INFORMATION TO USERS

The most advanced technology has been used to photo­ graph and reproduce this manuscript from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are re­ produced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. These are also available as one exposure on a standard 35mm slide or as a 17" x 23" black and white photographic print for an additional charge.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

University Microfilms International A Bell & Howell Information Com pany 300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA 313/761-4700 800/521-0600 Order Number 9002034

Reactive processing of polyureas and polyurethane-polyester hybrids

Wang, Kuan-Jong, Ph.D.

The Ohio State University, 1989

UMI 300 N. ZeebRd. Ann Arbor, MI 48106 REACTIVE PROCESSING OF POLYUREAS AND

POLYURETHANE-POLYESTER HYBRIDS

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of the Ohio State University

by

Kuan-Jong Wang, B.S., M.S.

*****

The Ohio State University

1989

Dissertation Committee: Approved by

Dr. L. James Lee

Dr. Jacques L. Zakin Adviser Dr. Shang-Tian Yang A Department of Chemical Engineering Dr. James D. Cawley To

my parents

and my wife Yin-Yiin

ii ACKNOWLEDGMENTS

I wish to thank my advisor, Dr. L. James Lee, for his assistance and guidance in the completion of this work. Thanks go to the other members of my dissertation committee, Dr. Jacques L. Zakin, Dr. Shang-

Tian Yang, and Dr. James D. Cawley, for their suggestions and comments.

Also, I would like to express appreciation to Mr. Michael B.

Kukla, departmental designer, and Mr. Roy Renshaw, departmental instrument designer, for their contribution to this study. Mr. Ou

Jiang, whom I have worked with in the RIM project, is gratefully acknowledged.

To my wife, Yin-Yiin, I offer sincere thanks for her understanding, sacrifice, and encouragement associated with this work.

She also typed the entire manuscript with great patience. Finally, I would like to thank my parents for their confidence and encouragement through all the years of my education. VITA

October 17, 1956 ------Bom, Kaohsiung, Taiwan, Republic of China

June, 1980 ------B.S., Chemical Engineering, National Tsing-Hua University

1980 - 1982------Military Service

1982 - 1983------Teaching Associate, Soochow University, Taiwan

June, 1985 ------M.S., Chemical Engineering, The Ohio State University

1983 - present ------Research Associate and Teaching Associate, The Ohio State University

PUBLICATIONS

1. "Rheological Behavior of Dispersed Multiphase Polymer Melts," K.J. Wang, M.S. Thesis, The Ohio State University, June (1985).

2. "Rheological and Extrusion Behavior of Dispersed Multiphase Polymeric Systems," K.J. Wang and L.J. Lee, J. Appl. Folyro. Sci.. 33, 431 (1987).

3. 1'Structure-Property-Processing Relationships of Polyurethane- Polyester Interpenetrating Polymer Networks," K.J. Wang, T.J. Hsu, and L.J. Lee, SPE ANTEC. 46. 507 (1988); Polym. Eng. Sci.. 29. 397 (1989).

4. "Rheo-Kinetic Studies in Reaction Injection Molding of Polyureas," K.J. Wang, Y.J. Huang, and L.J. Lee, SPE ANTEC. 47, 572 (1989).

iv FTRIDS OF STIJDY

1. Polymer Rheology

2. Polymer Chemistry and PhyBics

3. Extrusion Process

4. Reaction Injection Molding Process

v TABLE OF CONTENTS

DEDICATION ...... ii

ACKNOWLEDGMENTS ...... iii

VITA ...... iv

LIST OF TABLES ...... x

LIST OF FIGURES ...... xi

CHAPTER Page

CHAPTER I INTRODUCTION ...... 1

1.1 Reactive Polymer Processing...... 1

1.2 Scope of the Present Work ...... 9

CHAPTER II REACTIVE POLYMER PROCESSING ...... 18

2.1 Processes ...... 18 2.1.1 Reaction Injection Molding ...... 19 2.1.2 Transfer Molding ...... 29 2.1.3 Resin Transfer Molding ...... 36

2.2 Material Chemistry and Reaction Kinetics ...... 38 2.2.1 Polyurethanes ...... 38 2.2.2 Polyurethane-ureas/Polyureas ...... 40 2.2.3 Nylons ...... 42 2.2.4 Epoxies ...... 45 2.2.5 Interpenetrating Polymer Networks ...... 48

2.3 Rheological Changes ...... 50

2.4 Microstructure and Property Relations ...... 55 2.4.1 Structure and Property ...... 57 2.4.1.1 Intermolecular Interactions ...... 58 2.4 1.2 Molecular Structure ...... 61 2.4.1.3 Molecular Weight and Crosslinking ... 65

vi 2.4.2 Phase Separation ...... 69 2.4.2.1 Dynamic Mechanical Measurement ..... 71 2.4.2.2 Stress-Strain Measurement ...... 74 2.4.2.3 Differential Scanning Calorimetry ... 75 2.4.2.4 Infrared Spectroscopy ...... 78 2.4.2.5 Small Angle X-ray Scattering ...... 83 2.4.3 Morphology and Domain Structure ...... 84 2.4.3.1 Hard Segment Structure ...... 85 2.4.3.2 Polyurethane and Polyurea Morphology...... 87

CHAPTER III REACTION INJECTION MOLDING OF POLYUREAS: I. RHEO- KINETIC STUDIES IN SOLUTION POLYMERIZATION ...... 91

3.1 Previous Work on Kinetics, Rheology, and Phase Formation of Polyurea Reaction ...... 91 3.1.1 Chemical Systems ...... 91 3.1.2 Reaction Kinetics ...... 94 3.1.3 Rheological Changes ...... 97 3.1.4 Phase Formation ...... 100

3.2 Experimental ...... 104 3.2.1 Materials ...... 104 3.2.2 Instrumentation and Procedure...... 108 3.2.2.1 Haake Viscometer ...... 108 3.2.2.2 Fourier Transform Infrared Spectroscopy (FTIR) ...... 109 3.2.2.3 Rheoroetrics Dynamic Analyzer...... Ill

3.3 Results and Discussion ...... Ill 3.3.1 Influence of Solvent ...... 113 3.3.2 Rheological Measurements ...... 115 3.3.3 Kinetic Measurements ...... 125 3.3.4 Gelation and its Dependence onConcentration and Temperature ...... 132 3.3.4.1 Solvent Concentration Effect...... 138 3.3.4.2 Reaction Temperature Effect ...... 143 3.3.5 Moduli G* & G" Changes in Urea and Urethane Reactions ...... 148

3.4 Summary and Conclusions ...... 151

CHAPTER IV REACTION INJECTION MOLDING OF POLYUREAS: II.BULK POLYMERIZATIONS AND SIMULATION ...... 159

4.1 Introduction...... 159

4.2 Experimental ...... 160

vii 4.2.1 Material ...... 160 4.2.2 Instrumentation and Procedure...... 161 4.2.2.1 Reaction Injection Molding ...... 161 4.2.2.2 Adiabatic Temperature Rise Measurement ...... 164 4.2.2.3 Viscosity Rise Measurement ...... 167

4.3 Results and Discussion ...... 169 4.3.1 Reaction Injection Molding of Polyurea ..... 169 4.3.2 Rheo-kinetic Models for Polyurea Reaction .... 175 4.3.2.1 Reaction Kinetics ...... 175 4.3.2.2 Energy Balance ...... 176 4.3.2.3 Rheological Changes .... 178 4.3.3 Estimation of Model Parameters ..... 180 4.3.3.1 Kinetic Parameters ...... 180 4.3.3.2 Rheological Parameters ...... 187 4.3.3.2.1 Aliphatic Triamine/ Diisocyanate Reaction .... 187 4.3.3.2.2 Aromatic Diamine/ Diisocyanate Reaction .... 193 4.3.4 Model Prediction ...... 195

4.4 Summary and Conclusions ...... 198

CHAPTER V STRUCTURE-PROPERTY-FROCESSING RELATIONSHIPS OF POLYURETHANE-POLYESTER INTERPENETRATING POLYMER NETWORKS ...... 200

5.1 Previous Work on Morphology and Physical Properties of IPNs ...... 200 5.1.1 Interpenetrating Polymer Netwoeks (IPNs) ..... 200 5.1.1.1 General I P N s ...... 200 5.1.1.2 Polyurethane-based IPNs ...... 207 5.1.2 Morphology of IPNs ...... 209 5.1.2.1 Compatibility ...... 210 5.1.2.2 Crosslinking ...... 212 5.1.2.3 Synthetic Process and Condition ... 214 5.1.2.4 Composition ...... 216 5.1.3 Glass Transitions and MechanicalProperties of IPNs ...... 218

5.2 Experimental ...... 228 5.2.1 Materials ...... 228 5.2.1.1 Materials Used for Transfer Molding . 228 5.2.1.2 Materials Used for Reaction Injection Molding (RIM) ...... 232 5.2.2 Sample Preparation ...... 233 5.2.2.1 Transfer M o l d i n g ...... 233 I viii 5.2.2.2 RIM ...... 235 5.2.3 Methods of Analysis ...... 235 5.2.3.1 Transmission Electron Microscopy (TEM) ...... 235 5.2.3.2 Dynamic MechanicalAnalysis (DMA) ... 235

5.3 Results ...... 237 5.3.1 Effect of Various Processes...... 238 5.3.2 Effect of Molding Temperature and IPN Composition ...... 242 5.3.3 Effect of Postcure ...... 247 5.3.4 Effect of Initiator ...... 250

5.4 Discussion and Modeling ...... 253 5.4.1 Interpenetration and Continuity of Phases .... 253 5.4.2 Modeling of Modulus Behavior...... 256

5.5 Summary and Conclusions ...... 263

CHAPTER VI REOOMMEMDATIONS ...... 266

REFERENCES ...... 269

APPENDIX A FTIR MACRO Programs ...... 287

APPENDIX B ACSL Programs for Rheo-kinetic Modelling of Polyurea Reaction ...... 308

APPENDIX C Derivation of a Mechanical Model for 120 °C-Transfer Molded IPN Based on Takayanagi’s Model II ...... 312

ix LIST OF TABLES

TABLE PAGE

1.1 Reactive polymer processing...... 4

1.2 Revolution of RIM materials based on the isocyanate/polyether systems ...... 11

1.3 Experimental scheme ...... 17

2.1 Molar cohesive energies for common groups in urethane and urea systems ...... 59

3.1 Materials used in polyurea systems ...... 105

4.1 Parameters used for modelling of polyureareaction (T5000/ I1305/DWTDA) ...... 186

5.1 Materials used for transfer molding ...... 229

x LIST OF FIGURES

FIGURE PAGE

1.1 A schematic diagram of the reactive processing of thermoset polymers ...... 5

1.2 Effect of microscopic changes of reactive resin on the performance of each unit operation in processing...... 7

1.3 Research scheme of reactive polymer processing in this s t u d y ...... 15

2.1 Schematic diagram of the reaction injection molding process ...... 20

2.2 Comparison of energy consumption for RIM process and other material manufacturing processes (Sweeney, 1987) ...... 22

2.3 Basic components of a RIM processing equipment (Ngugen, 1984) ...... 24

2.4 Schematics of commercial mixing heads: (a) Krauss-Maffei and EMB, (b) Henneke, (c) Battenfeld, (d) Cincinnati Milacron (cross-section), (e) Cannon and Accuratio (Sweeney, 1979) ... 26

2.5 Schematic diagram of the transfer molding process ...... 30

2.6 Integral-pot transfer molding: (a) mold open, pot loaded; (b) mold closed; and (c) mold open, parts ejected, sprue on force plug (Hull, 1984) ...... 33

2.7 Plunger transfer molding: (a) mold closed, pot loaded; (b) mold closed, plunger down; and (c) mold open, molded part, cull, and runners ejected (Hull, 1984) ...... 35

2.8 Schematic diagram of the resin transfer molding process 37

2.9 Nylon block polymer segment (Sweeney, 1987) ...... 43

2.10 Changes in viscosity and modulus during the RIM cycle (Lee, 1987) ...... 51

xi 2.11 Schematic representation of the segmented polyurethane microstructure (Camargo, 1984) ...... 56

2.12 Trend of melting points in homologous series of polymers (Hill and Walker, 1948) ...... 60

2.13 Effect of molecular weight on the physical properties of a polymer (Saunders, 1960) ...... 66

2.14 Tensile modulus vs. molecular weight/crosslink (calculated) (Pigott et al., 1960) ...... 68

2.15 Typical dynamic mechanical response for block copolymers with different degrees of phase compatibility (Camargo, 1984) ...... 72

2.16 Hydrogen bondings of N-H group with different groups in polyurethanes and polyureas ...... 80

3.1 Chemical structures of aromatic diamines: DETDA, TBTDA, and DMTDA ...... 106

3.2 Schematic diagram of polyurea reaction ...... 112

3.3 Effect of solvent on the viscosity rise andconversion in the reaction of I1305/DWTDA (r=l) at 25 °C ...... 114

3.4 Viscosity rises of 85%-diluted T5000/I1305 and 11305/diamine polyureas at temperature 25 °C ...... 116

3.5 Viscosity rises of DETDA-based polyureas in 85% nitrobenzene at temperature 25 °C ...... 119

3.6 Viscosity rises of TBTDA-based polyureas in 85% nitrobenzene at temperature 25 °C ...... 121

3.7 Viscosity rises of EWTDA-based polyureas in 85% nitrobenzene at temperature 25 °C ...... 122

3.8 Viscosity rises of DETDA-, TBTDA-, DMTDA-, and DETDA/DMTDA(50/50)-based polyureas with segment ratio soft/hard=70/30, in 85% NB, at 25

3.9 Viscosity rises of DETDA/DMTDA(50/50)-based polyureas with segment ratio soft/hard=70/30, 50/50, and 20/80, in 85% NB, at 25 °C ...... 126

3.10 FTIR spectra of I1305/DMIDA polyurea reactions in 85% NB at 25 °C ...... 127

3.11 Comparison of chemical structures and IR spectra between TBT and TBTDA ...... 129

3.12 Conversion vs. time plots of TBTDA-based polyureas in 85% NB at 25 °C ...... 131

3.13 Conversion vs. time plots of DMTDA-based polyureas in 85% NB at 25 °C ...... 133

3.14 Effect of chain extender on the reaction rate and limiting conversion of polyurea reaction ...... 134

3.15 Mapping procedure to obtain viscosity as a function of conversion in reactive polymermixtures (Castro, 1980) ..... 135

3.16 Viscosity vs. conversion plots of TBTDA-based polyureas in 85% NB at 25 °C ...... 137

3.17 Viscosity vs. conversion plots of DMIDA-based polyureas in 85% NB at 25 °C ...... 139

3.18 Effect of chain extender on the gel conversion of polyurea reaction ...... 140

3.19 Viscosity and conversionvs. time plots of I1305/DMTDA polyureas in 85, 75, and65% NB at 25 °C...... 141

3.20 Viscosity vs. time and conversion plots of I1305/DMTDA polyureas in 85, 75, and 65% NB at 25 °C...... 142

3.21 Viscosity and conversion vs. time plots of DMIDA-based polyureas with segment ratio soft/hard=70/30 in 85, 75, and 65% NB at 25 °C ...... 144

3.22 Viscosity vs. time and conversion plots of DMTDA-based polyureas with segment ratio soft/hard=70/30 in 85, 75, 70, and 65% NB at 25 °C ...... 145

3.23 Viscosity and conversion vs. time plots of I1305/DMIDA polyureas in 85% NB at 25, 47, and 67 °C ...... 146

3.24 Viscosity vs. time and conversion plots of I1305/DMIDA polyureas in 85% NB at 25, 47, and 67 °C ...... 147

3.25 Viscosity vs. time plots of DMIDA-based polyureas in 85% NB at 25, 47, and 67 °C ...... 149

xiii 3.26 Viscosity, G ’, and G" vs. time for T0305/I143L and I143L/DMTDA reactions in 80% NB at 33 °C (legend symbols are same as in Figure 3.27)...... 150

3.27 Viscosity, G*, and G" vs. conversion for T0305/I143L and I143L/DMTDA reactions in 80% NB at 33 °C ...... 152

3.28 Viscosity, G ’, and G" vs. time for T5000/I143L/DWTDA reaction in 80% NB at 23... °C ..... 153

3.29 Viscosity, G ’, and G" vs. conversion for T5000/I143L/DMTDA reaction in 80% NB at 23... °C ...... 154

3.30 Viscosity, G*, and G" vs. time for T0305/I143L/BDO reaction in 80% NB at 23 °C ...... 155

3.31 Viscosity, G', and G" vs. conversion for T0305/I143L/BDO reaction in 80% NB at 23... °C ...... 156

4.1 Schematic diagram of the laboratory scale RIM machine (Nelson, 1987) ...... 162

4.2 Schematic diagram of data acquisition system (Nelson, 1987) . 163

4.3 Schematic diagram of the "Free-Table rheometer" ...... 168

4.4 Adiabatic temperature and viscosity rises of hard segment reaction measured by RIM ...... 170

4.5 Viscosity vs. conversion for the same data shown in Figure 4.4 ...... 171

4.6 Adiabatic temperature and viscosity rise rise in polyurea RIM reaction ...... 173

4.7 Viscosity vs. conversion for the same data shown in Figure 4.6 ...... 174

4.8 Conversion vs. time measured by FTIR for I1305/DWTDA reaction in 85% NB ...... 181

4.9 Plots of logtdflCg/dt) vs. logfl-oCg) ...... 183

4.10 Plot of G2 vs. 1 / T ...... 185

4.11 Viscosity as a function of %NB for the reaction of T5000 and 11305 ...... 188

xiv 4.12 Viscosity vs. 1/T for the same data shown in Figure 4.11 .... 189

4.13 The effect of r ratio on the viscosity rise of the reaction of T5000 and 11305 ...... 191

4.14 Viscosity vs. shear rate for the reaction of T5000/I1305/NB at different weight ratios ...... 192

4.15 Viscosity vs. conversion for hard segment in solution polymerization (85% NB) ...... 194

4.16 A vs. a* for hard segment in solution polymerization (85% NB) ...... 196

4.17 A vs. a'/ociT for the same data shown in Figure 4.16 ...... 197 £

5.1 A comparison of various polymeric composites ...... 202

5.2 Synthesis of IPNs ...... 204

5.3 Four Takayanagi's models for two-phase polymer system; A: plastic, B: rubber ...... 223

5.4 Three-phase model of a composite ...... 225

5.5 Schematic diagram showing the reaction mechanism of PU/PES IPN ...... 230

5.6 Schematic diagram of the transfer mold; (A) before molding, (B) after molding (Hsu, 1987) ...... 234

5.7 Schematic diagram of dynamic mechanical analysis system 236

5.8 Transmission electron micrographs of 120 °C-molded IPNs (PU/PES=50/50) by (a) RIM, and (b) transfer m o l d i n g ...... 240

5.9 G* and tan& vs. temperature for IPNs by RIM and transfer molding ...... 241

5.10 Transmission electron micrographs of 8 0 °C-transfer molded PU/PES IPNs with compositions of (a) 75/25, (b) 50/50, and (c) 25/75 ...... 243

5.11 G ’ and tanS vs. temperature for 8 0 °C-transfer molded IPNs with various compositions ...... 244

5.12 Transmission electron micrographs of 120°C-transfer molded

xv PU/PES IPNs with compositions of (a) 75/25, (b) 50/50, and (c) 25/75 ...... 246

5.13 G ’ and tani vs. temperature for 120 °C-transfer molded IPNs with various compositions ...... 248

5.14 G ’ and tanS vs. temperature for 8 0 °C-transfer molded IPNs before and after 6-hr postcure ...... 249

5.15 Transmission electron micrographs of 8 0 °C-transfer molded IPNs (PU/PES=50/50) initiated by FDO at (a) 1.38% and by MEKP/amine /Co-8 at (b) 0.67%/0.22%/0.22%, and (c) 2.0%/0.67%/0.67% ...... 251

5.16 G ’ and tanS vs. temperature for 8 0 °C-transfer molded IPNs initiated by various initiators ...... 252

5.17 G* vs. composition plots at 2 0 °C, (a) 120°C-transfer molding, (b) 8 0 °C-transfer molding ...... 255

5.18 Comparison of model prediction and experimental results of G* and tanS for 120°C-transfer molded IPN (FU/PES=25/75), V ^ O . l , Vc=0.82 ...... 258

5.19 Comparison of model prediction and experimental results of G ’ and tani for 120°C-transfer molded IPN (PU/PES=50/50), VgC=0.5, Vc=0.8 ...... 260

5.20 Comparison of model prediction and experimental results of G ’ and tanS for 8 0 °C-transfer molded IPN (PU/PES=50/50), V^rO.5, Vc=0.8 ...... 262

5.21 Comparison of model prediction and experimental results of G ’ and tani for 80 °C-transfer molded IPN (PU/PES=75/25), Vfi=0.75 ...... 264

C.l Mechanical model based on Takayanagi’s model II for 120°C- transfer molded IPN (PU/PES=50/50) ...... 314

C.2 Simplified mechanical model for 120°C-transfer molded IPN (FU/PES=50/50) ...... 315

xvi CHAPTER I

INTRODUCTION

SYNOPSIS

The objectives of this study and the motivation behind this work are stated in this chapter. The goal of this thesis is to analyze the reaction kinetics, rheological changes and structure formation of two new materials, polyurea and polyurethane-based IPN, in reactive polymer processing. Introduction to the reactive polymer processing is described first followed by the outline of the research scheme.

1.1 REACTIVE POLYMER PROCESSING

Thermosets and thermoplastics, both offsprings of polymer technology, have been nurtured in one diverse environment, that of a marketplace seeking innovative, cost-effective products with functionalities that would meet everyday requirements and enhance lifestyles. In many ways, the crosslinked thermosets can be viewed as the elder technology reliable in countless applications where levels of temperature and corrosion resistance, load-carrying ability, and comparatively low cost were the proven virtues, and product aesthetics were almost an affectation. Then the thermoplastics sparked a revolution of new plastic applications answering an increasing need

1 for greater productivity and a much more aggressive demand for high appearance levels. With advancements in reactive polymer processing, one can now also combine thermoplastic with thermosetting elements to optimize the product. For example, with the marriage of thermoset and thermoplastic technologies in a reactive extrusion, it has been demonstrated that a material with a low crosslinked density can be reconstituted to a useful new polymer. In essence, the tremendous changes in reactive chemistry and processing are wiping out the easy categorization of thermoplastics and thermosets. Also, it is now possible to provide different performance profiles in a single part.

With polyurethanes, for example, through use of multiple mixing heads containing independent compositions and ratios of polyols and isocyanates, different properties can be selected and controlled.

Therefore, the last two decades of polymer research has focused more on the improvement of processing technology and the modifications of existing polymers than on the development of new polymers. Among the newly developed processes, reactive polymer processing has been proved to be an efficient, productive, and energy-saving process.

Also, reactive processing with innovations both in material tailorability and productivity has moved the thermosets into an increasingly competitive position. Their use is one of the fastest growing areas in the plastics industry. In the meantime, modifications of existing polymers have attracted a great deal of research attention since the existing polymers do not seem to possess all the desired properties. The advent of polymeric composite materials such as fiber-

reinforced plastics (FRP) and interpenetrating polymer networks (IPN) are among the recent developments in the polymer industry.

The operation of reactive polymer processing innovatively involves polymerization and fabrication in a single step, i.e., the polymer is

formed after the monomer mixture is in the desired shap>e. Today, more and more polymer products are produced through reactive polymer processing. The applications of reactive polymer processing in various areas are summarized in Table 1.1. Sheet molding compound (SMC) and bulk molding compound (BMC), in which unsaturated polyester and styrene are usually the primary components, are two of the well-known products which are processed by this technology. Other examples include compression molding of rubber and reaction injection molding of polyurethane. The former is a popular process in the rubber industry; the latter is a relatively new area with continuously growing potential. One of the major application of reactive polymer processing is in the electronic industry where electric charge plates are encapsulated by thermosetting polymers like epoxies and polyurethanes.

Bulk state, less viscous resin, high reaction rate, high exotherm and fast cycle time are the characteristics of most reactive polymer processing processes. In most cases, physical properties of the finished products depend not only on raw materials used and the microstructure of product, but also on the reaction kinetics and rheological change during processing, as shown in Figure 1.1. In each Table 1.1 Reactive Polymer Processing

Process Resin

Reaction Injection Molding Polyurethane, Nylon Epoxy, Polyester

Transfer Molding Thennosets

Compression Molding Sheet Molding Compound Bulk Molding Compound

Injection Molding Bulk Molding Compound

Electric Encapsulation Hiennosets

Casting,Potting Acrylate, Epoxy Embedding Nylon REACTIVE PROCESSING OF TI-IERMOSET POLYMERS

Resins Rheo — Kinetics Curing

Processing Microstructu re Physical Conditions F ormation Properties

Figure 1.1 A schematic diagram of the reactive processing of thermoset polymers process, raw materials go through a series of unit operations such as mixing, mold filling, and mold curing. The performance of each unit operation is governed by the microscopic changes of the reactive resin, which include reaction kinetics, molecular diffusion, rheological changes and flow pattern as shown in Figure 1.2. For example, reaction rate may influence viscosity rise, which then affects the flow pattern in the mixing step and the mold filling stage. Flow pattern and molecular diffusion may determine the reaction rate and the final conversion in the curing stage.

Basically the rheo-kinetics of reactive resin involves both chemical reactions and physical changes. The chemical reactions involved can be categorized as chain growth polymerization (styrenes, unsaturated polyesters, etc.) and step growth polymerization (epoxies, polyurethanes, etc.) Hie physical changes involved include phase separation and inversion, gelation, crystallization, and glass transition, which play influential roles in determining the properties of the finished product. Phase separation and inversion are induced by domain formation due to thermodynamic incompatibility of the resin components. Gelation takes place when polymer resin changes from a viscous fluid to a network structure with chemical or physical crosslinking. Crystallization occurs when the molecular chains of polymer are arranged in a patterned order. Glass transition occurs when polymer changes from the rubbery state to the glassy state.

In some cases, the reactive resin contains two or more chemical reactions which may influence each other. The physical changes during EFFECT OF MICROSCOPIC CHANGES ON PROCESSING

Reaction Molecular Rheological Flow Kinetics Diffusion Changes Pattern

Mold Mold Mixing Fil ling Curing

Figure 1.2 Effect of microscopic changes of reactive resin on the performance of each unit operation in processing polymerization and processing further complicate the chemical

reactions. These interactions will affect the properties of the

finished product. For example, phase separation and inversion can

influence the mechanical properties of polymers such as urethane

elastomers because mechanical properties are dominated by the matrix

phase of polymers and the degree of phase separation. In chain growth polymerization, the gel effect can cause the thermal runaway problem due to excessive temperature rise from rapid conversion. In addition, the chemical reactions become diffusion-controlled after gelation.

Although crystallization makes the polymer stronger, it may impose a processing problem such as that in the reaction injection molding of nylon. The glass transition effect can freeze a polymer reaction and make polymerization incomplete because molecular diffusion becomes much more difficult in the glassy state than in the rubbery or fluid state. As a result, only limited conversion can be reached.

Polyurethanes, polyesters, phenolics and epoxies are major resins used in reactive polymer processing, such as reaction injection molding, transfer molding and compression molding. By the selection of ingredients it is possible to formulate different kinds of raw resins.

Hie selection usually depends on product specifications and the process applied. Catalysts, initiators and inhibitors are added to control the reaction rate. Fillers like calcium carbonate, mica, glass fiber are often combined into the reactive system to enhance the mechanical properties. Other components such as low profile agent

(e.g., PVAc for sheet molding compound) and foaming agent (e.g., water and methylene chloride for polyurethane) are added for different purposes. In developing reactive polymer processing technologies,

research efforts in resin compounding have paralleled the effort in process design and modifications.

1.2 SCOPE OF THE PRESENT WORK

More than a decade of active industrial research has centered on developing polymeric material systems for the reaction injection molding (RIM) process. The success'of RIM as a fabrication technique for producing polyurethane and polyurethane-urea elastomers is evident by its growth during the last decade. The first polymers to be used were polyurethane materials. Polyurethane systems contain polyether polyols, glycol chain extenders and catalysts; the isocyanate component is MDI(4,4*-diphenylmethane diisocyanate)-based. These ingredients lead to certain limitations, such as entrapping air bubbles during mold filling, long cycle time, difficulty in demolding, and poor dimensional stability of parts at high temperatures. Now, polyurethane materials above are not commonly used in RIM processing.

They were replaced by polyurethane-urea RIM materials. Polyurethane- urea systems consist of polyether polyols, diamine chain extenders such as diethyl-toluenediamine (DETDA), and catalysts; the isocyanates are aromatic polyisocyanates. The use of aromatic diamine to substitute for the diol chain extender reduces overall cycle time, and the resulting lower labor cost and higher productivity increase the competitiveness of RIM materials against other polymeric materials 10 such as SMC and engineering thermoplastics. Catalysts are still necessary to increase the reactivity of polyol/isocyanate to a level similar to that of amine/isocyanate. These RIM materials still have some processing limitations similiar to those of polyurethanes: poor mold release and poor dimensional stability at high temperatures.

Recently, polyureas, difficult to process by injection molding, have come into their own as a new material for the evolving RIM industry. Polyurea systems consist of polyether polyamines, diamine chain extenders such- as DETDA, tert-butyl-toluenediamine (TBTDA) dimethylthio-toluenediamine (DWTDA), and MDI-based isocyanates. The evolution of RIM material technology is sumnaried in Table 1.2. It is important to note that polyurea RIM technology uses polyether polyamines instead of polyether polyol, and no catalysts. Catalyst- induced degradation can be eliminated due to the absence of catalysts for urea polymerization. One of the critical characteristics of polyurea RIM materials is their excellent mold release. Polyether polyamines are good solvents for internal release agents such as acidic dimethyl-siloxanes and zinc stearate (Grigsby et al., 1987).

Internal mold release agents do not interfere with the urea reaction, so it is possible to eliminate external mold release agents for these systems (Dominguez et al., 1983, Dominguez; 1985). Rapid viscosity increase in urea systems prevents excess air entrapment during the mold-filling stage of RIM cycle (Dominguez, 1985). The polyurea polymers do not require postcure, yet yield thermal properties that are better than polyurethanes containing as much as 20% reinforcing 11

Table 1.2 Revolution of RIM Materials Based on the iBocyanate/Polyether Systems

Generation Polymer Isocyanate Polyether

First Polyurethane MDI-based Glycol chain extenders Ethylene oxide-capped polyols Catalyst

Second Polyurethane- MDI-based Aromatic diamine chain urea extenders Ethylene oxide-capped polyols Catalyst

New Polyurea MDI-based Aromatic diamine chain extenders Polyether polyamines

* Another new generation of RIM material is the polyurethane-based IPN 12 fillers (Even, 1985). In a property comparison between polyureas and polyurethane-ureas, the former have higher thermal stability and low water absorption than the latter (Dominguez, 1985; Even, 1985). Better thermal properties in the ureas may be important for implementation of on-line painting or E-coating of polyurea body panels in automotive construction (Hemphill and Vanderhider, 1987). Two other key properties, namely excellent impact resistance and high-temperature dimensional stability are characteristics of polyurea RIM materials.

Due to its superior thermal and mechanical properties mentioned above, polyurea is a desirable RIM material. However, the system reacts extremely fast with some of the reactions taking place during mixing and mold filling. In many cases, this presents processing difficulties such as insufficient mixing, premature gelation and low conversion. To avoid this problem, a thorough understanding of the reaction kinetics,

Theological changes, and structure formation is required. With this information in hand, formulating and mold design improvements can be made more readily.

The reaction kinetics of polyurea has been studied by several groups (Pannone and Macosko, 1985, 1987, 1988; Vespoli et al., 1985,

1986). Previous effort in our group (Hsu and Lee, 1988) indicated that the reaction rate of polyurea in solution polymerization was greatly reduced compared to the reaction in bulk polymerization. The kinetic parameters were determined by FTER data from solution polymerization.

Combined with a heat transfer model, the data successfully predicted the adiabatic temperature rise of bulk polyurea reaction in RIM. In 13

this study, experimental and theoretical investigations were

thoroughly conducted to understand the reaction kinetics, Theological change, phase separation and their interactions in polymerization and processing of polyurea. These investigations were carried out using both solution polymerization and bulk polymerization in RIM.

Interpenetrating polymer network (IPN) based on a polyurethane and an unsaturated polyester is another new material (Table 1.2) developed recently for RIM process. Basically, this IPN is a modification of polyurethane RIM material, in which the second reactive polymer

(unsaturated polyester) is introduced into the polyurethane reaction to make up the deficiencies of the existing material. The addition of glassy polyester to polyurethane can internally reinforce the elastomeric properties of polyurethane so that structural applications for automobiles are possible. The dual reactions in an IPN system also offer some advantages in processing. For example, the addition of a less viscous material to the urethane resin can reduce the resin viscosity and, consequently, facilitate the mixing and mold filling.

Furthermore, a mixing-activated step growth polymerization, such as polyurethane, can be used as an internal heat source to initiate a thermally-activated chain growth polymerization. Most IPNs have been developed for slow processes such as coating and casting. For fast processes like RIM, there are only a few commercially available IPN compounds. Ashland Chemical developed an acrylamate polymer (Wilkinson et al., 1983; Kelly, 1986) that is basically a polyurethane with unsaturation on the polyol chain, which forms a second network with a 14 crosslinking agent, acrylic monomer. Amoco Chemical developed a series of polyurethane-polyester hybrids which can be used in various reactive processes (Edwards, 1986). Most research efforts on IPNs have been on the synthesis method, morphology and mechanical properties

(Allen et al., 1973 and thereafter; Frisch et al., 1974 and 1975;

Sperling, 1985). Little is known about the structure-property and processing-property relations on these materials. There is a need for some detailed studies from processing aspects on IPN systems to gain a fundamental understanding of them. Additionally, the IPN polymerization route represents an unique way of blending two crosslinked polymers with a minimal degree of covalent bond and phase separation between the networks (Matsuo et al., 1970; Frisch et al.,

1974).

In this study, experimental and theoretical investigations were conducted to figure out the structure-property-processing relationships of polyurethane/polyester IPNs. The IPN samples were molded by both RIM and transfer molding.

To achieve the goal of this thesis, a comprehensive research scheme is designed in an effort to obtain a better understanding of the polyurea and polyurethane/polyester IPN applied in reactive polymer processing. Resin composition, reaction kinetics, theological changes, moldability, morphology and mechanical properties are the major factors to be studied. These factors and their interactive relationships are shown in Figure 1.3. The reaction kinetics were followed by differential scanning calorimetry and Fourier transfer 15

KINETICS COMPOSITION RHEOLOGY

MORPHOLOGY MECHANICAL MOLDABILITY PROPERTY

THEORETICAL MODEL

Figure 1.3 Research scheme of reactive polymer processing in this study 16

infrared spectroscopy. Hie rheological changes were investigated using

Haake, Brookfield and "free-table" viscometers. The "free-table" viscometer was designed and built in our laboratory. The moldability

studies were carried out on lab-scale RIM and transfer molding machines. The morphology and dynamic mechanical properties were analyzed using transmission electron microscopy, the Weissenberg

Rheogoniometer and the Rheometrics Dynamic Analyzer. These experimental techniques are listed in Table 1.3 and discussed in detail in later chapters.

Based on the above mentioned research scheme, this thesis is broken down into six chapters. Chapter 1 gives the objectives of this study and the motivation behind this work. Chapter 2 outlines the basics of RIM, transfer molding, and RTM processes, including process equipment, polymeric systems either in current use or under active research, reaction kinetics, rheological changes, microstructure formation and properties. Chapter 3 deals with rheo-kinetics of polyureas in solution polymerizations. Emphasis is placed on the interactions of reaction rate and viscosity rise for soft and hard segments. Chapter 4 is designed to study the bulk polymerization of polyurea in RIM both experimentally and theoretically. A rheo-kinetic model is proposed to simulate the polyurea RIM process. Chapter 5 presents the analysis of structure-property and processing-property relations of polyurethane/polyester IPN. Mechanical models are developed to explain the domain structure and phase inversion.

Finally, recommendations are given in Chapter 6. Table 1.3 Experimental Scheme

Kinetics DSC, FTIR

Rheology Haake Viscometer Brookfield Viscometer "Free-table" Viscometer

Molding Lab-scale RIM Machine Lab-scale Transfer Mold

Morphology TEN

Mechanical Rheometrics Dynamic Analyzer Properties Weissenberg Rheogoniometer Instron Tensile Test Machine CHAPTER II

REACTIVE POLYMER PROCESSING

SYNOPSIS

The relevant literature on reaction injection molding (RIM) and transfer molding is reviewed in this chapter. Following a discussion of process equipment and processing parameters encountered in RIM and transfer molding, the emphasis is placed on materials, processing variables, properties, and their interactions in the RIM process.

2.1 PROCESSES

In reactive polymer processing, there are many types of processes such as: reaction injection molding of polyurethane, nylon, and polyester; resin transfer molding of thermosetting polymers; compression molding of sheet molding compound, casting of acrylate sheet, and electronic encapsulation of thermosetting polymers. They, generally, can be categorized into two typos: fast processes such as reaction injection molding, and slow processes such as resin transfer molding and casting. Only reaction injection molding and transfer molding are considered in this study.

18 19

4 2.1.1 REACTION INJECTION MOLDING

Reaction injection molding (RIM) is a process in which two or more

liquid intermediates are metered separately to a mixing head where

they are combined by high-pressure impingement mixing and subsequently

flow into a mold where they polymerize to form a molded part. A

schematic diagram of the RIM process is shown in Figure 2.1. Detailed

review of the RIM process is available elsewhere (Sweeney 1979, 1987;

Lee, 1980; Macosko, 1983, 1988). Since the first commercialization of

RIM in 1974, the major developments of RIM technology have been in the

automotive industry. Front and rear bumper fascia covers were first

produced for several General Motors’ model vehicles in 1975 (Lee,

1980; Poole, 1985). Fenders, spoilers, trunk lids, and door were manufactured on limited model vehicles using RIM technology in recent years (Wood, 1976; Raia, 1977; McQuiston, 1979; Lee, 1980; Poole,

1985; Wigotsky, 1986). Although automotive components comprise the

largest market, a wide variety of other applications are growing rapidly such as foams (Sweeney, 1987).

RIM has three major characteristics: low pressures, low temperatures, and the use of reactive liquid intermediates.

Low pressures: Although the pressures in the mixing head can be as high as 3500 p.s.i., the in-mold pressures are significantly lower than those in other molding processes. When comparing a typical RIM in-mold pressure of 50 p.s.i. with the 5000 p.s.i. or more required for thermoplastic injection molding, it becomes apparent why RIM is particularly suitable for larger parts. Automotive bumpers are ratio mixer control

isocyanate

Figure 2.1 Schematic diagram of the reaction injection molding process co © 21

routinely produced on RIM presses with 100 to 150 tons of clamping

force, while comparable injection molded parts require presses of 3500

tons or more.

Low temperatures: The temperatures used in RIM are also

significantly lower. With polyurethanes, the intermediates normally

are processed at temperatures between 75 and 120 °F and the mold

temperature is usually between 130 and 170°F. These lower temperatures

obviously require significantly less energy consumption.

Liquid intermediates: The use of low viscosity liquid

intermediates has additional benefits beyond the low pressures and

temperatures involved. A tremendous amount of design flexibility is

possible with RIM. Since the mold is filled with low viscosity liquid,

very complex part configurations can be produced. Ribs, mounting bosses, slots, and cut-out areas are all possible. RIM parts are being molded with wall sections as thin as 0.1 in. and as thick as 1.5 in.

Furthermore, moldings can incorporate variations in thickness within the same part. Incorporation of inserts for mounting or reinforcement

is also practical. Since the mold is filled before polymerization occurs, there are no molded-in stresses to cause part warping or cracking after demold.

Because of its low temperature and pressure requirements, the energy cost of a RIM product can be considerable lower than that of a typical thermoplastic part. Figure 2.2 shows energy costs for making

RIM parts compared with energy costs for making parts in other ways.

It can be seen that RIM is indeed an energy saving process in the 22

| | m RIM/Baydur polyurethane

H LDPE

HDPE

Polystyrene | | Feedstock PVC K Fuel ABS

Polypropylene

Acrylic

Polycarbonate

Polyester

Nylon 66

Nylon 6

Modified PPO

Acetal D ie-cast aluminum X X 1,000 2.000 3,000 4,000 5,000 6,000 7,000 8,000 Btu/in.3 (feedstock and fuel)

Figure 2.2 Comparison of energy consumption for RIM prooess and other material manufacturing processes (Sweeney, 1987) 23

fabrication of polymeric products.

RIM processing equipment generally consists of a material

conditioning system, a high-pressure metering and injection system, a

mixhead, and a molding system. A basic structure is shown in Figure

2.3. Accurate stoichiometry, efficient mixing and high throughput

rates are the most important requirements for a RIM machine. Since

stoichiometry balance is extremely important for the polymer

properties, metering must be highly accurate through the entire

injection period. Because of the fast reaction rate and slow molecular

diffusion, high efficiency and short residence time are needed for

mixing. This also means higher flow rates through the metering and

injection system. There are currently many RIM machines available

conmercially in the United States (Sweeney, 1979, 1987) and in Europe

(Schaper, 1976). The machine size obviously depends on its particular

application. Hie larger ones have flow rates of 15 to 20 lbs/sec while

the smaller lab-scale ones can deliver up to a tenth of that capacity.

All, however, posess the following basic components.

Material conditioning system; Since the RIM process involves a

chemical reaction in the mold after the intermediates have been mixed,

the production of consistent parts requires that the material

delivered to the mixhead be consistent from shot to shot. Hie material

conditioning system is designed to ensure such consistency. It

typically includes tanks to hold the intermediates, agitators to

ensure that the material in the tanks is of homogeneous temperature, and a nucleation control system that keeps the level of dissolved HEAT HEAT EXCHANGER DISPLACEMENT EXCHANGER CYLINDERS MONOMER A MONOMER B + + CATALYST REINFORCEMENT

MIX HEAD

PUMP PUMP ~ r RECIRCULATION RECIRCULATION LOOP LOOP MOLD

Figure 2.3 Basic components of a RIM processing equipment (Ngugen, 1984)

>£kto 25

gases in the polyol component at the desired level. The tanks can

range in size from 15 to 160 gal. or larger depending on the

consumption rate. Jackets on the tanks as well as heat exchangers on

circulating loops are used for temperature control.

Metering and injection system: The metering system takes the

conditioned intermediates from the supply tanks and delivers them to

the mixing head at the desired rate and pressure. There are two basic

types of metering systems: high pressure axial or radial piston pumps,

and lance displacement cylinders. The piston pumps are hydraulic ptmps

that have been modified to handle chemicals. They are capable of

continuously metering at pressures up to 3500 p.s.i. Lance pistons, which are driven by a separate hydraulic pump, displace the reactants

from a high pressure metering cylinder. In addition to more precise metering, they have the capability of processing filled systems.

