide - 500 73.6 -7.2 7.4 11.0 - 1200 72.6 -7.6 6.8 9.3 - 10,000 68.5 -8.9 8.2 12.1 - 100 69.0 -11.4 3.3 -3.3 - 200 69.0 -11.7 4.8 0.4 MXDA - 500 73.5 -7.6 11.4 20.3 - 1200 71.7 -9.4 14.6 27.0 - 10,000 68.9 -10.8 11.6 18.8

127

The results for all of the PET samples containing alpha-cyclodextrin have been omitted from Table 4.17 because every sample had some brownness in its appearance. The relative prominence of this color altered the L, a, and b values and inherently yielded a false yellowness index for each alpha-cyclodextrin/PET blend sample. It can also be seen, in Table 4.17, that processing alone affects the b value and the yellowness index. As the number of passes through the twin-screw extruded increases, these two values also increase. The virgin PET resin, which had not been processed, has a b value of -2.3 and a

YI of -7.2; while, the values for the “one-time processed” sample are 0.7 and 1.4, respectively. A second and then third pass through the extruder further increases these values.

In terms of the anthranilamide/PET blends and the MXDA/PET blends, the general trend in Table 4.17 shows that both the b value and yellowness index increase with scavenging concentration. Reasonable yellowness indexes, for both anthranilamide and MXDA, were achieved when the scavenger concentration was decreased below 500 ppm. While their b values are higher, the yellowness indexes at the 100 and 200 ppm level for both of these scavengers are lower than that of the PET control sample that was only extruded once. The raw data from this analysis are shown in Appendix G.

4.3.2.2 Appearance of 2-Liter Bottles

As previously mentioned, one PET control sample and six AA scavenger/PET blend samples were injection molded into preforms. These preforms were then stretch-blow-

128

molded into 2-liter bottles. Figure 4-28 shows a representative sample for each of the seven concentrations that were produced: 1200 ppm alpha-cyclodextrin/PET blend, 500 ppm alpha-cyclodextrin/PET blend, 200 ppm MXDA/PET blend, 100 ppm MXDA/PET blend, 200 ppm anthranilamide/PET blend, 100 ppm anthranilamide/PET blend, and pure

PET resin. As mentioned in Section 4.3.2.1, the bottles containing alpha-cyclodextrin have a brownish tint. The two MXDA/PET blends and the two anthranilamide/PET blends all have an appearance that is indistinguishable from that of the pure PET bottle.

The similar appearance of these five bottles confirms the color results seen in Table 4.17.

Figure 4-28: 2-liter blow-molded PET bottles (from left to right: 1200 ppm alpha- cyclodextrin, 500 ppm alpha-cyclodextrin, 200 ppm MXDA, 100 ppm MXDA, 200 ppm anthranilamide, 100 ppm anthranilamide, pure PET)

4.3.3 Thermal Properties

A polymer’s thermal properties are important characteristics that can be critical in dictating its end uses. These properties also determine how the material should be

129

processed. The high melting temperature of PET requires that it be extruded or injection molded at or above 265oC. During stretch-blow-molding, the optimal operating window is above the glass transition temperature (Tg) and below the onset of crystallization. For

o o PET, this window is approximately between 80 C and 130 C. The moderate Tg of PET allows this semi-crystalline polymer to be predominantly amorphous at room temperature, when it is rapidly quenched as it cools from melting conditions. Crystallinity, however, can occur during cooling from the melt if the PET article is properly oriented. This is known as strain induced crystallization.

The glass transition (Tg) temperature, crystallization behavior, and melting temperature of the various PET samples were studied by differential scanning calorimetry (DSC). The results for the assorted AA scavenger/PET blend samples were compared relative to the results for the PET control samples; a virgin PET resin that was not extruded and an extruded PET sample. Figure 4-29 shows an example of a typical DSC curve obtained when a sample is heated at a rate of 10oC per minute; following a rapid quenching to remove any crystallinity. The glass transition temperature (Tg) is shown as the small endothermic step change in the baseline, around 80oC, in the scan. Crystallization of the material is shown as the only exothermic peak; occurring around 140oC. Finally, the melting behavior is also shown the prominent endothermic peak that occurred at about

235oC. Cooling the samples at a rate of 10oC per minute gave indication of their crystallization behavior when cooled from the melt. Figure 4-30 shows an example of this analysis, with the dominant feature being the exothermic crystallization peak that is shown at around 180oC.

130

2 Heat Flow Endo Up (mW)

0 30 80 130 180 230 280

-2 Temperature (oC)

Figure 4-29: DSC heating curve of the 5 weight % alpha-cyclodextrin/PET blend

0 Heat Flow Endo Up (mW) Up Flow Endo Heat 30 80 130 180 230 280

-2

-4 Temperature (oC)

Figure 4-30: DSC cooling curve of the 2.5 weight % alpha-cyclodextrin/PET blend

131

4.3.3.1 Glass Transition Temperature (Tg)

Table 4.18 shows the glass transition temperatures (Tg) for all of the experimental PET samples. It should be noted that the Tg of each sample was measured during the reheating step after being quenched from the melt. The results indicate that the addition of AA scavenging agents have no effect upon the Tg of PET. No significant change

o occurred as the Tg values ranged between 76 and 79 C.

Table 4.18: Glass transition temperature (Tg) data o Scavenger / Sample Concentration (ppm) Tg ( C) PET resin - 77.5 Extruded PET - 79.1 100 78.2 200 77.4 Anthranilamide 500 77.6 1200 77.3 10,000 78.2 500 78.9 1200 79.0 5000 77.9 Alpha-Cyclodextrin 10,000 76.3 25,000 77.3 50,000 76.0 100 78.4 200 78.6 MXDA 500 79.0 1200 78.5 10,000 78.4

4.3.3.2 Crystallization Behavior When Heating from the Glassy State

While there were only minor differences in the glass transition temperatures (Tg) among the various PET samples, some changes did occur with the samples’ crystallization

132

behavior when heated from the glassy state. As shown in Table 4.19, the heat of crystallization (delta H) values of all the processed materials are higher than that of the virgin PET resin. This shows a greater tendency of lower molecular weight PET to crystallize, while being reheated from the glassy state. The same conclusion can be stated for the “peak – onset of peak” values. The “peak – onset of peak” value is a measure of how quick a sample crystallizes and is obtained by subtracting the location of the crystallization curve’s onset from the location of its peak. The smaller the value, the quicker crystallization occurs. All of these numbers are lower than that of the virgin PET resin. In terms of the crystallization peak temperature, the general trend is that the peak temperature increases with decreasing AA scavenger concentration. This shows that the addition of AA scavengers may act as nucleating agents, making crystallization easier.

Table 4.19: Crystallization behavior data when heating from the glassy state Scavenger / Concentration Peak Delta H of Peak - Onset Sample (ppm) (oC) Peak (J/g) of Peak PET resin - 168 -18 27 Extruded PET - 160 -33 15 100 155 -34 13 200 152 -32 13 Anthranilamide 500 149 -31 11 1200 147 -32 11 10,000 150 -33 11 500 158 -33 12 1200 153 -33 12 Alpha- 5000 149 -34 10 Cyclodextrin 10,000 144 -36 10 25,000 141 -44 10 50,000 137 -22 14 100 156 -32 12 200 153 -33 13 MXDA 500 155 -34 12 1200 151 -33 11 10,000 151 -37 14

133

4.3.3.3 Melting Behavior

Analysis of the melting behavior, when heated from the glassy state, indicates that the addition of AA scavengers to PET has little affect upon the melting peak. Table 4.20 shows that all of blended resins have similar melting peaks to that of the extruded PET resin; due to the fact that they have all been extruded. Table 4.20 also shows that the heat of fusion values (delta H) increase slightly with increasing AA scavenger concentration, similar to the trend seen with the crystallization behavior that was previously discussed.

Table 4.20: Melting behavior data when heating from the glassy state Scavenger / Sample Concentration (ppm) Peak (oC) Delta H of Peak (J/g) PET resin - 223 17 Extruded PET - 229 31 100 231 32 200 227 34 Anthranilamide 500 226 32 1200 228 36 10,000 231 35 500 229 31 1200 230 32 5000 232 35 Alpha-Cyclodextrin 10,000 232 39 25,000 234 45 50,000 234 46 100 229 28 200 230 31 MXDA 500 230 32 1200 229 33 10,000 228 34

134

4.3.3.4 Crystallization Behavior When Cooling from the Melt

Table 4.21 shows the results of the crystallization behavior when cooled from the melt for the various PET samples. The analysis of this data shows some noticeable differences among the AA scavenger/PET blend samples. With respect to the crystallization peak temperatures, there appears to be a general trend of increasing peak temperature with increasing AA scavenger concentration. As discussed previously, this trend shows that the AA scavengers may act as nucleating agents because crystallization is generally occurring at a faster rate.

Table 4.21: Crystallization behavior data when cooling from the melt Scavenger / Sample Concentration (ppm) Peak (oC) Delta H of Peak (J/g) PET resin - 166 -5 Extruded PET - 162 -15 100 162 -25 200 167 -25 Anthranilamide 500 161 -15 1200 166 -28 10,000 170 -20 500 156 -15 1200 167 -24 5000 172 -29 Alpha-Cyclodextrin 10,000 176 -33 25,000 185 -44 50,000 191 -45 100 161 -17 200 167 -25 MXDA 500 161 -13 1200 162 -17 10,000 188 -26

It can also be seen, from Table 4.21, that the addition of AA scavengers has an effect upon the heat of crystallization values. Here is where there is a definite difference among

135

the AA scavengers. With respect to the anthranilamide and the MXDA, the data for both samples are very sporadic. No broad conclusions can be made, other than to say the overall addition of either scavenger increases the heat of crystallization. The alpha- cyclodextrin samples show that there is a definite trend of increasing delta H with increasing scavenger concentration.

4.3.4 Oxygen Film Permeation

Oxygen permeability is a measure of how much oxygen will permeate through a sample for a set of specified conditions. Table 4.22 shows the oxygen film permeability values obtained from sidewall samples cut from 2-liter bottles of varying AA scavenger/PET blends: pure PET, 500 ppm alpha-cyclodextrin/PET, 1200 ppm alpha-cyclodextrin/PET,

100 ppm anthranilamide/PET, 200 ppm anthranilamide/PET, 100 ppm MXDA/PET, and

200 ppm MXDA/PET. The data in Table 4.21 show that the oxygen permeability does not change due to the addition of these three AA scavenging agents.

Table 4.22: Oxygen Film Permeability Scavenger / Concentration Oxygen Permeability Sample (ppm) cc × STP × mil ( ) 100in 2 × day × atmospheres PET - 5.3 100 5.3 Anthranilamide 200 5.2 Alpha- 500 5.3 Cyclodextrin 1200 5.1 100 5.6 MXDA 200 5.4

136

4.4 Optimal AA Scavenger/PET Blends

The fourth objective of this work was to comprehensively evaluate the previously shown data and determine the optimal AA scavenger amounts to add to PET. As previously stated, reducing the amount of detected AA is a concern, however, not at the cost of sacrificing the desirable properties and physical appearance of PET. The goal is to determine the amount of AA scavengers that both minimizes the amount of detectable

AA and minimizes any possible negative effects upon the properties of PET. Evaluations of the various AA scavenger/PET blends samples were based upon the overall analysis of this research: AA generation rates, residual amount of AA, color, thermal properties, intrinsic viscosities (I.V.), and oxygen permeation.

4.4.1 Anthranilamide/PET Blends

A review of the data presented in Sections 4.2 and 4.3 revealed that optimal benefits of anthranilamide addition to PET are seen at concentrations of 200 ppm or less. As shown in Table 4.8, the 100 ppm anthranilamide/PET blend has lower AA generation rates than the extruded PET control sample at 280 and 300oC. The 200 ppm blend has a lower rate than the control at 300oC. Both of these blends, however, have rates higher than that of the control sample at 290oC. The 200 ppm blend also has a higher rate at 280oC. These anomalies are attributed to the sample blending practice which resulted in these two blend samples having portions of up to three thermal histories; as shown in Table 3.2.

137

These 100 and 200 ppm blends also showed benefits in reducing residual AA content.

For the pelletized samples, Table 4.11 shows that adding 100 ppm and 200 ppm of anthranilamide can have a 41.6% and 60.7% improvement, respectively, upon residual

AA concentration. When these same concentrations were achieved through the injection molding of preforms, shown in Table 4.12, the improvement in residual AA was 45.8% and 55.4%.

The melt-blending of 100 and 200 ppm of anthranilamide into PET has been shown to reduce the detection of AA. Additionally, these addition levels have shown little to no effect upon the physical properties of PET. Table 4.15 shows that the addition of up to

200 ppm of anthranilamide does not further reduce the resulting I.V. during injection molding of preforms. There was, however, a significant decrease in the measured I.V. for the pelletized 100 and 200 ppm/anthranilamide/PET blends; as shown in Table 4.14.

As previously mentioned, this phenomenon is due to the fact that these anthranilamide/PET blends samples have increased thermal histories relative to the control sample; which has only been extruded once.

Further analysis of the 100 and 200 ppm blends indicate that these anthranilamide addition levels result in little to no affect upon the color, thermal properties, and oxygen permeation of the evaluated samples. Table 4.17 shows that the pelletized 100 and 200 ppm anthranilamide/PET blends actually have lower yellowness indexes than the “one- time processed” control sample. Similarly, Figure 4-28 shows no differences among the appearances of the 100 and 200 ppm anthranilamide/PET blend 2-liter bottles and the

138

appearance of the pure PET 2-liter bottle. Tables 4.18, 4.19, and 4.20 demonstrate that these low anthranilamide addition levels do not alter the Tg, crystallization behavior, or melting behavior. The only apparent influence on PET thermal properties is observed for the heat of fusion values when the polymer blends are cooled from the melt. Table 4.21 shows that adding anthranilamide increases the delta H values in comparison to the results obtained for the extruded control sample. Finally, Table 4.22 shows that 100 and

200 ppm of anthranilamide addition has no affect upon the oxygen permeation of PET.

4.4.2 MXDA/PET Blends

Similar to anthranilamide, the optimal benefits of melt-blending MXDA into PET resin are seen at concentrations of 200 ppm or less. The data in Table 4.8 shows that the 200 ppm MXDA/PET blend has lower AA generation rates than the control sample at each of the three evaluated temperatures. A comparison between the 100 ppm blend and the control sample reveals that the experimental sample has a lower rate only at 300oC. As previously discussed in Section 4.4.1, this abnormality is attributed to the sample blending process. As shown in Table 3.3, this sample contains portions that have up to three thermal histories; the same can be said for the 200 ppm alpha-cyclodextrin/PET sample as well. The 100 and 200 ppm MXDA/PET blend samples contain portions which have been pass through the twin-screw extruder once, twice, and even three times.

The control sample, by comparison, has only been extruded once. As shown in Table 4.9, increasing the number of passes through an extruder not only increases a sample’s

139

thermal history, it also increases the degradation within that sample. This shortens the polymer chains and increases the number of byproducts, such as AA, which are present.

The addition of 100 and 200 ppm of MXDA to PET not only reduced the apparent AA generation rates, it also reduced the residual AA concentrations. Table 4.11 shows that the residual AA content of the pelletized samples was lowered by 53.9% and 62.9%, respectively, through the addition of MXDA. Analysis of the injection molded preform samples, as shown in Table 4.12, reveals that adding 100 ppm of MXDA reduced the residual AA content by 36.1%; adding 200 ppm resulted in a 44.6% reduction.

As discussed in Section 4.3, reducing AA content of PET is of concern, but not if it comes at the detriment of the physical properties and appearance of PET. The result of the melt viscosity measurements for the preform samples, shown in Table 4.15, indicate that the I.V. is not further reduced through the addition of up to 200 ppm of MXDA. The reason for the considerable I.V. decrease show in Table 4.14, for similarly concentrated pelletized samples, was addressed previously addressed in this section and in Section

4.4.1. The 100 and 200 ppm MXDA/PET blend samples are composed of portions which have up to three thermal histories. It has been shown in Table 4.14 that an increase in a sample’s thermal history results in a decrease in its I.V.

Further analysis of the physical properties of the 100 and 200 ppm MXDA/PET blends show that these additive levels cause little to no change in the color, thermal properties, and oxygen permeation. As shown in Table 4.17, these two MXDA samples have lower

140

yellowness indexes than that of the “one-time processed” control sample. Figure 4-28 illustrates that there is no visible differences among the appearances of the 100 and 200 ppm MXDA/PET blend 2-liter bottles and that of pure PET 2-liter bottle. As mentioned for the anthranilamide samples, Tables 4.18, 4.19, and 4.20 indicate that adding small amounts of MXDA will not affect the Tg, crystallization behavior, or melting behavior of the PET. The only influence that adding MXDA to PET may have is the heat of fusion value when the polymer is cooled from the melt. While Table 4.21 shows that the addition of MXDA does increase the delta H values relative to the extruded PET control sample, this data is sporadic and no general trend truly exists. As stated in Section 4.3.4, the addition of MXDA shows indication of affecting the oxygen permeation of PET.

4.4.3 Alpha-Cyclodextrin/PET Blends

For alpha-cyclodextrin, the optimal benefits of its addition to PET are seen at concentrations of 500 ppm or less. Table 4.8 shows that the 500 and 1200 ppm alpha- cyclodextrin blends have lower AA generation rates at 280 and 300oC than the extruded

PET control sample. Both of their AA generation rates at 290oC, however, are higher than that of the control sample. Previously mentioned for the anthranilamide and MXDA samples, this is due to the blending practice which resulted in portions of these samples having up to three thermal histories; as shown in Table 3.1.

Reducing the residual amount of AA was also achieved through the addition of alpha- cyclodextrin. The addition of 500 ppm lowered the residual AA content by 43.8%, when

141

compared to the results of the control sample. Table 4.11 also shows that the 1200 ppm alpha-cyclodextrin/PET blend reduced the residual AA content 62.9%. Table 4.12 shows the results obtained from the residual AA analysis of injection molded preform samples.

The addition of 500 ppm of alpha-cyclodextrin reduced the AA content by 42.2%; adding

1200 ppm limited the detectable AA concentration by 43.4%.

Based solely on the AA generation rates and the residual AA concentrations, it appears that increasing the amount of alpha-cyclodextrin in PET will yield the maximum benefits.

This preliminary conclusion is again solely based on the results from the two AA detection techniques and makes studying the changes in the physical properties and appearance of PET a vital step toward understanding the overall threshold for melt- blending alpha-cyclodextrin into PET.

Analysis of the thermal properties and oxygen permeation results for the alpha- cyclodextrin/PET blends revealed that neither of these properties was significantly affected by the addition of this scavenging agent. The conclusion based on Tables 4.18,

4.19, and 4.20 indicate that adding small amounts of alpha-cyclodextrin will not affect the Tg, crystallization behavior, or melting behavior of PET. According to Table 4.21, alpha-cyclodextrin may influence the heat of fusion value when PET is cooled from the melt. In fact, the data presented in that table indicate a general trend of increasing delta

H with increasing scavenger concentration. The oxygen permeation results, shown in

Table 4.22, signify that adding up to at least 1200 ppm of alpha-cyclodextrin will not affect this property.

142

Prominent changes in the physical properties and appearance of PET become an issue when examining the I.V. and color analysis results. For anthranilamide and MXDA, it was previously shown that the addition of these scavengers did not further reduce the I.V. of their preform samples in comparison to the I.V. of the control sample. For alpha- cyclodextrin, however, Table 4.15 shows that the addition of 500 and 1200 ppm of this scavenger additionally reduced preform I.V. by 0.03 and 0.06 dL/g, respectively.

As also observed with the other two scavengers, the addition of alpha-cyclodextrin appeared to result in a significant decrease in the I.V. for the pelletized samples. For

MXDA and anthranilamide, this result was attributed to the sample blending process; which caused the blend samples to have increased thermal histories in comparison to the control sample. The 500 and 1200 ppm alpha-cyclodextrin/PET blend samples were prepared in a similar manner and had similar thermal histories to those of anthranilamide and MXDA. The results from the preform I.V. analysis, however, indicate that the sample blending process is not the sole reason for the decrease in the pelletized samples’

I.V. It appears that the addition of as little as 500 ppm of alpha-cyclodextrin can affect the resulting I.V. of a PET blend sample.

As previously discussed in Section 4.3.2.1, the color results for each of the alpha- cyclodextrin/PET blend samples were omitted from Table 4.17. The reason for this omission was due to the fact that each sample had a brownish appearance that altered the

L, a, and b values and subsequently yielded a false yellowness index for each alpha- cyclodextrin/PET blend sample. This is further illustrated by Figure 4-28, where both of

143

the alpha-cyclodextrin/PET 2-liter bottles (500 ppm and 1200 ppm) shown in this figure have a slight brownish tint to their appearance. While the brownish appearance of the

500 ppm bottle is less than that of the 1200 ppm bottle, the appearance of the 500 ppm bottle would still not be acceptable for a commercial application.

The resulting color that emerges when alpha-cyclodextrin is melt-blended with PET appears to be a limiting factor towards its usage as an AA scavenging agent. The resulting color for even the 500 ppm alpha-cyclodextrin/PET blend had a noticeable amount of brown color to its appearance. This result indicates that either the concentration of alpha-cyclodextrin needs to be lowered beyond 500 ppm or further work is required to limit the formation of this color when alpha-cyclodextrin is melt-blending into PET resin.

4.5 Modeling

The fifth and final objective of this work was to develop a model, representing a multi- cavity injection molding system, capable of predicting the effectiveness of adding AA scavengers to PET. This initial attempt utilizes an existing AA generation modeling program, developed by the University of Toledo’s Polymer Institute, and modifies it to account for the various AA scavenging mechanisms discussed in this research. This modification is the addition of kinetic terms, determined experimentally, to characterize the ability of these scavengers to reduce the amount of detectable AA in PET.

144

Through the use of this program the addition of AA scavengers to PET can be studied in at least two manners. The first is to compare the AA generation results for a pure PET resin and an AA scavenger/PET blend, under identical processing conditions, to determine the effectiveness of the AA scavengers in reducing the detection of AA in a simulated production setting. The second way is to predict the amount of AA scavenger needed to meet the residual AA requirements for a particular packaging application. In either instance, the use of such a model could save time and effort by limiting the amount of experimental work that is required to generate this information.

4.5.1 Predictive AA Generation Program

The original predictive AA generation program was developed by the University of

Toledo’s Polymer Institute. This program models the accumulation of AA within PET preforms that results from the melting and shear hearing experienced during an injection molding process. It does this by using predetermined rheology and AA generation constants (for the particular PET resin being analyzed) to solve the momentum and the thermal energy equations.

Several assumptions are made in order to simply the momentum equation and thermal energy equation into the forms shown in Equations 11 and 12, respectively. This system assumes one directional flow, the z-direction, within the injection molder’s cylindrical channels. It also assumes that there is a constant pressure gradient and that the viscosity of the viscous, non-Newtonian PET melt is shear dependent. The program models both

145

steady flow and intermittent flow; whichever best represents the modeled system. Radial diffusion of AA is assumed to be negligible. The residence time within the flow channels is too short, on the order of seconds, for the diffusion of AA to become significant. The program works by modeling the accumulation of AA and neglects the transport of AA from that volume of material. Appendix H presents the derivation of the thermal energy equation, through the use of these assumptions, from the general form to the one shown in Equation 12.

∂P  1  δ  ∂v  =   r ×η ×  (Equation 11) ∂z  r  δr  ∂r 

Table 4.23: Explanation of the terms in Equations 11 Terms Meaning P Pressure z Longitudinal direction r Radial direction η Viscosity v Velocity ∂P Pressure forces

∂z  1  ∂  ∂v  Viscous forces   r ×η ×   r  ∂r  ∂r 

2  ∂T ∂T   1  ∂  ∂T   ∂v  ρ × C p  + v  =   k × r ×  +η  (Equation 12)  ∂t ∂z   r  ∂r  ∂r   ∂r 

146

Table 4.24: Explanation of the terms in Equation 12 Terms Meaning ρ Density Cp Heat capacity T Temperature t Time k Boltzmann’s constant ∂T Accumulation of heat over time ∂t ∂T Convection v ∂z  1  ∂  ∂T  Conduction   k × r ×  ( is the thermal  r  ∂r  ∂r  k conductivity) 2 Viscous dissipation  ∂v  η   ∂r 

As previously mentioned, the momentum and thermal energy equations, shown in

Equations 11 and 12, are used along with experimentally determined rheology and AA generation constants to predict the amount of AA that accumulates within PET preforms; for a particular PET resin. These two equations are continuously solved for finite volumes of space along the flow channels of a multi-cavity injection molder. Each individual section is treated as an isothermal system. The temperature, however, does change from section to section as heat accumulates. The momentum equation is solved to obtain the velocity profile. This is then used in the thermal energy equation to determine the viscosity term. The viscosity term is subsequently used within the momentum equation to restart this process for the next section. During this process, the generation of AA is calculated as a shear and temperature dependent product. The accumulation of AA is carried to each section, throughout the modeling of this system.

147

The rheology constants are determined through the curve-fitting of the capillary rheometry data. The trends for this data are determined by using the general forms shown in Equation 13 and 14. Table 4.25 provides an explanation for each term in these equations.

