THE DEVULCANIZATION OF UNFILLED AND CARBON BLACK FILLED
ISOPRENE RUBBER VULCANIZATES BY HIGH POWER ULTRASOUND
A Dissertation
Presented to
The Graduate Faculty of the University of Akron
In Partial Fulfillment
of the Requirements for the Degree
Doctor of Philosophy
Ximei Sun
May, 2007
THE DEVULCANIZATION OF UNFILLED AND CARBON BLACK FILLED
ISOPRENE RUBBER VULCANIZATES BY HIGH POWER ULTRASOUND
Ximei Sun
Dissertation
Approved: Accepted:
Advisor Department Chair Dr. Avraam I. Isayev Dr. Sadhan C. Jana
Committee Member Interim Dean of the College Dr. Ernst D. von Meerwall Dr. George R. Newkome
Committee Member Dean of the Graduate School Dr. Sadhan C. Jana Dr. George R. Newkome
Committee Member Date Dr. Erol Sancaktar
Committee Member Dr. Michael Cheung
ii
ABSTRACT
The effects of ultrasound on virgin gum isoprene rubber (IR) and on the devulcanization of unfilled and carbon black (CB) filled IR were studied. Ultrasonic treatment altered the structure and properties of gum IR by creating low molecular weight tails which broadened the molecular weight distribution and improved processability.
Ultrasonic devulcanization of IR vulcanizates resulted in a reduction of gel fraction and crosslink density. Increasing the ultrasonic amplitude yielded a further reduction, regardless of CB loading, in the IR vulcanizates. This is contrary to the previous work on natural rubber (NR), the natural counterpart of IR which showed a minimum gel fraction and crosslink density at an intermediate ultrasonic amplitude.
The devulcanization of filled IR resulted in more main chain scission than in unfilled IR due to the immobility of bound rubber at the filler surface which leads to lower properties in revulcanized rubbers than in virgin rubber. Upon blending the devulcanized IR with virgin IR, properties comparable to those of virgin rubber were obtained at certain blending ratios.
A cure kinetics model with reversion adequately predicted the evolution of state of cure in curing and reversion stages under isothermal and non-isothermal conditions.
The higher reversion observed in filled IR than in unfilled IR was consistent with the difference of reversion rate constant obtained in simulation.
iii NMR proton transverse relaxation technique was unable to differentiate the contribution of short component mobility between physically entangled (heavy sol) and chemically crosslinked (gel) networks. Ultrasound severed both the chemical crosslinks and the main chain, creating dangling chain ends, with no generation of additional fragments of oligomeric species.
Simulation of network structures using the Dobson-Gordon theory of network statistics indicated crosslinks were easier to break than main chains under ultrasonic exposure. Unfilled IR and NR had similar rate constant ratios of main chain scission and crosslink rupture. An increase of CB loading increased this ratio for both IR and NR with higher ratio in IR. The addition of processing oil in the filled IR compounds reduced this ratio.
iv
ACKNOWLEDGEMENTS
I would like to express my sincere gratitude to many people without whom this work could have never been accomplished.
First of all, I would like to thank Dr. Avraam I. Isayev for being my mentor for the past years. I am very grateful to work under his thorough guidance and consistent encouragement. I would also like to thank Dr. Ernst von Meerwall not just for his serving on my committee but more for his inspiring discussions on NMR work with me. My sincere appreciation extends to the remaining committee members: Dr. Sadhan C. Jana,
Dr. Erol Sancaktar and Dr. Michael Cheung for their thoughtful advice and suggestions.
I cherish the tremendous help from Theresa Schillig, Tayba Tahir and Marcile
Pendleton of the Akron Polymer Training Center (APTC, Univ. of Akron) who made my life and work in Akron a whole lot easier. I appreciate the help provided to me by Bob
Seiple (Polymer Science Dept.) and Henry Pawlowski (Alpha Technologies ) in using the APA 2000 and by Jon Page (Polymer Science Dept.) for GPC experiments. Many thanks to Mr. Tirtha Joshi (Physics Dept.) for the NMR experiments we did together.
My earnest gratitude goes to my parents for giving me the wisdom and devotion for work. Especially I want to thank my husband, Guofeng Huang, and my son, Roy
Huang, for their love and keeping me motivated to try my best all the time.
Last but not the least, thanks go to the Goodyear Tire and Rubber Company and the Akrochem Corporation for the material support in this research work.
v
TABLE OF CONTENTS
Page
LIST OF TABLES ……………………………………………………………………...xii
LIST OF FIGURES …………………………………………………………………….xiv
CHAPTER
I. INTRODUCTION...... 1
II. LITERATURE SURVEY AND BACKGROUND ...... 4
2.1 Natural Rubber (NR)...... 4
2.2 Synthetic Isoprene Rubber (IR) ...... 7
2.3 Comparison of NR and IR...... 10
2.4 Carbon Black (CB)...... 11
2.5 Carbon Black Filled Rubber...... 14
2.6 Vulcanization ...... 16
2.6.1 Sulfur Vulcanization ...... 17
2.6.2 Mechanism of Accelerated Sulfur Vulcanization ...... 22
2.7 Rubber Recycling...... 26
2.7.1 Landfills and Waste Rubber Utilization...... 27
2.7.2 Grinding Methods...... 27
2.7.3 Rubber as a Fuel Source and Pyrolysis ...... 28
2.7.4 Chemical Method...... 29
2.7.5 Microwave Method ...... 30
vi 2.7.6 Biotechnological Method ...... 30
2.7.7 Ultrasonic Method...... 32
2.7.8 Summary of the Recycling Methods...... 33
2.8 Recycling of Isoprene Rubber – Current Studies...... 34
2.9 Application of Ultrasound in Polymers...... 36
2.9.1 Ultrasonic Treatment of Polymers ...... 39
2.9.2 Ultrasonic Devulcanization Mechanism ...... 40
2.9.3 Modeling of Ultrasonic Devulcanization Process...... 41
2.10 Molecular Mobility Analysis by Solid-State NMR...... 45
1 2.10.1 Proton Transverse Relaxation ( H T2)...... 46
2.10.2 Pulsed Gradient Spin-Echo (PGSE) Diffusion...... 47
III. EXPERIMENTAL ...... 50
3.1 Materials...... 50
3.2 Compounding...... 52
3.3 Vulcanization and Vulcanizates Grinding ...... 54
3.4 Revulcanization...... 55
3.5 Ultrasonic Treatment and Devulcanization Equipment ...... 55
3.5.1 Ultrasonic Treatment of the Virgin Gum IR ...... 58
3.5.2 Devulcanization of the Vulcanizates...... 59
3.6 Characterization Methods...... 59
3.6.1 Vulcanization Kinetics...... 60
3.6.2 Dynamic Properties...... 60
3.6.3 Gel Fraction...... 62
3.6.4 Crosslink Density...... 62
vii 3.6.5 Mechanical Properties...... 67
3.6.6 Molecular Weight Determination...... 68
3.6.7 Thermal Properties...... 69
3.6.8 T2 Relaxation and PGSE Diffusion...... 69
IV. CURE KINETICS STUDY OF THE UNFILLED AND THE CB FILLED IR..... 71
4.1 General ...... 71
4.2 Cure Kinetics Experiments ...... 73
4.3 General Kinetic Model Equations...... 74
4.3.1 Isothermal Cure Kinetics...... 76
4.3.2 Non-isothermal Cure Kinetics...... 78
4.3.3 The Modeling Procedure for Cure Kinetics Study...... 79
4.4 Isothermal Cure Kinetics of the Unfilled IR ...... 79
4.4.1 Curing and the determination of experimental induction time ...... 79
4.4.2 Determination of the experimental state of cure α...... 82
4.4.3 Modeling of the state of cure α at the individual temperatures ...... 84
4.4.4 The simultaneous modeling of the isothermal state of cure α ...... 84
4.5 Non-isothermal Cure Kinetics of the Unfilled IR ...... 87
4.5.1 Determination of the non-isothermal induction time tI ...... 88
4.5.2 Non-isothermal modeling of the state of cure α ...... 89
4.6 Isothermal Cure Kinetics of the CB Filled IR...... 92
4.7 Non-isothermal Cure Kinetics of the CB Filled IR...... 98
4.8 Conclusions...... 102
V. STRUCTURE AND PROPERTIES OF THE ULTRASONICALLY TREATED GUM ISOPRENE RUBBER ...... 104
5.1 General ...... 104 viii 5.2 Preparation of the Samples...... 106
5.3 Ultrasonic Treatment: Die Pressure and Power Consumption...... 107
5.4 Curing...... 108
5.5 Gel Fraction and Crosslink Density ...... 110
5.6 Molecular Characteristics of the Ultrasonically Treated Gums...... 110
5.7 Rheological Properties ...... 113
5.8 Mechanical Properties...... 119
5.9 Thermal Properties...... 122
5.10 Conclusions...... 125
VI. ULTRASONIC DEVULCANIZATION OF THE UNFILLED IR...... 127
6.1 General ...... 127
6.2 Experimental ...... 128
6.3 Curing and Revulcanization...... 129
6.4 Power Consumption and Die Pressure...... 132
6.5 Gel Fraction and Crosslink Density ...... 135
6.6 Molecular Characteristics of the Devulcanized Sol...... 139
6.7 Rheological Properties ...... 141
6.8 Mechanical Properties...... 143
6.9 Thermal Properties...... 148
6.10 Conclusions...... 150
VII. ULTRASONIC DEVULCANIZATION OF THE CB FILLED IR ...... 152
7.1 General ...... 152
7.2 Experimental ...... 153
7.3 Vulcanization of the Virgin Filled IR without the Processing Oil...... 155
ix 7.3.1 Curing...... 155
7.3.2 Gel Fraction and Crosslink Density ...... 159
7.3.3 Mechanical Properties...... 160
7.4 Devulcanization of the Filled IR without the Processing Oil ...... 163
7.4.1 Power Consumption and Die Pressure...... 163
7.4.2 Gel Fraction and Crosslink Density ...... 165
7.4.3 Rheological Properties ...... 167
7.4.4 Revulcanization...... 171
7.4.5 Mechanical Properties...... 172
7.5 Devulcanization of the Filled IR Containing the Processing Oil...... 174
7.5.1 Power Consumption and Die Pressure...... 178
7.5.2 Gel Fraction and Crosslink Density ...... 179
7.5.3 Rheological Properties ...... 181
7.5.4 Revulcanization...... 183
7.5.5 Mechanical Properties...... 186
7.6 Effect of Retarder on Vulcanization and Revulcanization of Filled IR ...... 188
7.7 Blending of the Devulcanized IR (dIR) with the Virgin IR...... 191
7.8 Conclusions...... 197
VIII. MOLECULAR MOBILITY OF ULTRASONICALLY TREATED GUM IR, UNFILLED AND CB FILLED IR...... 199
8.1 General ...... 199
8.2 Preparation of the Samples...... 200
8.3 NMR Experiments ...... 201
8.3.1 Proton T2 Data Analysis...... 203
8.3.2 PGSE Diffusion Data Analysis ...... 204 x 8.4 Molecular Mobility of Treated Gum IR and their Vulcanizates...... 205
8.4.1 Molecular Weight and Glass Transition Temperature ...... 205
8.4.2 NMR Proton Relaxation...... 206
8.4.3 NMR Diffusion...... 212
8.5 Molecular Mobility of the Unfilled IR...... 215
8.5.1 Sol Fraction and Glass Transition Temperature...... 215
8.5.2 NMR Proton Relaxation...... 216
8.6 Molecular Mobility of the CB Filled IR ...... 222
8.7 Conclusions...... 231
IX. SIMULATION OF THE NETWORK STRUCTURES FOR THE DEVULCANIZED UNFILLED AND FILLED IR...... 235
9.1 General ...... 235
9.2 Development of the Modeling Equations ...... 236
9.3 The Unfilled IR ...... 239
9.4 The CB Filled IR...... 242
9.5 Conclusions...... 245
X. SUMMARY ...... 247
REFERENCES...... 254
xi
LIST OF TABLES
Table Page
2.1 Different polymerization processes lead to different products ...... 12
2.2 Accelerator type comparison79 ...... 18
2.3 Several common accelerators used in sulfur vulcanization81...... 19
2.4 Sulfur and accelerator of different vulcanization systems for NR82 ...... 20
2.5 Typical NR vulcanizate structures and properties at optimum cure time82... 21
3.1 Materials used in this study...... 51
3.2 Compounding recipes for the vulcanization of IR ...... 52
3.3 Compounding recipe for the revulcanization of the devulcanized IR...... 56
4.1 Induction time and cure kinetic constants obtained from individual isothermal fitting of the unfilled IR...... 86
4.2 Cure kinetic constants obtained from simultaneous isothermal fitting of the unfilled IR...... 87
4.3 Heating rates and step temperature profiles in nonisothermal curing of the unfilled IR ...... 88
4.4 Induction time and cure kinetic constants obtained from individual isothermal modeling of the 35phr CB filled IR...... 96
4.5 Cure kinetic constants obtained from simultaneous isothermal fitting of the 35phr CB filled IR...... 98
4.6 Heating rates and step temperature profiles in nonisothermal curing of the 35phr CB filled IR...... 99
5.1 Rheological parameters of the modified Cross model for the virgin and ultrasonically treated IR gums...... 119
xii 5.2 Tg of the virgin and the ultrasonically treated IR gums and their vulcanizates ...... 124
6.1 Tg of the virgin, cured and devulcanized IRs...... 149
8.1 The samples used in the NMR relaxation analysis of the filled IR...... 202
8.2 Molecular weight and glass transition temperature of the virgin and the ultrasonically treated IR gum ...... 206
8.3 Sol fraction and glass transition temperature of the virgin gum, vulcanized, and devulcanized IR at the flow rate of 0.63 g/s...... 216
9.1 Physical and chemical parameters of the unfilled IR used in the simulation...... 240
ξ (g) 9.2 Initial gel fraction g 0 and crosslink density 0 of the filled IR used in the simulation ...... 242
xiii
LIST OF FIGURES
Figure Page
2.1 Isoprene monomer, Natural rubber (cis-1, 4-addition) and Gutta-percha rubber (trans-1, 4-addition) ...... 5
2.2 Schematic drawing for the production of cis-1, 4 polyisoprene38...... 9
2.3 Structures formed during sulfur vulcanization of elastomers77...... 20
2.4 Typical cure curves for accelerated sulfur vulcanization...... 22
2.5 Possible formation of zinc complex with amine ligands under CBS accelerated sulfur vulcanization (NR1 is the amine ligand, the amine moiety of a sulfenamide accelerator) 87...... 24
2.6 Structural features of sulfur-vulcanized NR92...... 25
2.7 Ultrasound cavitation bubble growth and collapse140 ...... 38
2.8 Effect of diffusion on the Hahn spin echo179...... 48
2.9 The pulsed-gradient spin-echo (PGSE) sequence ...... 49
3.1 Schematic drawing of the coaxial ultrasonic reactor ...... 57
3.2 Schematic drawing of the effective volume of the die opening...... 58
3.3 Biconical die sample cavity185...... 60
3.4 Soxhlet extraction apparatus ...... 63
3.5 Determination of the evaporation time for the ground unfilled rubber...... 66
4.1 Sulfur reaction scheme (s: sulfur; c1 and c2: stable and unstable sulfur- containing crosslink, respectively; p: product of reversion reaction) ...... 74
4.2 The modeling procedure for the cure kinetics study...... 80
4.3 Isothermal cure curves of the unfilled IR (strain amplitude: 4.2%, frequency: 100 cpm)...... 81
xiv 4.4 Determination of the experimental induction time ...... 82
Γ − Γ 4.5 Finding the max,0 min,0 by Equation 4.20 for unfilled IR...... 83
4.6 Kinetic constants and induction time versus the reciprocal temperature for the unfilled IR...... 85
4.7 Isothermal state of cure for the unfilled IR: experimental (symbols) and fit (lines)...... 86
4.8 Non-isothermal induction time for the unfilled IR: experiments and predictions. Step cure refers to the step temperature profile in Table 4.3..... 89
4.9 Non-isothermal state of cure for the unfilled IR: experiments (starting from 80oC) and predictions ...... 90
4.10 Non-isothermal state of cure for the unfilled IR: experiments (starting from 120oC) and predictions ...... 91
4.11 Non-isothermal state of cure for the unfilled IR: experiments (step temperature profile) and predictions ...... 92
4.12 Isothermal cure curves of the 35 phr CB filled IR (strain amplitude: 4.2%, frequency: 100 cpm)...... 94
Γ − Γ 4.13 Finding max,0 min,0 by Equation 4.20 for 35phr CB filled IR...... 95
4.14 Kinetic constants and induction time versus the reciprocal temperature for the 35phr CB filled IR...... 96
4.15 Isothermal state of cure for the 35phr CB filled IR: experiments (symbols) and fits (lines)...... 97
4.16 Non-isothermal induction time for the 35phr CB filled IR: experiments and predictions. Step cure refers to the step temperature profile in Table 4.6...... 99
4.17 Non-isothermal state of cure for the 35phr CB filled IR: experiments (starting from 80oC) and predictions...... 100
4.18 Non-isothermal state of cure for the 35phr CB filled IR: experiments (starting from 120oC) and predictions...... 101
4.19 Non-isothermal state of cure for the 35phr CB filled IR: experiments (step temperature profile) and predictions ...... 102 xv 5.1 Die pressure and power consumption as a function of the ultrasonic amplitude during the ultrasonic treatment of gum IR ...... 108
5.2 Cure curves of the virgin and the ultrasonically treated IR gums at 160oC, a strain amplitude of 4.2% and a frequency of 100 cpm...... 109
5.3 Gel fraction and crosslink density as a function of the ultrasonic amplitude for the vulcanizates of virgin (symbols shown on the ordinate axes) and ultrasonically treated IR gums ...... 111
5.4 Molecular weight distribution of the virgin and ultrasonically treated IR gums at various amplitudes (a), and amplitude dependence of the number (Mn), weight (Mw) average molecular weight and polydispersity (Mw/Mn) of the virgin and treated IR gums (b) ...... 112
5.5 Complex viscosity versus frequency for the virgin and ultrasonically treated IR gums and their vulcanizates at 120oC and a strain amplitude of 4.2%...... 114
5.6 Loss tangent versus frequency for the virgin and ultrasonically treated IR gums and their vulcanizates at 120oC and a strain amplitude of 4.2% ...... 115
5.7 Storage (a) and loss (b) modulus of the virgin and ultrasonically treated IR and their vulcanizates as a function of frequency at 120oC and a strain amplitude of 4.2% ...... 116
5.8 Complex viscosity versus frequency for the virgin IR and the IR ultrasonically treated at the amplitude of 5, 7.5 and 10 µm at various temperatures (symbols: experiments; curves: the modified Cross model fittings) ...... 118
5.9 Stress-strain curves for the vulcanizates of the virgin and ultrasonically treated IRs ...... 120
5.10 Amplitude dependence of the tensile strength, elongation at break (a) and modulus at 100% and 300% (b) of the virgin (symbols shown on the ordinate axes) and ultrasonically treated IR vulcanizates ...... 121
5.11 TGA curves for the virgin and ultrasonically treated IR gums and their vulcanizates under the nitrogen atmosphere ...... 123
5.12 DSC curves for the virgin and ultrasonically treated IR and their vulcanizates under the nitrogen atmosphere ...... 124
xvi 6.1 Cure curves at 160oC for IR (a) (APA) and NR (b)181 (ODR) of the virgin rubber and devulcanized rubbers obtained at various ultrasonic amplitudes with a die gap of 2.54 mm, a flow rate of 0.63 g/s and a barrel temperature of 120oC ...... 130
6.2 Power consumption for IR (a) and NR (b)181 rubbers devulcanized at various flow rates at a barrel temperature of 120oC and a die gap of 2.54 mm...... 133
6.3 Die pressure for IR (a) and NR (b)181 rubbers devulcanized at various flow rates at a barrel temperature of 120oC and a die gap of 2.54 mm...... 134
6.4 Gel fraction of the devulcanized (solid symbols) and revulcanized (open rectangle, for flow rate of 0.63g/s only) IR (a) and NR (b)181 rubbers as a function of ultrasonic amplitude obtained at various flow rates, a die gap of 2.54 mm and a barrel temperature of 120oC...... 136
6.5 Crosslink density of the devulcanized (solid symbols) and revulcanized (open rectangle, for flow rate of 0.63g/s only) IR (a) and NR (b)181 rubbers as a function of ultrasonic amplitude obtained at various flow rates, a die gap of 2.54 mm and a barrel temperature of 120oC...... 137
6.6 Molecular weight distribution at various amplitudes (a), amplitude dependence of weight average molecular weight, Mw, and polydispersity, Mw/Mn (b) of the sol parts in devulcanized IR obtained at a gap of 2.54 o mm, a flow rate of 0.63 g/s and a barrel temperature of 120 C. Mw and Mw/Mn of the virgin IR are also shown...... 140
6.7 The complex viscosity |η*| and tan δ of virgin IR, the vulcanizates and the devulcanizates obtained at a gap of 2.54 mm, flow rate of 0.63 g/s and a barrel temperature of 120oC as a function of frequency ω (a) and |η*|ω (b), respectively, at a strain amplitude of 4.2%...... 144
6.8 Storage (a) and loss (b) modulus of the virgin IR, the vulcanizates and the devulcanizates obtained at a gap of 2.54 mm, a flow rate of 0.63 g/s and a barrel temperature of 120oC as a function of frequency ω at 120oC at a strain amplitude of 4.2%...... 145
6.9 The stress-strain curve for IR and NR vulcanizates and revulcanizates obtained at different ultrasonic amplitudes, a die gap of 2.54 mm, a flow rate of 0.63 g/s and a barrel temperature of 120oC...... 146
6.10 Amplitude dependence of the tensile strength (a), elongation at break (b) and modulus at 100% and 300% (c) of IR and NR vulcanizates and revulcanizates at a die gap of 2.54 mm, a flow rate of 0.63 g/s and a barrel temperature of 120oC ...... 147 xvii 6.11 DSC curves for the virgin, cured and devulcanized IR under the nitrogen atmosphere ...... 149
7.1 Cure kinetics of the filled IR at various CB loadings at 160oC (strain amplitude: 4.2%, frequency: 100 cpm) ...... 156
7.2 Isothermal induction time versus the CB loading at 160oC ...... 157
7.3 Complex viscosity versus the frequency for the virgin IR and IR processed in the Banbury mixer (strain amplitude: 4.2%)...... 158
7.4 Cure curves of various IRs at 160oC (strain amplitude: 4.2%, frequency: 100 cpm)...... 159
7.5 Gel fraction and crosslink density of virgin IR vulcanizates versus carbon black loading ...... 160
7.6 Stress-strain curves of the virgin IR vulcanizates ...... 161
7.7 Tensile strength σB, elongation at break εB (a) and moduli at 100 and 300% strain, E100 and E300 (b) as a function of carbon black loading for the virgin IR vulcanizates...... 162
7.8 Power consumption as a function of ultrasonic amplitude during the devulcanization of the filled IR at various CB loadings ...... 164
7.9 Die pressure as a function of ultrasonic amplitude during the devulcanization of filled IR at various CB loadings ...... 165
7.10 Gel fraction of the devulcanized (open symbols) and revulcanized (solid triangle, for 35 phr only) filled IR as a function of ultrasonic amplitude obtained at flow rate of 0.63 g/s, a die gap of 2.54 mm and a barrel temperature of 120oC ...... 166
7.11 Crosslink density of the devulcanized (open symbols) and revulcanized (solid triangle, for 35 phr only) filled IR as a function of ultrasonic amplitude obtained at flow rate of 0.63 g/s, a die gap of 2.54 mm and a barrel temperature of 120oC ...... 167
7.12 Complex viscosity |η*| and tan δ of virgin 35 phr filled IR, vulcanizates and devulcanizates obtained at a gap of 2.54 mm, a flow rate of 0.63 g/s and a barrel temperature of 120oC as a function of frequency ω (a) and |η*|ω (b), respectively, at a strain amplitude of 4.2%...... 169
xviii 7.13 Storage (a) and loss (b) modulus of the virgin 35 phr filled IR, vulcanizate and devulcanizates obtained at a gap of 2.54 mm, a flow rate of 0.63 g/s and a barrel temperature of 120oC as a function of frequency ω at a strain amplitude of 4.2%...... 170
7.14 Revulcanization curves at 160oC for the devulcanized 35 phr CB filled IR obtained at three ultrasonic amplitudes...... 172
7.15 Stress-strain curves of the virgin cured and the revulcanized 35 phr CB filled IR obtained at three ultrasonic amplitudes ...... 173
7.16 Vulcanization of the unfilled and the 35 phr CB filled IR containing 0 and 10 phr processing oil at 160oC, a strain amplitude of 4.2% and a frequency of 100 cpm...... 175
7.17 Stress-strain curves of the unfilled and the 35 phr CB filled IR vulcanizates containing 0 and 10 phr processing oil...... 176
7.18 Tensile strength σB, elongation at break εB (a) and moduli at 100 and 300% strain E100, E300 (b) of the unfilled and the 35 phr CB filled IR vulcanizates containing 0 and 10 phr processing oil...... 177
7.19 Power consumption as a function of ultrasonic amplitude for the 35 CB filled IR containing 0 and 10 phr processing oil ...... 178
7.20 Die pressure as a function of ultrasonic amplitude for the 35 CB filled IR containing 0 and 10 phr processing oil...... 179
7.21 Gel fraction and crosslink density of devulcanized (open symbols) and revulcanized (solid symbols) 35 phr CB filled IR containing 0 and 10 phr processing oil as a function of ultrasonic amplitude obtained at flow rate of 0.63 g/s, a die gap of 2.54 mm and a barrel temperature of 120oC...... 180
7.22 Complex viscosity (a) and loss tangent (b) of the virgin 35 phr CB filled (containing 0 and 10 phr processing oil) uncured, the cured and the devulcanized IR as a function of frequency at 120oC, a strain amplitude of 4.2% ...... 182
7.23 Storage (a) and loss (b) modulus of the virgin 35 phr CB filled (containing 0 and 10 phr processing oil) uncured, the cured and the devulcanized IR as a function of frequency at 120oC, a strain amplitude of 4.2% ...... 184
7.24 Revulcanization curves of the devulcanized 35 phr CB filled IR (containing 0 and 10 phr processing oil) at 160oC, a strain amplitude of 4.2% and a frequency of 100 cpm...... 185 xix 7.25 Tensile strength σB, elongation at break εB (a) and modulus (b) at 100% strain E100 (c) as a function of ultrasonic amplitude for the revulcanizates of 35 phr CB filled IR containing 0 and 10 phr processing oil ...... 187
7.26 Cure curves of the virgin and devulcanized 35 phr CB filled IR containing 10 phr oil with and without the retarder SAFE at 160oC, a strain amplitude of 4.2% and a frequency of 100 cpm...... 189
7.27 Tensile strength σB, elongation at break εB (a) and modulus at 100% strain E100 (b) of the revulcanized 35 phr CB filled IR containing 10 phr processing oil without and with the retarder SAFE as a function of ultrasonic amplitude ...... 190
7.28 Cure curves for blends of devulcanized (dIR) and virgin IR containing 35 phr CB and 10 phr oil using the recipe in Table 3.2 at 160oC. (Curatives were added with respect to the total rubber content. IR was devulcanized at an amplitude of 10 µm) ...... 193
7.29 Cure curves for blends of ground IR vulcanizates (gIR) and virgin IR containing 35 phr CB and 10 phr oil using the recipe in Table 3.2 at 160oC. (Curatives were added with respect to the total rubber content)..... 194
7.30 Tensile strength σB (a), elongation at break εB and modulus at 100% strain E100 (b) of recycled and virgin IR blends (gIR/IR and dIR/IR) as a function of virgin IR concentration...... 196
1 o 8.1 Spectral decomposition of the H T2 relaxation decay at 70.5 C for the vulcanizate of IR ultrasonically treated at the amplitude of 5µm ...... 207
8.2 Transverse 1H relaxation decay at 70.5oC for the vulcanized IR in which the gum was ultrasonically treated at the amplitude of 5 µm fitted with the three-component model containing the Weibull exponent...... 208
8.3 Proton T2 for the gum IRs ultrasonically treated at different amplitudes ... 209
8.4 Proton T2 for the vulcanizates of ultrasonically treated gum IR at different amplitudes...... 210
8.5 Proton T2 fractions for the gum IR ultrasonically treated at different amplitudes ...... 211
8.6 Proton T2 fractions for the vulcanizates of ultrasonically treated gum IR at different amplitudes...... 211
8.7 Diffusion spin-echo attenuation of the gum IR extruded at the flow rate of 0.63 g/s, a barrel temperature of 120oC and a gap of 2.54 mm ...... 212 xx 8.8 Diffusion coefficient for the ultrasonically treated IR gum and their vulcanizates as a function of ultrasonic amplitude...... 213
8.9 Fast-diffusing echo fraction for the ultrasonically treated gum IRs and their vulcanizates as a function of ultrasonic amplitude ...... 214
8.10 Fast-diffusing sample fraction for the ultrasonically treated gum IRs and their vulcanizates as a function of ultrasonic amplitude ...... 215
8.11 Transverse 1H relaxation decay at 70.5oC for the virgin IR fitted with the two-component model with the exclusion of the Weibull exponent...... 217
1 8.12 Spectral decomposition of the H T2relaxation decay for the virgin IR at 70.5oC...... 218
8.13 Transverse 1H relaxation decay at 70.5oC for the virgin IR fitted with the two-component model containing the Weibull exponent ...... 219
8.14 Chemically extracted sol fraction as a function of ultrasonic amplitude for the unfilled IR devulcanized at a barrel temperature of 120oC, a gap of 2.54 mm and varied flow rates...... 220
8.15 Proton T2 for the virgin, the cured and the devulcanized IR as a function of extracted sol ...... 221
8.16 Proton T2 fractions for virgin, cured and devulcanized IR as a function of extracted sol...... 222
8.17 Diffusion spin-echo attenuation of the processing oil at 70.5oC...... 223
8.18 Transverse 1H relaxation decay for the 35 phr CB filled gum IR at 70.5oC fitted with the two-component model ...... 224
1 8.19 Spectral decomposition of the H T2 relaxation decay for the 35 phr CB filled gum IR at 70.5oC ...... 225
8.20 Chemically extracted sol fraction as a function of ultrasonic amplitude for the 35 phr CB filled IR (containing 0 and 10 phr processing oil) devulcanized at a barrel temperature of 120oC, a gap of 2.54 mm and a flow rate of 0.63 g/s...... 226
8.21 Proton T2S for the virgin uncured, cured and devulcanized 35 phr CB filled IR without and with oil as a function of extracted sol...... 227
8.22 Proton T2L for the virgin uncured, cured and devulcanized 35 phr CB filled IR without and with oil as a function of extracted sol...... 229 xxi 8.23 Proton T2 long component fraction fL for the virgin uncured, cured and devulcanized 35 phr CB filled IR without and with oil as a function of extracted sol...... 231
9.1 Experimental (symbols) and fitted (lines) values of normalized gel fraction as a function of normalized crosslink density for the devulcanized unfilled IR and NR...... 241
9.2 Experimental (symbols) and fitted (lines) values of normalized gel fraction as a function of normalized crosslink density for the devulcanized filled IR and NR...... 244
xxii
CHAPTER I
1.INTRODUCTION
Natural rubber (NR) is widely used in the tire industry and is produced naturally from the biosynthesis in Hevea Brasiliensis Trees1. Its stress-induced crystallization behavior has made it unique among all the other synthetic elastomers as far as mechanical properties are concerned. NR is the standard by which the performance of synthetic rubbers are judged2. Synthetic isoprene rubber (IR) is the artificial equivalent of NR since they both share the same basic chemical structural unit – cis 1, 4-isoprene. The most critical difference between these two rubbers is that NR consists almost exclusively of cis
1, 4-isoprene unit (~99%) and a small portion of non-rubber components such as protein, sugar, amino acid, fatty acid and other substances3. While IR contains a lower content of cis 1, 4-isoprene unit and is a 100% chemical product. Different polymerization methods lead to IRs with somewhat different cis 1, 4-isoprene contents4. The slightly different contents of 1, 4-isoprene unit could contribute to the large difference in the rate and degree of crystallization and the mechanical properties. Compared with its natural counterpart, IR is inferior in mechanical strength, anti-aging and crystallization.
However, it exceeds NR in the consistency of product, cure rate uniformity, ease of processing (mixing, extrusion, molding and calendering), purity and particularly that it does not undergo storage hardening3, 5. 1 Recycling of the vulcanized elastomers, especially tires, is of significant concern to the industrial world as the increasing stockpiles of the used products fill up the limited landfills. The United States, like the rest of the industrialized world, uses a tremendous number of tires. Approximately one tire is discarded per person each year. It has been reported that over 1,165 million tires were produced worldwide in 2001, with 250 million scrap tires generated in the United States alone6. Scrap tires not only waste landfill space, they can damage the linings put in place to keep groundwater and surface water from mixing with landfilled contaminants. They also pose great threats to public health and safety. Tire dumps provide excellent breeding grounds for mosquitoes, and elevated incidents of mosquito-borne diseases have been noted near large tire piles. Tire pile fires have been an even greater environmental problem. Tire pile fires release toxic chemicals into the air and surrounding water supplies. Current scrap tire markets can utilize approximately 70% of the scrap tire production7; however this still leaves a substantial amount left over. Therefore, tire waste creates a huge threat to humankind. This urges scientific and engineering communities to develop effective recycling strategies to accommodate the ever-increasing scrap tires generated.
Rubber recovery is not an easy business since we are dealing with three- dimensional chemical crosslinks. The formation of this network is a non-reversible chemical process. It would be reasonable to think that breaking this network inevitably needs some form of energy (mechanical, thermal, electric, chemical, biological, irradiation etc.). There have been many attempts to recycle cured rubbers in the past several decades8. Generally the modern approaches can be divided into two categories.
One approach involves grinding the rubber mechanically to a particle size on the order of
2 ten microns9 without significant severing the chemical bonds. The second approach attempts to devulcanize the waste rubber by breaking the intermolecular bonds of the chemical network10. The current devulcanization techniques include chemical11, mechanical12, cryo-mechanical13, biotechnical14, 15, microwaves16, 17 and ultrasonic devulcanization18, 19. Particularly the last method, due to its advantages, is considered as a promising method to the rubber industry. This technology provides a rapid breakage of three-dimensional network in time on the order of seconds; it is a continuous process without the involvement of any chemicals and the rubbers treated by ultrasound are soft, remoldable and can be reprocessed very similar to virgin rubber20.
Therefore, in this research the application of applying high power ultrasound to devulcanize unfilled and carbon black filled IR vulcanizates will be examined. In order to obtain the effective devulcanization of IR, this study emphasizes on to: 1) study the effects of ultrasound on the molecular structure and properties of the virgin gum IR in order to examine the stability of the carbon-carbon backbone under exposure of high power ultrasound; 2) investigate the capability of high power ultrasound to devulcanize unfilled and carbon black filled synthetic IR vulcanizates; 3) understand the mechanism of ultrasonic devulcanization of the IR by solid-state NMR technique and the simulation of breakup of network structures.
3
CHAPTER II
2.LITERATURE SURVEY AND BACKGROUND
2.1 Natural Rubber (NR)
Before the emergence of synthetic rubbers, there was only one type of rubber – natural rubber. Natural rubber is obtained from the liquid latex secreted by a wide variety of rubber bearing plants, although the most widely cultivated is the rubber tree, Heavea
Brasiliensis21, native to the jungles of the Amazon. This tree grows best in hot and wet climates1. The natural rubber used in tire industry mostly comes from plantations in
Southeast Asia with additional production from Latin America and Africa22. In a modern plantation, when a tree is about seven years old, it is tapped for latex. As the tree matures, it produces from one to four gallons of latex each year. This latex is about 55-
60% water, 30-40% rubber, 1.5-3.0% resins, 1-1.5% proteins, 0.8-1.0% carbohydrates and 0.7-0.9% minerals23. The rubber is suspended in the latex. The collected latex is taken to a latex processing plant, where it is concentrated by centrifugation or coagulated by the addition of acid. The coagulated rubber is washed, dried and formed into conventional types of rubber such as ribbed smoked sheets (RSS), pale crepes, brown crepes or technically specified rubbers (TSR)23.
Nature does not produce natural rubber inside the tree by direct polymerization of isoprene monomer as such: biogenesis follows an extremely complex path which is not 4 yet fully understood but which certainly does not resemble the polymerization methods used by man24. Freshly prepared NR has a low gel content of about 5-10%. On storage, the gel content increases and may reach 50%, or higher depending on storage time. The increase in gel content mainly involves storage hardening but may be so partly due to free radical reactions.
Natural rubber is a substance mainly composed of long entangled chain molecules of cis-1, 4 polyisoprene, which are considerably branched25, 26. There are two ways of arranging the configuration of the isoprene polymer molecule according to the disposition of the polymer chain about the double bonds. As shown in Figure 2.1, one arrangement, the cis form, is the natural rubber. The other is the trans form which is the structure of another naturally-occurring polymer, gutta-percha.
Figure 2.1 Isoprene monomer, Natural rubber (cis-1, 4-addition) and Gutta-
percha rubber (trans-1, 4-addition)
Due to its high stereoregularity, natural rubber crystallizes spontaneously when stored at low temperatures or when stretched. Unstrained rubber has a maximum rate of crystallization at -25oC27. But even at 0oC, NR can crystallize in a few weeks. The maximum degree of crystallinity reached is only about 25-35%22. Crystallization leads to a stiffening of the rubber and is reversible upon heating. Rapid crystallization during
5 stretching gives NR the unique high tensile strength and tear resistance in pure gum or non-reinforced vulcanizates.
NR is usually considered to have good processing properties. Although it is tough and nervy at temperatures well below 100oC23, it breaks down easily to a usable plasticity. With its wide range of useful properties, NR can be used in a wide variety of applications. These include versatile products such as hoses, conveyor belts, rubber linings, gaskets, seals, rubber rolls, rubberized fabrics, etc. Because of its high elastic deformability, it is also used in dynamic applications such as springs, antivibration mountings, bushings and so forth. High fatigue resistance, good strength, durability and low heat build-up make NR the favored material in tires for passenger-cars, trucks and aircrafts, in both the carcasses and side walls. A small amount of oil-extended NR is used in winter-tire treads. In commercial vehicles, the amount of NR used increases with the size of the tire. In large earthmover tires, for example, almost 100% NR is used due to the requirements of low heat generation and maximum cutting resistance3. Products made from NR are less likely than most other elastomers to fail from excessive heat buildup or fatigue when exposed to severe dynamic conditions. This has secured the place of NR as the preferred sidewall elastomer in radial tires28.
Natural rubber’s principal weakness is the lack of inherent resistance to environmental damage due to the unsaturated unit in its macromolecular chain. The atmospheric oxygen (especially at high temperatures) and ozone easily attack the unsaturated sites causing weathering cracks29.
6 2.2 Synthetic Isoprene Rubber (IR)
Synthetic isoprene rubber is mostly composed of cis-1, 4 polyisoprene. The basic monomeric unit of polyisoprene is isoprene (2-methyl-1, 3-butadiene). The origins of synthetic isoprene rubber can be traced back to the first half of the 19th century when attempts were made to elucidate the composition and structure of natural rubber with the goal of reproducing NR30. The Frenchman Georges Bouchardat, with the aid of hydrogen chloride gas and prolonged distillation, converted isoprene to a rubberlike substance in
187931. Researchers tried that trick with isoprene isolated from other materials, but those early synthetic versions of rubber were not only inferior to natural rubber due to their low stereoregular structure but also very expensive to make.
During World War II, scientists extensively studied the polymerization of isoprene with the hope of replicating NR since the United States was temporarily cut off from sufficient NR supplies for military needs. These efforts were not successful until the
1950s. After World War II, increasing sophistication in synthetic chemistry led to the synthesis of many new polymers and elastomers. In 1953–54 two chemists, Karl Ziegler of Germany and Giulio Natta of Italy developed a family of organometallic catalysts that were able to precisely control the placing and arrangement of units along the polymer chain and could thus produce regular (stereospecific) structures32. With the use of such catalysts, isoprene was polymerized in such a manner that each unit in the chain was linked to its predecessor in a cis configuration, virtually identical to the structure of natural rubber. In this way almost 100 percent cis-1, 4 polyisoprene, “synthetic natural rubber”, was made.
7 In 1954 the B.F. Goodrich Company was successful in preparing a synthetic cis-1,
4-polyisoprene through the use of the newly discovered Ziegler transition-metal halide coordination-type catalyst (consisting of a trialkylaluminum, AlR3, and titanium
33 tetrachloride, TiCl4) . Soon afterwards, Firestone Tire & Rubber Corporation revealed a synthesis of cis-1, 4-polyisoprene with a catalyst based on lithium metal and yielded a polyisoprene with ~ 92% cis-1, 4 structure4, 34, 35.