Mixing head: The mixheads in today’s RIM machines are versatile

(Sweeney, 1979, 1987; Lee, 1980). Five types of mixhead designs are shown in Figure 2.4. Each conmercial mixhead has its own merits. Their basic function is, however, the same. The mixing head contains a cylindrical chamber where the intermediates are mixed by direct

impingement at pressures ranging from 1500 to 3500 p.s.i.. It also contains a cylindrical cleaning piston which, after the shot is complete, moves forward to wipe the remaining materials out of the mixing chamber. In addition, there is a valving mechanism to shift the material flow between recirculation back to the tank and flow into the mixing chamber. This allows the circulating materials to reach an ss^'

aMT ( z y ^ ^ m 1 Tsssa

®

Figure 2.4 Schematics of conmercial mixing heads: (a) Krauss-Maffei and EMB, (b) Henneke, (c) Battenfeld, (d) Cincinnati Milacron (cross-section), (e) Cannon and Accuratio (Sweeney, 1979) 27 (C)

1

ce >

( d )

I. (D

® > c i . © • 2

Figure 2.4 (Continue) 28 equilibriiin at the proper temperature, pressure, and flow rate before shifting into the mixing position. On many occasions, an L-shaped mixhead is found more efficient than the conventional straight mixhead

(Schluter, 1979; Florentini, 1982; Molnar and Lee, 1988). The L-shaped mixhead is formed with an extension chamber which opens in one direction into the mold with the mixing chamber opening in another direction at right angles. By choosing the proper size and location of the extension chamber, the L-shaped mixhead is able to provide a higher efficiency in mixing and a more laminar flow in the runner, probably because the fluid has to change the flow direction from the mixing chamber to the extension chamber. The extension chamber can also be used for the addition of a third component, such as pigments or additives, which does not require intensive mixing (Schneider,

1983).

Molding system: The molding system encompasses the runner and gate, the mold cavity and the carrier which provides clamping and opening functions. The mold design is crucial to the quality of a RIM product. Mold making generally follows the standard practices for conventional injection molds. There are, however, some differences

(Wallner, 1978; Von Hassel, 1981; Maier and Menges, 1986). A number of parameters which have to be taken into careful consideration to assure a successful RIM molding are: (1) runner and gate design, (2) carrier design, (3) venting, (4) temperature control, (5) proper material for the mold, and (6) deraolding. The mold carrier holds the tool in the proper orientation for molding, provides enough clamping force to 29

overcome the in-mold pressure, opens and closes the mold, and

positions the open mold in an accessible position for demolding,

cleaning, and preparing it for the next shot.

Polyurethanes are well developed as conmercial products for the

RIM process. Nylon RIM products are relatively new, while epoxy RIM is

still in an early stage of commercial development. Those and other

materials applied in the RIM process will be discussed in more detail

in Section 2.2.

2.1.2 TRANSFER MOLDING

In the transfer molding process, a charge of therrooset material is

placed in a separate chamber, usually referred to as a pot, then

forced out of the pot and into a closed cavity, or cavities, where polymerization (curing) takes place. Channels, called sprues and

runners, direct the flow of material from the pot to the cavities, passing through a restriction, or gate, just before entering the cavity. Many cavities can be fed by a single pot. Air in the cavities displaced by the incoming material must be expo lied through

strategically placed vents. Figure 2.5 shows a schematic diagram of

the transfer molding process.

When the material, usually a measured charge in compacted form,

is introduced into the pot, it is preheated to a tempjerature approaching that of polymerization. Only sufficient material for a single shot is loaded at one time. The force which moves the charge of preheated material out of the pot is transmitted through a plunger 30

PLUNGER

START OF FILL

4 ’r^ s j R y ZZ

PLUNGER END OF FILL

VA % V.

V/)y/iy/7P7777r///7777/77

Figure 2.5 Schematic diagram of the transfer molding process 31

which is closely fitted to the pot to prevent leakage of material

through clearances between the plunger and the sides of the pot.

Sealing grooves usually are cut into the plunger to further reduce

leakage. The pot, plunger, sprues, runners, and cavity surfaces are

maintained at a temperature suitable for rapid curing of the material.

When a transfer mold reaches the end of its cure cycle, the entire

shot is ejected, including the gates, runners, sprues, and the cushion

of cured material (called the cull) formed in the pot.

Preheating of material is very important in transfer molding. Cold

material flows slowly, with the result that the first material to

enter the cavity may polymerize before reaching its destination. Poor-

quality moldings can then be expected from the standpoint of both

appearance and mechanical properties. Preheating can generally be

accomplished by heat lamps or ovens, but an efficient and more widely

used method is by dielectric heaters.

Formulations for transfer molding generally are of a softer plasticity than compression-grade materials. Types of material are

almost limitless within the thermosetting family, with the reservation

that high-impact grades of 1.0 ft-lb and higher (Izod test) usually will flow only at extreme pressures.

Transfer molding is one of the conventional manufacturing methods

in the rubber industry and the electronic industry for molding parts from thermosetting polymers. The development of transfer molding is parallel to that of compression molding. Different from compression molding where materials are charged into the mold before mold closing, 32

in transfer molding, the mold is closed before the compound is

injected in. Two types of resin transfer molding processes are available in the plastic industry: the integral-pot transfer molding, and the plunger transfer molding (Tadmor and Gogos, 1979; Schwartz and

Goodman, 1982; Hull, 1984; O ’Brien and Lenosky, 1986).

Integral-pot transfer molding: It was the first to be used and is so called because the pot and plunger are built as an integral part of the mold. Round pots are the most common, but other shapes also can be used to accommodate particular sprue locations with minimum material waste.

The integral-pot mold frame is a three-plate type with the pot contained in the middle section. The transfer plunger is mounted in the top section of the mold frame and the cavity in the bottom section. The area of the pot must be at least 10% larger than the total clamping area (horizontal surfaces that are in contact with polymer material) in the cavity section. This is to prevent the unclamping force from overcoming the clamping force and causing the mold to flash.

After the material has cured, the molded parts are ejected by the action of the press moving the mold ejector bar, but the cull and sprue are held to the bottom of the plunger by one or more molded dovetail. A lateral blow with a wooden stick or a soft hammer releases the cull and thereby clears the mold. Figure 2.6 schematically describes this procedure. 33

Plunger

M olding m aterial

Pot

Sprue

Cavity

Ejector pins

U) (b)

+ r TIT r1 u ti

(c)

Figure 2.6 Integral-pot transfer molding: (a) mold open, pot loaded; (b) mold closed; and (c) mold open, parts ejected, sprue on force plug (Hull, 1984) 34

Plunger transfer molding: It uses an auxiliary press ram to force the plunger into the pot (or cylinder), moving the material out of the pot and into the cavities. Transfer pressure and speed of transfer are readily controllable, independent of clamping pressure.

Pot size (hence cull size) in a plunger mold needs only to be large enough and deep enough to accommodate the full charge of material. Maximum pot area, on the other hand, is established by determining the force in tons which can be delivered by the auxiliary ram of the press and dividing that figure by 3.5. This will assure that 3.5 tons/sq.in. (O'Brien and Lenosky, 1986) will be available as molding pressure, which is sufficient for most transfer-grade material formulations.

The auxiliary rams usually are atop the upper, fixed platen of the press and are down-acting. The clamping ram moves the lower platen upward to close the mold. After the mold is clamped, the material is loaded into the pot, and the auxiliary ram is energized. Ratio of the clamp ram to the auxiliary ram is generally 3 to 1 or 4 to 1 (O'Brien and Lenosky, 1986).

When the material has cured, the auxiliary ram retracts, and the press is opened. The molded parts, cull and runners are ejected simultaneously by the action of the mold ejector bar. The process is described schematically in Figure 2.7.

A variation of this mold configuration is the three-plate plunger- transfer mold. In this method a floating runner plate distributes the material to sprues that can feed material directly into the cavities. 35

Transfer plunger

Force

Cavity

-4- r “4“ Ejector pin L + l

FT ft

(a) (b)

Cull R unners

Molded part -e-kv\ HA*4- 11iii I • i• - i i ,

(c)

Figure 2.7 Plunger transfer molding: (a) mold closed, pot loaded; (b) mold closed, plunger down; and (c) mold open, molded part, cull, and runners ejected (Hull, 1984) 36

This method is useful where parting line gating would be objectionable or where sliding cores or severe irregularity at the parting line level would interfere with runner layouts.

2.1.3 RESIN TRANSFER MOLDING

The resin transfer molding (KIM) process was first introduced in the fiberglass reinforced plastic industry in the 1970’s, and was used to produce parts from polyester resins with glass mat reinforcements.

A diagram of the RIM process is shown in Figure 2.8.

The reinforcing mat is placed in a two part mold which is tightly sealed around the edge and a catalyzed resin is injected into the cavity through properly positioned gates. Typically, injection is carried out at the top center of the mold and air venting is allowed for around the top perimeter of the mold. Once the resin has completely filled the mold, the injection process is stopped, the parts are cured at a desirable temperature and subsequently the mold is opened and the finished parts removed. A gel coat is typically applied to the molds prior to the injection process to allow for good resin flow, to improve the finished product’s surface quality, and to allow for the easy removal of finished parts. Fiberglass boats, swimning pools, and bathtubs are typical products that are produced using the KIM process. Reinforcement

Figure 2.8 Schematic diagram of the resin transfer molding process 38

2.2 MATERIAL CHEMISTRY AND REACTION KINETICS

Polymerization in the RIM process is more complicated than in

conventional polymerization methods. For example, in order to combine

with the processing step, reaction has to occur in the bulk state.

Therefore, reaction should not give off any by-products, such as water

of condensation, unless it can be quickly absorbed in the system (say

by a filler) or used for foaming to compensate for the polymerization

shrinkage. For the purpose of competing with the thermoplastic

injection molding process, the reaction rate has to be very high to

reduce the cycle time (i.e., reaction time is in minutes or seconds

instead of hours common in typical polymerization reactors). Most

polymerizations are also highly exothermic, which often results in

large temperature gradients inside the part because polymers are poor

conductors of heat. Temperature rises in excess of 100°C above wall

temperature have been recorded in many RIM parts. RIM materials must

therefore be stable over fairly wide temperature ranges (Lee, 1987).

2.2.1 POLYURETHANES

RIM started with polyurethanes and all of the production equipment was designed for than. It remains the standard by which other RIM materials are judged. The reaction kinetics of polyurethanes is a

typical step-growth polymerization. Detailed urethane chemistry can be

found in several reviews by Saunder and Frisch (1962), Sweeney (1979,

1987), and Lee (1980). The basic reaction which forms the urethane

linkage is: 39

O

R-N=C=0 + HO-R’ -- > R-N-C-O-R’ (2-1) I H where an isocyanate group -N00 reacts with a hydroxyl group -OH. The materials used for the production of urethane elastomers mainly consist of polyether diols or triols, generally polyoxypropylene polyol capped with polyoxyethylene groups to provide terminal primary hydroxyl groups, ranging in molecular weight from about 2000 to 7000.

Low molecular weight chain extenders, primarily short diols such as

1,4-butanediol (1,4-BDO) or ethylene glycol (EGO), are employed to provide good compatibility with the long chain polyols. The isocyanate used is a liquid form of 4,4 ’-diphenylmethane diisocyanate (MDI). Both tertiary amines and metal catalyst (e.g., dibutyltin dilaurate) are employed to provide fast reaction rate (Kuryla et al., 1966; Gerkin and Critchfield, 1976; Lee, 1980; Frisch, 1980; Steinle, 1980).

Urethanes with catalyst are very reactive at 40 to 50 °C the temperature at which they typically are mixed by impingement. Mold walls usually are controlled at 60 to 80°C. At such temperatures, urethanes may reach gelation within 5 seconds and have a cure time less than 60 seconds. In most theoretical studies of the RIM process

(Broyer and Macosko, 1976; Lee and Macosko, 1980; Castro et al., 1982;

Castro and Macosko, 1982), the urethane reaction was generally expressed as a simple n-th order reaction with an Arrhenius temperature dependence, such as: 40

- = Aq exp(-E/RT) c" (2-2 ) where the reaction order n varies from 1.5 to 3.0 depending on the catalyst level and may change at different conversion levels.

2.2.2 POLYURETHANE-UREAS/POLYUREAS

Despite the overwhelming success of polyurethane RIM in the production of bumper fiscias and others in the automotive industries, the generally soft elastomeric polyurethane can not meet various requirements if it is to be expanded to other structural applications.

To meet the demands, current research efforts are focusing on the development of improved chemical systems of polyurethanes with faster reaction rate and better physical properties. Faster reaction leads to reduced demolded time and cycle time. Better physical properties mean more applications of RIM in automotive and non-automotive industries.

In order to increase the reaction rate (i.e., the productivity) and to provide products with better physical properties, many RIM products, especially for automotive applications, have shifted from polyurethanes to polyurethane/ureas (Steinle, 1980). The main difference is that the latter uses a low molecular weight diamine Ets a chain extender instead of BDO and EGO. Diamine is much more reactive than diols. Aliphatic diamines reEtct instantaneously with isocyanates, while Euxxnatic diamines, in particular, hindered diamines such els methylene bisorthochloroaniline (MOCA) and 3,3*-dichlorobenzidine

(DCB) reEtct considerably slower. Because of the proven health hazetrds, 41

MOCA and DCB have been discontinued in some countries and restricted in the United States (Frisch, 1980). Today, diethyl-toluenediamine

(DETDA) is the major chain extender used in the polyurethane-urea RIM systems. Its reactivity is much higher than that of MOCA and DCB

(i.e., pot life in seconds rather than in minutes). Diamine extended

RIM systems build viscosity rapidly after mixing, leading to less turbulence during mold filling and often to faster and cleaner demolding.

The reaction kinetics of polyurea is also a typical step-growth polymerization. The detailed chemistry of polyurea can be found elsewhere (Schwartz and Goodman, 1982; Ewen, 1985, Baumann et al.,

1986; Nalepa and Eisenbraun, 1987). The basic reaction which forms the urea linkage is:

0 II R-N=C=0 + H„N-R’ ----- > R-N-C-N-R’ (2-3) * I I H H where an isocyanate group -N00 reacts with a primary amine group -NHg.

Since urea-linkage is thermally more stable than urethane-linkage, many researchers have been working on total polyurea resins for the

RIM applications, i.e., amine-terminated polyether resins with amine chain extenders (Casey et al., 1984; Wood, 1984; Dominguez, 1984;

Grigsby and Dominguez, 1985; Vespoli et al., 1985; Ewen, 1985;

Sneller, 1986). These systems, however, react too fast to fill large, complex molds (Vespoli et al., 1985). Currently, a major effort in the development of polyurea RIM is to slow down the reactivity of the diamine chain extender by modifying its chemical structure. Even without any catalyst, the urea formation is much faster than the catalyzed urethane formation (i.e., activation energy 3 - 5 Kcal/g- mole vs. 10 - 15 Kcal/g-mole) (Lee and Macosko, 1980; Macosko, 1983;

Vespoli, 1985). Processing conditions of polyurea RIM are similar to those of polyurethane RIM except that mold temperature must be higher

(i.e., T«100°C) in order to achieve desirable physical properties

(Ferber, 1986). Reaction kinetics are generally modeled as n-th order with an Arrhenius temperature dependence (Ewen, 1985).

2.2.3 NYLONS

Nylon RIM is probably the most studied RIM system except polyurethanes. Monsanto and Allied Chemical are the two American companies most active in developing nylon RIM. Nylon RIM is a block polymer made by reacting caprolactam with polymeric polyols, using a couple/initiator (bisacyllactam). The reaction occurs in two steps as shown in Figure 2.9. This chemical equation looks complicated, and it is. However, supplied as a package system, the two components react to form a polymer having distinct and repeatable physical characteristies.

In the RIM system, both reactant streams A and B are solutions in molten caprolactam. To keep the components fluid and homogeneous, liquids in the system should be maintained at 80 to 85°C. The melting ff HO “+ w CTO POLYOL BIS ACYLLACTAM I ) 0 Q H H 0 I I I I • f v I 11 11 :-N-C-R-C-NN-

I 0 H 0 0 0 I r?T Hi? 1 C-R-C N-(CH,)C 0 -n -c A B ' -‘C b POLYETHER POLYAMIDE N ^_XXnXXXXOOOOOOOOOOXX*T**XXXXOOOOOOOY?XX*XTYXOOOOOOOOOOCM (onnonoonoo

NYLON BLOCK POLYMER SEGMENT

Figure 2,9 Nylon bloc* polymer segment (Sweeney, 1987) 44

point of caprolactam is 69°C. In stream A, diol and bisimide are

dissolved in caprolactam in the stoichiometric portions needed for

polymerization. For stream B, ethyl magnesium bromide is reacted with

and dissolved in caprolactam to form bromomagnesium caprolactam. Since

the reaction is mainly a chain growth polymerization, mixing is less

critical for nylon-RIM than it is for polyurethanes. While urethane

streams have a relatively high viscosity, nylon RIM streams have a very low viscosity - 15 to 50 mPa»s - about one tenth that of urethane. This reduces the power needed to pump these materials, but

it also makes mold sealing more critical in order to prevent flashing.

Polymerization occurs quite rapidly at 140°C. The mixed reactants are shot rapidly into a hot mold at 130 to 140°C. The polymerization proceeds from the surface in contact with the hot mold to the center of the part itself. The reaction may be completed in 2 to 3 minutes.

An intrinsic difference between nylon RIM and other chemical systems used in RIM is the crystallization of nylon. In other systems, modulus build up is achieved by crosslinking and/or phase separation.

Demolding of these systems is in general limited by the rate of polymerization. For nylon RIM, molding temperatures are usually above the glass transition temperature of nylon and crystallization is essential to build up the modulus necessary for demolding. Both reaction and crystallization kinetics need to be considered in determining the optimum processing conditions. A desirable processing condition is for polymerization to be complete before any 45

crystallization, and the delay between completion of polymerization

and onset of crystallization should be short.

Kinetic models for the nylon RIM polymerization have been studied

by several researchers (Sibal et al., 1983; Lin et al., 1985). It is

found that an auto-catalytic type model originally proposed by Malkin

and co-workers (Malkin et al., 1979) fits the data well and can be

expressed in a relatively simple form:

= Aq exp(-E/RT) (1-C*)n (1+BqC*) (2-4)

where Bq is called the intensity of autocatalysis. The reaction order, n, is found to be very near unity.

2.2.4 EPOXIES

Epoxies are more expensive than other RIM materials, but they have better mechanical, thermal and electrical properties. In spite of its slow reaction rate and large exothermic heat of reaction (i.e., «30

Kcal/g-mole compered to «20 Kcal/g-mole for urethanes and wlO kcal/g- mole for nylon (Macosko, 1983; Osinski, 1985)), epoxy has long been recognized as a candidate for a commercial RIM process. Development work by Shell Chemical has provided a line of epoxy RIM packages with the trade name "Epon" which show that products are available to meet a range of property specifications (Sweeney, 1987).

Common epoxy RIM systems consist of a diglycidyl ether of bisphenol A (DEGBA) epoxy with amine curing agents. Epoxy RIM systems are relatively insensitive to minor variations in stoichiometry compared to the polyurethane counterpart. For the epoxy resin cured with a primary amine, the sequence of reactions has been suggested

(Schechter and Wynstra, 1956; Bidstrup and Macosko, 1984) as follows:

1) reaction of the primary amine with an epoxy group to form a secondary amine

0 H / \ K I R-NH2 + CH2- CH----- > R-N-CHg-CH- (2-5)

OH

2) reaction with another epoxy group to form a tertiary amine

OH H 0 I ^CHg-CH- I / \ *2 R-N-CH2-CH- + CH2- CH ■> R-N; (2-6) CH„-CH- OH Z I OH

3) reaction of the hydroxyl so formed with an epoxy group

(etherification),

0 / \ K„ -CH- + CH„- CH -- > -CH-0-CH„-CH- (2-7) I I OH OH

In general, reaction rate constants Kg and Kg are much smaller than Kj

(Gupta et al., 1983). 47

Like nylon RIM systems, epoxy systems have relatively low- viscosity reactants. The reactants are conditioned and recirculated in the RIM machine at temperatures ranging from 40 to 70°C. They are mixed by impingement at mixing head pressures in the range of 1000 to

1500 p.s.i. and flow into a hot mold where the polymerization is completed. Mold temperature ranges from 125 to 175°C. The reaction forming the polymer and giving the part its characteristic physical properties is completed in the mold, so postcure is unnecessary.

Reaction time is quite short, so production cycles are economical.

Very large parts can be molded. Nonwoven and woven mats can be placed in the mold to produce reinforced laminated composites, and the nonwoven reinforcement can be used for three-dimensional reinforced parts.

Since tertiary amines and hydroxyl groups act as catalytic sites in the cure reaction, epoxy kinetics is often represented as an autocatalytic reaction with an n-th order Arrhenius expression (Kamal and Sourour; 1973; Willey, 1981), i.e., Equation (2-8):

"5?" = A0 exP<-E/RT) d-C*)n (c*>m (2-8)

Osinski (1983), however, found that activation energy changed at different temperatures for epoxy reactions. Since kinetic characteristics are often measured at low temperatures, while molding is at high temperatures, he suggested that a dual rate expression should be used, which can be expressed as Equation (2-9): 48

1 A ri-i “En 4 Hn ~ a r = A1 exP<-Rf ) (l-C ) 1 + A2 exp(-^-) (l-C ) ^ (2-9) where the second term on the right hand side of the equation becomes important at low temperatures, but decreases in significance as the reaction temperature increases.

2.2.5 INTERPENETRATING POLYMER NETWORKS (IPNs)

Because of their high thermal expansion coefficient and low rigidity, polyurethanes cure considered inappropriate for structural applications. To improve the physical properties of polyurethanes, there have been many investigations on developing new RIM resins such as nylon, epoxy, and polyester, or externally reinforcing the existing materials by compounding with various reinforcement agents. There is, however, another approach to material modification which uses internal reinforcement to enhance the physical properties of existing materials. This is done by introducing a second reaction system into the RIM process to make up for deficiencies of the existing material.

This approach is essentially an application of interpenetrating polymer network (IPN) to the RIM process.

According to many researchers (Thomas and Sperling, 1976), IPN is a polymer alloy of two polymers that have been crosslinked or synthesized in the presence of each other. Since there are two components in an IPN, it can be classified according to the polymerization sequence or polymer structure. There are two types of 49

IPN according to the polymerization sequence: simultaneous and sequential IPNs. According to the polymer structure, there are two different types: full and semi IPNs. A more detailed discussion is given in Section 5.1.

Many IPNs have been developed in the past (Frisch et al., 1974;

Kim et al., 1976; Thomas and Sperling, 1976; Sperling, 1985). The two reaction systems are often one step growth type and one chain growth type. For instance, epoxy resins (step growth) can be added to acrylic solutions (chain growth) (Touhsaent et al., 1974), while polyester, acrylates, styrenic monomers and other vinyl systems (chain growth) can be blended into urethane resins (step growth) (Frisch et al.,

1974a,b; Kircha et al., 1984). Two step growth polymerizations may also be combined together to form an IPN such as polyurethane and epoxy (Frisch et al., 1982; Pemice et al., 1982).

Most IPNs are developed for slow processes such as casting and coating. For RIM applications, there are only a few commercially available IPN compounds. Ashland Chemical developed an acrylamate polymer (Wilkinson et al., 1983; Kelly, 1986) that is basically a urethane with unsaturation on the polyol chain, which forms a second network with a crosslinking agent - acrylic monomer. Amoco Chemical developed a series of polyester-urethane hybrids which can be used in the RIM process (Edwards, 1986). A similar polyester-urethane IPN has also been studied by Nguyen and Suh (1985, 1986), and Lee and coworkers (Hsu and Lee, 1985, 1988; Yang and Lee, 1987). Hiey found that the processing conditions in the RIM process such as impingement 50 pressure, stream Reynolds number, temperature, compound composition, and reaction kinetics can all affect the morphology and, subsequently, the physical properties of the IPN. The reason that IPN is more sensitive to the processing conditions than the constituent polymers is probably due to the strong interaction between the two polymerizations (Hsu and Lee, 1985).

2.3 RHEOLOGICAL CHANGES

In the RIM process, the reaction system generally starts at a low viscosity. The viscosity increases with conversion and reaches an infinite level when the material solidifies either by chemical crosslinking or physical changes such as phase separation or crystallization. Further reaction occurs in the solid state, which may increase the polymer modulus to a desired level before demolding.

During the process, mold filling must be finished before the system viscosity reaches too high a value. Ibis process is then followed by a curing stage in which the material is reacted to a sufficient mechanical strength before being ejected. For some materials, such as polyurethanes, a post cure stage is needed so that the product physical properties can reach an optimal level. A schematic diagram, as given in Figure 2.10, shows the relationship between rheological changes (viscosity and modulus) and kinetic change (polymer formation) during the RIM process.

A detailed understanding of rheological changes before the gelation point is essential in determining the mold filling S h u t E ject

Mold Mold Post Filling C uring Curt

M il ■■■■ Goo i

rj Max M odulus

V isco sity

t CycU REACTION TIME

Figure 2.10 Changes in viscosity and modulus during the RIM cycle (Lee, 1987) 52 characteristics and a similar study after the gelation point is

important in determining the demold time, polymer structure, and mechanical properties of the final product.

For epoxies, the liquid-solid transition is caused primarily by chemical crosslinking. The reaction temperature has to be above the glass transition temperature, Tg , of the network polymer for complete conversion to be achieved. The material will therefore be in a rubbery state in the mold at the end of the reaction. Generally, a high level of filler reinforcement is needed to provide enough green strength to the product at demolding. Although there is some crosslinking in urethane RIM systems, structure develop® mainly through domain formation. During polymerization, the urethane hard segments associate into domains that can crystallize and form spherulites. The domains act like rigid particles, which raise the viscosity of the system rapidly to a gelation very similar to that occurring in crosslinking polymerizations. In nylon RIM, it appears that viscosity increases initially by molecular weight build up until a critical value is reached for crystallization to occur. After that, the growth of crystallization is the main reason for the onset of physical gelation and modulus development (Sibal et al., 1983)

Viscosity is the most important material property in pelymer processing operations involving flow. In RIM, since flow is coupled with chemical reaction, one needs to understand the viscosity increase with reaction. The viscosity rise in RIM systems is of fundamental 53 importance for material characterization and process modelling. There are numerous studies regarding the viscosity-molecular weight relationship in polymer solutions and melts (Debye and Bueche, 1948;

Flory, 1953; Graessley et al., 1967, 1976; Graessley, 1974, 1977).

Viscosity changes for reactive polymeric systems, however, have not been studied in detail until recent years. Those studies have been reported by Lipshitz and Macosko (1976), Valles and Macosko (1979),

Han et al. (1983), Gonz'alez-Romero and Macosko (1985), and Sack and

Sheu (1985). The literature regarding the measurement of viscosity rise in various reactive systems often shows quite inconsistent results.

Modelling the viscosity changes is critical to the processing of thermoset polymers. Roller (1986) reviewed some recent studies regarding the rheology of curing thermoset polymers. In those studies, viscosity (H) changes during processing are expressed as a function of time (t) and temperature (T) such as in Equation (2-10):

)T| = exp(E^/RT) exp(k(T)t) (2-10)

Where is pre-exponential factor in initial viscosity expression and

E^ is viscosity activation energy. This model is useful for correlating experimental data measured at different temperatures. It lacks kinetic justification, however, and contains no fundamental information about the molecular structure. The next level of modelling is to convert viscosity versus

-reaction time, i.e., Equation (2-10), to viscosity versus conversion

using independently measured kinetic data. A typical expression

(Castro, 1985) is Equation (2-11):

(2-11)

Where oc is extent of reaction, oc^ is apparent gel conversion, and a

and b are constants. Such a correlation corrects for the effect of

reaction rate due to various catalyst concentrations. The model,

however, still does not incorporate any molecular structure.

From conversion data, the corresponding molecular weight growth

can be calculated using appropriate theories. Thus, the viscosity changes during polymerization can be expressed as a function of weight-average molecular weight, M , and an Arrhenius temperature w dependence such as Equation (2-12) (Lipshitz and Macosko, 1976; Valles and Macosko, 1979; Richter and Macosko, 1980’s; Macosko, 1985):

(2-12)

This model brings the structure parameter, Mw » into the viscosity correlation. The function f, however, needs to be fitted empirically.

For different systems, the maximum f values differ from 2.1 - 2.6 for an HDI based polyurethane (Lipshitz and Macosko, 1976), 2.6 for 55 polydimethylsiloxanes (Valles and Macosko, 1979), and 3.4 - 6.7 for a

MDI-based polyurethane (Richter and Macosko; 1980’s).

Today, most theoretical modelling regarding the rheological changes of reactive polymeric systems is based on polymer formation.

The effect of physical changes, such as phase separation, crystallization and vitrification, on rheology has not been modelled in detail.

2.4 MICROSTOUCTURE AND PROPERTY RELATIONS

Polyurethane and polyurea elastomers are multiblock copolymers of the type [AB]n » n>2. The A blocks, known as the hard blocks or hard segments are made of glassy or semicrystalline polymer, with T or Tm above the use temperature of'the polymer. In polyurethanes, polyurethane-ureas, or polyureas, the hard segments are mainly made of diisocyanate-extender sequences. The B blocks, or soft blocks

(segments) are flexible and rubbery with T or Tm (if they are crystallizable) below use temperature. The lengths of the A and B blocks in these polymers are generally shorter than those in A-B-A or

A-B block copolymers (Schollenberger, 1979). Moreover, there is, in general, a distribution of block lengths.

Thermodynamic incompatibility between the hard and soft segments causes phase separation and tends to generate a two phase morphology in these segmented block copolymers. A simplified representation of the segmented polyurethane structure is shown in Figure 2.11. The 56

Schematic Representation ( \ of the Segmented / f HAFtD S£GME?*T Polyurethane Microstructure

HARO SEGMENT ^

Figure 2.11 Schematic repreBentation of the segmented polyurethane microstructure (Camargo, 1984) 57

association of hard segments from many different chains into rigid

regions produces a network structure in the polymer. Upon deformation,

the polymer has a rubber-like response. The hard segments act as

"virtual crosslinks" that connect the flexible regions of the polymer,

and prevent them from flowing, up to very large elongations

(Schollenberger, 1979).

2.4.1 STRUCTURE AND PROPERTY

It is widely accepted that the engineering properties of the [AB]n

type elastomers are determined to a large extent by phase separation

(Zdrahala et al., 1979; van Bogart et al., 1979, 1983; Gibson, 1982;

Abouzahr and Wilkes, 1985). The quality of the microphase separation depends on a number of variables such as block lengths, starting materials and other factors affecting the relative compatibility of the blocks. The resulting microdomains can form lamellar or spherulitic superstructures. The thickness of the domain interface is also important. It may be affected by domain size distribution or chain packing within the domains. Excellent overviews of the different variables affecting structure-property relations in polyurethanes have been given by Saunder (1960), Saunder and Frisch (1962), Allport and

Janes (1973), and Noshay and MacGrath (1977). These variables can be briefly classified in three groups: intermolecular interactions, molecular structure, and molecular weight and crosslinking. 58

2.4.1.1 INTERMOLECULAR INTERACTIONS

Polyurethanes, polyurethane-ureas, and polyureas contain a high

concentration of polar groups. Specific interactions between these

entities are important to polymer properties. These interactions are

contributed by several factors: hydrogen bonding, aromatic it-electron

interactions, chain flexibility, and chain packing. One method to

assess the relative magnitude of these interactions is to compare the

molar cohesive energy, Ec , of the various groups found in the polymer.

Values reported in the literature are not always in agreement but

general trends can be established. Table 2.1 lists the values of some

important groups reported by Bunn (1955) and Fedors (1974). Since E c

decreases with increasing temperature, the values reported by Bunn,

although more accurate, are not as useful as Fedor's in predicting

cohesive energy of polymers (van Krevelen, 1976). The effects of the

molar cohesive energy of the groups on the properties of fibers have been summarized previously (Saunders, 1960). The melting point, as

shown in Figure 2.12, decreases as the content of strongly attracting groups decreases. These data also illustrate the greater molar

cohesive energy of urea groups compared to urethane and amide groups.

The effect of chain flexibility which accounts for the low Ec

value of the ester group can also be observed from Figure 2.12. As

shown in the polyester series, decreasing the content of ester groups

slightly increases the melting point. This behavior is considered as

evidence for the flexible character of the C-O-C groups of the ester, Table 2.1 Molar Cohesive Energies (KJ/mole) for Common Groups in Urethane and Urea Systems

t t t Chemical Group Bunn Fedors

-CHg- (methylene) 2.9 4.9

-O- (ether) 4.2 3.4

-COO- (ester) 12.1 18.0

-CgH4~ (o,p,m phenyl) 22.6 31.7

-NHOO- (amide) 35.6 33.4

-NHOOO- (urethane) 36.6 27.0

-NH00NH- (urea) - 50.2

* At the boiling point ** At 298% iue21 Trend of meltingFigure 2.12 points in homologous series of O O MELTING POINT 300 200 250 100 150 50 UBR FCAN TM I REPEATING UNIT IN ATOMS OFCHAIN NUMBER 4 6 8 0 22 20 18 16 14 Polyesters polymers (Hill and Walker, 1948) oymides Polyam Polyethylenes Polyureas yrt nes an lyureth o P 26

60 61 which tends to offset the moderately strong cohesive energy of the ester groups. The flexibility of the C-O-C group in urethanes can also account in part for the low melting point of urethanes relative to amides.

The effect of inter- or intra-domain specific interactions has been illustrated in detail by the work of Cooper and coworkers on polyurethane ionomers (Hwang et al., 1981; Miller et al., 1983; Yang et al.f 1983). Using dynamic mechanical data and thermal analysis, these authors have shown that the polybutadiene-based polyurethanes have better phase separation than the polyether-based polyurethanes.

This result, they claimed, is due to the absence of hard-soft segment hydrogen bonding in the former system, which facilitates phase separation. Introduction of the ionomeric functionality creates strong electrostatic interactions between the hard segments. In the polybutadiene-based systems, these interactions strengthen hard segment cohesive energy without significantly affecting phase separation; in the polyether-based systems, the strong electrostatic interactions generate an extra driving force for demixing and more complete phase separation.

2.4.1.2 MOLECULAR STRUCTURE

In addition to cohesive forces and chain flexibility discussed in the last section, the geometric fit of the polymer molecules limits the effectiveness of certain interactions and affects p»lymer performance. This effect is classically illustrated by the "zig-zag" 62

behavior of melting points with the number of carbon atoms in the

monomer (Saunder, 1960).

Generally, polyurethanes of higher tensile strength, tensile

modulus and tear strength are associated with rigid, symmetric

isocyanates. Bayer and coworker (1950, 1953) showed that MDI (4,4*—

diphenylmethane diisocyanate) conferred excellent properties on water

extended elastomers. In studying the influence of different

isocyanates on the properties of polyester-based polyurethane, Pigott

et al. (1960) found that polymers based on 2,4-TDI (2,4-toluene diisocyanate) have poorer mechanical performance due to the presence

of substituents and decreased symmetry of the molecule. This effect has been best illustrated by the work of Paik Suing and Schneider

(1975, 1977) and Senich and MacKnight (1979). The polymers based on

the symmetric 2,6-TDI have been shown to have better phase separation

than the polymers produced from the asymmetric 2,4-TDI. Similar

results were reported by Seefried et al. (1975) in a comparison between MDI and TDI based polyurethanes.

Ethylene glycol (EDO) and 1,4-butanediol (BDO) are the most widely used diol chain extenders. This is mainly due to the superior modulus and elongation that can be achieved with these materials because of their regular symmetric structure. N-methyldiethanolamine (MDEA) gives polymers with poor phase separation and poor tensile properties. This

is caused by the pending methyl group in the MDEA molecule, which

interferes with the degree of hard segment ordering. Seefried et al. 63

(1975) have shown similar effects when cycloaliphatic chain extenders are used.

Diamine chain extenders give better tensile properties and tear resistance than glycol chain extenders. They have been used successfully with asymmetric isocyanates such as TDI. Clearly, the improvement in mechanical performance in polyurethane-ureas is controlled by the much stronger mutual interactions of the urea linkages rather than by chain packing.

The effect of polyester structure on the physical properties of polyester-based polyurethanes has been studied by Pigott and coworkers

(1960). Tensile strength and modulus, for instance, are more affected by the presence of side chains than by ester group separation. In general, greater ester group separation yields improved low temperature flexibility and lower tear strength. Closer ester group spacing reduces flexibility at low temperatures and increases permanent set after elongation. Polyethers have weaker interchain attractive forces than polyesters and generally give elastomers somewhat inferior in physical properties at high temperatures, but the low glass transition temperatures of polyethers confer good performance at low temperatures. Among the polyethers, polytetramethylene adipate polyol (PINO) gives elastomers with the best tensile properties. This is due in part to its regularity and the ability to crystallize upon extension. The side methyl group in polypropylene oxide polyol (PPO) based systems increases interchain separation and contributes to lower levels of mechanical properties. 64

Nevertheless, FPO-EO (polypropylene oxide end-capped with ethylene oxide polyol) glycols are most widely used in RIM. Tensile properties in this case have been improved by the use of polymeric polyols or by diamine extension.

The molecular weight of the soft segment precursor has a large effect on the phase separation and thus on the physical properties.

Typical functionalities of polyesters and polyethers used are 2 or 3 with equivalent weights (molecular weight/functionality) in the range of 400 to 2500. Optimum molecular weight depends on the nature of the soft segment and on the application being considered (Allport and

Janes, 1973). Several studies (Seefried et al., 1975; Senich and

Macknight, 1979; Zdrahala et al., 1980) have shown that increasing polyol molecular weight increases the phase separation all other factors being unchanged. This in turn affects properties such as flexural modulus, hardness and heat sag, increasing molecular weight increases room temperature flexural modulus and hardness, but decreases heat sag. There is a remarkable improvement in low temperature flexural properties as the polyol molecular weight increases. The observed changes have usually been interpreted in terms of different reinforcing characteristics of the hard domains that form when different M soft segments are used (Zdrahala et al., 1980). 65

2.4.1.3 MOLECULAR WEIGHT AND CROSSLINKING

When discussing the influence of molecular weight on the

properties of a polymer, attention is frequently called to the

generalized behavior shown in Figure 2.13 (Saunders, 1960): properties

are enhanced by an increase in molecular weight up to a certain level

above which further improvement is small.

In the work of evaluating polyurethane properties as a function of

molecular weight, Schollenberger and Dinbergs (1973, 1979) prepared

two series of polymers in solution and in bulk. Molecular weight of

the first series was controlled by quenching the reaction with 1-

propanol; molecular weight of the second series was adjusted by

varying the NCO/OH concentration. Both series were prepared from MDI,

BDO and a polytetramethylene adipate diol (Mn=1000). The two studies

showed similar effects for most properties evaluated as a function of

molecular weight. Results showed that as polyurethane molecular weight

increased, the following properties increased: soft segment glass

transition temperature, tensile strength, abrasion resistant and room

temperature modulus. The following properties decreased with

increasing molecular weight: melt index, hysteresis, extension set,

stress relaxation time. In both studies, a typical average molecular weight after which no appreciable property change took place was

reported. For the bulk polymerized materials, this value was 100,000

to 200,000 (Mw ). iue21 Effect of Figuremolecular 2.13 weight on the physical properties Tensile Strength (MPa) 0 of a polymer (Saunders, 1960) Mol. Wt. / Crosslink x 10 ~3 10 x Crosslink / Wt. Mol. 5 10 15 20 25

CT>cn 67

The development of a limited degree of crosslinking in

polyurethane elastomer is often desired since increased molecular

weight, permanent set and tensile properties may be obtained.

Increased chemical crosslinking, however, can interfere with the

formation of the domain structure which is responsible for the rubber-

like properties of polyurethanes. Pigott and coworker (1960) showed

the existence of a minimum Mc , molecular weight/crosslink calculated,

in the tensile modulus of polyester-based polyurethanes as a function

of crosslinking density (see Figure 2.14). Only below M , does c

crosslinking improve mechanical properties. Above this value, modulus

improvement is due to a better segmented structure in the polymer.

Pigott et al. also observed a decrease in the plateau modulus in

torsional modulus-temperature curves with crosslinking. Ibis behavior

was explained by Cooper and Tobolsky (1967) in terms of domain

disruptions affecting the physical crosslinking in these systems.

Parallel studies in polyether-based polyurethanes were carried out by

Smith and Magnussen (1960).

Weisfeld et al. (1962) presented calculations for the contribution

that covalent and polar crosslinks make to the modulus of polyurethane-ureas. The dependence of the modulus on temperature was divided into contributions from a covalent network conforming to the

statistical theory of rubber elasticity, and contributions from polar links assumed to be governed by an Arrhenius law. <0 h a as *o 4> h- ** > K o • ^

2w Qj£ Polymer Molecular Weight

Figure 2.14 Tensile modulus vs. molecular weight/crosslink (calculated) (Pigott et al., 1960) 69

Morphological evidence of the effects of crosslinking on the domain formation have been considered among others by Slowikowska and

Daniewska (1975), Ophir and Wilkes (1979, 1980), and Russo and Thomas

(1983). Slowikowska and Daniewska showed that increased degree of crosslinking decreased polymer crystallization. This factor is very important to phase separation in glycol extended materials. Ophir and

Wilkes used small angle x-ray scattering (SAXS) and mechanical measurements to show that crosslinking of the hard segments (by peroxides) reduced the extent of phase separation and the Young’s modulus of the samples. Russo and Thomas examined two series of polyester-base polyurethanes by electron microscopy, thermal and dynamic mechanical methods. They suggested that soft segment crosslinking reduces the scale of phase separation and domain organization in these systems.

2.4.2 PHASE SEPARATION

It is possible to classify the study of phase separation in two categories. One approach is to pick a suitable technique and carry out dynamic studies. That is, to monitor the development of the polymerization and any observable physical changes occurring during this event. The main limitation of this approach is speed. RIM polymerizations are generally very fast and this limits both the time available for loading the sample into the reaction compartment and the time resolution of the signal being measured in some cases. The second approach to the problem is to produce polymer samples under a well 70 defined set of conditions and use a battery of techniques to evaluate the characteristics of the final specimen. Using this information, it

is possible to make certain hypotheses about the dynamic behavior of the polymerization.

It was not until 1966 that an understanding of the two phase structure of hard and rubbery regions was achieved. Based on the torsional modulus-temperature behavior of some urethane elastomers,

Cooper and Tobolsky (1966) suggested that the plateau region observed above the of the polyester-based soft segments could be explained by the existence of a second phase resulting from the intermolecular association of the aromatic segments, producing a subsidiary high temperature T . Since that year, a large research effort has been dedicated to understanding the morphology, domain interactions and structure property relations of segmented polyurethane elastomers.

Noshay and MacGrath (1977) and Camargo (1984) have given a comprehensive review of the urethane literature. The number of techniques used in this research effort is large. For this reason, it is difficult to bring all the results into a comprehensive picture and derive general conclusions. The following review of these results is categorized according to the experimental techniques: dynamic mechanical measurements, stress-strain measurements, differential scanning calorimetry, infrared spoctroscopy, and small angle x-ray scattering. 71

2.4.2.1 DYNAMIC MECHANICAL MEASUREMENT

Linear dynamic mechanical analysis gives very useful information on the degree of heterogeneity of polymer blends and block copolymers.

The elastic and viscoelastic properties of polymers obtained from small amplitude cyclic deformations yield information on the transitions occurring on the molecular scale. Data obtained over a broad temperature range can be used to ascertain the molecular response of the different phases. Figure 2.15 shows a typical dynamic mechanical response for block copolymers with different degrees of phase compatibility (Camargo, 1984).

In a highly phase-separated material, the transitional behavior of each phase will be that of the corresponding pure component. The modulus region between the transitions is called the rubbery plateau as shown in curve 1 of Figure 2.15. The value of the modulus in these systems is proportional to the volume fraction of the reinforcing phase (i.e., the one with a high glass transition or melting temperature).