2 ln(τ ) = R5 + (R3 × X ) − (R4 × X ) (Equation 13)

R X = ln(γ ) − R + 2 1 T (Equation 14)

Table 4.25: Explanation of the terms in Equations 13 and 14 Terms Meaning Units τ shear stress Pascal (Pa) γ shear rate second-1

R1, R2, R3, R4, and R5 curve fitting constants - T temperature Kelvin (K)

Figure 4-31 shows the capillary rheometry trends for the Voridian CB12 PET resin used in this work. The individual results are listed in Table 4.26. Evaluations were conducted at three different temperatures (260, 270, and 280oC) so that these results could be extrapolated to predict the appropriate viscosity at any temperature. A separate program, also developed by the University of Toledo’s Polymer Institute, was used to analyze the data in Table 4.26. This program determined the constants needed to run the AA generation modeling program. These results are shown in Table 4.27.

148

1000

260 C 270 C 280 C Viscosity (Pa*s)

100 10 100 1000 10000 Shear Rate (1/s)

Figure 4-31: Viscosity versus shear rate curves for the Voridian CB12 PET resin

Table 4.26: Capillary rheometry results Temperature Crosshead Force Apparent Shear Apparent (oC) Speed (lbs.) Shear Rate Stress (Pa) Viscosity (inches/min.) (1/sec.) (Paxsec.) 0.1 133.2 67 62,694 939 0.3 347.0 200 163,325 816 260 1.0 766.0 667 360,538 540 3.0 1360.0 2002 640,120 320 10.0 3400.0 6674 1,600,299 240 0.1 110.2 67 51,869 777 0.3 286.0 200 134,613 672 270 1.0 686.0 667 322,884 484 3.0 1216.0 2002 572,342 286 10.0 2645.0 6674 1,244,939 187 0.1 93.4 67 43,961 659 0.3 234.0 200 110,138 550 280 1.0 595.0 667 280,052 420 3.0 1088.0 2002 512,096 256 10.0 1735.0 6674 816,623 122

149

Table 4.27: Rheology constants for the predictive AA generation program R1 R2 R3 R4 R5 17.38 6600.0 0.766 0.066 14.0

The AA generation rates for this PET resin were previously shown in Table 4.8; Section

4.2.1. For this sample, an additional data point was generated that is not listed in Table

4.8. At 270oC, the Voridian CB12 PET resin generated AA at a rate of 0.3 ppm per minute. Equation 15 describes the generation of AA in a PET. This has been determined to be a zero order reaction, dependent only upon temperature.15 Table 4.28 provides an explanation for the terms listed in this equation. As previously discussed, the AA generation rates can be plotted using the derived Arrhenius equation (Equation 3) to extrapolate these results to make predictions for temperatures beyond the ones which were evaluated. Figure 4-32 shows the Arrhenius plot for the Voridian CB12 PET resin.

d [ AA ] = R dt G (Equation 15)

Table 4.28: Explanation of the terms in Equation 15 Terms Meaning Units [AA] Concentration of AA ppm d[AA] Change in AA with respect to change ppm/minute

dt in time or AA generation rate RG AA generation rate for pure PET resin ppm/minute

150

2.50

2.00

1.50 y = -33365x + 60 R2 = 0.9632 1.00

0.50

ln Rate ln 0.00 0.00172 0.00174 0.00176 0.00178 0.00180 0.00182 0.00184 0.00186 -0.50

-1.00

-1.50

-2.00 1/Temperature (1/K)

Figure 4-32: Arrhenius plot for the Voridian CB12 PET resin

The previously mentioned AA generation constants are determined through the results shown in Figure 4-32. The Arrhenius plot for Voridian CB12 PET resin shows that the slope of the trend line is 33,365 and the y-intercept is 60. These two values are used as the constants shown in Equation 16. The y-intercept is constant aa1 and the slope is aa2.

Table 4.29 defines the various terms used in Equation 16.

aa ( aa − 2 ) 1 T RG = e (Equation 16)

151

Table 4.29: Explanation of the terms in Equation 16 Terms Meaning Units RG AA generation rate for pure PET resin ppm/minute

aa1 Constant from the Arrhenius Equation ppm/minute

aa2 Constant from the Arrhenius Equation Kelvin × ppm min.

T Temperature Kelvin

As previously discussed, the predictive AA generation program models the accumulation of AA within PET preforms that results from the melting and shear hearing exhibited during an injection molding process. As molten PET moves through the flow channels in the manifold, shear heating predominantly occurs at the surface of these cylindrical tubes.

This is further illustrated by Figure 4-33, which shows the modeling program’s prediction of temperature distribution as a function of radial distance from the center of a flow channel; for ten intervals of time within a two second period. It can be seen that the elevation of temperature is greatest near the wall surface, due to shear heating of the viscous polymer.

152

Figure 4-33: Temperature profile as a function of radial distance from the center of a flow channel over a two second period of time

Proximity to the flow channel’s sidewall not only affects the PET’s temperature, it also affects its AA concentration. It has been previously discussed that shear heating and temperature have a great influence the generation of AA within PET. As shown in Figure

4-34, this was verified using the AA generation modeling program. The results from

Figures 4-33 and 4-34 both show that the greatest effects are seen at the flow channel’s sidewalls. Minimal influence upon temperature and AA concentration are seen at the center, where shear heating is minimized.

153

Figure 4-34: Distribution of AA as a function of radial distance from the center of the flow channel

Within the manifold of a multi-cavity injection molder, material exits from the main flow channel and is then split off into multiple directions. This allows several cavities to be filled at once. This division may create an even distribution of material; however, it creates an uneven temperature and AA distribution. This effect is illustrated by Figures

4-35 and 4-36.

Figure 4-35 shows how material is distributed to fill the four cavities on the right-hand- side of an eight-cavity mold. The information in Figure 4-33 indicates that the material closest to the tube’s wall builds up the most heat. This effect diminishes the closer one gets to the center of flow channel. As PET is injected into the manifold and flows down

154

the main flow channel, material to fill cavity D is the closest to the side wall. Material to fill cavity C is the next closest to the sidewall, then followed by the material to fill cavity

B. Material to fill cavity A comes from the center of the main flow channel and has the lowest temperature. This statement was verified by the AA generation modeling program, as shown in Figure 4-36.

Figure 4-35: Distribution of material to fill four cavities within an eight-cavity mold

155

Figure 4-36: Temperatures for the various cavities as a function of filling times

The management of this material distribution is what makes this predictive AA generation program unique. The effects exhibited by a multi-cavity injection molder can be simulated, with this program, in order to optimize machine and manifold design to yield the best AA results possible. Evaluations can be made for a 16, a 24, a 32, or a 48 cavity system. The previously identified variables needed to run this program are summarized in Table 4.30.

156

Table 4.30: Variables needed to run the predictive AA generation program Variable Meaning Reason Used to Determine 5 Rheology Constants Determined by Parameters for Viscosity Constants Rheology Program Equations 13 and 14 aa1 Arrhenius Plot Constant (y- Parameter for RG intercept) Equation 15 aa2 Arrhenius Plot Constant Parameter for RG (slope) Equation 15 T Temperature (Kelvin) Parameter for RG Equation 15

4.5.2 Modified AA Generation Program

The original AA generation modeling program was developed to predict the amount of

AA that will accumulate within PET preforms due to the melting and shear heating that occurs during injection molding. In the previous section it was shown that this is achieved by solving the momentum and thermal energy equations along with rheology and AA generation models. The model constants were determined explicitly for the particular PET resin used in this analysis.

The original modeling program is useful for predicting the amount of AA that will be generated within PET preforms produced from pure resin. One limitation of this system, however, is that it is not able to account for the addition of AA scavengers added to PET resin. The addition of scavenging agents to PET has been analyzed, discussed, and shown to be effective in reducing the amount of detectable AA. The ability to account for and predict the effectiveness of their addition to PET would further improve upon the value of this AA generation modeling program.

157

To begin this process, the original AA generation rate equation, shown in Equation 15, must be modified. This equation states that the change in AA with respect to time is only a function of temperature. This is a zero order kinetic relationship.15 It has been previously shown, however, that the addition of AA scavengers to PET will affect this relationship.

When scavenging agents are added to PET, they are able to interact with the AA that is generated during processing. This interaction does not alter the amount of AA that is generated by the particular PET resin. It does, however, decrease the amount of AA that is detectable by either an analytical technique or a consumer. This new relationship leads to the development of Equation 17, which states that the change in the amount of detectable AA, with respect to time, is now a function of the resin’s AA generation rate minus the rate of reaction between AA and the scavenging agents.

The modification of Equation 15 is represented by the addition of the AA and AA scavenger reaction rate expression shown in Equation 17. This group of terms contains a reaction rate constant, k1 which is a function of temperature. Also accounted for are the concentrations of AA ([AA]), and scavenger interaction sites ([S]), and the reaction orders (a and b) for each of these components. k1, a, and b are generic kinetic terms of unknown quantities. For each scavenger, these terms will be determined experimentally and fit to the rate model shown in Equation 17. Table 4.31 provides an explanation for each of the terms shown in Equation 17.

158

d [ AA ] = R − k [ AA ]a [S ]b dt G 1 (Equation 17)

Table 4.31: Explanation of the terms in Equation 17 Terms Meaning Units d[AA] Change in AA with respect to change in time or AA ppm/minute

dt generation rate RG AA generation rate for pure PET resin ppm/minute

k1 Reaction rate constant for reaction between AA and dependant on the scavenger site quantity of a and b [AA] Concentration of AA at any point in time ppm [S] Concentration of AA scavenger sites at any point in ppm time a Reaction order for [AA] unitless b Reaction order for [S] unitless

It was previously shown, in Equation 16, that the AA generation rate (RG) can be expressed as an Arrhenius equation. The same is true for the reaction rate constant, k1 as shown in Equation 17. The terms in Equation 18 are also defined in Table 4.32.

bb ( bb − 2 ) 1 T k 1 = e (Equation 18)

Table 4.32: Explanation of the terms in Equation 18 Terms Meaning Units k1 Reaction rate constant for reaction between AA dependant on the and scavenger site quantity of a and b T Temperature Kelvin -1 -1 bb1 Constant from the Arrhenius Equation ppm min. bb Constant from the Arrhenius Equation Kelvin 2 ppm × min.

159

A similar reaction rate expression can be used to describe the consumption of the AA scavenger reaction sites; as shown in Equation 19. Both Equations 17 and 19 assume that one molecule of AA reacts with one functional site, on an AA scavenger, to produce one molecule of product. The 1H NMR results, discussed in Section 4.1, have shown that one molecule of anthranilamide and alpha-cyclodextrin can only react with one molecule of

AA. Each of these scavengers has only one functional site. MXDA, however, has two functional sites; two amine groups. 1H NMR results, Figure 4-15, indicate that one molecule of MXDA can react with up to two molecules of AA. In other words, MXDA has two functional sites.

One noticeable difference between these two equations is their respective reaction rate constant. The difference between these two constants is their units. In Equation 17, k1 is a reaction rate constant expressed in ppm (parts per million) units of mass concentration.

In Equation 19, however, k1’ is expressed in ppm units of molar concentration.

Converting between these two reaction rate constants can be achieved by accounting for the difference in molecular weight between AA and the AA scavenger being analyzed.

This conversion is shown in Equation 20. Table 4.33 provides further explanation of the terms used in both Equations 19 and 20.

d[S ] = −k '[ AA ]a [S ]b dt 1 (Equation 19)

160

MW S k1 ' = k1 ( ) (Equation 20) MW AA

Table 4.33: Explanation of the terms in Equations 19 and 20 Terms Meaning Units d[S] Change in amount of scavenger sites with respect to ppm/minute

dt change in time k1’ Derived reaction rate constant (based on weight) for dependant on the reaction between AA and scavenger site quantity of a and b [AA] Concentration of AA at any point in time ppm [S] Concentration of AA scavenger sites at any point in ppm time a Reaction order for [AA] unitless b Reaction order for [S] unitless

k1 Reaction rate constant (based on moles) for reaction dependant on the between AA and scavenger site quantity of a and b MWS Molecular weight of AA scavenger grams/mol

MWAA Molecular weight of AA grams/mol

Table 4.34 is shown to review the preform residual AA results previously presented in

Table 4.12 (Section 4.2.2.2). This data is used to compare the initial AA scavenger concentrations versus the amount of AA that is generated within the pure PET preforms; which is the theoretical concentration of AA generated within each sample. This comparison indicates that during an injection molding process the AA scavenger concentration remains essentially constant in comparison to the concentration of generated AA within each preform sample. Fundamentally, this means that the AA scavenger concentration within Equation 17 can be treated as a constant, which only varies with respect to change in temperature.

161

Table 4.34: Review of the residual AA data for preform samples Scavenger / Concentration Sample’s Theoretical AA Scavenger Sample (ppm) Residual Concentration Concentration – AA (ppm) of AA Theoretical Generated Concentration of within Each Generated AA Sample (ppm) (ppm) PET - 8.3 - 100 4.5 91.7 Anthranilamide 200 3.7 191.7 8.3 Alpha- 500 4.8 491.7

Cyclodextrin 1200 4.7 1191.7 100 5.3 91.7 MXDA 200 4.6 191.7

Establishing the AA scavenger concentration as a constant, which only varies with temperature, allows this term to be combined with any other constants that have the same dependency. As previously stated, the reaction rate term, k1, fits this criteria. The combination of these terms is shown in Equation 21; yielding a new reaction rate constant, k2. The implementation of k2 into Equation 17 produces a new mathematical relationship

(Equation 22) to describe the appearance of AA when AA scavengers are added to PET.

Table 4.35 provides further explanation of the terms used in both Equations 21 and 22.

b k 2 = k 1 [S 0 ] (Equation 21)

d [ AA ] = R − k [ AA ]a dt G 2 (Equation 22)

162

Table 4.35: Explanation of the terms in Equations 21 and 22 Terms Meaning Units k2 Reaction rate constant for reaction between AA and dependant on the scavenger site, assuming constant amount of AA quantity of a and b scavenger sites k1 Reaction rate constant for reaction between AA and dependant on the scavenger site quantity of a and b [S0] Initial, constant amount of AA scavenger sites with ppm time b Reaction order for [S] and [S0] unitless d[AA] Change in AA with respect to change in time or AA ppm/minute

dt generation rate RG AA generation rate for pure PET resin ppm/minute [AA] Concentration of AA at any point in time ppm a Reaction order for [AA] unitless

Within the modified AA generation modeling program, the original relationship describing the change in AA concentration with time (Equation 15) has been replaced with one that accounts for the addition of AA scavengers to PET. The formation of

Equation 22 simplifies this new relationship from the one previously presented in

Equation 17. The development of Equation 22 also simplifies the process to solve for the desired variables.

The variables needed to run the original version of this AA generation modeling program have been previously discussed and are listed in Table 4.30. The ones needed to run the modified program are listed in Table 4.36. This list includes the terms previously identified in Table 4.30 and the new variables needed to simulate the effects of AA scavenger addition. Ultimately, the value of these terms will be used within the modified

AA generation modeling program to solve for the respective variables they represent.

163

Table 4.36: Variables needed to modify the predictive AA generation program Variables Meaning Reason Used to Determine 5 Rheology Constants Determined by Parameters for Viscosity Constants Rheology Program Equations 13 and 14 aa1 Arrhenius Plot Constant (y- Parameter for RG intercept) Equation 16 aa2 Arrhenius Plot Constant Parameter for RG (slope) Equation 16 T Temperature (Kelvin) Parameter for RG and k1 Equation 16 and 18 AA Scavenger Initial Amount of AA Initial Parameter for d[AA] and k2 Concentration Scavenger Added to PET Equation 17 or 21 dt Resin a Reaction Order for Parameter for d[AA]

Concentration of AA ([AA]) Equation 17 or 22 dt b Reaction Order for Parameter for d[AA] and k2 Concentration of AA Equation 17 or 21 dt Scavenger ([S] or [S0]) bb1 Arrhenius Plot Constant (y- Parameter for k1, k1’, and k2 intercept) Equation 18 bb2 Arrhenius Plot Constant Parameter for k1, k1’, and k2 (slope) Equation 18

4.5.2.1 Numerical Analysis

The determination of the variables used to run the original predictive AA generation program, listed in Table 4.30 (5 rheology constants, aa1, and aa2), have been previously discussed in Section 4.5.1. These terms are also shown in Table 4.36; which lists all the variables needed to run the modified AA generation program. The remaining terms (a, b, bb1, and bb2) listed in this table have been added to the original modeling program to account for the addition of AA scavenging agents to PET; creating the modified AA generation program. These variables need to be quantified in order to run the modified

AA generation program discussed in Section 4.5.2.

164

4.5.2.1.1 Determination of k2 and a

To determine k2 and a, a rearrangement of Equation 22 was needed. As previously discussed, Equation 22 describes how the change in AA concentration, with respect to time, is a function of the AA generation rate of the pure PET resin minus the rate of reaction between AA and the AA scavenger; assuming the concentration of the AA scavenger is constant with time. These terms are slightly reorganized to produce

Equation 23 and then expressed as a natural log function (ln) to give Equation 24.

d[AA] Plotting ln(R − ) versus ln([AA]) (both sets of terms are shown in Equation 24) G dt yields a as the slope and ln(k2) as the y-intercept.

d [ AA ] R − = k [ AA ]a G dt 2 (Equation 23)

d [ AA ] ln( R − ) = ln( k ) + a × ln([ AA ]) (Equation 24) G dt 2

To obtain the desired sets of terms shown in Equation 24, new AA generation studies were conducted for each evaluated scavenger: anthranilamide, alpha-cyclodextrin, and

MXDA. These new experiments were conducted because the previously presented AA generation data, discussed in Section 4.2.1 and listed in Appendix B, contain only three data points obtained over a short range of time (9, 13, and 17 minutes). Using these

165

parameters, it is difficult and ambiguous to fit these data sets with anything besides a linear model. Equation 17, however, indicates that when AA scavengers are added to

PET the concentration of AA will not be linear with time.

These new AA generation studies evaluated the 10,000 ppm AA scavenger/PET blend samples, for each scavenger, at 290oC for 60 minutes. Measurements were made in five minute increments beginning at 5 and continuing up to 60 minutes. Twelve data points in all were obtained. These evaluations provided the increased number of data points and increased length of time necessary to better evaluate the time dependency of the rate model; Equation 17. The results of these experiments are shown in Figures 4-37, 4-38, and 4-39. Additionally, the data corresponding to these figures are listed in Appendix I.

Within each figure a solid line is drawn to represent the rate of AA generation for the pure Eastman Chemical Voridian CB12 PET resin that was used in this work. Each set of experimental data was fit with a polynomial trend-line; which will be used in subsequent calculations. As previously seen in Section 4.2, the results in Figures 4-37,

4-38, and 4-39 indicate that the presence of AA scavengers reduces the amount of detectable AA with PET.

The polynomial trend-lines, determined for each plot, are used to determine the instantaneous change in AA concentration with respect to time; d[AA] . This is dt determined by calculating the derivative of this curve and evaluating it for each established 5 minute interval. Equation 25 depicts this calculation.

166

d [ AA ] = −(2 × c × time ) + c dt 1 2 (Equation 25)

180.0

160.0 Experimental Data 140.0 Generation of AA for CB12 PET Resin 120.0

100.0

80.0

60.0 Acetaldehyde (ppm) y = -0.0084x2 + 0.9771x - 5.0688 2 40.0 R = 0.9657

20.0

0.0 0 102030405060 Time (minutes)

Figure 4-37: 60 minute AA generation curve for the 10,000 ppm anthranilamide/PET blend

167

180.0

160.0 Experimental Data 140.0 Generation of AA for CB12 PET Resin 120.0

100.0

80.0

60.0 Acetaldehyde (ppm) y = -0.0074x2 + 1.3078x - 6.9309 2 40.0 R = 0.9938

20.0

0.0 0 102030405060 Time (minutes)

Figure 4-38: 60 minute AA generation curve for the 10,000 ppm alpha- cyclodextrin/PET blend

180.0

160.0 Experimental Data 140.0 AA Generation for CB12 PET Resin 120.0

100.0

80.0

60.0 Acetaldehyde (ppm)

40.0 y = -0.0192x2 + 2.8387x - 13.252 R2 = 0.9961 20.0

0.0 0 102030405060 Time (minutes)

Figure 4-39: 60 minute AA generation curve for the 10,000 ppm MXDA/PET blend

168

Beginning at the first data point (5 minutes) and continuing up to the 60 minute mark, these d[AA] values are then subtracted from the 290oC AA generation rate of the “one- dt time” processed PET sample; which was previously shown to be 2.9 ppm/minute. This sample was chosen as the control because its thermal history most closely matches those of the experimental samples, while free from the addition of AA scavenging agents. The

d[AA] result of this calculation yields the various R − values. This group of terms is G dt then expressed as a natural log function to yield the left-hand-side of Equation 24. The final set of terms, ln([AA]), is obtained by expressing the average AA concentration as a natural log function.

Respectively, Figures 4-40, 4-41, and 4-42 show the graphs needed to determine the k2, at

290oC, and a values for anthranilamide, alpha-cyclodextrin, and MXDA. The data corresponding to these plots are located in Appendix J. As previously mentioned, the slopes of these figures equal the respective a values and the y-intercepts equal the natural

 d[AA]  log of k2. Equation 24 predicts that the relationship between ln RG −  and  dt  ln([AA]) should be linear. Within each of these figures, however, the experimental data shows that at high AA concentrations the trend is not linear; linearity is observed at lower

AA concentrations. One possible reason for this phenomenon is the “plateau” effect that occurs in each of the 60 minute AA generation studies; Figures 4-37, 4-38, and 4-39. To an extent, this effect is observed in each of these figures, but is most prominent for

 d[AA]  anthranilamide. Consequently, the ln RG −  vs. ln([AA]) fit, Figure 4-40, is the  dt 

169

least linear among the scavengers. The best fit is for MXDA, Figure 4-42, and its 60 minute AA generation results level-off less than anthranilamide or alpha-cyclodextrin.

The results from Figures 4-40, 4-41, and 4-42 are compiled into Table 4.37; where k2 is shown as its true value and not as ln(k2). It can be seen that the reaction rate constant (k2) and the reaction order (a) for anthranilamide and alpha-cyclodextrin are fairly similar.

The respective terms for MXDA, however, differ greatly compared to those of the other two scavengers. This would appear to indicate that anthranilamide and alpha- cyclodextrin have similar AA scavenging efficiencies; while MXDA is much less effective compared to the other two.

1.20

1.00

0.80

0.60 - (d[AA]/dt))

G y = 0.1052x + 0.6569 R2 = 0.7476 ln (R ln 0.40

0.20

0.00 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 ln ([AA]) d[AA] Figure 4-40: Plot of ln(R − ) versus ln([AA]) for the 10,000 ppm G dt anthranilamide/PET blend at 290oC

170

1.00

0.90

0.80

0.70

0.60

0.50 - (d[AA]/dt)) G 0.40 y = 0.1042x + 0.4394 R2 = 0.8017 ln (R ln 0.30

0.20

0.10

0.00 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 ln ([AA]) d[AA] Figure 4-41: Plot of ln(R − ) versus ln([AA]) for the 10,000 ppm alpha- G dt cyclodextrin/PET blend at 290oC

1.00

0.50

0.00 0.00 1.00 2.00 3.00 4.00 5.00

-0.50 - (d[AA]/dt)) G

ln (R ln -1.00 y = 0.6324x - 2.1881 R2 = 0.9408 -1.50

-2.00 ln ([AA]) d[AA] Figure 4-42: Plot of ln(R − ) versus ln([AA]) for the 10,000 ppm MXDA/PET G dt blend at 290oC

171

o Table 4.37: Calculated k2, for 290 C, and a value for each scavenging agent o AA Scavenger k2 at 290 C a Anthranilamide 1.93 0.1052 Alpha-Cyclodextrin 1.55 0.1042 MXDA 0.11 0.6324

Through the use of Polymath® 6.10 software, the k2 and a values listed in Table 4.37 were used to examine how well these calculated terms fit with the experimentally determined data points previously shown in Figures 4-37, 4-38, and 4-39. To perform this task, the software package was set-up to solve Equation 22 for each of the three AA scavenging agents. The respective results are shown as a dashed, green line within

Figures 4-43, 4-44, and 4-45. For each figure, there appears to be a good correlation between the experimental data points and this predicted d[AA] curve. dt

180.0

160.0 Experimental Data

140.0 No Scavenger Fitted Curve with a and k2 120.0

100.0

80.0

60.0 Acetaldehyde (ppm)

40.0

20.0

0.0 0 102030405060 Time (minutes)

Figure 4-43: 60 minute AA generation data for the 10,000 ppm anthranilamide/PET blend fitted with Equation 22, using a and k2 values

172

180.0

160.0 Experimental Data 140.0 No Scavenger Fitted Curve with a and k2 120.0

100.0

80.0

60.0 Acetaldehyde (ppm)

40.0

20.0

0.0 0 102030405060 Time (minutes)

Figure 4-44: 60 minute AA generation data for the 10,000 ppm alpha-cyclodextrin/PET blend fitted with Equation 22, using a and k2 values

180.0

160.0 Experimental Data 140.0 No Scavenger

120.0 Fitted Curve with a and k2

100.0

80.0

60.0 Acetaldehyde (ppm)

40.0

20.0

0.0 0 102030405060 Time (minutes)

Figure 4-45: 60 minute AA generation data for the 10,000 ppm MXDA/PET blend fitted with Equation 22, using a and k2 values

173

4.5.2.1.2 Determination of k1, bb1, bb2, and b

The determination of k1 and b is achieved through a rearrangement of Equation 21. This modified equation is shown in Equation 26. Equation 26 shows three dependent variables (b, k1, and k2) and two independent variables. The reaction rate order, b, is dependent only on the type of AA scavenger; it is independent of AA scavenger concentration and temperature. The reaction rate constant k1 is a function of temperature

94 and independent of AA scavenger concentration. As shown in Equations 21 and 26, k2 is a function of temperature, initial scavenger concentration, and the type of scavenger.

ln( k 2 ) = ln( k1 ) + b × ln([ S 0 ]) (Equation 26)

Since Equation 26 contains two independent variables, temperature and initial AA scavenger concentration ([S0]), k1 and b need to be determined through multiple linear regression. In order to perform this analysis, various initial scavenger concentrations

([S0]) and temperatures must be evaluated. To achieve this, the original AA generation data that was discussed in Section 4.2.1, and shown in Appendices B and C, was used.