Initial commercialization of a stereoregular, low cis-1, 4 IR (90% to 92%) was realized in 1960 by Shell Chemical Company36 with the introduction of Shell Isoprene
Rubber, produced with an alkyl lithium catalyst (Li-IR). However, the cis-1, 4 content of
Li-IR was insufficient to achieve the important crystallization properties of natural rubber because of its relatively low content of cis-1, 4 isoprene unit.
In 1962, Goodyear36 introduced NATSYN®, a Ziegler-Natta (titanium-aluminum) catalyzed IR (Ti-IR) with a cis-1, 4 content of 98.5%4, finally allowing the benefits of strain-crystallization to be realized. Goodrich-Gulf introduced another Ti-IR polymer three years later, but subsequently withdrew from the market in 1978. The manufacture of high cis IR has since been undertaken elsewhere, primarily in Russia and Japan37.
Figure 2.2 depicts a simplified flow diagram for an isoprene polymerization process38. Before entering the reactors, the solvent, catalyst, and isoprene monomer must be free of chemical impurities, moisture, and air—all of which are catalyst poisons. The purified streams first enter a series of chain of reactors into which the catalyst is injected, and the polymerization begins. After the desired extent of polymerization has been attained, a shortstop or catalyst deactivator is added to the reaction mixture so no further linkage of monomer or polymer takes place. A non-staining antioxidant is then added to
8 protect the polymer during finishing and storage. In the next step, the mixture is put through a stripping operation whereby the solvent is recovered and the reaction product containing the polymer is converted to a crumb by hot water and steam. The crumb slurry is processed through extruders to remove water before it is cooled, baled, packaged, and placed in storage ready for shipment.
Figure 2.2 Schematic drawing for the production of cis-1, 4 polyisoprene38
Stress crystallization in cis-1, 4 polyisoprene leads to important physical properties such as green strength (the ability of an incompletely cured material to undergo removal from the mold and handling without distortion), tear strength, and gum tensile strength. Research has shown39 that there are major differences in the ability of the cis-1, 4 IRs to crystallize depending on the level and the nature of the cis microstructure found in the polymer. For Li-IR, while x-ray diffraction patterns have indicated some 9 crystallinity in stretched specimens, no crystallinity is seen in the unstretched state. Ti-IR and NR both undergo crystallization in the unstretched state at low temperatures (the maximum rate of crystallization occurs at -25oC27), but the rate is always faster for NR.
Both synthetic and natural rubbers undergo crystallization at room temperature on stretching, and NR stress crystallizes at a lower elongation than Ti-IR40.
Currently, synthetic polyisoprene is being used in a wide variety of industrial applications. Roughly 60% of cis-1, 4 polyisoprene is used in tires4. Similar to the application of natural rubber, isoprene rubber goes largely into truck tires, off-the-road tires, aircraft tires, and carcass and sidewall compounds of passenger car tires.
Replacement of natural rubber with up to 20% of the 92% cis-1, 4 polyisoprene or up to
40% with 98% cis-1, 4 polyisoprene is possible without significant differences in manufacturing or tire performance. In some cases, a 100% replacement of natural rubber in passenger tires and truck tire treads has proven to be satisfactory4. The remaining 40% of isoprene rubber goes into automotive bushings and motor mounts, belting, gaskets, footwear, battery separators, adhesives and flooring. In addition, recent concerns about allergic reactions to proteins present in natural rubber have prompted the increased usage of synthetic polyisoprene in medical applications.
2.3 Comparison of NR and IR
Natural rubber is a general-purpose elastomer. The high resilience, low heat build-up and excellent dynamic properties ensure natural rubber’s place as a major source of elastomer for automotive tire products. The polymer contains almost exclusively 100% cis-1, 4 polyisoprene41, and has a high molecular weight. Synthetic polyisoprene has a slightly lower percentage of the cis 1, 4 polyisoprene, and lower average molecular
10 weights. These properties impart higher crystallinity to natural rubber40, which in turn lead to enhanced heat loss properties desirable for truck and aircraft tires.
Although synthetic isoprene rubber demonstrates lower green strength, slower cure rates, lower hot tear, and lower aged properties than its natural counterpart42, the commercial synthetic polyisoprene exceeds the natural type in color, low odor, possible elimination of pre-mastication, faster breakdown and mixing, consistency of product, better extrusion and molding and lower hysteresis4. The very specific nature of synthetic polyisoprene provides a number of factors that differentiate it from natural rubber. There is minimal variation in physical properties from batch to batch. Polymerization conditions are narrowly controlled to assure that the polymer is highly stereospecific in chemistry and has a narrow molecular weight distribution. Natural rubber contains 6-8% naturally occurring non-rubber materials23. The allergenicity of plant proteins retained in commercial natural rubber products has been detrimental for its use in certain medical and consumer applications.
The comprehensive comparison of natural rubber and synthetic isoprene rubber is summarized by us from relevant literatures 4,43 in Table 2.1.
2.4 Carbon Black (CB)
Carbon black refers to a group of industrial products involving thermal, furnace, channel and acetylene blacks44. It is essentially elemental carbon in the form of fine amorphous particles. Each particle is composed of randomly oriented microcrystalline layered arrays of condensed carbon rings45. Individual round carbon black particles do not exist as discrete entities but form aggregates, which may be clumps or chains of various sizes and configurations. The functional carbon black “particle”, is thus the
11 aggregate46. The major differences among commercial grades result from the control of average sizes of aggregates. Carbon black is produced by the incomplete combustion or thermal pyrolysis of hydrocarbons44. In the former case the oxygen diffuses into the gaseous hydrocarbon stream after leaving the burner (diffusion flames); while in the latter, the hydrocarbon and air are mixed before leaving the burner (premixed flames).
Table 2.1 Different polymerization processes lead to different products
Product Synthetic IR Natural Rubber (NR)
Catalyst Lithium Ziegler-Natta –
Polymerization Mechanism Ionic Coordination Bio-synthesis
cis-1,4 Content 90+% 96+% 98+%
MWD Narrow Broad Broad
Chain Structure Linear Branched Branched
Purity, Rubber Content >99% >99% ~94%
Unstretched State Crystallizability No Yes Yes
Storage Hardening No No Yes
Until 1968, carbon black nomenclature was informal and inconsistent. It was just based on a variety of characteristics, including level of abrasion resistance, level of reinforcement, vulcanizates modulus, processing properties, general usefulness, particle size, and electrical conductivity. In 1968, the ASTM committee on carbon black established a common nomenclature system ASTM D176547 consisting of a prefix followed by a three-digit number. The prefix is either N, for normal curing, or S, for slow curing. All current rubber grade carbon blacks carry the N prefix. The first of the three 12 digits indicates a range of average particle size in nanometers. The second and third digits are assigned by the ASTM Committee to new products when they are developed. In general, lower structure carbon blacks are assigned lower numbers and higher structure carbon blackshigher numbers, although there are some exceptions.
Carbon black has found wide spread applications as a pigmenting agent for inks, paints and plastics, a conducting material and more importantly as a reinforcing filler for cured rubber compounds, such as tires, belts and hoses. It provides economics in that it reduces the amount of more costly polymer needed for a rubber product, yet it also adds mechanical strength and wearability due to polymer-filler interactions and Van der Waals attraction between carbon aggregates48. It is recognized that the main parameters of carbon blacks which govern their reinforcing ability in rubber are the following49:
1) The size and distribution of primary particles joined by fusion into randomly
arranged aggregates. The particle size and its distribution directly determine the
surface area of the carbon blacks.
2) The size, shape and distribution of aggregates. These parameters are generally
termed carbon black “structure”.
3) Surface activity. It is related to the reactivity of the chemical groups on the carbon
black surface, and in terms of physical chemistry, is referred to as adsorption
capacity. This capacity is determined by carbon black surface energy.
All these parameters together play a role in carbon black reinforcement of rubber through different mechanisms, such as interfacial interaction between rubber and carbon black, occlusion of the polymer in the internal voids of the aggregate, and the agglomeration of carbon black aggregates in the polymer matrix.
13 2.5 Carbon Black Filled Rubber
Reinforcement of elastomers is characterized by the increase in modulus
(stiffness) and the improvement of fracture properties such as tensile, tear and abrasion resistance50. By incorporating a filler, the dimensional stability of the rubber compound is also improved and the cost of the product is reduced. The most common filler used for reinforcing rubber is carbon black due to its unique ability to enhance the dynamic mechanical properties of elastomers44, 51, although the research of applying silica in tires also came up in recent years52, 53. The fast development of carbon black came after the introduction of synthetic rubbers, particularly styrene-butadiene rubber (SBR) during
World War II because its strength as a pure gum vulcanizate is so low that SBR needed carbon black to bring out its potential for practical usefulness54.
The properties of carbon black filled rubber heavily depend on the nature and the concentration of carbon black. The nature of carbon black is determined by its particle size or specific surface area, geometrical arrangement of the carbon black unit and the nature of the surface. Usually the larger the particle size, the smaller the surface area, and consequently the reinforcement declines rapidly55, 56. The viscosity of the compound is affected not only by the filler content but also by its structure and particle size. The higher the viscosity, the higher the shearing forces created during the mixing and processing, and the better the dispersion achieved57. Increasing the amount of carbon black increases the stiffness of the vulcanizates58, 59 and generally decreases the elongation at break. The modulus increases with increasing filler content, but the relationship is not linear60. Unfilled gum vulcanizates have the highest resilience. It decreases substantially with increasing filler content61. Hardness increases with 14 decreasing particle size and increasing structure and carbon black loading. Also, carbon black is an additive with a decisive effect on the wear resistance62.
The reinforcing mechanism of carbon black has been studied. With the addition of carbon black, shorter molecular chains share the stress with longer ones so that the rubber does not break easily. Carbon black reinforces the rubber by adding many points of friction within the molecular system. It inhibits rupture by absorbing and dissipating strain energy at the friction points where the rubber molecules slip across the carbon black surface49, 63. Due to the interaction between rubber and carbon black, the rubber molecules can be absorbed onto the filler surface either chemically or physically. This is related to the restriction of the segmental movement of polymer molecules.
Carbon black influences vulcanization kinetics. Generally compounds filled with carbon black have a lower scorch time (premature vulcanization) than unfilled rubber, and have an increased rate of vulcanization is increased due to the addition of carbon
64 black . Carbon black facilitates the opening of the S8 rings even in the absence of accelerators and it enhances the formation of H2S, which activates most sulfur curing system. Carbon black supports the vulcanization reactions without changing their nature substantially25.
The addition of carbon black also influences the crosslink structure of vulcanized rubber. Variations in size of the rubber network caused by the filler can be studied either by means of the equilibrium swelling method65 or by measurements of elastic modulus of filled rubber66. The first method is based on the assumption that the interaction parameter between the rubber and solvent χ is independent of the filler content and thus its value is the same for filled and unfilled vulcanized rubber. Kraus67 determined the crosslink 15 density for a wide series of unfilled and carbon black filled vulcanizates and developed a theory to account for the restricted swelling in solvents of crosslinked elastomers containing reinforcing fillers. According to Kraus67 and Bueche68, the increase in the effective number of chains of the network caused by the filler depended on the crosslink density of the unfilled vulcanizates prepared with the same vulcanization system.
2.6 Vulcanization
Vulcanization is the chemical process of crosslinking the chain-like rubber molecules to form an elastic three-dimensional network which prevents permanent deformation after removal of the deforming force. In order to develop the full potential of rubbers, they must be crosslinked. For rubbers containing unsaturated carbon-carbon bonds in the backbone, a variety of curing agents are available to choose with some common ones being sulfur69, 70, peroxides71, phenolic resin72, and metallic oxides73.
Among them, sulfur is the most popular one employed in rubber vulcanization industry.
In most cases, the cure reaction is achieved by a chemical reaction between the rubber and curing agent74. The first method of vulcanization was based on the discovery of
Charles Goodyear in U.S. and Thomas Hancock in England75. It was found that addition of sulfur to rubber, followed by heating, led to an improvement in the properties.
Vulcanization results in the increase of tensile strength, modulus (stiffness), hardness, abrasion resistance and rebound, and the decrease of elongation, hysteresis (heat buildup), compression set and solubility. It also makes rubber impermeable to gases and resistant to heat, electricity and chemical action. The improved frictional properties of rubber by vulcanization are highly desired for pneumatic tire application. Tensile and tear strength usually show a specific optimum crosslink density. Vulcanization-induced
16 changes are proportional to the number of crosslinks and their length. Excessive crosslinking can convert the elastomer to a hard and brittle solid.
2.6.1 Sulfur Vulcanization
Sulfur vulcanization is, by far, the most commonly used method for general- purpose diene rubbers (NR, IR, SBR, butadiene rubber [BR], nitrile-butadiene rubber
[NBR] and chloroprene rubber [CR]) owning to its low cost, low toxicity, broadly compatibility with other compounding additives, availability, ease of processing, properties, and adaptability to diverse methods of heating media, compounding ingredients and temperatures76.
Sulfur is available in two forms: amorphous and rhombic. The amorphous form, also known as insoluble sulfur, is a metastable high polymer that is insoluble in rubber and most solvent. Rhombic sulfur, a ring of eight sulfur atoms, S8, usually in its normal crystalline state, is the form normally used for vulcanization77. It is thermally very stable, but upon heating, ring opening occurs at the activation energy of 270 kJ/mol77. Due to the high activation energy for sulfur ring opening, prolonged heating at high temperature is required.
Vulcanization of rubber by sulfur alone is a slow and inefficient process. The rate at which sulfur reacts with unsaturated polymer can be accelerated by certain chemical substances, or accelerator. Initially, the acceleration phenomenon was discovered by G.
Oenslager75 in Germany in 1906. It was found that the addition of aniline to a rubber/sulfur formulation greatly increased the rate of vulcanization and improved the final vulcanizate properties. It was then quickly realized that the use of accelerators resulted in improved properties and significantly reduced curing times75. Accelerators 17 function best when accompanied by metallic oxides such as lead, zinc or magnesium oxide along with a fatty acid. The most common combination is zinc oxide and stearic acid, with the latter solubilizing the zinc in the elastomer. It is believed that78 the sulfur changes into a cation and reacts, in the presence of the metal, at the double bond. This reaction results in charged and uncharged polysulfides, the latter of which could form free radicals.
Presently, there is a wide range of accelerator systems available for elastomers, providing a range of cure rates, scorch time and final properties. The choice of vulcanization accelerator will affect the scorch safety, the cure rate, and the length and number of crosslinks formed. These properties, which are generally related to the speed at which the accelerator is converted to its active salt form, are compared in Table 2.279.
Table 2.2 Accelerator type comparison79
Accelerator Type Scorch Safety Cure Rate Crosslink Length
None Very slow Very long
Guanidines Moderate Moderate Medium-long
Mercaptobenzothiazoles Moderate Moderate Medium
Sulfenamides Long Fast Short-medium
Thiurams Short Very fast Short
Dithiocarbamates Least Very fast Short
Currently available accelerators include the thiurams, sulfenamides, mercaptobenzothiazoles and amines80. Several common accelerators are listed in Table
2.381. 18 Table 2.3 Several common accelerators used in sulfur vulcanization81
Accelerated sulfur vulcanization systems can be classified into three types82: (1) conventional (CV) systems with low accelerator / sulfur ratios, (2) efficient (EV) systems having high accelerator / sulfur ratios, and (3) semi-efficient (semi-EV) systems with
19 accelerator / sulfur ratios in between (1) and (2). A commonly used division of the three systems for NR is given in Table 2.482. The vulcanization leads to a variety of crosslink structures with different lengths and distributions77 as shown in Figure 2.3. CV systems result in longer (polysulfide) crosslinks. Therefore, CV systems provide the vulcanizates with excellent initial properties such as strength (tensile, tear), resilience and resistance to fatigue and abrasion, and are satisfactory for most applications. However, due to reversion (cleavage of polysulfide crosslinks resulted from over heating), their heat-aging resistance, creep and stress-relaxation properties were reduced. In contrast, EV systems lead to shorter crosslinks (mono- or di-sulfide) which are more stable (less prone to scission) and thus provide better oxidative and thermal stability and lower compression set. The effect of semi-EV system is usually in between the CV and EV and it may be chosen as a compromise between the cost and performance. Typical NR vulcanizate structures at the optimum cure time and some properties are shown in Table 2.582.
Table 2.4 Sulfur and accelerator of different vulcanization systems for NR82
Vulcanization System Sulfur, phr Accelerator, phr Accelerator-Sulfur Ratio
Conventional (CV) 2.0-3.5 1.2-0.4 0.1-0.6
Efficient (EV) 0.4-0.8 5.0-2.0 2.5-1.2
Semi-EV 1.0-1.7 2.5-1.2 0.7-2.5
Figure 2.3 Structures formed during sulfur vulcanization of elastomers77
20 Table 2.5 Typical NR vulcanizate structures and properties at optimum cure time82
CV Semi-EV EV
Poly- and disulfidic crosslinks, % 95 50 20
Monosulfidic crosslinks, % 5 50 80
Cyclic sulfidic concentration, % High Medium Low
Low temperature crystallization High Medium Low resistance
Heat-aging resistance Low Medium High
Reversion resistance Low Medium High
Compression set, 22 hr. at 70oC 30 20 10
The extent of cure is measured as a function of cure time by using a cure meter.
The oscillating disc rheometer (ODR) is the oldest type of rheometer used to characterize the vulcanization kinetics at a characteristic of vulcanization temperature 83, 84. Resistance to this oscillation is measured and recorded as a function of time in order to generate the so-called rheometer charts. Rheometer charts (also called cure curves, see Figure 2.4) can characterize both the induction period and the rate of vulcanization or cure, once it starts.
The cure curve can be obtained not only by ODR, but also by modern rheometers such as the Advanced Polymer Analyzer (APA2000) developed by Alpha Technologies®. The difference between these two types of rheometers is that: ODR can only detect torque curves; APA not only measures torque changes with time, but also converts torque into dynamic properties such as storage modulus G’, loss modulus G”, loss tangent tanδ and dynamic viscosity η*. 21 Reversion Torque, Nm
Induction Curing Overcure /scorch
Time, min
Figure 2.4 Typical cure curves for accelerated sulfur vulcanization
2.6.2 Mechanism of Accelerated Sulfur Vulcanization
Sulfur vulcanization involves a combination of zinc oxide, fatty acid, sulfur, and at least one accelerator. There are a lot of studies focused on the vulcanization mechanism in various elastomers. Accelerated sulfur formulations are the most popular vulcanization systems used in commercial and industrial applications. Although it has been studied since the 1950s, the exact mechanism of accelerated sulfur vulcanization remains unresolved, including whether it proceeds by a radical or ionic process. The studied mechanisms74, 81, 84, 85, 86 for accelerated sulfur system include ones based on ionic, radical, combination of ionic and radical, and intermediate formation. It was suggested81 that the nature of the reaction may change, depending on the polarity of the particular polymer and whether or not zinc oxide is present. However, there is much controversy over the nature of the chemical reactions.
22 Nevertheless, there is some agreement on the general mechanism. In detail, induction or scorch-delay period is a common phenomenon in cure curves, although the induction period for a particular rubber varies in a wide range depending on the types of accelerator used. Zinc oxide and stearic acid produce soluble zinc stearate (a kind of zinc soap) at certain temperature during this period25. A zinc complex can form an accelerator complex (Acc-Zn-Acc; Acc is a moiety of the accelerator) through the interaction of fatty acid and other ligands. For the CBS accelerator system, Krebs87 proposed accelerator complex with amine ligands which donate their electrons to the zinc ion, thereby weakening the bonding force between the zinc and accelerator moiety. Under this condition, the elemental sulfur ring can be split and the sulfur chain be put between the zinc ion and the accelerator moiety. The possible formation of zinc complex with amine ligands under CBS accelerated sulfur vulcanization is shown in Figure 2.587. It was reported84, 88 that accelerators reacted with sulfur to form monomeric polysulfides of the type Acc-Sx-Acc, where, Acc is an accelerator-derived moiety. These polysulfides reacted with rubber to give polymeric polysulfides of the type rubber-Sx-Acc.
After the induction period, crosslinks of the type rubber-Sx-rubber are formed.
Curing torque increases with the degree of crosslinking, which can be considered as the magnitude of elastic modulus of the cured elastomer. This modulus is proportional to the crosslink density89. During the vulcanization period, the rubber’s elastic modulus increases up to a certain value called the plateau modulus. Accelerators shorten the time to reach plateau modulus during the reaction. Therefore, the rate of crosslink formation depends on the type of accelerator90. Morrison91 found that CBS accelerator affected the cure kinetics in both delay period and curing stage. During the induction period CBS
23 hinders the reaction, and during curing stage it promotes the cure reaction rate with its cyclohexylamine complexes.
NR1 NR1
Acc Zn Acc + S8 Acc Zn Acc
SS NR1 NR1 S6
NR1 NR1
Acc Sa Zn Acc Acc Sb Zn Sc Acc
NR1 NR1
Figure 2.5 Possible formation of zinc complex with amine ligands under CBS
accelerated sulfur vulcanization (NR1 is the amine ligand, the amine moiety of a
sulfenamide accelerator) 87
Some of the structural features that are generally accepted as occurring in sulfur vulcanized NR are shown in Figure 2.692. Zaper and Koenig85, 86 studied sulfur- vulcanized NR and accelerated sulfur-vulcanized NR systems using solid-state NMR technique for the detection of crosslinks and other structural modifications. They reported that the use of sulfur as a crosslinking agent by itself generates a considerable amount of main-chain structural modifications. Cyclic sulfide structures and cis-to-trans chain isomerization are detected in addition to polysulfidic crosslinks which are found attached to different carbons of the NR repeat unit. When the amount of accelerator is high 24 proportional to the amount of sulfur used, the network structure appears to be simpler with less crosslinking, less main chain structural modification, and fewer cyclic sulfide structures. Polysulfidic crosslinks have been detected in addition to accelerator terminated polysulfides.
CROSSLINKS OTHER MODIFICATIONS
C C Sx S C R C SA monosulfidic pendant side group CC
CC C SB S S C cyclic sulfides S multifunctional vicinal C - C=C-C=C-C=C-C=C -
disulfidic conjugated unsaturation CC
C SA SB C = C C = C
Sx C C C cis/trans isomerization vicinal polysulfidic
Figure 2.6 Structural features of sulfur-vulcanized NR92
The final stage of vulcanization is the over cure period of elastomers. At this stage, an initially linked network can be matured by thermal dissociation and 25 redistribution of sulfur bonds93. Polysulfide crosslinking rubbers are subsequently degraded to have a lower number of sulfur atoms in the network. The dissociation of sulfur bonds can promote the reaction or release of the sulfur94. The released sulfur can combine with a double bond complex or many other low molecular weight cyclic structures74.
2.7 Rubber Recycling8
Vulcanization is the reason that gum rubber is useful in the tire industry.
Unfortunately, it has also created a serious environmental problem as tremendous amount of waste rubbers dumped and stockpiled. Unlike the thermoplastic polymers which can be easily reprocessible by heating, the thermoset polymers, such as vulcanized elastomers can not be simply reused once they form the three-dimensional network. However, the recycling of such materials is demanding due to the environmental and economic factors resulted from the ever-increasing amount of waste rubbers, especially scrap tires. There have been many efforts to recycle waste rubbers during the past several decades8. Those methods are generally divided into two categories. One is at the physical level. It involves grinding the material mechanically into smaller pieces without breaking the chemical bonds. The end result is the size reduction. Another category of methods attempts to break the three-dimensional network chemically with the aid of various forms of energy
(mechanical, thermal, chemical, biological interaction, microwave, ultrasound, etc.). It has to convert the three-dimensional, insoluble and infusible thermoset into a soft, tacky, reprocessable and revulcanizable product simulating the properties of the virgin rubber.
Recovery and recycling of rubber from the used rubber products will save precious petroleum resources as well as solve the waste rubber disposal problem.
26 2.7.1 Landfills and Waste Rubber Utilization
Landfills have been, over the years, an easy way to dispose of waste rubber.
However, fires at dumpsites, lack of space, and the increasing costs associated with the land-filling operations have made this a non-viable solution95. Landfills provide breeding grounds for mosquitoes and rodents when waste tires are stockpiled or illegally dumped.
Another problem associated with the dumping of waste rubber is the leaching of toxic chemicals into the surrounding soil, which makes it a huge threat to agriculture and human health. Consequently, a majority of the states in America banned whole tires from landfills since 2003 due to fire hazards and human health hazards96.
2.7.2 Grinding Methods
Grinding tire rubber has been one of the most common methods used for recycling97. The end result of the process is the size reduction of the waste rubber, with the ability to control the average particle size. There are many grinding processes that have been developed to attain the particulate form of the rubber, namely, ambient grinding, cryogenic grinding and solution grinding.
Ambient grinding, unlike its name, does result in the generation of heat. In this process, vulcanized rubber is placed in a serrated grinder, reducing rubber to particles of size 10-30 mesh. The surface quality of the end product is highly dependant on the size of the particles. If an attempt is made to increase particle size, the smoothness of the particles decreases. The generation of heat during the process is higher for aged rubber, or rubber with a higher modulus, which in turn leads to increased degradation of the polymer chains95.
27 In the cryogenic grinding process, small pieces of vulcanized rubber are placed in liquid nitrogen for a period of time and are then transferred into a ball mill, in the presence of liquid nitrogen, to form a fine powder98. The size of the particles is controlled by the change in the immersion time of the polymer in the liquid nitrogen95. An advantage of this process is the lack of heat generation, which avoids any likelihood of the degradation of polymer chains in the vulcanized products. Secondly, the ease of separation of the fiber and steel from the rubber, the resulting increase in the yield of rubber, and finally the decreasing costs of liquid nitrogen have led to a significant increase in the use of the cryogenic grinding process.
Wet or solution grinding is another grinding process that reduces the particle sizes by grinding in a liquid medium. The process requires the use of coarse ground particles, approximately between 10-20 mesh in size, which are ground between two closely spaced grinding wheels in a liquid medium95. An advantage of the process is the improved heat transfer during shearing and size reduction, which avoids degradation of the polymer chains; secondly, the sizes that can be obtained are as small as 400-500 mesh.
2.7.3 Rubber as a Fuel Source and Pyrolysis
The high calorific value of rubber, 32.6 mJ/kg, compared to that of coal, 18.6-
27.9 mJ/kg99, and the fact that rubber contains over 90% organic materials, makes burning scrap rubber a great resource for fuel95. However, burning scrap rubber for fuel brings with it the problem of air pollution due to emissions.
Environmental concerns led to the development of a process that recycles rubber using an oxidation process, which results in the breakdown of the polymer by selective 28 oxidative decoupling of C-C, C-S and S-S bonds by water as a solvent near its supercritical temperature100. Another method of using waste rubber has been developed that adequately recovers oil, steel, and carbon black. This method involves heating the rubber at 700oF for 10 minutes to obtain the resulting by-products101.
Pyrolysis provides another route for reuse of the waste tire rubbers due to their high potential energy values stored in the hydrocarbons. It involves in the thermal decomposition of organic materials such as rubbers in the absence of air and oxygen to produce valuable gases and oils to be reusable97. In this process, the carbon black and steel can be recovered. However the separation of the components of gases and oils could be an expensive operation and it also inevitably releases the toxic substances into the atmosphere102.
2.7.4 Chemical Method
Research in using chemical method to devulcanize crosslinked rubber occurred at
1960-70’s. The original purpose of this method was to apply some specific chemical reagent as a probe to determine the structure and types of the sulfur crosslinks103. The chemical method was able to distinguish among the polysulfide, disulfide and monosulfide bonds in the sulfur cured rubber vulcanizates by measuring the crosslink density before and after the treatment of rubber with different chemical reagents specifically breaking particular types of crosslinks. For example, Saville et al104 used propane thiol/piperidine to cleave polysulfide linkages while leaving the mono-, disulfide and carbon-carbon linkages intact. Campbell105 found that hexane-1-thiol was more reactive and could cleave both poly- and disulfide links while leaving the monosulfide and carbon-carbon bonds intact. Selker et al106, 107, 108 reported the importance of methyl 29 iodide to break only the monosulfide bonds in rubber vulcanizates. There were also other effective chemicals to cleave particular crosslinks. However, it was found that some of the chemical reagents such as methyl iodide were carcinogenic.
Although the chemical method is very powerful to distinguish among the different types of crosslinks and analyze the chemical structures, this process is very slow and it creates additional problems such as the removal of solvents and operation safety.
2.7.5 Microwave Method
By carefully choosing the dosage of microwave energy, at a certain frequency, and at the desired energy level, elastomers can be devulcanized and thus be reprocessed.
Hence, this is a process where the elastomer can be reclaimed to a material that is capable of being re-compounded and has the properties that are equivalent to the original vulcanizate. It was presumed that breakdown of chemical bonds in the vulcanizates only occured in carbon-sulfur and sulfur-sulfur bonds and it resulted in no significant main chain (carbon-carbon bonds) degradation16, 17. However, one critical requirement of the microwave process is the presence of polar groups in the polymer. The availability of polar groups in the polymer results in an increase in the temperature of the material once it has been exposed to the microwave energy, which would result in the severing of the crosslinks.
2.7.6 Biotechnological Method
There are also some trials to use microorganisms to devulcanize waste rubbers. It was reported109 that microorganisms were able to break sulphur-sulphur and sulphur- carbon bonds by digestion and thus they could be used to devulcanize waste rubber in
30 order to make polymer chains on the surface more flexible and facilitate increased binding upon vulcanization.
Biodegradation of NR was achieved by Tsuchii and coworkers110, 111. They used bacteria from the genus Nacardia and the process led to a substantial weight loss of different types of NR vulcanizates. A recent approach involves the utilization of a type of fungus to degrade the vulcanized NR sheets on a wood medium112. The fungus decreased the total sulfur content of the rubber by 29% in 200 days, accompanied by the cleavage of sulfide bonds between polyisoprene chains. Dipolar decoupling/magic angle spinning
(DD/MAS) solid-state 13C NMR revealed that the fungus preferentially decomposed monosulfide bonds linked to a cis-1, 4-isoprene backbone but the cleavage of polysulfide bonds was also observed.
Biotechnological method has advantages over the mechanical and chemical processes since it consumes little energy and does not require hazardous chemicals.
However, an obstacle with the biotechnological processes is that it involves living organisms, which are affected by the environment they are cultivated in. Chemicals involved in vulcanization process (accelerators, antioxidants, etc.) might suppress the growth of microorganisms and thus inhibit the biodegradation of rubber materials113. Zinc oxide, mercaptobenzothiazole, dithiocarbamate accelerators and para-phenylenediamine type antioxidants are particularly strong anti-metabolites114. Besides, rubber biodegradation is a slow process, and the growth of bacteria utilizing rubber as a sole carbon source is also slow113. Therefore, incubation periods extending over weeks or even months are required to obtain enough cell mass or degradation products of the polymers for further analysis. 31 2.7.7 Ultrasonic Method
The most promising method of recycling elastomers is the ultrasonic method.
Ultrasonic technology was first reported in 1973 by Pelofsky115, in which rubber particles were immersed in a liquid medium, and then were exposed to ultrasonic energy, which resulted in the disintegration and dissolution of the polymer in the liquid medium.
Ultrasonic irradiation with a frequency of 20 kHz, and a power intensity greater than 100
W was used in the process. The next step in the development of ultrasonic technology was the development of a process by Okuda and Hatano116, in which a natural rubber vulcanizate was subjected to 50 kHz ultrasonic energy for approximately 20 minutes to achieve devulcanization. The researchers claimed that the properties attained after the revulcanization process were similar to that of the original vulcanizate. Mangaraj and
Senapati117, in their patent on ultrasonic vulcanization noted the possibility of degradation of rubber and the crosslinks by ultrasonic energy.
The development of ultrasonic technology to devulcanized rubbers continued until Isayev et al18, 19 developed a continuous devulcanization process. In their process design, the ultrasonic reactor was attached at the end of the extruder allowing the continuous processing and devulcanization of rubber. This process is the most recent approach for the recycling of rubber20, 118, 119, 120 and is now considered to be one of the most promising techniques for the recycling of waste rubber. The experiments have been carried out on various types of elastomers including ground rubber tire (GRT)118, 121,
SBR122, NR123, ethylene-propylene-diene monomer (EPDM)124 polyurethane rubber
(PU)125 and silicone rubber126. The various studies127 have shown that the ultrasonic waves, at a designed level of pressure and temperature, can rapidly break up the three 32 dimensional network of a vulcanized elastomer. The most desirable results lead to the reprocessing and revulcanization of the rubber, giving the end products, in some cases, mechanical properties similar to those of the original elastomers. The ultrasonic devulcanization process is an environmentally friendly process, free of any chemicals harmful or unharmful. The process can be operated continuously, which makes this method very attractive to the rubber industry.
2.7.8 Summary of the Recycling Methods
Recycling of waste rubbers is of a growing importance for the industries worldwide not only because of their high hydrocarbon resource but also due to the ever- increasing environmental problem created by the waste rubbers. As shown above, all kinds of recycling methods are developed to solve this problem. It is recognized that landfilling is not a desirable solution to handle the tremendous amount of waste rubber and therefore it should be prohibited ultimately. Mechanical grinding was the most popular method to treat waste rubbers until other methods involving the breakage of three-dimensional network were developed. It resulted in only the size reduction without significant breakage of chemical bonds. Thus, the waste rubber particles created by this method are especially useful in road filling applications but they generally can not be reprocessed in the same way that virgin rubber does. The use of scrap tires for fuel and the pyrolysis offer good alternatives for reusing waste rubber when fuel costs increase and tire disposal problems become more serious. However, these two methods inevitably bring about air pollution from emissions. Using the chemical probes selectively to cleave certain types of crosslinking bonds provides an effective solution of quantitative structural characterization of the networks and the waste rubber can be significantly
33 devulcanized, but this process is very slow and it creates additional problems of solvent removal and potential hazard resulted from the toxicity or carcinogenicity of the chemicals used. Biotechnological method seems to be a better alternative, compared with mechanical and chemical methods, since it consumes little energy and does not involve any dangerous chemicals. However, the cultivation of suitable strains to digest the sulfur containing chemical bonds in the cured rubber could be time consuming and the microorganisms are usually vulnerable to the additives used in rubber compounding and processing. Microwave method has the advantage of specifically breaking down the sulfur containing bonds without introducing significant main chain degradation. But one critical requirement for this method is the presence of polarity for the treated polymers.
Finally, the ultrasonic method offers a fast and continuous way of handling a huge amount of waste rubbers. Particularly this method can break down the three-dimensional network within seconds. It leads to a preferential breakage of sulfidic crosslinks with a simultaneously minor main chain scission in the vulcanized rubbers. This process involves no chemicals. The ultrasonically treated rubber becomes soft and remoldable.
Consequently the materials can be reprocessed and revulcanized in a manner similar to the virgin rubber. Due to these benefits, the ultrasonic method treatment of IR is the emphasis of this research.
2.8 Recycling of Isoprene Rubber – Current Studies
There are very limited ways of recycling synthetic isoprene rubber reported so far.
Currently the available methods include three types: high temperature pyrolysis, chemical devulcanization and biodegradation.
34 Cataldo128 studied the pyrolysis of synthetic cis-1, 4 polyisoprene in a 50 ml flask with 2 g of material was heated under reduced pressure. However, the pressure, temperature and atmosphere applied were not identified. The distillation product
(pyrolyzate) contains about 96% of dipentene and about 3.5% isoprene according to the density measurement. Chen and Qian 129 investigated the thermal pyrolysis reaction of cis-1, 4 polyisoprene by using pyrolysis-gas chromatography under an inert atmosphere at different temperatures ranging from 330 to 600oC. A small amount of polymer about
20 mg was put into the quartz tube in the pyrolyzer and heated with a rate of 10oC min-1 from room temperature to the desired temperature. The gaseous products were collected using a condensing tube frozen with liquid nitrogen. Their results indicate that at different temperature ranges the major pyrolyzates are similar but with relatively different yields.
They contain small gaseous molecules such as dipentene, isoprene, trimeric isoprene, benzene, toluene, xylene, C2-C4 hydrocarbons, among which dipentene is the major product except at temperatures above 431oC. Clearly, separation of these components will be costly.
Kojima et al.130, 131 used thiol-amine reagent to devulcanize unfilled polyisoprene rubber vulcanizates with the aid of supercritical CO2 (scCO2). ScCO2 was used as a swelling solvent for the vulcanizates and it helped the devulcanizing reagents penetrate and diffuse into the vulcanizates more efficiently than ordinary gaseous CO2. As a result of the devulcanization, the molecular weight of the sol component and crosslink density of the gel component was substantially decreased from the initial samples. The sol fraction increased with the increase in the scCO2 pressure. It was concluded that the partial degradation of main chain took place in addition to the scission of crosslinks. 35 However, the latter played a dominant role in yielding the sol component in the devulcanization.
Besides the pyrolysis and chemical methods, there are also some trials using certain bacteria to degrade isoprene rubber. For example, Bode et al14 used Gram-positive and Gram-negative types of bacteria to degrade natural rubber, synthetic poly cis-1, 4 isoprene and crosslinked NR latex gloves. The measurement of the average molecular weight of synthetic rubber before and after the bacteria treatement showed a time- dependent shift to lower values. It was assumed that degradation of the polymer backbone was initiated by an oxidative cleavage of the isoprene double bond132, 133.
2.9 Application of Ultrasound in Polymers
The range of human hearing is from about 16 Hz to 16 kHz. Ultrasound is generally defined as any sound with a frequency beyond the limit which the human ear can respond134. The upper limit of ultrasonic frequency is not sharply defined but is usually taken to be 5 MHz for gases and 500 MHz for liquids and solids.
The uses of ultrasound within this large range may be divided broadly into two areas134. The first area occurs at the physical level. It involves low amplitude (high frequency) ultrasound and is concerned with the physical effect of medium on the wave and is commonly referred to as “low power” or “high frequency” ultrasound. Typically, low amplitude waves are used for analytical purpose to measure the velocity and absorption coefficient of the wave in a medium in the 2 to 10 MHz range. Information from such measurements can be used in medical imaging, chemical analysis and the study of relaxation phenomena.
36 The second area is usually applied to influence the chemical reactivity and it involves high amplitude (low frequency) waves, known as “high power ultrasound”, and lies between 20 and 100 kHz. It is used for cleaning, plastic welding and, more recently, for sonochemistry. In fact the range available for sonochemistry has been extended to 2
MHz with the development of high power equipment capable of generating cavitation within liquid systems at these high frequencies135.
It is recognized that136, 137 high power ultrasound influences chemical reactivity through an effect known as cavitation. Cavitation is the production of microbubbles in a medium when a large negative pressure is applied to it. When an acoustic field is applied to a liquid, the pressure waves of the sonic vibrations create an acoustic pressure (Pa) which travel through the medium. This acoustic pressure is applied to a system in addition to the ambient hydrostatic pressure (Ph) which is already present in the medium.
138 Pa is time dependent and can be represented by
Pa = PA sin 2πft (2.1) where f is the frequency of the wave (f ≥ 20 kHz for ultrasound) and PA is the pressure amplitude of the wave. Like any sound wave, the ultrasonic wave will alternately compress and stretch the molecules of the medium through which it passes. If the rarefaction wave is sufficiently powerful, it can develop a negative pressure (Pc = Ph - Pa) that is large enough to overcome the intermolecular forces binding the liquid. The liquid breaks down and thus voids (cavitational bubbles) are created. The process of cavitation consists of three steps: nucleation, growth, and collapse of the bubbles139, as demonstrated in Figure 2.7140.
37 When a sufficiently intense sound wave is applied to a medium, bubbles are formed during the expansion portion of the wave. These bubbles then undergo repeated expansion and compression. The intensity experienced by individual bubbles is not constant due to surrounding bubbles forming and resonating around them. This then causes some bubbles to suddenly reach an unstable size. These unstable bubbles then violently collapse, creating extreme heat and pressure141, 142.
compression compression compression compression
+
Pressure
- amplitude wavelength
rarefaction rarefaction rarefaction rarefaction
5000oC 1000 atm Bubble Bubble grows in Reaches Undergoes forms Successive cycles unstable size Violent collapse
Figure 2.7 Ultrasound cavitation bubble growth and collapse140
Based on the assumption of bubble collapse, Noltingk and Neppiras143 showed that the final temperature and pressure generated, Tf and Pf, are:
γ/(γ-1) Tf = To [Pm (γ - 1)/P] and Pf = P[Pm (γ - 1)/P] (2.2) where To is the bulk temperature of the liquid, Pm is the pressure in the bubble after collapse, P is the pressure before collapse and γ is the ratio of specific heats of the vapor
38 or any dissolved gas. Depending on the conditions, the predicted values from these
144 equations are a value of 4000 – 6000 K for Tf and 1000 – 2000 bar for Pf . It is the energy generated on the collapse of these bubbles which is the underlying reason for the tremendous chemical transformations achieved in sonochemistry. In addition to the heat and pressure generated during ultrasonic exposure, numerous other beneficial effects are produced: cavitation, agitation, diffusion, acoustic streaming, interface instability and mechanical rupture145.