For material with partial phase mixing, the transitions move closer to each other as shown in curve 2 of Figure 2.15. If partial microphase separation occurs due to diffuse phase boundaries or a broad size domain distribution, the plateau may not be well defined as shown in curve 3 of Figure 2.15. Hashimoto et al. (1983) have considered the effects of phase mixing and interphase broadening on the linear dynamic mechanical response of a block copolymer. Using a composite model proposed by Takayanagi (1963) and a technique first iue21 Typical dynamicFigure mechanical2.15 response for block Log Storage Modulus BLOCK COPOLYMEft(AB) n COPOLYMEft(AB) BLOCK T) Tmeaue (Tg)h Temperature (Tg)s compatibility(Camargo, 1984) copolymers with different degrees of phase \2 M 3/\ \ ' 3 Hgl Compatible Highly 3. \ N. 2. Partially Incompatibla Incompatibla Partially 2. .Hgl Incompatible Highly 1. \

Loss Tangent

73

described by Kraus and Roliman (1976), they calculated the temperature

dependence of the storage and loss moduli, G ’ and G", for different

situations representing both mixing-in-domain and domain-boundary

mixing. Their results indicate that both mixing-in-domain and domain-

boundary mixing increase the temperature dependence of the storage

modulus, G ’. The effect on loss modulus, G", however, is quite

different. Domain-boundary mixing causes small shifts on the position

of the G" transitions and peak broadening. Mixing-in-domain causes

larger shifts in the G" peaks toward each other. Little or no

broadening of the transitions is predicted for this situation.

Finally, if the system is miscible, a single transition occurs at a

temperature that is between the transition temperatures of the pure

phases. This is illustrated by curve 3 of Figure 2.15.

Huh and Cooper (1971) confirmed the existence of two phase

structures in polyether-based systems with MDI/BDO hard segments.

Decreasing the polyol molecular weight increased domain mixing and

increased the glass transition temperature of the soft segments.

Higher hard segment content increased the rubbery plateau modulus as

in other block copolymer systems. They also showed that thermal

treatment improved phase separation, as concluded from an increase in

the value of the rubbery plateau modulus upon annealing. Critchfield

and coworkers (Critchfield et al., 1973; Seefried et al., 1975a,b,c,d)

have presented a systematic study of the variables affecting dynamic mechanical response and phase separation in polyurethane elastomers.

These variables included chain extender, polyol molecular weight, 74 isocyanate structure, and hard segment content. In the work of Russo and Thomas (1983), dynamic mechanical data indicated that soft segment crosslinking reduces somewhat the scale of phase separation. The most significant effect, however, was the existence of a modulus after the hard segment melting point, as expected from the presence of chemical crosslinks.

2.4.2.2 STRESS-STRAIN MEASUREMENT

The effect of phase separation on the tensile properties of block copolymers is not as readily apparent as in the case of dynamic modulus-temperature behavior. The high moduli and strength exhibited by polyurethane elastomers is interpreted in terms of hard segment reinforcement (Estes et al., 1970a,b). Since the hard segments may be disrupted by the application of stress, a hysteresis effect is commonly observed. At' subsequent deformations, lower moduli and stress levels result.

Modulus and stress increase with increasing hard segment. At the same time, the elongation at break is decreased (Ferguson and

Patsavoudis, 1972a,b; Fertuson and Ahmad, 1977; Ferguson and Kumar,

1981; Zdrahala, 1980; Turner et al., 1982). Chang et al. (1982) showed that there is an optimum hard segment content for which toughness

(measured as the total work of deformation) is a maximum. Hysteresis also yields useful results. The area under the stress-strain curve represents the work that has been performed during extension. Because of stress-softening, part of the work is dissipated upon recovery. The 75

amount of energy dissipated increases with hard segment content. In

general, it has been found that polyether-based polyurethanes have

lower hysteresis than polyester-based polyurethanes (Estes et al.,

1970).

The effect of segment distribution on the mechanical properties

was studied by Harrell (1969) in polyether-based nonhydrogen bonded

polyurethanes. Narrowing the hard segment size distribution

dramatically increased modulus, tensile and extension set. Increasing

the hard segment size distribution decreased packing order and thus decreased the crosslinking efficiency. Narrowing the soft segment size distribution had little effect on modulus and elongation but a large effect on extension set.

2.4.2.3 DIFFERENTIAL SCANNING CALORIMETRY

Differential scanning calorimetry, DSC, is widely used in the study of structure property relations in many polymeric materials

(Turi, 1981). This method detects heat capacity changes whenever the temperature reaches a point where reorganization of the polymer chains occurs. These changes called thermal responses or thermal transitions are either first or second order. Melting or crystallization transitions are associated with the disruption (melting) or creation

(crystallization) of order in the material. Glass transitions are associated with a change of mobility in the polymer chains. In the study of heterophase systems, the method relies on the identification of multiple transitions that can be associated with different phases 76

in the system. As in the case of dynamic mechanical data, the

transitions are modified or shifted as the system becomes more compatible. The applications of thermal analysis to the study of polyurethane elastomers have been reviewed by Schollenberger (1979).

Polyurethane elastomers display multiple transitions. Generally, they can be classified into four groups: (1) second order transitions up to room temperature, which are associated with the glass transition temperature (T ) of the soft segments; (2) endothermic transitions in aS the range 20 - 60°C, which are assigned to the melting of polyester or polyether structures resulting from soft segment crystallization; (3) second order transitions in the range 50 - 150°C, which are associated with the glass transition temperature (T of the hard segments (some polyurethane-ureas have even higher V s) ; (4) endothermic transitions in the range 80 - 260°C, which are generally assigned to the melting of hard segment crystalline or paracrystalline domains.

The last group of transitions is least understood. The location and intensity of the high temperature endotherms depends not only on the chemical nature of the material but also on thermal histories. It is common to find multiple endotherms in this region (Seymour and Cooper,

1973).

Macknight and coworkers (Macknight et al., 1968; Kajiyama and

Macknight, 1970) studied a series of diisocyanate-diol systems.

Isocyanates used were hexamethylene diisocyanate (HDI), toluene 77

diisocyanate (TDI), and 4,4’-diphenyllmethane diisocyanate (MDI);

diol lengths were changed between 2 and 10. The MDI series showed a

higher than the TDI and HDI series and a marked dependence on the

diol length. The T^ for the TDI, HDI series did not show any major

trend. In the MDI-BDO series studied by Camberlin and coworker

(1982a,b), melting temperature (Tm ) increased up to 3 MDI units, and

decreased after that, except for the octadecyl terminated samples. The

decrease in melting temperature above 3 units was ascribed to poor

molecular arrangement at longer chain lengths. It was suggested that

in the octadecyl sample the larger T values for 4 or more MDI units

could be due to phase separation between the hard segments and the

long chain ends.

Multiple endothermic transitions, as mentioned earlier, above

100°C have often been observed in polyurethanes. Samuels and Wilkes

(1973a,b) suggested that these multiple endotherms could be explained

by short range interactions. During annealing, certain strains are

eliminated by short-range hard segment motions, resulting in the

reordering of the hard segments into structures having higher

disruption temperatures. Ng et al. (1973) expanded the work in

Harrell's (1969) materials. They found that samples with narrow hard

segment length distributions had sharp endotherms at 160°C. These

endotherms were insensitive to annealing and were assigned to the melting of hard segment crystallines. Samples with broad hard segment distributions showed broad transitions below 160°C. Annealing could 78 gradually shift the peaks to produce a sharp endothermic transition at

160°C. Seymour and Cooper (1973) proposed short-range interaction mechanisms to explain endotherm shifts by annealing. Using polyether- or polyester-based polyurethane with MDI/BDO hard segment, they identified three kinds of endotherms, all morphological in origin. The highest temperature peak was assigned to well-ordered hard segment crystallines. This peak varied slightly from sample to sample but was typically between 200 and 230°C. Between the amorphous and crystalline states different ordered segmented morphologies exist. The two peaks observed below 200 °C were assigned to disordering of hard segment clusters, which can be continuously improved by annealing. These ideas presented by Seymour and Cooper have been confirmed in more recent work by Cooper’s group (Hesketh et al., 1980; Van Bogart et al., 1980) in a large number of urethane systems.

Brunette et al. (1982) were able to correlate changes in hydrogen bonding organization with the degree of structural order evidenced by

DSC. Using MDI/BDO and 2,6-TDI/BDO hard segment-like copolymers they suggest that the N-H*•*0=C bond is stronger and more uniform after annealing. The change in bond distance presumably results from the chain reorganization that takes place at high temperatures.

2.4.2.4 INFRARED SPECTROSCOPY

Segmented polyurethanes and polyurethane-ureas are suited to IR investigations because they contain functional groups with characteristic absorption frequencies that are expected to reside in 79 specific domain locations. The N-H group in polyurethanes and polyureas forms hydrogen bridges with different groups. Some possibilities are shown in Figure 2.16 (Boyarchuk et al., 1965; Tanaka et al., 1968a; Ishihara et al., 1974): (1) is possible in all urethane systems; (2) is only possible in urea or urethane-urea systems; (3) can occur in polyether-based urethane or urea systems; and (4) can occur in polyester-based urethane or urea systems. Assignments for bonded and nonbonded absorption bands have been done mainly by studying model compounds (Boyarchuk et al., 1965; Tanaka et al.,

1968a,b; Nakayana et al., 1969; Ishihara et al., 1974; Yokoyama,

1978). For example, band assignments of N-H and C=0 in polyether-based

_ i polyurethanes (Srichatrapimuk and Cooper, 1978) are: 3420 cm for free N-H; 3320 cm for bonded N-H; 1733 cm for free C=0; and 1703

_ i cm for bonded C=0.

Bonded groups undergo spectral changes. The donor group has shifts

_ i to lower frequencies of 100 - 150 cm , with band broadening and a large increase in extinction coefficients. The proton acceptor

_ i undergoes smaller frequency shifts of 20 - 50 cm , with minimal band broadening or changes in the extinction coefficients (Pimentel and

MacClellan, 1960).

Measuring the relative amounts of bonded and nonbonded N-H groups in pxjlyamides and p>olyurethanes, Trifan and Terenze (1958) concluded that these groups were essentially all hydrogen bonded. Boyarchuk et 80

\ \ O N-H • • • \ / \ / N -H • • • 0=C N -H • • • 0= C / \ / \ N-H • • • N-H • • • / /

(1) URETHANE (2) UREA

\ \ C O \ / \ / N-H • • • O N-H • • • 0=C / \ / \ C C / /

(3) ETHER (4) ESTER

Figure 2.16 Hydrogen bondings of N-H group with different groupe in polyurethanes and polyureas 81

al. (1965) confirmed that most N-H groups formed hydrogen bonds, but

failed to recognize the importance of the hydrogen linkage to the ether oxygen in polyether urethanes. Seymour et al. (1970) proposed

that IR hydrogen bonding studies could be used to determine the degree of phase segregation in polyurethane elastomers. In their polyether- based polyurethane, the most important hydrogen bond acceptors are urethane carbonyls and ether oxygens. Assuming that most interurethane hydrogen bonds occur within the hard phase, and that the bonded N-H groups not associated with the carbonyls must be associated with a polyether in the soft segment, the degree of phase mixing is measured by the difference between the percent of bonded N-H and the percent of bonded C=0. With polyester-based urethanes, this analysis is much more difficult due to the overlapping of up to four peaks: bonded and nonbonded urethane and ester carbonyls.

Paik Sung and Schneider (1975) investigated the relationship between hydrogen bonding and phase separation in polyether-based polyurethanes produced from 2,4- and 2,6-TDI and extended with EDO. At room temperature, almost complete hydrogen bonding of the N-H groups was reported. The urethane carbonyl, on the other hand, showed 50% and

80% hydrogen bonding for the 2,4-TDI and 2,6-TDI polymers respectively. Independent DSC studies indicated that the 2,4-TDI based systems had a larger degree of phase mixing as judged from increases in T with increased hard segment concentration. gs 82

Few reports concerning IR studies on segmented polyurethane-urea elastomers have been published. Ishihara and coworkers (1974) studied elastomers synthesized from MDI, PTMO (polytetramethylene oxide polyol) and several diamine chain extenders. Using infrared spectroscopy and deformation studies, these authors showed that hydrogen bonding played an important role in the physical crosslinking of segmented polyurethane-urea elastomers.

Paik Sung et al. (1980a,b) studied the hydrogen bonding of polyurethane-ureas prepared from 2,4-TDI, EDA (ethylene diamine), and

PIWO (Mn=1000 and 2000). Almost complete hydrogen bonding of the N-H

_ i groups was indicated by a strong peak at 3330 cm . Likewise the urea carbonyl groups showed complete hydrogen bonding as indicated by the

_ i absence of a bond at 1690 cm corresponding to the free urea carbonyls (Ishihara et al. 1974). Following a simple group balancing argument, these authors suggested bonding of one urea C=0 to two urea

N-H groups in a 3-dimensional array. Similar 3-D hydrogen bonding was suggested without experimental data by Bonart et al. (1974) for MDI- based polyurethane-ureas. Nevertheless, it was confirmed recently by the work of Wang and Cooper (1983) on MDI/EDA/PIMO polyurethane-ureas, and by the work of Camargo (1984) on polyether-based polyurethane- ureas with a MDI/DETDA hard segment. This 3-D hydrogen bonding can provide an explanation for the large differences in cohesiveness of urea and urethane hard segments. It can also explain the large differences in molar cohesive energy presented before. 83

2.4.2.5 SMALL ANGLE X-RAY SCATTERING

Clough et al. (1968) first considered the use of small angle x-ray

scattering (SAXS) to study phase separation in polyurethanes.

Differences in the degree of phase separation were related to SAXS

intensity as a function of scattering angle. They concluded that

polyether-based polyurethanes had a higher degree of phase separation

than polyester based materials. This qualitative approach was also

used by Bonart et al. (1969), Wilkes and Yusek (1973), Schneider et

al. (1975), Wilkes and Emerson (1976), and Ophir and Wilkes (1980).

Bonart and Muller (1974) first presented a theoretical treatment

of SAXS curves based on Porod’s theory of small angle x-ray scattering

(Guinier and Foumet, 1955). Quantitative analysis of the scattering

curve gives information on domain size, domain spacing, thickness of

the interphase boundaries and relative purity of the phases.

Although small angle x-ray techniques have made some contribution

to the understanding of polyurethane morphology, there are inherent

disadvantages in the SAXS method. The validity of interphase

thickness, domain sizes, phase mixing, etc. are based on certain

assumptions. Porod’s law was derived for isotropic samples while

lamellar and other anisotropic structures have been seen (Chang et

al., 1982). Neumuller and Bonart (1982) have recently made modifications to the theory to deal with ideal anisotropic structures.

SAXS measurements are often carried out using slit collimation of the

incident radiation. An inherent disadvantage of this method is the

smearing of the intensity. Koberstein et al. (1980) gave 84 approximations to work with the smeared intensity in the Porod

relations, avoiding numerical desmearing.

A fundamental hypothesis of the theory is the existence of two phases. SAXS theory for systems with more than two phases is still under development (Koberstein, 1979). Thus, the direct application of

SAXS to evaluate phase separation in certain polyurethane or polyurea systems is simply an approximation. In fact, some polyurethane systems may have as many as four phases with different electron densities: amorphous or poracrystalline hard segments, crystalline hard segments, amorphous soft segments, and crystalline soft segments. Application of

SAXS to these systems is bound to give an average of all possible situations and the results should be treated with caution.

2.4.3 MORPHOLOGY AND DOMAIN STRUCTURE

The shapo of the microdomains in A-B and A-B-A block copolymers is well understood (e.g., Allport and Janes, 1973; Shen and Kawai, 1978;

Wilkes et al., 1979). At low hard segment compositions, the hard segment sequences aggregate to form spherical domains that act as physical crosslinks in a soft segment matrix. At increasing hard segment content, these domains elongate to more stable cylindrical shapos. At roughly equal volume fractions of the blocks, a bicontinuous lamellar structure exists. Higher hard segment composition causes a phase inversion and reverses the morphological process. At very high hard segment content, the soft domains are dispersed spheres that act as rubber modifiers in a rigid matrix. 85

Segmented block copolymers of the type (AB)^ are more complex.

Statistical distribution of the short, block length (Peebles, 1974;

Lopez-Serrano et al., 1980) and the presence of inter-domain and

intra-domain specific interactions suggest that segmented

polyurethanes and polyurethane-ureas be classified as intermediate

between random copolymers and truly block copolymers. The position of

a given system within this classification depends on the composition

and on the history of the material (Kimura et al., 1974).

2.4.3.1 HARD SEGMENT STRUCTURE

Bonart and coworkers (Bonart, 1968; Bonart et al., 1969; Bonart et

al., 1974) studied several polyester-based polyurethane-ureas and

polyester-based polyurethanes by wide angle x-ray diffraction (WAXS).

Bonart (1969) proposed a structural arrangement for the hard segment

chains, which could explain property fluctuations with amine chain

length (cf. Heikens et al., 1969; Bonart et al., 1970; Bleijenberg et

al., 1972). In later studies, these ideas were applied to a glycol- extended system (Bonart et al., 1969). They suggested that the 7.9 A paracrystalline hard segment reflection arose from hydrogen bond containing planes inclined 300 to the chain axis. Upon annealing, new crystalline reflections could appear in MDI/BDO hard segment materials. The model proposed by Bonart and coworkers did not take

into account molecular stereochemistry and group conformations. 86

Blackwell and Gardner (1979) proposed a model for the chain conformation of packing of the MDI/BDO hard segment. The model was based on the structure of methanol-capped MDI, determined by single crystal WAXS method. The model describes the chains as linked in stacks through N-H*•*0=C bonds involving half of the urethane groups.

The remaining urethanes are bonded to adjacent stacks, and thus the structure is stabilized by hydrogen bonding in two directions normal to the chain axis. A triclinic cell was proposed for this structure.

Cell dimensions are: a=5.2 A, b=4.8 A, c=35 A, oc=115°, 0=121°, and

Y=85°. Later work by Blackwell and Ross (1979) confirmed the proposed hard segment cell structure by WAXS analysis of highly stretched and annealed films (700% elongation, 130°C/24hr).

In subsequent work, Blackwell and coworkers (Blackwell et al.,

1981, 1982; Blackwell and Nagarajan, 1981) showed that the BDO hard segments had higher crystallinity than the PDO hard segments. This was explained by the ability of the MDI-BDO units to form an effective hydrogen bond network in their lowest energy extended conformation.

The MDI/PDO hard segments needed to rearrange to higher energy contracted conformations to form effective hydrogen bonding networks.

The resultant structure had higher energy and less driving force for phase separation resulting from crystallization. The MDI/EDO hard segments were expected to pack in an extended conformation because of the even number of CHg units. Instead, it was found that the measured

WAXS spacings corresponded to the predictions of a contracted 87

conformation. Apparently, this is due to packing difficulties because

of the short length of the EDO molecule.

2.4.3.2 POLYURETHANE AND POLYUREA MORPHOLOGY

Electron microscopy results thus far in polyurethane elastomers

have been ambiguous. The hard segment domains were first observed as

equiaxial grains of 3.0 - 50.0 nm in solvent etched iodine stained

polyether- and polyester-based polyurethanes with MDI/BDO hard segment

(Koutsky et al., 1970). Similar structures were reported by Wilkes et al. (1974), Lagasses (1977), and Lunardon et al. (1980).

Wilkes’ group also reported that spherulitic structures could be

induced in their systems by controlling casting conditions. Evidence of spherulitic structures from small angle light scattering (SALS) studies had been given earlier by Samuels and Wilkes (1973b) and

Kimura et al. (1974). Nevertheless, the results of Wilkes, Samuels and

Crystal were the first direct observations of a second kind of morphological feature of the order of microns. The presence of spherulitic superstructures in polyurethane elastomer morphologies was rapidly confirmed by other groups working in this area (e.g., Chang and Wilkes, 1975; Schneider et al., 1975a; Saunder and coworkers,

1975). Schneider and coworkers (1975a,b) suggested that polyurethanes with high urethane content were best described as three phase systems consisting of a featureless soft segment rich matrix and two phase spherulites consisting of hard segment rich fibrils and dispersed mixed domains. 88

Fridman and Thomas (1980) used OsO^ staining to study

MDI/BEDO/PPO-EO (BEDO = 2-butene-l,4-diol, PPO-EO = polypropylene

oxide end-capped with ethylene oxide polyol) polyurethanes. Solvent

cast films consisted of volume filling fibrillar spherulites. The

fibrils were hard segment rich surrounded by a matrix made of soft

segment material and disorganized hard segments. On the basis of these

observations, Fridman and Thomas proposed a morphological model

similar to Schneider's (1975). In their model, branching was

introduced to allow for soft segment accommodation.

Fridman et al. (1980) used WAXS and TEM to study the morphology of

MDI/BDO/PCL (PCL = polycaprolactone polyol) polyurethanes produced by

RIM as 3 mm thick plaques. The RIM samples exhibited four types of morphological superstructures: (1) spherulitic hard segment crystalline domains; (2) soft segment crystalline domains; (3) globular structures; (4) a noncrystalline matrix. All these structures displayed a strong dependence on mold position (thickness direction).

Different thermal histories were used to explain variations in morphology across the molded part. The globules, a new morphological feature in polyurethanes, were proposed to be pockets of glassy hard segment rich material which was unable to crystallize due to low temperature. As a part of the same effort to understand the morphology of bulk polymerized samples, a series of MDI/BDO/PFO-EO polyurethanes with hard segment contents ranging from 20 to 80% at 10% increments were prepared in bulk and cured at 100°C. They were characterized by 89

several techniques (Chang et al., 1982). Detailed optical and electron

microscopy revealed that all samples were heterogeneous, displaying a

variety of structures depending on the hard segment content and cross-

sectional location.

Polyureas, like polyurethanes, are segmented multiblock copolymers

of the type (AB)^. There are important differences between urethane

and urea phase separation. Urethane hard segments are often

crystalline, while in ureas both phases are amorphous. Also, since

urea gelation occurs in a few seconds, a highly nonequilibrium phase

structure can be locked in. Willkomm et al. (1988) used SAXS to study

the domain structures of polyureas made from MDI, DETDA, and D2000 (a

linear polypropylene oxide chain Mn=2000, with an aliphatic amine at

each end). Comparing the measured lengths of the MDI-DETDA unit to the

calculated values, the results show that while the hard domain chord

length does increase with hard segment content, it does not increase

as rapidly as the length of the fully extended hard segment. The 70

percent hard segment solution and RIM samples have 92 A and 68 A of

MDI-DETDA units, respectively, while the average fully extended hard block length is 192 A. This indicates that either the long hard

segment blocks are not fully extended in the hard domains or they do not completely participate in the hard domains. The low degree of phase separation in the high hard segment content polyureas indicates

that some portion of the hard segment chains are not segregating into hard domains, favoring the latter choice. The absolute magnitudes of 90 the values of hard and soft segment chord lengths are similar to those found by Chen-Tsai et al. (1986) in a solution-polymerized all- amorphous polyurethane system consisting of a toluene diisocyanate and butanediol hard segment and a hydroxyl terminated polybutadiene

(Mn=2200) soft segment. Chen-Tsai reported values of 260 A and 116 A, respectively, for the hard and soft segment chord lengths of the sample with 74 percent hard segment content. CHAPTER III

REACTION INJECTION MOLDING OF POLYUREAS: I. RHEO-KINETIC STUDIES IN SOLUTION POLYMERIZATIONS

SYNOPSIS

Rheological and kinetic changes of the polyurea reaction in solution were studied. The viscosity rise and conversion profile were measured using a Haake viscometer and an FTIR, respectively. In a three-component polyurea system, it was found that the liquid-solid transition is dominated by the formation of hard segments. The formation of soft segments mainly contributes to an initial viscosity build-up. A series of experiments were carried out to study the effects of resin composition, solvent concentration and reaction temperature on the viscosity rise and gelation of polyureas.

3.1 PREVIOUS WORK ON KINETICS. RHEOLOGY AND PHASE FORMATION

OF POLYUREA REACTION

3.1.1 CHEMICAL SYSTEMS

Since 1974 when the reaction injection molding (RIM) process was first used to produce automotive parts in the United States, rapid changes in both RIM machinery and formulations have taken place. With improved equipment, machinery has become simpler to operate while at the same time machine functions and controls have become more

91 92 sophisticated. At first almost all RIM materials were segmented polyurethanes of which formulations were based primarily on butanediol or ethylene glycol as chain extenders. To increase the productivity of the RIM process and to improve the properties of RIM products, most

RIM materials have shifted from polyurethanes to polyurethane-ureas.

Hie main difference is that the latter system uses a low molecular weight diamine as a chain extender, instead of a diol, to form urea hard segments and polyether soft segments with urethane end groups.

Currently, there is growing interest in total-urea systems in which hydroxyl-terminated polyethers used in the polyurethane-urea system are replaced by amine-terminated polyethers. The reactivity of amines with isocyanates, even with no catalyst, is much higher than that of alcohols with isocyanates (i.e. activation energy 1 - 8 Kcal/g-mole vs. 10 - 15 Kcal/g-mole). Aliphatic diamines or triamines react almost instantaneously with isocyanates, while aromatic diamines react relatively slower. Because of the high reactivity of aliphatic amines, polyurea RIM systems have the potential for improving productivity.

Since the urea-linkage has higher thermal stability and better physical properties than the urethane-1inkage, many researchers have been working on total polyurea resins for RIM applications (Dominguez,

1984; Wood, 1984; Casey, 1985; Even, 1985; Grigsby and Dominguez.,

1985; Vespoli et al., 1985-1987; Sneller, 1986, Hsu and Lee, 1988;

Chen et al., 1988; Willkomm et al., 1988). These systems, however, have difficulty filling large complex molds (Vespoli et al, 1985) due to the fast urea formation. Several hindered diamines such as bis- orthochloroaniline (MOCA), 3,3’-dichlorobenzidine (DCB), and 4,4’- methylene-bi8-2-chloroaniline (MBOCA) were used to slow down the reaction of polyurea because they react with isocyanates considerably slower. However, because of health concerns, the use of MOCA, DCB, and

MBOCA has been discontinued in some countries and restricted in the

United States (Frisch, 1980; Voelker and Balling, 1986). Today, diethyl-toluenediamine (DETDA) is the major chain extender used in the polyurethane-urea and total polyurea RIM systems. Its reactivity is much higher than that of MOCA, DCB, and MBOCA (i.e., pot life in seconds rather than in minutes). Tert-butyl-toluenediamine (TBTDA, Air

Products) and dimethylthio-toluenediamine (DMTTDA, Ethyl Chemicals) are also used as chain extenders in RIM systems. Their reactivities are slightly lower, but of the same order, than that of DETDA. Other diamines commercially available include Dytek (DuPont), methylene-bis- chloro-diethylaniline (MCDEA, Lonza Ltd.), ditert-butyl- ethylenediamine (Virginia Chemical), XPA series (UOP Inc.), and

Unilink 4100, 4200 (UOP Inc.). Aromatic amines generally react rapidly with aromatic isocyanates. Baumann et al. (1986) slowed down this reaction to a controllable rate through the use of secondary amines,

Unilink 4200. They found that the addition of secondary aromatic diamines provided formulators the versatility of controlling the degree of crosslinking and the capability of adjusting the polymer structure so that the desired product properties can be obtained. 94

Currently, a major effort in the development of polyurea RIM is to slow down the reactivity of aliphatic amines and diamine chain extenders by modifying its chemical structure. Schlotterbech et al.

(1989) have developed a new class of soft-segment macromolecules for

RIM systems. This new macromolecule family is called polyether- ketimines and has adjustable reactivities and low viscosities. Casey et al. (1985) applied the Hammett sigma correlation to study structure-reactivity relationships of several aromatic diamines. They reported that 2,4-diamino-anisole has about half the reactivity of

2,4-toluenediamine, while 3-methyl-6-tert-butyl-2,4-diamino-anisole has only one-sixth of the reactivity of 2,4-toluenediamine due to the steric hindrance of the tert-butyl group. Similar structure-reactivity relationships between DETDA and DWTDA were also reported (Nalepa and

Eisenbrain, 1987).

3.1.2 REACTION KINETICS

Typically, commercial polyurea RIM formulations contain three components: a diisocyanate, a low molecular weight aromatic diamine, and a high molecular weight aliphatic triamine. Hie formation of the urea linkage is represented by Equation (2-3). Even without any catalyst, urea formation is much feus ter than the catalyzed urethane formation (Lee and Macosko, 1980; Macosko, 1983; Vespoli and Alberino,

1985; Hsu and Lee, 1988). Hie formation of the urethane linkage is described by Equation (2-1). Hie mechanism of the reaction between isocyanates and alcohols or amines has been investigated by many 95

researchers. Several mechanisms have been proposed involving a series

of intermediates and they have been thoroughly reviewed in the

literature (Farkas and Mills, 1962; Saunders and Frish, 1962; Entelis

and Nesterov, 1966; Lyman, 1966, 1972; Reegen and Frisch, 1970;

Lipatova, 1974; Camargo, 1984; Pannone, 1985). In spite of all this

effort, the mechanism of reaction is not well understood. Moreover,

much of the work published thus far has been carried out in solution

with model monofunctional compounds. Only a few studies involve

compounds actually used in RIM.

Since polyurea reaction is a step-growth polymerization, the

reaction kinetics are generally modeled as an n-th order reaction with

an Arrhenius temperature dependence. Davis and Ebersole (1934)

conducted reactions where n-butylamine and aniline reacted

competitively with phenyl isocyanate. Based on the composition of the

product mixture, relative overall reaction rates were calculated.

Pannone and Macosko (1987, 1988) used adiabatic temperature rise

measurements to monitor the extent of reaction of an aliphatic diamine

and an aromatic diamine reacting competitively with a single

diisocyanate. The reactions of primary aliphatic amines and aromatic

isocyanates were too rapid to be monitored in the batch apparatus.

Therefore, the reaction of the aliphatic amine was considered

instantaneous, and only the reaction kinetics of the aromatic amine

were studied using an inert polyether as a solvent to replace the

aliphatic amine. The reaction rate of aromatic amines and isocyanates was described by a third-order model (first order in isocyanate, 96 second order in amine) to fit the adiabatic temperature rise for a two-component polyurea system. Kinetic parameters determined from the

RIM experiments predicted lower reaction rates in solution polymerization. This discrepancy, they claimed, is due to the polarity of the solvent (dimethylacetamide) which might have a strong effect on polyurea reaction. However, the overall temperature rise for all RIM experiments was in good agreement with the values predicted by heats of reaction measured in solution polymerization.

Adiabatic temperature rise methods have been used by several researchers for polyurethane RIM (Lipshitz and Macosko, 1977; Richter and Macosko, 1978; Steinle et al., 1980; Camargo et al., 1983). One recent study of commercial polyureas by Vespoli and coworkers (1986,

1987) included measurements of in-mold pressures and temperatures in addition to adiabatic temperature rise. They also found that the aliphatic amine reacted much faster with isocyanate than the aromatic amine. When injecting a two-component polyurea (aliphatic triamine and isocyanate) into the mold, the RIM machine shut down due to excessive back-pressure from the gelled material trying to fill the mold.

Therefore, the aliphatic triamine was diluted with a slowly reacting triol to slow down the reaction. A second-order kinetic model was used for the polyurea reactions by assuming that the aliphatic amine reacted instantaneously and could be ignored in the model. Correlation of the model prediction with the experimental data was fairly good.

In a kinetics and heat transfer study of polyurea RIM done by Hsu and Lee (1988), the solution polymerization data were successively 97 applied to the bulk polymerization. A kinetic model was proposed to treat the aliphatic triamine reaction and the aromatic diamine reaction separately. The parameters of the kinetic model were determined by FTIR using data from solution polymerizations. Combined with a heat transfer model, the kinetic model predicted fairly well the adiabatic temperature rise of bulk polyurea reaction in RIM.

3.1.3 RHBOLOGICAL CHANGES

Viscosity is the most important p h y s i c a l property in plastics processing. For thermoplastic melts, the viscosity is influenced primarily by temperature and shear rate. But, for the reactive systems, the viscosity change becomes more complicated because of the increase of molecular weight and morphological changes. Viscosity rise is very important in RIM, especially during mold filling. RIM components usually have low viscosities. Low viscosity means low pressure is needed for mixing and mold filling. As soon as the components mix, they begin reacting and the viscosity starts to increase. In polyurethane RIM, the viscosity rises slowly first followed by a sharp increase to reach gelation. Beyond this point the mixture behaves as a solid. Thus, viscosity as a function of reaction extent and temperature is essential for predicting filling behavior.

Useful RIM products must be processed in such a way that voids, air entrapment and knit lines are minimized. These problems can be avoided if the Theological characteristics of the RIM system are better understood. 98

In general, viscosity (H) can be expressed as a function of

reaction conversion (a), temperature(T), and shear rate (?):

r) = h «x,T,t) (3-1)

Lipshitz and Macosko (1976) and Richter and Macosko (1980) found

experimentally that the viscosity for polyurethane systems is

independent of the shear rate up to at least 1000 poise. For the

polyurethane RIM systems with phase separation, however, preliminary

experimental evidence indicated that viscosity is shear rate

independent up to at least 20 fold viscosity increase. Temperature

influences viscosity in two opposing ways. Raising the temperature will cause the viscosity to drop at a given reaction conversion but it

will also raise the reaction rate to cause an increase in reaction

conversion and viscosity. In order to separate these effects, the

kinetics must be measured independently.

The viscosity rise for two-component polyurethane systems has been

studied in detail by Lipshitz and Macosko (1976) and Richter and

Macosko (1980). They showed that the viscosity is a function of molecular weight which is in turn a function of reaction extent given by branching theory (Macosko and Miller, 1976). To date, few data have been reported on viscosity rise for three-component polyurethane or

polyurea RIM systems where phase separation plays a major role

(Tirrell et al., 1979). Castro and Macosko (1980) carried out

viscosity rise measurements of commercial urethane RIM systems on a 99

Rheometrics Mechanical Spectrometer. The materials used were without catalyst to allow enough time for injecting them into the rheometer.

They found that if the reduced viscosity was plotted against reaction conversion, all data collapsed into a single curve. A simple but useful rheological model was proposed, which did a good job in modelling the experimental data. Due to the rapid reaction rate of urea formation, it is not practical to use commercial viscometers for

Theological measurements. Those viscometers include isothermal cone and plate (Castro, 1980) and adiabatic Couette rheometers (Blake,

1987). To study mold filling of fast polyurea RIM systems, VespxxLi and

Alberino (1985) designed an experimental mold to monitor the pressure rise as the material filled a long runner and mold cavity. Viscosity parameters were then extracted from a mold filling model by fitting the measured molding pressure.

In order to conduct a quantitative analysis of polyurea RIM, a thorough understanding of the reaction kinetics and rheological changes during reaction is required. In this study, a series of solution polymerizations of p>olyureas are first conducted to slow down the reaction rate such that rheological and kinetic information can be obtained. TTiis information is helpful in understanding the roles which soft and hard segments play in the reaction kinetics and rheological changes of polyurea RIM. 100

3.1.4 PHASE FORMATION

Nesterov et al. (1976) used light scattering and electron

microscopy to study a urethane network formation. They showed that

heterogeneities developed during the reaction and grew in size up to

the gel point; in addition they observed that if the reaction was

accelerated by the addition of dimethylformamide, the nonhomogeneities

were suppressed. Tirrell et al. (1979) suggested that phase separation

in linear polyurethanes occurred during the reaction. Their conclusion

was supported: (1) by gel permeation chromatography measurements

showing MWD is much broader than expected from a statistical step-

growth copolymerization; (2) by viscosity measurements showing an

apparent gel point at a conversion lower than predicted from

functionality calculations; (3) by an optical transition in the

reaction medium from clear to opaque; and (4) by variation in sample

morphology with position and temperature history in RIM polymerized

plaques.

Castro et al. (1981) followed the amount of light transmitted by a

reactive mixture of polyurethane monomers. The number-average sequence

length of hard segments at the onset of phase separation was estimated

to be fairly constant at Nn=1.3 and independent of temperature and

sample composition. Wolf (1980) extended some of the turbidity studies to slightly catalyzed samples. His results were similar to those previously reported by Castrol et al.. 101

Hager et al. (1981) used differential scanning calorimetry (DSC)

to follow the polymerization kinetics and phase development in

polyurethanes. A second peak was observed in the reaction exotherm

obtained by this method and it was associated with the onset of phase

separation. Calculated number averaged sequence lengths showed

excellent agreement with those reported by Castro and coworkers

(1981). Recently, Thomas and a coworker (Chen et al., 1983) used an

optical microscope interfaced with a video camera to follow the

morphological changes during a urethane polymerization. With a system

similar to that used by Castro et al. (1981) and Hager et al. (1981),

they observed the formation of birefringent spherulites early in the

reaction, followed by a second kind of structure showing in the

background, which apparently was nucleated later, and did not show

biref ringence.

Camargo (1984) reported unusual behavior of the complex modulus when the polymerization was carried out in a cone and plate geometry with sinusoidal oscillations. This behavior was suggested to be

related to phase separation occurring during the polymerization. In

the study of dynamic properties during a urethane polymerization, the presence of a maximum in dG*/dt (i.e., downward curvature) was first observed by Camargo (1984). Mussetti (1975) and Valles and Macosko

(1979) have shown that the modulus development in homogeneous step-

* growth network formations has an upward curvature (i.e., dG /dt increases monotonically with time). Thus the results obtained by 102

Camargo can only be explained by a change in the kinetics of the

system or the formation of structures that impede the normal development of the modulus. It was found, moreover, that a critical

t point defined from the shape of the G curve correlated surprisingly well with the onset of turbidity in the materials.

The above described experimental techniques for phase separation can not detect changes at the molecular level. Infrared spectroscopy can follow the hydrogen bonding of the N-H and C=0 groups of the urethane and urea linkages and thus determine the degree of urethane- urethane and urea-urea hydrogen bonding and the extent of phase separation (Seymour et al., 1970). Using this interpretation, the development of hydrogen bonding in the carbonyl groups formed during the polymerization was followed by infrared spectroscopy. With the advent of Fourier transform infrared spectroscopy (FTIR), it was possible to obtain complete spectral information at time intervals of the order of seconds.

Using the technique of FTIR, Camargo (1984) prepared a series of soft and hard segment model polyurethanes in bulk to study the N-H and

C=0 regions in amorphous and highly crystalline systems. His soft segment models showed basically a simple carbonyl absorption at 1730

_ i cm . Because little or no phase separation was expected in these materials and the ratio of ether oxygens to urethane carbonyls was over 20, interurethane hydrogen bonding became unlikely. The fact that the absorption frequency in the N-H groups was similar in both soft 103

_ i and hard segment models (i.e., 3330 vs. 3320 can ) confirmed the capacity of the N-H groups to form hydrogen bonding with the backbone oxygen in the polyether-based soft segments. Interurethane hydrogen bonding in the hard segment materials is expected to be high due to crystalline lattice arrangements (Blackwell and Gardner, 1979; Hocker and Bom, 1979). Camargo's infrared spectra of hard segment models

_ i showed a single carbonyl band at 1705 cm and an N-H band at 3330

_ i cm . Similar N-H and C=0 bands were reported by Hocker and B o m

(1979) for the MDI/MeOH (methanol) material. Combining the results of soft and hard segment models, Camargo concluded that the low frequency

_ i band (1705 cm ) did indeed correspond to hydrogen bonded carbonyls

_ i while the high frequency band (1730 cm ) corresponded to free carbonyl groups.

In polyether-based polyurethane-ureas with MDI/EDA (EDA = ethylene

_ i diamine), Ishihara et al. (1974) assigned a band at 1640 cm to the

_ i hydrogen bonded urea carbonyls and one at 1695 cm to free urea carbonyls. For the RIM polymerized polyurethane-ureas, Camargo (1984) found that most urea carbonyl groups were hydrogen bonded as indicated

_ i _ i by a strong band at 1640 cm and no detectable peak at 1695 cm . In addition, examination of the N-H region indicated essentially complete hydrogen bonding of these groups. 104

3.2 EXPERIMENTAL

3.2.1 MATERIALS

The basic ingredients of polyureas used in this study consist of an aromatic diisocyanate, an aliphatic triamine, and an aromatic diamine. These materials are suiraaried in Table 3.1. The aromatic diisocyanate used is a liquid form of 4,4’-diphenylmethane- diisocyanate (MDI) (11305, Dow Chemical). 11305 is a blend of a diisocyanate monomer (50% by weight ) and a high molecular weight polymer (50% by weight). The equivalent weight of 11305 is 210. The aliphatic triamine is an amine-terminated polyether with a functionality of 3 and a molecular weight of about 5000 (Jeffamine

T5000, Texaco Chemical). T5000 is 85% aroinated with some secondary hydroxyl groups. The major aromatic diamine studied is an isomeric mixture of dimethylthio-toluenediamine (DMTDA, Ethyl). DMTDA has a functionality of 2 and an equivalent weight of 107. For comparison, two other aromatic diamines were also studied: diethyl-toluenediamine

(DETDA, Mobay) with a functionality of 2 and an equivalent weight of

89, and tert-butyl-toluenediamine (TBTDA, Air Products) with a functionality of 2 and an equivalent weight of 89. Figure 3.1 shows the chemical structures of DETDA, TBTDA, and DMTDA. The solvent used for solution polymerization was nitrobenzene which is inert to polymerization of polyurea (Lyman, 1960; Hsu and Lee, 1988).

The soft segment of polyurea is based on the reaction of 11305 with T5000; and the hard segment is based on the 11305 reacting with 105

Table 3.1 Materials Used in Polyurea Systems

Ingredient Code Compound

Di isocyanate 11305 4,4’-diphenylroethane-di isocyanate

Triamine T5000 amine-termined polyether

Diamine DETDA diethyl-toluenediamine

TBTDA tert-butyl-toluenediamine

DMTDA dimethylthio-toluenediamine

Solvent NB nitrobenzene 106

DETDA

H N CH CH H N CH CH a a a a a

”*c_^C^~nh* H C

CH CH H N CH CH a a a a a 3.5 —diethyl — 2.4>—toluenediomin* 3.5—diethyl—2,6—toluenediamine

TBTDA CH I • H N H C-C-CH

tert—butyl—toluenediamine

DMTDA H N SCH H N SCH

H C

SCH H N SCH 3.5—dlmethylthio—2,*—toluenediamine 3.5—dlmethylthio—2.6—toluenediamine

Figure 3.1 Chemical structures of aromatic diamines: DETDA, TBTDA, and DMTDA 107

one or more of diamines. To study the composition effect on

rheological changes, the ratio of aliphatic amine to aromatic amine

was varied from 70/30 to 20/80 (by weight) with equivalent moles of

isocyanate. In addition, pure soft-segmented polyurea and corresponding pure hard-segmented polyureas were studied as

references. Several selected polyurea systems were studied at dilution

levels of 85%, 75%, and 65% nitrobenzene to study the influence of

solvent concentration. The temperature effect on gel time and conversion was investigated at reaction temperatures of 25, 47, and

6 7 °C for some polyurea systems.

All ingredients were degassed and demoisturized under vacuum at room temperature for at least 12 hours to remove water and air. To get polyurea resins ready for kinetic and rheological measurements in solution polymerization, homogeneous solutions of isocyanate with nitrobenzene and of amines with nitrobenzene were first prepared in

3 two beakers of 150 cm , respectively. The desired amount of each of the solutions was poured into a 50 ml disposable syringe; then discharged into a beaker or the cup of the Haake viscometer for better mixing before measurements. 108

3.2.2 INSTRUMENTATION AND PROCEDURE

3.2.2.1 HAAKE VISCOMETER

A Haake viscometer, Model MVII, was used to measure the system viscosity rise before gelation. The model MVII has a cup with an

inside diameter of 42 nan, and a rotor with an outside diameter of 36.8 ran and a length of 60 ran. Water of desired temperature was circulating

in a heating jacket outside the measuring cup to control the system temperature at a fixed value which was checked occasionally by reading the digital thermometer with a small thermocouple inserted into the gap between the cup and the rotor during reaction.

During polymerization, different shear rates, ranging from 3.3 to

176.3 (1/sec) were applied to the system by changing the gear ratio.