This data was established for each AA scavenger at temperatures of 280, 290, and 300oC and at heating times of 9, 13, and 17 minutes.

SigmaPlot® 2000 software was used to perform the multiple linear regressions. The regression analysis was set-up to evaluate the various ln(k2) and ln([S0]) terms as functions of temperature. For each evaluated AA scavenger/PET blend sample, the

174

temperature, used during AA generation analysis, and [S0] are readily available values. k2, however, is a set of values that needs to be calculated.

To begin these calculations, Equation 22 was used. This equation states that the change in AA concentration, with respect to time, ( d[AA] ) is a function of the pure PET resin’s dt

AA generation rate (RG) minus the rate of reaction between AA and the AA scavenger

a (k2[AA] ). This relationship in Equation 22 has been rearranged into the form shown in

Equation 27; which is more suitable for calculating k2. To calculate the k2 values for each

AA scavenger and each of the three evaluated temperatures, d[AA] was assigned to be dt the respective AA generation rate for the studied AA scavenger/PET blend. RG was the

AA generation rate of the pure, “one-time” processed PET resin. Values for both of these terms can be found in Table 4.8.

d [ AA ] ( RG − ) k = dt 2 [ AA ]a (Equation 27)

The [AA]a term was obtained by raising the AA concentration, determined for each time interval, to the appropriate a value listed in Table 4.37. For each set of conditions, AA scavenger/PET blend concentration and temperature, the [AA]a values were averaged among the 9, 13, and 17 minute data to obtain one value to be used in Equation 27.

An averaged value was used because, for the most part, the disparity among the [AA]a values for the 9, 13, and 17 data was minimal and this simplified the process.

175

These series of calculations, to determine k2, were carried out for each AA scavenger at each of the three evaluated temperatures. Once k2 is determined, it is expressed as a natural log function, ln(k2), to be used in the multiple linear regression analysis. The compilation of this tabulated data is listed in Appendix K.

The multiple linear regression results are shown in Table 4.38. This table shows the b, bb1, and bb2 terms for each AA scavenger. It was previously shown in Equation 18, that the bb1 and bb2 terms can be used to calculate k1 values for any desired temperature.

Table 4.38: b, bb1, and bb2 values for each scavenging agent determined through multiple linear regression AA Scavenger b bb1 bb2 Anthranilamide 0.5421 43.93 27,080 Alpha-Cyclodextrin 0.3466 26.08 16,310 MXDA 0.2898 50.60 31,000

Figures 4-46, 4-47, and 4-48 were prepared to examine the how well the determined a, b,

o bb1, and bb2 values fit their respective 290 C, 60 minute AA generation data. Similar to the preparation of those previous figures (4-43, 4-44, and 4-45), Polymath® 6.10 software was used to solve the predictive modeling equation, Equation 17, by using the a values, listed in Table 4.37, and b, bb1, and bb2 values, shown in Table 4.38.

176

180.0

160.0 Experimental Data

140.0 No Scavenger Fitted Curve with a, b, and k1 120.0

100.0

80.0

60.0 Acetaldehyde (ppm)

40.0

20.0

0.0 0 102030405060 Time (minutes)

Figure 4-46: 60 minute AA generation data for the 10,000 ppm anthranilamide/PET o blend fitted with Equation 17; using the a, k1 at 290 C, and b values

180.0

160.0 Experimental Data No Scavenger 140.0 Fitted Curve with a, b, and k1 120.0

100.0

80.0

60.0 Acetaldehyde (ppm)

40.0

20.0

0.0 0 102030405060 Time (minutes)

Figure 4-47: 60 minute AA generation data for the 10,000 ppm alpha-cyclodextrin/PET o blend fitted with Equation 17; using the a, k1 at 290 C, and b values

177

180.0

160.0 Experimental Data No Scavenger 140.0 Fitted Curve with a, b, and k1 120.0

100.0

80.0

60.0 Acetaldehyde (ppm)

40.0

20.0

0.0 0 102030405060 Time (minutes)

Figure 4-48: 60 minute AA generation data for the 10,000 ppm MXDA/PET blend fitted o with Equation 17; using the a, k1 at 290 C, and b values

A modified form of the kinetic model is shown in Equation 28. This form is a combination of Equations 18, 21, and 22. There are three fundamental variables for which Equation 28 studies: temperature (bb2), the concentration of AA (a), and the AA scavenger concentration (b). The 60 minute AA generation data was used study how the concentration of AA changes with time; producing the a values. The original 9 to 17 minute AA generation data, discussed in Section 4.2.1 and listed in Appendix B, indicates the effect of temperature and initial scavenger concentration. This set of data was used to determine the b and bb2 terms through multiple linear regressions. Since the bb1 term does not depend on any other variable, it is assumed to be a constant or a scaling factor. By modifying this independent term (bb1), for each AA scavenger, the modeled fit of Figures 4-46, 4-47, and 4-48 can be improved.

178

 bb 2  d [ AA ]  bb 1 −  = R − e  T  [ AA ]a [S ]b dt G (Equation 28)

The reason the bb1 terms needed to be optimized is due to the use of multiple data sets to determine these kinetic terms. The initial AA generation data, discussed in Section 4.2.1, contained only three data points that were measured at between 9 and 17 minutes. This data was fit a linear trend, indicating the sample’s AA generation rate as the slope. This information was generated for each AA scavenger/PET blend sample at three different temperatures.

This approach works well for a PET sample with no AA scavenger because its measured

AA generation rate is constant; as illustrated by Equation 15. Generating this information requires only a few data points. When AA scavengers are added to PET resin, however, this approach may need to be extended. As shown by Equation 22, the rate of change in

AA concentration changes over time. To observe this effect, more data points and longer evaluation time was needed. To achieve this, a 60 minute AA generation study, at 290oC, was conducted for only one concentration of each AA scavenging agent.

It is apparent that the 60 minute AA generation study has two advantages over the data presented in Section 4.2.1. Not only do the evaluated AA scavenger/PET blend samples have similar thermal histories, this study generated 12 data points for each sample. It is therefore believed that the data obtained during the 60 minute AA generation study is the

179

more accurate representation to model against. Modifications to the bb1 values were achieved through the use of these 60 minute AA generation data sets.

Once again, the Polymath® 6.10 software package was used to evaluate the governing predictive modeling equation, represented as Equation 17. With the values listed in

Table 4.38 as starting references, alterations were made to each AA scavenger’s bb1 value until the most desired modeled curve was obtained. It is important to note that throughout this process the a, b, and bb2 values remained unchanged. Respectively,

Figures 4-49, 4-50, and 4.51 show the results of these modifications relative to anthranilamide’s, alpha-cyclodextrin’s, and MXDA’s experimental data points. Table

4.39 lists the final a, b, bb1, and bb2 values for each AA scavenging agent evaluated in this work.

180.0

160.0 Experimental Data

140.0 No Scavenger Fitted Curve with a, b, and k1 120.0

100.0

80.0

60.0 Acetaldehyde (ppm)

40.0

20.0

0.0 0 102030405060 Time (minutes) Figure 4-49: 60 minute AA generation data for the 10,000 ppm anthranilamide/PET nd blend fitted with Equation 17; using the a, b, and bb2 values and 2 iteration bb1 value

180

180.0

160.0 Experimental Data No Scavenger 140.0 Fitted Curve with a, b, and k1 120.0

100.0

80.0

60.0 Acetaldehyde (ppm)

40.0

20.0

0.0 0 102030405060 Time (minutes) Figure 4-50: 60 minute AA generation data for the 10,000 ppm alpha-cyclodextrin/PET nd blend fitted with Equation 17; using the a, b, and bb2 values and 2 iteration bb1 value

180.0

160.0 Experimental Data No Scavenger 140.0 Fitted Curve with a, b, and k1 120.0

100.0

80.0

60.0 Acetaldehyde (ppm)

40.0

20.0

0.0 0 102030405060 Time (minutes) Figure 4-51: 60 minute AA generation data for the 10,000 ppm MXDA/PET blend fitted nd with Equation 17; using the a, b, and bb2 values and 2 iteration bb1 value

181

Table 4.39: Final a, b, bb1, bb2, and k1 values for each AA scavenging agent k AA Scavenger a b bb bb 1 1 2 280oC 290oC 300oC Anthranilamide 0.1052 0.5421 43.72 27,080 0.00532 0.01269 0.02937 Alpha- 0.1042 0.3446 26.15 16,310 0.03559 0.06008 0.09958 Cyclodextrin MXDA 0.6324 0.289850.15 31,0000.00276 0.00747 0.01951

4.5.3 Modeling Results

The previously determined reaction orders (a and b) and reaction rate constant terms (bb1 and bb2), shown for each AA scavenging agent in Table 4.39, have been used within the modified AA generation program to simulate the effects of adding AA scavengers to PET resin. As described earlier, this program predicts the amount of AA that will accumulate within PET preforms during a multi-cavity injection molding process. For each AA scavenger, a set of evaluations was conducted to study if the amount of cavities plays a role in how effective the scavengers are in reducing the appearance of AA. Similarly, a second set of experiments was also performed to investigate the effect of melt temperature. The raw data for each of these studies are listed in Appendix L.

Figures 4-52, 4-53, and 4-54 respectively show the predicted AA generation results for various AA scavenger/PET blend concentrations for varying number of cavities within an injection molder. Each evaluation was conducted under a standard set of conditions; the only variable was the number of cavities filled during each simulated run. Within each of these three figures, it can be seen that the number of cavities within an injection molder does not affect the AA sequestering ability of any of these three scavengers.

182

70.00

60.00

50.00

40.00

16 Cavity - Average AA 30.00 16 Cavity - Maximum AA 24 Cavity - Average AA 24 Cavity - Maximum AA 20.00 32 Cavity - Average AA

Percent Decrease in AA (%) in Percent AA Decrease 32 Cavity - Maximum AA 48 Cavity - Average AA 10.00 48 Cavity - Maximum AA

0.00 0 2000 4000 6000 8000 10000 12000 AA Scavenger Concentration (ppm)

Figure 4-52: Predicted injection molding results for various anthranilamide/PET blends and for various manifold designs; modeled at 280oC

80.00

70.00

60.00

50.00

40.00 16 Cavity - Average AA 16 Cavity - Maximum AA 30.00 24 Cavity - Average AA 24 Cavity - Maximum AA

Percent Decrease in AA (%) in Percent AA Decrease 20.00 32 Cavity - Average AA 32 Cavity - Maximum AA 10.00 48 Cavity - Average AA 48 Cavity - Maximum AA

0.00 0 10000 20000 30000 40000 50000 60000 AA Scavenger Concentration (ppm)

Figure 4-53: Predicted injection molding results for various alpha-cyclodextrin/PET blends and for various manifold designs; modeled at 280oC

183

1.40

1.20

1.00

0.80

16 Cavity - Average AA 0.60 16 Cavity - Maximum AA 24 Cavity - Average AA 0.40 24 Cavity - Maximum AA

Percent Decrease in (%) AA 32 Cavity - Average AA 32 Cavity - Maximum AA 0.20 48 Cavity - Average AA 48 Cavity - Maximum AA 0.00 0 2000 4000 6000 8000 10000 12000 AA Scavenger Concentration (ppm)

Figure 4-54: Predicted injection molding results for various MXDA/PET blends and for various manifold designs; modeled at 280oC

The modeling results presented in Figures 4-52 and 4-53 show that the addition of anthranilamide and alpha-cyclodextrin can significantly reduce the detection of AA in

PET preforms. According to Figure 4-54, however, the addition of MXDA to PET may not have any effect at all upon AA concentrations. The results seen in this figure contradict those discussed in Section 4.2; where reductions in both residual AA and the apparent AA generation rates were observed. According to Table 4.12, as little as 100 ppm of MXDA can reduce AA concentrations in PET preforms by 36%.

The reason these two observations do not agree with one another is because the kinetic parameters (a, b, bb1, and bb2) determined for MXDA/PET blends are such that there is very little reaction initially This can be seen by comparing Figure 4-51 with 4-49 and

184

4-50. For any given point in time, the difference between the AA generation curves for the pure PET resin and the 10,000 ppm AA scavenger/PET blends is less for MXDA than for either of the other two scavengers.

To further illustrate this point, Polymath® 6.10 software was used to simulate the generation of AA for times between zero and 60 seconds. Additionally, this modeling was performed at three different temperatures; as shown in Figures 4-55, 4-56, and 4-57.

For each evaluated temperature, it can be seen that the pure PET resin and the 10,000 ppm MXDA/PET blend have very similar AA generation paths for the first 40 seconds.

This is significant because the residence time of the PET blends within the manifold of the modeled injection molder is about 20 seconds.

1.8

1.6 10,000 ppm Anthranilamide 10,000 ppm Alpha-Cyclodextrin 1.4 10,000 ppm MXDA 1.2 No Scavenger 1

0.8 AA (ppm) AA

0.6

0.4

0.2

0 00.20.40.60.81 Time (Minutes)

Figure 4-55: One minute simulated AA generation at 280oC

185

3.5

10,000 ppm Anthranilamide 3 10,000 ppm Alpha-Cyclodextrin 10,000 ppm MXDA 2.5 No Scavenger

AA (ppm) 1.5

0.5

0 0 0.2 0.4 0.6 0.8 1 Time (Minutes)

Figure 4-56: One minute simulated AA generation at 290oC

8 10,000 ppm Anthranilamide 10,000 ppm Alpha-Cyclodextrin 7 10,000 ppm MXDA 6 No Scavenger 5

4 AA (ppm)

0 0 0.2 0.4 0.6 0.8 1 Time (Minutes)

Figure 4-57: One minute simulated AA generation at 300oC

186

To study the effect of melt temperature, predictive AA generation results were obtained by varying the temperature between 270 and 300oC; while the remaining variables were constant throughout this process. The outcomes of this study are shown in Figures 4-58,

4-59, and 4-60. For both anthranilamide and alpha-cyclodextrin, Figures 4-58 and 4-59 show that their addition is most effective at lower melt temperatures. This effect is more pronounced for alpha-cyclodextrin than anthranilamide. The reason for this phenomenon is due to the activation energies (EA) of each AA scavenger. Since alpha-cyclodextrin has the lowest EA (136 kJ/mol), it is able to react more than the other two scavengers at lower temperatures. The EA for anthranilamide is 225 kJ/mol; while it is 258 kJ/mol for

MXDA.

70.00

60.00

50.00

40.00

270 C - Average AA 30.00 270 C - Maximum AA 280 C - Average AA 280 C - Maximum AA Percent Decrease in AA (%) in AA Percent Decrease 20.00 290 C - Average AA 290 C - Maximum AA 300 C - Average AA 10.00 300 C - Maximum AA

0.00 0 2000 4000 6000 8000 10000 12000 AA Scavenger Concentration (ppm)

Figure 4-58: Predicted injection molding results for various anthranilamide/PET blends, studied as a function of temperature; modeled for a 48 cavity process

187

100.00

90.00 270 C - Average AA 80.00 270 C - Maximum AA 280 C - Average AA 70.00 280 C - Maximum AA 290 C - Average AA 60.00 290 C - Maximum AA 300 C - Average AA 50.00 300 C - Maximum AA 40.00

30.00 Percent Decrease in AA (%) in AA Decrease Percent 20.00

10.00

0.00 0 10000 20000 30000 40000 50000 60000 AA Scavenger Concentration (ppm)

Figure 4-59: Predicted injection molding results for various alpha-cyclodextrin/PET blends, studied as a function of temperature; modeled for a 48 cavity process

2.50 270 C - Average AA 270 C - Maximum AA 280 C - Average AA 2.00 280 C - Maximum AA 290 C - Average AA 290 C - Maximum AA 1.50 300 C - Average AA 300 C - Maximum AA

1.00 Percent Decrease in AA (%) in AA Decrease Percent 0.50

0.00 0 2000 4000 6000 8000 10000 12000 AA Scavenger Concentration (ppm)

Figure 4-60: Predicted injection molding results for various MXDA/PET blends, studied as a function of temperature; modeled for a 48 cavity process

188

The results for MXDA, however, oppose those observed for anthranilamide and alpha- cyclodextrin. Figure 4-60 shows that the greatest benefits of MXDA’s addition are seen at higher processing temperatures. The scale, however, is so small that these temperature differences are very small. Again, the predictive modeling of MXDA/PET blends shows little benefit in apparent AA reduction.

While these predicted AA generation results are modeled for a multi-cavity injection molding system, the observed trends can still be compared to the single-cavity residual

AA results obtained in Section 4.2.2.2. Through this comparison, it is apparent that a complete correlation was not achieved. For example, the trend observed in the predicted results indicates that the addition of MXDA to PET will not greatly reduce the concentration of AA. The single-cavity results shown in Table 4.12, however, reveal that the addition of 100 and 200 ppm of MXDA can reduce AA concentrations by 36.1% and

44.6%, respectively. For anthranilamide and alpha-cyclodextrin, both sets of results show that their addition will reduce residual AA content within PET preforms.

There are two suspected reasons as to why these data sets do not completely agree with each other. The first, as previously stated, is that the predicted results are modeled for a multi-cavity injection molder; while the residual AA data presented in Section 4.2.2.2 was obtained from a single-cavity injection molder. While the exact data values may not match each other, there should still be good correlation among the observed trends from each set of experiments.

189

The second reason as to why these data sets do not completely agree with one another is possible error in the approach to model these AA scavengers. In other words, there are potential improvements that can be made to enhance these predicted AA generation modeling results. These include:

1. Better AA scavenger/PET blending procedure to eliminate the varying

thermal histories among samples

2. Additional 60 minute AA generation data sets. Preferably, all AA

generation data sets, for each AA scavenger/PET blend sample at each

evaluated temperature, should be evaluated for 60 minutes and include at

least 12 data points. These larger data sets will be used in the

determination of the a, b, bb1, and bb2 terms. This should eliminate any

possible scattering and therefore obtain the best possible results

Despite the differences between the modeling results and the measured single-cavity results, there are two statements that can be made about this initial attempt toward developing a predictive AA generation model. First, a model was developed to examine the addition of AA scavengers to PET for a multi-cavity injection molding system.

Second, a detailed method was laid out in Section 4.5.2.1 to determine the identified terms (a, b, bb1, and bb2) to describe the interactions that occur upon AA scavenger addition to PET.

190

Chapter 5

Conclusions and Recommendations

5.1 Conclusions

The broad purpose of this work was to comprehensively study the overall effects of melt- blending acetaldehyde (AA) scavengers in poly(ethylene terephthalate) (PET). This included studying the reactions by which these additives scavenge AA, their effectiveness in reducing the concentrations of AA in PET articles, and any changes in the physical properties of PET due to the addition of these scavengers. The material properties that were studied include: thermal properties and stability, material strength and intrinsic viscosity (I.V.), barrier properties, color, and physical appearance.

Through the knowledge obtained from these first three goals, a greater understanding and the overall benefit of adding AA scavengers to PET was achieved. The compilation of this data provided the information needed to determine the most optimal concentrations to melt-blend these AA scavengers into PET resins. The objective was to establish AA scavenger/PET blends that lessened the detection of AA without adversely affecting the desirable properties of the PET.

191

The fifth and final outcome of this project was the creation of an initial, predictive model by which theoretical AA concentrations can be forecasted for various AA scavenger/PET blend systems. This development provides a tool to analyze and compare the amounts of

AA that are generated, for both virgin PET resins and resins melt-blended with AA scavengers, within a multiple cavity injection molder. Comparing these two scenarios yields a method to theoretically quantify the effectiveness of the AA scavengers in reducing the detection of AA for a simulated injection molding operation. This program could also be used to predict the amounts of AA scavengers that are needed to melt-blend with a particular PET resin in order attain the AA concentration requirements of a particular packaging system.

5.1.1 Chemical Mechanisms of AA and AA Scavenger Interactions

The chemical mechanisms by which anthranilamide, meta-xylenediamine (MXDA), and alpha-cyclodextrin sequester AA were spectroscopically studied by 1H NMR and mass spectrometry. Combining data from these two techniques revealed that anthranilamide scavenges AA through a reaction mechanism. In this reaction, the aldehyde group of AA reacts with the amide and amine groups of anthranilamide to produce a two-ring, organic structure. This reaction also generates water as a byproduct.

This described reaction mechanism was proven by both 1H NMR and mass spectra to be the primary AA scavenging mechanism for anthranilamide. There was, however, mass spectroscopy evidence that indicated a further, de-saturation reaction can also occur.

192

Additional experiments were conducted to in an attempt to observe this reaction. While these were unsuccessful, a review of literature revealed that other researchers89-92 have made this observation. To achieve this final product, however, elevated temperatures and a catalyst were needed. It is possible that PET processing temperatures and presence of residual catalysts, within PET resins, could provide the necessary conditions to initiate this reaction and produce this ultimate product.

Similar to anthranilamide, MXDA also scavenges AA by means of a reaction mechanism.

1H NMR analysis provided evidence to confirm this mechanism. In this scavenging reaction, the aldehyde group of AA reacts with a primary amine of MXDA to create an imine group and water, as a byproduct. The fact that MXDA possesses two primary amine groups allows for up to two molecules of AA to react with each MXDA molecule.

If this occurs, the result is the generation of two imine groups and two molecules of water.

The generation of imine groups, from a similar AA scavenging reaction between MXD6 and AA, was proven by Bandi, et al48 to lead to the formation of a yellowish color when melt-blending polyamides with PET. The imine formation from the MXDA and AA reaction also lead to the creation of color. Upon mixing a solution of dissolved MXDA with a solution of dissolved AA a solid, orange product was formed. Through further dilution, the color was altered to a dark yellow, slight greenish tint. This final appearance was similar to that of the 1 weight % MXDA/PET blend sample that was produced through twin-screw extrusion.

193

The AA scavenging mechanism for alpha-cyclodextrin was determined through an NMR titration experiment. For this work, several samples of varying alpha-cyclodextrin and

AA concentrations were prepared and analyzed by 1H NMR. Even though the mixing of these two solutions did not result in any chemical reactions, changes still occurred in the

1H NMR spectra. For each respective sample, the position changes of the two proton groups of AA were monitored to observe these changes.

The trends of this plot show that as the concentration of alpha-cyclodextrin increases, the chemical shift of AA’s protons, and the equilibrium product it forms with deuterium oxide (D2O), increase until a saturation point is reached. The saturation point for the alpha-cyclodextrin and AA complex was found to be reached at a one to one ratio. This means that each molecule of alpha-cyclodextrin can only scavenge one molecule of AA.

These results confirm that alpha-cyclodextrin sequesters AA through a hydrogen bonding/size-enclosure mechanism; which has also been reported for other host/guest complexes.61, 71-74, 93

5.1.2 Effectiveness of AA Scavengers’ Apparent Reduction in Generated

The ability of anthranilamide, MXDA, and alpha-cyclodextrin to reduce AA in PET was evaluated by two gas chromatography techniques. One technique quantifies the apparent rate of AA generation and the other technique determines the concentration of AA that remains residually trapped in the polymer’s matrix. The results from both of these

194

techniques indicated that as the concentration of the AA scavenger increases, the amount of detectable AA decreases.

Both tables and plots were presented to confirm the aforementioned trend that states increasing AA scavenger content results in the decreasing of apparent AA generation rates for PET blend samples. Within these data sets there are irregularities that deviate from this generalized trend. These have been shown to be the result of the sample blending process. The thermal history of each sample was discussed and in some cases it was shown that portions of these samples to have had up to three thermal histories.

One common feature among these three plots is that eventually the slope becomes zero.

This means that there is a point at which the AA generation rate becomes independent of

AA scavenger concentration. For anthranilamide and MXDA, this appears to happen around 1200 ppm; for alpha-cyclodextrin the slope appears to flatten in the region of

10,000 ppm, or 1 weight %. The difference between these values has to do with the molecular structure and interaction mechanisms of these scavenging agents.

On the surface, a relative comparison among these three scavengers reveals that anthranilamide is the most efficient at reducing the apparent rate of AA generation. It was shown that, generally, at the 1200 and 10,000 ppm addition levels, by weight, the anthranilamide/PET blend samples have lower AA generation rates than the alpha- cyclodextrin or MXDA samples. Anthranilamide (136.15 g/mol)86, however, is about one-seventh the weight of alpha-cyclodextrin (972.402 g/mol)85. Since the concentration

195

of AA scavengers in PET is based upon weight, and not on the number of moles, anthranilamide should therefore be about seven times as effective as alpha-cyclodextrin.

This, however, is not the case.