2.9.1 Ultrasonic Treatment of Polymers
Most of the work regarding the ultrasonic treatment of polymers was focused on the degradation in solutions146, 147, 148, 149, 150. These early works were important in the theoretical development because they established that: (a) the higher the molar mass of the original polymer, the faster the degradation rate; and (b) there was a lower molar mass limit beyond which the polymer resisted further degradation. Ultrasonic degradation of polymers in solution is different from chemical, thermal or photodegradation in the site of cleavage. The cleavage through thermal, chemical or photodegradation is random and it occurs at the points of inherent weakness within the polymer backbone. In contrast, the cleavage by ultrasound is Gaussian about the mid-point of the macromolecular chain151,
152. Wherever the site of cleavage, the result must be the production of macroradicals153,
154, the existence of which has been confirmed spectroscopically by the use of radical scavengers such as diphenyl picrylhydrazyl (DPPH). Obviously, in the absence of scavengers the macroradicals are free to combine by either disproportionation or combination termination, the former leading to smaller-sized macromolecules while the latter will give a distribution dependent on the size of the combining fragments. 39 In addition to degradation, ultrasound technique have also been applied to the polymerization of monomers due to the fact that radicals are created upon exposure of ultrasound149. It was reported that high intensity ultrasound assisted the emulsion copolymerization of methyl methacrylate (MMA) and styrene monomers. The rate of polymerization and the final conversion were higher compared to the condition without ultrasound. The ultrasonic effect was more pronounced at lower temperatures and low initiator concentration155.
2.9.2 Ultrasonic Devulcanization Mechanism
In the past two decades, extensive research work on the ultrasonic devulcanization of a wide variety of rubbers has been carried out by Isayev and co-workers at the
University of Akron. Their work demonstrates that ultrasound is a fast and effective technique for recycling and reprocessing waste rubber products to reduce the environmental hazard.
As mentioned earlier, cavitation is the reason leading to the polymer degradation in solution. It is also believed that ultrasonic devulcanization is due to the same mechanism. In this case, the cavitation occurs in a solid body. The occurrence of cavitation bubbles in solid polymers can be verified experimentally156. This is because the cavitation can be induced by cavities, voids, and density fluctuations existing in solid polymers. When a rubber sample is exposed to the high power ultrasound, during the half negative pressure cycle, the microvoids presented in the material are subjected to a tensile stress, and during the half positive pressure cycle, they experience compression. This pulsating force is capable of leading to the intermolecular bond rupture.
40 Structural studies of ultrasonically devulcanized rubber show that the rupture of crosslinks is accompanied by the partial main chain scission20, 118, 125, 126, 157, 158, with the former action being the preferential cause in the system. The exact mechanism of ultrasonic devulcanization of rubbers is currently not thoroughly understood.
Nevertheless, it has been shown that devulcanization of rubber with high power ultrasound requires local energy concentration, since uniformly distributed ultrasonic energy among all the chemical bonds is not capable of rubber devulcanization119, 159.
2.9.3 Modeling of Ultrasonic Devulcanization Process
Theoretical modeling of rubber devulcanization by high power ultrasound has been developed by Isayev and co-workers. Ultrasonic devulcanization is a very complex process. Many process details are to be considered during the simulation. On the one hand, the material behaviors (such as viscosity, molecular weight, network structures) are not constant throughout the process and they are coupled with the process parameters, temperature and pressure, which vary with time. On the other hand, the devulcanization process involves all sorts of bond breakage (Sx, S-S, S-C and C-C) at different degrees and rates. Consequently, many individual processes should be accounted for during the theoretical modeling.
2.9.3.1 Elastic Model
A simplified theoretical model is the elastic model119. This model was based on the description of the elastic characteristics of the material by neo-Hookean models160.
The rate of breaking up the various bonds was modeled with a 1st order kinetic equation, with time taken to be the current average residence time at a certain location. The results of simulation gave the distributions of velocity, shear rate, temperature, hydrostatic and 41 acoustic pressure, rate of breakage of various chemical bonds (monosulfidic, disulfidic, polysulfidic, and carbon-carbon), gel fraction of the devulcanized rubber and crosslink density of the gel component. The simulation is qualitatively correct in predicting the variation of gel fraction and crosslink density of devulcanized rubber with increasing ultrasonic amplitude. However, the theory predicted a lower degree of devulcanization at low ultrasonic amplitudes compared with the experimental observation157. The simulation also showed an opposite trend for all types of crosslink concentration compared with experimental results. This contradiction was due to the complex chemical transformation of polysulfidic crosslinks into di- and monosulfidic crosslinks, which was not accounted for in the theoretical modeling. In addition, the numerical simulation of continuous ultrasonic devulcanization was based on the cavitation collapse model119 and was in disagreement with experimental data157, 161. The disagreement was caused by the incapability of the cavitation collapse theory to describe devulcanization at high ambient pressure when no collapse occurs. This means the microvoids do not necessarily collapse in order to induce the bond breakage. Therefore, the experiments revealed that acoustic cavitation collapse is not a necessary condition for the ultrasonic devulcanization.
2.9.3.2 Viscoelastic Model
It was found that viscoelasticity tremendously affects the dynamics of cavitation, causing a reduction in the amplitude of oscillations and to suppress violent collapse contractions inherent in the pure elastic solid119. Therefore, ultrasonic devulcanization based on a viscoelastic model162 was developed to describe the dynamics of cavitation.
Standard linear viscoelastic solid and Rouse viscoelastic solid models were used for modeling rubber elasticity in numerical simulations. Significant differences were 42 observed for the cavity pulsation between the standard viscoelastic solid and the Rouse solid. No collapse-like effects were revealed during cavitation in the Rouse solid at low ambient pressures. High amplitude stable oscillations around the dynamically swollen state were observed. It can be concluded that cavitation collapse is not the primary mechanism of the ultrasonic devulcanization. The degradation is possible if acoustic pressure is high enough to cause a high level of strain during cavitation. A simple model of an ultrasonic reactor was used to estimate the acoustic pressure. The analysis showed that very high acoustic pressure could be produced in a cured rubber layer compressed in a narrow gap between the ultrasonic horn and die. In the mean time, the concept of overstressed chains163 was introduced. It provided a simple way to estimate the number of chains involved in mechanochemical degradation at certain amplitudes of ultrasonic pulsation. The calculation resulted in a logarithmic dependence of the fraction of the overstressed chains on the concentration of the cavities and high amplitude cavitation with no collapse-like effects is able to reduce the crosslink density significantly.
Furthermore the Dobson-Gordon classical theory of rubber network statistics164,
165 was employed to interpret the experimental data of the dependence of gel fraction in the devulcanized rubbers on crosslink density in the gel. Based on the assumption of randomness of network breakage and main chain scission, it led to a fairly good agreement between experimental and theoretical data for all types of rubbers ultrasonically treated under the continuous process166. The simplest model based on the random ruptures of chains and crosslinks and adequately describes all the data by fitting the parameter kp/kα for each type of rubber, with kp and kα being the rate constants of the rupture of main chains and crosslinks, respectively. The curve fitting was achieved by 43 taking the least-square method. There were two limiting cases for this model in the continuous process: with kp=0 indicating only crosslink rupture and kα=0 indicating only main chain rupture. All the experimental data for the continuous process were located within these two limits. In addition, in the continuous process, the influence of filler was
166 also investigated . It was found that the ratio kp/kα increased with the increase of filler content. This indicates that the addition of filler promotes the rupture of main chains due to the reduced mobility of rubber chains resulting from certain physical-chemical interactions between the polymer and the filler.
2.9.3.3 Radical Depolymerization Induced by Thermal Degradation
As mentioned in the previous section, a simple model based on the Dobson-
Gordon theory of rubber network statistics was able to describe the experimental gel fraction and crosslink density in the continuous process. All of the experimental data were within these two theoretical limits. However, this is not the case for the static ultrasonic treatment156. The experimental data for the rubber exposed under the ultrasonic device without flow fell below the curve with kα=0. This surprising outcome was explained by the inclusion of radical depropagation167, 168, 169 which is triggered by a significant temperature buildup in the static process. It was postulated that five basic reaction steps were involved in static ultrasonic treatment170: (a) Random crosslink scission (kα); (b) Random main chain scission (kp); (c) Depropagation (kd); (d) Radical
st transfer with chain scission (ktr); and (e) 1 order radical termination (kt). It was assumed that the primary breakup of crosslinks (a) and main chain scission (b) was only initiated by ultrasound and therefore the rate constants kp and kα were independent of temperature and their ratio was taken to be the same value as the one found in the continuous process. 44 For simplicity, the temperature dependence of kt was neglected, and temperature dependence of kd and ktr were taken to have the Arrhenius form. The calculation results showed that experimental data could be described by the proposed mechanism of adding the thermal degradation. The normalized gel fraction and crosslink density of the gel, after averaging over the sample thickness and initial values, were found to have linear relationships. Compared with the continuous process, the static process led to a lower gel fraction at a given crosslink density. This can be explained by the thermal degradation caused by the high temperature buildup in the static ultrasonic process. The radical depropagation results in the elimination of dangling chains containing radicalized ends created by main chain scission127.
2.10 Molecular Mobility Analysis by Solid-State NMR.
NMR is the direct investigation of nuclear energy levels based upon the angular momentum of nuclei. It is an effective tool in the analysis of chemical structures of polymers and small molecules. However, due to the inability of three-dimensional vulcanized elastomers to dissolve in any solvents, many techniques were developed to examine them in the solid-state, thus gaining a better insight into the dynamics and morphology171, 172. Unlike solid samples that are rigid and exhibit wide spectral lines requiring the use of special techniques to try to obtain useful information, vulcanized elastomers exhibit liquid-like behavior, and when combined with these special techniques, result in favorable conditions of spectral resolution.
In the investigation of sulfur cross-linked elastomers, solid-state 1H NMR lacks adequate chemical shift resolution and focuses on molecular mobility and probing the relaxation of protons near the sulfur cross-links. The use of varied temperature can assist 45 to improve resolution and sensitivity, but the minor signals close to the sulfur cross-links still tend to be lost in the baseline noise or are overlapped by the stronger signals.
Nevertheless, 1H NMR is still a powerful tool in the study of molecular dynamics and has an abundant amount of information to offer172, 173, 174, 175, 176.
1 2.10.1 Proton Transverse Relaxation ( H T2)
The complete exponential decay of magnetization is determined by two relaxation factors: the spin-lattice or longitudinal relaxation rate T1, and the spin-spin or transverse relaxation rate T2. T2 relaxation concerns itself with the transverse magnetization decay.
This magnetization dephases in the x-y plane due to the differences in the individual precession (a comparatively slow gyration of the rotation axis of a spinning body about another line intersecting it so as to describe a cone) frequencies of the adjacent nuclei.
The adiabatic energy transfer takes place between the adjacent spins, thus called spin- spin relaxation. Spin-spin relaxation rates provide information about local fields experienced by the nuclei, which relate to the structure and nature of the environment.
Liquids with increased motions average the majority of local magnetic fields to zero, thus producing a long T2. In contrast, solids have short T2 values due to their ability to rapidly lose coherence in the presence of spatially fixed neighbors.
When observing nuclei of elastomers by solid-state NMR, there are two important factors that regulate relaxation rates: molecular motion and magnetic interaction between nuclei. Relaxation in elastomers predominantly occurs through segmental motion resulting in a randomly varying field at the observed nuclei. Other contributions to these local fields arise from spin rotation, scalar coupling, quadrupoles (for I>1/2 nuclei only) and chemical shielding anisotropy. 46 2.10.2 Pulsed Gradient Spin-Echo (PGSE) Diffusion
Elastomers, either in the solution or in the melt state can change position and shape randomly by thermal agitation. This Brownian motion dominates various time- dependant events such as viscoelasticity and diffusion. Of these phenomena, in which the effect of Brownian motion appears, diffusion is the most visible. The process of diffusion is described by Fick’s first law177, which has the one-dimensional form
∂c ∂ 2 c = −D (2.3) ∂t c ∂x 2
178 where c is the concentration of the diffusing species and Dc is the diffusion constant . In three dimensions, we have
∂c = −D ∇ 2 c (2.4) ∂t c
∂ ∂ ∂ where ∇ = i + j + k ∂x ∂y ∂z
In experiments measuring self-diffusion, there is no need for a net concentration gradient.
By replacing the concentration gradient with the probability P of finding a particle with displacement r’ at time t, we obtain
∂P = D∇ 2 P (2.5) ∂t which is referred to as Fick’s second law, with D representing the molecular self- diffusion constant. The probability function is not dependent on starting position, but rather on net displacement (r’-r), and it is this displacement which pulsed-gradient spin- echo (PGSE) NMR measures. Experimentally, based upon the simple spin-echo pulse sequence shown in Figure 2.8179, the effects of diffusion can be seen and described by
47 A(2τ ,G ) 2 ln 0 = − γ 2 DG 2τ 3 (2.6) A(2τ ,0) 3 0
where A is the echo height as a function of time, G0 is the constant gradient and τ is the time between pulses. The echo or spin-echo, is the result of the rephasing of the spin vectors. After the initial 90o pulse, the transverse magnetization in the x-y plane decays during the first τ delay, due to the loss in phase coherence. The following 180o pulse flips the spins causing them to refocus at a time of 2τ.
Figure 2.8 Effect of diffusion on the Hahn spin echo179
To amplify the effects of diffusion on echo attenuation, large pulsed field gradients G are applied as shown in Figure 2.9, typically in addition to the steady gradient G0. This method is referred to as pulsed-gradient spin-echo (PGSE) NMR and results in an attenuated spin echo with amplitude described by the Stejskal and Tanner expression180
A(2τ , X ) = exp(−γ 2 DX ) (2.7) A(2τ ,0) where X takes into account the parameters and timings of the gradients,
δ δ 2 = δ 2 2 ∆ − − δ θ 2 + 2 + δ + + − τ 2 X G GG0 (cos )(t1 t 2 ) (t1 t 2 ) 2 2 (2.8) 3 3
48 t1
Figure 2.9 The pulsed-gradient spin-echo (PGSE) sequence
49
CHAPTER III
3.EXPERIMENTAL
3.1 Materials
Synthetic isoprene rubber (IR) used in this research is NATSYN®2200. It has a cis 1, 4-isoprene content of 98.0% and the number-average molecular weight Mn is
810,000 while the weight-average molecular weight Mw is 2,490,000 (provided by
Goodyear and measured by Thermal Field Flow Fractionation - ThFFF). NATSYN®2200 is polymerized following the coordination mechanism with a Ziegler-Natta catalyst in solution and it is stabilized with a non-staining antioxidant. As claimed by Goodyear,
NATSYN®’s brand has many advantages including:
• Ease of processing and improved flow
• Tight dimensional tolerances
• Uniform cure rate
• No contamination or proteins
• Reproducible physical properties
• Clean, light color
• Does not require premastication
• May be one-pass mixed
50 • Low water swell
• Low creep and compression set
The information for all the materials used including the compounding ingredients sulfur, N-cyclohexylbenzothiazole-2-sulphenamide (CBS), zinc oxide (ZnO), stearic acid, processing oil, retarder and carbon black as well as the solvents benzene and tetrahydrofuran (THF) are listed in Table 3.1.
Table 3.1 Materials used in this study
Material Symbol Grade Source
Synthetic isoprene rubber IR NATSYN®2200 Goodyear Tire & Rubber
Company
Sulfur S Rubber maker Akrochem Corporation
Zinc oxide ZnO Powder form Akrochem Corporation
N-cyclohexylbenzothiazole-2- CBS Powder form Akrochem Corporation
sulphenamide
Stearic acid Bead form Akrochem Corporation
Carbon black CB HAF, N330 Sid Richardson Carbon,
Fort Worth, TX
Processing oil Plasticizer Naphthenic oil of Akrochem Corporation
LN low viscosity
Retarder SAFE Powder form Akrochem Corporation
Benzene Analytic grade Aldrich Chemical
Company
Tetrahydrofuran THF Analytic grade Aldrich Chemical
Company
51 3.2 Compounding
Compounding is the process of formulating rubber and the additives for the final application. Both the unfilled and carbon black filled synthetic isoprene rubber samples were used. The recipe for compounding has been taken from previous research181, 182 for natural rubber and is given in Table 3.2.
Table 3.2 Compounding recipes for the vulcanization of IR
Materials Unfilled system (phr) Filled system (phr)
IR 100 100
CB - 15-60
Plasticizer LN - 0 or 10
ZnO 5 5
Stearic Acid 1 1
CBS 1 1
Sulfur 2 2
For unfilled IR, in the case of a small amount of rubber ( ~ 100 g), compounding was done in a two-roll mill (Dependable Rubber Machinery Co., Cleveland, OH) at ambient temperature. The shear force of the rubber was controlled by the gap between the two rolls and the ratio of the roll speed. A gap of 3 to 5 mm was used. The front roll speed was 14 mpm (meters per minute) and the back roll speed was 11 mpm and the roll diameter and length are 15 cm and 30.5 cm, respectively. The gap setting determines the thickness of the sheet that passes between the rolls. The higher the ratio of the roll speeds, the greater is the shearing force and more intensive is the mixing. 52 Before adding the cure ingredients, gum IR was experienced 3 to 5 passes in the two-roll mill. After the incorporation of ZnO, stearic acid and CBS, sulfur was the last to be added. The total processing in the two-roll mill was about 40 passes. In the case of a large amount of rubber (e.g., 2000 g), compounding was made in a 3L Moriyama D 3-7.5 internal mixer (Osaka, Japan). ZnO, stearic acid and CBS were added in the internal mixer containing gum IR and mixed for 7 minutes. Sulfur was then added into the mixer and mixed for an additional 3 minutes.
In the case of the filled rubber, the mixing of carbon black and gum IR was done in a 1200 cc, water-cooled counter rotating Farrel Banbury internal mixer (Model
86EM9804, Banbury USM Corp., Ansonia, CT) at ambient temperataure. The rotor speed during mixing was set at 30 rpm. The mixing time was 10 minutes. In the case of compounds containing processing oil, the oil was added together with the carbon black and the mixing conditions were the same as for the compounds without oil. The carbon black agglomerates break down only upon reaching a certain yield stress and this type of mixing is called dispersive mixing. Dispersive mixing in polymer processing involves the rupture of clumps and agglomerates of solid particles like carbon black and pigments in a deforming viscous liquid. It is accomplished by forcing the mixture to pass through high shear zones generated in narrow clearances such as the gap between the rolls of a roll- mill or in the clearance between the blades and the shell in internal mixers. The filling volume was about 80% of the total volume. The mixture of gum IR and carbon black was compounded with the curatives, in the same order as in the unfilled system, on a two-roll mill. It needs to be cautious that during mixing the temperature of the material is as close to room temperature as possible. This will prevent any premature curing of the material. 53 3.3 Vulcanization and Vulcanizates Grinding
Once the compounding process is finished, the rubber has to be cured. The curing time is taken to be the time at which the maximum torque occurred, measured by the
Advanced Polymer Analyzer (APA2000) (Alpha Technologies , Akron, OH), and is
o denoted as Tmax. A strain amplitude of 4.2% (0.3 ) and a frequency of 100 cpm were used in the curing experiment done by APA 2000.Various curing temperatures ranging from
110oC to 180oC were used in the cure kinetics study. For the devulcanization study, the curing and the revulcanization were completed at 160oC.
Vulcanization takes place when the compounded rubber is subjected to a particular temperature and pressure. This process involves three stages which are shown in the cure curves. The initial stage is the induction period, representing the time when no significant crosslinking occurs. In the beginning, the compounded material is stiff, so the curve shows a high initial torque value. Upon heating, the rubber softens and the resistance to the oscillation of the rotor decreases and the torque value decreases to a minimum. Following this, the crosslinking or curing occurs at a rate dependent on temperature, type of rubber and type of curatives. This is indicated by a relatively steep increase of torque. The crosslinking makes the rubber stiff and the torque increases. As the curatives and the number of crosslink sites are consumed, the curing reaction slows down as it goes to completion and the optimum time is reached at the onset of maximum torque. This represents a full cure. When vulcanization is complete, further heating may cause an overcure leading to what is named “reversion”. Reversion is caused by the irreversible destruction of the polysulfidic crosslinks183, 184.
54 The bulk samples were vulcanized in a compression molder (Model 20-1212-
2TMB, Wabash Corp., Wabash, IN) into slabs (260 × 260 × 12 mm3) at a temperature of
160°C and a pressure of 27.6 MPa (6000 psi) for the time of Tmax, predetermined in the
APA 2000. After molding, the vulcanizates were placed in the freezer for 24 hours and then ground into particles with a 5 mm sieve in a Nelmor grinding machine (01012M, N.
Uxbridge, Massachusetts).
3.4 Revulcanization
The devulcanized rubber (devulcanization will be described in Section 3.5.2) was compounded with curatives in the two-roll mill according to the recipe shown in Table
3.3. A gap of 3 to 5 mm was used. The front roll speed was 14 mpm (meters per minute) and the back roll speed was 11 mpm and the roll diameter and length are 15 cm and 30.5 cm, respectively. Devulcanized rubbers were firstly processed in the two-roll mill for 3 to
5 passes, then zinc oxide and stearic acid were added, and finally the sulfur. The total mixing took about 40 passes. Revulcanization was carried out in the Wabash compression molding press using a mold with dimensions of 127 × 127 × 2 mm3. Curing time was again taken as the time to reach the maximum revulcanization torque measured by APA 2000.
3.5 Ultrasonic Treatment and Devulcanization Equipment
The ultrasonic treatment of virgin IR and the ultrasonic devulcanization of vulcanized rubber were both carried out in a coaxial reactor. Generally it is a single screw rubber extruder attached with a set of high power ultrasound supply. The schematic drawing of the reactor was shown in Figure 3.120. The screw has a L/D ratio of 11 and D is 38.1mm. The cone-tipped ultrasound horn of 76.2 mm diameter was mounted coaxially 55 to the extruder die. The convex tip of the horn matches the concave surface of the die, so that the clearance between the horn and the die is uniform. A uniform die gap of 2.54 mm
(0.1 inch) was used in all the experiments. The flow rate was set between 0.47 and 2.55 g/s corresponding to 21.3 and 3.9 seconds of residence time in the devulcanization zone.
Table 3.3 Compounding recipe for the revulcanization of the devulcanized IR
Materials Unfilled or filled system (phr)
Devulcanized IR 100
ZnO 2.5
Stearic Acid 0.5
Sulfur 2
SAFE 0 or 1
A schematic drawing of the effective volume of the die opening is given in Figure
3.2158. The barrel has three temperature controll zones equipped with electrical heaters and fans. The ultrasound unit is composed of a 3.0 kW ultrasonic power supply, an acoustic converter and a 1:1 booster. The water-cooled horn vibrates longitudinally with a frequency of 20 kHz and amplitude ranging from 5 to 10 microns. The rubber particles from the extruder flow into the gap between the horn and the die plate where the rubber was subjected to the longitudinal (compressive) wave perpendicular to the flow direction.
A flush-mounted thermocouple and a pressure gauge were inserted into the barrel to measure the temperature and the pressure of the rubber at the entrance of the die. The ground rubber particles are fed into the extruder by a conveyor belt feeder (FA 50/500
VBR, Germany) with an adjustable output. Starve feeding to the barrel was applied so 56 that the flow rate of the rubber particles was controlled by the feeding rate. The ultrasonic energy consumed during the experiment was measured by a wattmeter (Model A410A,
Branson Ultrasonic Corporation, Danbury, CT) attached to the ultrasound unit. This reactor was built by the National Feedscrew and Machining (NFM), Inc., in a co- operation with our laboratory.
ULTRASONIC FE E DER POWER SUPPLY WAT TM E T E R
CON V E R T E R EX T R U D E R
TEMPERATURE AND PRESSURE GAUGES DIE HORN BOO S T E R
Figure 3.1 Schematic drawing of the coaxial ultrasonic reactor
Both the die and the horn have sealed inner cavities for cooling water. The die and horn are cooled down with tap water in order to reduce the heat build-up caused by the dissipation of the ultrasonic energy in the rubber so that the degradation of rubber due to high temperature can be minimized. The cooling water flow rate for both the die and the horn was 0.09 m3/hr.
57
D1: Die entry diameter, 25.4 mm
D2: Horn diameter, 76.2 mm
Dh: Diameter at constant gap clearance, 38.1 mm
d: Die gap
V: Effective volume
Figure 3.2 Schematic drawing of the effective volume of the die opening
3.5.1 Ultrasonic Treatment of the Virgin Gum IR
Virgin gum IR was preprocessed in a two-roll mill first in order to control the thickness at 3 mm and then it was manually cut into long strips leaving a 1cm width. The strips were fed into the ultrasonic treatment reactor (Figure 3.1) at a flow rate of 0.63 g/s.
The screw speed was set at 17 rpm. Above this speed, the conveying of gum IR into the extruder was not possible. The temperature of the extruder barrel and the ultrasonic attachment was set at 120°C. The ultrasonic treatment was carried out at a frequency of
20 kHz and varying amplitude of 5, 7.5, and 10 µm, with a 3.0 kW power supply. The gap between the horn and the extruder die was 2.54 mm. The treated rubber was collected and cooled in freezer overnight and it was ready for the characterization.
58 3.5.2 Devulcanization of the Vulcanizates
The ground IR vulcanizates were loaded into the hopper. The feeder, providing
“starved feed” to the extruder, controlled the output. The gap between the horn and the extruder die was 2.54 mm. In the extruder, rubber vulcanizates were compressed and conveyed by the screw to the devulcanization zone. The devulcanization of the rubber occurred in the gap between the horn and the extruder die in the reactor. After reaching the steady state conditions indicated by the pressure transducer and the ultrasonic power wattmeter, the devulcanized rubber sample was collected. The entrance pressure before the devulcanization zone and the ultrasonic power consumption were measured. The devulcanized rubber exiting from the reactor was collected and cooled in the freezer overnight.
3.6 Characterization Methods
A wide variety of characterization methods are applied in this study. This includes measurements of cure kinetics of the virgin and devulcanized IR, dynamic properties of the virgin uncured, cured and devulcanized IR, network structures (gel fraction and crosslink density) of the devulcanized and revulcanized IR, molecular weight determination of the treated gum IR and the sol part of the devulcanized unfilled IR, molecular mobility of the treated gum IR and their vulcanizates as well as the unfilled and filled IR, thermal properties of the virgin and devulcanized IR, and mechanical properties of the cured and revulcanized IR. These methods are combined together in order to characterize the devulcanized IR.
59 3.6.1 Vulcanization Kinetics
Vulcanization behaviors of the virgin and the devulcanized samples were studied by means of the APA 2000. The resistance to oscillation was measured by the complex torque and recorded on the rheometer chart as a function of time and the Tmax values were obtained accordingly.
3.6.2 Dynamic Properties
Dynamic tests of the original uncured, cured, and devulcanized samples were carried out using the APA 2000 at temperatures of 60, 90 and 120oC within a frequency range of 0.02 – 200 rad/s and a strain amplitude of γo = 0.042 (0.3 degrees). Figure 3.3 shows the biconical die cavity used in the APA 2000185. The sample was covered with a thin Nylon sheet at both sides and was placed on the lower die. The upper die was then brought down squeezing out the excess sample. The lower die was oscillated sinusoidally at a specified frequency and strain programmed into the computer. The upper die was attached to a torque transmitter to record the torque evolution with time.
Figure 3.3 Biconical die sample cavity185
The die dimensions, die gap, frequency of oscillation and strain were all used with appropriate rheological equations to convert the torque to shear modulus and dynamic
60 viscosity. A Fourier Transform of the torque and strain data separates the complex torque signal into an elastic and viscous component. These torque signals were converted to G’,
G”, η’ and η”. The recorded torques (S*) were directly converted to the shear modulus by multiplying the appropriate die form factor and divided by the strain as given by the following equation:
2πR3 S * = G *θ (3.1) 3Φ 0 where S * is the measured torque values, R is the die radius (=20.63 mm), Φ is the angle
(= 0.1251 radians) between the cone and the plate, G * is the absolute value of the
complex shear modulus of the sample and θ0 is the angular amplitude. Plugging in the above geometric parameters, Equation (3.1) is reorganized to obtain G * :
S * G * = (3.2) []−4 3 θ 1.47x10 m / rad 0
The measured torque is divided by 1.47 x 10-4 m3/rad and by the strain θ (in radians) to give |G*|. The strain amplitude is given by:
θ γ = 0 (3.3) o Φ
For the APA 2000, Φ = 2 x 3.58o where the cone angle = 3.58o. Thus for a movement of
1o, the strain amplitude in the lower die is:
1o γ = = 0.140 (3.4) o 7.16o
In the experiment, γo = 0.042 (θ0 = 0.3 degrees).
Accordingly, the shear rate amplitude is given by: 61 • . γ = γ ω o (3.5) where ω is the frequency of oscillation.
3.6.3 Gel Fraction
Typically a vulcanized rubber sheet contains a large amount of gel, which is insoluble in any solvent and some amount of sol, which is soluble and can be extracted out of the network. The gel fraction of the virgin, vulcanized, devulcanized, and revulcanized samples were measured by the Soxhlet extraction method using benzene as the solvent for isoprene rubber. A Kimax extraction set (Fisher Co.) was used. As shown in Figure 3.4, the Soxhlet extraction apparatus consists of an extractor, a condenser, a glass flask and a heater. Approximately 2-5 grams of sample (weighed accurate to 0.1 mg) was put into a Whatman cellulose extraction thimble in the extractor. Benzene in the flask under the extractor was heated and vaporized and the vapors were condensed at the top of the apparatus. The total extraction time was 24 hours. After that, the weight of the swollen sample was measured after removing the free solvent on the sample surface using a clean tissue paper. Then the sample was dried in an oven at 65°C for 24 hours and then allowed to stay at room temperature for another 24 hours before the weight was taken again. The gel fraction, g, can be determined from the following equation:
weight of dry gel after extraction g = (3.6) weight of rubber sample before extraction
3.6.4 Crosslink Density
The crosslink density was characterized by the average molecular mass of the network chain, i.e., the average chain length between two network junctions. This may be expressed by the average molecular weight between crosslinks (Mc) or the number of 62 network chains per unit volume of gel (ξ). Generally, the crosslink density of the vulcanizate is measured by the swelling technique or by using the Mooney-Rivlin plots186, 187. In our case, the latter was not viable for the devulcanized samples due to the inability to perform mechanical testing on the ground samples.
Figure 3.4 Soxhlet extraction apparatus
A crosslinked polymer absorbs a finite amount of a solvent, depending on the nature of the polymer and the solvent. The swelling process is controlled by the entropy of the dilution of the polymer/solvent caused by the polymer chains assuming elongated configurations, but constrained by the crosslinks.
The crosslink density and gel fraction of the virgin, vulcanized, devulcanized, and revulcanized samples used in this study were determined from equilibrium volume 63 swelling in benzene. Samples of each test piece was preweighed and then placed in a
Soxhlet extractor containing benzene and allowed to swell for 24 hours. After extraction, weights of the swollen wet samples were recorded after removing the excessive solvent, drying in an oven at 65oC for 24 hours and then allowed to stay at room temperature for another 24 hours. In this procedure, the handling of specimens was extremely critical.
Therefore, care was taken to ensure that the specimens were thoroughly dried before weights were recorded. The crosslink density was determined according to the Flory-
Rehner equation188:
ln() 1−++VVχ V2 ξ ()g =− rr r (3.7) 1/3 − VV1 ()rr V/2
(g) where ξ is the effective number of chains in a real network per unit volume of gel, V1 is the molar volume of the solvent (the molar volume of the solvent is the weight of the solvent/density of solvent), Vr is the volume fraction of the polymer in the swollen network in equilibrium with the pure solvent and χ is the interaction parameter between the solvent and the polymer. Benzene has a value of V1= 88.838 cc/mole. The specific gravity of the isoprene rubber is 0.92 and the interaction parameter χ has been taken as
0.42189. The volume fraction of the rubber network in the swollen phase was calculated from the equilibrium swelling data:
weight of dry rubber = density of rubber Vr (3.8) weight of dry rubber + weight of solvent density of rubber density of solvent
In order to find out how long the ground samples need to be kept in air in order to ensure that the proper evaporation of the free solvent takes place, the evaporation time
64 had to be found. Approximately 1 gram of swollen sample was removed from the thimble after extraction. It contains the excessive solvent trapped in the granules, which can cause errors in the measurement of crosslink density and it had to be removed. For a thin sheet
(about 3 mm) sample, the excessive solvent stays on the surface and can be easily removed by a piece of tissue paper. For the granules of virgin vulcanized or for the gel of the devulcanized sample, the excessive solvent could not be completely removed. Hence it was assumed that the ground vulcanized particles had the same crosslink density as they were in the sheet form. The crosslink density of ground samples was measured as a function of the solvent evaporation times in air. Figure 3.5 depicts the crosslink density of ground samples as a function of evaporation time for the unfilled IR as an example showing how to determine the evaporation time. Based on the fitted equation and the crosslink density of the vulcanized rubber sheet, the evaporation time for the unfilled ground rubbers was experimentally determined to be 4 hours. Similarly, the evaporation time for the 15phr, and 35phr filled rubber were found out to be 2 hours and 38 minutes, respectively. So the wet samples were not weighed until the evaporation time was reached after they were taken out of the Soxhlet apparatus.
In reinforced filler filled rubber systems, the Flory-Rehner equation can not be directly used. Due to the presence of filler, the Flory-Rehner equation was modified using the Kraus Correction Factor67:
V mΦ ro = 1− (3.9) − Φ Vr (1 )
= ()− 1/3 + − m 3C 1 Vro Vro 1 (3.10)
65 volume of rubber gel V = (3.11) r volume of rubber gel + volume of solvent where, Vro and Vr are the volume fraction of rubber in the swollen gels of unfilled and filled vulcanizates, C is a universal constant for a given filler, and Φ is the volume fraction of the filler in the dry gel. In using this equation, it was assumed that the filler does not swell in the solvent. Also, the greater the value of C, the less swelling there will be. The value of C was taken to be 1.17 for HAF (high-abrasion furnace) carbon black67.
The detailed procedure for calculating the gel fraction and crosslink density are described elsewhere182.
300 unfilled y = 0.512x + 115.163 R2 = 0.998
250 sheet sample 3 200
150
100 Crosslink density, mol/m
50
0 4 hours 0 50 100 150 200 250
Evaporation time, min
Figure 3.5 Determination of the evaporation time for the ground unfilled rubber
66 3.6.5 Mechanical Properties
The mechanical properties of the virgin vulcanized and revulcanized samples were measured at an elongation rate of 500 mm/min using an Instron 5567 tensile test machine (Canton, MA) with a 1kN load cell at room temperature following ASTM D
412-92 (Type C). Type C dumbbell samples were punched out from the compression molded rubber sheets (180×130×2 mm3). At least five samples of a gauge length of 65 mm, thickness of 2 mm, and the width of 3 mm were used for each measurement. From the force and displacement, the engineering stress-strain curves were constructed and then the 100% and 300% modulus, the tensile strength and the elongation at break were obtained. The results of the virgin vulcanizates and revulcanizates were compared. The engineering stress is defined as:
F σ = (3.12) A0 where F is the instantaneous force and A0 is the initial cross-sectional area of the sample.
The elongation, ε, is defined as:
l − l ε (%) = 0 ×100 (3.13) l0 where l is the observed distance between the benchmarks on the extended specimen and l0 is the extensometer gauge length.
If measured accurately, the slope at the origin is directly related to the rigidity of the material. Unlike metals or most plastics, rubber does not exhibit a yield point so the stress continues to increase until rupture takes place. For stress-induced crystallizing materials such as isoprene rubber, the slope of the stress-strain curve becomes very high
67 near the breaking point. The ultimate elongation value is an indication of fatigue life. A compound with higher elongation at failure will give better fatigue resistance. The strength of the material is its ability to carry load and is equal to the breaking or ultimate stress.
The main change in the stress-strain behavior of rubber due to the incorporation of carbon black is observed with regard to stiffness which is scaled with modulus, i.e. the ratio between stress and strain. In rubber technology, for quasi-static measurements, it is common to refer to the stress at a particular strain as the “modulus”49. The incorporation of reinforcing carbon black is a major source of energy dissipation thereby raising the strength of the filled rubber. Certain mechanisms for the reduction of stress magnification at the tips of inherent flaws in black filled materials have been proposed190.
Rubber in tension generally fails at a flaw. The flaw may be caused by the die used to click out the specimens, porosity, bubbles, inclusion of foreign matter, by discrepancies encountered during compression molding. So even though there may be one rubber sheet, its composition may not be uniform at each point due to these flaws. So different specimens will have different degrees of flaws and similarly the breaking of the tensile test samples will also occur randomly.
3.6.6 Molecular Weight Determination
The molecular weight (MW) and molecular weight distribution (MWD) of ultrasonically treated and untreated gum isoprene rubber was carried out by GPC (Gel
Permeation Chromatography) using a Waters 410 Differential Refractometer, a Waters
486 Tunable Absorbance Detector and a Waters 510 HPLC pump. THF solvent was run at room temperature and conventional calibration was used against the polystyrene 68 standards. The range of light intensity for the Absorbance Detector was between 196-600 nm with the primary use in the UV range (190 – 380 nm). The sensitivity range was
0.001 – 2.0 absorbance units full scale. The pump flow rate was set at 1ml/min. The specifications of the optical component was a path length of 10 mm (standard, analytical) and a cell volume of 8 µL (standard, analytical). Experiments were carried out on the gum (virgin and treated with ultrasound, which are directly soluble in THF) and on the sol part of the devulcanized samples of rubber obtained after extraction with THF for 96 hours in the Soxhlet apparatus.
3.6.7 Thermal Properties
A Dupont 2100 Differential Scanning Calorimeter (DSC) was used to determine the glass transition temperature of the ultrasonically treated rubber and their correspondent vulcanizates. The instrument was calibrated with indium. Before running, the instrument was cooled down below –100oC using liquid nitrogen. Then the experiment ran from –100 to 100ºC at the heating rate of 10ºC/min under nitrogen atmosphere. The sample weight was between 5 to10 mg.
The thermal stability of the ultrasonically treated rubber and their vulcanizates was evaluated using a Dupont 951 Thermogravimetric Analysis (TGA) at a heating rate of 20ºC/min, from room temperature to 600ºC under nitrogen atmosphere. Each time a 10 to 15 mg of sample was used.
3.6.8 T2 Relaxation and PGSE Diffusion
In this research, the T2 relaxation and diffusion measurements of ultrasonically treated gum IR and devulcanized IR were conducted by solid-state NMR at 70.5oC. The elevated temperature was chosen to accelerate molecular/segmental motions and 69 accentuate the differences between the gel and sol of rubbers. The 33 MHz modified
Spin-Lock CPS-2 spectrometer and the data interpretation methods were comprehensively described elsewhere191, 192. The pulse sequence employed for the relaxation experiments was the standard principal Hahn two-pulse sequence conducted on-resonance, 90o-τ-180o-τ-echo while for the Pulsed-Gradient Spin-Echo (PGSE)
o o o experiments the stimulated spin echo sequence 90 -τ1-90 -τ2-90 -τ1-echo was used off- resonance. The pulse spacing τ in the T2 experiments was adjusted from 0 to at least 60 ms in 30 or more steps, until the signal-to-noise ratio fell below 0.2% of its initial value.
The PGSE experiments were performed at a fixed magnitude of the pulsed magnetic field gradient, G, by varying the duration δ of each of the pair of gradient pulses coordinated with the stimulated-echo radio frequency (rf) pulse sequence. The experiments were conducted in Dr. Ernst D. von Meerwall’s laboratory at the Physics department.
70
CHAPTER IV
4.CURE KINETICS STUDY OF THE UNFILLED AND THE CB FILLED IR
4.1 General
Even though sulfur vulcanization of elastomers was discovered two centuries ago, the exact cure reaction mechanism is still not fully understood today due to the complicated nature of cure chemistry. Sulfur cure reaction involves the complex mixtures of elastomers, sulfur, accelerators, activators and retarders. Accordingly, the reaction produces a wide variety of reaction intermediates and final products. As a result, a complete understanding of the vulcanization process is quite challenging for rubber chemists. Nevertheless, there is still considerable research involved in this field. There are a number of phenomenological models that have been developed today for modeling the sulfur vulcanization. The commonly used model is the nth order kinetic model193:
dα = k(1−α) n dt where α is the state of cure and is a measure of the conversion of reactives such as sulfur, k is the rate constant and n is the order of reaction. More complicated models have also been developed. The most popular one is the Kamal-Ryan model194:
dα = (k + k α m )(1− α) n dt 1 2
71 This model relates the state of cure to two rate constants k1 and k2 and two adjustable parameters m and n. If one sets k1 to zero, the Kamal-Ryan model is then reduced to the model proposed by Piloyan195:
dα = kα m (1− α) n dt
Thereafter, Isayev and Deng196, 197 proposed another kinetic model:
dα n − + dα − + = α 2t (1 n) and = nk1/ nα (n 1) / n (1− α)(n 1) / n dt k dt
They were derived from the following equation proposed by Kamal et al198 by taking the derivative of α with respect to the cure time t:
kt n α = 1+ kt n
All of above proposed kinetic models accurately describe the curing portion of the vulcanization process; however these models poorly curve-fit the reversion stage of the process. Recently Han et al199 proposed a cure kinetic model that can predict the reversion and induction period commonly occurred in the vulcanization of many rubber compounds. Their model was based on the phenomenological study of cure reversion at high temperatures investigated by Bielstein183, Shankar184 and Rimondi200 earlier. In our previous research, this model was used for modeling GRT and GRT/SBR blends201. The prediction of the evolution of state of cure resulted in a reasonable good agreement with the experimental data. The slight departure from the prediction was probably due to the complexity of the system in GRT. In this study, Han’s model was applied to a single rubber component system (IR) and its applicability to the unfilled and the carbon black filled IR was examined.