The system viscosity was monitored continuously by using a chart recorder. Accuracy of the viscosity measurement depended on the sensitivity of the cable used in the Haake viscometer and also on the

Weissenberg effect of the reacting polymer solutions when they became viscoelastic. As the viscosity began rising sharply, material tended to climb the rotor, even at the lowest shear rate, due to the

Weissenberg normal force effect. Consequently, the accuracy of any further viscosity measurement was severely reduced and the measurement was terminated. The gel point is defined by extrapolating the viscosity rise curve to infinity. 109

3.2.2.2 FOURIER TRANSFORM INFRARED SPECTROSCOPY (FTIR)

A FTIR spectrometer (Nicolet 20DX) with a resolution of 4 cm in the transmission mode was used for kinetic measurements. After the reactants were mixed, about 0.05 ml of mixture was pasted between two sodium chloride (NaCl) plates that were then mounted on a sample holder located in the FTIR sample chamber. Since the functional group of isocyanate has a very strong IR absorbance, the gap of about 0.1 mm was sufficient and no spacer was used between the two NaCl plates. A temperature chamber with controller was designed to maintain the reaction temperature constant. Three consecutive one-second scans were taken, averaged, and stored in a floppy disk at each sampling time.

The sampling interval was 1 minute during most of the reaction, but was longer at high conversions because the reaction was very low in these regions. Measurement was ended at a preset time. All IR spectra in this study are expressed in absorbance. FTIR macro programs for spectra scanning and data analysis are described in Appendix A.

Infrared absorption is based on the fact that each chemical group in a sample absorbs infrared radiation at some characteristic frequencies. The amount of light intensity transmitted relative to the amount of light intensity incident on the sample can be related directly to the concentration of the absorbing species by Beer's law

(Kendall, 1966):

A. = ft. I C. (3-2) ill 110

where is the absorbance of species which can be determined from the

peak height or peak area, (L is the absorptivity that is characteristic of the absorbing species, t is the sample length, and

(L is the concentration of the absorbing species i. To compensate for thickness changes in the sample during the polymerization, a ratio is taken between the absorbance of the group of interest and that of an

internal standard, i.e., a group whose concentration does not change

_ i during reaction. In this study, the C-H peak at 2942 cm was chosen as the internal standard and the peak area is used to calculate the absorbance. Reaction conversion can then be determined from the change of the normalized absorbance:

a = l - ----^t (3-3)

where Aq and A^ are normalized absorbances of the monomer functional group of interest before and after a reaction time, t, respectively.

Before Beer's law is applied for any quantitative analysis, the absorptivities of reacting species need to be determined. Calibration curves of the isocyanate peak based on the changes of both the peak height and peak area have been established in our laboratory (Yang and

Lee, 1987). Hie calibration curves were established by preparing mixtures of isocyanate monomer and dichloromethane of known concentration. For the isocyanate peak, the calibration curves based Ill on both peak height and peak area formed straight lines. In this study, the change of the peak area of isocyanate peak was followed to determined the reaction kinetics of polyurea.

3.2.2.3 RHEOMETRICS DYNAMIC ANALYZER

Isothermal dynamic-mechanical measurements were performed using a

Rheometrics Dynamic Analyzer, RDA-700, in a parallel plate mode. The plate was a serrated disk 1.25 can in diameter. The frequency used was

1 Hz. The percent strain was adjusted between 1% to 25% to obtain a suitable torque signal.

Sample mixing and loading procedures have been described elsewhere

(e.g., Section 3.2.1; Cheng, 1988). Measurements were taken for 30 to

60 minutes depending on the reaction rate. The storage modulus G* and , loss modulus G" were recorded on a floppy disk for later data analysis.

3.3 RESULTS AND DISCUSSION

The reaction mechanism of the three-component polyurea is schematically described in Figure 3.2. The branched long chains represent polyether-based aliphatic triamines which react with isocyanates to form soft-segmented polyurea. The linear short chains represent aromatic diamines which react with isocyanates to form hard- segmented polyurea. Grafting between the two phases may occur through the reaction of urea linkages, of which one end is attached to soft segment and the other end is attached to hard segment. Such a TRIAMINE

\ DIAMINE

Figure 3.2 Schematic diagram of polyurea reaction 112 113 multiphase system may be considered as a block polymer.

3.3.1 INFLUENCE OF SOLVENT

One major concern in solution polymerization is how the solvent affects the chemical reaction. In order to figure this out, three solvents were tried in this study and the best one was selected for polyurea solution polymerization.

The three solvents are nitrobenzene, 2-methoxyethyl ether, and

N,N-dimethylacetamide. Their effect on viscosity rise and conversion of I1305/DMTTDA reaction is shown in Figure 3.3. N ,N-dimethylacetamide is a polar solvent and shows a very strong interaction with the urea system. It prevents the formation of hard segments. During the reaction, there is almost no viscosity rise and the reaction rate is very fast. The conversion of I1305/DMTDA reaction in 75% N,N- dimethylacetamide can reach 100% at room temperature. On the other hand, when the solvent was changed to nitrobenzene, there was a very sharp viscosity rise in a short time period and the reaction conversion could only reach 60% at the same conditions. When the system reaches gelation, the conversion curve tends to level off. When

2-methoxyethyl ether was used, an intermediate viscosity response was found.

With all three solvents tried, the nitrobenzene was found to have the least amount of interaction with the urea system. Therefore, it was chosen as the solvent in the solution polymerization experiments. Figure 3.3 Effect of solvent Effect of on the viscosity rise conversionand 3.3Figure

CONVERSION (ex) VISCOSITY (p o is e ) 0 in 0 o 0 0 0 0 M 0 0 •

. 0 2. 0 4. 50. 40. 30. 20. 10. 0. 1. 0 3. 0 50. 40. 30. 20. 10. 0 • °+ + . ° fW^ipo £l£ i— u_S u_S i— £l£ fW^ipo in the reaction I1305/DMTDAof / + „ .+

0 0 jf j V ise ise V + 0 0 o + I I 1 I'll I y I ■ I | I I | 0 + 0 V SOLVENT +

-t- 0

0 1 - vs. TIME & & TIME vs.

NITROBENZENE + 0 0

N.N-DIMETHYLACETAMIDE 0 -

1 2 MTOYTY TE D ETHER -METHOXYETHYL IE ) n i m ( TIME 1

0 1 111

1

1 □ A CONV Q □ □ IN 111Q<<< 1 T (r=l) 11305 75 □ s TIME vs. % SOLVENT % 1

/DMTDA /DMTDA a °C 5 2 at 6 □ • —T" I " T I— 0 1

1 114

115

3.3.2 RHEOLOGICAL MEASUREMENTS

The gel time (or flow time) is a critical parameter in the RIM process. Since this parameter is very difficult to measure in the polyurea RIM process, the polyurea systems were diluted with' nitrobenzene and the Theological changes during reaction were studied using a Haake viscometer.

In order to understand the roles that the soft and hard segments play in the Theological changes during polyurea reaction, viscosity measurements of pure soft and hard segments diluted in nitrobenzene were first investigated. Figure 3.4 shows the results of 85%-diluted

T5000/I1305 polyureas with two different stoichiometric ratios (amine to isocyanate) as well as 85%-diluted I1305/DETDA, I1305/TBTDA, and

I1305/DMTDA polyureas with equal stoichiometry. As shown in the upper figure of Figure 3.4, the viscosity rise of the T5000/I1305 reaction was extremely fast and reached the gel point in less than 18 seconds when the stoichiometric ratio is equal to one. Ibis gelation results from the chemical crosslinking formed by the reaction of triamines with diisocyanates. However, when the ratio is not equal to one (such as r=l/4.5), the system showed a limited viscosity rise and never gelled. These observations are consistent with the results predicted by Flory’s theory of gelation (Flory, 1953). According to Flory’s theory, the gel conversion of a branched system Ag + Bg can be calculated from the following equation: Figure 3.4 Viscosity rises of 85%-diluted T5000/I1305 and T5000/I1305 85%-diluted of rises Viscosity 3.4 Figure VISCOSITY (poise) VISCOSITY (p o fs e ) D. 1 OO. 2.00. 0 . 1 OO. 200. Mi . 5. 0. . . 0 1. 0 25. 20. 15. 10. 5. 0. □ b 3 Q 0 » ' ] 3 0 I 7 0 0 7 v o o v 0 V 3) o ) (3 v 0 0 v V i li 1 i i A 11305/diamine polyureas at temperature 25 25 ®C temperature at polyureas 11305/diamine 0 12) 2 (1 ICST v. TIME vs. VISCOSITY .n , ) I BL (sec) BULK IN t ) ( 0 11305/DMTDA 11305/DMTDA 0 0 11305/TBTDA 11305/TBTDA 0 V 11305/DETDA 11305/DETDA V IE ) n l m ( TIME 0 15. 10. , 9 N 5 NB 85% IN r 1 r« 0 r 1/ .5 /4 1 r= A I 05 5 30 /I1 0 0 0 5 T N 5 NB 85% IN 0 0 0 24) 4 (2 0 0 2 20. 0 0 0 117

As f=3 and r=l, Equation (3-4) predicts a gel conversion of 71% for

T5000/I1305 polyurea. However, no gelation is predicted for

T5000/I1305 polyurea with r=l/4.5.

Although the I1305/DETDA, I1305/TBTDA, and I1305/DWTDA polyureas are linear systems formed by reactions of diamines with diisocyanates, their viscosity rises behave like the crosslinking systems because of phase formation as shown in the lower figure of Figure 3.4. This kind of phase formation in urea reactions is due to the interactions such as hydrogen bonding, aromatic Tf-electron interactions, chain flexibility, and chain packing. The stoichiometric ratio between amine and isocyanate is equivalent to one in each 11305/diamine system.

After mixing of the components, the viscosity of the mixture began increasing and the system solidified in a certain period of time.

Compared to the chemical crosslinking, this solidification in

11305/diamine polyurea can be considered as a physical crosslinking.

Among the three 11305/diamine polyureas, the DETDA-based system showed the earliest onset of viscosity rise as well as the shortest gel time.

These differences are largely due to the hindrance effect of side groups on diamines, which reduces the activity of the adjacent amine group. As shown in Figure 3.1, the tert-butyl group in TBTDA has a stronger steric hindrance than the ethyl group in DETDA so that the

TBTDA has a lower reactivity then DETDA. Compared with DETDA, the 118

DOTDA has a much lower reactivity. This is because that methylthio

group has a higher capability of electron-withdrawing than the ethyl

group, which reduces the reactivity of the adjacent amine.

Also, the number in the parentheses indicates the gel time (in

seconds) of each 11305/diamine system in bulk polymerization. The bulk polymerizations were carried out using a powerful mixer for blending and the gel point was defined as the time when the mixer was stopped due to the solidification of system. The order of gel times are: DETDA

< TBTDA < DMTDA-based 11305/diamine polyureas. The results from bulk polymerizations are parallel to those from solution polymerizations.

The viscosity rises of all three 11305/aromatic diamine polyureas are slower than that of 15000/11305 polyurea under the condition that the stoichiometric ratio of amine to isocyanate is equal to one. This observation indicates that the reaction rate of aliphatic amine with

isocyanate is faster then that of aromatic amine with isocyanate.

Similiar results were also found by other researches in kinetic studies (Pannone and Macosko, 1987, 1988; Hsu and Lee, 1988).

Figure 3.5 shows the Theological changes of DETDA-based polyurea systems in 85% nitrobenzene. The stoichiometric ratio of amine to isocyanate is equivalent to one in each system. Shown in the upper figure of Figure 3.5 is the effect of soft/hard segment ratio (by weight) on the viscosity rises during the reactions of the three- component T5000/I1305/DETDA polyureas. The viscosity rises of corresponding T5000/I1305 polyurea and I1305/DETDA polyurea are shown in the lower figure of Figure 3.5. As shown in the upper figure, the Figure

VISCOSITY (poise) VISCOSITY (p o is e ) 0 0 0 N 0 0 0 0 N 0 0 0 • • 3.5 Viscosity riseB of DETDA-based polyureas in 85% in polyureas DETDA-based of riseB Viscosity 3.5 . . . . . 5. 4. 3. 2. 1. 0. 0. , , ° 4 S ■ VII .ft . 0 () 0 V 0 0 0 nitrobenzene at temperature 25 °Ctemperature at nitrobenzene V 0 V 0 0 A A * 0 I ,|M . 2. 1. 0 A 7 □ □ A ICST s TIME vs. VISCOSITY □ V . V "T r" T T" "

IE mi ) in (m TIME □ A

i i i i i i i | i i i | i i 11305/DETDA0 T5000/113050 V 20/80 V 0 A 50/50 A 0 3. 4. 4. 3. T5000/11305/DETDA ° D INNB 85% INNB 85% 70/30 0 a tri □ / di □

‘

119 three-component polyurea system with segment ratio soft/hard=70/30

(70% triamine and 30% diamine by weight with equivalent moles of isocyanate) has a longer gel time and a slower viscosity rise than the systems with segment ratios 50/50, and 20/80. In other words, with more aromatic diamine in a three-component polyurea, the viscosity rise tends to take place earlier. This observation seems to be contrary to the results which one would expect from the two-component polyurea systems shown in the lower figure in which the viscosity rise of the T5000/I1305 system takes place earlier than that of the

I1305/DETDA system. The same experimental results were also found for

TBTDA-based and DWTDA-based polyurea systems as shown in Figures 3.6 and 3.7, respectively. These results can be explained by the unequal reactivities of the aliphatic amine and the aromatic amine with isocyanates, as well as the physical crosslinking of the hard segment in polyurea formation. In the polyurea formation of a three-component system like that shown in Figure 3.2, the reactions of aliphatic and aromatic amines with isocyanate are competitive. Since aliphatic amines react much faster than aromatic amines (i.e., activation energy

1.5 Kcal/g-mole vs. 4 - 8 Kcal/g-mole, Pannone and Macosko, 1987; Hsu and Lee, 1988), one would expect that chemical crosslinking by the aliphatic triamine with diisocyanate should dominate both the viscosity rise and the gel formation in the reaction system. However, gel formation in an actual polyurea reaction is often not determined by the reaction of aliphatic triamine since, although the stoichiometric ratio of total amine functioned groups to isocyanate VISCOSITY vs. TIME 0 0 1 ] 1 1 1 1 1 0 N 9 I) A 9 0 D a A 9 ^ . 0 o TRI/DI >- o - 9 A D 7 0 /3 0 h <- □ A 5 0 /5 0 i/i * A 9 2 0 /8 0

0 D T5000/11305/TBTDA ■ V A 0 (0 V A ° IN 85% NB / 4 D > d a 9B9 iA 4, n n ° i .

0 . 2. 4. 6 . 8 . 1 0 o 0 1 | 1 1 1 | ■ T-

c 0 III

0 0 0 T5000/11305 a 0 V . 0 11305/TBTDA 0 0 0 y o h <- 0 0 w 0 0 0 o 0 0

O. 2. 4. 6. 8. 10. TIME ( m i n )

Figure 3.6 Viscosity rises of TBTDA-faased polyureas in 85% nitrobenzene at temperature 25 °C VISCOSITY vs. TIME 0

V 1) T5000/11305/DMTDA

0 TRI/DI a Q 7 0 /3 0 + v , A a A 5 0 /5 0 0 v 2 0 /8 0 v y o • + 1 0 /9 0 +’ A h *■ IN 85% NB +7 i/i A +7 I 0 0 0 , +7 ' 0 (/) ■fy A □ +7 □ > d

• 0. 10. 20. 30. 40. 5 n 0 O M • 0

0 0 0 T5000/I1305 laQ. v , 0 11305/DMTDA 0 > 0 >- o • I- *- 0 W > 0 0 0 > 0 (/) > 0 IN 85% NB 0 > 6

0. 10. 20. 30. 40. 50. TIME (m ln )

Figure 3.7 Viscosity rises of DMTDA-based polyureas in 85% nitrobenzene at temperature 25 “C 123

functional groups is kept near one, the "effective" stoichiometric

ratio during the reaction of aliphatic triamine and diisocyanate can be much less than that because aromatic amines have much lower reactivity than aliphatic amines. For the three-component system shown

in Figure 3.7, the molar ratio of T5000/I1305/DMTDA is nearly

1/7.7/6.7, which means that the "effective" stoichiometric ratio during the the reaction of aliphatic triamine and diisocyanate is about 1/7.7. Under this condition, no chemical crosslinking will be formed based on the Flory-Stockmayer theory (Flory, 1953) and the evidence shown in Figure 3.4. Therefore, the formation of soft segment in this polyurea system only contributes a certain amount of viscosity build-up, instead of gelation. The viscosity rise and gelation of this system are mainly dominated by the formation of hard segment, i.e., when the physical crosslinking occurs the system tends to gel.

Since the viscosity rise and gel time of polyurea reaction are dominated by the formation of hard segment, modifying the chain extender is a way to slow down the reaction in the development of polyurea RIM. Current suppliers of polyurea resins provide many types of chain extenders (the aromatic short chain amines). DETDA, TBTDA, and DMTDA are three of them. The chemical structures of these three chain extenders used in this study are shown in Figure 3.1. Their effects on the viscosity rise and gel time of the polyurea reaction are displayed in Figure 3.8. Each polyurea system has the same segment ratio soft/hard=70/30 by weight with equivalent moles of isocyanate, but a different chain extender. The DMTDA-based system has a very slow iue38 icst ie fDTA, BD- ETA, and EMTDA-, TBTDA-, DETDA-, of rises Viscosity 3.8 Figure VISCOSITY (p oise) o CO O (0 O o N o o. DETDA/DOTDA othr=03, n8%N, t25°C 5 2 at NB,85% in soft/hard=70/30, io. ICST v.TIME vs. VISCOSITY (50/50) .0 ■ I E ml ) ln (m TIME 0 30. 20. ) S t w ( 0 3 / 0 8 / 0 7 - A D T B T / 5 1 0 1 3 / O O O S T — -based polyureas with segment ratio segment with polyureas -based O 305/DETDA-70/80/30

< 124 viscosity rise and a very long gel time. A DFIDA/DMTDA-based system

(i.e., chain extender is composed of 50% DETDA and 50% DMIDA by weight) has a viscosity rise and gel time between those of DETDA-based and DMTDA-based systems. The viscosity rise of the mixture is similar to that of the TBTDA-based system. Figure 3.9 shows the viscosity rises of DETDA/DMTDA-based (DETDA/DMTDA=50/50) polyurea systems with segment ratios soft/hard=70/30, 50/50, and 20/80. The observed results are similar to those from DETDA-based or DMIDA-based systems, i.e., with more aromatic diamine in a three-component polyurea, the viscosity rise tends to take place earlier.

3.3.3 KINETIC MEASUREMENTS

To further study the interactions between soft and hard segments during polyurea reaction, kinetic information is needed. The reaction kinetics of polyureas were measured using Fourier transform infrared spectroscopy (FTIR). Figure 3.10 shows a portion of the infrared spectra (i.e., wavenunbers 2000 to 3600) for II305/DMTDA polyurea reaction diluted with 85% nitrobenzene. The FTIR analysis is based upon the peak change of functional groups or characteristic linkages during the reaction period. Therefore, there are more than one peak which may change when the polyurea reaction take place. In principle,

_ i _ i the isocyanate peak (2270 cm ), amine peak (3338 cm ), and the amide

_ i peak (NH stretching, around 3286 cm ) can be followed during urea formation in which both isocyanate and amine monomers are consumed and Figure 3.9 Viscosity rises of DETDA/DMTDAf50/50)-based polyureas DETDA/DMTDAf50/50)-based of rises Viscosity 3.9 Figure VISCOSITY (p o ise ) 40. 80. 120. 160. 200. . . . . 2 15. 12. 9. 6. 3. O. + . 9 . - - "■ " . -T* o .. + * ° o -- with segment ratio Boft/hard=70/30, 50/50, and 20/80, and 50/50, Boft/hard=70/30, ratio segment with in 85% NB, at 25 at NB,°C85% in t, o ____ '---r1 A o 1 o ___ . o . O _ ___ .. ICST v. TIME vs. VISCOSITY O 1 1 i 1 • * O — + 0 4 ■ -- 0 o I E ml ) ln (m TIME + A 1 -- . --- — + --- + A TSOOO/11 305/(DETDA/DMTDA) '---1 + A --- 70/(15/15) + 20/(40/40) 0 A TRI/C DETOA/OMTDA) N S NB IN 6SX 50/(25/25) - + +

- 126 ftSOS/ONTOA ONI) IN Mftl Ml 2 2 7 0

. . A -

s

Figure 3.10 FTIR spectra of I1305/BMTDA polyurea reactions in 85X NB at 25 °C 127 128 an amide peak is formed. However, the amine peak and the amide peak are found to be strongly affected by hydrogen bonding and also tend to interfere with each other. The isocyanate peak can be more precisely monitored since it is located in an isolated area and its absorbance is much higher than the amine and amide peaks. Previous studies (Hsu and Lee, 1988) in our laboratory indicate that the polyurea conversion data based on the peak of the second NH-streaching of amide (3286

_ i cm ) scatter more than those based on the NCO-peak because the amide peak is strongly affected by hydrogen bonding. Nevertheless, following the changes of both peaks gave the same results. The conversion profiles reported in this study were based on the change of the NCO- peak .

Since TBTDA starts to react with isocyanate after mixing, the reference spectrum (i.e., at oc=0 ) can not be easily obtained because

40 seconds is needed for sample preparation and IR stabilization before the first scanning is taken. Therefore, 4-tert-butyl-toluene

(TBT), a chemical similar to TBTDA but without an amine group, was selected to replace the TBTDA for the determination of the reference spectrum. The chemical structures and IR spectra of TBT and TBTDA are expressed in Figure 3.11. Except for the NHg group, TBT and TBTDA have similiar chemical structures and IR spectra. The C-H peak area of TBT is very close to that of TBTDA after normalization based on the peak

_ t of tert-butyl group (1351 cm ). Hence, to determine the reference spectrum for any TBTDA-containing polyurea system, equivalent moles of IH

K C - I - O r

w.e- ^ Q ^ - ww.

O 1000.0 1400.0 1000.0 V A V t N U M K M CCM— 1>

Figure 3.11 Comparison of chemical structures and IR spectra between TBT and TBTDA TBT were used to substitute TBTDA. Figure 3.12 showB conversion vs.

time plots of TBTDA-based polyureas in 85% nitrobenzene at 25 °C. Shown

in the tipper figure of Figure 3.12 is the effect of soft/hard segment

ratio (by weight) on the conversion profiles of the three-component

T5000/I1305/TBTDA polyureas. The conversion profiles of corresponding

T5000/I1305 and II305/TBTDA polyureas are also shown in the lower

figure of Figure 3.12. The T5000 and 11305 reaction was so fast that

even at 85% dilution with nitrobenzene, the conversion profile still could not be measured. The conversion profile of T5000 reacting with

11305 in Figure 3.12 is estimated based on a kinetic model developed by Hsu and Lee (1988). The reaction of II305/TBTDA polyurea was much

slower and has a limiting conversion at about 0.65. The conversion profiles of both 70/30- and 20/80-segmented polyureas (i.e.,

T5000/I1305/TBTDA polyurea systems with segment ratios soft/hard=70/30, and 20/80) resemble that of 11305/TBTflDA polyurea. In the early stage of the reaction, 70/30-segmented polyurea has higher conversions, but lower reaction rates than 20/80-segmented polyurea.

This is because the 70/30-segmented polyurea contains more T5000 than

20/80-segmented polyurea. Nevertheless, both 70/30- and 20/80- segmented polyureas have limiting conversions close to that (i.e.,

0.65) of I1305/TBTDA polyurea. It seems that the limiting conversions of TBTDA-containing polyureas are strongly affected by the formation of hard segments, i.e., when the physical crosslinking occurs the conversion curve tends to level off. CONVERSION vs. TIME 0 i 1 I I | I I ~ T " ■' | I I I |— 1T" I I | I — ■T-— r

#

Z I n ofO 0 o 0 o o 0 0 in D ° 5) 6 TRI/DI n Dv □ 70/30 u » 20/80 > V z T5000/11305/TBTDA 0 IN 8 5 % NB 0 0 1 * ■ l i i -I— I i i i i i i i I i i . 0 0. 2. 4. 6. 8. 10. 0 - V - I , , -- p.. ,

> «

z 0 0 0 m ooo0 ^ ° i/i o 0 ■ k 0 u 0 T5000/I1305 > 0 11305/TBTDA z 0 IN 8 5 % NB 0 o o 0. 2. 4. 6. 8. 10.

TIME (m in)

Figure 3.12 Conversion vs. time plots of TBTDA-based polyureas in 85% NB at 25 «C Figure 3.13 shows conversion vs. time plots of DMTDA-based

polyureas in 85% nitrobenzene. There is no difficulty in obtaining the

reference spectrun for any DMTDA-containing polyureas because of lower

reactivity of DMTDA. No changes were observed in the IR spectra of

II305/DMTDA polyurea during the first two minutes of reaction. The

results found in Figure 3.13 are very similar to those in Figure 3.12.

A significant conversion jump at the beginning of the 70/30-segmented polyurea reaction is due to a high content of T5000 contained in this system. All three DMTDA-containing polyureas have limiting conversions between 0.50 and 0.55.

For comparison, the conversion profiles of TBTDA-based and DMTDA- based polyureas with segment ratio 70/30 are reploted in Figure 3.14, along with the profiles of T5000/I1305, I1305/TBTDA, and I1305/DMTDA polyureas. TBTDA-based polyurea systems possess faster reaction rates and higher limiting conversions when compared with DMTDA-based polyurea systems. This is because TBTDA has a higher reactivity than

DMTDA. Apparently, there is a strong effect of chain extender on the reaction rate and the limiting conversion of polyurea reaction .

3.3.4 GELATION AND ITS DEPENDENCE ON CONCENTRATION AND TEMPERATURE

In the analyses of rheology and kinetics of RIM systems, it is useful to express the viscosity as a function of extent of reaction.

The mapping procedure (Castro, 1980) for combining the viscosity-time data and conversion-time data is shown in Figure 3.15; the viscosity­ time data were obtained with the Haake viscometer and the CONVERSION vs. TIME 0 ■ ■ ■ "[ i* i ' i | i i i | i i""i | i i i

# T5000/I1305/DMTDA

TRI/DI Z 0 7 0 /3 0 V 2 0 /8 0 O a

i/i 6 o 0 D °

u > z 0 7 □ V 0 0 IN 85% NB 0 o I l l I I I ■ ■ I ■ ■ ■ I ■ I I 0 0. 10. 20. 30. 40. 50. 0 r « 0 T5000/I1305

z 0 11305/DMTDA 0 . i/i 6

O o ° 0 ° u 0 0 > 0 0 z 0 0 0 IN 85% NB 0 o q. ■ I ■ ■ . I ■ ■ ■ I . ■ ■ I ■ ■ ■ 0 0. 10. 20. 30. 40. 50. TIME ( m i n )

Figure 3.13 Conversion vs. time plots of DMTDA-baaed polyureas 85% NB at 25 «C CONVERSION vs. TIME 0 I I | I I I | I I—"I— "T™ 1™ 'T'1 | I I I

8 0 T5000/11305/TBTDA=70/80/30(wt%)

A T5000/11305/DMTDA«70/68/30(wt%) Z □ 0 in □ to 0 . * 1 ‘ ‘ u. A u A A > A z A A 0 IN 85% NB 0 o ■ ■ ■ 1 ■ ■ *__I ,i 1.1. . i I i.i i I i i i 0 0. 10. 20. 30. 40. 50. 0

8 V T5000/I1305 0 11305/TBTDA 0 11305/DMTDA Z Ofl ■ t " i/i d •o 0 0 K y « 0 0 0 0 0 > u0 ° 0 z 0 0 0 IN 85% NB 0 o « 0 i i 0 0. 10. 20. 30. 40. 50.

TIME (m in)

Figure 3.14 Effect of chain extender on the reaction rate and limiting conversion of polyurea reaction a

t

Figure 3.15 Mapping procedure to obtain viscosity as a function of conversion in reactive polymer mixtures (Castro, 1980) 135 136

conversion-time data were obtained from FTIR.

Figure 3.16 showB viscosity vs. conversion plots of TBTDA-based polyureas in 85% nitrobenzene. The gelations of 70/30- and 20/80- segroented polyurea reactions took place around conversions between

0.60 and 0.70, which are very close to the gel conversion of the

I1305/TBTDA polyurea reaction. According to Flory’s gelation theory, the predicted gel conversion for the 70/30-segmented polyurea is 0.949 under the assumption r=l. The measured gel conversion of 70/30- segmented polyurea, however, is feu? below the predicted value by

Flory’s theory. This can be attributed to the following reasons:

First, the large discrepancy of reactivity between aliphatic and aromatic amines violates the equal reactivity assumption in Flory’s theory. Secondly, the physical crosslinking of hard segments results, in a sharp increase of system viscosity. Compared to the molecular size of aliphatic triamine, the size of aromatic diamine is so small that its contribution to the increase of molecular weight during polyurea formation is relatively small. In addition, the reactivity of aliphatic amine is much greater than that of aromatic amine (Vespoli et al., 1986; Pannone and Macosko, 1987; Hsu and Lee, 1988), which further reduces the influence of aromatic diamine c h i the molecular weight growth of polyurea. Therefore, the viscosity rise of the polyurea reaction may reflect the formation of hard segment between aromatic diamine and diisocyanate, i.e., the gel conversion is dominated by the phase formation of hard segment. Whenever physical crosslinking occurs during polyurea reaction, the system tends to gel. 137 VISCOSITY vs. CONVERSION 0 0 Qv N M • T5000/11305/TBTDA 7 0 TRI/DI □ a D 70/30 7 ^ , V 20/80 0 o 7 >• O - IN 85% NB h •" 0 Si 7

0 7 0 0 w 7 0 0 > d , , , 1 , ff , 1 7fl , itP i .

0.0 0.2 0.4 0.6 0.8 1 0 0 l | 1 1 l

® 0 0 T5000/I1305 0) tv 0 0 11305/DMTDA 0 0 a v . IN 85% NB 0 o 0 >- o 0 h - \f) 0 0 0 0 0 0 0 (/) 0 0 0 >6 ft ft, 1 1 , 0..$° 0.0 0.2 0.4 0.6 0.8 1.0

CONVERSION (a )

Figure 3.16 Viscosity vs. conversion plots of TOTDA-based polyureas in 85% NB at 25 °C 138

Similar phenomena were also observed in DMTDA-based polyureas as shown

in Figure 3.17.

For comparison, the viscosity vs. conversion plots of TBTDA-based

and DMTDA-based polyureas with segment ratio 70/30 are reploted in

Figure 3.18, along with the plots of T5000/I1305, I1305/TBTDA, and

I1305/DMTDA polyureas. TOTDA-based polyurea systems possess faster

viscosity rises, but higher gel conversions when compared with DMTDA-

based polyurea systems. It means that the gel conversion or physical

crosslinking are strongly dependent on the chain extender used in the

polyurea reaction.

3.3.4.1 SOLVENT CONCENTRATION EFFECT

In order to apply the solution polymerization data to bulk polymerization of polyurea in RIM, the influence of solvent concentration and reaction temperature on the gelation of polyurea needs to be understood.

Figures 3.19 and 3.20 show the effect of solvent concentration on the gel time and gel conversion of I1305/DMTDA polyureas diluted with

85, 75, and 65% nitrobenzene. As shown in Figure 3.19, a polyurea system with a lower solvent concentration has an earlier onset and faster increase of viscosity. In addition, such a system has a higher reaction rate and higher limiting conversion. It seems that when a system reaches the gel point, its reaction tends to stop, i.e., its conversion curve tends to level off. The gel conversion of the

I1305/DMTDA polyurea system is also affected by the solvent » VISCOSITY vs. CONVERSION 0 0 cv N I) T5000/11305/DMTDA

0 TRI/DI 0V a + □ 7 0 /3 0 v . V 2 0 /8 0 o V + 1 0 /9 0 >- o "b - h if W 0 .|V 0 □ 0 +7 □ (/) +V D IN 85% NB

J, 0 > d 7 7. 4-Wfi.s u L ni_n_in_li_Qi— .— ,—

• 0.0 0.2 0.4 0.6 0.8 1.0

i | i i i—^ i i i | r— i "T

n C T5000/11305 9 m 0 0 11305/DMTDA 0 a V , IN 85% NB 0 0 > o h - 0

0 0 0 w 0 0 0 0 .0 s n .n . n O . i .1 . i . i i i i i i t ^

0.0 0.2 0.4 0.6 0.8 1.0

CONVERSION (a )

Figure 3.17 Viscosity vs. conversion plots of DMTDA-based polyureas in 85% NB at 25 °C VISCOSITY vs. CONVERSION 0 0 I ■ i 'T I I I" I I " I ■ i ■ C9 N Q T5000/11305/TBTDA 9 •w ■70/80/30(wl%) 0 A T5000/I1305/DMTDA □ a -70/68/30(wt%) ^ . □ o IN 85% NB >■ 0 h ’’ w A 0 A 0 A 0 (/) A 0 > d

D.O 0.2 * 0.4 0.6 0.8 1.01 0 0 1 1 1 1 1" 1 1 i | J 9 0 9 ? T5000/I1305 0 11305/TBTDA Q 0 0 V a 0 11305/DMTDA

IN 85% NB 0 0 o V > O 0 h ’■ 0 V i/i 0 0 0 0 V o 0 0 (/) 0 V V 0 0 _v >6 6 6, 1— i 1 1., 1 T. T- 0.0 0.2 0.4 0.6 0.8 1.0

CONVERSION (a )

Figure 3.18 Effect of chain extender on the gel conversion of polyurea reaction Figure 3. Figure

CONVERSION (cx) VISCOSITY (p o is e ) o n 0 0 0 8 1. 4 3. 40. 32. 24. 16. 8. 0. „ 0 0 N 0 0 0 19 Viscosity and conversion vs. time plots of I1305/DMFDA of plots time vs. conversion and Viscosity 19 • . . 6 2. 2 40. 32. 24. 16. 8. 0. polyureas in 85, 75, and 65% NB at 25 at NB °C65% and 75,85, in polyureas , + , 4 i ° 0 0 II 305/DMTDA 4 o + 5 NB 65% 0 + A 0 0 se v. IE CN v. TIME vs. CONV & TIME vs. e is V * a

TM: 5 C 25 TEMP: + A 4 -T . i I i i i I ■ ■ ■ !■ ■ ■ IE min) (m TIME + + + 8% NB 85% + 7% NB 75% A

141 01

0 Ql ^ . 0 > 0 h <-

5i 0 0 A 0 + (/) o ,* + 0 J4 , + ■+. T ■ .— i— I , i. i- i. I i i— 1_

0. 5. 10. 15. 20. 25.

TIME (m in) 0 0

0 A V)

0 + a 0 A v-' . + 85% NB 0 + A 75% NB >■ 0 A h *- 0 + 0 65% NB in A + o 0b 11305/DMTDA o + (/) TEMP: 2 5 C / + +

0.0 0.2 0.4 0.6 0.8 1.0

CONVERSION (a ) Figure 3.20 Viscosity vs. time and conversion plots of I1305/DMTDA polyureas in 85, 75, and 65% NB at 25«C 143

concentration as displayed in Figure 3.20. The system with a higher

solvent concentration tends to gel at a higher conversion. Similar

results were also found in DMTDA-based three-component polyurea

systems in 85, 75, 70 and 65% nitrobenzene as shown in Figures 3.21

and 3.22. For the 70%- and 65%-diluted polyureas, there exists an

initial viscosity jump at the very beginning of the reaction. This

initial viscosity jump results from the reaction of T5000 with 11305.

The initial viscosity jvxnp is followed by a viscosity plateau, then a

rapid rise of viscosity due to the physical crosslinking of hard

segment. On the other hand, for the soft segment only, i.e., DMTDA is

replaced by an equal amount of NB as shown in 70%-diluted polyurea in

Figure 3.22, there is only an initial viscosity jump followed by a viscosity plateau without any further rise of viscosity. In other words, the formation of soft segment, leading to a molecular weight growth, only contributes to a sudden viscosity build-up. The gelation

is determined mainly by the formation of hard segment, i.e., the physical crosslinking of hard segment determines the gel time and gel conversion of a segmented polyurea system.

3.3.4.2 REACTION TEMPERATURE EFFECT

Figures 3.23 and 3.24 show the effect of reaction temperature on the gel time and gel conversion of I1305/DMTDA in 85% nitrobenzene at

25, 47, and 6 7°C. As shown in Figure 3.23, at a higher reaction temperature, the system has an earlier onset and fast increase of viscosity rise. In addition, such a system has a higher limiting Figure 3.21 Viscosity and conversion vs. time plots of EMITA-based of plots time vs. conversion and Viscosity 3.21 Figure

CONVERSION (ex> VISCOSITY (p o is e ) O O O.S 1.0 0 . 1 O O . 2 0 0 . 0 2. 0 4. 50. 40. 30. 20. 10. 0. . 0 2. 0 4. 50. 40. 30. 20. 10. 0. 0 0 ° 4 >0000° - and 65% NB at 25 at 75, NB 85, 65% «Cin and soft/hard=70/30 ratio segment with polyureas O + 1 + 0 a -- T , 0 se v. IE OV s TIME vs. CONV & TIME vs. e is V 0 + + 0 1 -- + 0 . 4 0 A 0 0 ° n 1 + o | — 0 4 o a --

+ 4 A * 0 1 'T | "T T + A k

IE min) (m TIME 0 -- A A 1 -- 1 -- + ±

| 1 |1 0 * 5000/11305/DMTQA T -- + + ♦ 68/ wt) (w 0 /3 8 /6 0 7 = 1 | ' 1 1 1 1 1 + ; + + * E P 5C ■ TEMP:25 C 0 0 A 1 i ... i i 75% NB 75% NB 85% 65% NB 65% + +

(j 145 0 i ■ i i "| i i ■ i [ -11 i i | i i i | i i 'J N 7 0 A 0 T5000/11305/NB II 0 =70/68/30 (wt) 0 7 a IN 70% NB a 0 V , 7 0 * 0 >■ 0 177777 h «- 0 VI 0 0 jorawtfboooooooo 0 0 W A 0 □ > o ■ rvA .w ■ aii . n. n i ri ,n l n 1 » 1 1 1 ._i_ 1.

0. 10. 20. 30. 40. 50.

TIME (min) 0 0 7 A « TEMP: 25 C

0 7 0 85% NB A a A 75% NB V 0 70% NB o 7 > 0 7 65% NB - h ■■ » » * f 4

i/i 0 o A° T5000/I1305/DMTDA o 0 i/) A° = 7 0 /6 8 /3 0 (wt) □

0.0 0.2 0.4 0.6 0.8 1.0

CONVERSION (a )

Figure 3.22 Viscosity v b . time and conversion plots of EMTOA-based polyureas with segment ratio soft/hard=70/30 in 85, 75, 70, and 65% NB at 25«C Figure 3.23 Viscosity and conversion vs. time plots of II305/DMTDA of plots time vs. conversion and Viscosity 3.23 Figure

CONVERSION Cot) VISCOSITY (p o is e ) 0 O 0.5 1 .0 0 . 1 OO. 2.0 0 . 0 . polyureas in 85% NB at 25, at NB 85% in polyureas Vise 8 s TM & CN v. TIME vs. CONV & TIME vs. . IE min) (m TIME 16. 16. 24. 47, and 67 and *C INNB 85% II 305/DMTDA 32. 32.24. 40. 40.

146 Figure 3.24 Viscosity vs. time and conversion plots of I1305/EWIDA of plots conversion and time vs. Viscosity 3.24 Figure

VISCOSITY (poise) VISCOSITY (poise) 0 0 M 0 0 0 0 0 N 0 . 02 . 06 0.8 0.6 0.4 0.2 0.0 .5. 0. 1.40- 1-4.1 polyureas in 85% NB at 25, 47, and 67 and 47,25, <€ at NB 85% in polyureas 1 0 + A * ,i i -4L_L r- r-i . . 1 . 1 ,kt\ 25 C 47 C 67 C / $ / ° 0 A 0 0 0 * 0 0 0 A A + A OVRIN a) (a CONVERSION A A + A6 + A0 + TIME (min) 0 15. 10. A0 + . ‘0 ‘ + “o “ + + to + A0 A + 0 + IN 85% N8 1 1 3 0 5 /D M T D A 0 2 20. + + 1.0 ■

147 conversion. Because of very tight 3-D hydrogen bonding in urea formation, when the hard segment forms, it will slow down, sometimes stop, the reaction. Therefore, a higher mold temperature in the RIM process is recommended for complete reaction of polyurea. The gel conversion of the I1305/DMTDA polyurea system is also influenced by the reaction temperature as displayed in Figure 3.24. The system at a higher reaction temperature tends to gel at a higher conversion.

However, when the reaction temperatures are between 47 and 67 °C, the gel conversions of II305/DMTDA polyureas seem less sensitive to temperature; and have values between 0.60 and 0.65. Similiar results were also found in the DMTDA-based three-component polyurea systems in

85% nitrobenzene at 25, 47, and 6 7 °C as shown in Figure 3.25. Only viscosity-time data are displayed. Apparently, the viscosity rise and gel time of DMTDA-based segmented polyureas are affected by the reaction temperature.

3.3.5 MODULI G» & G" CHANGES IN UREA AND URETHANE REACTIONS

Figure 3.26 shows viscosity rise, storage modulus G*, and loss modulus G" as a function of time for T0305/I143L and I143L/DMTDA reactions in 80% NB at 3 3 °C. T0305 (Union Carbide) is a polyol with a functionality of 3 and an equivalent weight of 180. I143L (Dow

Chemical) is an MDI-based aromatic diisocyanate with an equivalent weight of 144. Therefore, the T0305/I143L and I143L/DMTDA reactions represent the triol-based urethane (soft segment) and amine-based urea

(hard segment) reactions, respectively. Their viscosity rise, G ’, and VISCOSITY (p o is e ) + O +. O iue .5 icst s tm lt fDTAbsdplues in polyureas DMTDA-based of plots time vs. Viscosity 3.25 Figure 40. 80. 120. 160. 200. . 0 2. 0 4. 50. 40. 30. 20. 10. O. 305/DMTDA A D T M D / 5 0 1 3 1 / O O O S T N 5! B N 85?! IN C 7 6 O C 5 2 + 47 C C 7 4 A (wt) 0 3 / 8 6 / 0 7 - A A H- .O 85% NB at 25, 47, and 67 67 °C and 47, 25, at NB 85% T ♦ + A O ICST v. TIME vs. VISCOSITY O TIME (min) o 4 + A + Hi 149 VISCOSITY (poise) o o 40 . 80 . 120 . 160. Figure 3.26 Viscosity, G', and G" vs. time for T0305/I143L and T0305/I143L for time vs. G" and G', Viscosity, 3.26 Figure /-s CO \ If) w n 0 o o v o o -J o 0 C C * N . * <0 K)

O. are same as in Figure 3.27) in Figure as same are 13/MD ecin n8XN t33°C symbols (legend 3 3 at NB 80X in reactions I143L/EMTDA VISCOSITY, o o 8 fc O A o o & ir* o ^/patatA 6 . V $ ooooooooooooooooooo ooooooooooooooooooo AAAAAAAAAAAAAAAAAAAAjt I E (TIME m in ) * c ’* G 8c G* 1

2 . 7 V 1 □ n _

vs. 8 . o 0 0 + D O O ° ° n O 24 V V 7 TIME .

01 30 >

. 150 151

G" as a function of conversion is shown in Figure 3.27. As shown in

both Figures 3.26 and 3.27 for the reaction of I143L/DMTDA, when the

system reaches gelation, both G* and G" curves, versus either time or

conversion, tend to level off. However, for the reaction of

T0305/I143L, both G ’ and G" curves still develop after the system

reaches gelation. This seems to indicate that as long as the hard

segment forms in polyurea or polyurethane-urea RIM systems, the

structure and mechanical properties can not be changed. Therefore, the

molding condition in the polyurea RIM process has to be selected

carefully.