When the molecular weight and the functionality of the scavengers are factored in, alpha- cyclodextrin is actually the most efficient at reducing the generation of AA. This is based on the fact that for an equivalent ppm concentration, the number of moles of alpha- cyclodextrin is one-seventh that of anthranilamide or MXDA. Again, anthranilamide and/or MXDA were not shown to be seven times more efficient than alpha-cyclodextrin at scavenging AA. Since both MXDA (136.2 g/mol)47 and anthranilamide (136.15 g/mol) have approximately the same molecular weight, their chemical structures or functionality determine the next most efficient scavenger. It was shown that MXDA has two primary amines for AA can react with; making it di-functional. Anthranilamide, however, can only react with one molecule of AA. This implies that MXDA should be twice as effective as anthranilamide; which is not what was observed. This indicates that anthranilamide is the second most efficient at sequestering AA, followed by MXDA.

The trend of decreasing residual AA content, with increasing scavenger concentration, was shown for both pelletized and injection molded preform samples. As with the AA generation rate data, anomalies to this general trend exist and are the result of the sample blending process. According to the data for pelletized samples, it would again appear that anthranilamide is the most successful at reducing the residual AA content. When mole percentage is used in replacement of weight percentage, it is again clear that alpha-

196

cyclodextrin is the most proficient of the three scavengers at reducing AA; anthranilamide is second and MXDA is last among these three additives.

5.1.3 Physical Properties of AA Scavenger/PET Blend Samples

Beyond just the reductions in AA, the addition of anthranilamide, MXDA, and alpha- cyclodextrin were studied to determine the overall effects upon the physical properties and appearance of PET. The properties that were studied include: intrinsic viscosity

(I.V.), color, thermal properties, and oxygen permeation. Determining any changes in these properties is critical to understanding the overall benefit of adding these AA scavenging agents to PET.

The I.V. data for the pelletized PET blend samples indicates that as the AA scavenger is increased, the I.V. decreases. AA scavenger addition, however, is not the only reason for viscosity reductions. As previously mentioned, these AA scavenger/PET blend samples are composed of portions that have been extruded once, twice, and up to three times.

Examination of the control samples show that a sample’s thermal history plays a role in its I.V. reduction.

The roles of thermal histories and AA scavenger addition on I.V. reduction may be better illustrated by the PET preform I.V. data. These injection molded preform samples do not have the varying thermal histories that the pelletized samples possess. Examination of the anthranilamide and MXDA samples indicate that the minimal addition of these two

197

scavengers do not further reduce I.V.; relative to the control sample. For alpha- cyclodextrin, however, it appears that the addition of at least 500 ppm can affect the I.V. of the resulting PET blend sample.

The melt-blending of alpha-cyclodextrin into PET resin may not only affect the material’s I.V., it can also affect its color. All of the alpha-cyclodextrin/PET blend samples possessed a brownish tint that, when analytically measured, altered their L, a, and b values and yielded a false yellowness index for each sample. The brown appearance of the 500 and 1200 ppm alpha-cyclodextrin/PET blend samples can be seen in 2-liter bottle samples that were manufactured. The color is especially noticeable in the neck of the bottles, where more material is present.

The color analysis data for the anthranilamide and MXDA samples appear to indicate that as scavenger concentration increases, the b value and yellowness index also increase.

The impact on color, however, cannot be attributed to AA scavenger addition alone. The color analysis data shows that increasing thermal histories will affect these values. While their b values are higher, the yellowness indexes at the 100 and 200 ppm level for both anthranilamide and MXDA are lower than that of the PET control sample; which was only extruded once. This is further illustrated by the 2-liter bottle samples, as the appearance of the 100 and 200 ppm anthranilamide or MXDA bottles are indistinguishable from the appearance of the pure PET 2-liter bottle.

198

It was shown that the addition of these three scavenging agents did not impact the glass transition temperature (Tg) or melting behavior of PET. The only suspected influence that these additives have on the thermal properties of PET is that they appear to act as nucleating agents, making crystallization easier. Finally, it was shown that the addition of anthranilamide, alpha-cyclodextrin, or MXDA does not alter the oxygen permeability of PET.

5.1.4 Optimal AA Scavenger/PET Blends

The optimal amounts of AA scavengers to melt-blend into PET were determined through the evaluation of the data presented in Section 4.2 and 4.3. This involved balancing the reduction of detectable AA with any negative effects that result due to the addition of the

AA scavengers. Assessment of the various AA scavenger/PET blends were based on:

AA generation rates, residual amount of AA, color, thermal properties, intrinsic viscosities (I.V.), and oxygen permeation.

Results showed that increasing the AA scavenger concentration decreases both the apparent AA generation rates and residual amount of AA with PET. The data points that lay outside of this general trend have been shown to be the result of the sample blending process. This methodology caused portions of these samples to have up to three thermal histories.

199

Although results indicate that the addition of scavenging agents to PET do not greatly affect the thermal properties or oxygen permeation, it was found that I.V. and color can be impacted. It was shown that the addition of up to 200 ppm of anthranilamide or

MXDA does not further reduce the I.V. during injection molding. For alpha-cyclodextrin, however, even as little as 500 ppm reduced the I.V. by an additional 0.03 dL/g; compared to the control sample.

Color analysis showed that beyond 200 ppm of either anthranilamide or MXDA the yellowness index of the PET blend sample becomes an issue. The addition of alpha- cyclodextrin has an even greater impact. All of the alpha-cyclodextrin/PET blend samples had enough of a brownish tint that they were not accurately measurable. This is further illustrated by the appearance of the 2-liter bottle samples. While the brownish appearance of the 500 ppm bottle is less than that of the 1200 ppm bottle, the appearance of the 500 ppm bottle would still not be acceptable for a commercial application. The appearance of the 100 and 200 ppm anthranilamide/PET, 100 and 200 ppm MXDA/PET, and the pure PET 2-liter bottles are indistinguishable to one another.

The data presented in this work suggests that the generation of color and decrease in I.V. are factors that limit the addition of AA scavengers to PET. For these reasons, it is suggested that the most optimal addition amounts for both anthranilamide and MXDA are

200 ppm or less. When adding alpha-cyclodextrin to PET, its concentration should be no more than 500 ppm.

200

5.1.5 Modeling

An existing modeling program, which simulates the AA generation that occurs during multi-cavity injection molding, was modified to account for the addition of AA scavenging agents to PET. To quantify the interactions between the scavengers and AA, a new modeling equation, Equation 17, was developed to describe the appearance of AA over time. This new, governing equation utilizes four kinetic terms (a, b, bb1, and bb2) to describe the AA and scavenger interactions. This equation contains two reaction orders, a and b, and a reaction rate constant, k1. The reaction rate constant, a function of temperature, is further broken down into an Arrhenius function and described by the terms bb1 and bb2. The addition of these terms to the original modeling program allow for multi-cavity injection molding simulations, describing the accumulation of AA within

PET preforms, to be performed.

The methodologies to quantify the a, b, bb1, and bb2 terms for each AA scavenging agent was also described. These values were determined through a combination of graphical and numerical analysis techniques. The final a, b, bb1, and bb2 values have been listed for each of the evaluated AA scavengers.

Using these terms, modeling simulations were conducted to analyze the effects that melt temperature and number of cavities, within an injection molding set-up, have upon the scavengers’ ability to reduce detectable AA concentrations. It was determined that the number of cavities has no impact upon the effectiveness of the AA scavengers. It was

201

shown, however, that temperature does play a role in the scavenger’s capability to reduce

AA within PET preforms.

The results from the multi-cavity modeling were then compared to the residual AA data obtained from a single-cavity injection molding system. Since this predictive program is set-up to model a different machine, the actual values were not expected to match; only the observed trends. Both sets of experiments showed that anthranilamide and alpha- cyclodextrin are effective in reducing AA concentrations within PET preforms; the greater the scavenger concentration, the greater the impact.

For MXDA, however, the modeling results did not correlate with the observed trends from the single-cavity injection molding results. The residual AA results, obtained from the PET preforms manufactured through single-cavity injection molding, indicate that

MXDA is an effective AA scavenger; while the modeling results show only minor changes in AA concentrations as a result of its addition. This result points toward the conclusion that error may have occurred during the establishment of the four kinetic terms to describe the interaction between MXDA and AA. It is suspected that this is the result of the initial sample blending process and/or the establishment of only one set of 60 minute AA generation data to model against. Both of these limitations are recommendations to further improve upon this initial attempt at modeling the AA scavenger/PET blend systems for multi-cavity injection molding systems.

202

5.2 Recommendations

While this work provided a broad investigation toward understanding the overall effects of melt-blending AA scavengers into PET, further work should be conducted to expand upon this information and improve the fundamental understanding of AA sequestering systems for PET. For this, six recommendations are proposed:

1. One of the identified limitations of this work was the sample blending method

used to achieve the various AA scavenger/PET blend concentrations. Instead of

diluting a master-batch AA scavenger/PET blend sample, blends should be

prepared by directly melt-blending AA scavengers with PET resin to yield each

of the final, desired concentrations. This process would produce blends that all

have one thermal history and that can be compared directly to one another.

2. The scavengers that were studied included two which are patented

(anthranilamide and alpha-cyclodextrin) and one which is a monomer (meta-

xylenediamine, also known as MXDA) for a known AA scavenger, MXD6.

From another perspective, anthranilamide and MXDA are additives which

scavenge AA by reacting with the compound. Alpha-cyclodextrin interacts with

AA by a size-enclosing mechanism. Further AA scavengers, beyond these three,

should be evaluated. Of particular importance should be the scavengers that

sequester AA by mechanisms other than the ones studied in this work. These

203

include such scavengers that have been identified in this work, including

oxidation catalysts and hydrogenation catalysts.

3. The version of cyclodextrin that was examined in this study was alpha-

cyclodextrin; the same is true for the work by Suloff.24 It was mentioned that

three forms of cyclodextrin exist: alpha, beta, and gamma. The reason the

alpha version was chosen is because it possesses the smallest internal cavity and

should therefore provide the strongest attraction for AA to be housed in its

structure. Studying the AA scavenging abilities of beta- and gamma-

cyclodextrin would provide information such as the relative sequestering

strengths among this family of chemical compounds and evidence to see if

higher interaction ratios can be achieved by increasing the size of the

scavenging agent’s internal cavity. It could be important to determine if it is

possible for one molecule of beta- or gamma-cyclodextrin to sequester two or

more molecules of AA within either of their internal structures, at any given

point in time.

4. One of the effects observed when alpha-cyclodextrin was melt-blended into

PET was the generation of a brownish color. The appearance of this color could

be a limitation toward its use as an AA scavenger within commercial PET

packaging applications. The majority of PET packages in the marketplace are

of a clean, colorless appearance. Further work should be performed with the

intent of eliminating or managing this color formation. This could be done with

204

compatibilizers, alternate molecular structures to either increase the compound’s

melt/degradation temperature or limit the generation of color, other cyclodextrin

versions (i.e. beta or gamma), etc.

5. The results presented in Section 4.5 are only theoretical and have not been

verified by any experimental data. Some or all of the discussed AA

scavenger/PET blends should be processed into preforms by means of an

injection molding system similar to the one previously described in this work.

The residual AA content of these preforms should then be evaluated to

determine any disparities between these theoretical and experimental results.

6. The model discussed in Section 4.5 was developed to be an initial attempt at

predicting the effectiveness of AA scavengers in reducing the detectable AA

concentrations within processed PET articles. Recommendation five discussed

the experimental verification of the theoretical results obtained from this work.

This suggestion, however, is the evaluation of the approach discussed within

this work. The goal would be to improve upon the initial model and

consequently advance the accuracy of its results relative to experimental data.

205

References

1. Jabarin, S. A., PET Technology Course. University of Toledo: 2006.

2. Kenplas Industry Limited. PET Preform/Bottle Project.

http://www.kenplas.com/project/pet/ (January, 28 2005)

3. Jabarin, S. A., Poly(ethylene terephthalate). In Polymeric Materials Encyclopedia,

Salamone, J. C., Ed. CRC Press: 1996; Vol. 8, pp 6078-6123.

4. Rieckmann, T.; Volker, S., Poly(ethylene terephthalate) Polymerization -

Mechanism, Catalysis, Kinetics, Mass Transfer and Reactor Design. In Modern

Polyesters: Chemistry and Technology of Polyesters and Copolyesters, Scheirs,

J.; Long, T. E., Eds. John Wiley & Sons, Ltd.: 2003; pp 31 - 115.

5. Kezios, P. S.; Codd, H.; Harrison, K. R. Co-polyester Packaging Resins Prepared

without Solid-State Polymerization, a Method for Processing the Co-Polyester

Resins with Reduced Viscosity Change and Containers and Other Articles

Prepared by the Process. US Patent Application 20070248778, 2007.

206

6. DeBruin, B. Polyester Process Using a Pipe Reactor. U.S. Patent 6906164, 2005.

7. DeBruin, B. R. Polyester Process Using a Pipe Reactor. U.S. Patent 6861494,

2005.

8. DeBruin, B. R. Polyester Process Using a Pipe Reactor. U.S. Patent 7345139,

2008.

9. DeBruin, B. R.; Bonner, R. G. Polyester Process Using a Pipe Reactor. U.S.

Patent 7135541, 2006.

10. DeBruin, B. R.; Martin, D. L. Polyester Process Using a Pipe Reactor. U.S. Patent

7074879, 2006.

11. Nelson, G. W.; Cherry, C.; Olsen, E. G. Process for Producing Polyester Articles

Having Low Acetaldehyde Content. U.S. Patent 5597891, 1997.

12. Nelson, G. W.; Nicely, V. A. Process for Producing Polyester Articles Having

Low Acetaldehyde Content. U.S. Patent 5648032, 1997.

13. Al-AbdulRazzak, S.; Jarbarin, S. A., Processing Characteristics of Poly(ethylene

terephthalate): Hydrolytic and Thermal Degradation. Polymer International 2002,

51, 164-173.

207

14. Jabarin, S. A.; Lofgren, E. A., Thermal Stability of Polyethylene Terephthalate.

Polymer Engineering and Science 1984, 24, (13), 1056 - 1063.

15. Shukla, S. R.; Lofgren, E. A.; Jabarin, S. A., Effects of Injection-molding

Processing Parameters on Acetaldehyde Generation and Degradation of

Poly(ethylene terephthalate). Polymer International 2005, 54, 946-955.

16. Yoda, K.; Tsuboi, A.; Wada, M.; Yamadera, R., Network Formation in

Poly(ethylene terephthalate) by Thermooxidative Degradation. Journal of Applied

Polymer Science 1970, 14, (9), 2357 - 2376.

17. Goodings, E. P., Thermal Degradation of Poly(ethylene terephthalate). Society of

Chemical Industry (London) 1961, Monograph 13, 211.

18. Delaware Health and Social Services Division of Public Health. Acetaldehyde.

http://www.dhss.delaware.gov/dph/files/acetaldehydefaq.pdf (November 12,

2009)

19. Material Safety Data Sheet - Acetaldehyde, 99.5%. Fisher Scientific. 2005.

20. Controlling AA in Mineral Water Bottles. PET Planet 2003, (1).

208

21. van Aardt, M.; Duncan, S. E.; Bourne, D.; Marcy, J. E.; Long, T. E.; Hackney, C.

R.; Heisey, C., Flavor Threshold for Acetaldehyde in Milk, Chocolate Milk, and

Spring Water Using Solid Phase Microextraction Gas Chromatography for

Quantification. Journal of Agricultural and Food Chemistry 2001, 49, 1377 -

1381.

22. Khemani, K. C., A Novel Approach for Studying the Thermal Degradation, and

for Estimating the Rate of Acetaldehyde Generation by the Chain Scission

Mechanism in Ethylene Glycol Based Polyesters and Copolyesters. Polymer

Degradation and Stability 2000, 67, 91-99.

23. Kim, T. Y.; Jabarin, S. A., Solid-State Polymerization of Poly(ethylene

terephthalate). III. Thermal Stabilities in Terms of the Vinyl Ester End Group

and Acetaldehyde. Journal of Applied Polymer Science 2003, 89, 228-237.

24. Suloff, E. C. Sorption Behavior of an Aliphatic Series of Aldehydes in the

Presence of Poly(ethylene terephthalate) Blends Containing Aldehyde

Scavenging Agents. Virginia Polytechnic Institute and State University,

Blacksburg, 2002.

25. Marshall, I.; Todd, A., The Thermal Degradation of Polyethlyene Terephthalate.

Journal of the Chemical Society, Faraday Transactions 1953, 49, 67.

209

26. Ritchie, P. D., High Temperature Resistance and Thermal Degradation of

Polymers. Society of Chemical Industry (London) 1961, Monograph 13, 107.

27. Zimmermann, H., Degradation and Stabilization of Polyesters. In Developments

in Polymer Degradation, Grassie, N., Ed. Applied Science: London, 1984; Vol. 5,

p 79.

28. Kao, C.-Y.; Cheng, W.-H.; Wan, B.-Z., Investigation of Catalytic Glycolysis of

Polyethylene Terephthalate by Differential Scanning Calorimetry.

Thermochimica Acta 1997, 292, (1-2), 95-104.

29. Zimmermann, H.; Kim, N. T., Investigations of Thermal and Hydrolytic

Degradation of Poly(ethylene terephthalate). Polymer Engineering and Science

1980, 20, 680.

30. Granzow, A.; Ferrillo, R. G.; Wilson, A., The Effect of Elemental Red

Phosphorus on the Thermal Degradation of Poly(ethylene terephthalate). Journal

of Applied Polymer Science 1997, 21, (6), 1687-1697.

31. Bashir, Z.; Al-Uraini, A.-A.; Jamjoom, M.; Al-Khalid, A.; Al-Hafez, M.; Ali, S.,

Acetaldehyde Generation in Poly(ethylene terephthalate) Resins for Water Bottles.

Journal of Macromolecular Science, Part A - Pure and Applied Chemistry 2002,

A39, (12), 1407-1433.

210

32. Long, T. E.; Bagrodia, S.; Moreau, A.; Duccase, V. Process for Improving the

Flavor Retaining Property of Polyester/Polyamide Blend Containers for Ozonated

Water. U.S. Patent 6042908, 2000.

33. Igarashi, R.; Watanabe, Y.; Hirata, S. Thermoplastic Polyester Composition

Having Improved Flavor-Retaining Property and Vessel Formed Therefrom. U.S.

Patent 4837115, 1989.

34. Nelson, G. W.; Nicely, V. A.; Turner, S. R. Process for Producing Polyester PET

Articles with Low Acetaldehyde. U.S. Patent 6099778, 2000.

35. Bell, E. T.; Bradley, J. R.; Long, T. E.; Stafford, S. L. Polyester/Polyamide

Blends with Improved Color. U.S. Patent 6239233, 2001.

36. Andrews, S. M.; Odorisio, P. A.; Simon, D. Polyester Compositions of Low

Residual Aldehyde Content. U.S. Patent 6790499, 2004.

37. Mills, D. E.; Stafford, S. L. Copolyester/Polyamide Blend Having Improved

Flavor Retaining Property and Clarity. U.S. Patent 5266413, 1993.

38. Mills, D. E.; Stafford, S. Copolyester/Polyamide Blend Having Improved Flavor

Retaining Property and Clarity. U.S. Patent 5258233, 1993.

211

39. Mills, D. E.; Stafford, S. L. Polyamide Concentrate Useful for Producing Blends

Having Improved Flavor Retaining Property and Clarity. U.S. Patent 5340884,

1994.

40. Schumann, H.-D.; Thiele, U. Process for Direct Production of Low Acetaldehyde

Packaging Material. U.S. Patent 5656221, 1995.

41. Dalton, J. S. N.; Stafford, S. L.; Hewa, J. D. Polyester and Optical Brightener

Blend Having Improved Properties. U.S. Patent 5985389, 1998.

42. Odorisio, P. A.; Andrews, S. M.; Lazzari, D.; Simon, D.; King, R. E.; Stamp, M.;

Reinicker, R.; Tinkl, M.; Berthelon, N.; Muller, D.; Hirt, U. Polyester and

Polyamide Compositions of Low Residual Aldehyde Content. U.S. Patent

6908650, 2005.

43. Odorisio, P. A.; Andrews, S. M.; Lazzari, D.; Simon, D.; King, R. E.; Stamp, M.;

Reinicker, R.; Tinkl, M.; Berthelon, N.; Muller, D.; Hirt, U. Polyester and

Polyamide Compositions of Low Residual Aldehyde Content. U.S. Patent

7022390, 2006.

44. Turner, S. R.; Nicely, V. A. Polyester/Polyesteramide Blends. WO Patent

9728218, 1997.

212

45. Usilton, J. J.; Andrews, S. M. Polyester Compositions of Low Residual Aldehyde

Content. WO Patent 0123475, 2001.

46. Bell, E. T.; Long, T. E.; Stafford, S. L. Naphthalenedicarboxylate Containing

Polyester/Polyamide Blend Having Improved Flavor Retaining Property. WO

Patent 9839388, 1998.

47. Mitsubishi Gas Chemical Company, Inc. Nylon-MXD6.

http://www.gasbarriertechnologies.com/mxd6.html (March 30, 2007)

48. Bandi, S.; Mehta, S.; Schiraldi, D. A., The Mechanism of Color Generation in

Poly(ethylene terephthalate)/Polyamide Blends. Polymer Degradation and

Stability 2005, 88, 341-348.

49. Ferrari, G.; Sisson, E. A.; Knudsen, R. Compartmentalized Resin Pellets. U.S.

Patent 7550203, 2009.

50. Rule, M.; Shi, Y.; Huang, X. Method to Decrease the Acetaldehyde Content of

Melt-Processed Polyesters. U.S. Patent 6274212, 2001.

51. Speer, D. V.; Morgan, C. R.; Blinka, T. A.; Cotterman, R. L. By-Product

Absorbers for Oxygen Scavenging Systems. U.S. Patent 5942297, 1999.

213

52. Ching, T. Y.; Goodrich, J. L.; Katsumoto, K. Oxygen Scavenging System

Including a By-Product Neutralizing Material. U.S. Patent 6057013, 2000.

53. Brodie, V.; Visioli, D. L. Aldehyde Scavenging Compositions and Methods

Relating Thereto. U.S. Patent 5362784, 1994.

54. Brodie, V.; Visioli, D. Aldehyde Scavenging Compositions and Methods Relating

Thereto. U.S. Patent 5284892, 1994.

55. Brodie, V.; Visioli, D. L. Aldehyde Scavenging Compositions and Methods

Relating Thereto. U.S. Patent 5413827, 1995.

56. McNeely, G. W.; Woodward, A. J. Modified Polyethylene Terephthalate. U.S.

Patent 5250333, 1993.

57. Al-Malaika, S. Thermoplastic Molding Compositions and Polymer Additives.

WO Patent 0066659, 2000.

58. Al-Malaika, S. Thermoplastic Molding Compositions and Polymer Additives.

U.S. Patent 6936204, 2005.

59. Eckert, R.; Sela, M.; Munjal, S.; Nagel, M.; Voerckel, V.; Wiegner, J.-P.

Extruded Products from Polyethylene Terephthalate with Reduced Acetaldehyde

Content and Process of Their Production. WO Patent 0100724, 2001.

214

60. Odorisio, P. A.; Andrews, S. M. Polyester Compositions of Low Residual

Aldehyde Content. U.S. Patent 7138457, 2006.

61. Rekharsky, M. V.; Inoue, Y., Complexation Thermodynamics of Cyclodextrins.

Chemical Reviews 1998, 98, (5), 1875-1917.

62. Wood, W. In Plastic "Off-Taste" Control in Bottled Water Closure Compositions,

IFT Annual Meeting, 2003; 2003.

63. Wood, W. E.; Beaverson, N. J. Barrier Material Comprising a Thermoplastic and

a Compatible Cyclodextrin Derivative. U.S. Patent 5505969, 1996.

64. Wood, W. E.; Beaverson, N. Rigid Polymeric Beverage Bottles with Improved

Resistance to Permeate Elution. U.S. Patent 5837339, 2000.

65. Wood, W. E.; Beaverson, N. J. Rigid Polymeric Beverage Bottles with Improved

Resistance to Permeate Elution. U.S. Patent 6136354, 2000.

66. Wood, W. E.; Beaverson, N. J.; Lawonn, P.; Huang, X. Reducing Concentration

of Organic Materials with Substituted Cyclodextrin Compound in Polyester

Packaging Materials. U.S. Patent 6709746, 2004.

215

67. Wood, W. E.; Beaverson, N. J.; Lawonn, P.; Huang, X. Reducing Concentration

of Organic Materials with Substituted Cyclodextrin Compound in Polyester

Packaging Materials. U.S. Patent 6878457, 2005.

68. Wood, W. E.; Beaverson, N. J.; Lawonn, P.; Huang, X. Reducing Concentration

of Organic Materials with Substituted Cyclodextrin Compound in Polyester

Packaging Materials. U.S. Patent 7018712, 2006.

69. Wood, W. E.; Beaverson, N. J.; Lawonn, P. A.; Huang, X. Reducing

Concentration of Organic Materials with Substituted Cyclodextrin Compound in

Polyester Packaging Materials. U.S. Patent 6974603, 2005.

70. Wood, W. E.; Beaverson, N. J.; Seethamraju, K. V. Grafted Cyclodextrin. U.S.

Patent 7166671, 2007.