72 The purpose of cure kinetics modeling is to quantitatively predict the evolution of the state of cure during the isothermal and non-isothermal vulcanization process and to make a comparison of the predicted state of cure with the experimental data as measured from the cure rheometer APA 2000. This kinetic modeling is particularly suitable for the reversion type of cure reaction and it provides the needed assistance to optimize practical processing conditions such as the temperature and the time in the vulcanization.
4.2 Cure Kinetics Experiments
Before running the APA 2000 for the cure kinetics experiments, IR premixed with or without carbon black was compounded with cure ingredients. Compounding process for the unfilled and the black filled rubbers was described in Section 3.2. Around 4 to 5 grams of the premixed compounds were used in the measurement. In isothermal cure, the instrument was preheated to the desired curing temperature ranging from 130 to 180oC for the unfilled IR and 110 to 180oC for the filled IR. In non-isothermal cure, the instrument was preheated to the initial temperature such as 80 or 120oC. In order to obtain a uniform heating rate, the constant temperature increase was set up at each step
(for example, 4oC interval) and the value of heating rate was determined from the actual temperature history. When the desired temperature was reached, the sample was placed between the two thin sheets of Nylon film and loaded into the biconical die cavity (Figure
3.3) and the curing experiment was initiated immediately. The change in torque versus time during curing was automatically recorded in the instrument.
73 4.3 General Kinetic Model Equations
It was found that reversion during high temperature vulcanization was due to the irreversible destruction of polysulfidic crosslinks. Therefore the sulfur cure reaction with reversion was proposed to occur according to the following route183:
k1 c1
s k3 p k2 c2
Figure 4.1 Sulfur reaction scheme (s: sulfur; c1 and c2: stable and unstable sulfur-
containing crosslink, respectively; p: product of reversion reaction)
In Figure 4.1, k1 and k2 are the kinetic constants responsible for the formation of the stable and the unstable crosslinks, respectively; and k3 is the kinetic constant for the reversion reaction. The overall cure curve resulted from the summation of these three concurrent first order reactions: formation of stable crosslinks (k1), formation of unstable crosslinks (k2) and destruction of unstable crosslinks (k3). In addition to the assumption of the first order reaction, it is also assumed that the destruction of polysulfic crosslinks is an irreversible reaction and it is responsible for the reduction of torque in the post cure stage.
The initial amount of the sulfur present in the system is defined as s0. The instant amount of the sulfur, the stable and the unstable sulfur-containing crosslinks present in the reaction system are defined as s, c1 and c2, respectively. The relative amount of each species in the system, at any instant time is defined as199
74 cc C ==s αα12 = s , 12 , s00ss 0
α α = α + α Accordingly, the state of cure is defined as 1 2 .
Han et al199 proposed that the change of kinetic constants and isothermal induction time with temperature followed the Arrhenius type of relationship:
E kk=−expi , i = 1, 2, 3 (4.1) ii0 RT
E tt= expti (4.2) ii0 RT where ki0 and ti0 are the pre-exponent factors; Ei and Eti are the activation energies of the reaction; and R is the universal gas constant.
Based on the reaction scheme shown in Figure 4.1 and the assumption of the first
≥ order reactions, the cure kinetics equations after the induction period (t ti ) are expressed as:
dsC =−()kks + C (4.3) dt 12
dα 1 = ksC (4.4) dt 1
dα 2 =−ksC kα (4.5) dt 232
====αα α ≤ The initial conditions are: s 1,12 0, 0 for t ti
Equations (4.1) to (4.5) constitute the governing equations for the cure kinetics modeling. Particularly (4.3), (4.4) and (4.5) are in the differential form and have to be solved for both isothermal and non-isothermal conditions.
75 The experimental value of α with the reversion type of reaction was determined as199:
Γ (T ) − Γ (T) α(T,t) = t min (4.6) Γ − Γ max,0 min,0
Γ Γ where min (T) and t (T ) are the minimum and current torques at experimental
Γ − Γ temperature T, respectively. max,0 min,0 is the difference of the maximum and the minimum torques below a certain temperature T0, where reversion is negligible.
4.3.1 Isothermal Cure Kinetics
The solution for (4.3) under the isothermal condition is:
=−++[] s exp (kktC12 ) (4.7)
~ = = =+ C is the constant. Applying the initial conditiont ti , s 1, we haveCkkt()12i .
C =−+−[] So, s exp (kktt12 )(i ) (4.8)
Then (4.4) becomes:
dα 1 =−+−kkkttexp[] ( )( ) (4.9) dt 112i
The solution for (4.9) is:
k α =−1 exp[] − (kktt + )( − ) + C (4.10) 112+ i kk12
k Applying the initial condition: t = t , α = 0 ,, we have C = 1 . As a result, i 1 + kk12
k α =−−+−1 {}1exp([]kktt )( ) (4.11) 112+ i kk12
Accordingly (4.5) becomes:
76 dα 2 =−+−−kkkttkexp[] ( )( ) α (4.12) dt 21232i
(4.12) is the 1st order non-ordinary linear differential equation. To solve (4.12) involves two steps: the first step is to solve the correspondent 1st order linear ordinary differential equation (ODE):
dα 2 =−k α (4.13) dt 32
The solution for (4.13) is
α =−−[] 23Ckttexp (i ) (4.14)
In order to solve (4.12), replace the constant C in (4.14) with C(t):
α =−−[] 23Ct()exp k ( t ti ) (4.15)
Take the derivative of the above equation: ddCα 2 =−kCt()exp[][] − k ( t − t ) + exp − k ( t − t ) , and couple with (4.12) by plugging dt33ii 3 dt
(4.15) into (4.12): ddCα 2 =−kCt( )exp[][] −−+ ktt ( ) exp −− ktt ( ) = k exp[][] −+ ( k k )( tt −− ) kCt ( )exp −− ktt ( ) dt33ii 3 dt 21233 i i and it further develops into:
dC =−−−kkkkttexp[] ( )( ) (4.16) dt 2312i
k The solution isCt()=−−−+2 exp([] k k k )( t t ) C ', after combining it with −− 312 i kkk312
k (4.15) we have α =−−−+−−2 exp[][] (kkktt )( ) C ' exp ktt ( ) 23123−− ii kkk312
77 k Applying the initial condition:tt==,0α , it leads to C ' = 2 , i 2 +− kkk123
k So α =−+−−−−2 {}exp[][] (kktt )( ) exp ktt ( ) 2123−− ii kkk312
The final solution for α is
k αα=+= α 1 {}1exp( −[] −+kktt )( − ) (4.17) 12+ 12 i kk12 k +−−−−+−≥2 {}exp[][kt ( t ) exp ( k k )( t t ) ] fortt +− 312iii kkk123
Equation (4.17) represents the evolution of the state of cure when the curing takes place isothermally and it is the basis for the modeling of the isothermal cure kinetics.
4.3.2 Non-isothermal Cure Kinetics
196 Isayev and Deng proposed the definition of non-isothermal induction time tI as follows:
tI dt ∫ =1 (4.18) 0 tTi () where temperature T depends on time. Numerical integration techniques were used to obtain the tI value. The method to solve Equations (4.3) to (4.5) for the non-isothermal cure is similar to that of the isothermal case; however the kinetic constants k1 to k3 are variable with temperature and time. The analytical solution is
α =≤ 0 for t tI
≥ for t tI tttτ τ ατ=−++−∫∫kkkddkdkkkkdd()exp [ () θθθτ ()] exp ∫∫∫ () τττ ()exp −+−[] () θθθθτ () () 112 32123 ttII ttt III
78 In the modeling, the correspondent numerical form was applied where numerical double integration is adopted after the induction period tI:
α =≤ 0 for t tI
≥ for t tI
t τ tt τ ατ=−+∆∆+−∆⋅−+−∆∆∑∑kkkkkkkk()exp[] ( () θθθτ () exp ∑∑∑ () τττ ()exp [ ( () θθθθτ () () ] 11232123 ttII tttII I
(4.19)
Equation (4.19) represents the evolution of the state of cure when the curing temperature is varying with time and it forms the basis of non-isothermal kinetic modeling.
4.3.3 The Modeling Procedure for Cure Kinetics Study
The detailed procedure for the cure kinetics modeling study was shown in Figure
4.2. Γ(t) is the instant torque under the experimental conditions, α(t) is the state of cure, ti and tI are the isothermal and non-isothermal induction time, respectively.
4.4 Isothermal Cure Kinetics of the Unfilled IR
In this section the modeling of the isothermal cure kinetics was carried out for the unfilled IR.
4.4.1 Curing and the determination of experimental induction time
The experimental cure curves at various isothermal conditions are shown in
Figure 4.3 for the unfilled IR using the recipe in Table 3.2. The strain amplitude was
4.2% and the frequency was 100 cpm in APA 2000.
79 Run isothermal cure to obtain Γ(t)
Find experimental isothermal α(t) Find experimental isothermal ti
Obtain k1,2,3 by individual fitting α(t) with Eqn (4.17)
Obtain ti0 , Eti , k10, E1 , k20, E2 , k30, E3 with Arrhenius Eqn
α Obtain k10 , E1 , k 20 , E2 , k30, E3 by simultaneous fitting (t) with Eqn (4.17)
Run non -isothermal cure Find experimental Predict the non- isothermal αi to obtain Γ(t) non-isothermal α(t) with Eqn (4.19)
Comparison of α(t)
Determine experimental Predict the non-isothermal non- isothermal tI tI with Eqn (4.18) and T(t)
Comparison of tI
Figure 4.2 The modeling procedure for the cure kinetics study
Different degrees of reversion are observed when the curing temperature varies, as shown in Figure 4.3. Generally speaking, the lower the temperature, the less rigorous
80 the reversion. This indicates that the polysulfidic crosslinks are more stable at lower temperatures. This observation is consistent with Bielstein’s early work183. At higher temperatures, it is observed that both the maximum and the minimum curing torques are lower. The lower minimum torque at higher temperatures is due to the reduced viscosity of the materials at elevated temperatures. The lower maximum torque at higher temperatures is due to not only the reduced viscosity but also to more reversion at elevated temperatures.
6
140oC 5 o o 150 C 160 C o 170oC 130 C 4 180oC
3
Torque, dNm Torque, 2
1
0 0.1 1 10 100
Time, min
Figure 4.3 Isothermal cure curves of the unfilled IR (strain amplitude: 4.2%,
frequency: 100 cpm)
The evolution of the curing torque with time is observed in Figure 4.3: firstly it goes through a minimum, and then a gradual rise; after this delayed period the curing
81 torque increases much faster and almost occurrs at a constant rate. Based on this observation, it is appropriate to determine the experimental induction time as follows.
First, fit the linear equation for the time span where the curing torque increases linearly with time. Then extrapolate this straight line to the minimum torque line that is parallel to the time axis. The time where these two lines cross each other is taken as the experimental induction time. The detailed procedure is schematically shown in Figure 4.4.
The determined experimental induction time ti was used in the isothermal kinetic modeling by subtracting this value from the real time scale according to Equation 4.17.
reversion Torque
experimental induction time
post cure induction curing period
Vulcanization Time
Figure 4.4 Determination of the experimental induction time
4.4.2 Determination of the experimental state of cure α
After the experimental induction time is determined, it is necessary to find the experimental evolution of the state of cure, α, from the available torque data Γ(t). To do
Γ − Γ so according to Equation 4.6, one has to find the hypothetical value max,0 min,0 . This 82 Γ − Γ value is experimentally unattainable. However, the value of max min at each isothermal condition is available. These data are utilized to extrapolate to some low
Γ − Γ temperatures in order to find the max,0 min,0 value. Therefore, a proper description of
Γ − Γ max min as a function of temperature is desirable. Figure 4.5 shows the decrease of
Γ − Γ max min with the increase of temperature. Based on the observation of the line shape shown in this figure, it is appropriate to assume that the relationship follows the mathematical equation:
(Γ − Γ ) [Γ (T ) − Γ (T )] = max,0 min,0 (4.20) max min (Γ − Γ )T 1+{ max,0 min,0 }1−n τ
4.2
4.0
3.8
3.6 ), dNm min Γ - 3.4 max
(Γ 3.2
3.0 experiment curve fitting 2.8 400 410 420 430 440 450 460 o T, K Γ − Γ Figure 4.5 Finding the max,0 min,0 by Equation 4.20 for unfilled IR 83 Sigmaplot® 8.0 was employed to fit the torque difference as a function of temperature using the above equation. The curve fitting is based on a non-linear least-
Γ − Γ square regression technique and it leads to the max,0 min,0 value of 4.41 dNm for the unfilled IR. Then the isothermal vulcanization data (torque versus time) at six different temperatures shown in Figure 4.3 are converted to the experimental state of cure as a function of time using Equation 4.6.
4.4.3 Modeling of the state of cure α at the individual temperatures
The experimental state of cure at each temperature is fitted individually according to Equation 4.17 in order to obtain the kinetic rate constants k1(T), k2(T) and k3(T).
Accordingly, the natural logarithm of the rate constants versus the reciprocal temperature is plotted in Figure 4.6 and fitted with the Arrhenius equation (4.1) to obtain the activation energies (E1, E2, E3) and the pre-exponents (k10, k20, k30). The obtained values are summarized in Table 4.1. These constants will be employed for the initial guesses in the simultaneous fitting of the experimental state of cure versus time at all available isothermal temperatures.
4.4.4 The simultaneous modeling of the isothermal state of cure α
The simultaneous fitting of experimental state of cure from all the available isothermal conditions serves two purposes: one is to obtain more accurate values of activation energies and the pre-exponents for the prediction of non-isothermal state of cure (Equation 4.19); the other is to examine the stability of the kinetic model beyond the individual modeling. This is made by combining the state of cure data from all the isothermal temperatures and a non-linear least-square regression technique is applied.
The predictions of isothermal state of cure by the simultaneous modeling are shown as 84 5 k 4 1 k 2 4 k3 2 ti -1 3 0 , ln min i , ln min i
2 ln t ln k -2 1 -4 0 2.1x10-3 2.2x10-3 2.3x10-3 2.4x10-3 2.5x10-3 -1 -1 T , K
Figure 4.6 Kinetic constants and induction time versus the reciprocal temperature for
the unfilled IR the lines in Figure 4.7 by using the parameters in Table 4.1 as the initial guesses. It is observed from Figure 4.7 that there is some deviation of state of cure between the experimental and fitted results in the induction period. The value of α >0 in the induction period does not represent the real curing. These non-zero values of α are resulted from the temperature increase leading to softening of the material. On the other hand, the deviation of α after the minimum torque is due to the use of the first order kinetic model.
It is unable to capture the pre-curing period during which the complex reactions of accelerator and activator occur. Nevertheless, under all the available isothermal conditions, the kinetic model is able to predict an accurate progress of the state of cure
85 during the curing and reversion period compared to the experimental values seen in
Figure 4.7. In the mean time, it shows the stability of the kinetic model in use.
Table 4.1 Induction time and cure kinetic constants obtained from individual
isothermal fitting of the unfilled IR
ti0 Eti k10 E1 k20 E2 k30 E3
(min) (J/mol) (min-1) (J/mol) (min-1) (J/mol) (min-1) (J/mol)
9.57×10-12 9.86×104 6.46×108 7.74×104 6.22×1011 1.06×105 6.55×1013 1.21×105
R& R& R& R& R& R&
α
$9,9041.:70
7LPHPLQ
Figure 4.7 Isothermal state of cure for the unfilled IR: experimental (symbols) and fit (lines)
86 This is in contrast to the GRT isothermal cure where the same kinetic model was employed and it either over predicted or under predicted the progress of state of cure202.
The modeling result for GRT was due to the complexity of the sample under use, i.e.
GRT, which was a mixture of different elastomers (BR, SBR, NR, etc.) underwent complex chemical reactions at varied reaction rates that can not be well described by the simplified model being use. Additionaly, the isothermal kinetics modeling of the unfilled
IR suggests that the proposed Equation (4.20) represents an appropriate description of
Γ − Γ max min as a function of temperature. The constants from the simultaneous modeling of the unfilled IR isothermally are summarized in Table 4.2 and were employed to predict the evolution of non-isothermal state of cure as a function of time and temperature. In the following section, the non-isothermal cure kinetics modeling of the unfilled IR is investigated.
Table 4.2 Cure kinetic constants obtained from simultaneous isothermal fitting of
the unfilled IR
-1 -1 -1 k10 (min ) E1 (J/mol) k20 (min ) E2 (J/mol) k30 (min ) E3 (J/mol)
0.91×109 7.88×104 1.74×1011 1.01×105 7.54×1014 1.31×105
4.5 Non-isothermal Cure Kinetics of the Unfilled IR
The non-isothermal cure kinetic experiments were also performed using the rheometer APA 2000. Several non-isothermal runs are listed in Table 4.3. The non- isothermal curing experiments were carried out in order to examine the applicability of the kinetic model to a wide variety of non-isothermal conditions.
87 Table 4.3 Heating rates and step temperature profiles in nonisothermal curing of the
unfilled IR
Constant heating rate starting from 80oC 1.93 2.13 2.54 3.16 5.81
(oC/min)
Constant heating rate starting from 120oC 2.25 2.78 3.35 7.35
(oC/min)
Step temperature profile (oC) 80 120 160 180
5 minutes stay at each T
4.5.1 Determination of the non-isothermal induction time tI
Before testing the applicability of the kinetic model for the non-isothermal cure, the non-isothermal induction time tI has to be determined both experimentally and theoretically. The procedure for finding the experimental tI is the same as that for the isothermal ti described in Figure 4.4. The prediction of tI is based on Equation 4.18. The
Arrhenius type of induction time function (Equation 4.2) coupled with the temperature history T(t) is plugged into Equation 4.18 so that the theoretical non-isothermal induction time tI is determined. Numerical integration technique was applied in using Equation 4.18.
The experimental and the predicted induction time as a function of heating rate are shown in Figure 4.8. Shorter induction times are observed at faster heating rates in both the experimental and the predicted data. This suggests that vulcanization reactions start earlier at higher temperatures. An excellent agreement of the experimental and the predicted induction times were reached at a wide variety of the non-isothermal conditions.
This result reveals that the Arrhenius type of relationship is an adequate description of 88 induction time function. It also indicates the validity of the non-isothermal induction time
196 concept proposed by Isayev and Deng . The predicted tI is used in the calculation of the non-isothermal state of cure.
80 o exp, 80-180 C 70 o exp, 120-180 C 60 prediction exp, step cure 50 prediction, step cure
40
30
Induction time, min 20
10
0 0246810 o Heating rate, C/min
Figure 4.8 Non-isothermal induction time for the unfilled IR: experiments and
predictions. Step cure refers to the step temperature profile in Table 4.3.
4.5.2 Non-isothermal modeling of the state of cure α
The prediction of the non-isothermal state of cure as a function of time and temperature is based on Equation (4.19) by taking the parameters in Table 4.2 and the predicted tI in Figure 4.8. Figure 4.9, Figure 4.10 and Figure 4.11 show both the experimental and the predicted state of cure under a wide variety of non-isothermal conditions listed in Table 4.3. In Figure 4.11, the variation of the intial torques reflects 89 the temperature change with higher temperature leading to reduced intial torque. From these figures one can observe that regardless of the non-isothermal history, the predictions always show a close agreement with the experimental results in the curing and reversion period. This is again different from the results of GRT non-isothermal cure where the kinetic model over-predicted the reversion especially at faster heating rates202.
The loss of accuracy in GRT for the non-isothermal modeling was not only attributed to the complicated system but also due to the accumulation of inaccuracy from the isothermal modeling. The successful simulation of the unfilled IR isothermally and non- isothermally reveals that the simplified kinetic model proposed in Figure 4.1 adequately describes the curing and reversion reaction of the unfilled IR.
1.0 exp-1.93oC/min exp-2.13oC/min 0.8 exp-2.52oC/min exp-3.16oC/min exp-5.81oC/min α 0.6 Prediction
0.4
Statecure of 0.2
0.0
0204060
Time, min
Figure 4.9 Non-isothermal state of cure for the unfilled IR: experiments (starting
from 80oC) and predictions
90 1.0 H[S4&PLQ H[S4&PLQ 0.8 H[S4&PLQ H[S4&PLQ
α 0.6 SUHGLFWLRQ
0.4
Statecure of 0.2
0.0
0 102030
Time, min
Figure 4.10 Non-isothermal state of cure for the unfilled IR: experiments (starting
from 120oC) and predictions
91 1.0 200 exp α α 0.8 prediction 180 Temperature 160 C α 0.6 o 140 0.4 120 State of cure,
0.2 Temperature, 100
0.0 80
60 0 5 10 15 20 25
Time, min
Figure 4.11 Non-isothermal state of cure for the unfilled IR: experiments (step
temperature profile) and predictions
4.6 Isothermal Cure Kinetics of the CB Filled IR
Carbon black is one of the most widely used fillers in the rubber industry. The addition of CB is important in enhancing the performance of rubbers properties such as modulus, hardness, tensile strength, abrasion and tear resistance as well as resistance to fatigue and cracking. CB not only affects the properties of rubbers but also influences the vulcanization kinetics. The focus of this section is to examine the feasibility of the cure kinetic models on the filled system by taking the 35phr CB filled IR as the example. The method is similar to that of the unfilled IR described in Section 4.4.
Figure 4.12 shows the isothermal cure curves of the 35 phr carbon black filled IR at various temperatures. Compared to the unfilled IR (Figure 4.3), the filled IR has a
92 shorter induction period. This result is consistent with previous research203, 204, 205 where it was found that the carbon black acted as a catalyst speeding up the cure reaction, reducing the induction time compared with the unfilled system. Not all carbon blacks have the acceleration effect on the vulcanization. The surface chemistry and the pH value of carbon black have a great influence on the extent of vulcanization206, 207. Channel blacks, which contain a number of surface oxygen functional groups, such as quinones, hydroquinones, phenolic hydroxyls, carboxylic acids, and lactones, have been reported to be acidic, which tend to retard cure. Furnace blacks which are characterized by a neutral pH or slightly alkali with low oxygen content have an acceleration effect208.
Consequently, the furnace black such as the N330 used in this study (HAF with pH
8.7206) accelerates the cure reaction and decreases the scorch time.
Compared to the unfilled IR, Figure 4.12 shows that the filled IR has more severe reversion. Bhowmick and De209 showed that the addition of carbon black enhanced the polysulfidic cross-links as well as the total cross-links in their study on cross-linking kinetics and network changes in unfilled and filled NR vulcanizates with a dithiodimorpholine-based accelerator system. Pal et al.210 also showed that both overall cross-link density and polysulfidic cross-links were increased when carbon black was added in the 0–5 phr range in the conventional system. Since the mono- and di-sulfidic crosslinks are relatively stable, the severe reversion was due to the increased formation of polysulfidic crosslinks formed due to the addition of carbon black were responsible for more severe reversion. On the other hand, the increased reversion in the 35phr CB filled
IR compared to that of the unfilled IR can be explained by the comparison of the
93 reversion kinetic constants k30 between the unfilled IR (Table 4.2) and the filled IR which will be shown at the end of this section.
12 o 120oC 110 C o o 140 C 130 C 150oC 10 160oC o 170 C 180oC 8
6
Torque, dNm 4
2
0 0.1 1 10 100
Time, min
Figure 4.12 Isothermal cure curves of the 35 phr CB filled IR (strain amplitude: 4.2%,
frequency: 100 cpm)
Γ − Γ Figure 4.13 shows the max min as a function of temperature for the 35phr filled
IR. The fitting of the experimental data using Equation 4.20 results in a value of 9.72
Γ − Γ dNm for the max,0 min,0 . Based on this hypothetical value, the isothermal vulcanization data (torque vs. time) shown in Figure 4.12 are converted to the experimental state of cure as a function of time using Equation 4.6.
94 10.0
9.5
9.0 ), dNm
min 8.5 Γ - max
Γ 8.0 (
7.5 experiment curve fitting 7.0 360 380 400 420 440 460 o T, K Γ − Γ Figure 4.13 Finding max,0 min,0 by Equation 4.20 for 35phr CB filled IR
The obtained kinetic constants from the isothermal curve fitting at the individual temperatures based on Equation 4.17 and the experimental induction times as a function of the reciprocal temperature are shown in Figure 4.14. The fitting parameters based on the Arrehnius relationship (the activation energies and the pre-exponent constants) from
Figure 4.14 are tabulated in Table 4.4. These parameters were employed for the initial guesses in the simultaneous curve fitting of the state of cure versus the time at all of the available isothermal conditions.
95 4 4
2 2
0 0 -1 -2 -2 , ln min , ln i , ln min , ln i -4 -4 ln t ln k ln -6 ti -6 k1 -8 k2 -8 k3 -10 -10 2.0x10-3 2.2x10-3 2.4x10-3 2.6x10-3 2.8x10-3 -1 o-1 T , K
Figure 4.14 Kinetic constants and induction time versus the reciprocal temperature for
the 35phr CB filled IR
Table 4.4 Induction time and cure kinetic constants obtained from individual
isothermal modeling of the 35phr CB filled IR
ti0 Eti k10 E1 k20 E2 k30 E3
(min) (J/mol) (min-1) (J/mol) (min-1) (J/mol) (min-1) (J/mol)
1.75×10-10 8.23×104 3.76×108 7.37×104 1.10×1017 1.46×105 5.39×1017 1.56×105
The simultaneous isothermal curve fitting results are shown in Figure 4.15 as the solid lines. Similar to the unfilled IR, the predicted isothermal state of cure in the filled
IR also shows excellent agreement with that of the experimental data. This suggests that
96 the kinetic model in use is not only adequate for the unfilled system but is also applicable for the filled system despite the deviation of kinetic constants from the linearity (Figure
4.14). The constants from the simultaneous modeling are summarized in Table 4.5 and they are to be used in the prediction of the non-isothermal state of cure for the 35phr filled system.
o 120oC 1.0 o 130 C o 150oC140 C 110 C 160oC 170oC 0.8 180oC α 0.6
0.4 State of cure cure State of 0.2
0.0
0.1 1 10 100 1000 Time, min
Figure 4.15 Isothermal state of cure for the 35phr CB filled IR: experiments (symbols)
and fits (lines)
By comparing the parameters in Table 4.2 for the unfilled IR with those in Table
4.5 for the 35phr CB filled IR one can observe that all of the constants are almost at the same order of magnitude except that the k30 of the filled IR is two orders of magnitude higher than that of the unfilled IR. Knowing that k30 is responsible for the destruction of
97 unstable crosslinks according to the kinetic model shown in Figure 4.1, the simulation results indicate that after achieving the maximum torque, reversion is faster for the filled
IR than for the unfilled IR. This conclusion is consistent with the experimental results shown earlier in this section by comparing Figure 4.3 for the unfilled IR and Figure 4.12 for the 35 phr CB filled IR.
Table 4.5 Cure kinetic constants obtained from simultaneous isothermal fitting of
the 35phr CB filled IR
-1 -1 -1 k10 (min ) E1 (J/mol) k20 (min ) E2 (J/mol) k30 (min ) E3 (J/mol)
4.01×109 8.24×104 0.83×1011 0.95×105 4.18×1016 1.47×105
4.7 Non-isothermal Cure Kinetics of the CB Filled IR
The designed non-isothermal cure runs are listed in Table 4.6 to examine the applicability of the kinetic model to the non-isothermal cure of CB filled IR. Figure 4.16 shows both the experimental (symbols) and predicted (lines) non-isothermal induction time. The predicted value is obtained by numerical integration of Equation 4.18 coupling with Equation 4.2. From Figure 4.16 the excellent agreement between the predicted and the experimental induction time suggests the non-isothermal induction time concept is universal and it is valid in the unfilled system as well as in the filled system.
The prediction of non-isothermal state of cure for the 35 phr CB filled IR is carried out according to Equation (4.19) by taking the parameters in Table 4.5. Figure
4.17, Figure 4.18 and Figure 4.19 show both the experimental (symbols) and the predicted state of cure (lines) under various non-isothermal conditions listed in Table 4.6.
98 In Figure 4.19, the variation of the intial torques reflects the temperature change with higher temperature leading to lower intial torque. In any case, the kinetic model predicts a value very close to the experimental value in curing and reversion period. This suggests that regardless of whether the rubber is filled with CB or not, the cure kinetic model is able to accurately characterize the curing and the reversion stage of sulfur vulcanization.
Table 4.6 Heating rates and step temperature profiles in nonisothermal curing of the
35phr CB filled IR
Constant heating rate starting from 80oC (oC/min) 1.80 2.32
Constant heating rate starting from 120oC (oC/min) 2.01 2.51
Step temperature profile (oC) 5 min stay at each T 80 120 160 180
50 exp, 80-200oC exp, 120-200oC 40 prediction exp, step cure prediction, step cure 30
20 Inductionmin time,
10
1234 o Heating rate, C/min
Figure 4.16 Non-isothermal induction time for the 35phr CB filled IR: experiments
and predictions. Step cure refers to the step temperature profile in Table 4.6
99 1.0 exp-1.80oC/min exp-2.32oC/min 0.8 prediction
α 0.6
0.4
Statecure of 0.2
0.0
0204060
Time, min
Figure 4.17 Non-isothermal state of cure for the 35phr CB filled IR: experiments
(starting from 80oC) and predictions
100 1.0
0.8
α 0.6
0.4
Statecure of 0.2
0.0 H[S4&PLQ H[S4&PLQ
0 10203040
Time, min
Figure 4.18 Non-isothermal state of cure for the 35phr CB filled IR: experiments
(starting from 120oC) and predictions
101 1.0 180 exp α α 0.8 prediction Temperature 160 C α 0.6 140 o
0.4 120 State of cure,
0.2 100 Temperature,
0.0 80
60 0 5 10 15 20 25
Time, min
Figure 4.19 Non-isothermal state of cure for the 35phr CB filled IR: experiments (step
temperature profile) and predictions
4.8 Conclusions
The simplified reversion type of cure kinetic model proposed by Han et al199 with nonisothermal induction time model197 was successfully utilized in the modeling of isothermal and non-isothermal cure kinetics for the unfilled and the 35phr carbon black filled IR. In this model the induction time function was introduced as an explicit kinetic parameter and it follows the Arrhenius dependence on the temperature. This cure kinetic model was able to predict the accurate evolution of the state of cure in isothermal as well as in non-isothermal conditions in the curing and reversion stage.
The proposed Equation 4.20 adequately described the dependence of isothermal
Γ − Γ torque difference max min on temperature. Isothermal modeling of the IR compounds 102 (CB filled and unfilled) either from the individual temperatures or from the simultaneous modeling of all the temperatures showed the excellent agreement between the experimental and the predicted data in the curing and reversion period. This indicated the stability of the kinetic model in use.
The non-isothermal induction time was determined both experimentally via the method proposed in this research (Figure 4.4) and theoretically via the concept proposed by Isayev and Deng197 in a conjunction with the Arrhenius equation. An excellent agreement was observed between the experimental and the predicted non-isothermal induction time. More severe reversion occurring in the CB filled IR than in the unfilled
IR was explained by comparing their reversion kinetic constant k30 from the isothermal cure kinetics modeling.
103
CHAPTER V
5.STRUCTURE AND PROPERTIES OF THE ULTRASONICALLY TREATED GUM
ISOPRENE RUBBER
5.1 General
IR is the artificial equivalent of NR since they share the same basic repeat unit – cis 1, 4-isoprene. It is mostly utilized in tire industry in the combination with or instead of
NR4. Compared with its natural counterpart, IR is inferior in mechanical strength, anti- aging and crystallization. However, it exceeds NR in the consistency of product, uniformity of cure rate, ease of processing (mixing, extrusion, molding and calendering) and purity. Particularly, IR does not undergo storage hardening3. Products made from NR or IR are less likely than most other elastomers to fail from excessive heat buildup or fatigue when exposed to severe dynamic conditions. This has secured the place of NR and IR as the preferred sidewall elastomers in radial tires28.
The degradation of polymers in solutions due to prolonged ultrasonic treatment has a long history. As early as 1933 Szalay211 described how ultrasonic waves depolymerized starch, gum Arabic and gelatin, as measured by a reduction in viscosity.
Schmid and Rommel212 also investigated the decrease of viscosity in synthetic polymers such as poly (acrylic acid), poly (vinyl acetate), and nitrocellulose. Besides the observation of the reduction in viscosity, they found that the depolymerization was 104 initially rapid but quickly slowed down and eventually ceased when a minimum molecular weight was reached. The chain cleavage due to the ultrasound exposure has several aspects that differentiate it from the thermal or photochemical processes and these may be regarded as the characteristics of the ultrasonic method. It proceeds faster at high molecular weights and slows down until, at some limiting value, the degradation ceases.
Additionally, unique to ultrasonic degradation is the preferable cleavage of the polymer chain near the center of the macromolecules151, 152. The mechanism of ultrasonic depolymerization is now thought to be closely linked to the cavitation phenomenon. The cavitation can produce sufficiently high local pressure and temperature143, 144 to induce homolytic breakage of the macromolecular chains to form macroradicals. The evidence of the presence of these radicals due to ultrasonic exposure was obtained by carrying out the sonication in the presence of unsaturated, polymerizable monomers by the trapping of the radicals using radical scavengers such as DPPH (a,a’-diphenyl picryl hydrazyl)213.
High power ultrasound has found wide spread applications throughout polymer sonochemistry. For instance, the radicals produced during sonication have been shown to initiate polymerization in a second unsaturated monomer within a wide range of choices.
It has been used in the production of copolymers from two or more different monomers
214, and in copolymerization during melt processing of immiscible blends215. In addition to the degradation and copolymerization, ultrasound can also be applied to the devulcanization of various types of rubbers such as GRT20, 121, NR123, 216, silicone rubber126, 217, SBR218 and PU125. This technique provides a rapid breakage of the three- dimensional rubber network within several seconds. It is a continuous process without the involvement of any chemicals. The rubber treated with high power ultrasound is soft, 105 moldable, and can be reshaped and revulcanized similarly to the virgin rubbers121. These advantages make the ultrasound technique unique and attractive to the rubber recycling industry. Undoubtedly, the application of ultrasound is safer and requires less energy since it avoids the high temperatures and pressures of thermal methods.
Recently, ultrasonic treatment was applied to several unfilled gum rubbers, such as EPDM rubber219, BR220 and butyl rubber221. Their structure and properties after treatment were investigated and the results showed that different elastomers had different responses after ultrasound exposure. For example, ultrasonic treatment of BR and EPDM led to measurable gel formation with the gel fraction dependent on the ultrasonic amplitude. In contrast, the treatment of butyl rubber did not produce any gel221.
In this chapter, the high power ultrasound was applied on the virgin gum rubber to examine the stability of the carbon-carbon main chain linkage under the ultrasonic exposure. The effect of high power ultrasound on the structure and properties of IR gum was investigated and compared with those of virgin IR gum. Finally, this research explored a new way of controlling the structure and properties of elastomers and improving their processability via high power ultrasonic treatment.
5.2 Preparation of the Samples
The gum rubbers were cut into thin long strips for the convenience of conveying them into the extruder. The ultrasonic treatment was done at a barrel temperature of
120oC. During the process, the samples were exposed of three different ultrasonic amplitudes: 5, 7.5 and 10 µm at the flow rate of 0.63 g/s. The power consumption and die pressure were recorded during the treatment. After extrusion, the samples were collected, labeled and placed in the freezer overnight. 106 The cure kinetics of the ultrasonically treated rubber by taking the recipe shown in Table 3.2 for the unfilled system was investigated by the APA 2000. The molecular weight and its distribution of the ultrasonically treated rubber were analyzed by the GPC.
The thermal properties (such as the degradation temperature, the glass transition temperature) of the ultrasonically treated rubber and their vulcanizates were analyzed by the TGA and DSC. The dynamic properties of the ultrasonically treated rubber and their vulcanizates were examined by the APA 2000. The mechanical properties of the vulcanizates were carried out by the Instron Tensile Tester at room temperature and an extension rate of 500 mm/min.
5.3 Ultrasonic Treatment: Die Pressure and Power Consumption
The extrusion of IR gum strips was carried out using the ultrasonic reactor at
120oC, a gap of 2.54 mm and varying amplitudes. During this process the die pressure and power consumption were recorded. Figure 5.1 shows the die pressure and power consumption as a function of the ultrasonic amplitude. The power consumption increases with the ultrasonic amplitude dramatically. The increase of power consumption is due to an increase of the strain and stress amplitude causing the degradation of the material.
This result is supported by the molecular weight shown in Section 5.6 and the rheological measurements shown in Section 5.7. On the other hand, die pressure drops with increasing the ultrasonic amplitude. When no ultrasound is applied, the die pressure is as high as 6.2 MPa. However, upon ultrasound exposure, the die pressure significantly drops.
With the increase of ultrasonic amplitude, the die pressure decreases further showing a linear dependence on the amplitude. This is previously explained as the combined effect
107 of the frictional reduction between the rubber and the die wall and the degradation taking place as rubber entered the ultrasonic treatment zone219.
7 1600
1400 6 1200 5 1000
4 800
3 600 400
Die pressure, MPa 2 200 Power consumption, watts 1 0 0 0 2 4 6 8 10 12 µ Amplitude, m
Figure 5.1 Die pressure and power consumption as a function of the ultrasonic
amplitude during the ultrasonic treatment of gum IR
5.4 Curing
After the ultrasonic treatment at different amplitudes, the virgin and treated IR gum samples were individually compounded with the same curing recipe. The cure kinetics measurements were carried out on the APA 2000 at 160oC using a strain amplitude of 4.2% and a frequency of 100 cpm. As shown in Figure 5.2, the initial torques of the treated IRs before the start of crosslinking are lower than that of the virgin
IR. Also with the increase of ultrasonic amplitude, the initial torque decreases. This again reveals the rubber degradation in the ultrasonic extrusion. The induction time of the
108 ultrasonically treated IRs was shorter than that of the virgin IR. However, the induction time does not depend on the amplitude. This is in contrast to the behavior of BR, where the induction time of the ultrasonically treated BR is shorter than that of the virgin BR, but decreases with increasing amplitude220. It is also found in Figure 5.2 that all the treated samples as well as the virgin IR show a reversion at the later stage of cure. This is probably due to the irreversible destruction of the polysulfidic crosslinks183, 184. It seems that at higher amplitude, the reversion is more severe and the final torque is lower. In addition, it is noticed that the curing rate and the maximum torque of the sample treated at 5 µm are slightly higher than that of the sample passing through the extruder without the imposition of the ultrasound (0 µm).
5 160oC cure IR5µm IR 4 IR0µm IR7.5µm IR10µm 3
2 Torque, dNm
1
0 0 5 10 15 20 25 30 Time, min
Figure 5.2 Cure curves of the virgin and the ultrasonically treated IR gums at 160oC,
a strain amplitude of 4.2% and a frequency of 100 cpm 109 5.5 Gel Fraction and Crosslink Density
After vulcanization of the ultrasonically treated IR samples, gel fraction and crosslink density of the vulcanizates were evaluated and the results are shown in Figure
5.3. It is worthwhile to mention that the extraction experiments show no measurable amount of gel detected in the virgin and ultrasonically treated IR gums at three amplitudes. This finding is similar to the behavior of butyl rubber221 but different with the observation made on EPDM219 and BR220 where the gel was produced after their exposure to the ultrasound. The gel fraction of the vulcanizates prepared from the ultrasonically treated IR is slightly lower than that of the virgin IR vulcanizate. This suggests that even though degradation occurs, molecular weight of the ultrasonically treated rubber is still very high such that a significant amount of double bonds survive allowing the treated samples to undergo vulcanization. However, these gels are different from the gel of the virgin rubber vulcanizate. They exhibit lower crosslink densities than that of the virgin rubber, as shown in Figure 5.3.
5.6 Molecular Characteristics of the Ultrasonically Treated Gums
The molecular characteristics of the ultrasonically treated IR gums were determined by GPC. The results are shown in Figure 5.4. The number (Mn) and weight
(Mw) average molecular weight of the virgin IR are 982,000 and 1,998,000, respectively.