Figures 3.28 and 3.29 show the viscosity rise, G ’ and G" versus

time and versus conversion respectively for the T5000/I143L/DMTDA

polyurea reaction in 80% NB at 2 3 °C. From both figures, it is observed

that both G* and G", especially for the hard segment formation,

develop before or during gelation. For comparison, same plots for the

T0305/I143L/BDO (BDO = 1,4-butanediol) polyurethane reaction are shown

in Figures 3.30 and 3.31. The G' and G" seem to develop after

gelation.

3.4 SUMMARY AND (INCLUSIONS

The solution polymerization technique was used to slow down the reaction rate of polyurea such that Theological and kinetic

information could be obtained. A thorough understanding of the reaction kinetics, rheological change, and their relationships during reaction is helpful to figure out the roles which soft and hard VISCOSITY (poise) N o 4-0. 80 . 120 . 160. o Figure 3.27 Viscosity, G * , and G" vs. conversion for T0305/I143L for conversion vs. G" , and * G Viscosity, 3.27 Figure \ w /-s N v 0 o * o o _! 0 o _J ® • c © . E . X 10 00 K) O ICST, * G vs G” & G* VISCOSITY,

0.0

icqi OV RIN (ex)CONVERSION O—I iofl o I— Oo— 0.4

0.6

1 — 2 I- CONVERSION 0.8 1 1 1 1 1

.0 152 VISCOSITY (poise) N o 40. 80 . 120 . 160. o Figure Figure \ v-/ /-s « N o o 0 0 -J *0 0 £ ® c * . N 3.28 3.28 K) 10 .

o. AAAAAAAAAAAA a ^ Viscosity, G ’, and G" vs. time for for time vs. G" ’, and G Viscosity, reaction in in reaction ICST, ' <5c G' VISCOSITY, 12 80% NB NB 80% . A p 0 * a* Oaa A ° a°° at at 1143L/ 70/ 30( t) (w 0 /3 6 /4 0 -7 A D T M /D L 1 3 4 /1 0 0 0 5 T IE (min) TIME

24 °C 3 2 -. A^ A** ” G 36 T5000/I143L/I)MrDA a vs. A a a a a a a . EP 23 C 3 2 TEMP: N 05 NB 8055 IN O VISCOSITY VISCOSITY O A A G* □ a A G” j a a a a a a a a a a a a a a 48 TIME . 60

. 153 VISCOSITY (poise) CM 40 . 80 . 120. 160. o o v-x 10 \ /**\ N l O _l O O * o 0 o _l O V Figure 3.29 Viscosity, G ’, and G" vs. conversion for conversion vs. G" ’, and G Viscosity, 3.29 Figure • ® 0 . E ® c N X . M . CM ®0 K) ICST, G* VISCOSITY,

0.0 T5000/I143L/DMTDA reaction in 8OX NB at 23 23 *C at NB 8OX in reaction T5000/I143L/DMTDA O VISCOSITY VISCOSITY O G’ □ A o o o o o o o o o A A A A A&a & A A A A A A A A A £ Q D ° o° ° o □ o g G” 0.2

CONVERSION («) c ” G Sc 0.4 1,1,1

O vs. & 0.6 T 5 0 0 0 /I1 43L/DMTDA 43L/DMTDA /I1 0 0 0 5 T 30(wt) w ( 0 /3 6 4 / 0 7 =

EP 23 C 3 2 TEMP: t * t » N 0 NB 80% IN CONVERSION 0.8 I ■ ■ ■ ■

0 . 1 154 VISCOSITY (p o is e) 40. 80 . 120. 180 . Figure 3.30 Viscosity, G ’, and G" vs. time for T0305/I143L/BDO for time vs. G" ’, and G Viscosity, 3.30 Figure \ o o o O V C

. 155 VISCOSITY (poise) N o 40. 80 . 120 . 160. o r- \ N J o _J 0 O o O O * O TJ « • C ® 0 £ Figure 3.31 Viscosity, G * , and G" vs. conversion for conversion vs. G" , and * G Viscosity, 3.31 Figure « N 00 H) IC ST, * 3 G v. CONVERSION vs. <3c G*VISCOSITY, G” . .

0.0 I I I I 1143L/ 152/ ) t) w ( 0 /3 2 /15 0 7 = O D /B L 1 3 4 /1 5 0 3 0 T T0305/I143L/BDO reaction in 80% NB at 23 23 at °C NB 80% in reaction T0305/I143L/BDO EP 23 C 3 2 TEMP: N 0 NB 80% IN O VISCOSITY VISCOSITY O G’ □ G” A 0.2 I I I■I

CONVERSION (ex) 0.4 T

0.6 I I I'I

0.8 T 1.0

156 157

segments play in the segmented polyurea RIM process. The rheological

changes of polyurea reactions were measured using a Haake viscometer

to follow the viscosity rise before gelation. The reaction kinetics of

polyurea were monitored using FTIR to obtain conversion profiles

during reaction. Combining viscosity and conversion data, the

viscosity rise can be expressed as a function of conversion through

the mapping procedure.

The reaction of T5000/I1305 polyurea (i.e., aliphatic triamine and

diisocyanate only) is almost instantaneous. This reactive system has

an extremely fast viscosity rise and short gel time when the

stoichiometric ratio is equal to one. Such gelation is caused by

chemical crosslinking through the reaction of triamines and

diisocyanates. However, in a three-component polyurea system

(including aliphatic amine, aromatic amine, and equal molar amount of

isocyanate), the effective stoichiometric ratio is far from one during

the reaction of aliphatic triamine and diisocyanate because of the

relatively lower reactivity of the aromatic amine. Therefore, the

formation of soft segment in a three-component polyurea only contributes to a sudden viscosity build-up, instead of gelation.

The 11305/diamine polyurea (aromatic diamine and diisocyanate only) is a linear polymer and has a relatively low reaction rate compared with the T5000/I1305 polyurea. However, its solidification behaves like a gel system. Such gelation is caused by physical crosslinking through the formation of 3-dimensional hydrogen bonding.

It was observed in the three-component polyurea systems that the 158

system with a higher hard Begment content has a fast viscosity rise

and shorter gel time theui the one with a higher soft segment content.

For a three-component polyurea diluted with 65% nitrobenzene, there

exists an initial viscosity jump at the very beginning of the

reaction, which is followed by a plateau, then a rapid rise of

viscosity. This initial viscosity jump results form the fast formation

of soft segment. However, the viscosity rise and gelation of the

polyurea system are dominated by the formation of hard segment, i.e.,

when the physical crosslinking occurs, the system tends to gel.

Since the gel conversion of 11305/diamine polyurea was found to be

a function of aromatic diamine, the gelation behavior of a three-

component polyurea is strongly dependent on its chain extender. The gel formation of a polyurea was also found to be affected by the

solvent concentration and reaction temperature. For a given composition, the more the polyurea system was diluted, the higher the gel conversion was. The reaction temperature influences the gelation of a polyurea in such a way that at a higher temperature, the system has a shorter gel time, but a higher gel conversion. When the reaction temperatures are between 47 and 6 7 °C, the gel conversions of

11305/DPfTDA polyurea seem less sensitive to the temperature, and have values between 0.60 - 0.65. CHAPTER IV

REACTION INJECTION MOLDING OF POLYUREAS: II. BULK POLYMERIZATIONS AND SIMULATION

SYNOPSIS

An experimental and theoretical study of reaction injection molding of polyurea was conducted systematically in this chapter. A lab- scale RIM machine was used to carry out the polyurea bulk polymerizations. A "free-table" viscometer was designed to measure fast rheological changes and liquid-solid transition. A mathematical model is proposed to simulate the fast reaction and rheological changes in the polyurea RIM process. The parameters of this model are determined based on the solution polymerization data from FTIR and a Haake rheometer. Combined with an appropriate heat transfer equation, this model predicts fairly well the adiabatic temperature and viscosity rises of bulk polyurea reactions in RIM.

4.1 INTRODUCTION

Reaction injection molding (RIM) provides a high speed and relatively inexpensive production method for large polymeric parts from low viscosity reactants. The success of polyurethane RIM in the automotive industry has led to the development of improved resin systems with faster cycle times and better physical properties. The total polyurea system is a good candidate for RIM applications because

159 160 polyureas have excellent thermal and mechanical properties with short demold time. However, the current polyurea systems readt extremely fast with significant reaction occurring during impingement mixing and mold filling. This often results in poor mixing and premature gelation during flow. To avoid these problems, a thorough understanding of the rheo-kinetics of polyurea RIM is needed because this information can be used for better resin design and process optimization.

In this study, a series of RIM experiments were carried out. A mathematical model was developed to simulate the rheo-kinetic changes of polyurea reaction in the RIM process. Model parameters were determined from solution polymerization data measured by several analytical instruments.

4.2 EXPERIMENTAL

4.2.1 MATERIALS

A T5000/I1305/DWTDA based polyurea system was used in this study.

Detailed information on T5000, 11305, and DWTDA are given in Section

3.2.1. Most of polyurea systems in this work have a weight ratio

T5000/I1305/DWTDA=70/68/30 (i.e., 13/100/87 by mole). For the reaction of aliphatic triamine and diisocyanate, an equivalent amount of nitrobenzene (NB) was used to replace DMIDA; while for the reaction of aromatic diamine and diisocyanate, the T5000 was replaced by the same amount of NB. 161

4.2.2 INSTRUMENTATION AND PROCEDURE

4.2.2.1 REACTION INJECTION MOLDING

A laboratory scale RIM machine was used to carry out the RIM

experiments. The machine design is shown in Figure 4.1 (Nelson, 1987).

Both the hydraulic drive system (item 1 to 7) and the material mixing

system (item 8 to 10) are attached to a common support frame. The

hydraulic drive unit consists of a 7.5-horsepower variable flow rate

hydraulic pump and a 10-gallon hydraulic oil reservoir. A three-

position directional solenoid valve is used to control the flow

direction. Flow rates into the individual drive cylinders are

controlled by uni-directional flow control valves with load pressure

compensation. The drive cylinders for this RIM machine are 8.255 cm

(3.25-inch) diameter hydraulic cylinders with 7.62-cm (3-inch) stroke

length. These cylinders are connected directly to the material

cylinders which have a smaller diameter but with the same stroke

length. This machine is capable of delivering up to 250 ml of liquid

at rates up to 125 ml/sec and a maximum pressure of 2,000 psi in the material cylinder. Using a 0.635-cm (1/4-inch) diameter mixing chamber with 0.0794-cm (1/32-inch) diameter nozzles, these flow rates are able

to produce nozzle Reynolds numbers (Re) in the order of 300 to 500 for

the reaction systems explored in this work. An IBM PC/XT equipped with a Metrabyte DASH-16/EXP-16 interface board was linked to the lab-scale

RIM machine for real time data aquisition. This system is shown in

Figure 4.2 (Nelson, 1987). The machine’s reasonably small size allows (T) control panel

(2) pump motor

@ hydraulic pump

(!) directional solenoid

(5) hyd. reservoir

(e) flow control valves

(7) drive cylinder

5) material cylinder

(5) m ixhead OSU Lab Scale RIM Machine © mat’l storage tanks

Figure 4.1 Schematic diagram of the laboratory scale RIM machine (Nelson, 1987) Data Acquisition Schematic Diagram

VAX B500

IBM XT

DAS1 H-16 A/D ca inverter ft m u Llnlexer

STA08 connection Exp—10 LVDT P ressure TC Amplifier signal signal board conditioner .conditioner

I

Figure 4.2 Schematic diagram of data acquisition system (Nelson, 1987) 164

it to be portable and it can be interfaced with other analytical devices. A more detailed description of this machine and its operation can be found elsewhere (Nelson, 1987).

T5000 and/or DMTDA were blended and loaded into one storage tank of the RIM machine, and 11305 was loaded into the other. All materials were degassed and demoistured under vacuum at 40 °C for at least 12 hours to remove water and air before loading. After loading, the tanks were blanked with nitrogen. No catalyst was added to the system. The measurements of adiabatic temperature rise and viscosity rise are described in the next two sections.

4.2.2.2 ADIABATIC TEMPERATURE RISE MEASUREMENT

To measure the adiabatic temperature rise of polyurea, an insulated beaker was used as the adiabatic reactor with a thermocouple inserted in the center and about 1 cm from the bottom of the beaker.

The reaction was so fast that the error due to the adiabatic assumption was negligible. After the center temperature reached a maximum, it cooled down at a rate less than 0.2°C/min.

Based on several assumptions, the temperature rise can be related to the extent of reaction. The assumptions used here are:

1. constant heat capacity, C , density, p, and heat of reaction, P

AH over the temperature range;

2 . homogeneous reaction mixture after an initial mixing

period; 165

3. no heat sources other than reaction;

4. no diffusion restrictions on the reaction rate.

The energy balance for a single irreversible reaction, including heat

loss is:

p c = -AH r - U (T - T ) (4-1) p at m

Where U is the heat transfer coefficient per unit volume, r is the

reaction rate, and is the ambient temperature. The rate can be

expressed in terms of the initial concentration of the limiting

reagent, Cq , and the extent of reaction, a:

p °p - I - = -1H co - u

To simplify data analysis, the temperature values were corrected for

heat loss prior to calculating conversion. The heat transfer

coefficient was determined using a non-reactive system. Without

reaction, the energy balance simplifies to:

P C - ™ = - U (T - T ) (4-3) p at m

With the assumption of constant p and Cp , Equation (4-3) integrates

to:

In

Where Tq is the initial temperature of the reactants and tg=0. The

value of U is calculated from the slope of a plot of In (T -

T ) vs. (t - tg). All temperature data are corrected by the addition

of the following factor through Simpson’s rule of integration.

<75"> Jtn (T(t> - V dt (4-6> p 0

Once heat loss has been compensated for, temperature and fractional

conversion are directly related to each other, i.e.,

P C (T - T ) a = P (4-6) -AH CQ ' '

The heat of reaction is calculated from the maximum temperature attained, i.e.,

0 " Tn) -AH = B— (4-7) L0

Combining Equations (4-6) and (4-7), the final expression for a is:

T - T oc = ° (4-8) ad " 0

Because of the extremely fast reaction of polyurea and the low thermal conductivity of polymer, the correction term, i.e., Equation (4-5), can be neglected. 167

4.2.2.3 VISCOSITY RISE MEASUREMENT

Two Brookfield viscometers, Models RVT and HBT, were used to carry out the viscosity measurements for the mixture of reacted T5000 and

11305 in NB depending on the systems* viscosities. Since about 8 seconds must be allowed for the viscometer to reach an apparent equilibrium, both Brookfield viscometers are not suitable for the systems having very short gel time or extremely sharp viscosity rise.

For three-component polyureas, a special device was designed to measure the fast viscosity rise and gel time. The schematic drawing of this device is shown in Figure 4.3. An insulated disposable beaker

(250 ml) is used as the material container. This container is fastened on a free table with a diameter of 13.5 cm. Along with the container, the free table is mounted on a fixed table. Underneath the fixed table, there is an attached potentiometer which is connected to the free table through a rod. The potentiometer serves as a RVDT (rotary variable displacement transducer) and its signal goes to an IBM PC/XT and a chart recorder. A coil spring with one end bound to the fixed table and the other to the free table is used to confine the movement of the free table. For different reaction systems, a suitable coil spring can be selected based on the system’s initial viscosity level.

A disposable thin glass stirrer is used as a spindle which is driven by a 1/4-HP motor through a gearbox. This set of motor and gearbox can keep the spindle at a constant rotation speed. The rotation speed of the spindle is set at 60 rpm which is similar to that used in the

Brookfield viscometer measurement. As the viscosity of the reaction 168

motor gearbox

glass stirrer

beaker insulator

free table

fixed table coil spring

to computer potentiometer

Figure 4.3 Schematic diagram of the "Free-Table rheometer" 169

system in the container increases, the amount of torque transforming

from spindle to free table is increased, which stretches the coil

spring. The extent of rotation of the free table can be detected

instantaneously by the potentiometer and can be calibrated using a

series of polystyrene solutions with known viscosities. When the viscosity of the system reached a very high level, the glass spindle would break, which protected the rest of the device. The entire device

is mounted on a portable support frame which can be moved easily to the lab-scale RIM machine in order to deliver the impinge-mixed resins to its container rapidly.

4.3 RESULTS AND DISCUSSION

4.3.1 REACTION INJECTION MOLDING OF POLYUREA

Figure 4.4 shows the adiabatic temperature rise and viscosity rise curves for two two-component, hard segment only systems,

NB/I1305/DMTDA=70/68/30 and 50/104/50 by weight, measured using the lab-scale RIM machine. The maximum temperature rises (AT^) were 81 °C and 113°C for NB/I1305/DMTDA=70/68/30 and 50/104/50 (wt), respectively. Hie time to reach solidification was about 80 seconds for the former case and 48 seconds for the latter case. The experimental results shown in Figure 4.4 are replotted in Figure 4.5 where viscosity is expressed as a function of conversion. Here, the conversion is determined based on Equation (4-8). For LOG VISCOSITY (poise) 0.5 2.0 3.5 5.0 Figure 4.4 Adiabatic temperature and viscosity rises of hard of rises viscosity and temperature Adiabatic 4.4 Figure d - I- Id 2 . Q O 5 N o <0 o o co O * • o. segment reaction measured by RIM by measured reaction segment ICST

240 TIME . 300

. 170 VISCOSITY (poise) o o O O T— Figure 4.5 Viscosity vs. conversion for the same data shown in shown data same the for conversion vs. Viscosity 4.5 Figure O O O ^ 0.00 --- ICST v. CONVERSION vs. VISCOSITY Figure 4.4 Figure 0.20 OVR IN (ot)CONVERSION NB/ DMTDA- 104-50( t); (w 0 -/5 4 0 /1 0 -5 A D T M /D 5 0 3 1 /1 B N O a 0.40 11305/ 70/ 30( ) : t) (w 0 /3 8 /6 0 -7 A D T M /D 5 0 3 1 /1 B N ------0.60 o a PREDICTION 0.80 1 172

NB/11305/DWTDA=50/104/50 (wt), it shows an earlier onset and faster rise in viscosity compared to the system of NB/I1305/DMTDA=70/68/30

(wt). The latter system did not show much viscosity rise until a conversion of 30%. However, for both systems the reaction reached solidification around 60% of total conversion.

Figure 4.6 shows the experimental results of adiabatic temperature rise and viscosity rise as a function of time for a three-component polyurea RIM reaction (i.e., T5000/I1305/DMTDA=70/68/30 by weight).

The maximum temperature rise is 123 °C at an initial material temperature of 50 °C. About 80% of the adiabatic temperature rise took

4 place within 10 seconds. A sudden jump in viscosity (i.e., 8x10 poise) at the beginning of the reaction can be easily detected, which is due to the fast reaction of aliphatic triamine with diisocyanate.

This sudden viscosity jump is followed by a plateau viscosity region

4 for about 3 seconds before a sharp increase in viscosity from 8x10 to

5 5.6x10 poise due to the phase formation of hard segments (i.e., the reaction of aromatic diamine and diisocyanate). The initial viscosity jump indicates that the reaction of aliphatic triamine and diisocyanate occurs right after mixing, which makes the mold filling a difficult task in polyurea reaction injection molding because a high viscosity, "thermoplastic melt"-like mixture, instead of a low viscosity mixture, needs to be punrped into the mold. Figure 4.7 replots the viscosity rise as a function of total conversion for the reaction. Again, the total conversion is calculated from Equation LOG VISCOSITY (poise) 4.0 4.4 4.8 5.2 5.6 6.0 o id CL 2 Figure 4.6 Adiabatic temperature and viscosity rise rise in rise rise viscosity and temperature Adiabatic 4.6 Figure O o O o 10 (0 o 4 t - • •

0.0

VISCOSITY VISCOSITY polyurea RIM reaction RIM polyurea 4.0

I 305/ 68/ wt) (w 0 /3 8 /6 0 7 = A D T M /D 5 0 3 /I1 0 0 0 5 T IE (sec) TIME 8c 8.0 EP s TIME vs. TEMP

------T - 5 0 C (BULK) (BULK) C 0 5 - T O VISCOSITY VISCOSITY O □ TEMP TEMP □ 12.0 PREDICTION

16.0

20.0 173 VISCOSITY (poise) Figure 4.7 Viscosity vs. conversion for the same data shown in shown data same the for conversion vs. Viscosity 4.7 Figure O o O 0.0 ICST v. CONVERSION vs. VISCOSITY Figure 4.6 Figure 0.2 O VRIN (a) CONVERSION I 305/ 68/ wt) (w 0 /3 8 /6 0 7 = A D T M /D 5 0 3 /I1 0 0 0 5 T 4. .4 0 ------T - 5 0 C (BULK) (BULK) C 0 5 - T XEIETL DATA EXPERIMENTAL * 0.6 PREDICTION 0.8 1 175

(4-8). The reaction reaches solidification around 60% of total

conversion.

4.3.2 RHEO-KINETIC MODEIS FOR POLYUREA REACTION

The feet reaction of urea formation allows for a shorter in-mold cure time and, consequently, improved productivity. While the rapid reaction rate is advantageous for speeding the curing rate, a major drawback is that mold filling becomes more difficult and problems such as premature gelation, weld lines and uneven flow may occur. To avoid these problems, a better understanding of the reaction kinetics and

Theological changes is needed such that better control of mixing and mold filling can be achieved. A relevant rheo-kinetic model, i.e., constitutive relationships among material and processing variables, would be very valuable because it can be used to predict the molding pressure, viscosity rise and resin conversion during mold filling and curing in the RIM process. This information can be used to optimize the compound design and the RIM process.

4.3.2.1 REACTION KINETICS

Since the polyurea reaction is a step-growth polymerization, a simple n-th order model reaction with Arrhenius temperature dependence was proposed in this study (Hsu and Lee, 1988). In the model, the reaction kinetics of isocyanates with aliphatic amines and aromatic amines are treated separately as follows: 176

-dC1 n 1 -E1 n 1 rl = “ a t " = K 1 C 1 °i = A 1 exP<--Rf-> C 1 C? (4-9a)

n« "Ep rirt r 2 = " d t ” = K 2 C 2 = A 2 exP<-g|-> C 2 ^ «4-9b» where r^ and rg are reaction rates of isocyanates with aliphatic and aromatic amines, respectively; , Cg, and are concentrations of aliphatic amine, aromatic amine, and isocyanate functional groups; and Kg are rate constants; and Ag are frequency coefficients;

and Eg are activation energies; R is the universal gas constant; n^ and ng are reaction orders of aliphatic and aromatic amines; and m is the reaction order of isocyanate functional groups. Major assumptions used in the kinetic model include:

1 . homogeneous and well-mixed system at t=0 ;

2 . reaction order is the same throughout the entire

polymerization;

3. equal reactivity for the same amines.

4.3.2.2 ENERGY BALANCE

For heat transfer in cylindrical coordinates, the following assumptions are made:

1 . one-dimensional heat conduction;

2 . homogeneous and well-mixed reaction system at t=0 ;

3. temperature independent physical properties such as density, p, heat capacity, Cp, heat of reaction, AH, and thermal

conductivity, k;

4. same heat of reaction, AH, for the reactions of aliphatic and

aromatic amines with isocyanates.

With these assumptions, the governing equations for heat transfer can be described as follows (Hsu and Lee, 1988):

» cp -|f -- k [<-”f>+<-f -|f>1 - r, - M2 r2 (4-10)

where T is temperature. The initial conditions are:

II £ H at t = o, for 0 V ± d (4-lla)

at t for 0 £ r ^ d (4—1lb) C 1 = C 1 0 ’ = o,

at t for 0 £ r ^ d (4-llc) C 2 = C2 0 ’ = o,

at t for 0 £ r £ d (4—1Id) Ci = Ci0 ’ = o, where CgQi and CLq are initial aliphatic amine, aromatic amine, and isocyanate concentrations respectively. The boundary conditions 178

The adiabatic reaction in the RIM process is a special case of the energy balance Equation (4-10). With gel time on the order of 2-5 seconds and low thermal conductivity of polymer, minimal heat is lost

from the extremely fast reacting polyurea. Therefore, the heat conduction term and the boundary conditions can be removed from the equation. Assuming no heat exchange with the surroundings, the energy equation becomes:

0 Cp - i - = - iHl rl - 4H2 r2 <4-13>

The temperature rise of the system results from the amount of heat generated by the reaction. The heat of polyurea reaction can be calculated from the maximum adiabatic temperature rise, AT^, assuming constant density and heat capacity, i.e.,

-AH = p Cp A T ^ / C0 (4-14)

Where Tq is the initial resin temperature and Cq is the initial concentration of the functional group.

4.3.2.3 RHEOLOGICAL CHANGES

In this work, the viscosity rise of polyurea RIM is analyzed based on a model similar to the one proposed by Castro and Macosko (1980).

This model has been used in the studies of polyurethane and 179 polyurethane-urea RIM systems (Castro and Macosko, 1980, 1982; Vespoli et al., 1987). The model is expressed as follows:

" = \ “ P < - B F > (- 5 ^ - ,f<<<'T, «4- 15' where a* = a - D i , a ’ = a - D, g g 1 1

: pre-exponential factor in initial viscosity expression

: viscosity activation energy

0Cg : apparent gel point

a : extent of reaction

Dj : molar fraction of aliphatic amine

Equation (4-15) implies that the system viscosity, F)» is afunction of temperature, T, and reaction conversion, a. As a approaches the gel point, 0Cg, n becomes infinite. At the beginning of the reaction (i.e., oc 0 ), n becomes:

E n = An exp (-£j;-) (4-16)

For polyurethane RIM systems, A^ and E^ can be determined from the weighed average of the viscosities of the unreacted components as a function of temperature, oc^ can be approximated using various branching theories (Flory, 1953; Macosko and Miller, 1976) for 180 gelation based on chemical crosslinking. The function, f, has to be determined by experimental results.

From the observations in solution polymerizations and RIM experiments (i.e., Figures 3.22 & 4.6), it has been found that the sudden viscosity jump at the beginning of the reaction results from the reaction of aliphatic triamines and diisocyanates. After a plateau region, a gradual increase in viscosity until the mixture solidifies is due to the reaction of aromatic diamines and the remaining diisocyanates. A superposition rule is used here to determine the parameters of the model. In other words, A^ and E^ in Equation (4-15) are determined from the reaction of soft segment (i.e. aliphatic triamine and diisocyanate), while and f(a,T) are determined from the reaction of hard segment (i.e. aromatic diamine and diisocyanate).

For the reaction of aliphatic triamine and diisocyanate, an equivalent amount of unreactive solvent, nitrobenzene (NB), is used to replace

DMTDA; while for the reaction of aromatic diamine with diioscyanate, the T5000 is replaced by the same amount of NB. Results from these two types of reactions are used to evaluate model parameters.

4.3.3 ESTIMATION OF MODEL PARAMETERS

4.3.3.1 KINETIC PARAMETERS

The kinetic parameters in Equation (4-9) were determined from solution polymerizations measured by FTIR. Figure 4.8 shows the CONVERSION ( a ) Figure 4.8 Conversion vs. time measured by Fi'JLK by II305/DMTDA measured time for vs. Conversion 4.8 Figure 0.00 0.25 0.50 0.75 1.00 . 8.0 0.0 reaction in 85% NB 85% in reaction 1 35DMT A TD M 11305/D IE mi ) in (m TIME 1

0 . 6 PREDICTION FITTINGCURVE - 11 305/DMTDA= 1 /1 (mol) (mol) 111 /1 305/DMTDA= __ _ 24.0 N 5 NB 85% IN

32.0 181 measured conversion vs. time for reactions of I1305/DMTDA=1/1 by mole in 85% NB at three temperatures, 25, 47, and 67°C. Equation (4-9) can be rewritten in the following form:

nj-l log(doc^/dt) = log(Aj C jq exp(-Ej/RT)] + nj log(l-

= Gj + nj log (1 —oc^)

nX-1 log(da2/dt) = logtAg C2Q exp(-E2/RT)] + n£ log(l-a2 ) (4-17b)

= G2 + n 2 log(l-a2 )

where and a2 are fraction conversions of isocyanates with aliphatic and aromatic amines respectively and nj and n£ are the overall reaction orders of soft and hard segments, respectively. To obtain the reaction rate, doc2/dt, at each reaction time, the conversion versus time data shown in Figure 4.8 are fitted by polynomial curves.

Plotting log(da2/dt) versus log(l-a2) for three temperatures yields three parallel straight lines as shown in Figure 4.9. Since the peak change of isocyanate is relatively small at high conversions, FTIR data tend to scatter. Those high conversion data are not included in the determination of kinetic parameters. The intercept in Figure 4.9 gives factor, Gg, and the slop>e gives the overall reaction order, n2< The Gg’s at three temperatures versus 1/T are then plotted as LOG(dcx /d t) o 0 o K) o * O o O O - .0 05 —.2 02 —.4 0.00 —0.14 —0.28 —0.42 —0.56 1-0.70 1 Figure 4.9 Plots of logfdttg/dt)of Plots logd-otg) vs. 4.9 Figure 11 305/DMTDA= 1 /1 (mol) (mol) 1 11 /1 305/DMTDA= N 5 NB 85% IN 1 35DMT A TD M 11305/D ) « — 1 ( G O L

2 183 184

shown in Figure 4.10. The frequency coefficient, Ag, and the

activation energy, E^, are obtained from the intercept and the slope

g of the straight line shown in Figure 4.10. The results are A 2=8 .5xl0 ,

E2=4.71 Kcal/g-mole, and n£=2.33. The same procedure is used to

determine A^ , E^, and nj for the reaction of T5000/I1305. The results

are Ag=3.14x10^, E2=1.60 Kcal/g-mole, and n ’=2.10 (Hsu and Lee, 1988).

The kinetic parameters and other physical properties used in the

model are listed in Table 4.1. For comparison, the kinetic parameters

of I1305/TBTDA reaction are also included (Hsu and Lee, 1988). The

reaction order of isocyanate is assumed to be 1 , and those of

aliphatic amine, T5000, and aromatic amine, EMTDA, are assumed to be

1.1 and 1.33, respectively. The calculated activation energies are

similar to those mentioned by others (i.e., 1 - 8 Kcal/g-mole)

(Vespoli and Alberino, 1985; Pannone and Macosko, 1987). Experimental data in Figure 4.9 show some deviations from linear regression. This seems to indicate that reaction order changes during polymerization

(Borkent, 1974). However, the prediction of the conversion profiles of

DMTDA reacted with 11305 based on these parameter values is fairly good as shown in Figure 4.8. 185 0.0034 IN IN 8555 NB 0.0032 11 305/DMTDA— 1 /1 305/DMTDA— 1 (mol)11 1/T 11 305/D11 M TD A 0.0030 Figure 4.10 Plot of vs. 1/T

o 0*0 O* t — O'Z 186

Table 4.1 Parameters Used for Modelling of Polyurea Reaction (T5000/11305/DMTDA)

SyBtem 1 System 2 System 3 T5000/I1305 II 305/DMTDA T5000/I1305/ DMTDA

A (app. unit) 3.14 x 10' 8.5 x 10° (2.0 x 105 )

E (kcal/gmole) 1.60 4.71 (4.21)

n ’ 2.10 2.33 (2.15)

0.13 0.87 (0.889)

-AH (kcal/gmole) 15.4 (16.2)

Cp (cal/g/ °C) 0.4

p (g/c.c.) 1.118 (1.081)

-8 A^ (app. unit) 46.6 1.963 x 10 46.6

(kcal/gmole) 4.776 4.672 4.776

a 0.734 0.734

b 2626 2626

-8.322 x 10' -8.322 x 10'

* values in ( ) are for TBTDA-based systems, i.e., the DMTDA was replaced by an equivalent amount of TBTDA 187

4.3.3.2 RHEOLOGICAL PARAMETERS

4.3.3.2.1 ALIPHATIC TRIAMINE/DIISOCYANATE REACTION

The parameters in Equation (4-15) are determined by rheo-kinetic measurements of solution polymerizations. For the reaction of aliphatic triamine and diisocyanate, Figure 4.11 shows the viscosities of reacted mixtures of T5000/I1305/NB=70/68/30 by weight at different temperatures and solvent concentrations. The viscosities were measured by a Brookfield viscometer, either Model RVT or Model HBT, after the reactants were mixed in a lab-scale RIM machine. The viscosity rise

5 reached 1.5x10 poise for the system equivalent to a RIM reaction. The same reaction in 35% nitrobenzene resulted in a much smaller viscosity jump (i.e. 20 poise).

When viscosity data shown in Figure 4.11 are plotted as a function of 1/T in the log-scale, a linear relationship is obtained for all systems at different solvent concentrations as shown in Figure 4.12.

This means that adding solvent does not change the temperature dependence of reacted soft segments. For reacted soft segments, the temperature dependence of viscosity is assumed to follow the Arrhenius relationship:

n = exp(Erj/RT) (4-16)

The A^ and E^ can be determined from the intercept and the slope of the linear regression line. The results are A^=46.6 and E^=4.776 VISCOSITY (p o is e )

Figure 4.11 Viscosity as a function of XNB for the reaction of reaction the for XNB of function a as Viscosity 4.11 Figure 10° 101 102 103 104 io5 10 . O 2. 30. 20. IO. O. ■ * * 1 ■ 1 ■ • ■ « 1 * ■ ■ 1 * * ■ * * ■ 1 T5000 and 11305 and T5000 ICST v. %NB vs. VISCOSITY O A . T -1 1 1 - t - j- 1 1 1 | - 1 - .T i. — 1 00I NB-70/ 30(wt) w ( 0 /3 8 /6 0 7 - B /N 5 0 3 5000/I1 T %NB a 0 £ □ A 6 C 67 O 2 C 25 O 4 C 47 A o = 7.7 r=» /7 1 11 r- 1 — 4-0. VISCOSITY vs. 1/TEMP © O

T5QOO/11 305/NB**70/68/30(wt)

o

r - 1 /7 .7

O IN BULK 10% NB 20% NB O 30% NB

O

O

o

0.0027 0.0029 0.0031 0.0033 0.0035 1/T (1 /°K)

Figure 4.12 Viscosity vs. 1/T for the same data shown in Figure 4.11 00 CO 190

Kcal/g-mole for the T5000/I1305 system.

Figure 4.13 shows the viscosity rise of soft segment as a function of stoichiometric ratio with r (aliphatic NHg/NCO) ranging from 1/7.7,

1/16.6 to 1/34.6. These systems are equivalent to the

T5000/I130f>/DMIDA=70/68/30, 50/104/50, 20/160/80 by weight, in which

DMTDA is replaced by an equal amount of nitrobenzene. There exists a linear relationship when log viscosity is plotted against the r ratio.

This means that by changing the composition of a three-component polyurea system, the initial viscosity rise can vary a lot. This would affect the molding pressure in the RIM process.

The viscosities measured by the Brookfield viscometer are the viscosities at very low shear rates. These may not be representative of the mold filling step in the RIM process because the shear rate during mold filling is in the range of 1000 sec Therefore, a capillary device was used to measure the viscosities at different shear rates for the system of T5000/I1305/NB=50/104/50 by weight. The schematic diagram of the device and experimental results are shown in

Figure 4.14. Basically, this capillary device has a material reservoir of 5.08 cm (2 inch) diameter by 7.62 cm (3 inch) length and a capillary of 0.127 cm (0.05 inch) diameter by 5.08 cm (2 inch) length.

Ibis device is designed in such a way that after the reacted mixture is injected into the reservoir from the RIM machine, it can be sealed and a pneumatic pressure up to 500 p.s.i. can be applied through a nitrogen cylinder. By measuring the pressure and flow rate VISCOSITY vs. r RATIO (0 O

O

O

O

T 5 0 0 0 /1 1 30 5 /N B ( DMTDA)

O

0.00 0.04- 0.08 0.1 2 0.1 r RATIO (NH /NCO) 2 Figure 4.13 Hie effect of r ratio on the viscosity rise of the reaction of T5000 and 11305 VISCOSITY (poise) 10 O O o T— ' 1 o 2 io ! 1 o'* 10 T— Figure 4.14 Viscosity vs. shear rate for the reaction of reaction the for rate shear vs. Viscosity 4.14 Figure — 1 ICST v. HA RATE SHEAR vs. VISCOSITY 11305/ = B /N 5 1 0 /1 3 0 0 0 5 T 11305/ = B /N 5 1 0 /1 3 0 0 0 5 T C 25 TEMP: 1 SO(wt) / 4 0 /1 0 5 30( t) (w 0 /3 8 6 / 0 7 o T5000/I1305/NB at different weight ratios weight different at T5000/I1305/NB H A RATE(1/sec) SHEAR 1 2 3 4

5

6 192 relationship, the apparent viscosity can be calculated (Wang, 1985).

At shear rates higher than 1000 sec *, the viscosity-shear rate

follows the power-law with an index of 0.24. The viscosity at a shear

rate of 1000 sec * is about 800 poise for the T5000/I1305/NB=50/104/50

(wt) system. Therefore, in polyurea RIM, a highly viscous melt-like mixture, instead of a low viscous liquid, is formed in the mixhead and needs to be pumped into the mold. This is one of the major differences between polyurethane RIM and polyurea RIM.

4.3.3.2.2 AROMATIC DIAMINE/DIISOCYANATE REACTION

The viscosity vs. time data for polyurea solution reactions were obtained from a Haake rheometer at three temperatures, 25, 47, and

6 7 °C. The corresponding conversion vs. time data were generated from

FTIR measurements. Figure 4.15 shows the viscosity rise as a function of conversion for I1305/DMTDA=1/1 by mole in 85% nitrobenzene. The and of Equation (4-15) for the reaction of I1305/DMTDA can be determined from the initial viscosity i.e., when a=0 (no reaction),

Equation (4-15) becomes

(4-18)

The initial viscosity is modeled assuming perfect mixing of the reactants, i.e., VISCOSITY vs. CONVERSION

PREDICTION IN 85% NB

O o

o m

11 305/DMTDA— 1 /1 (mol)

o

0.00 0.20 0.40 0.60 0.80 1.00 CONVERSION (ex)

Figure 4.15 Viscosity vs. conversion for hard segment in solution polymerization (85% NB) <£> 195 where is initial viscosity, w is weight ratio, and i is nitrobenzene, 11305, and DMTDA. In order to determine the function f(oc,T), Equation (4-15) can be rearranged to the following form:

A = [log(n/An)-En/RT]/log[a^/(a^-a')] = f(oc,T) (4-20)

The plots of A vs. a are shown in Figure 4.16. These curves can not form a single curve. However, when A is replotted as a function of oc’/oc’T as shown in Figure 4.17, the curves of A vs. a ’/oc’T tend to B O merge together and can be expressed roughly by a single curve for roost of the data points. This curve can be expressed by a second-order polynomial:

A = a + b a" + c a "2 (4-21) where a" = a ’/a^T

The calculated values of A^, E^, a, b and c are listed in Table 4.1.

Based on these parameter values, the viscosity rises in the solution polymerization of II305/DMTDA can be reasonably fitted as shown in

Figure 4.15.

4.3.4 MODEL PREDICTION

The kinetic and rheological parameters, determined from the solution polymerization data, combined with the heat transfer model

(i.e. Equation (4-13)) give a good prediction of RIM reactions based Figure 4.16 A A Figure4.16 .27 .54 .81 .18 0.7415 0.6118 0.4821 0.3524 0.2227 0.78 1.58 2.39 3.20 4.00 • ■ ' • ; ■ • ACLTO FR b & c & b, , a FOR CALCULATION (85% NB) . b v oc’ for segmenthard in solution polymerization □ D * °D A □ A D 5 C 25 D O < A* * X O A O 7 C 67 O O A 11 305/DMTDA= 305/DMTDA= A I 8% NB 85% IN O \ O A O A o A A O a o o

O O - C 4-7 1 / 1 mo) - ol) (m

CALCULATION FOR a, b, Sc c

O N K)

0) K) CM

00 in «

11 305/DMTDA-1 /1 (mol) IN 85% NB 00 N • O 0.0014- 0.0019 0.0024 0.0030 0.0035 CX’/ o t ’T 9

Figure 4.17 A vs. a ’/oCgT for the same data shown in Figure 4.16 on hard segment only as shown in Figure 4.5.

In the three-component polyurea reaction, the term A^exp(E^/RT) in

Equation (4-15) is determined by the reaction of soft segment (i.e.,

f (oc T ) aliphatic triamine and diisocyanate), while the term (otg/OLg-CL') ’ is determined by the reaction of hard segment (i.e. aromatic diamine and diisocyanate). Using the parameters determined from the preceding section, the model prediction of the adiabatic temperature rise and viscosity rise in a three-component polyurea RIM experiment

(T5000/I1305/DWTDA=70/68/30 (wt)) is presented in Figures 4.6 and 4.7.

The prediction of temperature and viscosity rises is fairly good. The slight deviation may result from the interactions between hard and soft segments, which is not considered in this model.

4.4 SUM4ARY AND CONCLUSIONS

Experimental results of a DWTDA-based three-component polyurea RIM reaction indicate that more than 80% of the reaction takes place within 10 seconds and the system reaches solidification within 5 seconds. Hie conversion to reach solidification is close to 60% of total conversion. A mathematical model is proposed in this study to simulate the fast reaction and Theological changes in the polyurea RIM process. Experimental techniques for kinetic and Theological measurements, data analysis and parameter determination sure explained systematically. The parameters of kinetics and Theological changes are determined based on the solution polymerization data from FTIR and the 199

Haake rheometer. Combined with an appropriate heat transfer equation, this model results in very good agreement with RIM experiments. CHAPTER V

STRUCTURE-PROPERTY-PROCESSING RELATIONSHIPS OF POLYURETHANE-POLYESTER INTERPENETRATING POLYMER NETWORKS

SYNOPSIS

Interpenetrating polymer networks of polyurethane and unsaturated polyester were prepared by reaction injection molding (RIM) and transfer molding. The structure of the molded samples was analyzed by electron microscopy and dynamic mechanical analysis. It was found that polymer morphology and dynamic mechanical properties depend strongly on the molding temperature, reaction rate and reaction sequence. Simplified structure models based on Takayanagi's model and sample morphology can predict the storage modulus reasonably well but not the tan5.

5.1 PREVIOUS WORK ON MORPHOLOGY AND PHYSICAL PROPERTIES OF IPNs

5.1.1 INTERPENETRATING POLYMER NETWORKS (IPNs)

5.1.1.1 GENERAL IPNs

Polymer blends can be divided into two groups: those with chemical bonding between the polymeric components and those without this bonding. Mechanical blends are the most important class of the polymer blends without chemical bonding. Because mechanical blends are

200 201

prepared after polymerization has been completed, only thermoplastic

polymers fit into this category. As some chemical bonding is

introduced in the polymer blends, the materials are considered as

graft copolymers. An important special case of the graft copolymers

are the block copolymers, in which the individual components are

joined at their ends. Interpenetrating polymer networks (IPNs),

involving polymerization of one polymer in the immediate presence of

the other, are a special class of the polymer blends parallel to

mechanical blends, graft copolymers, and block copolymers. Since

mixing starts at the beginning of polymerization, synthesis of IPNs is

the only way to combine two crosslinked thermosetting polymers

together. Figure 5.1 shows schematically the topologies of several

polymer blends of interest (Manson and Sperling, 1976; Thomas and

Sperling, 1976).