71. Hao, Y.-Q.; Wu, Y.; Liu, J.; Luo, G.; Yang, G., 1H NMR Study on the Inclusion

Complex of Glutathione with a Glutathione Peroxidase Mimic, 2,2'-ditelluro-

bridged Beta-cyclodextrins. Journal of Inclusion Phenomena and Macrocyclic

Chemistry 2006, 54, 171-175.

72. Lantz, A. W.; Rodriguez, M. A.; Wetterer, S. M.; Armstrong, D. W., Estimation

of Association Constants Between Oral Malodor Components and Various Native

and Derivatized Cyclodextrins. Analytica Chimica Acta 2005, 557, 184-190.

216

73. Inokuma, S.; Kimura, K.; Funaki, T.; Nishimura, J., Synthesis and Complexing

Ability of Azacrownophanes: The Cyclodextrin Catalysis of the Photochemical

Cyclization Reaction. Journal of Inclusion Phenomena and Macrocyclic

Chemistry 2001, 39, 35-40.

74. Tan, Z. J.; Zhu, X. X.; Brown, G. R., Formation of Inclusion Complexes of

Cyclodextrins with Bile Salt Anions As Determined by NMR Titration Studies.

Langmuir 1994, 10, (4), 1034-1039.

75. Bell, R. G. What are Zeolites? http://www.bza.org/zeolites.html (January 30,

2010)

76. Massey, F. L.; Callander, D. D. Color Improvement and Acetaldehyde Reduction

in High Molecular Weight Polyesters. U.S. Patent 4391971, 1983.

77. Mills, D. E.; Stafford, S. L.; Carico, J. Polyester/Zeolite Admixtures. WO Patent

Application 1994/029378 A1, 1994.

78. Go, S. W.; Burzynski, D. J. Polyester Stabilization and Composition. U.S. Patent

4357461, 1982.

79. Rule, M. Process for Reduction of Acetaldehyde and Oxygen in Beverages

Contained in Polyester-Based Packaging. U.S. Patent 6569479, 2002.

217

80. Rule, M. Process for Reduction of Acetaldehyde and Oxygen in Beverages

Contained in Polyester-Based Packaging. WO Patent 0130900, 2001.

81. Rule, M.; Shi, Y. Polyester Composition and Articles with Reduced Acetaldehyde

Content and Method Using Hydrogenation Catalyst. U.S. Patent 7041350, 2006.

82. Fox, M. A.; Whitesell, J. K., Organic Chemistry. Second ed.; Jones and Bartlett

Publlishers: 1997.

83. Silverstein, R. M.; Webster, F. X.; Kiemle, D. J., Spectrometric Identification of

Organic Compounds. Seventh ed.; John Wiley & Sons, Inc.: 2005.

84. Material Safety Data Sheet - MXDA. Mitsubishi Gas Chemical Company, Inc.

2005.

85. Material Safety Data Sheet - Alpha-Cyclodextrin. Fisher Scientific. 2005.

86. Material Safety Data Sheet - Anthranilamide, 99+%. Fisher Scientific. 2005.

87. ASTM Method D 1925 - Test Method for Yellowness Index of Plastics; 1988.

88. ASTM Method D 3985 - Standard Test Method for Oxygen Gas Transmission

Rate Through Film and Sheeting Using a Coulometric Sensor; 1981.

218

89. Abdel-Jalil, R. J.; Voelter, W.; Saeed, M., A novel method for the synthesis of

4(3H)-quinazolinones. Tetrahedron Letters 2004, 45, 3475-3476.

90. Hanumanthu, P.; Seshavatharam, S. K. V.; Ratnam, C. V.; Subba Rao, N. V.,

Studies in the Formation of Heterocyclic Rings Containing Nitrogen: Part XXIII.

Condensation of o-aminobenzamide with aldehydes and Schiff bases.

Proceedings of the Indian Academy of Sciences 1976, 84, (2), 57-63.

91. Suh, M. E., Synthesis of Quinazoline-4-one Derivatives from 2-Aminobenzamide.

Reaction with Aldehyde and Ketone. Journal of the Pharmaceutical Society of

Korea 1985, 29, (2), 103-106.

92. Smith, T. A. K.; Stephen, H., Synthesis in the Quinazolone Series - VI. The

Synthesis of 1:2:3:4-Tetrahydro-2-aryl-4-oxoquinaolines. Tetrahedron Letters

1957, 1, 38-44.

93. Chen, S.; Teng, Q.; Wu, S., Theoretical Studies on the Binding Affinities of B-

Cyclodextrin to Small Molecules and Monosaccharides. Central European

Journal of Chemistry 2006, 4, (2), 223-233.

94. Fogler, H. S., Elements of Chemical Reaction Engineering. Third ed.; Prentice

Hall International Series in the Physical and Chemical Engineering Sciences:

1999.

219

Appendix A

1H NMR Spectra of AA and Alpha-Cyclodextrin Titration Experiment

220

221

222

223

224

225

226

Table A.1: Location of the AA and Alpha-Cyclodextrin Protons for each of the AA and Alpha-Cyclodextrin NMR Titration Experiments Chemical Proton Various AA to Alpha-Cyclodextrin Ratios Compound Group 0.000 0.211 0.426 0.613 0.800 1.006 2.016 3.026 Proton #1 5.071 5.070 5.072 5.072 5.073 5.073 5.074 5.073 Proton #2 3.653 3.653 3.653 3.654 3.655 3.655 3.656 3.656 Alpha- Proton #3 3.999 3.997 3.999 3.998 3.997 3.997 3.996 3.994 Cyclodextrin Proton #4 3.602 3.602 3.603 3.603 3.604 3.604 3.606 3.606 Proton #5 3.869 3.867 3.868 3.869 3.869 3.868 3.868 3.868 Proton #6 3.921 3.920 3.921 3.922 3.923 3.922 3.923 3.922 Methyl 2.249 2.264 2.282 2.291 2.296 2.310 2.312 2.311 AA Aldehyde 9.685 9.699 9.719 9.724 9.738 9.744 9.747 9.748 AA’s Methyl 1.337 1.343 1.344 1.346 1.349 1.351 1.351 1.350 Equilibrium Product in Aldehyde 5.254 5.259 5.260 5.262 5.264 5.266 5.266 5.267 D2O

Table A.2: Change in Location of the Protons Representing AA and its D2O Equilibrium Product, Due to the Presence of Alpha-Cyclodextrin Chemical Proton Various AA to Alpha-Cyclodextrin Ratios Compound Group 0.000 0.211 0.426 0.613 0.800 1.006 2.016 3.026 Methyl 0.000 0.015 0.033 0.042 0.048 0.061 0.063 0.062 AA Aldehyde 0.000 0.014 0.034 0.039 0.053 0.059 0.062 0.063 AA’s Equilibrium Methyl 0.000 0.006 0.007 0.009 0.012 0.014 0.014 0.013 Product in Aldehyde 0.000 0.005 0.006 0.008 0.010 0.012 0.012 0.013 D2O

227

Appendix B

Raw Data from AA Generation Experiments

Table B.1: AA generation data for the Voridian CB12 PET resin Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.03590 10021.33 5.3 9 0.03565 10637.06 5.6 5.3 0.02995 7912.75 5.0 0.03640 11327.21 5.9 280 13 0.04045 15090.50 7.1 6.5 0.03520 11976.80 6.4 0.03945 19272.92 9.2 17 0.04375 21695.30 9.4 9.2 0.04060 19546.90 9.1 0.03940 11647.28 5.6 9 0.04250 11753.70 5.2 5.4 0.04330 12313.72 5.4 0.04390 42196.15 18.2 15 290 0.03935 38641.15 52912 18.6 18.8 0.04180 43367.87 19.6 0.03695 45658.21 23.4 17 0.03860 51691.67 25.3 24.6 0.03890 51522.70 25.0 0.03765 22114.05 11.1 9 0.04150 25685.65 11.7 11.6 0.03765 23797.18 11.9 0.04220 64826.03 29.0 13 300 0.03915 53881.40 26.0 27.4 0.04365 62594.79 27.1 0.03305 118835.97 68.0 0.03790 125983.47 62.8 17 62.6 0.03595 112324.94 59.1 0.04385 140493.70 60.6

228

Table B.2: AA generation data for the “one-time” processed PET sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.03675 21561.67 11.1 9 0.05290 26224.16 9.4 10.3 0.03320 18188.71 10.4 0.03660 32478.93 16.8 280 13 0.02845 23381.53 15.5 16.0 0.03210 26570.23 15.6 0.02765 31566.38 21.6 17 0.02845 33645.70 22.4 22.8 0.03160 41023.85 24.5 0.03480 30106.83 16.4 9 0.02995 25659.67 16.2 16.1 0.03080 25741.07 15.8 0.03355 39748.23 52912 22.4 290 13 0.02935 34547.64 22.2 23.1 0.04185 54840.06 24.8 0.02480 51687.63 39.4 17 0.02525 53062.66 39.7 39.6 0.02720 57303.06 39.8 0.02750 35911.24 24.7 9 0.03710 50612.11 25.8 25.9 0.02715 39078.79 27.2 300 0.03295 97288.47 55.8 13 56.4 0.03225 97192.73 57.0 17 0.02910 139276.83 90.5 90.5

229

Table B.3: AA generation data for the “two-times” processed PET sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.02680 19086.90 14.8 9 0.03125 19304.59 12.9 13.7 0.03080 19881.60 13.5 0.03740 37567.37 20.9 280 13 0.03690 30734.59 17.4 19.2 0.03455 32154.95 19.4 0.03585 45155.96 26.2 17 0.04010 56210.06 29.2 28.4 0.03060 43798.33 29.8 0.02675 20099.93 15.7 9 16.0 0.02720 21411.92 16.4 0.03505 41574.96 24.7 13 290 0.02985 39945.77 27.9 27.3 0.03650 51133.55 29.2 0.04175 84840.30 47986 42.3 17 0.03445 78842.81 47.7 43.9 0.02655 53093.88 41.7 0.04300 67648.51 32.8 0.03075 40882.68 27.7 0.02615 24352.52 19.4 9 27.7 0.02900 39893.46 28.7 0.02400 27161.42 23.6 0.03185 51644.79 33.8 300 0.02780 68678.28 51.5 0.03855 104591.71 56.5 13 53.8 0.03445 90257.62 54.6 0.03960 100011.75 52.6 0.03375 145962.51 90.1 17 0.03300 159866.00 101.0 92.9 0.03745 157664.24 87.7

230

Table B.4: AA generation data for the “three-times” processed PET sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.03480 19983.98 12.0 9 0.03705 18711.93 10.5 11.1 0.03515 18150.14 10.8 280 0.03395 34505.04 21.2 13 0.03070 27271.89 18.5 19.8 0.03575 33724.72 19.7 0.03875 55181.86 29.7 17 31.4 0.03840 61038.78 33.1 0.03335 28107.98 17.6 9 0.03530 31510.98 18.6 17.2 0.02395 17628.15 15.3 0.03410 53302.24 32.6 290 13 0.03455 49735.29 30.0 31.1 0.03675 54171.85 30.7 0.03525 83178.72 47986 49.2 17 0.03635 81792.24 46.9 48.1 0.03540 81691.81 48.1 0.03380 46176.30 28.5 0.03505 49510.63 29.4 0.03610 43349.75 25.0 9 25.5 0.03255 45173.80 28.9 0.03945 29147.50 15.4 0.03830 47533.72 25.9 300 0.03310 97304.28 61.3 13 0.03880 117635.76 63.2 60.5 0.03450 94438.06 57.0 0.03055 136156.96 92.9 0.03485 171052.89 102.3 17 96.5 0.03220 142404.74 92.2 0.03805 180394.53 98.8

231

Table B.5: AA generation data for the 10,000 ppm anthranilamide/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.02245 4249.97 3.6 9 3.7 0.02120 4245.73 3.8 0.02165 6291.43 5.5 13 5.4 280 0.02105 5852.28 5.3 0.01955 6051.81 5.9 17 0.02160 7280.81 6.4 6.1 0.02015 6508.65 6.1 0.02060 3627.16 3.3 9 4.1 0.02395 6130.99 4.8 0.02185 6725.36 5.8 13 5.7 290 0.02025 5997.03 5.6 0.02070 12793.24 11.7 17 0.02130 12423.49 52912 11.0 10.9 0.02120 11228.52 10.0 0.02040 5418.01 5.0 9 0.02085 8824.99 8.0 6.2 0.02130 6296.73 5.6 0.02140 17216.12 15.2 13 0.02230 18149.60 15.4 14.5 300 0.02210 15163.62 13.0 0.02205 25131.05 21.5 0.02005 22510.37 21.2 17 0.02175 26931.84 23.4 21.7 0.02395 25677.80 20.3 0.02265 26426.77 22.1

232

Table B.6: AA generation data for the 1200 ppm anthranilamide/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.02330 9896.61 8.0 9 0.02510 11507.79 8.7 8.5 0.02505 11733.47 8.9 0.02420 14406.13 11.3 280 13 0.02260 12812.98 10.7 10.6 0.02415 12607.13 9.9 0.02535 23442.67 17.5 17 0.02290 22603.12 18.7 16.6 0.02405 17490.64 13.7 0.02355 15553.14 12.5 9 0.02515 15674.32 11.8 12.1 0.02390 15151.83 12.0 0.02345 24561.60 19.8 13 290 0.02335 22359.01 52912 18.1 18.5 0.02285 21346.96 17.7 0.02425 36310.41 28.3 17 0.02165 32864.18 28.7 27.3 0.02365 30998.15 24.8 0.02180 22027.68 19.1 9 0.02220 19628.29 16.7 17.1 0.02220 18364.02 15.6 0.02260 35213.92 29.4 300 0.02205 28914.71 24.8 13 0.02485 35418.39 26.9 27.1 0.02420 35430.54 27.7 0.02545 36231.08 26.9 17 0.02105 47545.93 42.7 41.1 0.02310 48288.46 39.5

233

Table B.7: AA generation data for 500 ppm anthranilamide/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.02810 18021.34 12.1 0.03145 19407.38 11.7 9 0.02615 13790.78 10.0 11.0 0.03050 16692.38 10.3 0.02880 16376.51 10.7 280 0.03045 20993.01 13.0 0.02975 24129.28 15.3 13 15.7 0.02915 25945.29 16.8 0.02965 27567.94 17.6 0.02920 33708.79 21.8 17 0.02940 36425.88 23.4 22.4 0.03135 36548.79 22.0 0.03075 16768.51 10.3 11.3 9 0.02945 19494.97 12.5 0.02380 13985.96 11.1 0.02900 37315.08 24.3 0.02930 35571.38 52912 22.9 290 13 0.02810 34562.32 23.2 22.4 0.02875 31650.26 20.8 0.02775 30169.45 20.5 0.02895 48544.61 31.7 17 0.02775 43053.66 29.3 32.2 0.02885 54501.38 35.7 0.03080 33083.18 20.3 9 0.02860 34598.22 22.9 22.3 0.02870 36228.11 23.9 0.02950 60455.79 38.7 0.02890 74613.97 48.8 300 13 0.02960 62517.32 39.9 42.9 0.02985 68620.31 43.4 0.03075 71221.23 43.8 0.02810 120124.21 80.8 17 0.02925 93989.92 60.7 70.3 0.02860 105137.97 69.5

234

Table B.8: AA generation data for 200 ppm anthranilamide/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.03030 10225.45 6.4 9 0.03165 10644.51 6.4 6.6 0.03130 11481.15 6.9 0.03140 24122.73 14.5 280 13 0.03010 22480.27 14.1 13.9 0.03205 22101.00 13.0 0.03110 34760.10 21.1 17 0.03170 34558.37 20.6 22.0 0.02985 38158.87 24.2 0.02730 10873.25 7.5 9 0.03050 14021.67 8.7 8.0 0.02960 12310.55 7.9 0.03205 46925.65 27.7 290 13 0.03115 43623.08 52912 26.5 27.7 0.03175 48448.48 28.8 0.02925 71773.69 46.4 17 0.02960 63022.82 40.2 46.0 0.02895 78874.51 51.5 0.02920 28766.98 18.6 0.02990 31040.32 19.6 9 18.1 0.03015 27853.00 17.5 0.03105 27150.27 16.5 300 0.03080 77789.41 47.7 13 0.03135 82340.70 49.6 46.8 0.02975 67856.91 43.1 0.02985 113689.79 72.0 17 74.5 0.03045 124115.41 77.0

235

Table B.9: AA generation data for 100 ppm anthranilamide/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.03570 12735.22 6.7 9 0.03480 12492.80 6.8 6.9 0.03505 13341.74 7.2 0.03560 19565.94 10.4 280 13 0.03520 20124.78 10.8 10.7 0.03550 20461.39 10.9 0.03365 31584.91 17.7 17 0.03375 33297.00 18.6 18.2 0.03435 33074.75 18.2 0.03215 18921.91 11.1 9 0.03445 19842.30 10.9 11.3 0.03255 20607.85 12.0 0.03355 37765.31 52912 21.3 290 13 0.03360 39113.55 22.0 21.6 0.03385 38620.54 21.6 0.03400 70310.31 39.1 17 0.03385 74614.29 41.7 41.8 0.03420 81055.47 44.8 0.03345 36266.90 20.5 9 0.03295 33385.59 19.1 18.8 0.03355 29598.84 16.7 300 0.03390 87612.33 48.8 13 51.3 0.03540 100510.02 53.7 0.03230 133978.81 78.4 17 78.8 0.03540 148366.18 79.2

236

Table B.10: AA generation data for 50,000 ppm alpha-cyclodextrin/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.02660 6595.91 4.7 9 0.03120 7747.56 4.7 4.7 0.02905 7457.12 4.9 280 0.02680 8188.56 5.8 13 0.02860 8392.27 5.5 5.8 0.02605 8455.12 6.1 0.02965 10721.99 6.8 17 6.8 0.02970 10608.79 6.8 0.02980 10112.52 6.4 9 0.03550 14179.20 7.5 6.5 0.02895 8704.33 5.7 0.02900 13060.14 8.5 290 13 52912 8.7 0.02715 12847.98 8.9 0.02910 20739.10 13.5 17 0.02765 16413.42 11.2 12.7 0.02965 20896.71 13.3 0.03210 19118.94 11.3 9 0.02795 17774.72 12.0 11.4 0.02935 16961.20 10.9 0.02940 33958.58 21.8 300 13 19.8 0.03175 29992.48 17.9 0.02755 43588.17 29.9 17 0.02630 51193.84 36.8 35.9 0.03185 69159.29 41.0

237

Table B.11: AA generation data for 25,000 ppm alpha-cyclodextrin/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.03035 6540.76 4.1 9 3.9 0.02710 5377.78 3.8 0.02525 6910.67 5.2 13 5.3 280 0.02585 7437.02 5.4 0.02585 7675.39 5.6 17 0.02840 9225.79 6.1 6.1 0.02595 8795.84 6.4 0.02760 9190.52 6.3 9 0.02890 8589.87 5.6 5.9 0.02890 8802.35 5.8 0.02405 9342.69 7.3 290 13 0.02725 12520.92 8.7 7.8 0.02580 9909.52 52912 7.3 0.02505 14992.04 11.3 17 0.02895 14512.46 9.5 11.9 0.02790 21937.51 14.9 0.02605 8530.57 6.2 9 0.02865 11646.99 7.7 7.7 0.03155 15509.56 9.3 0.02765 24849.61 17.0 300 13 0.02685 28596.51 20.1 18.4 0.02665 25474.35 18.1 0.02980 56340.97 35.7 17 0.02985 61681.45 39.1 37.6 0.02435 49060.00 38.1

238

Table B.12: AA generation data for 10,000 ppm alpha-cyclodextrin/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.02940 5281.10 3.4 9 0.02920 6594.27 4.3 4.1 0.02805 6963.86 4.7 0.02985 7509.68 4.8 280 13 0.02960 10526.92 6.7 5.3 0.03175 7260.41 4.3 0.03125 11185.25 6.8 17 0.02835 12178.99 8.1 7.1 0.02960 10185.59 6.5 0.03030 6860.12 4.3 9 0.02990 6806.93 4.3 4.3 0.03070 7078.03 4.4 0.03035 10375.66 6.5 290 13 0.02840 12320.47 52912 8.2 7.2 0.03035 11328.21 7.1 0.03115 22331.10 13.5 17 0.02710 20352.17 14.2 13.4 0.02855 18745.94 12.4 0.02945 11015.65 7.1 9 0.02875 9707.83 6.4 6.7 0.03040 10746.55 6.7 0.02915 26449.12 17.1 300 0.02730 25143.91 17.4 13 19.5 0.02930 35513.64 22.9 0.02650 28967.23 20.7 0.03215 77981.30 45.8 17 41.9 0.02735 54976.22 38.0

239

Table B.13: AA generation data for 5000 ppm alpha-cyclodextrin/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.03370 4028.45 2.3 9 0.03180 4449.18 2.6 2.7 0.02635 4498.99 3.2 0.03410 8351.80 4.6 280 13 0.03115 8591.22 5.2 4.9 0.03045 7872.67 4.9 0.02915 9293.40 6.0 17 0.03275 11399.70 6.6 6.7 0.03075 12166.39 7.5 0.02830 8853.82 5.9 9 0.02935 8248.81 5.3 5.7 0.03090 9842.70 6.0 0.03220 19480.79 11.4 290 13 0.03045 19106.01 52912 11.9 11.3 0.03120 17479.80 10.6 0.02990 33492.25 21.2 17 0.03185 33031.97 19.6 20.0 0.03135 31897.86 19.2 0.03175 16590.18 9.9 9 0.03375 20325.06 11.4 9.7 0.03085 12915.65 7.9 0.03275 34816.97 20.1 300 13 0.03395 36154.95 20.1 20.7 0.03150 36298.49 21.8 0.03420 82093.43 45.4 17 0.02830 57303.69 38.3 43.3 0.03310 81164.35 46.3

240

Table B.14: AA generation data for 1200 ppm alpha-cyclodextrin/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.03275 10353.17 6.0 9 0.03450 9201.02 5.0 5.4 0.03110 8691.70 5.3 0.03280 20289.26 11.7 13 0.03115 18825.27 11.4 11.7 280 0.02920 18719.20 12.1 0.03305 22355.14 12.8 0.03210 21024.08 12.4 17 0.03250 19807.18 11.5 12.0 0.03015 18523.24 11.6 0.03235 19769.17 11.5 0.02990 12760.17 8.1 9 0.03165 13754.56 8.2 7.9 0.03315 13207.64 7.5 0.03170 22940.03 52912 13.7 290 13 0.02925 22020.73 14.2 14.6 0.03065 25651.33 15.8 0.03245 57244.16 33.3 17 0.03385 60454.90 33.8 33.1 0.03345 57115.96 32.3 0.03125 23378.92 14.1 9 0.03190 26013.59 15.4 14.3 0.02945 20654.78 13.3 0.03305 56562.87 32.3 0.03095 53728.78 32.8 300 13 34.5 0.03265 65121.31 37.7 0.03035 56663.26 35.3 0.03000 90427.19 57.0 17 0.03205 117884.86 69.5 63.7 0.03300 112926.56 64.7

241

Table B.15: AA generation data for 500 ppm alpha-cyclodextrin/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.03540 12693.84 6.8 9 0.03355 11745.19 6.6 6.6 0.03535 12102.76 6.5 0.03170 23279.11 13.9 280 13 0.03505 24231.11 13.1 13.3 0.03355 23261.63 13.1 0.03680 33651.79 17.3 17 0.03645 34374.17 17.8 17.5 0.03445 31610.63 17.3 0.03545 19266.52 10.3 9 0.03360 19989.25 11.2 10.3 0.03310 16684.28 9.5 0.03490 39173.70 21.2 0.03405 38066.63 21.1 13 22.5 290 0.03530 42430.24 52912 22.7 0.03405 44991.92 25.0 0.03455 77099.50 42.2 0.03570 82365.18 43.6 17 0.03435 74410.02 40.9 42.2 0.03350 76266.25 43.0 0.03470 75943.04 41.4 0.03445 31196.78 17.1 9 0.03130 33359.75 20.1 19.3 0.03235 35547.25 20.8 0.03650 83803.01 43.4 300 13 0.03365 76059.88 42.7 42.6 0.03530 78098.14 41.8 0.03575 134801.71 71.3 17 0.03440 155068.99 85.2 77.0 0.03085 121579.66 74.5

242

Table B.16: AA generation data for 10,000 ppm MXDA/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.01915 10754.44 10.6 9 0.02440 10904.45 8.4 9.5 0.02775 13664.13 9.3 0.02880 22799.00 15.0 280 13 0.02750 22376.95 15.4 15.6 0.02600 22565.43 16.4 0.03050 26850.58 16.6 17 0.03000 26629.59 16.8 16.5 0.02510 21336.00 16.1 0.01520 9012.27 11.2 9 0.02335 12347.19 10.0 10.9 0.02545 15505.40 11.5 0.02680 28529.24 20.1 290 13 52912 21.2 0.02675 31445.05 22.2 0.02605 40309.15 29.2 17 0.02795 45428.41 30.7 28.8 0.02320 32385.49 26.4 0.03255 31877.31 18.5 9 0.03005 34953.07 22.0 20.0 0.01910 19864.26 19.7 0.02755 57753.20 39.6 300 0.02650 55280.63 39.4 13 37.9 0.02870 52144.27 34.3 0.02395 48557.32 38.3 0.02735 89201.85 61.6 17 61.4 0.02845 91998.16 61.1