These values are somewhat lower than those supplied by the manufacturer and measured by the ThFFF method. In Figure 5.4a, curves of the molecular weight distribution of the ultrasonically treated gum rubbers are seen to be shifted to the lower molecular weight compared to that of the virgin IR. It is also found that low molecular weight tails are generated at various amplitudes. With the increase of amplitude, the molecular weight of 110 100 0.24
0.22 3 90 0.20
0.18 80 0.16 Gel fraction, % 0.14 70 Crosslink density, kmol/m 0.12
60 0.10 024681012 µ Amplitude, m
Figure 5.3 Gel fraction and crosslink density as a function of the ultrasonic amplitude
for the vulcanizates of virgin (symbols shown on the ordinate axes) and ultrasonically
treated IR gums
111 1.2 (a) 1.0
0.8 w 0.6 /dLog M f 0.4 dW IR5µm 0.2 IR 0.0 µ IR10µm IR7.5µm IR0 m
345678 Log M w 107 7 (b)
6 Mw/Mn Mw
w 5 n
106 /M w or M or n M
M 4 Mn
3
105 2 IR 0 µµ m 5 m 7.5 µµ m 10 m Material
Figure 5.4 Molecular weight distribution of the virgin and ultrasonically treated IR gums at various amplitudes (a), and amplitude dependence of the number (Mn), weight (Mw) average molecular weight and polydispersity (Mw/Mn) of the virgin and treated IR gums (b)
112 the tails is progressively shifted to a lower value. This indicates the degradation of the rubber main chain upon ultrasonic treatment. It is also observed that the sample passing through the extruder without exposure of ultrasound (0 µm) also shows degradation due to the mastication effect of IR occurring in the two-roll mill and in the extruder. This is in agreement with previous work on mastication in two-roll mill222, 223. Furthermore, both the number and weight average molecular weight decrease with increasing amplitude as shown in Figure 5.4b. However, the number average molecular weight decreases more significantly than the weight average value. Accordingly, the polydispersity (Mw/Mn) increases substantially with increasing ultrasonic amplitude. Obviously, the wider distribution of molecular weight is caused by the low molecular weight tails generated.
Therefore, the ultrasonic treatment of IR can be beneficial and could possibly be used as a means of improving the gum rubber processability.
5.7 Rheological Properties
The rheological properties of the virgin, and the ultrasonically treated IR gums and their vulcanizates were evaluated by the APA 2000 at 120oC at a strain amplitude of
4.2% and a frequency of 100 cpm. The complex viscosity of the gums and vulcanizates as a function of frequency is plotted in Figure 5.5. It is clear that the viscosity of the vulcanizates is substantially higher than that of the gums (untreated or treated) within the whole range of frequency. This is due to the formation of a three-dimensional network hindering the flow of the material. For the treated gums, it was found that the dynamic viscosity decreases with the increase of ultrasonic amplitude. This is another proof of degradation in addition to the reduction of molecular weight. Dynamic viscosity of all the vulcanizates does not show any measurable differences. This is in contrast to the findings 113 in the butyl rubber221. Different behaviors in the dynamic viscosity of the vulcanizates between the ultrasonically treated IR and the butyl rubber are due to the fact that IR contains a significant amount of the double bonds needed for vulcanization, while butyl rubber has a very little amount of the double bonds after the treatment.
109 IR 108 IR0µm IR5µm 107 IR7.5µm IR10µm 106 Vulcanizates *|, Pa-s 5
|η 10 Gums 104
103
102 10-2 10-1 100 101 102 103 ω , rad/s
Figure 5.5 Complex viscosity versus frequency for the virgin and ultrasonically
treated IR gums and their vulcanizates at 120oC and a strain amplitude of 4.2%
Figure 5.6 shows the loss tangent as a function of frequency for the virgin and ultrasonically treated IR gums along with their vulcanizates. The loss tangent increases as the ultrasonic amplitude increases for the treated gums indicating that after the treatment the material exhibits more viscous dissipation. However, the loss tangent of the vulcanizates is too small to differentiate among the various vulcanizates.
114 102 IR µ 101 IR0 m Gums IR5µm IR7.5µm 100 IR10µm δ 10-1 tan tan Vulcanizates 10-2
10-3
10-4 10-2 10-1 100 101 102 103 ω , rad/s Figure 5.6 Loss tangent versus frequency for the virgin and ultrasonically treated IR
gums and their vulcanizates at 120oC and a strain amplitude of 4.2%
Figure 5.7 shows the storage (G’) and loss (G”) modulus as a function of frequency for the virgin and ultrasonically treated IR gums and their vulcanizates at
120oC. Loss moduli of the vulcanizates are lower than those of the gums due to the network formed in the vulcanizates. In contrast, storage moduli of the vulcanizates are substantially higher than those of the gums. A tendency to the plateau modulus is observed for the virgin IR. The plateau region is also observed for all of the vulcanizates indicating the full cure is reached for all the treated and the untreated (0 µm) samples.
The storage modulus for the gums is dependent on the ultrasonic amplitude with higher amplitude corresponding to lower modulus. This is another indication of the rubber main chain degradation. However, the storage modulus for the vulcanizates is independent of
115 the ultrasonic amplitude. This suggests that full cure is achieved even when the degradation of the rubber main chain actually happened.
103 (a)
Vulcanizates
102
Gums G', kPa IR 101 IR0µm IR5µm IR7.5µm IR10µm 100 10-2 10-1 100 101 102 103 ω, rad/s 102 (b)
101 Gum
100 G", kPa
10-1 Vulcanizate IR IR-7.5µm IR-0µm IR-10µm IR-5µm 10-2 10-2 10-1 100 101 102 103 ω , rad/s Figure 5.7 Storage (a) and loss (b) modulus of the virgin and ultrasonically treated IR and their vulcanizates as a function of frequency at 120oC and a strain amplitude of 4.2%
116 In order to investigate whether or not branching took place during the ultrasonic treatment of IR gums, the complex viscosity was determined at three different temperatures for the virgin and ultrasonically treated IR gums. The modified Cross model224 was applied to fit the viscosity – frequency curve. The following two equations were used for this fitting:
|η * (T ) | T |η * |= 0 |η * (T ) |= Aexp b |η * (T) | ω 0 T 1+[ 0 ]1−n τ
τ η * with A, Tb, n and being the fitting parameters, 0 (T ) is the zero-frequency viscosity function based on the Arrhenius type of temperature dependence.
Figure 5.8 shows the experimental data (symbols) and the fitted curves for the virgin and the 5, 7.5, 10 µm IRs. The four fitting parameters (A, Tb, τ and n) obtained by the least-square regression method for the virgin and treated rubber gums are shown in
Table 5.1. It is well known that branched polymers usually have a higher viscous flow
225 activation energy E than the correspondent linear polymers . In terms of Tb, the value of activation energy is E = Tb×R with R being the universal gas constant. Table 5.1 shows that except that the rubber treated at the amplitude of 5 µm has a relatively high value of
Tb, the Tb values of the other treated gums and the virgin IR are lower and they show insignificant differences. Therefore, it is possible that branching occurred only for the sample treated at 5 µm while the other samples are only subjected to the degradation without any other additional structural changes (branching or crosslinking) during the ultrasonic treatment.
117 107
106
105
o *|, Pa-s *|, 4 IR 60 C |η 10 IR10µm 60oC IR 90oC o 103 IR10µm 90 C IR 120oC IR10µm 120oC 102 10-2 10-1 100 101 102 103 ω , rad/s 107
106
105
*|, Pa-s o
η 4 µ | 10 IR-5 m 60 C o IR-7.5µm 60 C o IR-5µm 90 C o 103 IR-7.5µm 90 C o IR-5µm 120 C o IR-7.5µm 120 C 102 10-2 10-1 100 101 102 103 ω , rad/s Figure 5.8 Complex viscosity versus frequency for the virgin IR and the IR
ultrasonically treated at the amplitude of 5, 7.5 and 10 µm at various temperatures
(symbols: experiments; curves: the modified Cross model fittings)
118 Table 5.1 Rheological parameters of the modified Cross model for the virgin and
ultrasonically treated IR gums
2 Gum Rubbers A (Pa-s) n Tb (K) τ (Pa) R
Virgin IR 693.72 0.1363 3094 89901 0.9934
IR0µm 269.29 0.1610 3208 71378 0.9919
IR5µm 14.55 0.1617 4253 67905 0.9881
IR7.5µm 171.56 0.1976 3109 46533 0.9892
IR10µm 119.50 0.2025 3192 40395 0.9884
5.8 Mechanical Properties
The stress-strain curves for the vulcanizates of the virgin and the ultrasonically treated IR at different ultrasonic amplitudes are shown in Figure 5.9. Even though the degradation occurrs during the ultrasonic treatment, it is clear that the vulcanizates of the treated IR gums, similar to the vulcanizate of the virgin rubber, show a high extent of strain-induced crystallization. This again indicates the destruction of the rubber main chain is minor. It is observed that for the sample treated at the amplitude of 5 µm, the slope of the stress-strain curve is slightly higher than that of the untreated sample only passing through the extruder (0 µm). Figure 5.10 shows the mechanical properties as a function of ultrasonic amplitude for the vulcanizates of the virgin and the ultrasonically treated IR. Generally, tensile strength, elongation at break and modulus of the vulcanizates decreases with the increase of ultrasonic amplitude, with the most reduction in the sample treated at the amplitude of 10 µm. The reduction is probably due to the
119 degradation of the rubber main chain. However, the elongation at break of the treated samples is higher than that of the virgin vulcanizate.
25 500mm/min
20 IR 0 µm 15 5 µm 7.5 µm , MPa
σ 10 10 µm 5
0 0 200 400 600 800 1000 1200 1400 ε , %
Figure 5.9 Stress-strain curves for the vulcanizates of the virgin and ultrasonically
treated IRs
120 25 1200 (a) ε B 1100 20 σ 1000 B
15 900 , % , MPa B B ε σ 800 10 virgin IR 700
5 600 024681012 Amplitude, µm
1.0 (b)
0.9
E100 0.8
0.7 , MPa E300 300
, E 0.6 100 E 0.5 virgin IR
0.4 024681012 µ Amplitude, m Figure 5.10 Amplitude dependence of the tensile strength, elongation at break (a) and
modulus at 100% and 300% (b) of the virgin (symbols shown on the ordinate axes) and
ultrasonically treated IR vulcanizates
121 5.9 Thermal Properties
The thermal stability of the untreated and the ultrasonically treated IR gums and their vulcanizates are evaluated by the TGA under nitrogen atmosphere. As shown in
Figure 5.11, generally there are no noticeable differences of the TGA curves for the treated and untreated gums as well as among their different vulcanizates. However, the difference is only evident between the gum rubbers and the vulcanizates: the gums degrade at a slightly lower temperature compared with the vulcanizates. This is reasonable because the vulcanization results in better thermal stability due to the crosslinking and the reduction of unsaturation in the rubber chain. At the final temperature, the gum rubbers show almost zero weight residues indicating the thermal degradation is complete (100%). The vulcanizates show about 5% weight residue. This exactly corresponds to the amount of ZnO used in the curing recipe.
Figure 5.12 shows the DSC curves of the virgin and the ultrasonically treated IR gums and their vulcanizates at low temperatures (-90 to -10oC) to determine the glass transition temperature Tg. The Tg was determined by the inflection point method. From the curves of gums no significant differences of Tg are observed for both untreated and treated IRs. Also, there are no observable differences of Tg for various vulcanizates. The
o only detectable difference is that the Tg values of vulcanizates are 3-4 C higher than those of gums. The reason is that the chain mobility is reduced due to the formation of three- dimensional network and higher temperature is necessary for the chain segment motion.
The Tg values are listed in Table 5.2. The Tg behavior of the ultrasonically treated IR gums is similar to that of the butyl rubber, but in contrast to that of the EPDM. There are
o substantial Tg differences among the treated EPDM gums, namely 4 to 6 C. The EPDM 122 gum treated at an amplitude of 5 µm shows a higher Tg value compared to the virgin untreated gums219. This is due to the measurable amount of gel formed during ultrasound exposure. In addition, the Tg value decreases with the increase of amplitude for EPDM.
The increased mobility is explained by the generation of shorter chains at higher ultrasonic amplitude.
100
80 IR IR-0µm 60 IR-5µm IR-7.5µm 40 IR-10µm Weight, % Weight, Vulcanizates 20 Gums
0 ~ 5.0%
0 100 200 300 400 500 600 700 o Temperature, C
Figure 5.11 TGA curves for the virgin and ultrasonically treated IR gums and their
vulcanizates under the nitrogen atmosphere
123 1.0
Gums IR
0.5 IR0µm IR5µm IR7.5µm µ 0.0 IR10 m
Vulcanizates IR Heat Flow (W/g) Heat IR0µm -0.5 IR5µm IR7.5µm IR10µm -1.0 -90 -70 -50 -30 -10 Exo up Temperature (oC) Universal V3.9A TA Instruments
Figure 5.12 DSC curves for the virgin and ultrasonically treated IR and their
vulcanizates under the nitrogen atmosphere
Table 5.2 Tg of the virgin and the ultrasonically treated IR gums and their
vulcanizates
o o Gum Rubbers Uncured Tg, C Cured Tg, C
Virgin IR -62.2 -59.0
IR0µm -63.0 -58.5
IR5µm -62.3 -58.0
IR7.5µm -63.9 -58.7
IR10µm -63.9 -59.3
124 5.10 Conclusions
Ultrasonic treatment altered the structure and properties of IR gum and the change was highly amplitude dependent. The reduction of die pressure and the increase of power consumption with the increase of ultrasonic amplitude were observed. The degradation of
IR gum under ultrasonic treatment was supported by the measurement of molecular weight, dynamic and mechanical properties. In particular, the molecular weight slightly decreased as the amplitude was increased. Ultrasound treatment created low molecular weight tails which broadened the molecular weight distribution. Therefore, it was beneficial to improve the processability of gum rubber.
The complex viscosity as well as the storage modulus of the treated rubber gums decreased as the amplitude increased. The fitting of the complex viscosity – frequency curves according to the modified Cross model indirectly indicated possible branching of
IR gum treated at the ultrasonic amplitude of 5 µm.
The mechanical properties (tensile strength, modulus, elongation at break) of the treated rubber vulcanizates decreased with the increasing amplitude. However, the elongation at break of the treated IR vulcanizates was higher than that of the virgin vulcanizate.
The cure curves of the treated rubber gums were similar to the virgin IR and showed the reversion. Because of the degradation, the initial and maximum torque of cure curves reduced with increasing amplitude. The vulcanization created a comparable amount of gel but a significantly lower crosslink density for the treated rubber gums compared with the virgin rubber. However, the thermal stability and glass transition temperature of the untreated and treated IR gums as well as their various vulcanizates 125 showed no significant differences. The difference in Tg as well as the thermal stability only occurred between the gums and the vulcanizates.
126
CHAPTER VI
6.ULTRASONIC DEVULCANIZATION OF THE UNFILLED IR
6.1 General
The emergence of synthetic rubbers such as IR, SBR, BR etc. was due to the scarcity of natural rubber (NR) during World War II. IR is the man-made substitute of
NR. It is a pure chemical product containing less stereoregular units than NR. The chemical name for IR is synthetic cis-1, 4 polyisoprene. Currently synthetic polyisoprene is being used in a wide variety of industrial applications requiring low water swell, high gum tensile strength, good resilience, high hot tensile and good tack. Unfilled IR vulcanizates are used in rubber bands, cut thread, baby bottle nipples, and extruded hose.
In addition, recent concerns about allergic reactions to proteins present in NR have prompted increased usage of the more pure synthetic polyisoprene in some applications4.
In order to solve the ever-increasing environmental problem due to the tremendous amount of waste rubber disposed, many rubber recycling methods have been developed. Extensive reviews on rubber recycling methods were given95, 226, 227. So far, the developed methods to recycle the waste rubber include chemical, mechanical, cryo- mechanical, biotechnical, microwaves and ultrasonic devulcanization. Among these methods, the application of high power ultrasound for the devulcanization of rubber is
127 one of the most promising techniques. Ultrasonic devulcanization is a continuous process, allowing one to recycle rubbers without the inclusion of any chemicals. The devulcanized rubber can be reprocessed, shaped, and revulcanized in the same way as the virgin rubber.
Extensive studies of ultrasonic devulcanization have been carried out on various rubbers including GRT20, 121, NR123, 216, silicone rubber126, 217, SBR122, 218, EPDM124, PU125, BR220,
228 and butyl rubber229.
In this chapter, an extensive investigation of the continuous ultrasonic devulcanization of unfilled IR vulcanizates is carried out using the co-axial devulcanization reactor. The devulcanized IR samples at varied ultrasonic amplitudes and their revulcanizates were investigated to compare the properties with those of the virgin vulcanizate. The network structures such as gel fraction and crosslink density, the mechanical properties, the rheological properties, and the molecular weight of the sol part of the devulcanized IR samples were investigated in order to elucidate the mechanism of the processes taking place during the ultrasonic devulcanization of the unfilled IR. In addition, some of the properties of IR were compared with those of NR181.
6.2 Experimental
IR compounded with curing ingredients, using the recipe listed in Table 3.2, were molded into slabs (260×260×12 mm3) at a temperature of 160°C and a pressure of 17.2
MPa using a compression molding press (Wabash Metal Products Co., Model 12-12-2T,
Wabash, Indiana). Cure time was taken when the maximum torque was reached using the cure rheometer APA 2000. After molding, the vulcanized samples were ground in the
Nelmor (01012M, N. Uxbridge, Massachusetts) grinding machine using a 5 mm screen.
128 Vulcanized sheets with dimensions of 127×127×2 mm3 were also obtained by compression molding at 160°C and they were used for the tensile test.
Ground rubber was fed into the co-axial ultrasonic reactor. Devulcanization experiments were performed at a barrel temperature of 120°C. The flow rate was in the range of 0.47 to 2.55 g/s. The gap size in the devulcanization zone was fixed at 2.54 mm.
The amplitudes of the ultrasonic waves were 5, 7.5 and 10 µm. The devulcanized samples were then mixed with the cure ingredients using the recipe in Table 3.3 in a two-roll mill.
Then it was revulcanized into sheets of 2 mm thickness at 160°C.
6.3 Curing and Revulcanization
Different rheometers were used when measuring the cure kinetics of IR in the present study and that of NR in earlier study123. Namely, the APA 2000 (Alpha
Technologies ) was used for IR and a Monsanto oscillating disc rheometer (ODR) was used for NR. Although these two rheometers typically show different levels of torques the general cure behavior would not be affected. Thus qualitative comparison of the cure curves can be made. Figure 6.1 shows the vulcanization and revulcanization curves for both rubbers. First, it is evident that both rubbers shared some similarity in the vulcanization and also in revulcanization. Particularly, both virgin rubbers have a comparative induction time of nearly 8 minutes and both of them also show a clear reversion at the later stage of curing which is due to the irreversible destruction of the polysulfidic crosslinks183, 184. Furthermore, in the revulcanization process, the induction period is completely absent because of the existence of accelerator residue in the system218, 230. This suggests that the revulcanization reaction starts immediately once the compounds have been heated to the desired temperature. However, some differences in 129 5 (a)IR
4
3
2
Torque, dNm virgin µ 1 5 m 7.5µm 10µm 0 0 5 10 15 20 25 Time, min 7 (b)NR 6
5
4
3 Torque, Nm Torque, 2 virgin 5µm 1 7.5 µm 10 µm 0 0 5 10 15 20 25 Time, min Figure 6.1 Cure curves at 160oC for IR (a) (APA) and NR (b)181 (ODR) of the virgin rubber and devulcanized rubbers obtained at various ultrasonic amplitudes with a die gap
of 2.54 mm, a flow rate of 0.63 g/s and a barrel temperature of 120oC
130 these two rubbers were also observed. For example, the torque of IR in revulcanization is lower than that in the vulcanization at all the amplitudes. It is also seen that with increase of the ultrasonic amplitude, the torque monotonously decreases, which suggests that the ability for the devulcanized rubber to undergo revulcanization has been weakened with the increase of ultrasonic amplitude. In contrast, the torque of NR in revulcanization only slightly drops compared to that in the vulcanization. Furthermore, the revulcanization torque of NR first drops with the initial increase of the ultrasonic amplitude from 5 to 7.5
µm, but it increases instead with the further increase of amplitude from 7.5 µm to 10 µm.
The torque value for 10 µm revulcanization is close to that in the vulcanization, which shows completely different trends with what is observed in IR. In addition, the dependence of initial torque on the amplitude for two rubbers is different. The initial torque decreases with the increase of amplitude for IR. In contrast, the initial torque has a minimum value at 7.5 µm for NR. The difference in the ability to undergo revulcanization for 10 µm devulcanizates is probably due to the fact that NR contains a higher concentration of cis-1, 4 isoprene units than IR does. Hence it is possible that upon devulcanization, NR retaining more unreacted C=C bonds than IR. This higher content of regular conjugated units could provide a higher probability of crosslinking contributing to better revulcanization. In addition, if one only considers the NR and IR vulcanization, one can observe that after the induction period the steeper slope and more severe drop of torque in NR indicates that the cure and reversion rate of NR is much higher than IR.
This may lead to the vulcanized IR sheets being more uniform than those of NR.
131 6.4 Power Consumption and Die Pressure
Generally, the amplitude of ultrasound, the pressure in the devulcanization zone and the residence time of rubbers in the reactor are the three major operating parameters which affect the degree and the rate of devulcanization. If the amplitude is too small or the pressure is too low, no devulcanization takes place. If the amplitude is sufficiently high, the devulcanization process is very fast and it usually takes place within several seconds. In fact, the mean residence time of rubber in the devulcanization zone is from
3.9 to 21.3 seconds based on the experimental flow rates of 2.55 to 0.47 g/s. Figure 6.2 and Figure 6.3 show, respectively, the power consumption and die pressure as a function of ultrasonic amplitude at different flow rates for IR (a) and NR (b) vulcanizates in the devulcanization process. During the devulcanization of both rubbers, die pressure drops with the increase of amplitude and a decrease of flow rate (Figure 6.3). It indicates that higher amplitude and longer residence times subject the materials to more intense treatment leading to a higher degree of devulcanization that in turn leads to a reduction of the viscosity. In contrast, for IR and NR vulcanizates the power consumption shows different behaviors as the amplitude increases. The power consumption of NR shows a maximum value at the intermediate amplitude of 7.5 µm and drops at both lower and higher ultrasonic amplitudes, except when the flow rate is as high as 2.52 g/s. Evidently, during the ultrasound extrusion, the devulcanization and revulcanization occur simultaneously, competing with each other. It was argued231 that unlike other rubbers such as SBR, BR and EPDM, the revulcanization of NR was dominated over the devulcanization at the highest amplitude and over a certain range of flow rates. At a very high flow rate, and hence the short residence time, it is possible that the broken NR 132 1200 (a)IR 1000
800
600 Flow rate, g/s
Power, watts 400 0.47 0.63 1.07 200 2.10 2.55 0 4567891011 µ Amplitude, m
1800 (b)NR 1500 Flow rate, g/s 0.32 1200 0.63 1.26 2.52 900
Power, watts 600
300
0 4 5 6 7 8 9 10 11 µ Amplitude, m
Figure 6.2 Power consumption for IR (a) and NR (b)181 rubbers devulcanized at
various flow rates at a barrel temperature of 120oC and a die gap of 2.54 mm
133 25 (a)IR Flow rate, g/s 0.47 20 0.63 1.07 2.10 15 2.55
10 Die pressure, MPa 5
0 024681012 Amplitude, µm
(b) Flow rate, g/s 20 NR 0.32 0.63 16 1.26 2.52
12
8 Die pressure, MPa Die pressure, 4
0 024681012 Amplitude, µm
Figure 6.3 Die pressure for IR (a) and NR (b)181 rubbers devulcanized at various flow
rates at a barrel temperature of 120oC and a die gap of 2.54 mm
134 chains do not have sufficient time to meet each other and to combine together
(revulcanization). Consequently, it leads to further devulcanization and results in a continuous increase of power consumption for NR at the high flow rate of 2.52 g/s.
However for IR, the maximum on the power consumption curve is not observed at all the flow rates applied. Instead, a continuous increase with increasing ultrasonic amplitude is observed. The difference in the behavior of power consumption with amplitude reveals that for NR the revulcanization could dominate over the devulcanization process at certain amplitude and over a certain range of flow rates. The exact reason for such a behavior is unclear. However, this is possibly due to the high content of regular structure of NR macromolecular chains. For IR, similar to other synthetic rubbers, devulcanization always prevails because of its lower content of regular structure than in NR. The explanation of the behavior of power consumption is also supported by the data obtained on crosslink density and gel fraction investigated in the following section.
6.5 Gel Fraction and Crosslink Density
The network structure of the vulcanizates, devulcanizates and revulcanizates was mainly characterized through the measurement of gel fraction and crosslink density.
Figure 6.4 and Figure 6.5 show, respectively, the gel fraction and the crosslink density of
IR and NR devulcanizates and revulcanizates as a function of the ultrasonic amplitude. It is seen that for IR devulcanizates the gel fraction and the crosslink density continuously drops with increasing amplitude. This is quite consistent with the dependence of power consumption on the amplitude for IR devulcanizates discussed in Section 6.4.
135 100 (a)IR
90
80
Flow rate, g/s Gel fraction, % 70 0.47 0.63 1.07 2.10 2.55 60 024681012 µ Amplitude, m
100 (b)NR 95
90
85
80
75
Gel fraction, % Gel fraction, Flow rate, g/s 70 0.32 0.63 65 1.26 2.52 60 024681012 µ Amplitude, m
Figure 6.4 Gel fraction of the devulcanized (solid symbols) and revulcanized (open rectangle, for flow rate of 0.63g/s only) IR (a) and NR (b)181 rubbers as a function of ultrasonic amplitude obtained at various flow rates, a die gap of 2.54 mm and a barrel temperature of 120oC
136 0.25 Flow rate, g/s
3 0.47 0.20 0.63 1.07 2.10 2.55 0.15
0.10
0.05 Crosslink density,Crosslink kmol/m (a) IR 0.00 024681012 µ Amplitude, m
0.25 Flow rate, g/s 0.32 3 0.63 0.20 1.26 2.52 0.15
0.10
0.05 Crosslink density, kmol/m (b)NR 0.00 024681012 µ Amplitude, m
Figure 6.5 Crosslink density of the devulcanized (solid symbols) and revulcanized (open rectangle, for flow rate of 0.63g/s only) IR (a) and NR (b)181 rubbers as a function of ultrasonic amplitude obtained at various flow rates, a die gap of 2.54 mm and a barrel temperature of 120oC
137 The experimental results of NR devulcanizates show that the gel fraction and the crosslink density initially decrease when the ultrasonic amplitude increases from zero to
7.5 µm but then increase with further increase of amplitude from 7.5 to 10 µm. This behavior is quite different from the occurrence in IR devulcanizates. However, it is still consistent with the dependence of power consumption on the amplitude in NR devulcanizates. The observed power consumption and network structure change with amplitude during NR devulcanization suggest that when the amplitude is increased from
7.5 to 10 µm, the revulcanization process dominates over the devulcanization, which is the dominating process at lower amplitudes. This unique phenomenon in NR could again be due to its highly regular main chain structure. The gel fraction and the crosslink density of both rubber devulcanizates obtained at the flow rate of 0.63g/s and revulcanized by the recipe shown in Table 3.3 are evaluated and shown in Figure 6.4 and
Figure 6.5 by open symbols). The NR revulcanizates exhibit substantially higher gel fraction and crosslink density than their respective devulcanizates. The IR revulcanizates at 5 and 7.5 µm show a gel fraction close to the corresponding devulcanizate. At an amplitude of 10 µm, IR revulcanizate shows higher gel fraction than its devulcanizates.
Generally, the crosslink density of the NR revulcanizates is higher than that of the respective devulcanizates. These findings indicate that NR exhibits a better ability to undergo revulcanization than IR. The results also indicate that in the presence of additional sulfur, the devulcanized rubbers are capable of undergoing revulcanization. It also indicates that the entire amount of C=C bonds is not consumed in the vulcanization of virgin rubbers. There remained a fair amount of C=C bonds left in the devulcanized samples. Further calculations of the relative amount of C=C bonds and sulfur point 138 towards this indication. For example, by taking the molecular weight of repeating units in the rubber and sulfur 68.11 and 32, respectively, in the recipe shown in Table 3.2, one obtains the content of the C=C bonds equal to 100(g)/68.11(g/mol) = 1.468 mol and that of the sulfur equal to 2(g)/32(g/mol) = 0.0625 mol. This result shows a large amount of unsaturated units is left in the vulcanizates. They preserve the ability of the devulcanized rubber to undergo revulcanization.
6.6 Molecular Characteristics of the Devulcanized Sol
The molecular characteristics of the sol parts of the devulcanized IR rubbers were determined by GPC. The results are shown in Figure 6.6. The measured number (Mn) and weight (Mw) average molecular weight of the virgin IR were 982,000 and 1,998,000, respectively. In Figure 6.6a, curves of the molecular weight distribution of the sol part extracted from the devulcanized rubber are shifted to the lower molecular weight compared to that of the virgin IR. This indicates that the molecular weight of sol generated during ultrasonic devulcanization is decreased significantly upon ultrasonic treatment. Furthermore, the molecular weight depends on the amplitudes of treatment.
For example, at amplitudes of 7.5 and 10 µm, the sol of the devulcanized IR shows both low and high molecular weight tails, compared with that of 5 µm (Figure 6.6a).
Accordingly, the polydispersity (Mw/Mn) also varies with ultrasonic amplitude (Figure
6.6b). The value of Mw/Mn for the 5 and 7.5 µm samples is lower than that of virgin IR.
At the same time, the value of Mw/Mn of the 10 µm is higher than that of virgin IR. This shows that the ultrasonic amplitude affects the molecular weight of sol and its distribution in a very complex way.
139 2.0 (a)
µ 1.5 5 m ) w IR 1.0 7.5µm /d(logM f 10µm dW 0.5
0.0 2345678 logM w 107 3.6 (b)
3.2 106 2.8 n w 5 /M
10 2.4 w M M
2.0 104 1.6
103 1.2 024681012 µ Amplitude, m Figure 6.6 Molecular weight distribution at various amplitudes (a), amplitude
dependence of weight average molecular weight, Mw, and polydispersity, Mw/Mn (b) of the sol parts in devulcanized IR obtained at a gap of 2.54 mm, a flow rate of 0.63 g/s and
o a barrel temperature of 120 C. Mw and Mw/Mn of the virgin IR are also shown.
140 It is speculated that at lower amplitudes, such as 5 µm, breaking of the macromolecular chains is less severe and thus smaller amounts of macroradicals are created compared with that at higher amplitudes; the result would be the molecular weight reduction. At higher amplitudes such as 7.5 and 10 µm, more radicals are produced due to the exposure of more intense ultrasound (intensity of the power is proportional to the square of the amplitude). Therefore, it is possible that these radicals terminated either by disproportionation leading to lower molecular weight tails or by coupling with several macroradicals leading to higher molecular weight tails.
6.7 Rheological Properties
The rheological properties of the devulcanized rubbers are very important as far as the processing is concerned. The dynamic properties of virgin IR, its vulcanizate and devulcanizates were obtained by APA 2000. The results are shown in Figure 6.7 and
Figure 6.8. Upon cure, the complex viscosity increases by at least one order of magnitude compared with the virgin uncured rubber (Figure 6.7a). After devulcanization, the complex viscosity is obviously lower than that of the cured rubber. With the increase of amplitude the complex viscosity drops even more. Particularly, the viscosity of the 5 µm sample is only slightly lower than that of the cured sample. However, it is still much higher than the viscosity of the uncured rubber over the entire frequency range. For the
7.5 and 10 µm samples, the viscosity is much lower than that of the cured sample. At a low frequency range, the viscosity of these samples is higher than that of the uncured rubber but at a relatively high frequency range, their viscosity is lower. This indicates that different degrees of shear thinning take place in the virgin uncured IR, the cured rubber and the devulcanized samples obtained at different amplitudes. The cured and 141 devulcanized rubbers exhibit greater shear thinning behavior due to the presence of a substantial amount of croslinking structure compared with virgin uncured IR consisting of 100% sol. This effect can be more clearly observed in Figure 6.7b, where the complex viscosity was plotted as a function of the product of the viscosity and frequency. From this figure, the slope of the curve changes with the level of crosslinking of the materials; with the virgin vulcanizates and the 5 µm devulcanized sample exhibiting the greatest slope and the virgin uncured rubber showing the smallest slope. Furthermore, the loss tangent of the virgin uncured IR decreases with frequency (Figure 6.7a). Among all the rubbers, its value is the highest over almost the entire frequency range. For the devulcanized rubber the loss tangent progressively increases with the increase of amplitude.
The storage (G’) and loss (G”) modulus versus frequency for virgin IR, its vulcanizate and devulcanizates are shown in Figure 6.8. Virgin IR has the highest G”.
The cured IR and the devulcanizates at 5 µm have the lowest G”. Due to the very low values of tan δ for the cured and devulcanized IR at 5 µm, it is difficult for the instrument to separate the imaginary part of G*, i.e., G”. Therefore, it leads to the waveness of G” for these two samples. The devulcanizates at 7.5 and 10 µm have the G” in between the gum and cured IR. The plateau modulus (G’) is observed in the virgin cured sample and the 5 µm devulcanized sample indicating the occurrence of the full cure for the virgin cured sample and only a slight de-crosslinking during the devulcanization at the ultrasonic amplitude of 5 µm. With a further increase of amplitude from 7.5 to 10 µm, more intense devulcanization takes place such that the storage modulus decreases and the
142 plateau modulus gradually disappears. From all of these figures (Figure 6.7a, b and
Figure 6.8) it is observed that the most significant change in properties takes place when the amplitude is increased from 5 to 7.5 µm, compared with the changes occurring when the amplitude increased from 0 to 5 or 7.5 to 10 µm. Therefore, from the rheological properties of these materials one can conclude that a higher degree of devulcanization can be achieved at higher ultrasonic amplitudes.
6.8 Mechanical Properties
The stress-strain curves for the IR and NR vulcanizates and revulcanizates at different ultrasonic amplitudes are shown in Figure 6.9. The cured virgin IR vulcanizates show the tensile strength (σB) to be slightly lower than that of the virgin NR vulcanizates.
This is attributed to the higher molecular weight and narrower molecular weight
182 distribution of IR than those of NR (Mn: 180,400; Mw: 1,116,000; Mw/Mn: 6.19) . Also
NR has a higher amount of cis- 1, 4 isoprene structure (>99%) than IR (98%). The higher degree of stereo-regularity would introduce a stronger ability for stress-induced crystallization which is evident by the rise of the tensile stress at lower strain. Evidently, stereo-regularity plays a more important role in the mechanical properties than the molecular weight and its distribution. From Figure 6.9, it is seen that both rubbers show a different extent of the stress-induced crystallization leading to a high tensile strength, which is a typical phenomenon for NR. It is also noticed that IR vulcanizates and revulcanizates show higher elongation compared with the respective NR vulcanizate and revulcanizates.
143 1010 1.0 (a) |η*| tan δ 109 virgin IR IR cured 0.8 µ 8 5 m 10 7.5µm 10µm 0.6 107 δ Pa-s 6 |, |, 10 0.4 * tan η | 105 0.2 104 0.0 103
102 10-2 10-1 100 101 102 103 ω , rad/s
108 (b)
107
106
105 *|, Pa-s η | 104 virgin IR IR cured 5µm 103 7.5µm 10µm 102 104 105 106 η ω | *| , Pa
Figure 6.7 The complex viscosity |η*| and tan δ of virgin IR, the vulcanizates and the devulcanizates obtained at a gap of 2.54 mm, flow rate of 0.63 g/s and a barrel temperature of 120oC as a function of frequency ω (a) and |η*|ω (b), respectively, at a strain amplitude of 4.2%
144 103 (a)
102 G', kPa G', virgin IR IR cured 5µm 7.5µm 10µm 101 10-2 10-1 100 101 102 103 ω , rad/s 102 (b)
101
100 G", kPa G", virgin IR IR cured -1 10 5µm 7.5µm 10µm 10-2 10-2 10-1 100 101 102 103 ω , rad/s
Figure 6.8 Storage (a) and loss (b) modulus of the virgin IR, the vulcanizates and the devulcanizates obtained at a gap of 2.54 mm, a flow rate of 0.63 g/s and a barrel temperature of 120oC as a function of frequency ω at 120oC at a strain amplitude of 4.2%
145 20 IR NR virgin
15 5 µm
10 , MPa
σ 7.5 µm
5
10 µm
0 0 200 400 600 800 1000
ε, %
Figure 6.9 The stress-strain curve for IR and NR vulcanizates and revulcanizates
obtained at different ultrasonic amplitudes, a die gap of 2.54 mm, a flow rate of 0.63 g/s
and a barrel temperature of 120oC
Figure 6.10 shows the mechanical properties of IR and NR vulcanizates and revulcanizates as a function of ultrasonic amplitude. In Figure 6.10 (a), when the amplitude increases from zero to 7.5 µm, the tensile strength (σB) decreases for both IR and NR revulcanizates. However, when the amplitude increases from 7.5 to 10 µm, the change of the tensile strength for these two rubbers is different. In particular, the tensile strength for IR revulcanizates continuously drops at the amplitude of 10 µm. In contrast, the strength for NR revulcanizates increases. This is attributed to revulcanization dominating over the devulcanization for NR at the ultrasonic amplitude of 10 µm as discussed earlier. This phenomenon is consistent with the power consumption, gel
146
22 1000 (a) 20 18 800 ε 16 B 14 600 , % , MPa B B 12 ε σ 10 σ 400 B 8 6 IR 200 NR 4 0 2 4 6 8 10 12 µ Amplitude, m 1.0 (b) 0.9
E100 0.8
0.7 , MPa E300 300
, E 0.6 100 E 0.5 IR NR 0.4 024681012 µ Amplitude, m
Figure 6.10 Amplitude dependence of the tensile strength (a), elongation at break (b) and modulus at 100% and 300% (c) of IR and NR vulcanizates and revulcanizates at a die gap of 2.54 mm, a flow rate of 0.63 g/s and a barrel temperature of 120oC
147 fraction and crosslink density measurements reported in Section 6.4 and 6.5. From Figure
6.10 (a), generally the elongation at break (εB) for the NR revulcanizates shows the value close to its virgin vulcanizate. However, for the IR revulcanizates, the εB values are lower than that of virgin vulcanizate. Particularly, the εB value decreases linearly with the increase of the amplitude. Overall the εB values of IR vulcanizate and revulcanizates are higher than those of respective NR samples. In Figure 6.10 (b), the change of modulus at
100% and 300% strain, E100 and E300 for the NR revulcanizates, with the amplitude shows the same trend as the change of tensile strength. Namely, the modulus initially drops when the amplitude increased from zero to 7.5 µm, but it increases when the amplitude increases from 7.5 to 10 µm. For IR revulcanizates, the modulus generally drops with the increase of amplitude except at the amplitude of 5 µm. The modulus at this amplitude is slightly higher than that of the virgin IR vulcanizate.
6.9 Thermal Properties
The thermal properties of the virgin, the cured and the devulcanized unfilled IR at various ultrasonic amplitudes were analyzed by DSC as shown in Figure 6.11. The glass transition temperature Tg was determined by the inflection point method. The results are summarized in Table 6.1. It is observed that the virgin IR has the lowest Tg value among those samples. This is reasonable since the virgin IR is the only sample free of chemical crosslinks and therefore it retains the highest molecular chain mobility. However, there is no significant difference of Tg between the devulcanized samples and the cured sample.
148 virgin IR
cured IR
devulc-5µm
devulc-7.5µm
devulc-10µm
Figure 6.11 DSC curves for the virgin, cured and devulcanized IR under the nitrogen
atmosphere
Table 6.1 Tg of the virgin, cured and devulcanized IRs
o Rubbers Tg, C
Virgin IR gum -62.2
Virgin IR cured -59.0
Devulcanized at 5 µm -59.4
Devulcanized at 7.5 µm -59.6
Devulcanized at 10 µm -59.3
149 6.10 Conclusions
Sulfur-cured vulcanizate of the unfilled synthetic isoprene rubber (IR) was devulcanized using a coaxial ultrasonic reactor in the same way as natural rubber was done earlier. The two rubbers showed some similarities in the vulcanization, devulcanization, revulcanization, and network structure. They had an induction period close to each other and both of them showed a reversion in the original vulcanization.
The induction period was absent in the revulcanization. The die pressure during devulcanization continually dropped with the increase of ultrasonic amplitude. At all the ultrasonic amplitudes, the gel fraction and crosslink density were lower than the original cured rubber.
There were some clear differences in the rate of vulcanization, the level of revulcanization and the dependence of the degree of devulcanization on the amplitude between IR and NR. In particular, the cure rate of IR was lower than that of NR. This could make the vulcanized sheets of IR more uniform than those of NR. Unlike the IR, the NR sample devulcanized at the highest amplitude 10 µm could reach a higher level of revulcanization than those of the 5 and 7.5 µm samples. This was revealed by the torque value attained upon revulcanization, the value of gel fraction and crosslink density, the value of power consumption and also the mechanical properties.