Interpenetration of two networks represents a novel approach to

combining polymers to produce improved properties and process

characteristics that cannot be achieved by conventional mixing

techniques. As an IPN is a material composed of two polymers which are

crosslinked or synthesized in the presence of each other, it can be

classified according to the method of synthesis or the structure of

the polymeric components (Manson and Sperling, 1976; Thomas and

Sperling, 1976). Depending on the method of synthesis, IPNs are

generally classified into two categories: simultaneous IPNs (SINs) and

sequential IPNs. The first starts with a mutual solution of both monomers along with their own crosslinkers and initiators. Both POLYMER BLEND GRAFT COPOLYMER

BLOCK COPOLYMER ip n

Figure 5.1 A comparison of various polymeric composites 203 reactions, which are either step-growth or chain-growth polymerization, proceed at the same time. The second starts with the preparation of crosslinked polymer 1. Monomer 2 and its own crosslinker and initiator are swollen into polymer 1 and then polymerized in situ. Depending on the polymer structure, IPNs can also be classified into two categories: full-IPNs and semi-IPNs. A full-IPN is obtained when both polymeric components are crosslinked (i.e., thermosetting polymers). If only one of the components is crosslinked, the product is referred to as a semi-IPN. For a given system, two different semi-IPNs may be distinguished. When polymer 1, first formed component, is crosslinked and polymer 2 is linear, the material is designated as a semi-l-IPN. When, however, polymer 1 is linear and polymer 2 is crosslinked, the material is designated as semi-2-IPN. If both polymerization sequence and polymer structure are taken into account, five different IPNs can be classified. Various IPNs are shown in Figure 5.2.

Historically, the concept of IPNs can be traced back to at least as far as 1951, when a British patent was granted to Staudinger and

Hutchinson (Staudinger and Hutchinson, 1951). Their patent of 1951, granted by the United States, explicates the use of an IPN topology to prepare an improved optically smooth plastic surface. The first use of the term "interpenetrating polymer network" began with a paper of

Milieu' in 1960 on homo-IPNs made from polystyrene (Millar, 1960). A similar system was also evaluated by Shibayama et al. (1962-1966).

Early sequential IPN systems have been investigated by Sperling et al. Monoaer II Initiator I Initiator II Croaalinkar I Croasllnkar II Q z

Monoasr I (A) Initiator Polyaar I

Croaalinkar I A (B) Initiator I A II

(C) Initiator I A II Croaalinkar I Synthesis of IPNs - ► (A) sequential-IFN (D) (B) seai-IPN (sequential) (C) simultaneous IPN (SIN) (D) seal-SIN

Figure 5.2 Synthesis of IPNs 204 (1969-1976) and by Bowden et al. (Allen et al., 1973-1974; Bowden et

al., 1974). These systems, with different polymer entities, consist of

different polymer combinations generally based on chain-growth

polymerization. SINs were first reported by Frisch et al. in 1974 by

simultaneously curing the mixture of a urethane prepolymer and a low

molecular weight epoxy resin (Frisch and Mukherjee, 1974). Other SINs

were also studied by Frisch et al. (1974's) and by Touhsaent et al.

(1974). A series of semi- and full-IPNs based on polyurethane-

polyacrylate system were reported recently by Hourston et al.

(Hourston et al., 1983-1986). Several academic laboratories interested

in IPNs have conducted intensive research in this area. These

laboratories include Sperling’s group at Lehigh University, Frisch’s

group at the University of Detroit and SUNY, Hourston’s group at the

University of Lancaster, England, Cohen’s group at MIT, and many

others. A few review papers have been published by these researchers

(Manson and Sperling, 1976; Thomas and Sperling, 1976; K1 emptier, 1978;

Sperling, 1980; Sperling, 1981; Sperling 1985).

As far as polymerization conditions are concerned, solution

polymerization is the most popular method to synthesize IPNs since it

is the easiest way to control the reaction rate. The reaction usually

requires several days for solution polymerization, while only a few

seconds are needed in some bulk polymerizations such as in the RIM process. Latex polymerization has also been explored and its product

is termed interpenetrating elastomeric network (IEN) (Klempner et al.,

1969-1970). Most IPNs, synthesized through solution or latex 206 polymerizations, are developed for slow processes like casting and coating. These are not the preferred methods for reactive polymer processing operations. In recent years, there has been increasing interest in using IPNs as reaction injection molding (RIM) materials.

For the RIM process, the only practical way is through bulk polymerization without any solvents being added to the system.

The synthesis of IPN involves two independent and non-interferring reactions carried out in the same reactor under the same reaction conditions. Both sequential and simultaneous syntheses have been mentioned in the literature (Manson and Sperling, 1976; Thomas and

Sperling, 1976). A sequential IPN is made by swelling a crosslinked primary polymer with a second monomer containing its own appropriate crosslinking agent and initiator, followed by polymerization of the second monomer. This process is not suitable for mass production because it takes too long for swelling and strongly depends on the mutual solubility of the individual networks. For simultaneous IPN

(SIN), a mutual mixture of monomers with their own croBslinking agents and initiators is prepared and poured into a mold for curing.

Therefore, SINs are more suitable for applications in RIM processes.

In the near future, SIN-RIM products may become an attractive alternative to SMC, BMC, fiber-reinforced RIM, and even sheet metals for exterior automobile body parts and structural panels. 207

5.1.1.2 POLYURETHANE-BASED IPNa

Polyurethanes are widely used in polymer industries. They are, however, considered inappropriate for structural applications because of their high thermal expansion coefficient and low rigidity at high temperatures. One approach to improve the material properties is by

introducing a second reactive polymer into the polyurethane reaction to make up for the deficiencies of the existing material. Ibis approach is essentially an application of the interpenetrating polymer network (IPN). The dual reactions in cm IPN system also offer some advantages in processing. For example, the addition of a less viscous resin to the urethane material cam reduce the resin viscosity and, consequently, can facilitate the mold filling. One may also use a mixing-activated step-growth polymerization, such as polyurethane, am an internal heat source to initiate a thermally-activated chain-growth polymerization.

The urethane-based IPN systems, usually composed of a rubbery network (elastomeric urethane) and a glassy network (polyester, polystyrene, polyacrylate, epoxy, etc.), provide a wide spectrum of composite materials. Although many IPNs and SINs have been studied before, only a few of than include polyurethanes. Kim et al. (1976a,b) investigated SINs of polyurethane/polystyrene systems and polyurethane/polymethyl methacrylate systems. Djomo et al. (1983) and

Morin et al. (1983) studied polyurethane/polymethyl methacrylate SINs.

Yoon et al. (1976) reported a series of semi-SINs based on polyurethane and polyacrylates. Most of the work concentrated on the 208 product morphology and mechanical properties. Hie reaction kinetics and processing feasibility have not been studied in detail. Recently,

Hsu et al. (1985, 1987) and Yang et al. (1987) investigated the reaction kinetics of polyurethane/polyester SIN using a differential scanning calorimeter (DSC) and a Fourier transform infrared spectrometer (FTIR). They found that these two polymerizations greatly interacted with each other and the reaction profile depended strongly on the compound composition, reaction sequence, type of catalyst, and crosslinking nature of the constituent polymers. Nguyen et al. (1984-

1986), Hsu (1987), and Nelson (1987) reported their works on processing a polyurethane/polyester SIN by the RIM process. Their results showed that the processing conditions in the RIM process such as impingement pressure, stream Reynolds number, and reaction kinetics can affect the morphology profoundly, and subsequently influence the physical and mechanical properties of the SIN. In order to apply polyurethane-based IPNs in reactive polymer processing, further research in processing-strueture and processing-property relationships is needed to have a better understanding of these composite materials for actual commercial applications.

Up to the present, most IPNs have been developed for slow processes such as coating and casting. For fast processes like reaction injection molding (RIM), there are only a few commercially available IPN compounds. Ashland Chemical developed an acrylamate polymer (Wilkinson et al., 1983; Kelly, 1986) that is basically a urethane with unsaturation on the polyol chain, which forms a second 209 network with a crosslinking agent, acrylic monomer. Amoco Chemical developed a series of polyurethane-polyester hybrids which can be used in various reactive processes (Edwards, 1986).

In this study, we investigate the effect of processing conditions on the morphology and dynamic mechanical properties of a polyurethane/polyester IPN prepared by both transfer molding and reaction injection molding processes.

5.1.2. MORPHOLOGY OF IPNs

Ideally, interpenetrating polymer networks are composed of two polymeric networks with one dissolved in the other on a molecular scale. However, roost IPNs and related materials synthesized to date tend to undergo phase separations. Major studies of phases in the past have been centered around the variations in amount, size, shape, sharpness of their boundaries, and degree of continuity. These aspects together form the morphology of the material. Some aspects of morphology, such as amount, size, and shape of phase, can be observed directly by transmission electron microscopy of very thin sample sections which are stained and ultramicrotomed based on Kato’s (1967) osmium tetroxide (OsO^) technique. Unfortunately, except for polymers containing carbon-carbon double bonds, the osmium tetroxide staining method is less successful for many materials. Other aspects of morphology, such as interface characteristics and phase continuity, are best determined by combining information from electron microscopy 210

and chemical and mechanical tests (Klempner et al., 1970's; Hourston,

1983; Hayashi et al., 1987; Karger-Kocsis and Kiss, 1987). ihese tests

include infrared spectrum, glass transition temperature (T^) from differential scanning calorimetry (DSC), swelling, dynamic mechanical analysis, small angle x-ray scattering and many others.

Some factors are now clearly understood to control the IPN's morphology, which in turn influences the physical and mechanical behaviors of the material. These factors include chemical compatibility of the polymers, crosslinking densities of the networks, synthesis process and conditions, and the IPN composition. While these factors may be interrelated, they can often be varied independently.

5.1.2.1 COMPATIBILITY

The thermodynamic compatibility of the polymeric constituents in

IPNs seems to be the most important factor in determining the extent of phase separation which generally occurs in the course of polymerization. Since IPNs are made from solutions of monomers or prepolymers of swollen networks, a degree of compatibility between polymers is necessary. The domain structures of IPNs were observed to have a fine size less than 100 A for a highly compatible system and have a coarse size around 1000 A for a less compatible system

(Klempner et al., 1971; Sperling et al., 1972b,c; Kim et al.,

1976a,b). Sperling et al. (1972c) investigated systematically the compatibility dependence of morphology in sequential IPNs, 211

polyethylacrylate(PEA)/copolymer of methyl methacrylate(M M ) and

styrene(S). While leas compatible with PS, PEA is isomeric and nearly

compatible with PMiA. They found that when S-mers are replaced by M4A-

mers in the PEA-based IPN, the cellular structure becomes smaller in

size and less distinct. As S-mers were completely replaced by M4A-

mers, the cell structure almost completely crumbled and because a

fine, more interpenetrating structure. However, in spite of the fact

that dispersed domains decrease with increasing compatibility of the

systems, fully thermodynamic compatibility has not been achieved.

One way to improve the compatibility of IPNs is by introducing

grafting between polymers. Djomo et al. (1981) and Hayashi et al.

(1983) reported the formation of the graft copolymer of styrene-

divinylbenzene onto PVC in the P(S-DVB)/PVC semi-IPN system, in which

PS has a network structure by copolymerization with DVB. Their

(Hayashi et al., 1983) electron microscopy and dynamic mechanical

tests indicated that P(S-DVB)/FVC systems have a two-phase nature with

a styrene-divinylbenzene copolymer as the continuous phase (P(S-

DVB)phase) and a PS/PVC composite as the dispersed phase (PS-DVB

phase), in which PS penetrates into the PVC domain. Kircher et al.

(1984) studied the SIN of polyurethane/polymethyl methacrylate

(PU/PM4A) starting from a mixture of polyol, polyisocyanate, and vinyl monomer. An opaque two-phase polymer alloy was observed. However, grafting FMMA onto the PU through incorporation of 2- hydroxyethylacrylate leads to a transparent product. 212

5.1.2.2 CROSSLINKING

Donatelli et al. (1976). studied the morphology of IPNs and semi-

IPNb synthesized from styrene-butanediene copolymers (SBR) as polymer

1 and polystyrene (PS) as polymer 2. By controlling the level of

crosslinking, they found that the polymer synthesized first forms the

more continuous phase and tends to control the morphology. The second

polymer forms a cellular structure whose size is determined primarily

by the degree of crosslinking of polymer 1 , with an increase in

crosslinking producting a finer structure. This effect has also been

rationalized by a semiempirical thermodynamic model (Donatelli et al.,

1977).

Allen et al. (1973a,b) observed the same effect of the first

network in polyurethane(PU)/polymethyl methacrylate(PMIA) materials, made by "interstitial polymerization". For a given PU/PMiA

composition, the mean domain size of F M M was found to be independent

of factors connected with the vinyl polymerization, but very sensitive

to the tightness of the PU network determined by the reaction conditions. Although the F W A was uncrosslinked, its domain size varied from about 650 A for tight PU networks to about 1800 A for

loose networks.

Similar results were reported by Kim et al. (1976a,b) in their

investigations on polyurethane/polymethyl methacrylate IPNs. They concluded that physical interlocking prevented the demixing (phase separation) of IPN, thereby producing a better mixing of the constituent networks. In the studies of polyurethane/polyethylacrylate 213

semi-2-IPNs, Hourston et al. (1983) found that the degree of crosslinking had a significant influence on the morphology and properties by controlling the amount of enforced mixing.

Yeo et al. (1981) derived the following equations to predict the domain diameter of polymer 2 in a fully sequential IPN:

4 r (5-la) R~f~TAV~+BV~)

(5-lb)

B = (ln4»2-34>^/3+3) (5-lc)

where "sphere" domain diameter of polymer 2 r : interfacial tension V : crosslinking density 4> : volume fraction of each volume R : universal gas constant T : absolute temperature

Thus, the domain diameter of polymer 2 may be predicted solely from knowledge of the networks and their interaction. Compared to the experimental data, the predicted results showed good agreement although the spheres were a poor approximation for many sequential

IPNs

Donatelli et al. (1976) proposed a semi empirical equation for the phase domain size of semi-l-IPN. Ibis equation can be written in a simplified form: 214

2 r W , 2 D, 2 (5-2) R T VX K - ^ w - - * 273 where Dg*. domain size of polymer 2 Mgj molecular weight of polymer 2 Vj: crosslinking density of polymer 1 r : interfacial tension Wg." weight fraction of polymer 2 R : universal gas constant T : absolute temperature

This equation shows that the domain size of polymer 2 decreases with increasing crosslinking density of polymer 1 , and follows a complex behavior with changes in . The case for the full-IPN can be approximated by taking ^ = 00. This equation has been applied reasonably well to three systems: SBR/PS (Donatelli et al., 1977), castor- oil/urethane/PS IPN (Yenwo et al., 1977), and PEA/poly(S-co-M4A)

(Siegfried et al., 1978).

5.1.2.3 SYNTHETIC PROCESS AND CONDITION

For sequential IPNs, the polymer synthesized first tends to control morphology and forms the more continuous phase when the polymerizing system is not stirred (Donatelli et al., 1976). However, when the IPNs are synthesized in inverse order, the new morphology is again controlled principally by the first network (Sperling et al.,

1972c). As a result of comparing the morphologies of the normal- and inverse-polymerized composites, it can be stated that the network synthesized first forms either the continuous phase or the more

continuous phase.

In simultaneous IPNs (SINs), the networks form during the same

time period, although not necessarily at the same rate, and more complex morphologies result. It seems that better materials result if the reactions are carried out more or less sequentially after the monomers are introduced simultaneously. Therefore, polymerization rate is an important factor in determining the morphology of SINs. If polymerization rates of constituent reactions are significantly different, the one that forms a network first will become the more continuous phase and subsequently "freeze" the phase formation of the other. This phenomenon has been observed by Touhsaent et al. (1974,

1976) in the studies of epoxy/poly-n-butyl-acrylate SINs. By controlling the relative reaction rates, it was found that the dimension of the dispersed phase, and the extent of molecular mixing between the two networks, depends on the relative rates of polymerization, or the gelation time. The minimum domain size and the greatest amount of molecular mixing occur when the closest approach to simultaneity is attained.

Effect of polymerization temperature on the morphology has been observed in the studies of poly(styrene-co-divinylbenzene)/poly(vinyl chloride) (P(S-co-DVB)/PVC) systems, prepared by copolymerization of S and DVB in the presence of fine PVC powder (Hayashi et al., 1983,

1987). The changes of the phase-separated structure of P(S-co-DVB)/FVC systems polymerized at various temperatures are explained by the 216 differences, based on the temperature dependences, of the diffusion constants of S and DVB into PVC particles.

5.1.2.4 COMPOSITION

The relative amounts of the two phases present after polymerization are determined by the IPN's composition. Increasing amounts of dispersed component generally lead to increasing domain size, but the effect also depends on the synthesis method.

Occasionally, phase inversion will occur at the time when the composition of the dispersed component is increased to a certain level.

For sequential IPNs based on polyethylacrylate/polystyrene, it was found that there was only a slight increase in dispersed domain size as the composition ranged from 75/25 to 25/75 (Sperling et al.,

1972c). Hie same result was found for styrene-co- polystyrene/polystyrene IPNs over a range limited to PS-rich materials

(Donatelli et al., 1976). These phenomena are due to the fact that an upper limit on the amount of the later-formed polymer is set by the equilibrium swelling of the second monomer plus crosslinker into the earlier-formed network.

In the studies of polyurethane/polystyrene IPNs, Kim et al. (1975) found that a phase inversion occurred at a composition of about 75% polyurethane , i.e., the continuous phase being polystyrene when the polyurethane composition is less than 75%. Their electron microscopy result showed phase separation with some chain interpenetration. According to similar investigations of polyurethane/polymethyl methacrylate IPNb , Kim et al. (1976a,b) also suggested a two-phase structure with the phase inversion occurring between 60 and 80% polyurethane concentration.

The optical microscopy of polyurethane/polysiloxane, studied by

Ebdon et al. (1984), showed the result that the urethane network was continuous and the polysiloxane was present as a dispersed phase from

90 to 50% polyurethane content. The situation, however, was reversed from 40 to 10% of polyurethane. Matsuo et al. (1970) also found phase inversion in polyacrylate/polyurethane-urea IPN at a composition of

30% polyacrylate. From electron microscopy, Hourston et al. (1986) observed that the IPN of polyurethane/polyvinylacrylate=20/80 is likely to form a structure of two-continuous phase.

Jordhamo et al. (1986) proposed a semiempirical expression to predict the phase continuity and inversion in polymer blends and simultaneous IPNs. Based on the volume fraction 4> and viscosity n » a

Theological model led to the equation:

as the criteria for a dual phase continuity of phases 1 and 2. This relation was evaluated for two systems: a castor oil polyester- urethane/polystyrene SIN, and a mechanical blend of polystyrene and polybutadiene. This prediction did not show good agreement with the 218 experimental data. However, the experimental technique designed to measure the volume fraction of phases by centrifugation was reliable.

5.1.3 GLASS TRANSITIONS AND MECHANICAL PROPERTIES OF IPNs

Polymers are widely used and rapidly growing because they have desirable mechanical properties at an economical cost. Their versatile mechanical properties cover the range from soft elastomers to rigid materials. For this reason, the mechanical properties may be considered the most important of all the physical and chemical properties of high polymers for most applications. The mechanical response (modulus) of a polymer depends upon molecular structure, morphology, composition, temperature, and time scale of the experiment. These experimental tests include impact, creep, fracture, toughness, stress relaxation, stress-strain, dynamic mechanical tests and many others.

Most polymers are either completely amorphous or have an amorphous-like component even if they are crystalline. While glasses below the glass transition temperature (T^) are hard and rigid, amorphous polymers above that temperature are soft and flexible like either elastomers or very viscous liquids. Their mechanical properties show profound changes in the region of the glass transition.

Accordingly, can be considered the most important material characteristic of a polymer as far as mechanical properties are concerned. One principal glass transition is usually exhibited in 219

simple homopolymers and random copolymers. When two incompatible

polymers are blended, the individual phase domains retain the glass

transitions of their respective parent homopolymers. As a result, most

blends, grafts, blocks and IPNs display two principal glass

transitions. If significant molecular mixing takes place, the

transitions will be broadened and/or their temperatures will be

shifted closer together. There are many methods by which the of a

polymer may be determined such as differential scanning calorimetry

(DSC), dynamic mechanical analysis (DMA), dilatometry and many others.

The interrelationships between morphology (from electron microscopy) and mechanical modulus (from mechanical tests) are always clearly visualized. Both experiments provide information about phase domain formation and the extent of molecular mixing, but in different ways. The morphology reveals the size and shape of phase domains, which cannot be deduced from the appearance of the glasa-rubber transition. Although some qualitative information based on electron microscopy shows the extent of molecular mixing, the glass-rubber transition behavior tells much more. Mechanical properties are controlled by morphology and knowledge of mechanical behavior that is easily detectable in mechanical tests, complements the information about mixing that is obtained by electron microscopy.

Dynamic mechanical analysis (DMA), in general, gives more information about a material than other tests can. DMA data taken over a wide temperature and frequency range are especially sensitive to the 220 physical and chemical structure of polymers. In many cases, DMA is the most sensitive test known for studying glass and secondary transitions in polymers as well as the morphology of crystalline polymers (Nielsen

1974).

In DMA experiments, a sinusoidal or other periodic stress is applied to a specimen and the responses are measured. Since the stress and Btrain are generally not in phase, two quantities can be determined: the storage modulus, G ’, a measure of the energy stored elastically, and the loss modulus, G", a measure of the energy dissipated. Exactly analogous to the notation of Young’s modulus, E, the complex shear modulus, G, composed of G ’ and G" is expressed through the relation

G = G ’ + iG" (5-4) where i=4=T. The angle which reflects the time lag between the applied stress and the strain is defined by

tan* = G"/G’ (5-5) which is a damping term and is a measure of the ratio of energy dissipated as heat to the maximum energy stored in the material during one cycle of oscillation. With DMA spectroscopy, the frequency may be varied at constant temperature, or, often more conveniently, the temperature may be varied at constant frequency. Results of the two 221 experiments may be interrelated by means of the well-established time- temperature superposition principle formulated by Williams, Landel, and Ferry (Williams, 1955). The (or T^’s) of a polymer is defined as the temperature corresponding to the maximum G" or tani at the main relaxation, which marks the onset of main chain segmental mobility corresponding to the glass-rubber transition (often designated as the oc relaxation in amorphous polymers). Since the tan5 is dependent on the frequency tested, its peak at a frequency of one cycle per second generally is at a temperature 5 to 10 °C above the the glass transition temperature as measured by dilatometry or differential thermal analysis (DTA). The maximum in the loss modulus, G", (also function of frequency) at low frequencies is very close to T .

Some theoretical models have been proposed to relate the mechanical moduli of a polymer blend to its composition if the mechanical characteristics of the components are known. Most assume that the adhesion between the phases is perfect and that the dispersed phase is uniformly distributed within the matrix.

Kemer (1956) and Hashin et al. (1963) proposed that the upper and lower bounds of the modulus of a composite could be determined from the volume fraction and moduli of its components. Similar to the simplified equations derived by Kemer and Hashin et al., more elementary bounds for any geometry of the dispersed phase were proposed as follows (Broutman and Krock, 1967; Lange, 1974):

upper bound: G = G^ + Gg (5-6) 222

lower bound: G = GA Gg / (VAGA+VgGfi) (5-7) where V represents the volume fraction and G represents the modulus.

These two equations are similar to Takayanagi’s models 1 and 2 described in the following paragraph.

Takayanagi et al. (1964) presented several mathematic models to better explain the mechanical behavior of polymer blends. By transforming the spring and dashpot relaxation models to plastic and rubber elements, four simple combinations of the Takayanayi’s models are proposed for two-phase polymer systems and are shown in Figure

5.3. The plastic phase is denoted by A and the rubber phase by B, while the quantities X and 4> ) are functions of the volume fractions of parallel and series elements, respectively. The mathematic expressions of their corresponding models are represented by the following equations:

Model 1 : G = (1-X) GA + X Gg (5-8)

Model 2 : ~ (5-9) Ga " + gb

Model 3 : G = (1-X) G. + X (- (5-10) A

Model 4 : ■n-xT_G~"+"x"Gg (5-11)

Combinations shown in models 1 and 2 form the basic parallel and 223

1-X

(1) (2)

T 1 -

1 T 1 ~ X ♦ T

( 3 ) ( 4 )

Figure 5.3 Four Takayanagi’b models for two-phase polymer system; A: plastic, B: rubber 224 series models, the equivalence of model 3 and model 4 has been proven earlier by Kaplan and Tschoegl (1974). Takayanagi’s models apply only to composites which have only one continuous phase. However, these models can be modified to make them applicable to oomposites with two continuous phases (Nielsen, 1977). Analogous to Takayanagi’s two-phase models, a three-phase model was presented by Rosovizky et al. (1979).

Based on their observations, the existence of an interfacial layer was introduced to interpret the loss of the distinct two-phase character of pxjlyurethane/pxjlyurethane-diacrylate IPN. Figure 5.4 shows this three-parameter model of a composite and its ma thematic expression is given by:

G = (1-X) Ga + X (--=-- + (5-12)

where X=l-VA+(Vc/2), 4>=VB-(Vc/2) and VC=1-VA -VB . The third phase

(phase C) was assured to have the character of a homogeneous mixture with the composition VA :Vg= 1:1 (Gc=GA/2+GB/2) and to have the values of in the range of 0VB ).

Finally, Davies (1971) has also developed a theory to predict the modulus behavior for composite systems containing two continuous phases. In his derivation, he assured that two composites with moduli

G+AG and G-AG respectively have been mixed to obtain a new composite with modulus, G. His equation can be expressed in the following manner: 225

1-0

Figure 5.4 Three-phase aodel of a composite 226

(5-13)

A general mixing equation which often successfully predicts certain properties of composites with two continuous phases has been presented by Nielsen (1977):

, -1 < n < 1 (5-14) where n is a function of the morphology of the system and possibly the property being measured. When n=l, the ordinary rule of mixtures results. When n=-l, the inverse rule of mixtures results. The Davies' principle results when n=l/5. The logarithmic rule of mixtures is obtained when n=0 (Nielsen, 1967, 1974). Ibis is presented below:

logG = logGA + Vg logGg (5-15)

Sometimes, the elastic modulus of block copolymers and polyblends follows, quite accurately, the logarithmic rule of mixtures, which is based on partial continuity of both phases (Nielsen, 1974).

Combining the DMA results with the phase structure revealed by transmission electron microscopy (TEM), Sperling et al. (1972c) obtained complementary information on the two-phase poly(ethylacrylate)/poly(styrene-co-methyl methacrylate) (PEA/P(S-co-

M4A)) IFNs. Hie TEM results showed that PEA/PS materials exhibited a cellular structure of about 1000 A diameter, together with a fine 227 structure. When S-mers were replaced by WMA-mers, a greater compatibility was indicated by decreases in cellular dimensions and a corresponding rise to prominence of the fine structure. It was suggested that phase separation took place later in the second component polymerization. The DMA results showed a slight inward shift of the transitions for the PEA/PS material, and a major buildup of multiple transitions accompanied by high damping in the region between the two G" peaks as S-mers were replaced by MiA-mers. This result confirms in a semiquantitative manner the conclusion that extensive but incomplete molecular mixing takes place in MMA-rich materials.

With the aid of DMA spectroscopy, Matsuo et al. (1970) clearly observed the effect of composition on the moduli, G* and G", for polyacrylate/polyurethane-urea IPNs. Two distinct transitions were found, with the polyacrylate softening at the higher temperature. The slight shifting and broadening of the moduli indicate a modest degree of molecular mixing. However, the large shifts in G ’ indicate changes in phase continuity. Thus, Takayanagi’s parallel and series models

(models 1 and 2) were employed to judge the relative continuity of the two phases. The parallel model for the mechanical behavior of a two- component polymer system corresponds to the case in which the stiffer component is continuous, while the series model corresponds to the case in which the softer component is continuous. A plot of storage modulus against composition at 25 °C, selected because this temperature is nearly halfway between the two transitions, revealed that a phase inversion takes place at about 30% polyacrylate content. This 228 confirmed the observations based on electron microscopy. Takayanagi’s two-phase model (model 3) was also used by Matsuo et al. to study the mechanical properties of IPNs and they found that the model fitted the

G' reasonably well with appropriately adjusted parameters X. and ♦ .

Rosovizky et al. (1979) applied the same model to predict the the dynamic Young's modulus of polyurethane/polyurethane-diacrylate IPNs and found a pronounced two-phase behavior in the concentration range of polyurethane-diacrylate about ^ 50%, confirming the heterogeneous character of the IPN structure.

Allen et al. (1974a,b,c) tested the DavieB* equation along with

Kemer's, and Haskin-Shtrikman's equations and others in the studies of polyurethane/polymethyl methacrylate (PU/FM4A) IPNs. The success of the Davies’ equation supports relaxation studies and suggests interactions between PU and PMiA in the interdomain regions.

5.2 EXPERIMENTAL

5.2.1 MATERIALS

5.2.1.1 MATERIALS USED FOR TRANSFER MOLDING

Hie ingredients of the polyurethane/polyester (PU/PES) IPN used in this study are listed in Table 5.1 and their reaction mechanism is schematically described in Figure 5.5. Hie recipe can be divided into 229

Table 5.1 Materials Used For Transfer Molding

Ingredients Percentage

Polyurethane (50 wt%) in IPN Part (wt%) in Polyurethane

MDI (I143L, Dow Chemical) 41

Polyol (T0NE0240, Union Carbide) 48

Butanediol (BDO, Aldrich) 11

Catalyst (T-12, M&T Chemical) 0.033

Polyester (50 wt%) in IPN Part (wt%) in Polyester

65% unsaturated polyester 67 in styrene (P325, OCF)

Styrene 33

Initiator (PDO, Lucidol) 1.38 STYRENE UNSATURATED POLYESTER HO

OH

O-C OH 6 h

\ \ \ O -C -N -*

"h«rd MtPMt" "»oft

polyurethane

Figure 5.5 Schematic diagram showing the reaction mechanism of FU/PES IPN 230 231

two parts, namely, a polyurethane and a polyeater. The polyurethane

chosen for this Btudy consists of a soft segment based on a poly(£- caprolactone diol) (T0NE0240, Union Carbide) and a hard segment based

on a liquid form of 4,4’-diphenylmethane diisocyanate (MDI) (I143L,

Dow Chemical Company) chain extended with 1,4-butanediol (BDO, Aldrich

Chemical Company). MDI was degassed and demoistured at room temperature for 20 minutes to remove water and air. The treated MDI solution was then filtered under vacuum. 70NE0240 is a long chain diol with a number average molecular weight of 2000 and is a solid at room temperature. A heating plate was used to melt this material. BDO is a low molecular weight diol with a viscosity slightly higher than that of water. The combination of the molten T0NE0240 and BDO was degassed for 40 minutes at 60°C using a heating plate and vacuum to remove water and air. The 100180: ratio of T0NE0240/I143L/BDO was set at 1/6/5 which is a typical recipe for RIM elastomers. Hie catalyst, dibutyltin dilaurate (T-12, M&.T Chemical) was used as received. The amount of T-

12 was 0.033% by volume of resin for a reasonable reaction time so that sample preparation by transfer molding was possible.

For the polyester part, styrene was used as a crosslinking agent for the unsaturated polyester resin (P325, OCF) which is a 1:1 propylene-maleate polyester combined with 35% by weight of styrene.

Styrene was not freed of inhibitor in all cases. Initiator PDO

(Lucidol) was used as received. PDO, t-butyl peroxy-2-ethyl hexanoate, is a diluted high temperature initiator. Hie amount of PDO used was

1.38% by volume of polyester resin. In the study of the effect of reaction sequence, initiator PDO was replaced by a combination of methyl ethyl ketone peroxide (MEKP), a tertiary amine and cobalt naphthenate. Hie combined reagent is a low temperature, reduction- oxidation initiator for unsaturated polyester resins. It was used as received without any further treatment. With PDO as the initiator of the polyester phase, the polyurethane reaction occurred much earlier than the polyester reaction, while with MEKP/amine/Co as the initiator of the polyester phase, both reactions occurred at the same time except that the polyurethane reaction had a higher initial reaction rate (Hsu and Lee, 1985; Yang and Lee, 1987). Hie MEKP/amine/Co ratio was set at 3/1/1. Two concentrations were tried: one sample had 1.1% of the combined initiator based on the total weight of the unsaturated polyester resin. Hie other sample had 3.3% by weight of the initiator.

The molar ratio of styrene to the double bonds of the unsaturated polyester was adjusted to 2:1. The ratio of polyurethane to polyester was fixed at 50/50 by weight for most IPNs prepared. In order to study the effect of compound composition on the morphology and mechanical properties of IPN, the ratio of polyurethane/polyester was also varied as 100/0, 75/25, 50/50, 25/75, and 0/100.

5.2.1.2 MATERIALS USED FOR REACTION INJECTION MOLDING (RIM)

For the reaction injection molding process, the amount of catalyst

T-12 used for polyurethane was 0.1% by volume of polyurethane resin and the amount of PDO for polyester resin remained at 1.38% by volume of polyester resin. Hie molar ratio of T0NE0240/I143L/BDO was set at 233

1/6/5. The polyester used was Q6585 from Ashland Chemical Company, which iB similar to P325 of OCF. The ratios of polyurethane/polyester chosen were 50/50.

5.2.2 SAMPLE PREPARATION

5.2.2.1 TRANSFER MOLDING

A laboratory scale transfer molding device with a single cavity was used to prepare samples. Figure 5.6 shows the schematic of such a transfer mold (Hsu, 1987). The mold has a single cavity. The sprue plate is 0.635 cm thick, and has four conical sprues with an entrance diameter of 0.635 cm and an exit diameter of 0.127 cm. The spacer just below the sprue plate is 0.3175 cm thick and has a rectangular cavity of 10.16 x 15.24 cm. The plunger diameter is 3.81 cm. IPN components were first mixed in a suction flask by a magnetic stirrer until no bubble was observed. This bubble-free mixture was then transported to the mold cavity through the transfer pot. Once it was in the mold, polymerization was allowed to proceed for two hours before demolding. Two molding temperatures were used: 8 0 °C and 120 °C.

Half of the 80 °C-molded samples were further postcured at 120 °C for six hours. The concentrations of catalyst and initiator were chosen in such a way that the mixture would not reach gelation in at least two minutes which was required for material preparation (i.e., mixing and transfer molding). The transfer molding technology was found to be I

PLUNGER SPRUE PLATE

POLYMER

SPRUES

MOLD CAVITY MOLD

(a) BEFORE (b) AFTER

TRANSFER MOLD

Figure 5.6 Schematic diagram of the transfer nold; (A) before

molding, (B) after molding (Hsu, 1987) 234 235

successful in the preparation of bubble-free samples.

5.2.2.2 RIM

A laboratory scale RIM machine was used to prepare samples. Figure

3.1 shows the machine design (Nelson, 1987). More detailed description

of this machine can be found in Section 4.2.2.1.

IPN samples from RIM were prepared using a rectangular mold of 15 x 3 x 0.3175 cm in size. The mold temperature was controlled at 120®C.

Hie samples were demolded after 2 hours in the mold.

5.2.3 METHODS OF ANALYSIS

5.2.3.1 TRANSMISSION FIMTTRON MICROSCOPY (TEM)

The morphology of IPN samples from transfer molding and RIM was observed using a Philips &1-300 transmission electron microscope

(TEM). The materials were prepared for the microscope by a LKB HT-2B ultramicro tone and stained with 1% osmiun tetroxide solution. A low temperature ultramicrotone unit (Ultracut E, Reichert-Jung) was used for soft samples like polyurethane and FU/PES=75/25 IPNs. Samples were examined at a power of 60 KV at various magnifications.

5.2.3.2 DYNAMIC MECHANICAL ANALYSIS (DMA)

The dynamic mechanical data of IPNs and their constituent polymers were obtained using a Weissenberg Rheogoniometer (Model R18). This system is shown schematically in Figure 5.7. Details of the stre ss transducer m eters strain heating media

D A SH -8 temperature monitor

IBM PC/XT cooling media

m o to r/ drive gearbox unit Tono

position control

Figure 5.7 Schematic diagram of dynamic mechanical analysis system 236 experimental set-up and procedure can be found in the literature

(Vang, 1985; Sangamo Controls Ltd.)* Tbe measurements were performed on solid IPN samples 0.09 x 0.30 x 2.5 cm in size. Three torsion bars with specification of 27 100°C), 98 (10 - 100°C), and 430 (-60 -

2 10 °C) dyne/cm /micron were used (depending on the glass transitions of the samples) for temperatures ranging from -60 to 130°C at a frequency of 1.2 Hz. The temperature was controlled using an aqueous ethylene glycol solution and liquid nitrogen as heating/cooling media and was monitored by an Onega 871 digital thermometer. An IBM personal computer was linked to the Weissenberg Rheogoniometer for real time data acquisition (Dash-8 , Metrabyte) with output voltage ranging from

-5 to 5 V. The temperature dependences of the shear storage modulus,

G*, shear lose modulus, G", and loss tangent, tanj, were measured and calculated.

5.3 RESULTS

To differentiate polyurethane from polyester in transmission electron microscopy, a staining technique using osmium tetroxide was applied. The osmium tetroxide technique was originally developed for rubberized polymers in which the C=C bonds of the rubber phase are stained. This technique gives excellent contrast for many multiphase polymers in electron microscopy (Kato, 1967). Generally speaking, it is possible to observe the microtexture of a two-phase polymer system under TEM if one component contains C=C, -OH, -NHg, or C-C-C- groups 238 which can react with osmium tetroxide. Others such as C=N and Btyrene can only be slightly stained (Ninomi et al., 1975). In FU/FES IPNs, previous research work has shown that polyurethane can be stained by

OsO^ (Matsuo et al., 1969; Ninomi et al., 1975; Kim et al., 1976b;

Kircher, 1979; Kircher et al., 1984), although the actual staining mechanism is still unknown. For a linear segmented polyurethane such as RIM elastomer, the hard segments which exhibit crystallinity, cannot be stained by OsO^ (Kim et al., 1976a,b). This is probably due to the rigid crystalline structure of the hard segments which prohibit any free diffusion of OsO^ vapor.

For polyesters, since most C=C bonds have been transformed into C-

C bonds, this phase cannot be greatly stained by OsO^. The -OH and

-000H groups at the end of the polyester molecules may be stained by

ObO^, but the amount of these functional groups is relatively small compared to the -OH groups in the polyurethane phase (less than 4% by mole) and they can also react with diisocyanates. Therefore, in analyzing the TEM pictures of FU/FES IPNs, the dark areas (OsO^ stained) represent the polyurethane phase and the bright areas represent the polyester phase, in general.

5.3.1 EFFECT OF VARIOUS PROCESSES

Shown in Figure 5.8 are micrographs of 50/50 IPNs processed by RIM and transfer molding. The transfer molding process represents a slow 239

process (>2 min. gel time) (Hsu and Lee, 1985), whereas the RIM

process represents a fast process («15 sec. gel time) (Yang and Lee,

1987). In Figure 5.8, the RIM-processed IPN has a more homogeneous

morphology, indicating that phase interpenetration is better achieved

in the RIM process. On the other hand, the transfer-molded IPN shows

phase separation. The continuous phase is relatively bright for the

120 °C-molded sample, indicating a high level of polyester. Both polyurethane and polyester are able to form dispersed droplets with diameters ranging from 1.0 to 3.0 Um for polyurethane and 0.1 to 0.6

Um for polyester. Within the polyurethane droplets, one can clearly see some dispersed polyester spheres.

Shown in Figure 5.9 are results of dynamic mechanical analysis

(DMA) of the 50/50 IPNs processed by transfer molding and RIM. Also included are the DMA spoctra of neat polyurethane and polyester processed by transfer molding. Plots of storage modulus vs. temporature in Figure 5.9 suggest that the RIM-processed IPN has better phase mixing than the transfer-molded IPN. Unlike the transfer- molded IPN which shows a sharp drop of G ’ at about 60 °C, the RIM-mixed

IPN shows a gradual transition of the storage modulus in the temperature range tested. The tanS vs. temperature plots shown in

Figure 5.9 indicate that there are two T^’s for the segmented polyurethane phase, one at 10°C (the soft segment) and the other at

67 °C (the hard segment). The pxjlyester phase has a single T^ at 9 5 °C.

For IPN processed by transfer molding, there are three T^'s located at 120 °C - MOLDED PU/PES=50/50

r n m m m

1 jum

(a ) RIM (b) TRANSFER MOLDING

Figure 5.8 Transmission electron micrographs of 120°C-molded IPNs (PU/PES=50/50) by (a) RIM, and (b) transfer molding 240 G’ vs. TEMP O TAN <5 vs. TEMP in

^ P E S 00 00 o PU

Z < o I-

MOLD TEMP K) (0 1 2 0 C PU o o PU/PES-50/50 PES RIM TM o o

-50.0 10.0 70.0 1 30. -50.0 10.0 70.0 1 3 0 . 0 TEMP (C ) TEMP (C )

Figure 5.9 G ’ and tanS vs. temperature for IPNs by RIM and transfer molding 242

10, 55, and 93 °C, respectively. These T^’s correspond to those of the constituent polymers. The RIM-molded IPN shows much less distinguishable T^’s. Instead, a broad tanS peak exists from 4 0 °C to

80 °C, which again implies a strong phase mixing between the two constituent polymers. These results suggest that the morphology and mechanical properties of IPN are greatly affected by the processing method.

5.3.2 EFFECT OF MOLDING TEMPERATURE AND IPN COMPOSITION

Shown in Figure 5.10 are micrographs of 8 0 °C-transfer molded IPN samples with various compositions. Hie micrograph of the 75/25 IPN shows a polyurethane-dominated matrix and a dispersed polyester phase with an average domain size around 2.0 Um. Increasing the polyester content to 50%, the 50/50 IPN shows a well dispiersed polyester phase with piarticle sizes ranging from 1.0 to 2.0 Um. As polyester content is further increased to 75%, the polyurethane phase tends to become dispersed with piarticle sizes of the order of 1.0 to several Um. The matrix is a mixture of polyurethane and polyester.

The temperature dependencies of G ’ and tani of three 80 °C-transfer molded IPNs (PU/PES=75/25, 50/50, and 25/75) are shown in Figure 5.11.

Plots of storage modulus vs. temperature show that the IPNs with more polyester content have a higher G ’ than those with more polyurethane.

The tanS curve of the 75/25 IPN shows a T at 10 °C and a T at about S g

53 °C; both are characteristic of segmented polyurethane. The T^ of the 8 0 °C - TRANSFER MOLDED P U /P E S

1 /um

(a) 75/25 (b) 50/50 (c) 25/75

Figure 5,10 Transmission electron micrographs of 80°C-transfer molded PU/PES IPNs with compositions of (a) 75/25, (b) 50/50, and (c) 25/75 tc -c> G* vs. TEMP o TAN <5 vs. TEMP n in CM O

CO 00

o o o in CM Z < o h*

MOLD TEMP K) (0 8 0 C o O PU/PES o 7 5 / 2 5 5 0 / 5 0 2 5 / 7 5 o O o O -50.0 10.0 70.0 1 30.0 -50.0 10.0 70.0 1 30.0 TEMP (C) TEMP (C)

Figure 5.11 O ’ and tanS vs. temperature for 80°C-transfer molded IPNs with various compositions 244 245 polyester phase in the 75/25 IPN could not be obtained since the sample was too soft to conduct DMA measurementt when the testing temperature exceeded 6 0 °C. The 50/50 IPN has T^’s at about 5 °C, 3 0 °C

(soft and hard domains of polyurethane), and 8 0 °C (polyester). The

25/75 IPN has two T^’s at - 6 °C and 2 5 °C (soft and hard domains), and a sharp T at 8 3 °C (polyester). In general, as polyurethane content in

IPN increases, the T ^ ’s of both soft and hard domains in polyurethane increase and approach the T 's of the pure polyurethane, while the T o o of the polyester phase decreases, and deviates more from the T^ of the pure polyester.