243

Table B.17: AA generation data for 1200 ppm MXDA/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.02640 6649.77 4.8 9 0.02620 7235.20 5.2 5.0 0.02750 7401.17 5.1 0.02680 18914.67 13.3 280 13 13.2 0.02530 17367.92 13.0 0.02735 21933.95 15.2 17 0.02460 21345.86 16.4 15.8 0.02470 20843.12 15.9 0.02430 13497.20 10.5 9 0.02510 14573.04 11.0 11.2 0.02670 17207.69 12.2 0.02710 23537.10 16.4 290 13 17.1 0.02605 24523.26 52912 17.8 0.02955 49215.59 31.5 17 0.02755 43674.35 30.0 30.5 0.02500 39667.18 30.0 0.03005 36184.83 22.8 9 0.02825 33937.10 22.7 22.8 0.02435 29587.95 23.0 0.02695 55417.39 38.9 300 13 0.02620 62362.74 45.0 40.8 0.02755 55980.98 38.4 0.02640 76308.92 54.6 17 0.02635 82921.32 59.5 63.0 0.02650 104928.60 74.8

244

Table B.18: AA generation data for 500 ppm MXDA/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.02765 13191.44 9.0 0.02820 12492.65 8.4 9 0.02525 13328.85 10.0 9.6 0.02555 14422.95 10.7 0.02705 14393.95 10.1 0.03065 20966.06 12.9 0.02830 20708.67 13.8 0.02745 20939.59 14.4 280 0.02665 17507.10 12.4 13 13.4 0.02790 19088.11 12.9 0.02780 19932.43 13.6 0.02835 21572.78 14.4 0.02645 17871.96 12.8 0.02885 36947.35 24.2 0.02780 30677.74 20.9 17 21.9 0.02780 31587.87 21.5 0.02790 31245.72 21.2 0.02590 16052.88 52912 11.7 0.02700 18553.96 13.0 9 12.0 0.02690 18183.49 12.8 0.02740 15284.85 10.5 290 0.02775 32819.58 22.4 13 0.02945 37924.79 24.3 23.7 0.03030 39276.05 24.5 0.02810 42378.80 28.5 17 0.02605 41239.44 29.9 30.6 0.02585 45655.78 33.4 0.02675 25835.95 18.3 0.02785 25942.89 17.6 9 20.5 0.02865 33708.15 22.2 0.02575 32832.29 24.1 300 0.02850 62757.82 41.6 13 40.8 0.02760 58393.91 40.0 0.02645 92609.84 66.2 17 0.02980 105440.67 66.9 63.9 0.02695 83682.10 58.7

245

Table B.19: AA generation data for 200 ppm MXDA/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.03445 16212.06 8.9 9 0.03310 15477.49 8.8 8.9 0.03495 16748.32 9.1 0.03720 28532.02 14.5 280 13 0.03495 26934.12 14.6 14.8 0.03345 27149.13 15.3 0.03345 41604.17 23.5 17 0.03515 34630.82 18.6 20.7 0.03335 35111.05 19.9 0.03395 21221.74 11.8 9 0.03465 22229.38 12.1 12.6 0.03410 24742.35 13.7 0.03270 43249.72 25.0 290 13 0.03470 39944.27 52912 21.8 24.0 0.03515 46892.40 25.2 0.03385 58241.69 32.5 17 0.03295 57959.91 33.2 33.0 0.03330 58616.22 33.3 0.03235 28788.19 16.8 9 0.03525 36024.56 19.3 18.5 0.03515 35902.69 19.3 0.03095 60134.01 36.7 300 13 0.03430 69062.89 38.1 39.3 0.03610 82593.73 43.2 0.03295 122668.66 70.4 17 0.03320 116086.15 66.1 67.4 0.03290 114399.97 65.7

246

Table B.20: AA generation data for 100 ppm MXDA/PET blend sample Temperature Time Sample Peak Calibration AA Concentration (oC) (minutes) Weight Area Factor (ppm) (grams) Measured Average 0.03710 20894.29 10.6 9 0.03535 20557.17 11.0 10.4 0.03525 17958.63 9.6 0.03460 33680.88 18.4 0.03390 33325.23 18.6 0.03235 32686.28 19.1 280 13 18.4 0.03315 32775.89 18.7 0.03440 32475.79 17.8 0.03295 31438.10 18.0 0.03695 51837.08 26.5 17 0.03650 47193.62 24.4 24.5 0.03450 40945.71 22.4 0.03190 22810.96 13.5 0.03580 23400.97 12.4 9 13.5 0.03335 25164.79 14.3 0.03690 27110.60 13.9 0.03285 42073.29 24.2 0.03790 50437.61 52912 25.2 0.03240 39126.95 22.8 290 13 23.8 0.03410 43681.92 24.2 0.03680 44574.51 22.9 0.03455 42487.90 23.2 0.03355 91214.95 51.4 0.03440 92538.91 50.8 17 50.5 0.03525 96388.32 51.7 0.03465 87926.34 48.0 0.03505 34816.07 18.8 9 0.03330 32006.28 18.2 18.7 0.03490 35129.44 19.0 0.03405 93445.80 51.9 300 13 0.03405 92382.33 51.3 51.3 0.03410 91450.67 50.7 0.03625 132755.61 69.2 17 0.03405 138677.16 77.0 73.1 0.03365 130307.69 73.2

247

Appendix C

AA Generation Plots

70.0

280 C 60.0 290 C

50.0 300 C

40.0 y = 6.3768x - 49.045 R2 = 0.954 30.0 y = 2.3578x - 15.977 R2 = 0.9968 Acetaldehyde (ppm) 20.0 y = 0.4915x + 0.6075 R2 = 0.9459 10.0

0.0 0 5 10 15 20 Time (minutes) Figure C-1: AA generation plots for the Voridian CB12 PET resin

248

100.0

90.0

80.0 280 C 290 C 70.0 300 C 60.0

50.0 y = 8.0708x - 47.346 R2 = 0.999 40.0 y = 2.941x - 11.937 R2 = 0.9486 Acetaldehyde (ppm) 30.0 y = 1.5688x - 4.0364 R2 = 0.9973 20.0

10.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-2: AA generation plots for the “one-time” processed PET sample

100.0

90.0 280 C 80.0 290 C

70.0 300 C

60.0 y = 8.1602x - 47.946 R2 = 0.987 50.0 y = 3.4842x - 16.226 R2 = 0.9876 40.0 y = 1.8384x - 3.4389 2

Acetaldehyde (ppm) R = 0.9794 30.0

20.0

10.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-3: AA generation plots for the “two-times” processed PET sample

249

100.0 280 C 290 C 80.0 300 C y = 8.8765x - 54.546 R2 = 0.9999 60.0 y = 3.8605x - 18.08 R2 = 0.9968 40.0

Acetaldehyde (ppm) y = 2.5396x - 12.258 R2 = 0.9932 20.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-4: AA generation plots for the “three-times” processed PET sample

25.0 280 C 290 C 20.0 300 C y = 1.9366x - 11.038 R2 = 0.9982 15.0

y = 0.8527x - 4.187 R2 = 0.9162 10.0

Acetaldehyde (ppm) y = 0.3034x + 1.1103 R2 = 0.9508 5.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-5: AA generation plots for the 10,000 ppm anthranilamide/ PET blend sample

250

90.0 280 C 80.0 290 C 70.0 300 C 60.0

50.0 y = 5.9991x - 32.787 R2 = 0.9933 40.0 y = 2.6164x - 12.04 R2 = 0.9989 30.0 Acetaldehyde (ppm) y = 1.4318x - 2.2536 R2 = 0.9898 20.0

10.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-6: AA generation plots for the 500 ppm anthranilamide/ PET blend sample

80.0

70.0 280 C 290 C 60.0 300 C

50.0

40.0 y = 7.0565x - 45.271 R2 = 0.9999 30.0 y = 4.7512x - 34.526 Acetaldehyde (ppm) R2 = 0.9996 20.0 y = 1.9258x - 10.9 R2 = 0.9992 10.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-7: AA generation plots for the 200 ppm anthranilamide/ PET blend sample

251

80.0

70.0 280 C 290 C 60.0 300 C

50.0

40.0 y = 7.5038x - 47.941 R2 = 0.9978 30.0

Acetaldehyde (ppm) y = 3.815x - 24.668 R2 = 0.9658 20.0 y = 1.4109x - 6.4099 R2 = 0.9652 10.0

0.0 0 2 4 6 8 101214161820 Time (minutes) Figure C-8: AA generation plots for the 100 ppm anthranilamide/ PET blend sample

45.0 280 C 40.0 290 C 35.0 300 C

30.0

25.0 y = 3.0638x - 17.446 R2 = 0.9687 20.0 y = 0.7651x - 0.6319 R2 = 0.9731 15.0 Acetaldehyde (ppm) y = 0.2561x + 2.4553 R2 = 0.9992 10.0

5.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-9: AA generation plots for the 50,000 ppm alpha-cyclodextrin/ PET blend sample

252

40.0

35.0 280 C 30.0 290 C

25.0 300 C y = 3.7375x - 27.343 R2 = 0.9734 20.0 y = 0.7491x - 1.2275 R2 = 0.9552 15.0 y = 0.2676x + 1.6111 Acetaldehyde (ppm) R2 = 0.9706 10.0

5.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-10: AA generation plots for the 25,000 ppm alpha-cyclodextrin/ PET blend sample

50.0

45.0 280 C

40.0 290 C

35.0 300 C

30.0 y = 4.4006x - 34.489 2 25.0 R = 0.976 y = 1.1339x - 6.4286 2 20.0 R = 0.9597 y = 0.3763x + 0.6121 2

Acetaldehyde (ppm) R = 0.9815 15.0

10.0

5.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-11: AA generation plots for the 10,000 ppm alpha-cyclodextrin/ PET blend sample

253

50.0

45.0 280 C 290 C 40.0 300 C 35.0

30.0

25.0 y = 4.2003x - 30.033 R2 = 0.961 20.0 y = 1.7815x - 10.812 R2 = 0.9839 Acetaldehyde (ppm) 15.0 y = 0.498x - 1.7025 R2 = 0.9964 10.0

5.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-12: AA generation plots for the 5000 ppm alpha-cyclodextrin/ PET blend sample

70.0

60.0 280 C 290 C

50.0 300 C

40.0 y = 6.1812x - 42.849 R2 = 0.9893 30.0 y = 3.1481x - 22.382 2

Acetaldehyde (ppm) R = 0.9306 20.0 y = 0.817x - 0.9064 R2 = 0.7758 10.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-13: AA generation plots for the 1200 ppm alpha-cyclodextrin/ PET blend sample

254

90.0 280 C 80.0 290 C 70.0 300 C

60.0

50.0 y = 7.2048x - 47.341 R2 = 0.9879 40.0 y = 3.9843x - 26.77 R2 = 0.9816 30.0 Acetaldehyde (ppm) y = 1.3576x - 5.165 R2 = 0.9813 20.0

10.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-14: AA generation plots for the 500 ppm alpha-cyclodextrin/ PET blend sample

70.0

60.0 280 C 290 C 50.0 300 C

40.0 y = 5.166x - 27.374 R2 = 0.994 30.0 y = 2.2346x - 8.7653 R2 = 0.9927 Acetaldehyde (ppm) 20.0 y = 0.8797x + 2.4067 R2 = 0.8454 10.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-15: AA generation plots for the 10,000 ppm MXDA/ PET blend sample

255

80.0

70.0 280 C 60.0 290 C 300 C 50.0

40.0 y = 5.0212x - 23.097 R2 = 0.9962 30.0 y = 2.4073x - 11.696 Acetaldehyde (ppm) Acetaldehyde R2 = 0.9521 20.0 y = 1.3516x - 6.233 R2 = 0.9218 10.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-16: AA generation plots for the 1200 ppm MXDA/ PET blend sample

70.0

60.0 280 C 290 C 50.0 300 C 40.0 y = 5.4201x - 28.709 R2 = 0.9986 30.0 y = 2.3245x - 8.1065 R2 = 0.9778 Acetaldehyde (ppm) 20.0 y = 1.5383x - 5.0164 R2 = 0.9529 10.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-17: AA generation plots for the 500 ppm MXDA/ PET blend sample

256

70.0 280 C 60.0 290 C

50.0 300 C y = 6.1134x - 37.74 R2 = 0.9928 40.0 y = 2.5574x - 10.064 R2 = 0.9954 30.0 y = 1.4682x - 4.2849 R2 = 1 Acetaldehyde (ppm) 20.0

10.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-18: AA generation plots for the 200 ppm MXDA/ PET blend sample

80.0

70.0 280 C 290 C 60.0 300 C

50.0 y = 6.8088x - 40.829 R2 = 0.9871 40.0 y = 4.6202x - 30.822 R2 = 0.938 30.0 y = 1.7549x - 5.0404 Acetaldehyde (ppm) R2 = 0.9933 20.0

10.0

0.0 0 5 10 15 20 Time (minutes)

Figure C-19: AA generation plots for the 100 ppm MXDA/ PET blend sample

257

*The AA generation plots for the 1200 ppm anthranilamide/PET blend sample are shown

in Figure 4-23.

258

Appendix D

Arrhenius Plots

2.50

2.00 y = -25926x + 47.253 R2 = 0.9795 1.50 ln Rate ln 1.00

0.50

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 1/Temperature (1/K)

Figure D-1: Arrhenius plot for the “one-time” processed PET sample

259

2.50

2.00 y = -23603x + 43.24 R2 = 0.9915 1.50 ln Rate ln 1.00

0.50

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 1/Temperature (1/K)

Figure D-2: Arrhenius plot for the “two-times” processed PET sample

2.50

2.00 y = -19796x + 36.649 R2 = 0.961 1.50 ln Rate ln 1.00

0.50

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 1/Temperature (1/K)

Figure D-3: Arrhenius plot for the “three-times” processed PET sample

260

1.00

0.50

y = -29401x + 51.988 R2 = 0.9969 0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 ln Rate ln -0.50

-1.00

-1.50 1/Temperature (1/K)

Figure D-4: Arrhenius plot for the 10,000 ppm anthranilamide/PET blend sample

1.20

1.00

0.80 y = -18593x + 33.577 R2 = 0.9865 0.60

ln Rate ln 0.40

0.20

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182

-0.20 1/Temperature (1/K)

Figure D-5: Arrhenius plot for the 1200 ppm anthranilamide/PET blend sample

261

2.00

1.80

1.60

1.40 y = -22687x + 41.332 2 1.20 R = 0.9897

1.00 ln Rate ln 0.80

0.60

0.40

0.20

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 1/Temperature (1/K)

Figure D-6: Arrhenius plot for the 500 ppm anthranilamide/PET blend sample

2.50

2.00

1.50 y = -20631x + 38.032 R2 = 0.9559 ln Rate ln 1.00

0.50

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 1/Temperature (1/K)

Figure D-7: Arrhenius plot for the 200 ppm anthranilamide/PET blend sample

262

2.50

2.00 y = -26518x + 48.332 R2 = 0.9902 1.50 ln Rate ln 1.00

0.50

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 1/Temperature (1/K)

Figure D-8: Arrhenius plot for the 100 ppm anthranilamide/PET blend sample

1.50

1.00 y = -39310x + 69.649 R2 = 0.9939 0.50

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182

ln Rate ln -0.50

-1.00

-1.50

-2.00 1/Temperature (1/K)

Figure D-9: Arrhenius plot for the 50,000 ppm alpha-cyclodextrin/PET blend sample

263

1.50

1.00 y = -41738x + 74.034 R2 = 0.9816 0.50

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182

ln Rate ln -0.50

-1.00

-1.50

-2.00 1/Temperature (1/K)

Figure D-10: Arrhenius plot for the 25,000 ppm alpha-cyclodextrin/PET blend sample

2.00

1.50

1.00 y = -38954x + 69.396 R2 = 0.9952 0.50

ln Rate ln 0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182

-0.50

-1.00

-1.50 1/Temperature (1/K)

Figure D-11: Arrhenius plot for the 10,000 ppm alpha-cyclodextrin/PET blend sample

264

2.00

1.50 y = -33837x + 60.536 R2 = 0.9896 1.00

0.50 ln Rate ln

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182

-0.50

-1.00 1/Temperature (1/K)

Figure D-12: Arrhenius plot for the 5000 ppm alpha-cyclodextrin/PET blend sample

2.50

2.00 y = -32138x + 58.003 R2 = 0.968 1.50

1.00 ln Rate ln

0.50

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182

-0.50 1/Temperature (1/K)

Figure D-13: Arrhenius plot for the 1200 ppm alpha-cyclodextrin/PET blend sample

265

2.50

2.00 y = -26500x + 48.287 R2 = 0.9759 1.50 ln Rate ln 1.00

0.50

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 1/Temperature (1/K)

Figure D-14: Arrhenius plot for the 500 ppm alpha-cyclodextrin/PET blend sample

1.80

1.60

1.40 y = -20787x + 37.85 R2 = 0.9937 1.20

1.00

ln Rate ln 0.80

0.60

0.40

0.20

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 1/Temperature (1/K)

Figure D-15: Arrhenius plot for the 1200 ppm MXDA/PET blend sample

266

1.80

1.60

1.40 y = -19922x + 36.371 R2 = 0.9579 1.20

1.00

ln Rate ln 0.80

0.60

0.40

0.20

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 1/Temperature (1/K)

Figure D-16: Arrhenius plot for the 500 ppm MXDA/PET blend sample

2.00

1.80

1.60 y = -22580x + 41.148 1.40 R2 = 0.9812 1.20

1.00 ln Rate ln 0.80

0.60

0.40

0.20

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 1/Temperature (1/K)

Figure D-17: Arrhenius plot for the 200 ppm MXDA/PET blend sample

267

2.50

2.00 y = -21544x + 39.602 R2 = 0.9471 1.50 ln Rate ln 1.00

0.50

0.00 0.00174 0.00175 0.00176 0.00177 0.00178 0.00179 0.00180 0.00181 0.00182 1/Temperature (1/K)

Figure D-18: Arrhenius plot for the 100 ppm MXDA/PET blend sample

*The Arrhenius plot for the virgin PET sample, Voridian CB12 PET resin, is shown in

Figure 4-32.

**The Arrhenius plot of the 10,000 ppm MXDA/PET blend sample is shown in Figure

4-27.

268

Appendix E

Raw Data from Residual AA Experiments

Table E.1: Residual AA data for the control PET pelletized samples Number of Sample Peak Calibration AA Concentration Processing Weight Area Factor (ppm) Times (grams) Measured Average 0.31795 262.89 997.76 0.8 0 0.8 0.37230 309.34 997.76 0.8 0.36160 3440.53 997.76 9.5 1 0.33355 2810.13 997.76 8.4 9.0 0.36225 3222.81 997.76 8.9 0.42160 3548.47 607.32 13.9 2 13.5 0.29300 2339.18 607.32 13.1 0.32455 2787.06 607.32 14.1 3 0.34670 3014.77 607.32 14.3 14.5 0.37735 3442.92 607.32 15.0

Table E.2: Residual AA data for the anthranilamide/PET blend pelletized samples AA Scavenger Sample Peak Calibration AA Concentration Concentration Weight Area Factor (ppm) (ppm) (grams) Measured Average

1200 0.37050 382.29 997.76 1.0 1.0 0.41515 418.46 997.76 1.0 500 0.30000 883.41 997.76 3.0 2.9 0.32855 943.95 997.76 2.9 200 0.31830 1121.57 997.76 3.5 3.5 0.32530 1101.01 997.76 3.4 100 0.31940 1759.95 997.76 5.5 5.2 0.30195 1467.89 997.76 4.9 *10,000 ppm anthranilamide/PET blend sample was crystallized prior to residual AA analysis and was therefore not analyzed.

269

Table E.3: Residual AA data for the alpha-cyclodextrin/PET blend pelletized samples AA Scavenger Sample Peak Calibration AA Concentration Concentration Weight Area Factor (ppm) (ppm) (grams) Measured Average 50,000 0.32105 850.28 997.76 2.7 2.6 0.33890 859.9 997.76 2.5 25,000 0.30835 771.16 997.76 2.5 2.7 0.32920 976.4 997.76 3.0 10,000 0.30890 825.7 997.76 2.7 2.6 0.29715 740.93 997.76 2.5 5000 0.33785 1045.93 997.76 3.1 2.8 0.32785 803.48 997.76 2.5 1200 0.32540 1034.51 997.76 3.2 3.3 0.31900 1059.25 997.76 3.3 500 0.29275 1456.23 997.76 5.0 4.9 0.31710 1539.92 997.76 4.9

Table E.4: Residual AA data for the MXDA/PET blend pelletized samples AA Scavenger Sample Peak Calibration AA Concentration Concentration Weight Area Factor (ppm) (ppm) (grams) Measured Average 10,000 0.32840 810.54 997.76 2.5 2.8 0.29355 893.1 997.76 3.0 1200 0.38100 1270.76 997.76 3.3 3.4 0.30530 1069.07 997.76 3.5 500 0.30215 1136.53 997.76 3.8 3.6 0.30175 1031.58 997.76 3.4 200 0.35285 1122.09 997.76 3.2 3.2 0.33590 1092.64 997.76 3.3 100 0.29875 1398.91 997.76 4.7 4.1 0.33440 1174.02 997.76 3.5

Table E.5: Residual AA data for the PET control preform samples Number Preform Sample Peak Calibration AA Concentration of # Weight Area Factor (ppm) Processing (grams) Measured Overall Times Average 0.23540 2016.35 997.76 8.6 1 0.28975 2220.50 997.76 7.7 0.31690 2497.23 997.76 7.9 1 2 0.25230 2113.16 997.76 8.4 8.3 0.26200 2076.94 997.76 7.9 0.23540 2115.34 997.76 9.0 3 0.27660 2573.11 997.76 9.3 0.26920 2030.63 997.76 7.6

270

Table E.6: Residual AA data for the anthranilamide/PET preform samples AA Scavenger Preform Sample Peak Calibration AA Concentration Concentration # Weight Area Factor (ppm) (ppm) (grams) Measured Overall Average 0.31950 1329.36 997.76 4.2 1 0.32495 1208.60 997.76 3.7 0.32155 1117.58 997.76 3.5 0.28760 1068.99 997.76 3.7 200 2 0.30770 1143.28 997.76 3.7 3.7 0.30220 1173.87 997.76 3.9 0.27245 835.33 997.76 3.1 3 0.34575 1269.13 997.76 3.7 0.28550 1083.24 997.76 3.8 1 0.28610 1226.11 997.76 4.3 0.29265 1253.49 997.76 4.3 2 0.35085 1984.18 997.76 5.7 100 0.29525 1367.02 997.76 4.6 4.5 0.31510 1167.27 997.76 3.7 3 0.27695 1244.44 997.76 4.5 0.34720 1364.31 997.76 3.9

Table E.7: Residual AA data for the alpha-cyclodextrin/PET preform samples AA Scavenger Preform Sample Peak Calibration AA Concentration Concentration # Weight Area Factor (ppm) (ppm) (grams) Measured Overall Average 1 0.28785 1187.25 997.76 4.1 0.29595 1138.16 997.76 3.9 2 0.33600 1698.11 997.76 5.1 1200 0.32930 1374.88 997.76 4.2 4.7 0.40820 2138.45 997.76 5.3 3 0.34405 2029.54 997.76 5.9 0.32190 1731.37 997.76 5.4 1 0.36810 1942.57 997.76 5.3 0.25510 1237.87 997.76 4.9 0.26760 1069.82 997.76 4.0 500 2 0.29785 1127.20 997.76 3.8 4.8 0.28585 1310.25 997.76 4.6 3 0.37415 1990.50 997.76 5.3

271

Table E.8: Residual AA data for the MXDA/PET preform samples AA Scavenger Preform Sample Peak Calibration AA Concentration Concentration # Weight Area Factor (ppm) (ppm) (grams) Measured Overall Average 0.27425 1599.40 997.76 5.8 1 0.31565 1619.12 997.76 5.1 0.32605 1664.68 997.76 5.1 0.36975 1391.97 997.76 3.8 200 2 0.30485 986.34 997.76 3.2 4.6 0.25450 793.49 997.76 3.1 0.34915 1905.83 997.76 5.5 3 0.29340 1172.77 997.76 4.0 0.34005 1992.79 997.76 5.9 1 0.33475 2207.49 997.76 6.6 0.25985 1241.14 997.76 4.8 0.28880 1226.73 997.76 4.3 2 0.29330 1018.32 997.76 3.5 100 5.3 0.31425 1982.32 997.76 6.3 0.32795 1902.30 997.76 5.8 3 0.29690 1170.65 997.76 4.0 0.25850 1673.71 997.76 6.5

272

Appendix F

Raw Data from Melt Viscosity Measurements to Determine I.V.