The dynamic properties of the IR devulcanizates indicated that at the higher amplitude, the complex viscosity became lower and the loss tangent became larger. These results were consistent with the gel fraction and crosslink density as well as the power consumption. A lower degree of shear thinning of samples devulcanized at higher ultrasonic amplitude was observed. Also, more significant changes of all the properties 150 took place when the amplitude went from 5 to 7.5 µm compared with those variations when the amplitude increased from 0 to 5 µm and 7.5 to 10 µm.
The glass transition temperature of virgin IR was lower than the cured and devulcanized unfilled IR. There was no significant change in the glass transition temperature after the cured rubber was devulcanized at various ultrasonic amplitudes indicating a slight change of the network chain mobility from the ultrasonic treatment.
151
CHAPTER VII
7.ULTRASONIC DEVULCANIZATION OF THE CB FILLED IR
7.1 General
Although the properties of gum rubbers can be improved through the vulcanization process, they alone still can not be fully used in the high performance applications such as tires. Instead, they have to be used in the combination with fillers such as carbon black or silica. The addition of carbon black is significant in improving the modulus, hardness, tensile strength, abrasion and tear resistance as well as resistance to fatigue and cracking. Black loaded IR compounds are used in a wide variety of applications, such as tires, motor mounts, pipe gaskets, shock absorber bushings and many other molded and mechanical goods.
The biggest waste rubber sources are the waste rubber tires. Tires are made up of complex components including several rubbers (NR, IR, BR, SBR, EPDM), fibers (such as PET, Nylon, cotton), inorganic substances (such as steel, glass) and fillers (such as carbon black, silica) and so on43. Unlike the plastics, rubbers such as tires are difficult to recycle due to the three-dimensional network structures formed during vulcanization.
Therefore, a direct reuse of the waste rubber tires is impossible. The crosslinking network has to be partially destroyed with the aid of some energy. A wide variety of recycling
152 methods based on different formats of energy were explored including chemical method, mechanical method, microwave method, biotechnical method, ultrasonic method etc. In this chapter the devulcanization of carbon black filled IR was developed in the co-axial ultrasonic reactor to help our understanding on the devulcanization process. The effect of carbon black on the cure kinetics, devulcanization process, the network structures (gel fraction and crosslink density), rheological and mechanical properties was investigated.
The influence of the processing oil and the retarder on the devulcanization and revulcanization process as well as the properties was also studied. The similarities and the differences between the devulcanization of filled IR and the filled NR studied earlier were also discussed.
Devulcanized rubber alone usually can not be used in tire application. It has to be blended with virgin rubber compounds. Another aspect of research in this chapter was to study the blending of the devulcanized filled rubber with the virgin filled rubber at varied blending ratios with the purpose to obtain the properties comparable to the original rubbers. The cure kinetics and mechanical properties were also investigated.
7.2 Experimental
NATSYN® 2200 IR was used in the filled rubber experiments. Various amount of
CB (HAF N330, Sid Richardson Carbon Company, Fort Worth, TX) and other cure ingredients were added as shown in Table 3.2. Virgin gum IR was mixed with CB in the
Banbury internal mixer (Model 86EM9804, Banbury USM Corp., Ansonia, CT) for 10 minutes in order to breakdown the CB agglomerates. In the case of oil containing compounds, the processing oil was added into the internal mixer together with CB and the mixing also took 10 minutes. The rotor speed in mixing was 30 rpm. Then the filled 153 rubbers were compounded with the curing ingredients: ZnO, stearic acid and CBS in the two-roll mill. Sulfur was the last ingredient to be added to the compounds. The mixing in the two-roll mill took about 40 passes. The compounds were molded into slabs
(260×260×12 mm3) at a temperature of 160°C and a pressure of 17.2 MPa using a compression molding press (Wabash Metal Products Co., Model 12-12-2T, Wabash, IN).
Cure time was taken when the maximum torque was reached using the cure rheometer
APA 2000. After molding, the vulcanized samples were ground in the Nelmor (01012M,
N. Uxbridge, Massachusetts) grinding machine using a 5 mm screen. Vulcanized sheets with dimensions of 127×127×2 mm3 were also obtained by compression molding at
160°C and they were used for the mechanical test.
The ground rubber was fed into the co-axial ultrasonic reactor via the hopper loaded with a feeder. The feeder, providing the “starved feed” to the extruder, controlled the output. Devulcanization experiments were performed at a barrel temperature of
120°C. The flow rate was 0.63 g/s. The gap size in the devulcanization zone was 2.54 mm. The ultrasonic treatment of the rubber was occurred in the gap. The amplitudes of the ultrasonic waves were 5, 7.5 and 10 µm. During the devulcanization process, the ground rubber was compressed and conveyed by the screw to the devulcanization zone.
After reaching the steady state conditions indicated by the pressure transducer and the ultrasonic power wattmeter, ultrasonically treated samples were collected. The entrance pressure before the ultrasonic treatment zone and the ultrasonic power consumption were measured. The devulcanized samples were processed on the two-roll mill for 3-5 passes and then were mixed with the cure ingredients using the recipe in Table 3.3. Thereafter, it was revulcanized into sheets of 2 mm thickness at 160°C. 154 The devulcanized rubber was also blended with the virgin filled IR at varied blending ratios (25/75, 50/50 and 75/25) in a two-roll mill. The cure kinetics experiment was carried out on the APA 2000 at 160oC, a strain amplitude of 4.2% and a frequency of
100 cpm. The mechanical properties were measured on the Instron Tensile Tester 5567 at room temperature and an extension speed of 500 mm/min.
7.3 Vulcanization of the Virgin Filled IR without the Processing Oil
Before the study of the ultrasonic devulcanization of filled IR, the characterization of the virgin filled IR has to be made. This includes the cure kinetics of the virgin filled
IR, the gel fraction and crosslink density of the cured IR and their mechanical properties.
7.3.1 Curing
Before studying the devulcanization of CB filled IR, curing has to be taken for the filled compounds. Figure 7.1 shows the cure curves of the unfilled IR and the filled IR at the CB loading ranging from 15 to 60 phr at 160oC. Clearly with the increase of CB loading, the induction time is shorter and the curing is faster. As discussed in Chapter IV, this is explained by the acceleration effect of CB on the vulcanization. The reversion is seen to be more severe in CB filled IR. The obtained induction time following the method described in Figure 4.4 as a function of CB loading is shown in Figure 7.2. An exponential reduction of induction time with the increase of CB loading is observed.
From Figure 7.1, a tremendous increase of the maximum torque particularly at high CB loadings is observed. The addition of CB causes the increase of modulus of the rubber creating more resistance to the moving die in the rheometer. However, this is not the case for the minimum torque. The minimum torque at 15 phr is surprisingly lower than that of the unfilled IR. It is possible that the main chain breakage of IR takes place 155 during mixing with CB in the Banbury internal mixer. In order to clarify this, the following experiments were designed. Firstly the virgin gum rubber was processed in the
Banbury mixer for 10 minutes so that the gum IR experienced the exact same mechanical treatment as the IR containing the 15 phr CB. Then the viscosity of the treated gum IR was measured and compared with the virgin untreated gum IR. This comparison is shown in Figure 7.3. The reduction of viscosity indicates the mastication effect that gum IR has experienced in the Banbury internal mixer.
16 60 phr 14
12
10 35 phr
8 25 phr 6 15 phr Torque, dNm Torque, 4 0 phr 2
0
0 5 10 15 20 25
Time, min
Figure 7.1 Cure kinetics of the filled IR at various CB loadings at 160oC (strain
amplitude: 4.2%, frequency: 100 cpm)
156 10 -0.06*CB(phr) o ti=8.43*e (160 C) 8
6
4 , min i t
2
0
0 20406080100
CB, phr
Figure 7.2 Isothermal induction time versus the CB loading at 160oC
157 7 10 o 160 C Virgin IR IR in Banbury 106
105 *|, Pa-s
η 4 | 10
103
102 10-2 10-1 100 101 102 103 ω , rad/s
Figure 7.3 Complex viscosity versus the frequency for the virgin IR and IR processed
in the Banbury mixer (strain amplitude: 4.2%)
Next, the gum IR processed in the Banbury mixer was cured using the recipe that was used for the unfilled rubber as shown in Table 3.2. The comparison with the virgin unfilled and the 15phr filled IR is made in Figure 7.4. In Figure 7.4, the minimum torque of the unfilled IR processed in Banbury mixer is lower than that of the virgin unfilled IR.
This observation is in accordance with the measurement of viscosity shown in Figure 7.3.
It also reveals that the rubber main chain degradation occurred in the Banbury internal mixer. In addition, it is expected that the minimum torque of 15 phr filled virgin IR would be higher than that of the unfilled IR processed in the Banbury mixer. However, it is not the case. The minimum torque of the unfilled IR processed in the Banbury mixer is 158 close to that of the 15phr filled virgin IR. Evidently, more degradation of IR occurs during the mixing of IR with CB in the Banbury mixer.
6 15phr virgin IR
5 unfilled IR in Banbury 4
3 unfilled virgin IR Torque, dNm 2
1
0 0 5 10 15 20 25
Time, min
Figure 7.4 Cure curves of various IRs at 160oC (strain amplitude: 4.2%, frequency:
100 cpm)
7.3.2 Gel Fraction and Crosslink Density
Figure 7.5 shows the gel fraction and the crosslink density of the rubber vulcanizates as a function of CB loading. Except the sample with the 15 phr carbon black, the gel fraction and the crosslink density of the filled IR vulcanizates are close to the values of the unfilled IR vulcanizate. It is generally recognized that the interaction of CB with the rubber matrix causes the formation of bound rubber layer at the surface of CB.
This portion of rubber is difficult to extract by the solvent. Accordingly, it contributes to 159 the increase of apparent crosslink density and gel fraction. However, after the Kraus correction, this effect is minimized. Therefore, the true crosslink density and gel fraction in Figure 7.5 shows a weak dependence on CB loading. This result is similar to that of
NR studied earlier182 and also in agreement with the results of Mori and Koenig232 who found that the total amount of sulfurization reaction occurring in NR is almost independent of the level of CB loading.
1.00 400 3 0.90 300
0.80 200 Gel fraction Gel
0.70 100 Crosslink density, mol/m
0.60 0 0 204060 CB, phr
Figure 7.5 Gel fraction and crosslink density of virgin IR vulcanizates versus carbon
black loading
7.3.3 Mechanical Properties
Figure 7.6 shows the stress-strain curves of the virgin unfilled and the filled IR vulcanizates at various carbon black loadings. Similar to NR, due the high content of stereo-regular structure in IR, it is able to undergo crystallization upon stretching leading to very high tensile strength and elongation at break. This indicates that the addition of 160
500mm/min CB, phr 30 0 15 25 35 60 20 , MPa σ
10
0 0 200 400 600 800 1000 1200 ε , % Figure 7.6 Stress-strain curves of the virgin IR vulcanizates carbon black does not deteriorate the ability of IR to undergo the stress-induced crystallization. Particularly, with the increased loading of carbon black, the strain- induced crystallization starts at lower strain. The obtained tensile strength, elongation at break and modulus as a function of carbon black loading are plotted in Figure 7.7. The elongation at break decreases and the modulus increases with the increase of carbon black loading. The increase of modulus with the carbon black loading is not linear59. In contrast, the tensile strength shows an optimum value at 35 phr CB loading. This is consistent with the result obtained earlier for NR182 where the optimum value of tensile strength was found at the same carbon black loading. With the further increase of carbon black above the 35 phr, the reinforcement of carbon black diminishes. This suggests that
161 35 1000 (a) 30 800 25
20 , %
, MPa 600 B B ε σ 15
10 400 σ 5 B ε B 0 200 0 10203040506070
CB, phr 8
E100 (b) 7 E300 6
5 , MPa
300 4 , E
100 3 E 2
1
0 0 10203040506070
CB, phr Figure 7.7 Tensile strength σB, elongation at break εB (a) and moduli at 100 and
300% strain, E100 and E300 (b) as a function of carbon black loading for the virgin IR
vulcanizates
162 the reinforcement is controlled by the volume effect of carbon black. Based on this result, the 35 phr CB filled IR is chosen for the further investigation on devulcanization, revulcanization, rheology, the effect of processsing oil and retarder.
7.4 Devulcanization of the Filled IR without the Processing Oil
The devuclanization of the filled IR was made in the coaxial ultrasonic reactor.
During the process, the power consumption and die pressure were measured. The gel fraction and crosslink density of the devulcanized and revulcanized rubbers were determined. The dynamic properties of the virgin uncured, cured and devulcanized rubbers were evaluated. Finally, the mechanical properties of the revulcanized rubber were determined.
7.4.1 Power Consumption and Die Pressure
Similar to the unfilled IR, the devulcanization of the black filled IR was also carried out in the co-axial ultrasonic reactor. During the extrusion, the flow rate was 0.63 g/s; the gap between the die and the ultrasonic horn was 2.54 mm and the barrel temperature was 120oC. Figure 7.8 shows the power consumption as a function of the ultrasonic amplitude during the devulcanization process. For the 60 phr filled IR, when the ultrasonic amplitude was increased from 7.5 µm to 10 µm, the ultrasonic system was overloaded. Therefore, the measurement of power consumption is not possible.
Accordingly, the material to be devulcanized at 10 µm for the 60 phr IR is not available.
Nevertheless, it is observed that at any CB loading, the power consumption increases with the increase of ultrasonic amplitude. This is in contrast with the existence of maximum power consumption that occurred at the amplitude of 7.5 µm in the filled
NR182. As was explained in Chapter V, the exclusively high streroregular structure of NR 163 main chains imparts the ability of NR to undergo prevailing revulcanization than devulcanization at 10 µm and thus leads to reduced power consumption when amplitude increases from 7.5 µm to 10 µm. However, due to the less content of cis-1, 4 isoprene units in IR, the devulcanization is always dominated over the revulcanization at any amplitude during ultrasonic devulcanization of IR.
1400 CB, phr 0 1200 15 25 1000 35 60 800
600
Power consumption, watts 400
200 4567891011 µ Amplitude, m
Figure 7.8 Power consumption as a function of ultrasonic amplitude during the
devulcanization of the filled IR at various CB loadings
Figure 7.9 shows the dependence of die pressure on ultrasonic amplitude during the devulcanization of the unfilled and the CB filled IR vulcanizates. The die pressure decreases as the amplitude increases. This trend is independent of the CB level. This has been previously explained as the combined effect of the rubber softening due to the devulcanization in the die gap and the reduction in the friction between the particles and
164 the die walls due to ultrasonic vibrations118. However, the die pressure is seen to increase with the increase of filler loading due to more flow resistance generated by CB.
35 CB, phr 30 0 15 25 25 35 60 20
15 Die pressure, MPa pressure, Die
10
5 024681012 µ Amplitude, m
Figure 7.9 Die pressure as a function of ultrasonic amplitude during the
devulcanization of filled IR at various CB loadings
7.4.2 Gel Fraction and Crosslink Density
The gel fraction and the crosslink density of the devulcanized IR (open symbols) as a function of ultrasonic amplitude at various CB loadings are plotted in Figure 7.10 and Figure 7.11, respectively. The results for devulcanized IR are different from what was observed in the study of the unfilled and the filled NR where a minimum of gel fraction and crosslink density was found at the intermediate amplitude of 7.5 µm182. The gel fraction and the crosslink density for both the unfilled and the filled IR continuously decreases as the ultrasonic amplitude increases. This difference in the devulcanization
165 behavior of IR and NR probably results from the stereoregular structure differences of the two rubbers. NR has a higher amount of cis- 1, 4 isoprene structure (>99%) than IR
(98%). The higher amount of stereoregular structure of NR imparts the ability of NR to undergo prevailing revulcanization than devulcanization during ultrasonic devulcanization at 10 µm. In contrast, the devulcanization is always dominated over revulcanization at any amplitude in ultrasonic devulcanization of IR. As a result, the increase of gel fraction and crosslink density at 10 µm was occurred in NR182. The decrease of gel fraction and crosslink density with the increase of amplitude is observed in IR (Figure 7.10 and Figure 7.11). This result is consistent with the measurement of power consumption in Section 7.4.1.
1.00
0.95
0.90
0.85 CB, phr Gel fraction 0.80 0 15 35 0.75 35 (revulc)
0.70 024681012 µ Amplitude, m Figure 7.10 Gel fraction of the devulcanized (open symbols) and revulcanized (solid triangle, for 35 phr only) filled IR as a function of ultrasonic amplitude obtained at flow rate of 0.63 g/s, a die gap of 2.54 mm and a barrel temperature of 120oC
166 300 CB, phr 0 250
3 15 35 200 35 (revulc)
150
100
Crosslink density, mol/m density, Crosslink 50
0 024681012 µ Amplitude, m
Figure 7.11 Crosslink density of the devulcanized (open symbols) and revulcanized
(solid triangle, for 35 phr only) filled IR as a function of ultrasonic amplitude obtained at
flow rate of 0.63 g/s, a die gap of 2.54 mm and a barrel temperature of 120oC
7.4.3 Rheological Properties
The rheological properties of the filled IR containing the 35 phr carbon black were studied. This experiment was carried out on the APA 2000 at a strain amplitude of
4.2% and a temperature of 120 oC for the virgin filled uncured, the cured and the devulcanized rubber. The complex viscosity and the loss tangent as a function of frequency are plotted in Figure 7.12a. It is observed that the viscosity of the cured rubber is substantially higher than that of the uncured within the whole range of frequency. This is due to the formation of a three-dimensional network creating the resistance to the flow of the material. Upon devulcanization at three different ultrasonic amplitudes, the viscosity decreases compared to that of the cured rubber. However, it is still higher than 167 that of the virgin uncured sample. This suggests that the ultrasonic devulcanization only partially disrupts the network structures. The viscosities of the samples devulcanized at
7.5 and 10 µm are close to each other. However, they are lower than that of the 5 µm sample. The above observation indicates that the viscosity is closely correlated with the degree of crosslinking. The dependence of the loss tangent on the degree of crosslinking of the samples is contrary to that of the viscosity on the crosslink density. The uncured sample which is free of chemical crosslinks has the largest loss tangent value. The cured sample which has the highest crosslink density among the five samples has the smallest loss tangent value. The loss tangent value of the devulcanized samples is in between these two limits. The loss tanglent is increased with the increase of ultrasonic amplitude.
Different degrees of shear thinning of these samples are shown in Figure 7.12b.
Generally the samples containing the chemical crosslinking structures (cured and devulcanized) show a higher degree of shear thinning than the one uncured. Three samples devulcanized at different ultrasonic amplitudes show similar degree of shear thinning. This is in contrast to the devulcanized unfilled IR discussed in Chapter VI
Section 6.7, where the 5 µm devulcanized unfilled IR exhibited higher degree of shear thinning. The similar degree of shear thinning in the devulcanized filled IR samples is possibly due to the same amount of CB added (35 phr) and also the slight differences in the crosslink density among these three samples shown in Figure 7.11.
The storage (G’) and loss (G”) modulus as a function of frequency is plotted in
Figure 7.13. In Figure 7.13 (a), the plateau modulus (G’) is only observed in the case of virgin cured 35 filled IR indicating the occurrence of the full cure in this particular sample. Upon ultrasonic devulcanization of the filled IR, not only the storage modulus 168 1010 1.0 (a) |η*| tan δ virgin IR/35CB 109 IR/35CB cured 5µm 0.8 108 7.5µm 10µm 0.6 107 δ Pa-s *|,
6 tan
|η 10 0.4
105 0.2 104
103 0.0 10-2 10-1 100 101 102 103 ω , rad/s 108 (b)
107
106
*|, Pa-s *|, 5
|η 10 virgin IR/35CB IR/35CB cured µ 104 5 m 7.5µm 10µm 103 105 106 |η ω *| , Pa Figure 7.12 Complex viscosity |η*| and tan δ of virgin 35 phr filled IR, vulcanizates and devulcanizates obtained at a gap of 2.54 mm, a flow rate of 0.63 g/s and a barrel temperature of 120oC as a function of frequency ω (a) and |η*|ω (b), respectively, at a strain amplitude of 4.2%
169 (a)
100 G', kPa 10-1 virgin IR/35CB IR/35CB cured 5µm 7.5µm 10µm 10-2 10-2 10-1 100 101 102 103 ω , rad/s
(b)
10-1 G", kPa G", virgin IR/35CB IR/35CB cured 5µm 7.5µm 10µm 10-2 10-2 10-1 100 101 102 103 ω , rad/s
Figure 7.13 Storage (a) and loss (b) modulus of the virgin 35 phr filled IR, vulcanizate and devulcanizates obtained at a gap of 2.54 mm, a flow rate of 0.63 g/s and a barrel temperature of 120oC as a function of frequency ω at a strain amplitude of 4.2%
170 decreases, but also the plateau modulus is not evident due to the breakage of crosslink structure. In Figure 7.13 (b), the dependence of the loss modulus on the frequency for the uncured rubber is the strongest among the five samples; while that for the cured rubber is the weakest. The dependence of the loss modulus on the frequency for the devulcanized rubbers is in between these two limits.
7.4.4 Revulcanization
The revulcanization was carried out on the APA 2000 at 160oC, a strain amplitude of 4.2% and a frequency of 100 cpm. Figure 7.14 shows the revulcanization curves of the
35 phr filled IR devulcanized at three different ultrasonic amplitudes. As a comparison, the cure curve of the virgin filled IR is also plotted in the same graph. A different curing behavior of the virgin and devulcanized filled IR is observed. In particular, the induction time is present in the vulcanization of virgin filled IR. In contrast, the induction time is absent in the revulcanization. This effect is independent of the ultrasonic amplitude in the devulcanization. This result is similar to that occurred in the revulcanization of the unfilled IR shown earlier in Chapter VI Section 6.3. Again, it is due to the presence of accelerator residue in the rubber after devulcanization218.
In addition, the final torque in revulcanization of the devulcanized rubber is lower than that in curing the virgin rubber. This possibly results from the dual effects: one is the partial main chain degradation occurring in the devulcanization, another is the network destruction indicated by the lower crosslink density of the revulcanizates compared with the vulcanizate (Figure 7.11). With the increase of the ultrasonic amplitude in the devulcanization, the final torque is lower.
171 14 35phr virgin 12 35phr 5 µm 35phr 7.5 µm 10 35phr 10 µm
8
6 Torque, dNm 4
2
0 0 5 10 15 20 25 Time, min
Figure 7.14 Revulcanization curves at 160oC for the devulcanized 35 phr CB filled IR
obtained at three ultrasonic amplitudes
7.4.5 Mechanical Properties
The stress-strain curves of the virgin cured and the revulcanized 35 phr CB filled
IRs are plotted in Figure 7.15. It is observed that the tensile strength of the revulcanized rubber is significantly lower than that of the virgin cured rubber. The reason is probably due to the scission of the main chains as well as the breakdown of the network structures during the devulcanization. However, the reduction of tensile properties in the filled rubber is more severe than that in the case of the unfilled rubber (Section 6.8). In the filled rubber, some of the rubber molecules are incorporated into the CB particles to form the bound rubber49. It is reported that the percentage of the bound rubber could be as high as 25% of the total rubber matrix at the CB loading of 35 phr64. Accordingly, the mobility
172 of bound rubber is restricted by the surface of the CB. As a result, upon the ultrasound exposure, these immobilized bound rubber chains would experience greater probability of severance compared to the relatively more flexible chains in the unfilled IR. Once the flaw is created locally in the filled rubber after the ultrasound treatment, the inferior mechanical properties are inevitable. This explanation will be supported by the simulation of the network structures – gel fraction and crosslink density in Chapter XI
Section 9.4 where it will be shown that the relative degree of breakdown of the main chain over the crosslinks in the filled rubber is higher than that of the unfilled IR.
500mm/min Amplitude, µm 30 5 7.5 10 0 20 , MPa σ 10
0 0 100 200 300 400 500 600 ε , %
Figure 7.15 Stress-strain curves of the virgin cured and the revulcanized 35 phr CB
filled IR obtained at three ultrasonic amplitudes
173 7.5 Devulcanization of the Filled IR Containing the Processing Oil
Processing oils are generally considered as plasticizers. They are used in the carbon black filled rubbers as a processing aid. At low levels of loading their function is to aid in the dispersion of fillers. At high amounts they reduce the uncured compound viscosity and the vulcanizate stiffness and hardness233. Petroleum oils are one of the major sources of plasticizers. They are broadly divided into three categories: aromatic, naphthenic and paraffinic. The naphthenic oil has good processability and compatibility in most rubbers. The processing oil used in this study was the Plasticizer LN from the
Akrochem Corporation, Akron, OH. It is a hydro-treated naphthenic oil of low viscosity234. In this section the processing oil was added in the 35 phr carbon black filled
IR compounds with the purpose to examine its effect on the devulcanization process, cure and revulcanization kinetics, rheological and mechanical properties. Generally the proportion of the processing oil used in passenger tire sidewall rubber formulation is 10 phr3. Therefore the amount of oil used in this study was fixed at 10 phr.
Vulcanization
Before studying the effect of oil on devulcanization, the 35 phr black filled IR containing the 10 phr processing oil was vulcanized first. Figure 7.16 shows the comparison of the cure curves among the unfilled, the 35 phr CB filled rubbers without and with the inclusion of oil. The addition of carbon black substantially increases the curing torque which is due to the creation of flow resistance by the filler. Due to the plasticization effect of the processing oil, the curing torque was somewhat lower than the filled rubber without oil. In Figure 7.16, the induction period is much shorter in the filled
IR than in the unfilled IR. This was explained by the acceleration effect of carbon black 174 in the vulcanization discussed earlier in Chapter IV Section 4.6. For the filled rubbers containing the 10 phr processing oil in the compounding recipe, it is observed a slight delay of induction period. This is probably due to the fact that the surface of carbon black is able to absorb some oil molecules and thus this action more or less suppresses the acceleration of carbon black on the vulcanization. However, this effect is not substantial.
In addition, it seems that adding the processing oil in the filled rubber does not affect the degree of reversion occurred in the post curing stage.
CB/oil, phr 12 0/0 35/0 10 35/10
8
6 Torque, dNm 4
2
0 0 5 10 15 20 25
Time, min
Figure 7.16 Vulcanization of the unfilled and the 35 phr CB filled IR containing 0 and
10 phr processing oil at 160oC, a strain amplitude of 4.2% and a frequency of 100 cpm
Figure 7.17 shows the stress-strain curves of the unfilled and 35 phr black filled
IR vulcanizates containing 0 and 10 phr processing oil. It is observed that after adding the
175 processing oil, the rubber vulcanizate is still able to undergo crystallization upon stretching. However it seems that the stress induced crystallization is delayed by the oil.
The obtained tensile strength, elongation at break and moduli at 100 and 300% elongation are plotted in Figure 7.18. The tensile strength and the moduli of the rubber vulcanizate are improved by incorporating the rigid filler carbon black into the compounds. The addition of processing oil somewhat reverses the strengthening effect of carbon black.
Probably the plasticizing effect of oil contributes to the lower strength and moduli. The elongation at break is shortened after the rubber is filled with carbon black. The addition of oil does not affect the elongation at break.
500mm/min CB/oil, phr 30 0/0 35/0 35/10
20 , MPa σ
10
0 0 200 400 600 800 1000 ε , %
Figure 7.17 Stress-strain curves of the unfilled and the 35 phr CB filled IR
vulcanizates containing 0 and 10 phr processing oil 176 35 1200 σ (a) B 30 ε B 1000 25 800 20 , % B ε , MPa 600 B
σ 15 400 10
5 200
0 0 IR IR/35CB IR/35CB/10oil
Vulcanizates 4 (b) E100
E300 3 , MPa
300 2 , E 100 E 1
0 IR IR/35CB IR/35CB/10oil
Vulcanizates
Figure 7.18 Tensile strength σB, elongation at break εB (a) and moduli at 100 and 300% strain E100, E300 (b) of the unfilled and the 35 phr CB filled IR vulcanizates containing 0 and 10 phr processing oil
177 7.5.1 Power Consumption and Die Pressure
The power consumption and the die pressure as a function of ultrasonic amplitude for the devulcanization of the 35 phr CB filled IR containing 0 and 10 phr processing oil are shown in Figure 7.19 and Figure 7.20, respectively. It is observed that the dependence of power consumption and die pressure on the ultrasonic amplitude shows the similar trend after the inclusion of the processing oil in the filled IR compounds. The power consumption increases and the die pressure decreases with amplitude. Generally, the power consumption is higher in the rubber containing the processing oil. This suggests that the addition of oil contributes to more dissipation of ultrasonic energy. As seen from
Figure 7.20, the die pressure is lower in the rubber containing the 10 phr processing oil.
This is partially due to the lubrication effect of oil on the CB causing the material flow much easier.
1600 CB/oil, phr 1400 35/0 35/10 1200
1000
800
600 Power consumption, watts 400
200 4567891011 µ Amplitude, m Figure 7.19 Power consumption as a function of ultrasonic amplitude for the 35 CB filled IR containing 0 and 10 phr processing oil
178 30 CB/oil, phr
25 35/0 35/10
20
15
10 Die pressure,MPa
5
0 024681012 µ Amplitude, m
Figure 7.20 Die pressure as a function of ultrasonic amplitude for the 35 CB filled IR
containing 0 and 10 phr processing oil
7.5.2 Gel Fraction and Crosslink Density
After devulcanization of the 35 phr filled rubber containing the processing oil, gel fraction and crosslink density were determined by the Soxhlet extraction method. The data as a function of amplitude are plotted in Figure 7.21. For the convenience of comparison, the results for the revulcanized samples as well as the rubbers without the oil are also plotted in the same graph. For the devulcanized samples, the gel fraction and crosslink density of the rubbers containing the oil are lower than those of the rubbers without oil regardless of the ultrasonic amplitude. This is also the case for the revulcanized samples with one exception that the crosslink density of the 10 µm samples containing oil is somewhat higher than that of the 10 µm sample without oil.
179 1.00
0.95
0.90
0.85 Gel fraction Gel 0.80 CB/oil, phr 35/10 (devulc) 35/10 (revulc) 0.75 35/0 (devulc) 35/0 (revulc) 0.70 024681012 µ Amplitude, m 300 CB/oil, phr 35/10 (devulc) 250 3 35/10 (revulc) 35/0 (devulc) 200 35/0 (revulc)
150
100
Crosslink density, mol/m density, Crosslink 50
0 024681012 µ Amplitude, m
Figure 7.21 Gel fraction and crosslink density of devulcanized (open symbols) and revulcanized (solid symbols) 35 phr CB filled IR containing 0 and 10 phr processing oil as a function of ultrasonic amplitude obtained at flow rate of 0.63 g/s, a die gap of 2.54 mm and a barrel temperature of 120oC
180 The solubility of the processing oil in benzene is experimentally examined as follows. About 1 ml oil is mixed with 10 ml benzene. The mixture is manually stirred for
1 min and then it is sealed and stored for a few days. Since the oil has the yellowish color and benzene is colorless, the observation of uniform yellowish mixture (no separated layers) after several days confirmed that the oil is soluble in benzene. More sol fraction in the rubber containing the oil than in the one without oil is due to the reason that a portion of oil is extracted out of the rubber sample by benzene. Generally, the gel fraction and crosslink density of the revulcanized samples are higher than those of the devulcanized samples regardless of the presence of oil. Similar to the occurrence in the unfilled devulcanized IR (Section 6.5), the filled devulcanized rubber is also able to undergo revulcanization provided that the additional amount of sulfur is incorporated into the devulcanized samples. The presence of processing oil does not alter the ability of devulcanized rubber to undergo revulcanization.
7.5.3 Rheological Properties
The rheolgoical properties of the 35 phr black filled rubber containing the 10 phr processing oil were investigated on the APA 2000 by taking the frequency sweep at
120oC and a strain amplitude of 4.2%. Complex viscosity and loss tangent as a function of frequency are plotted in Figure 7.22. For the convenience of comparison, the dynamic properties of the filled IR without the processing oil are plotted in the same figures. In
Figure 7.22 (a), after adding the 10 phr oil in the compounding recipe, it is observed a decrease of complex viscosity for the uncured, the cured and the devulcanized rubber at
5µm ultrasonic amplitude at the whole available frequency range. Apparently, the addition of oil makes the flow resistance less in these samples. However, the reduction of 181 108 oil (phr) 0 10 virgin IR/35CB 107 IR/35CB cured 5 µm 7.5 µm 10 µm 106 *|, Pa-s
η (a) | 105
104
103 10-2 10-1 100 101 102 103 ω , rad/s 101 oil (phr) 0 10 virgin IR/35CB IR/35CB cured 5 µm 7.5 µm 10 µm 100 δ
tan (b)
10-1
10-2 10-1 100 101 102 103 ω , rad/s
Figure 7.22 Complex viscosity (a) and loss tangent (b) of the virgin 35 phr CB filled
(containing 0 and 10 phr processing oil) uncured, the cured and the devulcanized IR as a
function of frequency at 120oC, a strain amplitude of 4.2%
182 viscosity in the presence of oil is not evident in the filled rubbers devulcanized at 7.5 and
10 µm. Nevertheless, the inclusion of oil does not change the degree of shear thinning for all the rubber samples. In Figure 7.22 (b), the dependence of the loss tangent on the presence of oil for the uncured rubber is weak. In contrast, the loss tangent for the cured and devulcanized rubbers becomes lower in the presence of oil. Storage and loss moduli as a function of frequency are shown in Figure 7.23. In Figure 7.23 (a), the storage modulus of the uncured, cured and devulcanized rubbers at 5 µm generally decreases with the addition of oil. The reduction of the storage modulus in the presence of oil for the rubbers devulcanized at 7.5 and 10 µm is not evident. From Figure 7.23 (b), it is seen that a reduction of loss modulus occurred with the inclusion of oil for the uncured, cured and devulcanized samples.
7.5.4 Revulcanization
The revulcanization experiment of the devulcanized rubbers containing the 10 phr oil was carried out under the same conditions as it was done in the rubbers without the oil
(Section 7.4.4). Figure 7.24 compares the revulcanization curves of the devulcanized rubbers containing 0 and 10 phr oil. First, it is found that the induction period is absent in the revulcanization regardless of the presence of oil. For the rubbers devulcanized at 5 and 7.5 µm, the torque upon revulcanization is lower when the 10 phr oil was included in the compounds. For the rubber devulcanized at 10 µm, the torque in the revulcanization is instead higher in the compound with oil than in the one without oil. This suggests that the dependence of revulcanization torque on the presence of oil highly depends on the ultrasounic amplitude in the devulcanization. Less network destruction occurred at lower ultrasonic amplitudes of 5 and 7.5 µm than at higher amplitude of 10 µm. Therefore, the 183 oil (phr) 0 10 (a) virgin IR/35CB IR/35CB cured 101 5 µm 7.5 µm 10 µm
100 G', Pa
10-1
10-2 10-1 100 101 102 103 ω , rad/s 100 oil (phr) 0 10 (b) virgin IR/35CB IR/35CB cured 5 µm 7.5 µm 10 µm
10-1 G", Pa G",
10-2 10-2 10-1 100 101 102 103 ω , rad/s
Figure 7.23 Storage (a) and loss (b) modulus of the virgin 35 phr CB filled (containing
0 and 10 phr processing oil) uncured, the cured and the devulcanized IR as a function of
frequency at 120oC, a strain amplitude of 4.2%
184 10
8 0-0 10-0 0-5 6 0-7.5 10-5 10-7.5 10-10
Torque, dNm 4 0-10
2 oil(phr)-amplitude(µm)
0 5 10 15 20 Time, min
Figure 7.24 Revulcanization curves of the devulcanized 35 phr CB filled IR
(containing 0 and 10 phr processing oil) at 160oC, a strain amplitude of 4.2% and a
frequency of 100 cpm rubber devulcanized at lower amplitudes has more heterogeneity as indicated by apprearance. This makes the dispersion and migration of revulcanization ingredients
(especially the sulfur) into the remaining unsaturated rubber main chain more difficult.
On the other hand, the majority of the CB may still be trapped in the crosslinking network due to less devulcanization. Therefore the dispersion of the reculcanization ingredients is difficult. In this case, the processing oil present in the devulcanized rubber may absorb the additional amount of revulcanization ingredients resulting in less effective revulcanization as indicated by a drop of revulcanization torque. In contrast, at the higher ultrasonic amplitude such as 10 µm, more devulcanization is occurred. This makes the material less heterogeneous and more CB may escape out of the broken 185 network. The presence of oil would help the dispersion of isolated CB. The less heterogeneity of the material would make the revulcanization ingredients easier to mix with the devulcanized rubber. Thus the revulcanization is more effective leading to increased revulcanization torque. As a matter of fact, experimentally, the compounding of the 5 and 7.5 µm samples in the two-roll mill is much harder than the 10 µm sample. The latter compounded sample is easily integrated into stable sheets after compounding; while the former two samples are only integrated into fragile sheets. In addition, the change of maximum revulcanization torque with the ultrasonic amplitude before and after the addition of processing oil in the rubber is consistent with the change of crosslink density of the revulcanizates with the amplitude (Figure 7.21).
7.5.5 Mechanical Properties
The tensile strength, the elongation at break and the modulus at 100% strain of the revulcanized 35 phr filled IR containing the 0 and 10 phr processing oil are shown in
Figure 7.25. For the tensile strength and the elongation at break, the revulcanizates of 5 and 7.5 µm without oil show higher than those with oil. In contrast, the revulcanizate of
10 µm containing oil exhibits better tensile properties. The variation of the tensile properties of the revulcanizates containing oil on the ultrasonic amplitude follows the similar trend as in the revulcanization torque (Figure 7.24). The reason for such behavior of tensile properties is similar to the explanation given in discussing the behavior of revulcanization torque. The modulus of the reculcanizates in the presence of oil is decreased with the increase of amplitude. The modulus behavior is consistent with that of the crosslink density (Figure 7.21).
186 10 500 (a)
8 400
6 300 , % B , MPa ε B
σ 4 oil, phr 200 σ 0 ( B) σ 10 ( B) 2 ε 100 0 ( B) ε 10 ( B) 0 0 0.0 2.5 5.0 7.5 10.0 µ Amplitude, m 2.5 (b)
2.0
1.5 , MPa 100
E 1.0
0.5 oil, phr 0 10 0.0 0.0 2.5 5.0 7.5 10.0 µ Amplitude, m Figure 7.25 Tensile strength σB, elongation at break εB (a) and modulus (b) at 100% strain E100 (c) as a function of ultrasonic amplitude for the revulcanizates of 35 phr CB
filled IR containing 0 and 10 phr processing oil
187 7.6 Effect of Retarder on Vulcanization and Revulcanization of Filled IR
As seen in the revulcanization kinetics of the devulcanized filled IR (Figure 7.24), the induction period is absent regardless of the presence of processing oil. It indicates that the revulcanization starts immediately without any delay. This is not desirable in the practical production of rubber articles since the materials have to be allowed a certain amount of time (scorch safety) to be shaped before the occurrence of vulcanization. In this section an attempt is made to combine the retarder into the compounding recipe of the devulcanized rubber with the purpose to improve the scorch safety of the revulcanized rubbers. The retarder used in this study is the retarder SAFE from the
Akrochem Corporation, Akron, OH. The composition of the SAFE is generally a treated aromatic sulfonamide. It is widely used in the natural and synthetic rubber compounds.
The common amount of SAFE added in the rubber is from 0.1 to 2.0 phr with the 1 phr as the preferred amount in natural rubber235. Therefore, in this investigation, 1 phr of SAFE is added in the compounding recipe of the devulcanizates. In addition, the effect of retarder on the mechanical properties is investigated.
The vulcanization and revulcanization curves of the 35phr CB filled rubber containing the 10 phr oil with and without 1phr retarder SAFE is shown in Figure 7.26. It is observed that the inclusion of retarder in the vulcanization recipe improves the scorch safety and reversion resistance. The final torque is also increased due to the presence of retarder. For the devulcanized samples, the presence of retarder in the revulcanization recipe causes the delay of the revulcanization reaction. Particularly, the scorch delay is more substantial at higher amplitude during the ultrasonic devulcanization. There is less reversion in the revuclanizated rubber containing the retarder. At any ultrasonic 188 amplitude, the final torque is higher in the samples containing the retarder. These observations indicate that the selected retarder is very effective in improving the scorch safety for the devulcanized samples and in improving the reversion resistance for the virgin and devulcanized rubbers. In addition, regardless of the presence of retarder, the final revulcanization torque is decreased with the increase of ultrasonic amplitude.