Electron micrographs of 120 °C-transfer molded IPN samples with three compositions are presented in Figure 5.12. The 75/25 IPN has a dark continuous phase with dispersed particles of about 0.5Um diameter. One can see that the dispersed phase consists of both polyurethane and polyester droplets. The composition of the continuous phase is not well defined. When the polyester content is increased to

50%, several large polyurethane droplets are formed. These particles have sizes ranging from 1.0 to 3.0 Um. One can also observe a lot of small polyester particles (0.5Um or less) distributed in the very large polyurethane particles. The 25/75 IPN has a polyester-dominated matrix. The dispersed phase consists of both polyurethane and polyester of about 0.5 Um diameter. The temperature dependence of G ’ and tanS of the three 120 °C-transfer molded IPNs are shown in Figure 1 2 0 °C - TRANSFER MOLDED P U /P E S

(a) 75/25 (b) 50/50 (c) 25/75

Figure 5.12 Transmission electron micrographs of 120 “C-transfer molded FU/PES IPNs with compositions of (a) 75/25, (b) 50/50, and (c) 25/75 246 247

5.13. Hie 25/75 IPN has a G ’ curve very close to that of neat polyester. The tani curves again reveal three characteristic peaks for each sample. Compared with the tani curves of the 8 0 °C-molded samples, one can clearly see that the polyester transition peak is much larger and tends to shift to higher temperature when samples are molded at higher temperatures. This is probably because more polyester is in the matrix when the molding temperature is increased.

5.3.3 EFFECT OF POSTCURE

Figure 5.14 shows the effect of 6 hour postcuring on the DMA spectra of an 80 °C-transfer molded IPN sample (PU/PES=50/50). The storage modulus, G ’, increases slightly and reveals a sharp transition around 60°C. The tanS vs. temperature plots show a significant difference before and after postcure. After a 6 hour postcure, the polyester transition peak is greatly depressed, while one of the polyurethane transition peaks (the hard domain) is greatly enhanced.

The other polyurethane transition peak (the soft domain) is also depressed and shifts to a lower temperature. A DSC measurement of the crystallinity of the polyurethane phase reported in a previous work

(Hsu and Lee, 1985) indicated that postcuring tends to change the crystalline structure of the hard domain, making it more stable at higher temperatures. This may be the reason why the tanS curve shows such a strong dependence on the postcure. G* vs. TEMP O TAN <5 vs. TEMP pi in M O

00 CO

o

in N Z < O H

MOLD TEMP K) 10 1 2 0 C O O PU/PES o 7 3 / 2 5 3 0 / 3 0 2 3 / 7 5 o O o -50.0 10.0 70.0 1 30.0 — 50.0 10.0 70.0 1 30.0 TEMP (C) TEMP (C)

Figure 5.13 G ’ and tanS vs. temperature for 120°C-transfer molded IPNs with various compositions 248 G \ G" v s. TEMP O TAN 6 v s. TEMP N in

MOLD TEMP 80 C 00 PU/PES-50/30 00 N r - E o O /K N2 ® o C r - in >s M V Z v/ o o» < - O I- b ” o ® BEFORE 6 —HR POSTCURE "

AFTER 6-HR POSTCURE

o o

—50.0 10.0 70.0 1 30.0 -50.0 10.0 70.0 1 30.0 TEMP (C) TEMP (C)

Figure 5.14 O ’ and tan5 vs. temperature for 80®C-transfer molded IPNs before and after 6-hr postcure 249 250

5.3.4 EFFECT OF INITIATOR

Shown in Figure 5.15 are micrographs of 8 0 °C-transfer molded IPN samples initiated by PDO and MEKP/amine/Co. The morphology of these samples initiated by MEKP/amine/Co is similar to that initiated by PDO except that the size of the dispersed polyester droplets becomes smaller at a higher initiator concentration. There are also some small dispersed polyurethane droplets (0.2 - 0.5 Um diameter) formed in the sample with the higher initiator concentration. Figure 5.16 shows the corresponding DMA results of these samples. The plots of G ’ vs. temperature show that MEKP/amine/Co initiated IPNs have a higher storage modulus, which implies that more polyester is included in the matrix. The tan5 vs. temperature plots indicate that the type and the concentration of the initiator have a profound effect on the polymer transition peaks. Both peak height and peak position change when the initiator changes. We are not able to explain these differences in detail. One observation which seems to be clear is that at a higher concentration of MEKP/amine/Co, the peaks become smaller and broader, which implies a better phase mixing. This may be the result of closer reaction rates of polyurethane and polyester phases when

MEKP/amine/Co, instead of PDO, was used as the initiator for the polyester reaction (i.e., simultaneous reactions instead of sequential reactions (Hsu and Lee, 1985; Yang and Lee, 1987). 8 0 °C - TRANSFER MOLDED P U / P E S = 5 0 /5 0

1 / u m

(o) PDO (b) MEKP/AMINE/CO —8 (c) MEKP/AMINE/CO-8 1.38% 0.67%/0.22%/0.22% 2.0%/0.67%/0.67%

Figure 5.15 Transmission electron micrographs of 8 0 °C-transfer molded IPNs (PU/PES=50/50) initiated by PDO at (a) 1.38% and by MEKP/amine /Co- 8 at (b) 0.67%/0.22%/0.22%, and (c) 2.0X/0.67%/0.67% G* vs. TEMP TAN <5 vs. TEMP

c m o MOLD TEMP 8 0 C

CO PU/PES-50/50 00

o

in CM

o

fO (0 PDO : 1.38% o o MEKP/AMINE/CO —8 : 0.6755/0.2255/0.22% 2.0%/0.67%/0.67% o o

50.0 10.0 70.0 130.0 — 50.0 10.0 70.0 130. TEMP (C) TEMP (C)

Figure 5.16 G* and tani vs. temperature for 80°C-transfer molded IPNs initiated by various initiators 253

5.4 DISCUSSION AND MODELING

5.4.1 INTERPENETRATION AND CONTINUITY OF PHASES

The studies of electron transmission micrographs can reveal

morphological information such as the size and shape of phase domains,

but cannot prove interpenetration and continuity of phases. The DMA

spectrum provides a sensitive indicator of the extent of molecular

mixing. Inward shifts of the transitions, broadening and merging of

the transitions, and increased intensity of the two transitions can

qualitatively tell how much mixing takes place. Further studies of the

relationships between storage moduli and composition can yield more

insight into the relative interpenetration and continuity of two

phases.

TEM micrographs of transfer-molded IPN samples reveal a

complicated two phase structure. The matrix can be polyurethane,

polyester or a mixture of the two constituent polymers depending on

the sample composition and processing conditions. If the glassy

component, polyester, is the continuous phase, a plot of storage

modulus vs. sample composition at a fixed temperature would show a

positive deviation from the logarithmic additivity rule, based on

Takayanagi’8 parallel model (Takayanagi et al., 1964). i.e.,

G* = (1-VB ) G» + VB G* (5-16)

where V is the volume fraction of component B and G*=G,+iG". 254

If the elastomeric component, polyurethane, is the continuous phase, a similar plot would show a negative deviation from the logarithmic additivity rule, according to Takayanagi’s series model, i.e.,

(5-17)

In case both components form a two-continuous phase, Davies’ model

(Davies, 1971) can be used to calculate the storage modulus as a function of sample composition, i.e.,

G (5-18)

Figure 5.17 shows such plots of IPNs prepared in this study. For

120°C-transfer molded samples, Figure 5.17a indicates that polyester tends to form the matrix for samples with a high concentration of polyester. A phase inversion seems to take place near 70% polyester, where the matrix changes from a polyester predominated phase to a two- continuous phase. For 8 0 °C-transfer molded samples, the results are quite different. Figure 5.17b implies that the matrix has a two- continuous phase for samples with a high concentration of polyester. A phase inversion takes place around 40% polyester, where the matrix changes from a two-continuous phase to a polyurethane predominated phase. Apparently, reaction temperature plays cun important role in determining the final morphology of the molded IPNs. At low molding O o O ( d y n e / c m o o Figure 5.17 G* vs. composition plots at 20 at plots ®C, 120°C-transfer (a) composition vs. G* 5.17 Figure 20. 40. 60. SO. 20. 40. 0 4 . 0 2 . O . O S . 0 6 . 0 4 . 0 2 . D P M E T D L O M C 0 2 1 (O)

odn, b 80°C-transfer molding 0 (b) 8 molding, TA KAYANAGI'S KAYANAGI'S TA E L U P ADDITIVITY AIS MODEL E D O M OAVIES* XP MENTAL A T N E IM R PE EX G* P M E T T S E T S L E D O M DATA C 0 2 ) % ( R E T S E Y L O P s COMPOSITION vs.

P M E T D L O M n S E P In U P U P S E P C 0 8

OO. O lO . 0 8 . 0 6 n U P In S E P S E P U P

255 256

temperatures, polyurethane reacts much faster (Hsu and Lee, 1985; Yang

and Lee, 1987) and tends to form the continuous phase. Even at a high

concentration of polyester, polyester still cannot form the

continuous phase, instead, a two-continuous phase is formed. At high

molding temperatures, both reactions have similar reaction rates (Hsu

and Lee, 1985; Yang and Lee, 1987). The crosslinking nature of the

polyester reaction and its early onset of gelation result in a two-

continuous matrix for IPNs with a low polyester content, and a

polyester dominated matrix for IPNs with a high polyester content. TEM

micrographs agree with the above discussion in general although the

two-continuous matrix structure and the phase inversion cannot be

easily identified in the micrographs.

5.4.2 MODELING OF MODULUS BEHAVIOR

Takayanagi’s models have been used widely to explain the dynamic

mechanical properties of multiphase polymers. The models assume that

the dispersed phase is uniformly distributed within the continuous

phase and the adhesion between the two phases is perfect. It is not

surprising that difficulties are involved in adequately applying these

simple models to describe the mechanical behavior of polymer

composites with complex morphology, such as IPNs.

In this study, Takayanagi’s models are modified to make them more

easily applicable to IPNs prepared by transfer molding. The derivation of one mechanical model is given in Appendix C. Other models follow

the same approach. Only four models will be discussed here. According to Figures 5.12c and 5.17a, it may be hypothesized that the 120°C- transfer molded FU/FES=25/75 IPN has a polyester-dominated matrix

(phase C), in which small droplets of both polyester (phase A) and polyurethane (phase B) are dispersed. A simplified structure of this sample and a simplified mechanical model are shown in Figure 5.18, where phase C has the characteristics of a mixture of A and B based on

Davies’ model. i.e.,

i/5 1/5 1/5 VC GC = VAC °A + VBC °B (5-19a) and

where is the volume fraction of phase i, i=A, B, C, V >c is the volume fraction of component j in phase C, j=A or B, G. is the complex 3 modulus of phase i. G ’ and G” of this structure can be calculated from the following equations:

(5-19b)

(5-19c) where G\ G” vs. TEMP o TAN 5 vs, TEMP in (N|

MOLD TEMP 1 2 0 C G* CO PU/PES-25/75 00 EXPERIMENTAL E DATA o — MODEL • o PREDICATION c »- *o in >. n v Z r < o 01 H - O - 0 .8 2 b *" K) (0 o / o ®

O O

-50.0 1 0.0 70.0 1 30.0 -50.0 1 0.0 70.0 1 30.0 TEMP (C) TEMP (C) Figure 5.18 Comparison of model prediction and experimental results of G ’ and tanS for 120 °C-transfer molded IFN

(FU/PES=25/75), ^ = 0.1, Vc=0.82 258 259

♦ G" (1-0) G"

AA = G !2 + G " 2 A A

CC = G^ 2 + gj:2 ♦ = vc/d-vB)

The calculated values of G ’ and tani are compared with

experimental data in Figure 5.18 where VgC=0.1V^ and Vc=0.82. These

values are chosen based on the morphology shown in the TEM micrograph

and the best fit of the G' curve. The results show that the

temperature dependence of G ’ can be fitted reasonably well, there is

reveals a large deviation between experimental data and model prediction for tanS.

According to Figures 5.12b and 5.17a, it can be hypothesized that the 120 °C-transfer molded PU/PES=50/50 IPN has a two-continuous matrix

(phase C), in which large droplets of polyurethane are dispersed within the polyurethane droplets and a small polyester subdomain exists. Shown in Figure 5.19 are a simplified structure of this sample and a simplified mechanical model. G* and G" of this structure can be calculated from the following equations:

(5-20a)

(5-20b) G’ vs. TEMP O TAN <5 vs. TEMP in

O MOLD TEMP 1 2 0 C 00 PU/PES—50/50 00

O

A : P E S

C : PU/PES Z V - 0 .5 < o V*" - 0 .8 H*

K) (0 o o

experimental d a ta O MODEL PREDICATION ' O

1 30.0 *50.0 10.0 70.0 1 30.0 -50.0 10.0 70.0 TEMP (C) TEMP (C)

Figure 5.19 Comparison of model prediction and experimental results of G' and tanS for 120 °C-transfer molded IPN (PU/PES=50/50), V0C=O.5, Vc=0.8 to 05 where

9

M = Q^ 2 + G^ 2

BB = Gg2 + G£2

V = 1 - V, C

♦ = VbA

The calculated values of G ’ and tanS are compared with experimental data shown in Figure 5.19 where Vnr,=0.5Vri and V_=0.8. The results show that the temperature dependence of G ’ can be fitted reasonably well, but tanS reveals a large deviation between experimental data and model prediction. Figure 5.20 shows the modeling results of 8 0 °C-transfer molded PU/FES=50/50 IPN (V =0.5V , V =0.8), based on Figures 5.10b d u L/ U and 5.17b. G ’ and G" are calculated from the following equations:

F|_____ G’ p»2 + f„2 (5-21a)

F" G" (5-21b) where G* v s. TEMP O TAN <5 vs. TEMP in N O MOLD TEMP 8 0 C 00 PU/PES-50/50 00

o

in B : PU N C : PU/PES Z V - 0 .5 < o V- - 0.8 H

K) (0 O o

EXPERIMENTAL DATA O MODEL PREDICATION O o -50.0 10.0 70.0 1 30.0 -50.0 10.0 70.0 1 30.0 TEMP (C) TEMP (C) Figure 5.20 Comparison of model prediction and experimental results of G* and tanS for 80 °C-transfer molded IPN (FV/PES=50/50), Vg^O.5, Vc=0.8 to 05 263

vc Gc + «1“vc> pM - ”CC "72_””2"

P‘ = (1-M G’ U G J

P" = (l-X) + )i G"

cc = g^2 + gj;2

k a

Similar to the previous modeling, the temperature dependence of G ’

shows good agreement between calculated values and experimental data.

But tan<5 reveals a large deviation between experimental data and model prediction.

According to Figures 5-10a and 5-17b, the 80 °C-transfer molded

PU/PES=75/25 IPN may be considered as polyester droplets dispersed in a continuous polyurethane matrix. Using Takayanagi's series model

(i.e., Equation 5-17), the predicted G ’ and tanS are shown in Figure

5.21. Again, the predicted G ’ agrees well with experimental data. The predicted tanS does show a larger polyurethane peak than the polyester peak, which agrees with the experimental result. But the deviation of the model prediction is still very large.

5.5 SUMMARY AND CONCLUSIONS

Morphology and dynamic mechanical properties of the molded polyurethane-polyester IPN depend strongly on the processing G ’ vs. TEMP O TAN 6 vs. TEMP in

o MOLD TEMP 8 0 C 00 00

o

in M A : P E S B : PU o 0 .7 5

K) 10 O ■ O PU/PES-75/25

EXPERIMENTAL DATA O MODEL PREDICATION ' O -50.0 10.0 70.0 130.0 50.0 10.0 70.0 130.0 TEMP (C ) TEMP (C )

Figure 5.21 Comparison of model prediction and experimental results of G* and tanJ for 80 *C-transfer molded IPN (FU/PES=75/25), VB=0.75 to conditions such as molding temperature, reaction rate, and reaction sequence. In most cases, all constituent polymers appear in both matrix and dispersed phases. Using several simplified structure models based on the morphology observed from the electron microscopy, the storage modulus of IPN samples can be predicted reasonably well. There are, however, large deviations in the predictions of tanS. The fine structures of these samples are apparently very complicated and cannot be explained by the existing models. CHAPTER VI

RECOMMENDATIONS

As a consequence of this study, the following areas are recommended for continuing work in this field.

1. The kinetic framework developed in Chapters 3 and 4 offers the

possibility of a more quantitative understanding of reaction

kinetics in polyurea systems being developed for RIM. By using

solution polymerization and FTIR technologies, it is possible to

conduct isothermal experiments and follow the aliphatic

triamine/diisocyanate and aromatic diamine/diisocyanate reactions

individually. The kinetic parameters determined based on the

solution polymerization results predict fairly well the adiabatic

temperature rise of bulk polyurea reactions in RIM. In this study,

the amine groups in aliphatic triamine are assented to have an equal

reactivity as are the amine groups in aromatic diamine. In order to

facilitate accurate process modelling, especially the rheo-kinetic

changes, in polyurea RIM the effects of unequal reactivity of

aliphatic amine groups or aromatic amine groups on phase separation

are needed. The study should include chemical systems

representative of commercial ureas.

266 267

2 . Only solution polymerizations of polyurea systems were carried out

by infrared spectroscopy in this study. With the continuous

progress in FTIR technology, it should be possible to handle

polyurea bulk polymerizations. For example, it is necessary to

facilitate the study of resin-fiber interface in glass fiber

reinforced polyurea or polyurea structural RIM. Current FTIR

instrumentation allows data collection with a good signal to noise

ratio at rates as high as five or even ten spectra per second, and

this should not be the main limitation in carrying out these

studies. More important perhaps is to develop RIM-like mixing and

injecting techniques that permit the mixtures to be loaded in the

FTIR sampling compartment before complete reaction has taken place.

3. The structure-property and processing-property relationships in

amorphous polyurea, especially the physical crosslinking or the

formation of hard segment, is an area which has not been thoroughly

explored. For example, why are the thermal properties of polyureas

superior to polyurethanes with crystalline hard domains? One

difference between ureas and urethanes which has been proposed to

answer this question is that the presence of 3-D hydrogen-bonding

ureas at the interface between hard and soft domains leads to the

superior properties of urea (Dominquez, 1985). Apparently, the way

polymer chains are built differs significantly between ureas and

urethanes. This fundamental difference in the phase formation

mechanism may actually be the primary reason for superior properties in ureas, and manipulation of this difference may improve processibility and properties further.

Research on the flow analysis of mold filling and curing of polyurea should be extended to investigate the possibility of polyurea RIM.

IPN seems to be a unique way to blend two thermosetting materials together. Therefore, its applications should be extended. At present, only a few commercial IPN compounds are available, such as polyurethane/acrylamate of Ashland Chemical and hybrids of Amoco

Chemical. Other thermosetting polymers like nylon, epoxy, polyurea, and thermoplastic polymers like methacrylate (M*1A) should be tried. REFERENCES

Abouzahr, S. and Wilkes, G.L., Ch. 5 in "Processing, Structure and Properties of Block Copolymers," Folkes, M.J., ed., Elsevier Science Publishing Co. Inc., New York, 1985.

Allen, G., Bowden, M.J., Blundell, D.J., Hutchinson, F.G., Jeffs, G.M., Vyvoda, J., and White, T., Polymer. 14, 597 (1973); Ibid., 14, 604 (1973).

Allen, G. Bowden, M.J., Lewis, G., Blundell, D.J., and Jeffs, G.M., Polymer. 15. 13 (1974a).

Allen G. Bowden, M.J., Lewis, G., Blundell, D.J., Jeffs, G.M., Vyvoda, J., and White, T., Polymer. 15. 19 (1974b); Ibid., 1.5, 28 (1974c).

Allport, D.C. and Janes, W.H., "Block Copolymers," John Wiley and Sons, New York, 1973.

Baker, J.W. and Bailey, D.N., J. Chem. Soc.. Part V I . 4649 (1957a); Part VII. 4652 (1957b); Part VIII. 4663 (1957c).

Baker, J.W., Davis, M.M., and Gaunt, J., J. Chem. Soc.. Part IV. 22 (1949).

Baker, J.W. and Gaunt, J., J. Chem. Soc.. Part II. 9 (1949a); Part III. 19 (1949b); Part V . 27 (1949c).

Bauer, P., Henning, J., and Schreyer, G., Angew. Makromol. Chem. 11, 145 (1970).

Baumann, W.M., Scott, D.W., Klempner, D., Frisch, K.C., Xiao, H.X., and House, R.V., J. Elast. Plast., 18, 233 (1986).

Bayer, 0., Miller, E., Peterson, S., Piepenbrink, H.F., and Windemuth, E., Rubber Chem. Technol.. 23. 812 (1953); translated from: Agnew. Chemie. 62. 57 (1950).

Bidstrup, S.A. and Macosko, C.W., in "Proceedings of Soc. Plast. Eng. Tech. Conf.," 278 (1984).

Blackwell, J. and Gardner, K.H., Polymer. 20. 13 (1979).

269 270

Blackwell, J. and Nagarajan, M.R., Polymer, 22, 202 (1981).

Blackwell, J., Nagarajan, M.R., and Hoitink, T.B., Polymer. 22, 1534 (1981).

Blackwell, J., Nagarajan, M.R., and Hoitink, T.B., Polymer. 23, 950 (1982).

Blackwell, J. and Ross, M., J. Polym. Sci., Polym. Letters, 17. 447 (1979).

Blake, J.W. et al., Polym. Eng, and Sci., 27, 1236 (1987).

Bleijenberg, C.A.M., Heikens, D., Meijers, A., Lampe, H.G.M., and von Reth, P.H., Brit. Polym. J.. 4, 125 (1972).

Bonart, R., J. Macromol. Sci., Phys.. B2, 115 (1968).

Bonart, R . , Morbitzer, L., and Hentze, G., J . Macromol. Sci.. Phys.. B3, 337 (1969).

Bonart, R., Morbitzer, L., and Muller, E.H., J . Macromol. Sci. Phys.. B9 , 447 (1974).

Bonart, R., Morbitzer, L., and Rinke, H., Kolloid-Z. u.Z. Polymer. 240. 807 (1970).

Bonart, R. and Muller, E.H., J. Macromol. Sci.. Phys.. B10, 177 (1974a).

Bonart, R. and Muller, E.H., J . Macromol. Sci.. Phys.. B10, 345 (1974b).

Borkent, G., Adv. Urethane Sci. Technol.. 3, 1 (1974).

Bowden, M.J., Blundell, D.J., Longman, G.W., and Wignall, G.D., Polymer, 15, 33 (1974).

Boyarchuk, Y.M., Rappoport, L.Y., Nikitin, V.N., and Apukhtina, N.P., Polym. Sci.. U.S.S.R.. 7, 859 (1965).

Broutman, L.J., and Krock, R.H., "Modem Composite Materials," Addison-Wesley, Massachusetts, 1967.

Broyer, E. and Macosko, C.W., AJChE J .. 22, 268 (1976).

Brunette, C.M., Hsu, S.L., and Macknight, W.J., Macroroolecules. 15, 71 (1982). 271

Bunn, C.W., J. Polym. Sci.. 16, 323 (1955).

Camberlin, Y., P&scault, J.P., Letoffe, J.M., and Claudy, P., J. Polym. Sci., Polym. Chem.. 20, 383 (1982a).

Camberlin, Y., Pascault, J.P., Letoffe, J.M., and Claudy, P., J. Polym. Sci., Polym. Chem., 20, 1445 (1982b).

Camargo, R.E., Ph.D. Dissertation, University of Minnisota, 1984.

Camargo, R.E., Gonzalez, V.M., and Macosko, C.W., Rubber Chem. Technol.. 56, 774 (1983).

Casey, J.P., Milligan, B., and Fasolka, M.J., in "Proceedings of the SPI 28th Annual Technical/Marketing Conference," San Antonio, Texas, October 1984.

Casey, J.P., Milligan, B., and Fasolka, M.J., J . Elast. Plast.. 17, 218 (1985).

Castro, J.M., Ch. 2 in "Development in Plastics Technology," Vol. 2 . Whelan, A. and Craft, J.L., ed., Applied Science, London, 1985.

Castro, J.M., Ph.D. Dissertation, University of Minnisota, 1980.

Castro, J.M., Lipshitz, S.D., and Macosko, C.W., AIChE J.. 28, 973 (1982).

Castro, J.M., Lopez-Seranno, F., Camargo, R.E., Macosko, C.W. and Tirrell, M . , J Appl. Polym. Sci.. 26. 2067 (1981).

Castro, J.M. and Macosko, C.W., SPE ANTEC Tech. Pap.. 26, 434 (1980).

Castro, J.M. and Macosko, C.W., AIChE J.. 28, 250 (1982).

Chang, A.L., Briber, R.M., Thomas, E.L., Zdrahala, R.J. and Critchfield, F.E., Polymer. 23, 1060 (1982).

Chen, C.H.Y., Briber, R.M., Thomas, E.L., Xu, M . , and Macknight, W. J., Polymer. 24, 100 (1983).

Chen, Z.S., Yang, W.P., and Macosko, C.W., Rubber Chem. Technol.. 61, 8 6 (1988).

Cheng, C.Y., M.S. Thesis, The Ohio State University, 1988.

Chen-Tsai, C.H.Y., Thomas, E.L., MacKnight, W.J., and Schneider, N.S., Polymer. 27, 659 (1986). 272

Clough, S.B. and Schneider, N.S., J . Macromol. Sci.. Phys.. B2. 553 (1968).

Cooper, S.L. and Tobolsky, A.V., J. Appl. Polym. Sci.. 10, 1837 (1966).

Cooper, S.L. and Tobolsky, A.V., J. Appl. Polym. Sci., 11, 1361 (1967).

Coulter, J.P. and Guceri, S.I., "Proceedings of Manufacturing International," Vol. 4 , 79 (1988).

Critchfield, F,E., Koleske, J.V., Magnus, G., and Dodd, J.L., Adv. Urethane Sci. Technol.. 2, 141 (1973); J . Elastoplast., 4, 22 (1972).

Davis, T.L. and Ebersole, F., J. Am. Chan. Soc.. 56. 883 (1934).

Davies, W.E.A., J. Phys. P.. 4, 1176, 1325 (1971).

Debye, P. and Bueche, A., J . Chem. Phys.. 16, 573 (1948).

Djomo, H., Colmenares, R., and Mayer, G.C., Eur. Polym. J., 17, 521 (1981).

Djomo, H., Morin, A., Damyanidu, M., and Meyer, G.C., Polymer. 24, 65 (1983).

Dominguez, R.J.G., J. Cell Plast.. Nov/Dec., 433 (1984).

Donatelli, A.A., Sperling, L.H., and Thomas, D.A., J. Appl. Polym. Sci.. 21, 1189 (1977).

Donatelli, A.A., Sperling, L.H., and Thomas, D.A., Macromolecules, 9 , 671, 676 (1976).

Ebdon, J.R., Hourston, D.J., and Klein, D.G., Polymer, 25. 1633 (1984).

Edwards, H.R., in "Proceedings of Soc. Pleat. Eng. Tech. Conf.," 1326 (1986).

Edwards, H.R., "Soc. Plast. Tech. Conf.," 1326 (1986).

Estes, G.M., Cooper, S.L., and Tobolsky, A.B., J . Macromol. Sci., Revs. Macromol. Chem.. C4, 313 (1970a).

Estes, G.M., Huh, D.S., and Cooper, S.L., P. 225 in "Block Polymers," S.L. Aggarwal, ed., Plenum Press, New York, 1970b.

Ewen, J.H., J. Elast. Pleat.. 17, 281 (1985). 273

Fara, R.A. Polymer, 13, 127 (1972).

Fedors, R.F., Polym. Eng. Sci.. 14. 147'(1974); Ibid., 14, 472 (1974).

Ferber, G., General Motors Company, Private Communication, 1986.

Ferguson, J. and Ahmad,N., European Polym. J..13. 859 (1977a).

Ferguson, J. and Ahmad,N., European Polym. J..13. 865 (1977b).

Ferguson, J. and Kumar, M., Plast. Rubber Proc.Appl.. !, 259 (1981).

Ferguson, J. and Patsavoudis, D., European Polym. J.. 8 , 385 (1972).

Florentini, C., U.S. Patent. 4,332,335 (1982).

Flory, P.J., "Principles of Polymer Chemistry," Cornell University Press, Ithaca, New York, 1953.

Fridman, I.D. and Thomas, E.L., Polymer. 21, 388 (1980).

Fridman, I.D., Thomas, E.L., Lee, L.J., and Macosko, C.W., Polymer. 21, 393 (1980).

Frisch, H.L., Foreman, R . , Schwartz, R . , Yoon, H., Klempner, D., and Frisch, K.C., Polym,. Eng. Sci.. 19. 284 (1979).

Frisch, K.C., Rubber Chem. Tech.. 53. 126 (1980).

Frisch, K.C., Klempner, D., Frisch, H.L., and Ghiradella, H., "Recent Advances in Polymer Blends, Grafts, and Blocks", Sperling, L.H., ed., Plenum Press, New York, (1974a).

Frisch, K.C., Klempner, D., Migdal, S., Frisch, H.L., and Ghiradella, H., Polym. Eng. Sci.. 14, 76 (1974b).

Frisch, K.C., Klempner, D., Antczak, T., and FriBch, H.L., J. Appl. Polym. Sci.. 18. 683 (1974c).

Frisch, K.C., Klempner, D., and Mukherjee, S.K., J. Appl. Polym. Sci.. 18, 689 (1974d).

Frisch, K.C., Klempner, D., Migdal, S., and Frisch, H.L., Polym. Eng. Sci.. 12, 885 (1974e).

Frisch, K.C., Klempner, D., Kim, S.C., and Frisch H.L., Paper B-42. SIM. IUPAC. Brazil (1974).

Frisch, K.C., Klempner, D., Migdal, S., Thomas, H.L., and Dunlop, A.P., J. Appl. Polym. Sci.. 19, 1893 (1975). 274

Frisch, K.C., Klempner, D., and Frisch, H.L., SAE paper 820422, Int. Cong. & Expo.. Detroit, MI, February 1982.

Fujino, K., Ogawa, Y . , and Kawai, H., J. Appl. Polym. Sci.. 8 , 2147 (1964).

Gerkin, R.M. and Critchfield, F.E., in "Proceedings of Soc. Auto Eng. Tech. Conf.," Paper No. 770837, 1976.

Gibson, P.E., Vallance, M.A., and Cooper, S.L., Ch. 6 in "Developments in Block Copolymers-1," I. Goodman, ed., Appl. Sci. Publishers, New York, 1982.

Gillham, J.K., Polym. Eng. Sci., 19, 676 (1979).

Gonza'lez-Romero, V.M. and Macosko, C.W., J. Rheology. 29. 259 (1985).

Graessley, W.W., Acc. Chem. Res.. 10. 332 (1977).

Greassley, W.W., Adv. Polym. Sci., 16. 48 (1974).

Graessley, W.W., Hazleton, R.L., and Lindeman, L.R., Trans. Soc. Pheol.. 11, 267 (1967).

Graessley, W.W., Masuda, T., Roovers, J.E.L., and Hadjichristidis, N., Macromolecules. 9, 127 (1976).

Grigsby, R.A. Jr. and Dominguez, R.J.G., in "Preceedings of the SPI 29th Annual Technical/Marketing Conference," Reno, Nevada, October 1985.

Grigsby, R.A. Jr., Rice, D.M., and Dominguez, R.J.G., Soc. Plast. Eng. RETEC Papers. 6 6 (1987).

Gupta, A., Cizmeciaglu, M., Coulter, D., Liang, R.H., Yarrouian, A., Tsay, F.D., and Maacanin, J., J. Appl. Polym. Sci.. 28. 1011 (1983).

Hager, S.L., McRury, T.B., Gerkin, R.M. and Critchfield, F.E., A.C.S. Symposium Series. Urethane Chemistry and Applications. 172. 149 (1981).

Han, C.D. and Lem, K.W., J. Appl. Polym. Sci.. 28. 3155 (1983).

Harrell, L.L. Jr., Macromolecules. 2, 607 (1969).

Hashimoto, T., Tsukahara, Y . , Tachi, K., and Kawai, H., Macromolecules. 16. 648 (1983).

Hashin, Z., and Shtrikman, S., J. Mech. PhvB. Solids. 11, 127 (1963). 275

Hayashi, T., Ito, J., Mitani, K . , and Mizutani, Y., J. Appl. Polym. Sci.. 28, 2867 (1983).

Hayashi, T., Ito, J., Mitani, K . , and Mizutani, Y., J. Appl. Polym. Sci.. 33, 375 (1987).

Heikens, D., Meijers, A., von Reth, P.H., Polymer. 9, 15 (1968).

Hemphill, J.J. and Vanderhider J.A., Soc. Plast. Eng. RETEC Papers. 70 (1987).

Hesketh, T.R., van Bogart, J.W.C., and Cooper, S.L., Polym. Eng. Sci.. 20, 190 (1980).

Hill, R. and Walker, E.E., J. Polym. Sci.. 3, 609 (1948).

Hocker, J. and Bom, L., J. Polym. Sci.. Polym. Letters. 17. 723 (1979).

Hourston, D.J., Huson, M.G., and McCLUSKEY, J.A., J. Appl. Polym. Sci.. 31, 709 (1986).

Hourston, D.J. and McCLUSKEY, J.A., J. Appl. Polym. Sci.. 30. 2157 (1985); 31, 645 (1986).

Hourston, D.J. and Zia, Y., J. Appl. Polym. Sci., 28. 2139 (1983); 28, 3745 (1983); 28, 3849 (1983); 29, 629 (1984); 29, 2951 (1984); and 29, 2963 (1984).

Hsu, T.J., Ph.D. Dissertation, Hie Ohio State University, Columbus, Ohio (1987).

Hsu, T.J., and Lee, L.J., 42nd Annual Conference, Composite Institute, The Society of Plastic Industry, Inc., Session 18-C, Cincinnati, Ohio, February (1987).

Hsu, T.J. and Lee, L.J., J. Appl. Polym. Sci..33. 793 (1987).

Hsu, T.J. and Lee,L.J., J. Appl. Polym. Sci.. 36. 1157 (1988).

Hsu, T.J. and Lee, L.J., Polym. Eng. Sci.. 25.951 (1985).

Hsu, T.J. and Lee, L.J., Polym. Eng. Sci.. 28.956 (1988).

Huang, Y.J., and Lee, L.J., AIChE Journal. 31, 1585 (1985).

Hull, J.L., in "Encyclopedia of Polymer Science and Technology," 2nd Ed., Mark, H.F., Bikales, N.M., Overberger, C.G., and Menges, G., ed., John Wiley & Sons, Inc., New York, 1984. 276

Huh, D.S. and Cooper, S.L., Polym. Eng. Sci., H , 369 (1971).

Hwang, K.K.S., Yang, C.Z., and Cooper, S.L., Polym. Eng. Sci.. 21, 1027 (1981).

Ishihara, H., Kimura, I., Saito, K . , and Ono, H., J. Macromol. Sci.. Phys.. BIO. 591 (1974).

Jordhamo, G.M., Manson, J.A., and Sperling, L.H., Polym. Eng. Sci.. 26, 517 (1986).

Kajiyaroa, T. and Macknight, W, J., Polym. J.. 1., 548 (1970).

Kamal, M.R. and Sourour, S., Polym. Eng. Sci.. 13, 59 (1973).

Kaplan, D., and Tschoegl, N.W., Polym. Eng. Sci.. 14, 43 (1974).

Karger-Kocsis, J., and Kiss, L., Polym. Eng. Sci.. 27. 254 (1987).

Kato, K., Polym. Eng. Sci.. 7, 38 (1967).

Kelly, W.L., Plast. Eng.. February, 29 (1986).

Kendall, D.N., P. 5 in "Applied Infrared Spectroscopy," Reinhold, London, 1966.

Kemer, E.H., Proc. Phys. Soc.. 69B, 808 (1956).

Kim, S.C., Klempner, D., Frisch, K.C., Frisch, H.C., and Ghiradella, H., Polym. Eng. Sci., 15, 339 (1975).

Kim, S.C., Klempner, D., Frisch, K.C., Radigan, W., and Frisch, H.C., Macromolecules. 9, 258 (1976a).

Kim, S.C., Klempner, D., Frisch, K.C., and Frisch, H.L., Macromolecules. 9, 263 (1976b).

Kimura, I., Ishihara, H., Ono, H., Yoshihara, H., Nomura, S., and Kawai, H., Macromolecules. 7, 355 (1974).

Kircha, K., Mrotzek, W., and Menges, G., Polym. Eng. Sci.. 24. 974 (1984).

Kircher, K., Angew. Makrom. Chemie.. 76/77, 241 (1979).

Kircher, K., Mrotzek, W., and Menges, G., Polym. Eng. Sci.. 24, 974 (1984).

Klempner, D., Angew. Chem. Int. Ed. Engl.. 17, 97 (1978). Klempner, D., Frisch, K.C., and Frisch, H.L., J. Polym. Sci.. Polym. Lett. Ed.. 7, 775 (1969).

Klempner, D., Frisch, K.C., and Frisch, H.L., J. Polym. Sci.. Part A- 2, 8 , 921 (1970a).

Klempner, D . , Krrei, T.K., Mastsuo, M., and Frisch, H.L., Polym. Eng. Sci.. 10, 327 (1970b).

Klempner, D., Frisch, K.C., and Frisch, H.L., J. Elastoplastics. 3, 2 (1971).

Koberstein, J.T., Ph.D. Dissertation, University of Massachusetts, 1979.

Koberstein, J.T., Morra, B., and Stein, R.S., J. Appl. Cryst.. 13, 34 (1980).

Koutsky, J.A., Hien, N.V., and Cooper, S.L., J. Polym. Sci.. Polym. Letters. 8 , 353 (1970).

Kraus, G. and Rollman, K.W., J. Polym. Sci., Polym. Phys.. 14. 1133 (1976).

Kuryla, W.C., Critchfield, F.E., Platt, L.W., and Stamberger, D., J. Cellular Plast., March, 84 (1966).

Lagasse, R.R., J. Appl. Polym. Sci.. 21, 2489 (1977).

Lange, F.F., in Fracture and Fatigue, "Composite Materials," Vol. 5 . Broutman, L.J., ed., Academic, 1974.

Lee, L.J., "Polymer Reaction Injection Molding," ERC report, Hie Ohio State University (1987).

Lee, L.J., Polym. Eng. Sci.. 21, 483 (1981).

Lee, L.J., Rubber Chem. Tech.. 53. 542 (1980).

Lee, L.J. and Macosko, C.W., Int. J. Heat Mass Transfer. 23. 1479 (1980).

Lee, Y.M., Yang, Y.S., and Lee, L.J., Polym. Eng. Sci.. 27, 716 (1987).

Lin, D.J., Ottino, J.M., and 'niomas, E.L., Polym. Eng. Sci.. 25. 1155 (1985).

Lipshitz, S.D. and Macosko, C.W., J. Appl. Polym. Sci.. 21, 2029 (1977). Lipshitz, S.D. and Macosko, C.W., Polym. Eng. Sci., 16. 803 (1976).

Lopez-Serrano, F., Castro, J.M., Macosko, C.W., and Tirrell, M . , Polymer. 21. 263 (1980).

Lunardon, G., Sumida, Y . , and Vogl, O . , Die Angew. Makromol. Chem.. 87, 1 (1980).

Lyman, D.J., J. Polym. Sci.. 45. 49 (1960).

Macknight, W.J., Yang, M . , and Kajiyama, T., P. 99 in "Analytical Calorimetry," Porter, R.S. and Johnson, J.F., ed., Plenum Press, New York, 1968.

Macosko, C.W., British Polym. J.. 17. 239 (1985).

Macosko, C.W., Plastics Engineering. April, 21 (1983).

Macosko, C.W., "Fundamentals of Reaction Injection Molding," 1988.

Macosko, C.W. and Miller, D.R., Macromol.. 9, 199 (1976).

Maier, U. and Menges, G . , in "Proceedings of Soc. Plast. Eng. Tech. Conf.," 1350 (1986).

Malkin, A.Y., Frolov, V.G., Ivanova, A.N., and Andrianova, Z.S., Polym. Sci.. USSR. 21, 691 (1979).

Manson, J.A., and Sperling, L.H., "Polymer Blends and Composites," Plenum Press, New York, 1976.

Matsuo, M., Kwei, T.K., Klempner, D., and Frisch, H.L., Polym. Eng. Sci.. 10, 327 (1970).

Matsuo, M., Nozaki, C., and Jyo, Y . , Polym. Eng. Sci., 9, 197 (1969).

McQuiston, H., Plastics Engineering, October, 22 (1979).

Miller, J.A., Hwang, K.K.S., and Cooper, S.L., J. Macromol. Sci., Phys.. B22. 321 (1983).

Millar, J.R., J. Chem. Soc., 1311 (1960).

Molnar, J.A. and Lee, L.J., J. Appl. Polym. Sci.. 37, 2295 (1988).

Morin, A., Djomo, H., and Meyer, G.C., Polym. Eng. Sci.. 23. 394 (1983).

Mussatti, F. G., Ph.D. Dissertation, University of Minnesota, 1975. 279

Nakayama, K . , Ino, T., and Matsubara, I., J. Macromol. Sci.. Chem., A3, 1005 (1969).

Nalepa, C.J. and Eisenbraun. A.A., J . Elast. Plast.. 19. 6 (1987).

Nelson, L.E., J. Appl. Polym. Sci., 21, 1579 (1977).

Nelson, M., M.S. Thesis, The Ohio State University, 1987.

Nesterov, A.E., Lipatova, T.E., Dusek, K., Pelzbauer, Z., Houska, M . , Hradil, J., and Lipatov, Y.S., Die Angew. Makromol. Chem., 52. 39 (1976).

Neumuller, W. and Bonart, R . , J . Macromol. Sci. Phys.. B21. 203 (1982).

Ng, H.N., Allegrezza, A.E., Seymour, R.W., and Cooper, S.L., Polymer. 14, 255 (1973).

Nguyen, L.T., Ph.D. Dissertation, Massachusetts Institute of Technology, 1984.

Nguyen, L.T. and Suh, N.P., ACS Organic Coatings & Appl. Polym. Sci. Proc.. 49. 599 (1983).

Nguyen, L.T. and Suh, N.P., Polym. Eng. Sci.. 26. 799 (1986); Ibid., 26, 843 (1986).

Nguyen, L.T. and Suh, N.P., Polym. Proc. Eng.. 3, 37 (1985).

Nguyen, L.T. and Suh, N.P., Proceedeing of Annula Conference of Amer. Chem. Soc., div. of Polym. Proc., August 1983.

Nielsen, L.E., J. Appl. Polym. Sci.. 21. 1579 (1977).

Nielsen, L.E., J . Compos. Mater♦. J., 100 (1967).

Nielsen, L.E., "Mechanical Properties of Polymers and Composites," Vol. I and II. Marcel Dekker, Inc., New York, 1974.

Nielsen, L.E., Rheol. Acta. 13. 8 6 (1974).

Ninomi, M . , Katsuta, T., and Kotani, T., J. Appl. Polym. Sci., 19, 2919 (1975).

Nissen, D. and Markovs, R.A., J . Elast. Plast.. 15, 96 (1983).

Noshay, A. and McGrath, J.E., "Block Copolymers: Overview and Critical Survey," Academic Press, New York, 1977. 280

O ’Brien, J.C. and Lenosky, T., in "Modem Plastics Encyclopedia," Vol. 63, McGraw-Hill Co., New York, October 1986.

Ophir, Z.H. and Wilkes, G.L., Adv. Chem. Series, 176, 53 (1979).

Ophir, Z.H. and Wilkes, G.L., J. Polym. Sci., Polym. Phys.. 18. 1469 (1980).

Osinski, J.S., Manzione, L.T., and Chan, C., Polym. Proc. Eng., 3, 97 (1985).

Osinski, J.S., Polym. Eng. Sci., 23, 756 (1983).

Paik Sung, C.S., Hu, C.B., and Wu, C.S., Macromolecules. 13. Ill (1980a).