Table F.1: Melt viscosity data for the control PET pelletized samples Resin Number of η* (Paxseconds) @ I.V. (dL/g) Processing Times 10 radians/second Calculated Average 853.900 0.80 1234.400 0.86 0 754.640 0.78 0.80 776.730 0.79 814.700 0.79 CB12 762.010 0.78 1 724.320 0.78 0.78 685.990 0.77 2 407.340 0.69 0.69 389.950 0.69 3 319.280 0.66 0.65 303.520 0.65

Table F.2: Melt viscosity data for the anthranilamide/PET blend pelletized samples Resin AA Scavenger η* (Paxseconds) I.V. (dL/g) Concentration @ 10 Calculated Average (ppm) radians/second

125.570 0.52 10,000 148.680 0.55 0.51 88.662 0.47 1200 324.320 0.66 0.66 299.250 0.65 CB12 500 363.900 0.68 0.68 375.300 0.68 200 373.130 0.68 0.68 339.770 0.67 100 319.240 0.66 0.67 360.030 0.68

273

Table F.3: Melt viscosity data for the alpha-cyclodextrin/PET blend pelletized samples Resin AA Scavenger η* (Paxseconds) I.V. (dL/g) Concentration @ 10 Calculated Average (ppm) radians/second 50,000 54.810 0.40 0.40 54.806 0.40 25,000 80.640 0.46 0.46 79.998 0.46 10,000 142.070 0.54 0.54 CB12 143.420 0.54 5000 186.590 0.58 0.58 186.380 0.58 1200 463.270 0.71 0.71 450.910 0.71 500 506.740 0.73 0.73 527.640 0.73

Table F.4: Melt viscosity data for the MXDA/PET blend pelletized samples Resin AA Scavenger η* (Paxseconds) I.V. (dL/g) Concentration @ 10 Calculated Average (ppm) radians/second 10,000 35.708 0.33 0.34 37.114 0.34 1200 360.350 0.68 0.68 352.600 0.67 CB12 500 432.210 0.70 0.70 395.860 0.69 200 352.630 0.67 0.67 318.130 0.66 100 372.120 0.68 0.68 353.780 0.67

Table F.5: Melt viscosity data for the control PET preform samples Resin Preform η* (Paxseconds) I.V. (dL/g) # @ 10 radians/second Calculated Overall Average 660.740 0.76 1 618.040 0.75 628.370 0.76 CB12 2 0.76 641.580 0.76 668.130 0.77 3 719.220 0.78

274

Table F.6: Melt viscosity data for the anthranilamide/PET blend preform samples Resin AA Scavenger Preform η* (Paxseconds) I.V. (dL/g) Concentration # @ 10 Calculated Overall (ppm) radians/second Average 1 729.520 0.78 688.350 0.77 200 2 727.550 0.78 0.78 734.040 0.78 3 871.040 0.80 CB12 715.910 0.78 1 689.730 0.77 698.670 0.77 100 2 768.700 0.79 0.78 763.020 0.78 3 889.980 0.81 724.990 0.78

Table F.7: Melt viscosity data for the alpha-cyclodextrin/PET blend preform samples Resin AA Scavenger Preform η* (Paxseconds) I.V. (dL/g) Concentration # @ 10 Calculated Overall (ppm) radians/second Average 1 501.820 0.72 379.990 0.68 1200 2 487.060 0.72 0.70 394.790 0.69 3 475.970 0.72 CB12 388.100 0.69 1 521.390 0.73 532.320 0.73 500 2 547.480 0.74 0.73 526.190 0.73 3 538.930 0.73 496.260 0.72

275

Table F.8: Melt viscosity data for the MXDA/PET blend preform samples Resin AA Scavenger Preform η* (Paxseconds) I.V. (dL/g) Concentration # @ 10 Calculated Overall (ppm) radians/second Average 1 632.750 0.76 574.410 0.74 200 2 627.680 0.76 0.76 672.820 0.77 3 704.760 0.77 CB12 634.310 0.76 1 818.820 0.80 560.270 0.74 100 2 741.200 0.78 0.77 604.420 0.75 3 979.010 0.82 633.530 0.76

276

Appendix G

Raw Data from Color Measurements

Table G.1: Color data for the Voridian CB12 PET control samples Number of Measured Average Sample Processing Y X% Z% # L a b L a b Times

1 75.0 -1.8 -2.4 0 2 74.4 -1.7 -2.2 74.8 -1.8 -2.3 56.0 55.2 58.4 3 75.0 -1.9 -2.3 1 68.5 -0.4 0.7 1 2 68.4 -0.4 0.7 68.5 -0.4 0.7 46.9 46.7 46.2 3 68.5 -0.3 0.7 1 71.5 -0.3 4.0 2 2 71.6 -0.5 3.9 71.5 -0.4 4.0 51.1 50.9 47.1 3 71.4 -0.4 4.0 1 71.1 -0.4 5.8 3 2 71.2 -0.4 5.6 71.2 -0.4 5.6 50.6 50.5 44.9 3 71.2 -0.5 5.5

277

Table G.2: Color data for the anthranilamide/PET blend samples AA Measured Average Scavenger Sample Y X% Z% Concentr- # L a b L a b ation (ppm) 1 67.0 -8.9 8.3 10,000 2 69.1 -8.9 8.1 68.5 -8.9 8.2 46.9 43.4 38.9 3 69.4 -8.8 8.3 1 72.5 -7.7 6.7 1200 2 72.1 -7.6 6.9 72.6 -7.6 6.8 52.7 49.6 45.7 3 73.2 -7.5 6.8 1 73.7 -7.2 7.4 500 2 73.9 -7.1 7.3 73.6 -7.2 7.4 54.2 51.1 46.4 3 73.2 -7.2 7.5 1 70.2 -10.3 2.9 200 2 69.9 -10.3 3.0 69.9 -10.3 3.0 48.8 44.7 45.9 3 69.5 -10.3 3.0 1 65.1 -10.1 2.6 100 2 66.3 -9.9 2.3 66.2 -10.0 2.4 43.8 40.0 41.5 3 67.1 -10.0 2.3

Table G.3: Color data for the MXDA/PET blend samples AA Measured Average Scavenger Sample Y X% Z% Concentr- # L a b L a b ation (ppm) 1 68.1 -10.7 11.3 10,000 2 68.8 -10.8 11.6 68.9 -10.8 11.6 47.5 43.2 36.1 3 69.9 -10.8 11.9 1 72.3 -9.6 14.5 1200 2 71.6 -9.2 14.7 71.7 -9.4 14.6 51.4 47.6 36.5 3 71.2 -9.4 14.6 1 73.7 -7.6 11.3 500 2 72.5 -7.6 11.5 73.5 -7.6 11.4 54.1 50.9 42.1 3 74.4 -7.6 11.4 1 68.3 -11.8 4.9 200 2 70.2 -11.7 4.8 69.0 -11.7 4.8 47.6 43.0 42.8 3 68.4 -11.6 4.8 1 68.9 -11.5 3.3 100 2 68.9 -11.4 3.3 69.0 -11.4 3.3 47.6 43.1 44.3 3 69.1 -11.4 3.3

278

Table G.4: Color data for the alpha-cyclodextrin/PET blend samples AA Measured Average Scavenger Sample Y X% Z% Concentr- # L a b L a b ation (ppm) 1 53.2 -36.7 12.8 50,000 2 52.0 -37.1 12.7 52.4 -36.9 12.8 27.5 16.4 17.9 3 52.0 -37.0 12.8 1 43.4 -47.6 12.7 25,000 2 43.8 -47.3 12.7 43.5 -47.3 12.7 18.9 7.2 11.0 3 43.3 -47.1 12.7 1 43.6 -47.7 12.3 10,000 2 43.7 -47.5 12.4 43.7 -47.6 12.4 19.1 7.2 11.4 3 43.8 -47.6 12.4 1 41.8 -44.7 10.5 5000 2 41.7 -44.6 10.6 41.8 -44.7 10.5 17.5 6.8 11.2 3 41.9 -44.7 10.5 1 58.6 -22.8 11.8 1200 2 59.1 -22.3 11.9 58.5 -22.6 11.9 34.2 26.7 24.3 3 57.8 -22.7 11.9 1 63.9 -11.7 10.0 500 2 64.3 -11.9 10.2 64.2 -11.9 10.1 41.2 36.8 31.9 3 64.3 -11.9 10.1 *The b value was so greatly altered that the resulting yellowness indexes were false values and therefore the data for the alpha-cyclodextrin/PET blend samples was not included in the results and discussion of this work.

279

Appendix H

Derivation of Thermal Energy Equation

 DT  ρ × C p   = ()()∇ • q − τ : ∇v (Equation 29)  Dt 

 DT  2 ρ × C p   = ()k∇ T − ()τ : ∇v (Equation 30)  Dt 

 DT   ∂T ∂T vθ ∂T ∂T  where:   =  + vr + + vz  (Equation 31)  Dt   ∂t ∂r r ∂θ ∂z 

 ∂ 2T 1 ∂T 1 ∂ 2T ∂ 2T  k∇ 2T = k + + +  where: () 2 2 2 2  (Equation 32)  ∂r r ∂r r ∂θ ∂z 

 ∂vr  1 ∂vr vθ   ∂vr  τ rr   +τ rθ  −  +τ rz    ∂r   r ∂θ r   ∂z 

 ∂vθ  1 ∂vθ vr   ∂vθ  +τθr   +τθθ  +  +τθz   where: ()τ : ∇v =  ∂r   r ∂θ r   ∂z  (Equation 33)

 ∂vz  1 ∂vz   ∂vz  +τ zr   +τ zrθ   +τ zz    ∂r   r ∂θ   ∂z 

280

 ∂T ∂T   1  ∂  ∂T   ∂vz  ρ × C p  + vz  =   k × r ×  +τ zr   (Equation 34)  ∂t ∂z   r  ∂r  ∂r   ∂r 

2  ∂vz   ∂vz  ∂vz   ∂vz  where: τ zr   = η   =η  (Equation 35)  ∂r   ∂r  ∂r   ∂r 

2  ∂T ∂T   1  ∂  ∂T   ∂vz  ρ × C p  + vz  =   k × r ×  +η  (Equation 36)  ∂t ∂z   r  ∂r  ∂r   ∂r 

281

Table H.1: Definition of Terms Listed in Equations 12 and 29 to 36 Terms Meaning ρ Density Cp Heat capacity T Temperature t Time ∇ Gradient q Heat generation v Velocity τ Shear stress k Boltzmann’s constant vr Velocity in the radial direction vθ Velocity in the rotation direction vz Velocity in the longitudinal direction r Radial direction θ Rotational direction z Longitudinal direction τrr Flux of radial momentum directed in the radial direction τrθ Flux of rotational momentum directed in the radial direction τrz Flux of longitudinal momentum directed in the radial direction τθr Flux of radial momentum directed in the rotational direction τθθ Flux rotational momentum directed in the rotational direction τθz Flux of longitudinal momentum directed in the rotational direction τzr Flux of radial momentum directed the longitudinal direction τzθ Flux of rotational momentum directed in the longitudinal direction τzz Flux of longitudinal momentum directed in the longitudinal direction η Viscosity

282

Appendix I

AA Generation 60 Minute Curve Study Data

Table I.1: AA Generation Data for the CB12 PET Resin Sample AA Scavenger Temperature Time AA Average Concentration (ppm) (oC) (minutes) (ppm) AA (ppm) 5 6.3 5.9 5 5.4 10 11.2 13.0 10 14.8 15 23.3 24.6 15 25.9 20 53.1 52.6 20 52.2 25 52.5 Extruded 54.6 25 56.6 Eastman 30 62.0 Voridian 66.7 0 30 CB12 290 71.4 35 70.8 PET 72.3 Resin 35 73.8 40 87.3 91.5 40 95.7 45 110.3 101.7 45 93.2 50 115.7 118.7 50 121.7 55 134.6 138.7 55 142.9 60 163.0 155.2 60 147.5

283

Table I.2: AA Generation Data for the 10,000 ppm Anthranilamide/PET Blend AA Scavenger AA Scavenger Temperature Time AA Average Concentration (oC) (minutes) (ppm) AA (ppm) (ppm) 5 1.4 1.0 5 0.6 10 3.0 2.6 10 2.2 15 8.9 8.6 15 8.3 20 12.1 11.3 20 10.5 25 13.5 10.4 25 7.3 30 16.4 16.6 Anthranilamide 10,000 30 290 16.8 35 22.9 20.4 35 17.8 40 19.3 20.9 40 22.5 45 22.0 23.8 45 25.6 50 28.8 22.5 50 16.2 55 22.2 23.6 55 24.9 60 20.9 22.2 60 23.4

284

Table I.3: AA Generation Data for the 10,000 ppm Alpha-Cyclodextrin/PET Blend AA Scavenger AA Scavenger Temperature Time AA Average Concentration (oC) (minutes) (ppm) AA (ppm) (ppm) 5 1.4 1.2 5 1.0 10 5.2 4.6 10 3.9 15 9.6 9.0 15 8.3 20 16.8 17.0 20 17.2 25 21.2 20.5 25 19.8 30 24.9 Alpha- 26.0 10,000 30 Cyclodextrin 290 27.0 35 30.4 29.2 35 28.1 40 33.5 34.4 40 35.2 45 38.6 36.5 45 34.4 50 41.2 41.7 50 42.2 55 39.0 42.8 55 46.6 60 42.4 43.4 60 44.4

285

Table I.4: AA Generation Data for the 10,000 ppm MXDA/PET Blend AA Scavenger AA Scavenger Temperature Time AA Average Concentration (oC) (minutes) (ppm) AA (ppm) (ppm) 5 3.2 3.5 5 3.8 10 12.5 9.6 10 6.8 15 24.7 23.8 15 22.8 20 34.7 36.6 20 38.5 25 39.8 45.9 25 52.1 30 54.0 53.9 MXDA 10,000 30 290 53.7 35 62.2 64.9 35 67.5 40 71.7 70.4 40 69.1 45 78.8 75.8 45 72.9 50 74.4 78.3 50 82.3 55 84.6 85.5 55 86.4 60 91.7 88.1 60 84.5

286

Appendix J

Data Used to Determine the k2 and a Values

Table J.1: Calculated Data Based on 60 Minute AA Generation Data, at 290oC, for the 10,000 ppm Anthranilamide/PET Blend Time Average d[AA]/dt RG – ln (RG – ln ([AA]) (minutes) AA (ppm) (d[AA]/dt) (d[AA]/dt)) 5 1.0 0.89 2.0 0.72 0.00 10 2.6 0.81 2.1 0.76 0.95 15 8.6 0.73 2.2 0.80 2.15 20 11.3 0.64 2.3 0.83 2.42 25 10.4 0.56 2.4 0.87 2.34 30 16.6 0.47 2.5 0.90 2.81 35 20.4 0.39 2.6 0.94 3.01 40 20.9 0.31 2.6 0.97 3.04 45 23.8 0.22 2.7 1.00 3.17 50 22.5 0.14 2.8 1.03 3.11 55 23.6 0.05 2.9 1.06 3.16 60 22.2 -0.03 3.0 1.09 3.10

287

Table J.2: Calculated Data Based on 60 Minute AA Generation Data, at 290oC, for the 10,000 ppm Alpha-Cyclodextrin/PET Blend Time Average d[AA]/dt RG – ln (RG – ln ([AA]) (minutes) AA (ppm) (d[AA]/dt) (d[AA]/dt)) 5 1.2 1.23 1.7 0.53 0.19 10 4.6 1.16 1.8 0.58 1.52 15 9.0 1.09 1.9 0.62 2.19 20 17.0 1.01 1.9 0.66 2.83 25 20.5 0.94 2.0 0.69 3.02 30 26.0 0.86 2.1 0.73 3.26 35 29.2 0.79 2.2 0.77 3.38 40 34.4 0.72 2.2 0.80 3.54 45 36.5 0.64 2.3 0.83 3.60 50 41.7 0.57 2.4 0.86 3.73 55 42.8 0.49 2.4 0.89 3.76 60 43.4 0.42 2.5 0.92 3.77 * The natural log of a value less than or equal to zero does not produce a real number

Table J.3: Calculated Data Based on 60 Minute AA Generation Data, at 290oC, for the 10,000 ppm MXDA/PET Blend Time Average d[AA]/dt RG – ln (RG – ln ([AA]) (minutes) AA (ppm) (d[AA]/dt) (d[AA]/dt)) 5 3.5 2.65 0.3 -1.22 1.26 10 9.6 2.45 0.5 -0.72 2.26 15 23.8 2.26 0.7 -0.39 3.17 20 36.6 2.07 0.9 -0.14 3.60 25 45.9 1.88 1.1 0.06 3.83 30 53.9 1.69 1.3 0.23 3.99 35 64.9 1.49 1.4 0.37 4.17 40 70.4 1.30 1.6 0.49 4.25 45 75.8 1.11 1.8 0.60 4.33 50 78.3 0.92 2.0 0.70 4.36 55 85.5 0.73 2.2 0.79 4.45 60 88.1 0.53 2.4 0.88 4.48

288

Appendix K

Data Used to Determine the k1, bb1, bb2, and b Values

] 0 ln 4.6 5.3 6.2 7.1 9.2 [S ) 2 k ) 2 * ln (k 0.0726 -2.0995 -2.1294 -0.8458

2 k 0.1225 0.1189 0.4292 1.0753 -0.2731 Calculated

a ] 0 1.3 1.3 1.3 1.3 1.2 [S C, Calculated for the Original ln( the Original C, Calculated for [AA] (ppm) o Average for Each

a 1.2 1.3 1.4 1.2 1.3 1.4 1.3 1.3 1.4 1.3 1.3 1.3 1.1 1.2 1.2 [AA] (ppm) equal to zero does not produce a real number a Value 0.1052 ) Plot ) Plot ] 0 6.9 6.6 8.5 3.7 5.4 6.1 [S 10.7 18.2 13.9 22.0 11.0 15.7 22.4 10.6 16.6 [AA] (ppm) Average 9 9 9 9 9 Anthranilamide/PET Blend Data, at 280 13 17 13 17 13 17 13 17 13 17 Versus ln( Time (min.) ] 0 100 200 500 [S 1200 (ppm) 10,000 Table K.1: * The natural log of a value less than or

289

] 0 ln ) 4.6 5.3 6.2 7.1 9.2 [S ] 0 [S ) 2 * * ln (k 0.5286 -1.4419 -0.2627 ) Versus ln( ) Versus 2 k

2 k 0.2365 0.7690 1.6965 -0.6312 -1.3058 Calculated for a ] (ppm) ] (ppm) 0 1.4 1.4 1.4 1.4 1.2 Each [S Average [AA] C, Calculated for the Original ln( the Original C, Calculated for o a 1.3 1.4 1.5 1.2 1.4 1.5 1.3 1.4 1.4 1.3 1.4 1.4 1.2 1.2 1.3 [AA] (ppm) equal to zero does not produce a real number a Value 0.1052 8.0 4.1 5.7 11.3 21.6 41.8 27.7 46.0 11.3 22.4 32.2 12.1 18.5 27.3 10.9 Average [AA] (ppm) 9 9 9 9 9 Anthranilamide/PET Blend Data, at 290 13 17 13 17 13 17 13 17 13 17 Plot Time (min.) ] 0 100 200 500 [S 1200 (ppm) 10,000 Table K.2: * The natural log of a value less than or

290

] 0 ln ) 4.6 5.3 6.2 7.1 9.2 [S ] 0 [S ) 2 ln (k 0.2930 1.2705 1.5336 -1.2144 -0.4855 ) Versus ln( ) Versus 2 k

2 k 0.2969 0.6154 1.3405 3.5627 4.6350 Calculated for a ] (ppm) ] (ppm) 0 1.5 1.5 1.5 1.4 1.3 Each [S Average [AA] C, Calculated for the Original ln( the Original C, Calculated for o

a 1.4 1.5 1.6 1.4 1.5 1.6 1.4 1.5 1.6 1.3 1.4 1.5 1.2 1.3 1.4 [AA] (ppm) a Value 0.1052 6.2 18.8 51.3 78.8 18.1 46.8 74.5 22.3 42.9 70.3 17.1 27.1 41.1 14.5 21.7 Average [AA] (ppm) 9 9 9 9 9 Anthranilamide/PET Blend Data, at 300 13 17 13 17 13 17 13 17 13 17 Plot Time (min.) ] 0 100 200 500 [S 1200 (ppm) 10,000 Table K.3:

291

] 0 ln 6.2 7.1 8.5 9.2 [S 10.1 10.8 ) 2 ) Versus ) Versus ln (k 0.0007 0.0989 0.0894 2 -1.8111 -0.4468 -0.0883 k

2 k 0.1635 0.6397 0.9155 1.0007 1.1039 1.0936 Calculated for a ] (ppm) ] (ppm) 0 1.3 1.3 1.2 1.2 1.2 1.2 C, Calculated for the Original ln( the Original for C, Calculated o Each [S Average [AA]

a 1.2 1.3 1.3 1.2 1.3 1.3 1.1 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 [AA] (ppm) a Value 0.1042 6.6 5.4 2.7 4.9 6.7 4.1 5.3 7.1 3.9 5.3 6.1 4.7 5.8 6.8 13.3 17.5 11.7 12.0 Average [AA] (ppm) ) Plot ) Plot ] 0 [S 9 9 9 9 9 9 Alpha-Cyclodextrin/PET Blend Data, at 280 Alpha-Cyclodextrin/PET 13 17 13 17 13 17 13 17 13 17 13 17 ln( Time (min.) ] 0 500 [S 1200 5000 (ppm) 10,000 25,000 50,000 Table K.4:

292

] 0 ln 6.2 7.1 8.5 9.2 [S 10.1 10.8 ) 2 * * ) Versus ) Versus ln (k 0.3810 0.5656 0.5483 2 -0.1024 k

2 k 0.9027 1.4637 1.7605 1.7303 -0.8163 -0.1552 Calculated for a ] (ppm) ] (ppm) 0 1.4 1.3 1.3 1.2 1.2 1.3 C, Calculated for the Original ln( the Original for C, Calculated o Each [S Average [AA]

a 1.3 1.4 1.5 1.2 1.3 1.4 1.2 1.3 1.4 1.2 1.2 1.3 1.2 1.2 1.3 1.2 1.3 1.3 [AA] (ppm) equal to zero does not produce a real number a Value 0.1042 7.9 5.7 4.3 7.2 5.9 7.8 6.5 8.7 10.3 22.5 42.2 14.6 33.1 11.3 20.0 13.4 11.9 12.7 Average [AA] (ppm) ) Plot ) Plot ] 0 [S 9 9 9 9 9 9 Alpha-Cyclodextrin/PET Blend Data, at 290 Alpha-Cyclodextrin/PET 13 17 13 17 13 17 13 17 13 17 13 17 ln( Time (min.) ] 0 500 [S 1200 5000 (ppm) 10,000 25,000 50,000 Table K.5: * The natural log of a value less than or

293

] 0 ln 6.2 7.1 8.5 9.2 [S 10.1 10.8 ) 2 ) Versus ) Versus ln (k 0.2261 1.0128 1.0033 1.1450 1.2789 2 -0.6389 k

2 k 0.5279 1.2537 2.7534 2.7272 3.1425 3.5927 Calculated for a ] (ppm) ] (ppm) 0 1.5 1.4 1.4 1.4 1.4 1.4 C, Calculated for the Original ln( the Original for C, Calculated o Each [S Average [AA]

a 1.4 1.5 1.6 1.3 1.4 1.5 1.3 1.4 1.5 1.2 1.4 1.5 1.2 1.4 1.5 1.3 1.4 1.5 [AA] (ppm) a Value 0.1042 9.7 6.7 7.7 19.3 42.6 77.0 14.3 34.5 63.7 20.7 43.3 19.5 41.9 18.4 37.6 11.4 19.8 35.9 Average [AA] (ppm) ) Plot ) Plot ] 0 [S 9 9 9 9 9 9 Alpha-Cyclodextrin/PET Blend Data, at 300 Alpha-Cyclodextrin/PET 13 17 13 17 13 17 13 17 13 17 13 17 ln( Time (min.) ] 0 500 [S 1200 5000 (ppm) 10,000 25,000 50,000 Table K.6:

294

] 0 ln 4.6 5.3 6.2 7.1 9.2 [S ) 2 ) Plot ] 0 * [S ln (k -3.9876 -4.4850 -3.0393 -2.0277

2 k ) Versus ln( ) Versus 2 0.0185 0.0113 0.0479 0.1316 k -0.0306 Calculated for a ] (ppm) ] (ppm) 0 6.1 5.4 5.5 4.5 5.2 Each [S Average [AA]

a 4.4 6.3 7.6 4.0 5.5 6.8 4.2 5.2 7.0 2.8 5.1 5.7 4.1 5.7 5.9 C, Calculated for the Original ln( the Original C, Calculated for [AA] (ppm) o equal to zero does not produce a real number a Value 0.6324 8.9 9.6 5.0 9.5 10.4 18.4 24.5 14.8 20.7 13.4 21.9 13.2 15.8 15.6 16.5 Average [AA] (ppm) 9 9 9 9 9 MXDA/PET Blend Data, at 280 13 17 13 17 13 17 13 17 13 17 Time (min.) ] 0 100 200 500 [S 1200 (ppm) 10,000 Table K.7: * The natural log of a value less than or

295

] 0 ln 4.6 5.3 6.2 7.1 9.2 [S ) 2 ) Plot ] 0 * [S ln (k -2.9295 -2.4697 -2.4900 -2.2343

2 k ) Versus ln( ) Versus 2 0.0534 0.0846 0.0829 0.1071 k -0.2053 Calculated for a ] (ppm) ] (ppm) 0 8.2 7.2 7.0 6.4 6.6 Each [S Average [AA]

a 5.2 7.4 5.0 7.5 9.1 4.8 7.4 8.7 4.6 6.0 8.7 4.5 6.9 8.4 11.9 C, Calculated for the Original ln( the Original C, Calculated for [AA] (ppm) o equal to zero does not produce a real number a Value 0.6324 13.5 23.8 50.5 12.6 24.0 33.3 12.0 23.7 30.6 11.2 17.1 30.5 10.9 21.2 28.8 Average [AA] (ppm) 9 9 9 9 9 MXDA/PET Blend Data, at 290 13 17 13 17 13 17 13 17 13 17 Time (min.) ] 0 100 200 500 [S 1200 (ppm) 10,000 Table K.8: * The natural log of a value less than or