10 0-0 1-0
8 1-5 1-7.5 6 0-5 1-10 0-10 4 Torque, dNm 0-7.5
2
SAFE(phr)-amplitude (µm) 0 0 5 10 15 20 25 Time, min Figure 7.26 Cure curves of the virgin and devulcanized 35 phr CB filled IR containing
10 phr oil with and without the retarder SAFE at 160oC, a strain amplitude of 4.2% and a
frequency of 100 cpm
The tensile strength, the elongation at break and the modulus of the 35 phr CB filled IR containing 10 phr oil with and without 1 phr SAFE as a function of ultrasonic amplitude of devulcanization are shown in Figure 7.27. Generally, the properties of vulcanizates are better than those of the revulcanizates due to the the crosslink rupture accompanied by the main chain scission during ultrasonic devulcanization. For the
189 25 700 (a) SAFE, phr 0 (σ ) B 600 20 σ 1 ( B) ε 0 ( B) 500 15 ε 1 ( B)
400 , % B , MPa ε B
σ 10 300
5 200
0 100 0.0 2.5 5.0 7.5 10.0 µ Amplitude, m 2.8 (b) SAFE, phr 2.6 0 1 2.4
2.2
, MPa 2.0 100
E 1.8
1.6
1.4
1.2 0.0 2.5 5.0 7.5 10.0 µ Amplitude, m
Figure 7.27 Tensile strength σB, elongation at break εB (a) and modulus at 100% strain
E100 (b) of the revulcanized 35 phr CB filled IR containing 10 phr processing oil without
and with the retarder SAFE as a function of ultrasonic amplitude 190 revulcanizates, the tensile strength and the elongation at break increase with the increase of ultrasonic amplitude regardless of the presence of retarder (Figure 7.27 a). The modulus at 100% strain decreases with the amplitude (Figure 7.27 b). Concerning for the effect of retarder on the tensile properties at the same ultrasonic amplitudes, the observation is somewhat complicated. At a low ultrasonic amplitude of 5 µm, the tensile properties are slightly improved after adding the retarder SAFE. However, at the higher amplitudes of 7.5 and 10 µm, tensile properties are deterioated in the case of adding the retarder (Figure 7.27). The exact reason for this complicated behavior is unknown.
However, it seems that the tensile properties in the presence of retarder SAFE at various amplitudes is in accordance with the relative magnitude of maximum revulcanization torque in samples with and without the retarder (Figure 7.26). The revulcanization time of the samples used for the tensile test measurement is chosen as the time when the maximum torque is reached in revulcanization curve.
7.7 Blending of the Devulcanized IR (dIR) with the Virgin IR
Since the revulcanized filled IR shows inferior tensile properties than the virgin filled IR vulcanizate, an attempt is made to blend the devulcanized IR with the virgin IR.
The purpose of this study is to obtain the properties comparable to the virgin rubber. This approach was shown to be successful in PUR236, NR237 and EPDM238. The devulcanized filled IR (dIR) was processed in the two-roll mill for 3-5 passes and then was blended with the virgin filled IR at varied ratios of 25/75, 50/50, 75/25. The front roll speed was
14 mpm (meters per minute) and the back roll speed was 11 mpm and the roll diameter and length are 15 cm and 30.5 cm, respectively. The blending in the two-roll mill took about 40 passes. The recipe shown in Table 3.2 is used for the vulcanization of the blends. 191 In addition to the ingredients listed in Table 3.2, 1 phr retarder SAFE is also included in the compounding recipe with the purpose of improving the scorch safety and minimizing the reversion. The curatives are all added with respect to the total rubber content in the blends. The dIR used for blending is the 35 phr black filled rubber containing the 10 phr processing oil devulcanized at 120oC and at an amplitude of 10 µm with a gap size of
2.54 mm. The cure kinetics and the mechanical properties are studied. Another objective of this study is to examine if the ultrasonic devulcanization is an effective way of recycling the waste rubbers. Therefore, for a comparative purpose, the ground cured filled IR (gIR) is also blended with virgin filled IR containing the 35 phr carbon black and the 10 phr processing oil at various ratios of 25/75, 50/50 and 75/25. The vulcanization of gIR/IR blends and their mechanical properties were studied using the same procedure as for the blends of dIR/IR.
The cure curves of dIR/IR blends with various ratios at 160oC are given in Figure
7.28. The curatives are added with respect to the total rubber content in blends. It is observed that when the amount of devulcanized rubber is reduced in the blends, the induction period becomes longer. It is recognized that the devulcanized rubber contains the accelerator residue218, 230. When the content of dIR in the blends is reduced, accordingly the amount of accelerator residue in the blends becomes lower. As a result, it takes longer time for the blend rubbers to start the crosslinking reaction. On the other hand, as shown in Figure 7.28, the final torque is increased when the dIR/IR blends contains less dIR. Particularly, in the dIR/IR blends of 75/25 ratio, the final torque is increased dramatically. This is due to the fact that dIR is a rubber obtained from the combined rupture of main chain and crosslink network of virgin IR vulcanizates. 192 10 5 4 3 8 2
6 1
Torque, dNm 4 dIR / virgin IR 1- 100 / 0 2- 75 / 25 2 3- 50 / 50 4- 25 / 75 5- 0 / 100 0 0 5 10 15 20 25
Time, min
Figure 7.28 Cure curves for blends of devulcanized (dIR) and virgin IR containing 35 phr CB and 10 phr oil using the recipe in Table 3.2 at 160oC. (Curatives were added with
respect to the total rubber content. IR was devulcanized at an amplitude of 10 µm)
In order to identify whether the ultrasound devuclanization technique has an advantage over recycling the rubber vulcanizates by grinding, the ground cured rubber particles (gIR) were blended with the virgin IR containing the 35 phr CB and the 10 phr processing oil. The vulcanization of gIR/IR blends was carried out in the way of similar to the dIR/IR blends. The cure curves of gIR/IR blends at 160oC are given in Figure 7.29.
In Figure 7.29, it is observed that with the increased amount of gIR in the blends, the induction period is shorter. This observation is in an agreement with the earlier study239 by using HPLC that the shortened induction time was due to the diffusion of accelerater 193 residue left in the ground vulcanizate into the virgin rubber matrix. It was also inferred from the fact that the ground vulcanizate particles originally cured with peroxide did not alter the induction time.
10 1 2 3 8 4
6
Torque, dNm 4 gIR / virgin IR 1- 75 / 25 2- 50 / 50 2 3- 25 / 75 4- 0 / 100
0 0 5 10 15 20 25
Time, min
Figure 7.29 Cure curves for blends of ground IR vulcanizates (gIR) and virgin IR
containing 35 phr CB and 10 phr oil using the recipe in Table 3.2 at 160oC. (Curatives
were added with respect to the total rubber content)
In Figure 7.29, it is also observed that with the increased amount of gIR, the final torque is higher, a phenomenon different from dIR/IR blends (Figure 7.28). In this case, since the gIR is fully vulcanized, the gIR particles act like the rigid fillers in the blends. It is expected that the flow resistance would be greater when the amount of gIR is increased in the gIR/IR blends. Therefore, it is not surprising to observe the increased torque in the 194 blends containing more gIR. In addition to the final torque, the minimum torque of the gIR/IR blends is also dramatically increased with the increased proportion of gIR in the blends. This observation is due to the same reason as explained in the final torque variation. In contrast, the minimum torque of the dIR/IR blends shows no difference. The almost same minimum torque observed in the dIR/IR blends indicates that the network destruction of IR vulcanizates occurs upon ultrasonic devulcanization. This effect improves the processability and the mixing of the two different types of rubbers: devulcanized IR and virgin uncured IR.
The tensile strength σB and the elongation at break εB of the two types of blends are shown in Figure 7.30 (a) and (b), respectively. In Figure 7.30 (a), it is observed that as the proportion of the virgin IR is increased in the blends, the tensile strength of the dIR/IR blends progressively increases. This increase almost exactly follows the rule of mixing. The tensile strength of gIR/IR blends is also increased with the addition of the virgin IR, however, at a much slower pace. Generally, the dIR/IR blends shows higher tensile strength (Figure 7.30 (a)) and elongation of break (Figure 7.30 (b)) than the gIR/IR blends at same blending ratios. Since the unsaturated carbon-carbon double bonds left in the ground rubber are already attached to the network before the blending, it is difficult for them to diffuse out of the network and to take part in the vulcanization in blends. The direct result would be little formation of interfacial sulfur crosslinks between the particles of gIR and virgin IR matrix. Consequently, the poor adhesion between gIR and the matrix of virgin IR leads to the inferior properties of the gIR/IR blends. In
195 20 (a)
15
10 , MPa B σ
5 dIR/IR gIR/IR 0 0 255075100
Concentration of virgin IR in blends, wt% 500 3.0 (b) 2.5 400
2.0 300
, %
1.5 , MPa B ε 100
200 E dIR/IR (ε ) B 1.0 ε gIR/IR ( B) 100 dIR/IR (E ) 100 0.5 gIR/IR (E100) 0 0.0 0 255075100
Concentration of virgin IR in blends, wt%
Figure 7.30 Tensile strength σB (a), elongation at break εB and modulus at 100% strain
E100 (b) of recycled and virgin IR blends (gIR/IR and dIR/IR) as a function of virgin IR
concentration 196 contrast, for the dIR/IR blends, due to the network rupture occurred in dIR during devulcanization, the carbon-carbon unsaturated bonds are able to diffuse into the virgin rubber matrix and form interfacial sulfur crosslinks with the rubber matrix contributing to better tensile properties. Therefore, this comparative study suggests that the mechanical properties of recycled gIR/IR blends can be improved by the ultrasonic devulcanization of gIR before the blending was carried out. The modulus at 100% elongation of the blends was given in Figure 7.30 (b). This result shows that, independent of the blending ratios, the difference of the modulus at 100% strain between these two types of blends is insignificant.
7.8 Conclusions
The continuous ultrasonic devulcanization of carbon black filled IR was carried out in the co-axial reactor using the similar procedures as used in the unfilled IR. Similar to the occurrence in the unfilled IR, the power consumption during the devulcanization of the filled IR continuously increased with the increase of ultrasonic amplitude. This is in a contrast with the observation in NR where the maximum power consumption was obtained at the intermediate amplitude of 7.5 µm. This suggested that extent of devulcanization increased with ultrasonic amplitude in IR. The extent of devulcanization decreases at 10 µm in NR. Furthermore, the trend of power consumption agreed with the change of other properties such as gel fraction, crosslink density and the mechanical properties for both rubbers. The different behaviors in two rubbers were probably resulted from the slight difference in the amount of stereoregular structures.
The effects of processing oil on the vulcanization, devulcanization and revulcanization process were examined. The processing oil slightly delayed the 197 vulcanization and lowered curing torque. However, it did not change the degree of reversion. In addition, the power consumption was increased, while gel fraction and crosslink density were decreased with the ultrasonic amplitude in the oil containing system during devulcanization. These results sugested that the addition of the oil led to more devulcanization. As a result of this, the dynamic viscosity and storage modulus of the uncured, cured and devulcanized filled IR (except 7.5 µm) was decreased. The revulcanization of the devulcanized filled IR did not have the induction period regardless of the presence of carbon black and processing oil.
The retarder introduced in the revulcanization recipe was effective to improve the scorch safety and minimize the reversion in revulcanization. However, it was ineffective to improve the tensile properties of the revulcanizates.
The ultrasonically devuclanized IR (dIR) was blended with the virgin IR and the properties were compared with the blends of fully cured ground IR (gIR) and virgin IR.
The increased proportion of the virgin IR resulted in the extension of induction period for both blends. The dIR/IR blends showed much better tensile strength and elongation at break than the gIR/IR blends indicating the advantage of ultrasonic devulcanization in recycling the waste rubbers. The dependence of the tensile strength of the dIR/IR blends on the proportion of virgin IR followed the rule of mixing.
198
CHAPTER VIII
8.MOLECULAR MOBILITY OF ULTRASONICALLY TREATED GUM IR,
UNFILLED AND CB FILLED IR
8.1 General
The recycling of the vulcanized elastomers, particularly the tire rubbers, has been of increasing concern to the industry. There are many attempts to reclaim or devulcanize the rubber by breaking the chemical bonds in the network. Among them, ultrasonic technique is the emphasis of our research. It is a continuous process without the involvement of any chemicals. Isoprene rubber is the artificial equivalent of natural rubber. It is heavily used in the tire components with the combination or instead of NR.
IR used in tire usually contains over 98% of cis-1, 4 structures in order to achieve the important stress-induced crystallization.
Solid-state NMR technique has been applied to study the molecular dynamics and diffusion of various types of elastomers including SBR122, 240 silicone rubber241, 242, PU243, butyl rubber244 and BR245. Proton NMR transverse magnetic relaxation reflects mainly
1 the intermolecular chain interaction. In order to analyze the H T2 relaxation spectrum, generally a two- or three- component model is employed to interpret the transverse decay data. For example, in the study of SBR molecular mobility, it is suggested that the devulcanized SBR consisted of two major components, namely, the extractable sol for 199 122 long T2 relaxation and the unextractable gel for short T2 relaxation . In contrast, in the study of BR proton T2 relaxation decay, it was shown that three-component model was more appropriate description245. The three components in devulcanized BR include the physically entangled and chemically crosslinked networks responsible for short T2 relaxation, the unentangled light sol and the dangling network chain ends for intermediate
T2 relaxation and the unreacted oligomers for the long T2 relaxation.
In this research high power ultrasonic technique was applied to treat the virgin gum IR. This led to the treated gums degraded at different ultrasonic amplitudes. These treated gums are vulcanizable. The ultrasound was also employed to devulcanize the unfilled and the CB filled cured IR. In an attempt to understand the effect of ultrasound on the degradation of gum IR and the devulcanization of cured rubber, the solid-state
NMR proton relaxation decay and molecular diffusion measurement were employed to examine the effects of intense ultrasound on the molecular mobility of the treated gum rubbers, vulcanized and devulcanized rubbers. Particularly the production of light sol, the destruction of network, network segmental mobility, sol mobility and diffusion were the focus of investigation.
8.2 Preparation of the Samples
Virgin gum IR or ground unfilled or 35phr CB filled IR vulcanizates with and without the processing oil was fed into the co-axial ultrasonic reactor for the treatment of the gum and the devulcanization of the cured rubber. The extrusion was carried out at a barrel temperature of 120oC and a gap of 2.54 mm. The screw speed was 17 rpm and the corresponding flow rate was 0.63 g/s for the virgin gum IR and the filled IR. Varied flow
200 rates ranging from 0.63 to 2.10 g/s were used in the devulcanization of the ground unfilled IR vulcanizates. The ultrasonic amplitude was varied from 5 to 10 µm.
Ten samples in total were used in the molecular mobility analysis of the ultrasonically treated gum IR and their vulcanizates. Firstly the virgin gum IR subjected to the ultrasonic treatment at four levels of amplitude: 0 (passing the extruder without the ultrasound), 5, 7.5 and 10 µm and this generated five gum samples including the virgin untreated gum. Thereafter the five gum samples were vulcanized according to the recipe in Table 3.2 and this procedure produced another five additional samples.
For the unfilled IR, totally eleven samples were investigated for the NMR proton relaxation. In addition to the virgin uncured gum IR and its vulcanizate, the nine devulcanizates obtained with the combination of three flow rates – 0.63, 1.07 and 2.10 g/s and three ultrasonic amplitudes – 5, 7.5 and 10 µm were included in the analysis.
For the filled IR, eleven samples were used in the NMR relaxation measurements.
The details for the sample composition and preparation are summarized in Table 8.1.
8.3 NMR Experiments
All the NMR measurements (T2 relaxation decay and molecular diffusion) were conducted non-spectroscopically at 33 MHz and at 70.5oC to accelerate the molecular/segmental motions and to accentuate the differences between the gel and the sol. The radio-frequency (rf) pulse sequence employed for T2 measurements was the standard Hahn spin echo sequence with only a steady magnetic field gradient shown in
Figure 2.8. The echo height A(2τ) was measured on-resonance using rf phase-sensitive detection, and signal averaging over six passes. Pulse spacing (τ) was adjusted between
0.1 ms and 60 ms in 30 or more steps, until the echo height was less than 0.5% of the 201 Table 8.1 The samples used in the NMR relaxation analysis of the filled IR
Composition, parts
Sample IR CB Oil Preparation
IR/CB 100 35 — Premixed in Banbury for 10 minutes
IR/CB-vulc 100 35 — Vulcanized with recipe in Table 3.2
IR/CB-5 µm 100 35 — IR/CB-vulc devulcanized at 5 µm
IR/CB-7.5 µm 100 35 — IR/CB-vulc devulcanized at 7.5 µm
IR/CB-10 µm 100 35 — IR/CB-vulc devulcanized at 10 µm
Oil — — 100 —
Virgin IR/CB/oil 100 35 10 Premixed in Banbury for 10 minutes
IR/CB/oil-vulc 100 35 10 Vulcanized with recipe in Table 3.2
IR/CB/oil-5 µm 100 35 10 IR/CB/oil-vulc devulcanized at 5 µm
IR/CB/oil-7.5 100 35 10 IR/CB/oil-vulc devulcanized at 7.5 µm
µm
IR/CB/oil-10 µm 100 35 10 IR/CB/oil-vulc devulcanized at 10 µm
initial height. For diffusion measurements the Hahn echo was mostly used, occasionally supplemented by the stimulated echo, both coordinated with a matched pair of pulsed magnetic field gradients of magnitude G and duration δ, separated by time ∆. Rf phase- sensitive detection of the echo signal recorded 3 kHz off-resonance was followed by windowed magnitude-Fourier transformation, integrating the peak area, and performing baseline correction. As a consequence of the gradient pulses the spin echo in a diffusing
202 substance was attenuated; the amplitude was recorded as a function of δ as part of the gradient parameter X (shown in Equation (2.8)). The implementation of this method has been described elsewhere191, 192.
In this investigation, the parameter settings used were fixed values of ∆ = τ = 15 ms; G = 634 Gauss/cm (for the ultrasonically treated IR) and 287 Gauss/cm (for the processing oil); and a small steady gradient G0 = 0.35 Gauss/cm, with δ varied in eight to twenty steps until the echo signal was attenuated to the background noise level, or until
17 ms was reached. Echo height measurements were signal-averaged over six to twelve passes. These procedures ensured that the tail of the echo decay, from which light sol fraction and sol mobility are inferred, is adequately characterized. It further ensured that the short T2 component, arising from non-diffusing chemical network or physically entangled network, would not have interfere with the direct measurement of sol or oligomer diffusion. Optimization of magnetic field homogeneity combined with the expected low sol diffusion coefficient confirmed that the T2 values measured with the two-pulse sequence were not falsified by diffusion artifacts246.
8.3.1 Proton T2 Data Analysis
1 The H T2 data, the echo height A(2τ) as a function of the rf pulse spacing, τ, were initially analyzed off-line with a Fortran-based computer code, T12FIT247, in terms of a two-component relaxation decay model, which has served well in previous investigations241, 242, 245. It will be continuously applied in the analysis of the unfilled and the 35 phr CB filled IR. However, for the ultrasonically treated gum IR, after more detailed analysis via spectral inversion241 it was found that this model was deficient in
203 capturing the decay shape coming from the additional components possibly generated by ultrasound treatment. Hence it was adjusted by a three-component model of the form
τ τ τ τ A(2 ) = − 2 E + − 2 + − − − 2 f S exp( ) f L exp( ) (1 f S f L )exp( ) (8.1) A(0) T2S T2L T2M
This model has six adjustable parameters, the three T2 components: long, medium and short, the two relative fractions of these, and the Weibull exponent E (1 8.3.2 PGSE Diffusion Data Analysis When the diffusion rates display an arbitrary distribution, they might be modeled as several components with fractional amplitude ai and diffusion coefficients Di. The diffusion echo attenuation assumes the form AX(2τ , ) =−∑expaDX (γ 2 ) (8.2) τ ii A(2 ,0) i where γ represents the gyromagnetic ratio of the nucleus at resonance (1H). In this study, a two-diffusion species model is considered due to the bimodal nature of the diffusion curve and the above equation can be simplified as AX(2τ , ) =−f exp(γγ22DX ) +−− (1 f )exp( DX ) (8.3) A(2τ ,0) fast fast fast slow where Dfast and Dslow represent the diffusion coefficients of the two species and ffast denotes the echo fraction of the fast-diffusing species. This model has three adjustable parameters, the two diffusion coefficients Dfast and Dslow and the relative echo amplitude ffast. The extent of the production of fast-diffusing oligomeric material by ultrasound is an important consideration in the recycling of various rubbers. By combining the information from the T2 measurements with those from diffusion, it is possible to extract 204 a close estimate of the proton fraction of the specimen Ffast consisting of oligomeric materials. Given that the diffusion pulse sequence timing eliminates the contribution of the network to the measured spin echo but retains a portion of the mid-component intensity, Ffast is obtained by multiplying ffast, the fraction of the fast-diffusing portion of the echo at 2τ, by the fraction of the echo from that species contributing a signal at 2τ: =×' τ Ffffast fast long (2 ) (8.4) 11 ' ττ=+−− − where fffflong(2 ) long (1 short long )exp 2 (8.5) TT22long med 8.4 Molecular Mobility of Treated Gum IR and their Vulcanizates The molecular mobility of the treated gum IR and their vulcanizates was investigated by diffusion measurements and proton T2 relaxation. The evaluation of the molecular weight and glass transition temperature of the virgin and treated gum IR was made as the complimentary study of the molecular mobility. 8.4.1 Molecular Weight and Glass Transition Temperature The molecular weight and the glass transition temperature for the virgin and ultrasonically treated rubber are summarized in Table 8.2. Ultrasonic treatment induced rubber main chain degradation and broadened the molecular weight distribution accordingly. However, the glass transition temperature shows no substantial change before and after the treatment. This will be supported by the measurement of proton T2S which will be discussed in the Section 8.4.2. This result was in a contrast with the previous studies on the EPDM219 and BR245 where both the rubbers showed a small increase in Tg after the ultrasonic treatment. The decreased mobility observed in EPDM 205 and BR is supported by the measurable amount of gel formed during the ultrasonic treatment. Table 8.2 Molecular weight and glass transition temperature of the virgin and the ultrasonically treated IR gum o Sample Amplitude (µm) Mn Mw Mw/Mn Tg ( C) Virgin IR - 982000 1998000 2.04 -62.2 0 444400 1391000 3.13 -63.0 5 431200 1484000 3.44 -62.3 Treated IR 7.5 294500 1293000 4.39 -63.9 10 164900 1057000 6.41 -63.9 8.4.2 NMR Proton Relaxation Figure 8.1 shows the spectral decomposition (component relative amplitude versus the log relaxation time) of the transverse 1H relaxation decay for the IR firstly ultrasonically treated at 5 µm and then vulcanized using the recipe in Table 3.2. This spectral decomposition shows a wide distribution of decay. The observed distribution of the relaxation rates reflects the differences of segmental and/or translational mobility among the motional species. From Figure 8.1, it is suggested that the T2 relaxation decay for the treated gum IR and the vulcanizates can be more suitably described by the three- component model than the two-component model possibly due to the reason that ultrasound may generate an additional component. The curve of proton T2 relaxation decay together with a three-component modeling is given in Figure 8.2. While the 206 1 o Figure 8.1 Spectral decomposition of the H T2 relaxation decay at 70.5 C for the vulcanizate of IR ultrasonically treated at the amplitude of 5µm 207 Figure 8.2 Transverse 1H relaxation decay at 70.5oC for the vulcanized IR in which the gum was ultrasonically treated at the amplitude of 5 µm fitted with the three- component model containing the Weibull exponent shortest component probably comes from the chemically crosslinked and physically entangled network, the longest component seems to come from the mobile oligomers making possible diffusion measurements. The intermediate component must arise from the remainder, i.e., the unentangled sol molecules and perhaps the dangling network chain ends. Figure 8.3 and Figure 8.4 show the proton T2 relaxation time as a function of the ultrasonic amplitude for the ultrasound treated gum IR and their vulcanizates, 208 respectively. From these two figures, it is observed that both similarities and differences exist in the gums and the correspondent vulcanizates as far as the T2 dependence on the ultrasonic amplitude is concerned. Particularly, the similarities are: T2S is independent of the amplitude and T2M showed positive change with the amplitudes. The difference is that the T2L is independent of amplitude before vulcanization; but it increases with the amplitude after vulcanization. In addition, comparing the correspondent T2 values before (Figure 8.3) and after the vulcanization (Figure 8.4), a great reduction of T2L, T2M and T2S is found due to the restricted mobility by crosslinking. Furthermore the T2S of the gum IR before and after the ultrasound treatment is unchanged (Figure 8.3). This observation seems to be consistent with the unchanged Tg (Table 8.2) discussed in Section 8.4.1. 100 T2L , ms 10 2 T T 2M T2S 1 024681012 µ Amplitude, m Figure 8.3 Proton T2 for the gum IRs ultrasonically treated at different amplitudes 209 100 T2L , ms 10 2 T T2M T2S 1 024681012 Amplitude, µm Figure 8.4 Proton T2 for the vulcanizates of ultrasonically treated gum IR at different amplitudes Figure 8.5 and Figure 8.6 shows the T2 fraction as a function of the ultrasonic amplitude for the ultrasonically treated gum IR and their vulcanizates, respectively. The intermediate T2 fractions for the treated gum IR slightly decreased. The short as well as the long T2 fractions slightly increased with the amplitude. It seems that the dependence of T2 fractions on the ultrasonic amplitude for the vulcanizates follow the same trend as observed in the treated gums. Some of the intermediate component radicals may either decompose upon ultrasonic treatment and then terminate to form smaller molecules or combine with short component radicals to form larger molecules. 210 100 80 f2S 60 fractions, % fractions, 2 40 20 Proton T f2M f 0 2L 024681012 Amplitude, µm Figure 8.5 Proton T2 fractions for the gum IR ultrasonically treated at different amplitudes 100 f2S 80 60 fractions, % fractions, 40 2 20 Proton T f2M 0 f2L 024681012 Amplitude, µm Figure 8.6 Proton T2 fractions for the vulcanizates of ultrasonically treated gum IR at different amplitudes 211 8.4.3 NMR Diffusion In this investigation, diffusion measurements were carried out on the virgin and ultrasonically treated gum IR and their vulcanizates. A sample diffusion echo attenuation is shown in Figure 8.7 for the gum IR passing through the extruder only. The delay 1 spacing τ chosen resulted in essentially the full intensity of the long H T2 component being observed with a minor contribution from the intermediate component, and avoiding any inclusion of the short component. Generally a bimodal diffusivity spectrum is observed in Figure 8.7. The data suggest that the fast-diffusing portion arises from the oligomeric material relaxing at T2L and that the slow-diffusivity portion originates from the unentangled molecular fragments, i.e., the non-oligomeric light sol relaxing at T2M. Figure 8.7 Diffusion spin-echo attenuation of the gum IR extruded at the flow rate of 0.63 g/s, a barrel temperature of 120oC and a gap of 2.54 mm 212 Figure 8.8 shows the diffusion coefficients, Dfast and Dslow plotted as a function of the ultrasonic amplitude. Clearly diffusing coefficients for the fast and slow species have a difference of about two orders of magnitude. It is also seen that ultrasonic amplitude does not alter the magnitude of the diffusing coefficient for each species. This is also consistent with the unchanged Tg of the treated gum IR shown in Table 8.2. -6.0 -6.5 /s) 2 -7.0 Dfast Dslow treated gum -7.5 vulcanizate Log (D, cm -8.0 -8.5 024681012 µ Amplitude, m Figure 8.8 Diffusion coefficient for the ultrasonically treated IR gum and their vulcanizates as a function of ultrasonic amplitude Figure 8.9 shows the drop of fast-diffusing echo fraction ffast with the increase of ultrasonic amplitude for the treated gum IR and the vulcanizates. This evaluation is based on the echo height at the delay time chosen. A more useful approach is the sample fraction of the fast-diffusing component Ffast. The desired quantity, Ffast, is obtained with 213 the cancellation of T2 effect as expressed in the Equation 8.4 and 8.5. Figure 8.10 shows the weak dependence of Ffast on the ultrasonic amplitude for the treated gum IR and their vulcanizates. For the gum and vulcanizates, the portion of oligomer which contributed to the fast-diffusion is only about 1 to 2.6% indicating ultrasound is inefficient to detach a significant amount of oligomeric species. Vulcanization incorporates a limited amount of oligomer molecules into the chemical network, with the rest of the light molecules left for diffusion. 80 , % fast f τ) 60 40 20 treated gum vulcanizate Fast-diffusing echo fraction (at t=2 (at fraction echo Fast-diffusing 0 024681012 Amplitude, µm Figure 8.9 Fast-diffusing echo fraction for the ultrasonically treated gum IRs and their vulcanizates as a function of ultrasonic amplitude 214 10 , % fast treated gum treated vulc 5 Fast-diffusing sample fraction F fraction sample Fast-diffusing 0 024681012 µ Amplitude, m Figure 8.10 Fast-diffusing sample fraction for the ultrasonically treated gum IRs and their vulcanizates as a function of ultrasonic amplitude 8.5 Molecular Mobility of the Unfilled IR Due to the high molecular weight of IR NATSYN®2200 containing traces amount of oligomers, the diffusion measurement was not possible to perform for the virgin uncured and cured unfilled IR. However, the measurement of T2 relaxation decay of these rubbers was possible to carry out. 8.5.1 Sol Fraction and Glass Transition Temperature The sol fraction and the glass transition temperature (Tg) of the virgin gum, vulcanized, and devulcanized IR at the flow rate of 0.63 g/s are summarized in Table 8.3. Clearly the ultrasonic devulcanization creates additional amount of sol. The increased amount of sol is observed at higher ultrasonic amplitude. The glass transition temperature of the virgin IR is increased after the vulcanization, which is due to the decreased 215 mobility by the crosslinking. However the glass transition temperature is not substantially changed by ultrasonic devulcanization. This observation is similar to what was occurred in BR245. But it is different from SBR122 and silicone rubber241 where the glass transition temperature was found to exceed those of the respective virgin vulcanizates. Table 8.3 Sol fraction and glass transition temperature of the virgin gum, vulcanized, and devulcanized IR at the flow rate of 0.63 g/s Sample Amplitude (µm) Sol fraction (%) Tg (°C) Gum 100 -62.2 Virgin IR Vulcanizate 1.93 -59.0 5 2.49 -59.4 Devulcanized IR 7.5 3.28 -59.6 10 10.91 -59.3 8.5.2 NMR Proton Relaxation Figure 8.11 shows the transverse 1H relaxation decay for virgin IR analyzed using a two-component model with only three adjustable parameters: T2S, T2L and fS. The line shape of the transverse decay indicates that this decay is in between the exponential (the Weibull exponent E corresponding to 1) and the Gaussian (E corresponding to 2) decay. Therefore, the two-component model without including the Weibull exponent is not 1 sufficient to describe the H T2 relaxation for virgin IR. In order to facilitate the further analysis, the decay data are analyzed by a spectral decomposition method shown in Figure 8.12. From Figure 8.12, it is convincing that the two-component model is a 216 suitable description of virgin IR proton transverse relaxation; however it is necessary to include the Weibull exponent as an additional adjustable parameter. The fitting result with this consideration is shown in Figure 8.13. Similar analysis is applied to cured IR and the other nine samples devulcanized at different flow rates and ultrasonic amplitudes. The fact that the virgin IR gum is equally well described by two component model with only quantitative differences from the cured and devulcanized samples indicates that the short T2 component characterizes all the network segmental mobility, both chemically crosslinked (gel) and physically entangled (heavy sol). The long T2 component is likely resulted from mobile unentangled sol and perhaps the dangling network chain ends. Figure 8.11 Transverse 1H relaxation decay at 70.5oC for the virgin IR fitted with the two-component model with the exclusion of the Weibull exponent 217 1 Figure 8.12 Spectral decomposition of the H T2relaxation decay for the virgin IR at 70.5oC 218 Figure 8.13 Transverse 1H relaxation decay at 70.5oC for the virgin IR fitted with the two-component model containing the Weibull exponent A plot of chemically extracted sol fraction as a function of ultrasonic amplitude for the cured and devulcanized rubber is shown in Figure 8.14. It is evident that an additional amount of sol is created by ultrasonic devulcanization. This effect is more significant at higher ultrasonic amplitude. With more sol produced, more mobility would be expected in the devulcanized rubber. This speculation could be easily verified in the T2 relaxation time change as shown below. The T2 values for short and long components as a function of sol fraction are shown in Figure 8.15. In this figure, the virgin IR which contains 100% sol, has the highest mobility among all the samples (i.e., T2 for the short and long components has the 219 highest value among all). In contrast, the cured IR which contains the lowest sol fraction, shows the lowest mobility. The T2 values for the devulcanized samples are located in between these two extremities. They showed positive changes with the sol fraction. The scattering of the data at low sol fraction area comes from the imperfection of the two- component model. The fact that devulcanization data are heavily located in the low sol fraction area (shown in Figure 8.15) suggests that the applied ultrasonic conditions (temperature, gap and amplitude) are ineffective to produce significant amount of sol. In other words, ultrasound only results in partial decrosslinking of the rubber networks. From Figure 8.15 it is clear that the vulcanization decreases the T2 for both long and short components. In contrast, ultrasonic devulcanization reverses this effect. And this result is consistent with the observations in many rubber systems240-245. 35 flow rate, g/s 30 0.63 1.07 25 2.10 20 15 Sol fraction, % fraction, Sol 10 5 0 024681012 µ Ultrasonic amplitude, m Figure 8.14 Chemically extracted sol fraction as a function of ultrasonic amplitude for the unfilled IR devulcanized at a barrel temperature of 120oC, a gap of 2.54 mm and varied flow rates 220 100 T2L (ms) 2 10 H T 1 T2S 1 0 20406080100 Extracted sol (%) Figure 8.15 Proton T2 for the virgin, the cured and the devulcanized IR as a function of extracted sol The relative amplitude contribution of the long and short components on the T2 1 delay is evaluated by H T2 fraction. Figure 8.16 shows the T2 fraction as a function of sol fraction for the virgin uncured, cured and devulcanized gum IR. It is seen a slight increase of short T2 fraction at the expense of long T2 fraction, but this change is not significant. The bold line labeled with fsol in Figure 8.16 represents the extractable sol fraction itself. The fact that the majority of the short T2 components is well above the sol fraction suggests that the short T2 components also comes from the chemically extractable sol and perhaps the physically entangled macromolecules in addition to the chemical network segment. 221 100 fS 80 60 fsol fractions, % fractions, 40 2 H T 1 20 0 fL 0 20406080100 Extracted sol, % Figure 8.16 Proton T2 fractions for virgin, cured and devulcanized IR as a function of extracted sol 8.6 Molecular Mobility of the CB Filled IR Among the eleven samples chosen (Table 8.1), the processing oil is the only one whose diffusion measurement was possible. The unavailabity of the NMR diffusion measurement for all the black filled rubber is possibly due to the great difference of the susceptibility between the carbon black and the rubber. Figure 8.17 shows the diffusion curve of the oil. The slight departure of the diffusion curve from the linearity suggests that the oil is not a single component hydrocarbon and the molecular weight of the oil has some polydispersity. The arithmetic average log value of the diffusion coefficient (cm2/s) is determined to be -5.73. This result suggests that the processing oil used in this study is highly mobile compared with the diffusion coefficient of IR (Figure 8.8). 222 Figure 8.17 Diffusion spin-echo attenuation of the processing oil at 70.5oC A typical T2 relaxation decay curve is shown in Figure 8.18 by taking the 35 phr CB filled IR as the example. From the correspondent spectral decomposition shown in Figure 8.19, it is convincing that a two-component model is an appropriate description for the proton transverse relaxation decay of the black filled IR. A plot of chemically extracted sol fraction as a function of ultrasonic amplitude for the cured and the devulcanized 35 phr CB filled IR containing 0 and 10 phr processing oil is shown in Figure 8.20. It is evident that the additional amount of sol is generated upon the ultrasonic devulcanization. This effect is more substantial at higher ultrasonic amplitude. This is similar to what is occurred in the unfilled IR (Section 8.5.1). It is also observed that the rubbers containing the processing oil have more chemically 223 extracted sol than those without oil. Particularly, in the virgin cured 35 phr black filled IR, the difference of the sol fraction between the rubber with oil and the one without oil is 7.57%. The oil content in the filled rubber is 6.49% according to the recipe shown in Table 3.2. Based on the verification that the processing oil in use is soluble in benzene, it can be indicated that the more extracted sol generated in the filled rubber containing oil could be partially from the extractable oil and partially from the less bound rubber49 in the oil containing compounds. The latter is due to the adsorption of oil to the surface of carbon black. Figure 8.18 Transverse 1H relaxation decay for the 35 phr CB filled gum IR at 70.5oC fitted with the two-component model 224 1 Figure 8.19 Spectral decomposition of the H T2 relaxation decay for the 35 phr CB filled gum IR at 70.5oC 225 30 oil, phr 25 0 10 20 15 Sol fraction, % 10 5 0 024681012 µ Ultrasonic amplitude, m Figure 8.20 Chemically extracted sol fraction as a function of ultrasonic amplitude for the 35 phr CB filled IR (containing 0 and 10 phr processing oil) devulcanized at a barrel temperature of 120oC, a gap of 2.54 mm and a flow rate of 0.63 g/s Figure 8.21 shows the comparison of the proton short T2 values for the 35 phr black filled rubber without and with 10 phr oil. The short T2 for the oil only is plotted on the right ordinate axis (marked as an open triangle) as the reference. The T2S values corresponding to the lowest and the 100% extracted sol are attributed to the vulcanized and the virgin filled uncured rubber, respectively. The T2S values for the devulcanized rubber at various ultrasonic amplitudes are located in between these two extremities. First of all, it is observed in Figure 8.21 that among all the eleven samples tested, the processing oil has the greatest mobility identified by the highest T2S value. After 226 inclusion of the carbon black and the IR, the mobility is less as indicated by the reduction of T2S by more than one order of magnitude. This is reasonable since the oil is composed of low molecular weight (typically at several hundred level) hydrocarbons and the molecular weight of virgin gum IR is over 106 (Section 5.6). Generally the lower molecular weight contributes to more mobility. Another reason leading to less mobility of the filled rubber than the pure oil is due to the addition of the filler carbon black. The surface of carbon black absorbs both the rubber (to form the bound rubber49) and the oil molecules. As a result, the mobility of the bound rubber and the oil is restricted by the presence of carbon black. 