Paik Sung, C.S., Smith, T.W., and Sung, N. H., Macromolecules. 13, 117 (1980b).

Paik Sung, C.S. and Schneider, N.S., Macromolecules. 8 , 6 8 (1975).

Paik Sung, C.S. and Schneider, N.S., Macromolecules. 10. 452 (1977).

Pannone, M.C., M.S. Thesis, University of Minnisota, 1985.

Pannone, M.C. and Macosko, C.W., J. Appl. Polym. Sci., 34, 2409 (1987).

Pannone, M.C. and Macosko, C.W., Polym. Eng. Sci., 28. 660 (1988).

Peebles, L.H. Jr., Macromolecules, 7, 872 (1974).

Pemice, R . , Frisch, K.C., and Navare, R., J. Cell. Plast., March/April, 121 (1982).

Pigott, K.A., Frye, B.F., Aallen, K.R., Steingiser, S., Darr, W.C. and Saunders, J.H., J. Chem. and Eng. Data, 5, 391 (1960).

Pimentel, G.C. and McClellan, A.L.,"The Hydrogen Bond," W.H. Freeman, San Francisco, 1960.

Poole, A., J. Cellular Plastics, July, 261 (1985).

Raia, E., Purchasing, December 1977.

Reegen, S.L., and Frisch, K.C., J . Polym. Sci., Part A-l. 8 , 2883 (1970).

Richter, E.B. and Macosko, C.W., Polym. Eng. Sci., 18, 1012 (1978). 281

Richter, E.B. and Macosko, C.W., Polym. Eng. Sci.. 20, 14 (1980).

Richter, E.B. and Macosko, C.W., Polym. Eng. Sci., 20, 921 (1980).

Roller, M.B., Polym. Eng. Sci.. 26, 432 (1986).

Rosovizky, V.F., Ilarsky, M . , Hrouz, J., Dusek, K . , and Lipatov, Yu.S., J. Appl. Polym. Sci., 24, 1007 (1979).

Russo, R. and Thomas, E.L., J. Macromol. Sci., Phys., B22(4). 551 (1983).

Sacks, M.D. and Sheu, R.S., in "Proceedings of the 2nd. Int. Conf. on Ultrastructure Processing of Ceramics, Glasses and Composites," Palm Beach, Florida, February 1985.

Samuels, S.L. and Wilkes, G.L., J. Polym. Sci., A-2, 11., 807 (1973a).

Samuels, S.L. and Wilkes, G.L., J. Polym. Sci., Polym. Symp.. 43, 149 (1973b).

Saunders, J.H., Rubber Chem. Technol., 33. 1259 (1960).

Saunders, J.H., Burroughs, J.A., Williamns, L.P., Martin, D.H., Southern, J.H., Ballman, R.L., and Lea, K. R . , J. Appl. Polym. Sci., 19, 1387 (1975).

Saunders, J.H. and Frisch, K.C., "Polyurethanes Chemistry and Technology," Vol. 1 & 2 , Wiley-Interscience, New York, 1962.

Schaper, S., Kunstoffe. 6 , 400 (1976).

Schechter, L. and Wynstra, J., Ind. Eng. Chem.. 48, 8 6 (1956).

Schluter, K., U.S. Patent, 3,975,128 (1976).

Schlotterbeck, D.G., Matzke, G., Horn, P., and Schmidt, H.U., Plastics Engineering, Junuary, 37 (1989).

Schnerder, F.W., in "Proceedings of ACS Polym. Mat. Sci. & Eng. Div.," 49, 629 (1983).

Schneider, N.S., Desper, C.R., Illinger, J.L., and King, A.O., J. Macromol. Sci.. Phys., Bll. 527 (1975a).

Schneider, N.S., Paik Sung, C.S., Matton, R.W., and Illinger, J.L., Macromolecules. 8 , 62 (1975b).

Schollenberger, C.S., Adv. Chan. Series. 176, 83 (1979). 282

Schollenberger, C.S. and Dinbergs, K., Adv. Urethane Sci. Technol.. 7, 1 (1979).

Schollenberger, C.S. and Dinbergs, K., J. Elastoplastics. 5, 222 (1973).

Schollenberger, C.S., Scott, H., and Moore, G.R., Rubber World. 137, 549 (1958).

Schwartz, S.S. and Goodman, S.H., "Plastics Materials and Processes," Van Nostrand Rhinhold Co., New York, 1982.

Seefried, C.G. Jr., Koleske, J.V., and Critchfield, F.E., J. Appl. Polym. Sci., 19, 2494 (1975a).

Seefried, C.G. Jr., Koleske, J.V., and Critchfield, F.E., J. Appl. Polym. Sci.. 19. 2503 (1975b).

Seefried, C.G. Jr., Koleske, J.V., and Critchfield, F.E., J. Appl. Polym. Sci., 19, 3185 (1975c). * Seefried, C.G. Jr., Koleske, J.V., and Critchfield, F.E., Polym. Eng. Sci., 15, 646 (1975d).

Senich, G.A. and Macknight, W.J., Adv. Chem. Series, 176, 97 (1979).

Seymour, R.W. and Cooper, S.L., Macromolecules, 6, 48 (1973).

Seymour, R.W., Estes, G.M., and Cooper, S.L., Macromolecules, 3, 579 (1970).

Shen, M. and Kawai, H., AIChE J.. 24, 1 (1978).

Shibayama, K., Kobunshi Kagaku, 19, 219 (1962).

Shibayama, K., Kobunshi Kagaku. 20, 221 (1963).

Shibayama, K. and Suzuki, Y., Kobunshi Kagaku, 23. 24 (1966).

Shibayama, K., Zairyo. 12, 362 (1963).

Sibal, P.W., Camargo, R.E., and Macosko, C.W., Polym. Proc. Eng., 1, 147 (1983).

Sieyfried, D.L., Manson, J.A., and Sperling, L.H., J. Polym. Sci. Phys. Ed.. 16, 583 (1978).

Silberberg, J. and Han, C.D., J. Appl. Polym. Sci.. 22, 599 (1978). 283

Slowikowska, I. and Danieweska, I., J. Polym. Sci., Polym. Symp.. 55, 187 (1975).

Smith, T.L. and Magnussen, A.B., J. Polym. Sci.. 42, 391 (1960).

Sneller, J.A., Modern Plastics, February, 55 (1986).

Sperling, L.H., "Interpenetrating Polymer Networks and Related Materials," Pieman Press, New York, 1981.

Sperling, L.H., Material Engineering, 66, September (1980).

Sperling, L.H., Polym. Eng. Sci., 25. 517 (1985).

Sperling, L.H., Chiu, T.W., and Thomas, D.A., J. Appl. Polym. Sci., 17, 2443 (1973).

Sperling, L.H., Chiu, T.W., Hartman, C.P., and Thomas, D.A., Int. J . Polym. Mater., 1, 331 (1972a).

Sperling, L.H., Thomas, D.A., Covitch, M.J., and Curtius, A.J., Polym. Eng. Sci.. 12, 101 (1972b).

Sperling, L.H., Thomas, D.A., and Huelck, V., Macromolecules, 5, 340 and 348 (1972c).

Sperling, L.H., Donatelli, A.A., and Thomas, D.A., J. Appl. Polym. Sci.. 21, 1189 (1977).

Sperling, L.H., Donatelli, A.A., and Thomas, D.A., Macroroolecules. 9, 671 & 676 (1976).

Sperling, L.H. and Friedman, D.W., J. Polym. Sci., Part A-2, 7, 425 (1969).

Sperling, L.H. and Mihalakis, E.N., J. Appl. Polym. Sci., 17, 3811 (1973).

Sperling, L.H. and Sarge III, H.D., J. Appl. Polym. Sci., 16. 3041 (1972).

Sperling, L.H., Taylor, D.W., Kirkpatrick, George, H.F., and Bardman, D.R., J. Appl. Polym. Sci., 14. 73 (1970a).

Sperling, L.H., George, H.F., Huelck; V., and Thomas, D.A., J. Appl. Sci., 14, 2815 (1970b).

Srichatrapimuk, V.W. and Cooper, S.L., J . Macromol. Sci., Phys., B15, 267 (1978). 284

Staudinger J.J.P. and Hutchinson, F., U.S. Patent, 2,539,377 (1951).

Steinle, E.C., Critchfield, F.E., Castro, J.M., and Macosko, C.W., J. Appl. Polym. Sci.. 25. 2327 (1980).

Sweeney, F.M., "Introduction to Reaction Injection Molding," Technomic Publishing Co., Inc., Westport, CT, 1979.

Sweeney F.M., "Reaction Injection Molding Machinery and Processes," Marcel Dekker, Inc, Basel, New York, 1987.

Tadmor, Z, and Gogos, C.G., "Principles of Polymer Processing," John Wiley & Sons Inc., New York, 1979.

Takayanagi, M., Harima, H . , and Iwata, Y . , J. Soc. Mater. Sci.. Jpn.. 12, 389 (1973).

Takayanagi, M., Vermura, S., and Minami, S., J. Polym. Sci. C . 5, 113 (1964).

Tanaka, T., Yokoyama, T., and Yamaguchi, Y., J. Polym. Sci.. A-l, 6, 2137 (1968a).

Tanaka, T., Yokoyama, T., and Yamaguchi, Y . , J . Polym. Sci.. A-l, 6, 2153 (1968b).

Thomas, D.A. and Sperling, L.H., "Polymer Blends," Vol. 2 . Ch. 11, Plenum Press, New York, 1976.

Tirrell, M . , Lee, L.J., and Maqosko, C.W., P. 149 in "Polymerization Reactors and Processes," ACS Symposium Series, Henderson, J.N. and Bouton, T.C., ed., American Chem. Soc., Washington D.C., 1979.

Touhsaent, R.L., Thomas, D.A., and Sperling, L.H., J. Polym. Sci. Polym. Symp.. 46. 175 (1974).

Touhsaent, R.L., 11100133, D.A., and Sperling, L.H., in "Toughness and Brittleness of Plastics" (Deanin, R.D., and Crugnola, A.M., eds.), Adv. Chem. Ser. 154 Amer. Chem. Soc.. Washington, D.C., 1976.

Trifan, D.S. andTerenzi, J.F., J. Polym. Sci.. 28. 443 (1958).

Turi, E.A., "Thermal Characterization of Polymeric Materials," Academic Press, New York, 1981.

Turner, R.B., Spell, H.L. and Vanderhider, J.A., "Reaction Injection Molding and Fast Polymerization Reactions," Kresta, J., ed., Plenun Press, New York, 1982.

Valles, E.M. and Macosko, C.W., Macromolecules. 12, 521 (1979). 285

Valles, E.M. and Macosko, C.W., Macromolecules, 12. 673 (1979).

van Bogart, J.W.C., Bluemke, D.A., and Cooper, S.L., Polymer. 22. 1428 (1981). van Bogart, J.W.C., Lilaonitkul, A., and Cooper, S.L., Adv. Chem. Ser.. No. 176, 3 (1979). van Krevellen, D.W., "Properties of Polymers," Elsevier Scientific Publishing Co., New York, 1976.

Vespoli, N.P. and Alberino, L.M., J . Elast. and Plast.. 17, 173 (1985).

Vespoli, N.P. and Alberino, L.M., Polym. Proc. Eng.. 3, 127 (1985).

Vespoli, N.P., Alberino, L.M., Peterson, A.A., and Ewen, J.H., in "Proceedings of the SPI 29th Annual Technical/Marketing Conference," Reno, Nevada, October 1985.

Vespoli, N.P., Alberino, L.M., Peterson, A.A., and Ewen, J.H., J. Elast. Plast.. 18. 159 (1986).

Vespoli, N.P. and Marken, C.C., "Heat Transfer and Reaction Effects During Mold Filling of Fast Reacting, Polyurea RIM Systems," 1987 Aannual AIChE Meeting, NYC, New York, 1987.

Voelker, T.H. and Balling, P., P. 133 in "30th Annual Polyurethane Technical/Marketing Conference," October 1986. von Bogart, J.W.C., Gibson, P.E., and Cooper, S.L., J. Polym. Sci.. Polym. Phys. Ed.. 21. 65 (1983). von Hassel, A., Plastics Technology. November, 69 (1981).

Wallner, J., Plastics Tech.. 24. 61 (1978).

Wang, C.B. and Cooper, S.L., Macromolecules. 16. 775 (1983).

Weisfeld, L.B., Little, J.R., and Wolstenholme, W.E., J. Polym. Sci.. 56, 455 (1962).

Wigotsky, V., Plastics Engineering. May, 19 (1986).

Wilkes, C.E. and Yusek, C.S., J . Macromol. Sci.. Phys.. B7. 157 (1973).

Wilkes, G.L. and Rnerson, J.A., J . Appl. Phys.. 47, 4261 (1976). 286

Wilkes, G.L., Mody, P.C., and Tant, M.R., Polym. Eng. Sci.. 19. 1029 (1979).

Wilkes, G.L., Samuels, S.L., and Crystal, R., J . Macromol. Sci.. Phys.. BIO. 203 (1974).

Wilkinson, T.C., Borgnaes, D., Chappel, S.F., and Kelly, W.L., in "Proceedings of ACS Polym. Mat. Sci., & Eng. Div." 49, 469 (1983).

Willey, S.J., in "Proceedings of the 2nd Would Congr. Chem. Eng., Montreal," VI, 453 (1981).

Williams, M.L., J. Phys. Chem.. 59. 95 (1955).

Williams, M.L., Landel, R.F., and Ferry, J.D., J. Am. Chem. Soc.. 77, 3701 (1955).

Willkomn, W.R., Chen, Z.S., Macosko, C.W., Gobran, D.A., and Thomas, E.L., Polym. Eng. Sci.. 28. 888 (1988).

Wood, A.S., M o d e m Plastics. 53. 35 (1976).

Wood, AS., M o d e m Plastics. October, 48 (1984).

Yang, C.Z., Hwang, K.K.S., and Cooper, S.L., Makromol. Chem.. 184, 651 (1983).

Yang, Y.S. and Lee, L.J., Macromolecules. 20. 1490 (1987).

Yenwo, G.M., Sperling, L.H., Pulido, J., Manson, J.A., and Conde, A., Polym. Eng. Sci., 17. 251 (1977).

Yeo, J.K., Sperling, L.H., and Thomas, D.A., Polym. Eng. Sci.. 21. 696 (1981).

Yokoyama, T., Adv. Urethane Sci. Technol.. 6, 1 (1978).

Zdrahala, R.J., Gerkin, R.M., Hager, S.L., and Critchfield, F.E., J. Appl. Polym. Sci.. 24. 2041 (1979).

Zdrahala, R.J., Hager, S.L., Gerkin, R.M., and Critchfield, F.E., J. Elastomers and Plast.. 12. 225 (1980). APPENDIX A

FTIR MACRO Programs

1. Scanning Program for Polyurea Reactions

2. Fourier Transfer Program

3. Spectrum Analysis Program

287 APPENDIX A.l Scanning Program for Polyurea Reactions

INTEGER RESOL INTEGER COUNT INTEGER STOP INTEGER FOOUNT INTEGER FSTOP INTEGER FDD INTEGER NUM INTEGER FIRST INTEGER SECOND INTEGER THIRD INTEGER FORTH INTEGER ID INTEGER YN INTEGER CHECK REAL TM REAL TIME STRING TITLE STRING TTLSAM STRING FILE

CRT SCANBKG ( For background scanning ) CR CR ">> SCANBKG <<" CR CR "> CHECK TOE FOLLOWING PARAMETERS FOR BACKGROUND SCANNING CR RESOL=4 CR "> CHECK RESOLUTION : " RESOL= " CM-1" GAN=1 CR "> CHECK GAN " GAN BDL=1 CR "> CHECK BDL " BDL NSB=3 CR "> CHECK NSB " NSB CR CR "> READY? PRESS ANY KEY TO START BACKGROUND SCANNING !" PAU CR CR "> SCANNING, PLEASE W A I T ...... " SCB DSB END

CRT SCANFST ( For fast reaction ) CR CR "»> SCANFST <<<" CR YN=0 CR ">D0 YOU WANT TO RUN BACKGROUND SCANNING CR " <0> YES <1> NO " CR "YN " YN IF YN EQZ THEN SCANBKG ENDIF CR CR "> INSERT 1ST BLANK DISK IN D1 AND 2ND BLANK DISK IN D2 CR "> AFTER INSERT DISKS, PRESS ANY KET TO MOUNT D1 & D2 !" PAU CR D2 D1 CR CR "> CHECK TOE FOLLOWING PARAMETERS FOR SAMPLE SCANNING !" CR RESOL=4 CR "> CHECK RESOLUTION RESOL= " CM-1" GAN=1 CR "> CHECK GAN " GAN BDL=3 CR "> CHECK BDL " BDL NSS=3 CR "> CHECK NSS " NSS CR TITLE="T5000/I1305/DMTDA" CR "> CHECK TOE MAIN TITLE CR "TITLE " TITLE TITLE=TITLE+" " EXT=RESOL TITLE=TITLE+EXT TITLE=TITLE+"CM-1" TITLE=TITLE+" GAN=" EXT=GAN TITLE=TITLE+EXT TITLE=TITLE+" BDL=" EXT=BDL TITLE=TITLE+EXT TIB=TITLE+" NSB=" EXT=NSB TIB=TIB+EXT TIS=TITLE+" NSS=" EXT=NSS TIS=TIS+EXT CR FILE="TIM111." CR "> CHECK TOE MAIN FILE NAME (*##*.) CR "FILE " FILE IF YN EQZ TOEN PDB FILE+"BKG" 290

ENDIF CR (X>UNT=0 CR "> CHECK THE STARTING FILE NUMBER COUNT STOP=110 CR ”> CHECK THE FINAL FILE NUMBER STOP FSTOP=38 PCOUNT=FSTOP FDD=0 CR CR "> READY? PRESS ANY KEY TO START SAMPLE SCANNING !" PAU CR CR "> SCANNING, PLEASE W A I T ...... " BEGIN CLI EXT=OOUNT PDI FILE+EXT CR CR "> THE CURRENT FILE FILE= EXT= OOUNT=OOUNT+l CHECK=STOP-COUNT FCOUNT=FCOUNT-l IF POOUNT EQZ THEN IF FDD EQZ THEN FDD=1 CR D2 CR ">>> REPLACE A NEW DISK IN D1 <— " ELSE FDD=0 CR D1 CR ">» REPLACE A NEW DISK IN D2 <— " ENDIF POOUNT=FSTOP ENDIF UNTIL CHECK LTZ CR CR ”> FINISHED! THE FINAL FILE FILE= FILE= EXT= CR D1 END ASG SCANFST 1

CRT SCANMDR ( For moderate reaction ) CR CR ”>>> SCANMDR <<<" CR YN=0 CR "> DO YOU WANT TO RUN BACKGROUND SCANNING CR " <0> YES <1> NO " CR "YN " YN IF YN EQZ THEN SCANBKG ENDIF CR CR "> INSERT A BLANK DISK IN D2 !" CR "> AFTER INSERT DISKS, PRESS ANY KEY TO MOUNT D2 !" PAU CR D2 CR CR "> CHECK THE FOLLOWING PARAMETERS FOR SAMPLE SCANNING CR RESOL=4 CR "> CHECK RESOLUTION RESOL= " CM-1" GAN=1 CR "> CHECK GAN " GAN BDL=1 CR "> CHECK BDL " BDL NSS=3 CR "> CHECK NSS " NSS CR TITLE="T5000/I1305/DMrDA" CR "> CHECK THE MAIN TITLE CR "TITLE " TITLE TTLSAM=TITLE+" RXN TIME : " TITLE=TITLE+" " EXT=RESOL TITLE=TITLE+EXT TITLE=TITLE+"CM-1" TITLE=TITLE+" GAN=" EXT=GAN TITLE=TITLE+EXT TITLE=TITLE+" BDL=" EXT=BDL TITLE=TITLE+EXT TIB=TITLE+" NSB=" EXT=NSB TIB=TIB+EXT CR FILE="TIM111." CR "> CHECK TOE MAIN FILE NAME (#####.) :" CR "FILE " FILE IF YN Et*Z THEN PDB FILE+"BKG" ENDIF CR COUNT=0 CR "> CHECK THE STARTING FILE NUMBER COUNT STOP=120 CR "> CHECK THE FINAL FILE NUMBER STOP CR CR "> READY? PRESS ANY KEY TO START SAMPLE SCANNING !" PAU CR CR "> SCANNING, PLEASE W A I T ...... " TIME=0.0 FIRST=60 SECOND=80 THIRD=100 F0RTH=110 (FIFTH=STOP-FORTH) ID=0 BEGIN IF ID NEZ THEN NUM=FIRST-COUNT+l IF NUM GTZ THEN BDL=28 (1 MIN.) ELSE NUM=SEOOND-OOUNT+1 IF NUM GTZ THEN BDL=66 (2 MIN.) ELSE NUM=THIRD-OOUNT+1 IF NUM GTZ THEN BDL=183 (5 MIN.) ELSE NUM=PORTH-OOUNT-1 IF NUM GTZ THEN BDL=376 (10 MIN.) ELSE BDL=1150 (30 MIN.) ENDIF ENDIF ENDIF ENDIF TM=1.55*BDL 1M=TM+17.02 TM=TM/60.0 TIME=TIME+TM ELSE BDL=1 ENDIF SCS DSS EXT=TIME TIS=TTLSAM+EXT TIS=TIS+" MIN" EXT=OOUNT PDS FILE+EXT 293

RAS ABS DSS DRAW CR CR "> THE CURRENT FILE FILE= EXT= ID=1 00UNT=00UNT+1 CHECK=STOP-OOUNT UNTIL CHECK LTZ CR CR "> FINISHED! THE LAST FILE FILE= EXT= CR " TOTAL RXN TIME TIME= " MIN." BDL=1 FXF=4600 LXF=600 CR D1 END ASG SCANMDER 2 APPENDIX A.2 Fourier Transfer Program

INTEGER IOOUNT INTEGER COUNT INTEGER STOP INTEGER FOOUNT INTEGER FSTOP INTEGER CHECK REAL TIME REAL SEC REAL DT REAL DTI STRING TITLE STRING TTLSAM STRING FILE

CRT FRTRSF (-- Fourier transfer ---) CR CR "»> FRTRSF <<<" CR CR "> INSERT THE INTERFEROGRAM DISK " CR " IN D1 AND THE EMITTANCE DISK IN D2 !" CR "> AFTER INSERT DISKS, PRESS ANY KEY TO MOUNT D1 & D2 !" PAU CR D2 D1 CR CR "> CHECK TOE FOLLOWING PARAMETERS FOR FOURIER-TRANSFER !" CR BDL=3 CR "> CHECK BDL " BDL NSS=3 CR "> CHECK NSS " NSS CR FILE="TIM111." CR "> CHECK TOE MAIN FILE MANE (#####.) CR "FILE " FILE GDB FILE+"BKG" DSB CR D2 PDB FILE+"BKG" CR TITLE="T5000/I1305/DMTDA" CR "> CHECK TOE MAIN TITLE CR "TITLE " TITLE TTLSAM=TITLE+" RXN TIME : " CR ICOUNT=0 CR "> CHECK TOE INITIAL INTERFEROGRAM FILE NUMBER TO TRANSF. CR "IOOUNT " IOOUNT 295

OOUNT=0 CR "> CHECK THE STARTING EMITTANCE FILE NUMBER TO WRITE ON !" CR "COUNT " COUNT STOP=110 CR "> CHECK THE FINAL FILE NUMBER TO TRANSFER !" CR "STOP " STOP SEC=0.0 CR "> CHECK THE STARTING TIME OF REACTION (SEC) :" CR "SEC " SEC FSTOP=38 CR "> CHECK THE FILES IN DISK 1 TO BE POURIER-TRANSFERED !" CR "FSTOP " FSTOP POOUNT=FSTOP CR CR "> READY? PRESS ANY KEY TO START FOURIER-TRANSFERING !" PAU CR CR "> TRANSFERRING, PLEASE W A I T ...... " DT=1.52*BDL DT1=1.83*NSS DT=DT+DT1 DT=DT+5.67 (4.26) SEC=SEC-DT BEGIN CR D1 EXT=IOOUNT GDI FILE+EXT FPS DSS CR "> THE CURRENT FILE FILE= EXT= SEC=SEC+DT TIME=SEC/60.0 EXT=TIME TIS=TTLSAM+EXT TIS=TIS+" MIN" CR EXT=OOUNT DSS D2 PDS FILE+EXT RAS ABS DSS DRAW IOOUNT=IOOUNT+1 OOUNT=OOUNT+l CHECK=STOP-IOOUNT FOOUNT=POOUNT-l IF POOUNT EQZ THEN CR " > » REPLACE THE DISK IN D1 <— " CR CR "> READY? PRESS ANY KEY TO MOUNT D1 !" PAU CR D1 FOOUNT=FSTOP CR "> CHECK TOE FILES IN DISK 1 TO BE TRANSFEREE : " CR "FSTOP " FSTOP ENDIF UNTIL CHECK LTZ CR CR "> FINISHED! THE LAST FILE FILE= EXT= CR " TOTAL RXN TIME TIME= " MIN." CR D1 END ASG FRTRSF 1

CRT ABTRSF (--- Absorbance transfer--- ) CR CR ">>> ABTRSF «<" CR CR "> INSERT THE EMITTANCE DISK " CR " IN D1 AND ABSORBANCE DISK IN D2 !" CR "> AFTER INSERT DISKS, PRESS ANY KEY TO MOUNT D1 & D2 !" PAU CR D2 FIL "SCRATCH" CR D1 FIL "SCRATCH" CR CR "> CHECK THE FOLLOWING PARAMETERS FOR ABSORBANCE-TRANSFER CR FILE="TIM111." CR "> CHECK TOE MAIN FILE NAME (#####.) CR "FILE " FILE GDB FILE+"BKG" DSB CR ICOUNT=0 CR "> CHECK TOE INITIAL EMITTANCE FILE NUMBER TO TRANSFER :" CR "IOOUNT " IOOUNT STOP=120 CR "> CHECK TOE TOTAL FILE NUMBER TO TRANSFER (110/120) :" CR "STOP " STOP COUNT=0 CR "> CHECK TOE STARTING ABSORBANCE FILE NUMER TO WRITE ON CR "COUNT " COUNT CR CR "> READY? PRESS ANY KEY TO START ABSORBANCE-TRANSFORRING ! PAU CR CR "> TRANSFERRING, PLEASE W A I T ...... " BEGIN 297

CR D1 EXT=IOOUNT GDS FILE+EXT DSS CR "> THE CURRENT FILE FILE= EXT= RAS ABS DSS PTS=9 MSD SMOD MOS DSS DRAW CR D2 EXT=OOUNT PDS FILE+EXT IOOUNT=IOOUNT+1 OOUNT=OOUNT+1 CHECK=STOP-IOOUNT UNTIL CHECK LTZ CR CR "> FINISHED! TOE LAST FILE :" FILE= EXT= CR D1 END ASG ABTRSF 2 APPENDIX A.3 Spectrum Analysis Program

REALDEL REALTEST REALCHK REAL XI REAL Y1 REAL X2 REAL Y2 REALXDEL REAL YDEL REALAA REALBB REAL X2941 REAL H2941 REAL X2273 REAL H2273 REAL X700 REAL H700 REALYSET REALTA REAL STDH1 REAL STDA1 REAL STDH2 REAL STDA2 REAL HI REAL A1 REAL H2 REAL A2 REAL H3 REAL A3 REALBASEH REAL BASEA REALSH REALSA REAL RH2941 REAL RA2941 REAL PKH2273 REAL PKA2273 REAL CVNH2273 REAL CVNA2273 REAL RH700 REAL RA700 INTEGER YN002 INTEGER MODE INTEGER REF INTEGER BSHA INTEGER STD 299

INTEGER ID INTEGER YNPN INTEGER COUNT INTEGER STOP INTEGER CHECK INTEGER DELAY STRING FILE

CRT FINDMAX (-- Find the max. point in a given section --- ) CR CR ">> FINDMAX « " DEL=LASER/NTP XCUR=XSP VALS PXA=XCUR PYA=YCUR BEGIN VALS TEST=YCUR-PYA IF TEST GEZ THEN PXA=XCUR PYA=YCUR ENDIF XCUR=XCUR-DEL CHK=XCUR-XEP UNTIL CHK LTZ END

CRT FINDMIN ( Find the min. point in a given section ) CR CR ">> FINDMIN « " DEL=LASER/NTP XCUR=XSP VALS PXA=XCUR PYA=YCUR BEGIN VALS TEST=YCUR-PYA IF TEST LEZ THEN PXA=XCUR PYA=YCUR ENDIF XCUR=XCUR-DEL CHK=XCUR-XEP UNTIL CHK LTZ END

CRT SETSEARCH (— Find a new setpoint for a negative peak — ) CR CR ">> SETSEARCH <<” CR DEL=LASER/NTP XCUR=XEP BEGIN VALS TEST=YCUR-YSET XCUR=XCUR+DEL CHK=XCUR-XSP IF CHK GTZ THEN TEST=-0.01 ENDIF UNTIL TEST LEZ XCUR=XCUR-DEL VALS PXA=XCUR PYA=YCUR END

CRT ALLOUTFUT LXF=600 FXF=4600 LYA=3.6 END

CRT PARTOUnvr LXF=800 FXF=1200 LYA=1.2 END

CRT CHBAND CR CR ">> CHBAND <<" CR CR "> CALCULATION OF CH-BAND 2941 XSP=3130 (3100) XEP=2800 (2900) FINDMAX X2941=PXA H2941=PYA XSP=3200 (3200) XEP=3050 (3050) FINDMIN X1=PXA Y1=PYA XSP=2800 (2900) XEP=2700 FINDMIN X2=PXA 301

Y2=PYA CR YDEL=Y2-Y1 XDEL=X2-X1 AA=YDEL/XDEL BB=AA*X1 BB-Y1-BB PCS=AA*X2941 FCS=PCS+BB H2941=H2941-PCS CR "> PEAK HEIGHT AT " X2941= " BAND IS " H2941= " !" CR FXF=X1 LXF=X2 SMS CR "> PEAK AREA FROM " Xl= " TO " X2= " IS " PCS= " !" ALLOUTPUT END

CRT NOOBAND CR CR ">> NOOBAND <<" CR CR "> CALCULATION OF NCO-BAND 2280 !" XSP=2320 (2290) XEP=2220 (2230) FINDMAX X2273=PXA H2273=PYA XSP=2200 XEP=2150 (1820) FINDMIN X2=PXA Y2=PYA (CR) YN002=1 (CR "> DOES THE NEGATIVE 002 PEAK EXIST ? (CR " <0> YES <1> NO") (CR "YNC02 " YN002) IF YN002 EQZ THEN YSET=Y2 XSP=2350 XEP=2280 SETSEARCH X1=PXA Y1=PYA ELSE XSP=2420 XEP=2350 FINDMIN 302

X1=PXA Y1=PYA ENDIF CR YDEL=Y2-Y1 XDEL=X2-X1 AA=YDEL/XDEL BB=AA*X1 BB=Y1-BB FCS=AA*X2273 PCS=PCS+BB H2273=H2273-FCS CR "> PEAK HEIGHT AT " X2273= " BAND IS " H2273= " !" FXF=X1 LXF=X2 SMS IF YNC02 EQZ THEN CR M0DE=0 CR "> SELECT A METHOD TO CALCULATE THE NOO PEAK AREA CR " <0> WHOLE AREA <1> SHOULDER AREA" CR "MODE " MODE IF MODE NEZ THEN XSP=2350 XEP=2260 FINDMIN LXF=PXA TA=FCS SMS FCS=TA-FCS ENDIF ENDIF CR CR 1"> PEAK AREA FROM " Xl= " TO " X2= " IS " FCS= " !" ALLOUTPUT END

CRT CH2R0CK CR CR ">> CH2R0CK <<" CR CR "> CALCULATION OF CH2-ROCK 700 !" XSP=714 XEP=690 FINDMAX X700=PXA H700=PYA XSP=725 XEP=714 FINDMIN 303

X1=PXA Y1=PYA XSP=690 XEP=680 (640) FINDMIN X2=PXA Y2=FYA CR YDEL=Y2-Y1 XDEL=X2-X1 AA=YDEL/XDEL BB=AA*X1 BB=Y1-BB FCS=AA*X700 PCS=FCS+BB H700=H700-PCS CR "> PEAK HEIC5HT AT " X700= " BAND IS " H700= ” !" CR FXF=X1 LXF=X2 SMS CR "> PEAK AREA FROM " Xl= " TO " X2= " IS " FCS= " !" ALLOUTPUT END

CRT PUACAL (--- Conversion = 1 - residure ---) CR CR ">>> PUACAL <<<" CR CR "> CALCULATION OF NOO CONVERSION !” CR CR "> INSERT TOE ABSORBANCE DISK IN D2 !" CR "> AFTER INSERT DISK, PRESS ANY KEY TO MOUNT D2 !" PAU CR D2 CR CR "> SET UP IRE FOLLOWING BASIC REFERENCE CONSTANTS CR " BAS EH BASEA " CR M> INPUT THE REFERENCE-BASED FILE (#####.###) CR "FILE (#####.) " FILE CR "REF (###) " REF EXT=REF GDS FILE+EXT DSS DRAW CHBAND STDH1=H2941 STDA1=PCS H1=H2941 A1=FCS NOOBAND H2=H2273 A2=FCS CH2R0CK STDH2=H700 STDA2=FCS H3=H700 A3=FCS CR BSHA=1 CR "> SELECT A CORRECTION METHOD :" CR " <0> BASED ON HEIGHT <1> BASED ON AREA ” CR "BSHA " BSHA CR STD=0 CR "> SELECT A INTERNAL STANDARD CR " <0> CH-BAND <1> CH2-ROCK " CR "STD " STD IF STD EQZ THEN BASEH=STDH1 BASEA=STDA1 ELSE BASEH=STDH2 BASEA=STDA2 ENDIF CR ID=0 CR ">CHECK THE CALCULATION STATUS CR " <0> START FROM REFERENCE (1st) FILE " CR " <1> RESTART FROM NON-REFERENCE (any) FILE " CR "ID " ID CR YNPN=0 CR "> SET THE PLOTTER ON ?" CR " <0> YES <1> NO " CR "YNPN " YNPN IF YNPN EQZ THEN XPN=1.6 CR "XPN " XPN YPN=7.0 CR "YPN " YPN PNSZ=96 IF ID EQZ THEN CR CR "> PLOTTER IS WRITING, PLEASE WAIT ...... PLOTON PEN PL! FILE= EXT= TE! DELAY=3400 BEGIN DELAY=DELAY-1 UNTIL DELAY EQZ YPN=YPN-0.15 PLOTON PEN PL! "EXT RH2941 RA2941 PKH2273 PKA2273" " CVNH2273 CVNA2273 RH700 RA700" TE! DELAY=5500 BEGIN DELAY=DELAY-1 UNTIL DELAY EQZ YPN=YPN-0.15 ENDIF ENDIF CR CR "> BEGIN TO CALCULATE THE CONVERSIONS OF SAMPLE FILES !" CR CR "> CHECK THE MAIN FILE NAME (#####.) CR "FILE " FILE CR OOUNT=EXT CR "> CHECK THE INITIAL FILE NUMBER TO CALCULATE : " COUNT STOP=120 CR "> CHECK THE FINAL FILE NUMBER TO CALCULATE : " STOP CR CR "> READY! PRESS ANY KEY TO START CALCULATION !" PAU CR CR "> CALCULATING, PLEASE W A I T ...... " BEGIN IF YPN LTZ THEN CR CR ">>> WARNING! PLOTTING PAPER IS RUN OUT OF ?!?" CR CR "> REPLACE A NEW PAPER, PRESS ANY KEY TO CONTINUE PAU YPN=7.0 PLOTON PEN PL! FILE= TE! DELAY=3000 BEGIN DELAY=DELAY-1 UNTIL DELAY EQZ YPN=YPN-0.15 ENDIF EXT=OOUNT GDS FILE+EXT DSS IF STD EQZ THEN CHBAND CH=H2941 306

SA=FCS ELSE CH2ROCK SH=H700 SA=PCS ENDIF IF BSHA EQZ THEN FCS=BASEH/SH CR CR "> HEIGHT INTEGRATING FACTOR = " FCS= ELSE FCS=BASEA/SA CR CR "> AREA INTEGRATING FACTOR = " PCS= ENDIF MUS DSS DRAW CHBAND H2941=H2941/H1 FCS=FCS/A1 RH2941=H2941 RA2941=FCS CR CR "> HEIGHT RATIO = " RH2941= ” <~" CR "> AREA RATIO = " RA2941= " <— " NOOBAND FKH2273=H2273 PKA2273=FCS H2273=H2273/H2 FCS=FCS/A2 H2273=l.0-H2273 FCS=1.0-FCS CVNH2273=H2273 CVNA2273=PCS CR CR "> HEIGHT CONVERSION = " CVNH2273= " CR ”> AREA CONVERSION = " CVNA2273= " CH2ROCK H700=H700/H3 PCS=PCS/A3 RH700=H700 RA700=PCS CR CR "> HEIGHT RATIO = " RH700= " <— " CR "> AREA RATIO = " RA700= " <— " CR CR "> TOE CURRENT FILE : " FILE= EXT= IF YNPN EQZ THEN CR CR "> PLOTTER IS WRITING, PLEASE WAIT . . . PLOTON 307

PEN PL! EXT= " " RH2941= " " RA2941= " " PKH2273= " " PKA2273= " " CVNH2273= " " CVNA2273= " " RH700= " " RA700= " " TE! DELAY=3500 BEGIN DELAY=DELAY-1 UNTIL DELAY EQZ YPN=YPN-0.15 ENDIF OOUNT=OOUNT+l CHECK=STOP-COUNT UNTIL CHECK LTZ CR CR "> FINISHED! THE LAST FILE : " FILE= EXT= CR D1 END ASG PUACAL 1 APPENDIX B

ACSL Programs for Polyurea Reaction

Reaction Kinetics

Reaction Kinetics and Rheological Changes APPENDIX B.l ACSL Program for Polyurea Reaction (Kinetics)

INTEGER INTV INTEGER MOOT INTEGER NM INTEGER NO CONSTANT A2=8.5E6 CONSTANT E2=4709.0 CONSTANT NN2=2.33 CONSTANT R=1.9872 CONSTANT C20=5.6398E-4 CONSTANT TP=320.15 CONSTANT 10=3 CONSTANT TSTP=30.0 CONSTANT INTV=10

INITIAL NM=0 NO=0 AFA2IC=0.0 VARIABLE TIME=0.0 CINTERVAL CINT=0.01 NSTEPS NSTP=1 NODT=INT(AINT(TSTP/CINT/INTV)) WRITE(10,100) NODT 100..P0RMAT(1X,I4) END $" OF INITIAL"

DYNAMIC DERIVATIVE PROCEDURAL AFA2D=A2*EXP(-E2/R/TP)*(C20**(NN2-1.0))*((1.0-AFA2)**NN2) AFA2=INTEG (AFA2D, AFA2IC) END $" OF PROCEDURAL" TERMT (TIME. GE. TSTP) END $" OF DERIVATIVE" IF ((NM.EQ.O).OR.(NM.BQ.INTV)) GO TO 300 GO TO 400 300..CONTINUE NO=NO+l WRITE(10,200) NO,TIME,AFA2 200..FORMAT(IX,14,2(2X,El2.5)) NM=0 400..CONTINUE NM=NM+1 END $" OF DYNAMIC"

END $" OF PROGRAM" 310

APPENDIX B.2 ACSL Program for Polyurea Reaction (Kinetics k Rheology)

INTEGER INTV INTEGER MODT INTEGER NM INTEGER NO CONSTANT Al=3.14E6 CONSTANT El=1600.0 CONSTANT NN1=2.10 CONSTANT A2=8.5E6 CONSTANT E2=4709.0 CONSTANT NN2=2.33 CONSTANT DN=1.118 CONSTANT CP=0.4 CONSTANT HRXN=15374. CONSTANT R=1.9872 CONSTANT AETA=4.6601E3 CONSTANT EETA=2074.0 CONSTANT AA=0.734 CONSTANT BB=2626.0 CONSTANT CC=-8.332E5 CONSTANT APHAG=0.580 CONSTANT C0=2.1321E-3 CONSTANT D1=0.130 CONSTANT D2=0.870 CONSTANT TP0=323.15 CONSTANT 10=1 CONSTANT TSTP=20.0 CONSTANT INTV=5

INITIAL NM=0 N0=0 TPIC=TP0 AFA1IC=0.0 AFA2IC=0.0 HR1=C0*D1*HRXN/DN/CP HR2=C0*D2*HRXN/DN/CP VARIABLE TIME=0.0 CINTERVAL CINT=0.01 NSTEPS NSTP=1 NODT=INT(AINT(TSTP/CINT/INTV)) WRITE(IO, 100) NODT 100..F0RMAT(1X,I4) END $" OF INITIAL"

DYNAMIC DERIVATIVE PROCEDURAL AFA1D=A1*EXP(-E1/R/TP)*(((C0*D1)**(NN1-1.0))/D1 ... *((1.0-AFA1)**(NN1-1.0))*(1.0-Dl*AFAl-D2*AFA2) AFA2D=A2*EXP(-E2/R/TP)*(((C0*D2)**(NN2-1.0))/D2 ... *((1.0-AFA2)**(NN2-1.0))*(1.0-Dl*AFAl-D2*AFA2) AFA1=INTEG(AFA1D,AFA1IC) AFA2=INTEG (AFA2D, AFA2IC) AFA=D1*AFA1+D2*AFA2 END $" OF PROCEDURAL" PROCEDURAL TPD=HR1*AFA1D+HR2*AFA2D TP= INTEG (TPD, TPIC) END $" OF PROCEDURAL" PROCEDURAL IF (AFA.GE.APHAG) GO TO 200 AFAP=AFA/APHAG/TP EP=AA+BB*AFAP+OC* (AFAP*AFAP) VISC=AETA*EXP(EETA/R/TP)*((APHAG/(APHAG-AFA))**EP) GO TO 300 200..CONTINUE VISC=-0.1 300..CONTINUE END $" OF PROCEDURAL" TERWT (TIME. GE. TSTP) END $" OF DERIVATIVE" IF ((NM.EQ.O).OR.(NM.EQ.INTV)) GO TO 400 GO TO 600 400..CONTINUE NO=NO+l WRITE(10,500) NO,TIME,AFA,TP,VISC 500..FORMAT(1X ,I4,4(2X,E12.5)) NM=0 600..CONTINUE NM=NM+1 END $" OF DYNAMIC"

END $" OF PROGRAM" APPENDIX C

Simplified mechanical model for 120 °C-transfer molded IPN (PU/PES=50/50)

312 APPENDIX C

Hie structure given in Figure 5.19 can be described by a 4- parameter mechanical model (i.e., Vg^ or V^, V^,, Xj or 4>j, Xg or shown in Figure C.l based on Takayanagi’s model II (Takayanagi et al.,

1964). Although the data from dynamic mechanical measurements may be better fitted using this model, it is difficult to determine the parameters independently. In this study, we applied a simplified mechanical model (Nelson, 1977; Rosovizky et al., 1979) by assuming that 4j=\2 =l» i.e., not considering the distribution state of the dispersed phase. The model is shown in Figure C.2 with only 2 independent parameters, Vg^ or V ^ , and V^,. The values of these parameters can be easily determined from the morphology shown in the

TEM micrograph and the best fit of the G ’ curve. \ A1 T c 1-0t I 1- 0 2

0 1

02 II

Figure C.l Mechanical model based on Takayanagi's model II for 120°C-transfer molded IPN (RJ/PES-50/50) 1- 4*2 1 1-^2

4*21 1

Figure C.2 Simplified mechanical model for 120 °C-transfer molded IPN (FU/PES=50/50) 315