296

] 0 ln 4.6 5.3 6.2 7.1 9.2 [S ) 2 ) Plot ] 0 [S ln (k -2.2544 -1.7061 -1.3925 -1.2627 -1.2499

2 k ) Versus ln( ) Versus 2 0.1049 0.1816 0.2484 0.2829 0.2865 k Calculated for a ] (ppm) ] (ppm) 0 11.2 10.3 10.4 10.5 10.0 Each [S Average [AA]

a 6.4 6.3 6.8 7.2 6.7 12.1 15.1 10.2 14.3 10.4 13.9 10.4 13.7 10.0 13.5 C, Calculated for the Original ln( the Original C, Calculated for [AA] (ppm) o equal to zero does not produce a real number a Value 0.6324 18.7 51.3 73.1 18.5 39.3 67.4 20.5 40.8 63.9 22.8 40.8 63.0 20.0 37.9 61.4 Average [AA] (ppm) 9 9 9 9 9 MXDA/PET Blend Data, at 300 13 17 13 17 13 17 13 17 13 17 Time (min.) ] 0 100 200 500 [S 1200 (ppm) 10,000 Table K.9: * The natural log of a value less than or

297

Table K.10: Multiple Linear Regression Data Used to Determine the b, bb1, and bb2 Values for the Anthranilamide/PET Blends ln (k2) ln ([S0]) Temperature Temperature 1/Temperature (oC) (oK) (1/oK) 0.0726 9.21 -0.8458 7.09 280 553.15 0.001808 -2.1294 6.21 -2.0995 4.61 0.5286 9.21 -0.2627 7.09 290 563.15 0.001776 -1.4419 6.21 1.5336 9.21 1.2705 7.09 0.2930 6.21 300 573.15 0.001745 -0.4855 5.30 -1.2144 4.61

Table K.11: Multiple Linear Regression Data Used to Determine the b, bb1, and bb2 Values for the Alpha-Cyclodextrin/PET Blends ln (k2) ln ([S0]) Temperature Temperature 1/Temperature (oC) (oK) (1/oK) 0.0894 10.82 0.0989 10.13 0.0007 9.21 280 553.15 0.001808 -0.0883 8.52 -0.4468 7.09 -1.8111 6.21 0.5483 10.82 0.5656 10.13 290 563.15 0.001776 0.3810 9.21 -0.1024 8.52 1.2789 10.82 1.1450 10.13 1.0033 9.21 300 573.15 0.001745 1.0128 8.52 0.2261 7.09 -0.6389 6.21

298

Table K.12: Multiple Linear Regression Data Used to Determine the b, bb1, and bb2 Values for the MXDA/PET Blends ln (k2) ln ([S0]) Temperature Temperature 1/Temperature (oC) (oK) (1/oK) -2.0277 9.21 -3.0393 7.09 280 553.15 0.001808 -4.4850 6.21 -3.9876 5.30 -2.2343 9.21 -2.4900 7.09 290 563.15 0.001776 -2.4697 6.21 -2.9295 5.30 -1.2499 9.21 -1.2627 7.09 -1.3925 6.21 300 573.15 0.001745 -1.7061 5.30 -2.2544 4.61

299

Appendix L

Raw Data from Modeling Program

– 0.419 0.419 0.401 0.385 0.364 0.333 0.169 0.323 0.282 0.196 0.152 0.117 0.106 0.418 0.417 0.417 0.416 0.414 Max G Initial (ppm) [AA] [AA] 0.363 0.363 0.349 0.334 0.316 0.289 0.149 0.282 0.246 0.172 0.136 0.104 0.088 0.362 0.361 0.361 0.360 0.358 Avg. 1.919 1.919 1.901 1.885 1.864 1.833 1.669 1.823 1.782 1.696 1.652 1.617 1.606 1.918 1.917 1.917 1.916 1.914 Max AA (ppm) 1.863 1.863 1.849 1.834 1.816 1.789 1.649 1.782 1.746 1.672 1.636 1.604 1.588 1.862 1.861 1.861 1.860 1.858 Avg.

4 Æ 1.919 1.919 1.901 1.885 1.864 1.833 1.669 1.823 1.782 1.696 1.652 1.617 1.606 1.918 1.917 1.917 1.916 1.914

3 G 1.887 1.887 1.871 1.855 1.836 1.808 1.657 1.799 1.762 1.682 1.643 1.607 1.588 1.886 1.885 1.885 1.884 1.882 [AA] 2 1.822 1.822 1.811 1.797 1.781 1.758 1.634 1.752 1.720 1.654 1.624 1.596 1.579 1.822 1.821 1.820 1.820 1.818 1 AA in Cavities (ppm) 1.822 1.822 1.811 1.797 1.781 1.758 1.634 1.752 1.720 1.654 1.624 1.596 1.579 1.822 1.821 1.820 1.820 1.818 C o 0 AA 100 200 500 500 100 200 500 1200 1200 5000 1200 (ppm) 10,000 10,000 25,000 50,000 10,000 tration Concen- Scavenger Predicted AA Generation Results for a 24 Cavity Injection Molding Process, Process, Molding a 24 Cavity Injection Results for AA Generation Predicted (AA from Resin and Extruder) = 1.5 ppm, Each Cavity Represents a Set of 4 Symmetric Cavities Cavities 4 Symmetric of Set a Represents Cavity Each ppm, 1.5 = Extruder) and Resin from (AA Alpha- Initial MXDA Sample / CB12 PET Scavenger Cyclodextrin Anthranilamide Table L.1: Modeled at 280 [AA]

300

– C 0.466 0.466 0.441 0.430 0.408 0.375 0.216 0.359 0.322 0.237 0.180 0.141 0.120 0.465 0.465 0.465 0.464 0.463 Max G o Initial (ppm) [AA] [AA] 0.379 0.379 0.359 0.350 0.333 0.307 0.184 0.294 0.265 0.199 0.154 0.119 0.102 0.378 0.378 0.378 0.377 0.376 Avg. 1.966 1.966 1.941 1.930 1.908 1.875 1.716 1.859 1.822 1.737 1.680 1.641 1.620 1.965 1.965 1.965 1.964 1.963 Max AA (ppm) 1.879 1.879 1.859 1.850 1.833 1.807 1.684 1.794 1.765 1.699 1.654 1.619 1.602 1.878 1.878 1.878 1.877 1.876 Avg. 6

1.966 1.966 1.941 1.930 1.908 1.875 1.716 1.859 1.822 1.737 1.680 1.641 1.620 1.965 1.965 1.965 1.964 1.963 G 5 [AA] 1.966 1.966 1.941 1.930 1.908 1.875 1.716 1.859 1.822 1.737 1.680 1.641 1.620 1.965 1.965 1.965 1.964 1.963 Æ 4 ity Represents a Set of 4 Symmetric Cavities Injection Molding Process, Modeled at 280 1.821 1.821 1.804 1.797 1.783 1.762 1.663 1.751 1.727 1.674 1.637 1.603 1.590 1.820 1.820 1.820 1.820 1.819 3 1.821 1.821 1.804 1.797 1.783 1.762 1.663 1.751 1.727 1.674 1.637 1.603 1.590 1.820 1.820 1.820 1.820 1.819 2 1.850 1.850 1.831 1.823 1.807 1.784 1.674 1.773 1.746 1.686 1.645 1.612 1.597 1.849 1.849 1.848 1.848 1.847 AA in Cavities (ppm) 1 1.850 1.850 1.831 1.823 1.807 1.784 1.674 1.773 1.746 1.686 1.645 1.612 1.597 1.849 1.849 1.848 1.848 1.847 0 100 200 500 500 100 200 500 1200 1200 5000 1200 (ppm) 10,000 10,000 25,000 50,000 10,000 AA Scavenger Concentration Predicted AA Generation Results for a 24 Cavity Results for AA Generation Predicted (AA from Resin and Extruder) = 1.5 ppm, Each Cav Initial Alpha- MXDA Sample / CB12 PET Scavenger Cyclodextrin Anthranilamide [AA] Table L.2:

301

– 0.384 0.384 0.363 0.354 0.336 0.310 0.172 0.297 0.267 0.195 0.148 0.120 0.109 0.383 0.382 0.382 0.382 0.381 Max G Initial (ppm) [AA] [AA] 0.348 0.348 0.329 0.322 0.306 0.282 0.159 0.271 0.244 0.179 0.138 0.110 0.097 0.347 0.347 0.347 0.346 0.345 Avg. C o 1.884 1.884 1.863 1.854 1.836 1.810 1.672 1.797 1.767 1.695 1.648 1.620 1.609 1.883 1.882 1.882 1.882 1.881 Max AA (ppm) 1.848 1.848 1.829 1.822 1.806 1.782 1.659 1.771 1.744 1.679 1.638 1.610 1.597 1.847 1.847 1.847 1.846 1.845 Avg. 8 1.848 1.848 1.829 1.822 1.806 1.782 1.659 1.771 1.744 1.679 1.638 1.609 1.595 1.847 1.847 1.847 1.846 1.845 7

1.848 1.848 1.829 1.822 1.806 1.782 1.659 1.771 1.744 1.679 1.638 1.609 1.595 1.847 1.847 1.847 1.846 1.845 G 6 [AA] 1.812 1.812 1.796 1.789 1.775 1.754 1.646 1.744 1.721 1.664 1.627 1.602 1.589 1.811 1.811 1.811 1.811 1.810 Æ 5 1.812 1.812 1.796 1.789 1.775 1.754 1.646 1.744 1.721 1.664 1.627 1.602 1.589 1.811 1.811 1.811 1.811 1.810 ity Represents a Set of 4 Symmetric Cavities Injection Molding Process, Modeled at 280 4 1.884 1.884 1.863 1.854 1.836 1.810 1.672 1.797 1.767 1.695 1.648 1.620 1.609 1.883 1.882 1.882 1.882 1.881 3 1.884 1.884 1.863 1.854 1.836 1.810 1.672 1.797 1.767 1.695 1.648 1.620 1.609 1.883 1.882 1.882 1.882 1.881 AA in Cavities (ppm) 2 1.848 1.848 1.829 1.822 1.806 1.782 1.659 1.771 1.744 1.679 1.638 1.609 1.595 1.847 1.847 1.847 1.846 1.845 1 1.848 1.848 1.829 1.822 1.806 1.782 1.659 1.771 1.744 1.679 1.638 1.609 1.595 1.847 1.847 1.847 1.846 1.845 0 AA 100 200 500 500 100 200 500 1200 1200 5000 1200 (ppm) 10,000 10,000 25,000 50,000 10,000 tration Concen- Scavenger Predicted AA Generation Results for a 32 Cavity Results for AA Generation Predicted (AA from Resin and Extruder) = 1.5 ppm, Each Cav Initial Alpha- MXDA Sample / CB12 PET Scavenger Cyclodextrin Anthranilamide Table L.3: [AA]

302

0.460 0.435 0.424 0.402 0.370 0.203 0.354 0.317 0.229 0.173 0.146 0.127 0.459 0.459 0.459 0.458 0.457 – Max Max G Initial (ppm) (ppm) [AA] [AA] 0.379 0.359 0.350 0.333 0.307 0.174 0.294 0.265 0.194 0.150 0.119 0.098 0.379 0.378 0.378 0.378 0.377 Avg. Avg. 1.960 1.935 1.924 1.902 1.870 1.703 1.854 1.817 1.729 1.673 1.646 1.627 1.959 1.959 1.959 1.958 1.957 C Max Max o AA (ppm) (ppm) AA 1.879 1.859 1.850 1.833 1.807 1.674 1.794 1.765 1.694 1.650 1.619 1.598 1.879 1.878 1.878 1.878 1.877 Avg. Avg. 12 1.907 1.885 1.876 1.857 1.828 1.684 1.815 1.782 1.706 1.658 1.621 1.592 1.906 1.906 1.906 1.905 1.904 11 1.907 1.885 1.876 1.857 1.828 1.684 1.815 1.782 1.706 1.658 1.621 1.592 1.906 1.906 1.906 1.905 1.904 10 1.824 1.807 1.800 1.786 1.764 1.653 1.753 1.729 1.670 1.633 1.606 1.588 1.823 1.823 1.823 1.823 1.822

9 G 1.824 1.807 1.800 1.786 1.764 1.653 1.753 1.729 1.670 1.633 1.606 1.588 1.823 1.823 1.823 1.823 1.822 8 [AA] 1.880 1.860 1.851 1.834 1.808 1.674 1.795 1.765 1.695 1.650 1.617 1.596 1.880 1.879 1.879 1.879 1.878 Æ 7 1.880 1.860 1.851 1.834 1.808 1.674 1.795 1.765 1.695 1.650 1.617 1.596 1.880 1.879 1.879 1.879 1.878 ity Represents a Set of 4 Symmetric Cavities Injection Molding Process, Modeled at 280 6 1.960 1.935 1.924 1.902 1.870 1.703 1.854 1.817 1.729 1.673 1.646 1.627 1.959 1.959 1.959 1.958 1.957 5 1.960 1.935 1.924 1.902 1.870 1.703 1.854 1.817 1.729 1.673 1.646 1.627 1.959 1.959 1.959 1.958 1.957 AA in Cavities (ppm) (ppm) Cavities in AA 4 1.824 1.807 1.800 1.786 1.764 1.653 1.753 1.729 1.670 1.633 1.606 1.588 1.823 1.823 1.823 1.823 1.822 3 1.824 1.807 1.800 1.786 1.764 1.653 1.753 1.729 1.670 1.633 1.606 1.588 1.823 1.823 1.823 1.823 1.822 2 1.880 1.860 1.851 1.834 1.808 1.674 1.795 1.765 1.695 1.650 1.617 1.596 1.880 1.879 1.879 1.879 1.878 1 1.880 1.860 1.851 1.834 1.808 1.674 1.795 1.765 1.695 1.650 1.617 1.596 1.880 1.879 1.879 1.879 1.878 0 100 100 200 500 500 100 200 500 1200 1200 5000 1200 (ppm) (ppm) 10,000 10,000 25,000 50,000 10,000 AA Scavenger AA Scavenger Concentration Predicted AA Generation Results for a 48 Cavity Results for AA Generation Predicted (AA from Resin and Extruder) = 1.5 ppm, Each Cav Initial amide Alpha- MXDA Sample / Anthranil- CB12 PET CB12 Scavenger Cyclodextrin Cyclodextrin Table L.4: [AA]

303

0.168 0.158 0.153 0.144 0.131 0.063 0.108 0.088 0.041 0.015 0.010 0.008 0.168 0.168 0.168 0.168 0.167 – Max Max G Initial (ppm) (ppm) [AA] [AA] 0.139 0.130 0.127 0.119 0.109 0.054 0.090 0.074 0.035 0.014 0.009 0.007 0.139 0.139 0.139 0.139 0.138 Avg. Avg. 1.668 1.658 1.653 1.644 1.631 1.563 1.608 1.588 1.541 1.515 1.510 1.508 1.668 1.668 1.668 1.668 1.667 C Max Max o AA (ppm) (ppm) AA 1.639 1.630 1.627 1.619 1.609 1.554 1.591 1.574 1.535 1.514 1.509 1.507 1.639 1.639 1.639 1.639 1.638 Avg. Avg. 12 1.649 1.640 1.636 1.628 1.616 1.557 1.597 1.579 1.537 1.514 1.509 1.507 1.649 1.649 1.649 1.649 1.648 11 1.649 1.640 1.636 1.628 1.616 1.557 1.597 1.579 1.537 1.514 1.509 1.507 1.649 1.649 1.649 1.649 1.648 10 1.618 1.611 1.608 1.602 1.593 1.547 1.578 1.564 1.531 1.513 1.508 1.506 1.618 1.618 1.618 1.618 1.618

9 G 1.618 1.611 1.608 1.602 1.593 1.547 1.578 1.564 1.531 1.513 1.508 1.506 1.618 1.618 1.618 1.618 1.618 8 [AA] 1.639 1.631 1.627 1.620 1.609 1.554 1.591 1.574 1.536 1.514 1.509 1.507 1.639 1.639 1.639 1.639 1.639 Æ 7 1.639 1.631 1.627 1.620 1.609 1.554 1.591 1.574 1.536 1.514 1.509 1.507 1.639 1.639 1.639 1.639 1.639 ity Represents a Set of 4 Symmetric Cavities Injection Molding Process, Modeled at 270 6 1.668 1.658 1.653 1.644 1.631 1.563 1.608 1.588 1.541 1.515 1.510 1.508 1.668 1.668 1.668 1.668 1.667 5 1.668 1.658 1.653 1.644 1.631 1.563 1.608 1.588 1.541 1.515 1.510 1.508 1.668 1.668 1.668 1.668 1.667 AA in Cavities (ppm) (ppm) Cavities in AA 4 1.618 1.611 1.608 1.602 1.593 1.547 1.578 1.564 1.531 1.513 1.508 1.506 1.618 1.618 1.618 1.618 1.618 3 1.618 1.611 1.608 1.602 1.593 1.547 1.578 1.564 1.531 1.513 1.508 1.506 1.618 1.618 1.618 1.618 1.618 2 1.639 1.631 1.627 1.620 1.609 1.554 1.591 1.574 1.536 1.514 1.509 1.507 1.639 1.639 1.639 1.639 1.639 1 1.639 1.631 1.627 1.620 1.609 1.554 1.591 1.574 1.536 1.514 1.509 1.507 1.639 1.639 1.639 1.639 1.639 0 100 100 200 500 500 100 200 500 1200 1200 5000 1200 (ppm) (ppm) 10,000 10,000 25,000 50,000 10,000 AA Scavenger AA Scavenger Concentration Predicted AA Generation Results for a 48 Cavity Results for AA Generation Predicted (AA from Resin and Extruder) = 1.5 ppm, Each Cav Initial amide Alpha- MXDA Sample / Anthranil- CB12 PET CB12 Scavenger Cyclodextrin Cyclodextrin Table L.5: [AA]

304

1.256 1.195 1.170 1.118 1.041 0.648 1.074 1.011 0.836 0.724 0.636 0.594 1.251 1.250 1.249 1.247 1.242 – Max Max G Initial (ppm) (ppm) [AA] [AA] 1.036 0.988 0.968 0.927 0.866 0.557 0.892 0.843 0.700 0.611 0.539 0.496 1.032 1.031 1.031 1.029 1.025 Avg. Avg. 2.756 2.695 2.670 2.618 2.541 2.148 2.574 2.511 2.336 2.224 2.136 2.094 2.751 2.750 2.749 2.747 2.742 C Max Max o AA (ppm) (ppm) AA 2.536 2.488 2.468 2.427 2.366 2.057 2.392 2.343 2.200 2.111 2.039 1.996 2.532 2.531 2.531 2.529 2.525 Avg. Avg. 12 2.610 2.558 2.536 2.491 2.425 2.088 2.454 2.400 2.246 2.149 2.070 2.010 2.606 2.605 2.604 2.603 2.598 11 2.610 2.558 2.536 2.491 2.425 2.088 2.454 2.400 2.246 2.149 2.070 2.010 2.606 2.605 2.604 2.603 2.598 10 2.387 2.348 2.331 2.297 2.247 1.996 2.269 2.229 2.109 2.034 1.977 1.956 2.384 2.383 2.383 2.382 2.378

9 G 2.387 2.348 2.331 2.297 2.247 1.996 2.269 2.229 2.109 2.034 1.977 1.956 2.384 2.383 2.383 2.382 2.378 8 [AA] 2.537 2.489 2.469 2.428 2.367 2.058 2.393 2.344 2.201 2.111 2.038 1.979 2.533 2.533 2.532 2.530 2.526 Æ 7 2.537 2.489 2.469 2.428 2.367 2.058 2.393 2.344 2.201 2.111 2.038 1.979 2.533 2.533 2.532 2.530 2.526 ity Represents a Set of 4 Symmetric Cavities Injection Molding Process, Modeled at 290 6 2.756 2.695 2.670 2.618 2.541 2.148 2.574 2.511 2.336 2.224 2.136 2.094 2.751 2.750 2.749 2.747 2.742 5 2.756 2.695 2.670 2.618 2.541 2.148 2.574 2.511 2.336 2.224 2.136 2.094 2.751 2.750 2.749 2.747 2.742 AA in Cavities (ppm) (ppm) Cavities in AA 4 2.387 2.348 2.331 2.297 2.247 1.996 2.269 2.229 2.109 2.034 1.977 1.956 2.384 2.383 2.383 2.382 2.378 3 2.387 2.348 2.331 2.297 2.247 1.996 2.269 2.229 2.109 2.034 1.977 1.956 2.384 2.383 2.383 2.382 2.378 2 2.537 2.489 2.469 2.428 2.367 2.058 2.393 2.344 2.201 2.111 2.038 1.979 2.533 2.533 2.532 2.530 2.526 1 2.537 2.489 2.469 2.428 2.367 2.058 2.393 2.344 2.201 2.111 2.038 1.979 2.533 2.533 2.532 2.530 2.526 0 100 100 200 500 500 100 200 500 1200 1200 5000 1200 (ppm) (ppm) 10,000 10,000 25,000 50,000 10,000 AA Scavenger AA Scavenger Concentration Predicted AA Generation Results for a 48 Cavity Results for AA Generation Predicted (AA from Resin and Extruder) = 1.5 ppm, Each Cav Initial amide Alpha- MXDA Sample / Anthranil- CB12 PET CB12 Scavenger Cyclodextrin Cyclodextrin Table L.6: [AA]

305

3.368 3.218 3.153 3.023 2.829 1.805 3.036 2.916 2.628 2.430 2.282 2.191 3.346 3.341 3.335 3.327 3.298 – Max Max G Initial (ppm) (ppm) [AA] [AA] 2.781 2.661 2.609 2.507 2.352 1.537 2.516 2.421 2.192 2.036 1.913 1.841 2.763 2.760 2.754 2.748 2.726 Avg. Avg. 4.868 4.718 4.653 4.523 4.329 3.305 4.536 4.416 4.128 3.930 3.782 3.691 4.846 4.841 4.835 4.827 4.798 Max Max C o AA (ppm) (ppm) AA 4.281 4.161 4.109 4.007 3.852 3.037 4.016 3.921 3.692 3.536 3.413 3.341 4.263 4.260 4.254 4.248 4.226 Avg. Avg. 12 4.477 4.347 4.291 4.180 4.011 3.127 4.190 4.087 3.838 3.667 3.518 3.472 4.458 4.454 4.446 4.442 4.418 11 4.477 4.347 4.291 4.180 4.011 3.127 4.190 4.087 3.838 3.667 3.518 3.472 4.458 4.454 4.449 4.442 4.418 10 3.887 3.788 3.746 3.661 3.533 2.857 3.669 3.590 3.401 3.272 3.179 3.121 3.873 3.870 3.866 3.861 3.843 9

3.887 3.788 3.746 3.661 3.533 2.857 3.669 3.590 3.401 3.272 3.179 3.121 3.873 3.870 3.866 3.861 3.843 G 8 [AA] 4.282 4.162 4.110 4.007 3.852 3.037 4.017 3.922 3.692 3.536 3.411 3.319 4.264 4.261 4.255 4.249 4.227 Æ 7 4.282 4.162 4.110 4.007 3.852 3.037 4.017 3.922 3.692 3.536 3.411 3.319 4.264 4.261 4.255 4.249 4.227 ity Represents a Set of 4 Symmetric Cavities Injection Molding Process, Modeled at 300 6 4.868 4.718 4.653 4.523 4.329 3.305 4.536 4.416 4.128 3.930 3.782 3.691 4.846 4.841 4.835 4.827 4.798 5 4.868 4.718 4.653 4.523 4.329 3.305 4.536 4.416 4.128 3.930 3.782 3.691 4.846 4.841 4.835 4.827 4.798 AA in Cavities (ppm) (ppm) Cavities in AA 4 3.887 3.788 3.746 3.661 3.533 2.857 3.669 3.590 3.401 3.272 3.179 3.121 3.873 3.870 3.866 3.861 3.843 3 3.887 3.788 3.746 3.661 3.533 2.857 3.669 3.590 3.401 3.272 3.179 3.121 3.873 3.870 3.866 3.861 3.843 2 4.282 4.162 4.110 4.007 3.852 3.037 4.017 3.922 3.692 3.536 3.411 3.319 4.264 4.261 4.255 4.249 4.227 1 4.282 4.162 4.110 4.007 3.852 3.037 4.017 3.922 3.692 3.536 3.411 3.319 4.264 4.261 4.255 4.249 4.227 0 100 100 200 500 500 100 200 500 1200 1200 5000 1200 (ppm) (ppm) 10,000 10,000 25,000 50,000 10,000 AA Scavenger AA Scavenger Concentration Predicted AA Generation Results for a 48 Cavity Results for AA Generation Predicted (AA from Resin and Extruder) = 1.5 ppm, Each Cav Initial amide Alpha- MXDA Sample / Anthranil- CB12 PET CB12 Scavenger Cyclodextrin Cyclodextrin Table L.7: [AA]

306