100 filled without oil filled with oil oil only 10 , ms 2S H T 1 1 0.1 020100 Extracted sol, % Figure 8.21 Proton T2S for the virgin uncured, cured and devulcanized 35 phr CB filled IR without and with oil as a function of extracted sol 227 In Figure 8.21, for the filled rubber (uncured, vulcanized or devulcanized), it is observed that the T2S of the compounds without oil is close to that of the comopounds with oil at low and 100% sol fraction. This is true because mostly the T2S is coming from the physically and the chemically entangled rubber network. And this technique is incapable of differentiating the contribution of the physical and the chemical network on the short T2. Nevertheless, the dependence of the T2S on the sol fraction relies on whether or not the oil is added in the filled compounds. For the rubber without adding any processing oil, the T2S becomes longer after the rubber is devulcanized and it further increases when more sol is generated at higher ultrasonic amplitude. It may be predicted that if the rubber was exposed to more intense ultrasound by further increasing the amplitude beyond 10 µm, more sol would be continuously generated; however it would not be expected that the short T2 would be able to reach that of the virgin uncured rubber. This is reasonable since the devulcanization by ultrasound would never recover the rubber to the original uncured state due to the simultaneous breakage of the crosslinking network and main chain which is indicated both experimentally by measuring the molecular weight of sol (Chapter VI) and theoretically by simulation of gel fraction and crosslink density (Chapter XI). It is revealed that main chain scission would be unavoidable when the ultrasound is applied to the rubber with the purpose of de- crosslinking the network. In contrast, for the compounds containing the oil, it seems that the short T2 shows independence on the vulcanization and ultrasonic amplitude. This phenomenon suggests that the addition of oil would significantly influence the mobility of short component in the rubber. It is possible that the oil would both soften the rubber 228 network chain and cover the carbon black and thus weaken the effect of carbon black on the short T2. Figure 8.22 shows the long T2 for the 35 phr filled IR with and without the oil. The T2L value for the oil is plotted on the right ordinate axis (open triangle). Again, the T2L corresponding to the lowest and 100% extracted sol belongs to the vulcanized and the virgin filled and uncured rubber, respectively. The T2L for the devulcanized rubber at various ultrasonic amplitudes are located in between the two extremities. filled without oil 100 filled with oil oil only , ms 2L 10 H T 1 1 0 20 100 Extracted sol, % Figure 8.22 Proton T2L for the virgin uncured, cured and devulcanized 35 phr CB filled IR without and with oil as a function of extracted sol Same as the T2S, the T2L value of the oil is much larger than that of the filled rubber. Different from what is observed in the T2S, at the low sol fraction, the T2L values 229 of the filled rubber without oil are lower than those of the filled rubber with oil. This is also reasonable since the hydrocarbons in the oil contain many protons which contribute to the long components transverse relaxation. These long component protons are more mobile compared with the protons from the dangling rubber network chains in the rubber. As a result, it takes longer time for the protons in the oil to interact with each other leading to a larger T2L than those of the rubber without oil. The dependence of the long T2 on the sol fraction is same as that of short T2 on the sol fraction. In detail, for the filled rubber without the oil, the long T2 has positive dependence on the sol fraction. While for the filled rubber added with the oil, the dependence of the T2L on the sol fraction is weak. Similar to the expection in T2S, upon more intense ultrasound treatment, it would expect more and more sol generated but the long T2 values would never reach the corresponding value of uncured filled rubber. Figure 8.23 shows the proton long component fractions fL as a function of the sol fraction for the 35 phr filled rubber with and without 10 phr oil. For the convenience of comparison, the value of fL for the oil is marked on the right ordinate axis shown as an open triangle. The bold line labeled with fsol represents the fraction of the extracted sol versus itself. First of all, it is observed from Figure 8.23 that whenever there is network existing in the rubber compounds (which is the case for the vulcanized and devulcanzied rubber), the fraction of long component is no exceptionally above the chemically extracted sol regardless of if the processing oil is added or not. This suggests that the long components come not only from the chemically extracted sol but also from the dangling network chains which can not be chemically extracted. Secondly, it is observed from 230 Figure 8.23 that regardless of whether or not the oil is added, the fraction of the long component increases after the devulcanization and it increases further with the increase of ultrasonic amplitude. However, it seems that the dependence of the long component fractions on the sol fraction is stronger in the compounds containing oil than in those without oil. filled with oil , % L 100 filled without oil oil only 80 60 40 20 long component fractions f fractions component long 2 0 H T fsol 1 0 20 100 Extracted sol, % Figure 8.23 Proton T2 long component fraction fL for the virgin uncured, cured and devulcanized 35 phr CB filled IR without and with oil as a function of extracted sol 8.7 Conclusions 1H NMR transverse relaxation decay measurements were performed in the ultrasonically treated gum IR and their vulcanizates as well as in the unfilled and the 35 phr CB filled IR. The transverse relaxation decay of the unfilled and filled IR was 231 conveniently described by the two-component model while that of the ultrasonically treated gum IR and their vulcanizates was described by the three-component model possibly due to the reason that ultrasound created additional relaxation component. For the ultrasonically treated gum IR and their vulcanizates, the shortest component came from the chemically crosslinked and physically entangled network, and the longest component arose from the mobile oligomers which made the sample diffusion possible. And the intermediate component came from the remainder, i.e., the unentangled sol molecules and perhaps the dangling network chain ends. After vulcanization, the proton T2 values were reduced due to the restricted motion by the crosslinking. The unchanged short component T2 values for the gum IR before and after the ultrasonic treatment were consistent with no variation in Tg. The NMR diffusion measurement of the ultrasonically treated gum IR and their vulcanizates indicated the bimodal nature of the diffusion: the fast-diffusing portion arises from the oligomeric material relaxing at T2L and that the slow-diffusivity portion is originated from the unentangled molecular fragments, i.e., the non-oligomeric light sol relaxing at T2M. Vulcanization incorporated a limited amount of oligomer molecules into the chemical network, with the rest of the light molecules left for diffusion. The diffusion rate and the component fractions among the treated gum as well as among their vulcanizates were not significantly affected by the ultrasonic treatment. Ultrasound was ineffective to produce a significant amount of oligomer in the degradation of gum IR. For the unfilled IR, the short T2 component not only came from all the gel segments (chemical network), but also from a portion of chemically extractable sol (physically entangled network). The remainder of the latter was probably the unentangled 232 sol and dangling network chain ends contributing to the long T2 decay. The vulcanization decreased the T2 for both long and short components, but ultrasonic devulcanization reversed this effect. Ultrasonic devulcanization increased the amount of chemically extractable sol generated from 2% up to 30% indicating ultrasound was relatively ineffective to significantly destroy the IR network and the breakage of networks only occurred locally. For the CB filled IR, the long components came not only from the chemically extracted sol but also from the dangling network chains which can not be chemically extracted. The filled IR containing the processing oil contained more chemically extracted sol than the rubber without the oil regardless of whether or not the network was severed. The more extracted sol could come from both the extractable oil and the more devulcanization. The addition of oil in the filled IR significantly altered the dependence of short and long T2 on the sol fraction. Namely, the T2S and T2L were increased with more sol generated in the rubber without oil, but this dependence was leveled off in the rubber containing oil. Analysis from three different rubber compounds (ultrasonically treated gums plus their vulcanizates, the unfilled and the CB filled IR) indicated that NMR relaxation technique was ineffective to distinguish the contribution of the physically entangled (heavy sol) and the chemically crosslinked (gel) network to the short component mobility. Ultrasound disrupted both the chemical network crosslinks and the rubber main chain. The latter created the dangling chain ends. Ultrasound detached the molecular fragments from the chemical or physical network and from the unattached longer chain molecules to generate more sol fraction. In addition, ultrasound was inefficient in generating additional 233 fragments of oligomeric species. Much of this information obtained by solid state NMR relaxation technique was unattainable in using other methods. 234 CHAPTER IX 9.SIMULATION OF THE NETWORK STRUCTURES FOR THE DEVULCANIZED UNFILLED AND FILLED IR 9.1 General The continuous devulcanization of rubbers by high power ultrasound has been proven to be an effective and environmental friendly technique to recycle waste rubbers especially the scrap tires. It has been shown that the ultrasonic waves of certain levels, in the presence of pressure and heat, without the aid of any chemicals, rapidly break up the three-dimensional networks leading to rubber devulcanization123, 126, 157, 158, 161. Gel fraction of the devulcanized rubber (g) and crosslink density in the gel (ξ) are the primary network structural parameters characterizing the devulcanization process. The results of many experimental studies reveal that for each specific vulcanized rubber, there is a unique correlation between the normalized gel fraction (g*) and the normalized crosslink density (ξ*). This function is independent of the processing conditions of the continuous ultrasonic treatment: compounding recipes, temperatures, flow rates and ultrasonic amplitudes. Processing conditions only affect the degree of devulcanization. It was reported that157, 248 the percolation concept and the Monte-Carlo technique were successfully applied in order to numerically simulate the sol - gel curve for the devulcanized SBR. In this study another theoretical approach, namely, the Dobson- 235 Gordon classical theory of branched polymers164, 165 will be employed to describe the devulcanized rubber network structures. The main assumption of this theory was a tree- like structure of the branched macromolecules (no loops). Strictly speaking, this assumption was not appropriate for the network polymers; however it was argued that the tree approximation could be applied for the vulcanized rubbers which have many crosslinks per polymer chain249. Below is the model development to be applied in network structures of devulcanized elastomers following the Dobson-Gordon theory. This method was already successfully applied to many other rubbers127, 229, 250. 9.2 Development of the Modeling Equations Firstly it was assumed that the primary polymer chains obey the Flory molecular weight distribution ∞ = − 2 n−1 = wn (1 p) np and ∑ wn 1 (9.1) n=1 where wn is the weight fraction of chains having n monomeric units, and p is the Flory distribution parameter ( 0 < p ≤ 1) and it characterizes the monomer conversion in the polymerization process. The number- and weight-average molecular weight N n and N w following the Flory distribution are the functions of p . Particularly, 1 1+ p N (p) = and N (p) = (9.2) n 1− p w 1− p During the vulcanization, some of the monomeric units form crosslinks with other chains. It was assumed that the vulcanization was achieved by random crosslinking. The degree of vulcanization is characterized by the average fraction (α ) of monomers 236 forming crosslinks. Based on the classical gel theory, if α exceeds a certain critical value: 1 α ≥ α = (9.3) crit − N w 1 an insoluble gel part starts to appear in the system164. In the case of vulcanization of elastomer chains with the Flory molecular weight distribution shown in (9.1), the fraction of the soluble part (sol) above the critical gel point, (9.3), can be found as164 2(1− p) + (2 − p)[αp − (α 2 p 2 − 4αp 2 + 4αp)1/ 2 ] S( p,α) = 1− g( p,α) = (9.4) 2αp 2 where g( p,α) is the gel fraction. Another characteristic of the vulcanized system, which is of interest for the simulation purpose, is the number of elastically active network chains per a primary polymer chain,ξ (elast) . For the Flory distribution, it was determined as follows165: ξ (elast) α = α −υ 1/ 2 α 3 + υ 1/ 2 α ( p, ) N n ( p)[1 ( p, )] [1 2 ( p, )] (9.5) The extinction probability υ( p,α) refers to the chance that a randomly chosen crosslink is not a tie with respect to either one of the two primary chains it links165. The solution for the extinction probability has the following form: υ = + − α −υ −2 {1 [Nn ( p) 1] (1 )} (9.6) The value of ξ (elast) is closely related to the experimentally measured crosslink density in gel, ξ ( g ) , which is the number of network chains per unit volume of gel after removing the sol part from a rubber sample. Conceptually, the relationship between ξ (elast) and ξ ( g ) has the following form: 237 ρ ξ (elast) ( p,α) ξ ( g ) ( p,α) = (9.7) α M 0 N n ( p)g( p, ) ρ where and M 0 are the rubber density and the molecular weight of a monomeric unit of the rubber, respectively. Another assumption was that the rubber devulcanization takes place following an irreversible scission of crosslinks and main chains. If these rupture processes proceed randomly in time and space, the theoretical description of the rubber devulcanization can be achieved. Namely, under the randomness assumption, the initial Flory distribution varies in such a way that it is of the same form as in Equation (9.1) but with a time dependent Flory distribution parameter p(t) 251. It follows from this observation that the model Equations (9.4) to (9.7) remain to be valid and can be employed to describe the evolution of rubber network structure during the devulcanization. Combination of these equations is applied below for the calculations of the normalized crosslink density in gel ξ* and the normalized gel fraction g*: ξ ( g ) ξ ( g ) ( p,α) ξ * ( p,α) = devulc = ξ (g ) ξ ( g ) α 0 ( p0 , 0 ) g 1− S( p,α) g * ( p,α) = devulc = (9.8) − α g 0 1 S( p0 , 0 ) where p0 and p are the initial and current values of the Flory distribution parameter and α α 0 and are the initial and current values of the fraction of crosslinked monomeric units. In the case of random bond scission, the parameters p and α exhibit the exponential decay251 = − α = α − p(t) p0 exp( k pt) and (t) 0 exp( kα t ) 238 where k p and kα are the scission rate constants for main chains and crosslinks, respectively. The above two equations can be combined together and re-written in the following form k / kα p(t) α(t) p = (9.9) α p0 0 M = − 0 The p0 value was determined as p0 1 , where M n is the number-average M n α molecular weight of the rubber. The 0 can be found using the experimental data on ξ (g ) crosslink density 0 and gel fraction g 0 in the virgin rubbers using Equation (9.7) with the aid of Equations (9.4) to (9.7), i.e., ρα ξ (g ) ( p ,α ) = 0 (1− υ )3 (1+ 2 υ ) 0 0 0 α 0 0 M 0 g 0 ( p0 , 0 ) −2 1 υ = 1+ −1α (1−υ ) 0 − 0 0 1 p0 Finally Equations (9.8) and (9.9) define the desirable function g* (ξ*) in a parametric form. They are the basis to simulate the network structures of the devulcanized IR with the aid of Equations (9.4) to (9.7). 9.3 The Unfilled IR The devulcanization of the ground rubbers with ultrasound not only led to the network rupture but also unavoidably caused the main chain scission. In order to investigate the relative degree of these two effects, a model based on the random ruptures of rubber main chains and crosslinks 163,164,165 was applied to interpret all the experimental data by fitting the parameter kp/kα for each rubber, with kp and kα being the 239 rate constants of the rupture of main chains and crosslinks, respectively. The material parameters of the unfilled IR and NR used in this simulation are shown in Table 9.1. Table 9.1 Physical and chemical parameters of the unfilled IR used in the simulation Rubber IR NR182 Density, kg/m3 920 920 Mn, g/mol 982,000 180,400 Mw, g/mol 1,998,000 1,116,000 Monomeric unit weight M0, g/mol 68.11 68.11 Initial gel fraction g 0 0.9801 0.9622 ξ ( g) 3 0.2356 0.1569 Initial crosslink density 0 , (kmol/m ) First, the dependence of the experimental normalized gel fraction on the normalized crosslink density of IR and NR devulcanizates was analyzed by taking the normalized quantities of original vulcanized rubber to be unity. There were two limiting cases in the model, with kp=0 indicating only crosslink rupture and kα=0 indicating only main chain rupture. The experimental data characterizing the actual partial crosslink rupture and partial main chain scission in the rubbers should lie in between these two limiting cases. The experimental (symbols) and fitted (lines) results for the unfilled rubbers are shown in Figure 9.1. It is found that the limiting case of kα=0 is independent of the rubber type and the molecular weight. The other limit of kp=0 and fitted curves to experimental data are both dependent on the molecular weight as well as the rubber type even though NR and IR share the same main chain structure. It should be noted that the 240 1.0 0.8 0.6 -3 kp/kα=4.1x10 (IR) -3 0.4 kp/kα=4.2x10 (NR) kα=0 kp=0 (IR) Normalized gelfraction 0.2 kp=0 (NR) IR NR 0.0 0.00.20.40.60.81.0 Normalized crosslink density Figure 9.1 Experimental (symbols) and fitted (lines) values of normalized gel fraction as a function of normalized crosslink density for the devulcanized unfilled IR and NR 252 number average molecular weight Mn of NR used was 180,400 , which is about 22% of the Mn of IR. With the higher molecular weight of IR, the curves shift in the direction of lower crosslink density. This is reasonable, since the rubber chain with higher number of monomer units is able to accommodate a higher gel content at the same crosslink density level. Nevertheless, the simulation results show that these two rubbers share close kp/kα -3 -3 values. For IR and NR, the kp/kα value of 4.1×10 and 4.2×10 are obtained, respectively. This suggests that the probability of main chain scission over the crosslink rupture is equivalent regardless of the molecular weight as long as the rubbers have the 241 same main chain structure and stereo-regularity. On the other hand, the very low kp/kα values found in the simulation of both rubbers indicates that the main chain containing C- C bonds is much more difficult to break down than the crosslink bonds under the exposure of ultrasound. The main chain (C-C) bond energy is about 346 kJ/mol being much higher than that of the crosslink network containing C-S (mono-sulfidic), S-S (di- 253 sulfidic) and Sx (multi-sulfidic) with the bond energies of 285, 268 and 251 kJ/mol , respectively. Therefore, low values of kp/kα resulting from the curve fitting are in an excellent agreement with the bond energy values of main chains and crosslinks. 9.4 The CB Filled IR In the simulation of gel fraction and crosslink density for the CB filled IR and NR, the same method is adopted as it was done in the unfilled rubbers. The material parameters such as the Mn, Mw, and M0 used in the simulation of the filled rubbers still follow those used in the unfilled rubbers listed in Table 9.1. However for the initial gel ξ ( g) fraction g 0 and crosslink density 0 , one has to adopt the values belonging to the vulcanizates of the correspondent CB level. The detailed values are listed in Table 9.2. ξ ( g) Table 9.2 Initial gel fraction g 0 and crosslink density 0 of the filled IR used in the simulation IR NR182 CB/oil (phr) g ξ ( g) 3 g ξ ( g) 3 0 0 (kmol/m ) 0 0 (kmol/m ) 15/0 0.9615 0.1727 0.9690 0.1765 35/0 0.9718 0.2126 0.9690 0.1710 35/10 0.8961 0.1775 − − 242 Figure 9.2 shows the experimental (symbols) and the calculated (lines) values of the normalized gel fraction versus the normalized crosslink density for the filled IR and NR at varied CB loadings. The result of black filled IR containing 10 phr oil is also included in the same figure for the convenience of comparison. Similar to the unfilled IR, the simulation of the filled IR results in a fairly good agreement with the experimental data. It is observed that in the filled rubber, the curves of the gel fraction versus the crosslink density shift in the direction of higher crosslink density compared with that of the unfilled IR. Accordingly the fitted value of kp/kα is larger. This suggests that the chance of disrupting the main chain relative to the crosslink is higher in the rubber loaded with CB than in the unfilled one. This effect is even stronger with the increasing CB loading as seen from Figure 9.2. This is probably caused by the immobilized bound rubber located in the vicinity of the carbon black surface254. The restricted motion of the bound rubber chains by the CB makes them experience more tension and thus they are more vulnerable under the ultrasound exposure compared to those relatively flexible chains in the unfilled rubber. Therefore it is logical to find that after adding the processing oil into the filled rubber this effect is weakened which is revealed by the drop of the kp/kα values. Possibly the processing oil has a plasticizing effect on the bound rubber. In the case of filled NR, similar to the occurrence in filled IR, it is also seen that kp/kα value is increased with the increase of CB loading. The main chain scission and crosslink rupture ratio of 15 phr CB filled IR and NR is at the same order of magnitude. The difference of kp/kα values for IR and NR at the CB loading of 15 phr is small. This is also probably due to the same stereoregular structure of two rubbers. In contrast, the kp/kα 243 1.0 0.8 0.6 kα=0 kp=0 (IR) kp=0 (NR) 0.4 -3 exp predic CB/oil (phr) kp/kα (x10 ) 15/0 5.3 Normalized gel fraction 0.2 IR 35/0 17.1 35/10 4.0 15/0 6.7 NR 35/0 8.1 0.0 0.00.20.40.60.81.0 Normalized crosslink density Figure 9.2 Experimental (symbols) and fitted (lines) values of normalized gel fraction as a function of normalized crosslink density for the devulcanized filled IR and NR ratio of the 35 phr CB filled IR is much larger than NR at the same CB loading. This suggests that the main chain of 35 phr filled IR would experience a greater probability of scission than that of NR at the same CB loading. It was established that higher molecular weight of polymer is easier to be degraded by ultrasound and there is a limiting molecular weight value below which the polymer resists further degradation146-150. By comparing the molecular weight and its distribution of NR (Mn: 180,400; Mw: 1,116,000; Mw/Mn: 182 6.19) used in earlier study and IR (Mn: 982,000; Mw: 1,998,000; Mw/Mn: 2.04) used in this study, it indicates that the more chain scission in 35 phr filled IR than that in NR at 244 the same black loading is due to the higher molecular weight and narrower molecular weight distribution of IR than those of NR. 9.5 Conclusions The Dobson-Gordon classical theory of rubber network statistics was employed to describe the rubber network structures. The normalized gel fraction in the devulcanized rubbers as a function of normalized crosslink density in the gel was described by a unique curve. This approach involved only one adjustable parameter describing the relative ratio of the rates in rupturing the network and main chain chemical bonds. The assumption of spatial-temporal randomness of rubber network rupture and main chain scission in the simulation allowed us to achieve a fairly good agreement between the experimental and theoretical values of network structures. This was true for both the unfilled and the CB filled IR. The kp/kα values obtained from the simulation of the experimentally measured normalized gel fraction and crosslink density of the filled and the unfilled IR suggested that the crosslink rupture dominated over the main chain scission. This result was consistent with the bond energy values of main chain and crosslinks. For the unfilled rubbers, the almost equivalent ratio of kp/kα values for IR and NR rubbers indicated that they had the same probability of main chain scission and crosslink rupture ratio, which was evidently determined by the same main chain stereoregular structures of both rubbers. The kp/kα value increased when the rubber was filled with carbon black and the value further increased at higher carbon black loading. The chance of severing the main 245 chain relative to the crosslink was higher in the highly filled IR. However this effect was weakened by adding the processing oil into the filled IR. 246 CHAPTER X 10.SUMMARY This research was aimed at the application of the high power ultrasound technique to devulcanize the unfilled and the carbon black filled isoprene rubber. At first, the isothermal and non-isothermal cure kinetics study of the unfilled and the filled IR were carried out. The cure kinetics study was followed by the investigation of the effect of ultrasound on the structure and properties of the gum rubber in order to examine the stability of the main chain linkage carbon-carbon backbone under the exposure of high power ultrasound. The applicability of the ultrasound to the devulcanization of the unfilled and the carbon black filled IR was intensively investigated by studying the effect of ultrasound on a wide variety of properties. The solid-state NMR technique was used to study the molecular mobility of ultrasonically treated gum IR and devulcanized unfilled and black filled IR. This NMR study along with the simulation of the network gel fraction and crosslink density based on the Gordon-Dobson theory of the elastic network was employed to understand the ultrasonic devulcanization mechanism of the unfilled and the filled IR. The reversion type of cure kinetic model was successfully utilized to predict the isothermal and non-isothermal evolution of the state of cure in curing the unfilled and carbon black filled IR after the induction period. The induction time function with the 247 Arrhenius dependence on the temperature was introduced as an explicit kinetic function. The mathematical model proposed in this study adequately described the dependence of Γ − Γ isothermal torque difference max min on temperature. Stronger reversion observed in the filled IR than in the unfilled IR was explained by comparing reversion kinetic rate constants k30 obtained from simulation. The stability of the kinetic model in use was verified by individual and simultaneous modeling of the isothermal state of cure data. The universality of the simplified kinetic model and the non-isothermal induction time concept was verified in the unfilled and the black filled IR. Ultrasonic treatment altered the structure and properties of gum IR and the change was highly amplitude dependent. The indication of severing the carbon-carbon backbone in gum IR was supported by the observation of die pressure drop and the measurement of molecular weight, dynamic and mechanical properties. Ultrasound treatment created low molecular weight tails which broadened the molecular weight distribution and thus improved the processability of gum rubber. The fitting of the complex viscosity – frequency curves according to the modified Cross model indirectly indicated possible branching of gum IR treated at the ultrasonic amplitude of 5 µm. The treated gum rubber vulcanizates retained the stess-induced crystallization although the main chain was severed. The tensile strength, modulus, elongation at break of the treated rubber vulcanizates were reduced with the increase of amplitude. However, the elongation at break of the treated IR vulcanizates was higher than that of the virgin vulcanizate. The cure kinetics of the treated rubber gums was similar to the virgin IR and showed the reversion. The vulcanization created a comparable amount of gel but a significantly lower crosslink density for the treated rubber gums compared with the virgin rubber. 248 Sulfur-cured vulcanizate of the unfilled IR was devulcanized using a coaxial ultrasonic reactor in the same way as NR was done earlier. The two rubbers showed some similarities in the vulcanization, devulcanization, revulcanization, and network structure due to the same main chain structure existed in NR and IR. They had an induction period close to each other and both of them showed a reversion in the vulcanization. The induction period was absent in the revulcanization. The die pressure during devulcanization continuously dropped with the increase of ultrasonic amplitude. At all the ultrasonic amplitudes, the gel fraction and crosslink density were lower than those of the original cured rubber. There were some differences in the rate of vulcanization, the level of revulcanization and the dependence of the degree of devulcanization on the amplitude between IR and NR. The cure rate of IR was lower than that of NR contributing to more uniform vulcanized sheets of IR than those of NR. Unlike the IR, the NR sample devulcanized at the highest amplitude 10 µm could reach a higher level of revulcanization than those of the 5 and 7.5 µm samples. This was supported by the torque attained upon revulcanization, the gel fraction and crosslink density, the power consumption and the mechanical properties. The dynamic properties of the unfilled IR devulcanizates indicated that the complex viscosity was proportional to the degree of crosslinking. Less degree of shear thinning of the samples devulcanized at higher ultrasonic amplitude was observed. More significant changes of all the properties took place when the amplitude went from 5 to 7.5 µm compared with those variations when the amplitude increased from 0 to 5 µm and 7.5 to 10 µm. 249 Similar to the unfilled IR, the continuous ultrasonic devulcanization of the carbon black filled IR was also carried out by using the co-axial reactor. The power consumption during the devulcanization of IR did not show a maximum value at certain ultrasonic amplitude which was a typical occurrence in the NR. This suggested that extent of devulcanization increased with ultrasonic amplitude in IR. Furthermore, the trend of power consumption and extent of devulcanization agreed with the change of other properties such as gel fraction, crosslink density and the mechanical properties for both rubbers. The different behaviors occurred in two rubbers were probably resulted from the slightly higher amount of stereoregular structures of NR than that of IR. More significant drop of the mechanical properties in the revulcanizates of the filled IR than those of the unfilled IR was resulted from more main chain scission over the network rupture in the bound rubber. This was verified in the simulation of the network structures. The inclusion of the processing oil in the filled IR compounds slightly lengthened the induction period of vulcanization and lowered curing torque. However it did not affect the degree of reversion. The addition of the oil led to more devulcanization as indicated by the lower gel fraction and crosslink density. The oil decreased the dynamic viscosity and storage modulus of the uncured, cured and certain devulcanized filled IR. The induction period was absent in the revulcanization of the devulcanized filled IR. The addition of the processing oil did not improve this situation. The retarder was combined in the revulcanization recipe and it was effective to improve the scorch safety and minimize the reversion in revulcanization. The ultrasonically devuclanized IR vulcanizates (dIR) were blended with the virgin IR and the properties were compared with the blends of fully cured ground IR 250 (gIR) and virgin IR. The increased proportion of the virgin IR led to the extension of induction period for both blends. The dIR/IR blends showed much better tensile strength and elongation at break than the gIR/IR blends. The dependence of the tensile strength of the dIR/IR blends on the proportion of virgin IR followed the rule of mixing. The improved tensile properties of the blends proved that ultrasonic devulcanization was an effective way of recycling the carbon black filled IR. NMR proton relaxation technique was employed to analyze the molecular mobility of the rubber with the purpose to understand the effect of ultrasound on the degradation of gum IR and on the devulcanization of the unfilled and the CB filled IR at the molecular level. The proton transverse relaxation decay of the unfilled and the black filled IR was described by the two-component model. Particularly the short T2 component came from chemically crosslinked (gel) and physically entangled network (heavy sol). The long T2 decay came from the unentangled sol and dangling network chain ends. The vulcanization decreased the T2 for both long and short components, but ultrasonic devulcanization reversed this effect. Addition of the processing oil in the filled IR significantly altered the dependence of short and long T2 on the sol fraction. The proton T2 relaxation decay of the ultrasonically treated gum IR and their vulcanizates was described by the three-component model. The shortest component came from the chemically crosslinked and physically entangled network, and the longest component came from the mobile oligomers which made the sample diffusion possible. And the intermediate component arose from the remainder, i.e., the unentangled sol molecules and perhaps the dangling network chain ends. 251 The NMR diffusion of the ultrasonically treated gum IR and their vulcanizates had the bimodal nature of the diffusion: the fast-diffusing portion arose from the oligomeric species relaxing at the longest T2 and that the slow-diffusivity portion originated from the unentangled molecular fragments, i.e., the non-oligomeric light sol relaxing at the intermediate T2. NMR relaxation technique was unable to differentiate the contribution of short component mobility between the physically entangled and the chemically crosslinked network. Ultrasound severed both the chemical network crosslinks and the rubber main chain. The latter created the dangling chain ends. However, it was inefficient to generate additional fragments of oligomeric species. Much of this information obtained by NMR relaxation technique was unattainable in using other methods. The Dobson-Gordon theory of network statistics was employed to describe the network structures in the devulcanized unfilled and black filled IR. The normalized gel fraction was expressed as a function of the normalized crosslink density in the gel. The assumption of spatial-temporal randomness of rubber network rupture and main chain scission resulted in a fairly good agreement between the experimental and the theoretical values of network structures. The magnitude of the kp/kα values obtained from the simulation led to the conclusion that the crosslinks were more vulnerable than the main chains under the exposure of ultrasound. This was supported by the bond energy values of main chains and crosslinks. The IR and NR had the same probability of main chain scission and crosslink rupture ratio, which was determined by the same main chain structures of both rubbers. The opportunity of severing the main chain relative to the 252 crosslink was greater in the highly filled IR. However, it was weakened by including the processing oil into the filled IR compounds. 253 REFERENCES 1 L.K. Opeke, “Tropical Tree Crops”, John Wiley and Sons, New York, 1982. 2 F.W. Barlow, “Elastomers: Natural Rubber”, in “Rubber Compounding: Principles, Materials and Techniques”, 2nd ed., F. W. Barlow ed., Marcel Dekker, New York, 1993. 3 A. Subramaniam, “Natural Rubber”, in “The Vanderbilt Rubber Handbook”, R. F. Ohm, ed., R. T. Vanderbilt Co., Norwalk, CT, 13th ed., 1990. 4 F.M. Herman, ed., “Encyclopedia of Chemical Technology, Polyisoprene”, vol. 8, 3rd ed., Wiley, New York, NY, pp. 582–592, 1978. 5 J.W. Mapp, P. W. Milner, H. I. Walsh, A. A. Watson, “Storage, Processing, and Vulcanization Properties of High cis-1, 4-polyisoprene”, Revue Generale des Caoutchoucs & Plastiques, 50, 823-828, 1973. 6 W. Kingensmith, K. Baranwal, “Recycling of Rubbers”, in “Handbook of Elastomers”, 2nd ed., A.K. Bhowmick, H.L. Stephens, eds., Marcel Dekker, New York, 2001. 7 M. Blumenthal, Proceedings of the Air & Waste Management Association’s Annual Conference & Exhibition, 94th, Orlando, FL, United States, June 24-28, 5361-5370, 2001. 8 S. K. De, A. I. Isayev, K. Khait (editors), “Rubber Recycling”, Boca Raton, FL.: Taylor & Francis/CRC Press, 2005. 9 C. Charlotte, M. Kenneth, R. Dexter, “The Chemistry and Technology of Rubber”, Noyes Data Corporation: Park Ridge, NJ., U.S.A., 1993. 10 M. Morton, “Rubber Technology”, 3rd ed., Van Nostrand Reinhold Co.: New York, 1987. 254 11 M. Toyoharu, S. Mie, S. Fumio, E. Takeshi, “Six-Membered Cyclic Carbonate Having Styrene Moiety as a Chemically Recyclable Monomer. Construction of Novel Cross-Linking-De-Cross-Linking System of Network Polymers”, Macromolecules, 38, 7944-7949, 2005. 12 A.A. Phadke, A.K. Bhattacharya, S.K. Chakraborty, S. K. De, “Vulcanization of Reclaimed Rubber”, Rubber Chem. Technol., 56, 726-736, 1983. 13 A.A. Phadke, A.K. Bhowmick, S.K. De, “Effect of Cryoground Rubber on Properties of NR”, J. Appl. Polym. Sci., 32, 4063-4074, 1986. 14 H.B. Bode, K. Kerkhoff, D. Jendrossek, “Bacterial Degradation of Natural and Synthetic Rubber”, Biomacromolecules, 2, 295-303, 2001. 15 A. Linos, R. Reichelt, U. Keller, A. Steinbuchel, “A Gram-negative Bacterium, Identified as Pseudomonas Aeruginosa AL98, is a Potent Degrader of Natural Rubber and Synthetic cis-1, 4-polyisoprene”, FEMS Microbiology Letters, 182, 155- 161, 2000. 16 S.R. Fix, “Microwave Devulcanization of Rubber”, Elastomerics, 112, 38-40, 1980. 17 D.S. Novotny, R.L. Marsh, F.C. Masters, D.N. Tally, US 4104205, “Microwave Devulcanization of Rubber”, Goodyear Tire & Rubber Co., USA, 1978. 18 A.I. Isayev, US 5258413, “Continuous Ultrasonic Devulcanization of Vulcanized Elastomers”, University of Akron, USA, 1993. 19 A.I. Isayev, J. Chen, US 5 284625, “Continuous Ultrasonic Devulcanization of Vulcanized Elastomers”, University of Akron, USA, 1994. 20 A.I. Isayev, J. Chen, A. Tukachinsky, “Novel Ultrasonic Technology for Devulcanization of Waste Rubbers”, Rubber Chem. Technol., 68, 267-280, 1995. 21 S. Kawahara, T. Kakubo, N. Nishiyama, Y. Tanaka, Y. Isono, “Crystallization Behavior and Strength of Natural Rubber: Skim Rubber, Deproteinized Natural Rubber, and Pale Crepe”, J. Appl. Polym. Sci., 78, 1510-1516, 2000. 22 A. Subramaniam, “Natural Rubber”, in “Rubber Technology”, 3rd ed., M. Morton, ed., Van Nostrand Reinhold Co., New York, 1987. 23 N.M. Mathew, “Natural Rubber”, in “Rubber Technologist’s Handbook”, S.K. De, J.R. White, ed., Shawbury, Shrewsbury, England: Rapra Technology Ltd., 2001. 255 24 B.L. Archer, B.G. Audley, F.J. Bealing, “Biosynthesis of rubber in Hevea Brasiliensis”, Plastics and Rubber International, 7, 109-111, 1982. 25 I. Franta, ed., “Elastomers and Rubber Compounding Materials: Manufacture, Properties and Applications”, Elsevier, New York, 1989. 26 K.N. Fuller, G.W.S. Fulton, “The Influence of Molecular Weight Distribution and Branching on the Relaxation Behavior of Uncrosslinked Natural Rubber”, Polymer, 31, 609-615, 1990. 27 K.N.G. Fuller, J. Gough, A.G. Thomas, “The Effect of Low-Temperature Crystallization on the Mechanical Behavior of Rubber”, J. Polym. Sci.: Part B: Polym. Phy., 42, 2181–2190, 2004. 28 A.C. Peter, H. Norman, “The Rubber Formulary”, Noyes Publications/William Andrew Publishing, LLC, 1999. 29 P.W. Allen, “Natural Rubber and the Synthetics”, John Wiley & Sons, New York, 1972. 30 F. Cataldo, “Thermal Depolymerization and Pyrolysis of cis-1, 4 polyisoprene: Preparation of Liquid Polyisoprene and Terpene Resin”, J. Analytical & Appl. Pyrolysis, 44, 121-130, 1998. 31 H.M. Lybarger, “Isoprene”, in “Kirk-Othmer Encyclopedia of Chemical Technology”, J.I. Kroschwitz (executive editor) M.H. Grant (editor), vol. 14, 4th ed., Wiley, New York, NY, pp. 934, 1991. 32 W. Cooper, T. Vaughan, in “Progress in Polymer Science”, Vol. I, A.D. Jenkins, ed., Pergamon Press, London, p.128, 1967. 33 S.E. Horne, US 3114743, “Method of Preparing Synthetic Rubber”, Goodrich-Gulf Chemicals, Inc., 1963. 34 L.E. Forman, US 3285901, “Polymerization of Isoprene in the Presence of Lithium Metal and an Inert Liquid Hydrocarbon”, Firestone Tire & Rubber Co. 1966. 35 L.E. Forman, US 3208988, “Polymerization of Isoprene Using a Hydrocarbon Polylithium Catalyst”, Firestone Tire & Rubber Co., 1965. 36 W. Hofmann, “Synthetic Rubber”, in “Rubber Technology Handbook”, W. Hofmann, ed., Hanser Publishers, New York, 1989. 256 37 S. Datta, “Synthetic Elastomers”, in “Rubber Technologist’s Handbook”, S.K. De, J.R. White, ed., Shawbury, Shrewsbury, England: Rapra Technology Ltd., 2001. 38 http://www.iisrp.com/WebPolymers/11POLYISOPRENE.pdf 39 M. Bruzzone, G. Corradini, F. Amato, “Comparison of the Technological Properties of Polyisoprenes of Different cis-1, 4 Content”, Rev. Gen. Cautchouc Plastiques, 42, 1017-1025, 1965. 40 A.N. Gent, S. Kawahara, J. Zhao, “Crystallization and Strength of Natural Rubber and Synthetic cis-1, 4-polyisoprene”, Rubber Chem. Technol., 71, 668-678, 1998. 41 M.A. Golub, S.A. Fuqua, N.S. Bhacca, “High Resolution Nuclear Magnetic Resonance Spectra of Various Polyisoprenes”, J. Am. Chem. Soc., 84, 4981-4982, 1962. 42 G.M. Bristow, J.I. Cunneen, L. Mullins, “Comparative Properties and Performance of Natural Rubber and Synthetic cis-polyisoprene”, NR Technology, 4(Pt. 2), 16-32, 1973. 43 K.S. Loganathan, ed., “Natural Rubber”, “Synthetic Rubbers” in “Rubber Engineering”, Indian Rubber Institute, 2000. 44 J.B. Donnet, A. Voet, ed., “Manufacture of Carbon Black”, in “Carbon Black, Physics, Chemistry, and Elastomer Reinforcement”, Marcel Dekker, New York, 1976. 45 http://www.continentalcarbon.com/downloads/whatis.pdf 46 S. Vieweg, R. Unger, G. Heinrich, E. Donth, “Comparison of Dynamic Shear Properties of Styrene–Butadiene Vulcanizates Filled with Carbon Black or Polymeric Fillers”, J. Appl. Polym. Sci., 73, 495–503, 1999. 47 ASTM D1765, Standard Classification System for Carbon Blacks Used in Rubber Products, 1968. 48 P.G. Maier, D. Goeritz, “Molecular Interpretation of the Payne Effect”, Kautschuk Gummi Kunststoffe, 49, 18-21, 1996. 49 S. Wolf, M. Wang, “Carbon Black Reinforcement of Elastomers”, in “Carbon Black: Science and Technology”, J.B. Donnet, R.C. Bansal, M. Wang, ed., New York, N.Y., Dekker, 1993. 257 50 B.B. Boonstra, “Role of Particulate Fillers in Elastomer Reinforcement: a Review”, Polymer, 20, 691-704, 1979. 51 G. Kraus, ed., “Reinforcement of Elastomers”, Wiley Interscience, New York, 1965. 52 O.J. Ogbebor, A.S. Farid, U.N. Okwu, “Properties of Silica/Clay Filled Heavy-Duty Truck Tire Thread Formulation”, J. Appl. Polym. Sci., 94, 1024–1028, 2004. 53 H. Ismail, R. Nordin, A.M. Noor, “The Effect of Filler Loading on Curing and Mechanical Properties of Natural Rubber/Recycled Rubber Powder Blends”, International Journal of Polymeric Materials, 54, 9-20, 2005. 54 C.E. Scott, V. Chirico, “Performance of SBR and BR Tread Compounds as a Function of Carbon Black Dispersion, Surface Chemistry, and Structure”, Rubber Chem. Technol., 42, 700-713, 1969. 55 A.K. Bagchi, B.G. Sharma, “Reinforcement and Physical Properties of Filled Rubber System”, Indian Journal of Technology, 19, 368-372, 1981. 56 H. Gong, “Development and Prospects in Studies of Reinforcement of Carbon Black on Rubber”, Xiangjiao Gongye, 40, 53-58, 1993. 57 T. Sajjayanukul, P. Saeoui, C. Sirisinha, “Experimental Analysis of Viscoelastic Properties in Carbon Black-Filled Natural Rubber Compounds”, J. Appl. Polym. Sci., 97, 2197-2203, 2005. 58 P.L. Cho, G.R. Hamed, “Green Strength of Carbon-Black-Filled Styrene-Butadiene Rubber”, Rubber Chem. Technol., 65, 475-487, 1992. 59 N.K. Dutta, D. Khastgir, D.K. Tripathy, “The Effect of Carbon Black Concentration on the Dynamic Mechanical Properties of Bromobutyl Rubber”, J. Mat. Sci., 26, 177-188, 1991. 60 A. Mongruel, M. Cartault, “Nonlinear Rheology of Styrene-Butadiene Rubber Filled with Carbon-Black or Silica Particles”, Journal of Rheology, 50, 115-135, 2006. 61 P.P. Kundu, T.R. Kukreja, “Surface Modification of Carbon Black by Vegetable Oil - its Effect on the Rheometric, Hardness, Abrasion, Rebound Resilience, Tensile, Tear, and Adhesion Properties”, J. Appl. Polym. Sci., 84, 256-260, 2002. 62 F. Findik, R. Yilmaz, T. Koksal, “Investigation of Mechanical and Physical Properties of Several Industrial Rubbers”, Materials & Design, 25, 269-276, 2004. 258 63 G. Kraus, “Reinforcement of Elastomers by Particulate Fillers”, in “Science and Technology of Rubber”, F. R. Eirich, ed., New York, Academic Press, 1978. 64 S. Choi, K. Hwang, B. Kim, “Influence of Bound Polymer on Cure Characteristics of Natural Rubber Compounds Reinforced with Different Types of Carbon Blacks”, J. Appl. Polym. Sci., 98, 2282-2289, 2005. 65 H.E. Adams, B.L. Johnson, “Cross-linking in Natural Rubber Vulcanizates”, J. Ind. Eng. Chem., 45, 1539-1546, 1953. 66 P.J. Flory, “Theory of Elasticity of Polymer Networks. The Effect of Local Constraints on Junctions”, J. Chem. Phys., 66, 5720-5729, 1977. 67 G. Kraus, “Swelling of Filler-reinforced Vulcanizates”, J. Appl. Polym. Sci., 7, 861- 871, 1963. 68 A.M. Bueche, “The Curing of Silicone Rubber with Benzoyl Peroxide”, J. Polym. Sci., 15, 105-120, 1955. 69 J.R. Wolfe, “The Chemistry of Sulfur Curing. I. Kinetics of Vulcanization of an EPDM Elastomer Accelerated by Zinc Dimethyldithiocarbamate”, J. Appl. Polym. Sci., 12, 1167-1181, 1968. 70 J.R. Wolfe, “The Chemistry of Sulfur Curing. II. Kinetics of Vulcanization of an EPDM Elastomer Accelerated by Tetramethylthiuram Monosulfide and Disulfide”, J. Appl. Polym. Sci., 12, 1183-1192, 1968. 71 M. Chen, N. Ao, B. Zhang, C. Den, H. Qian, H. Zhou, “Comparison and Evaluation of the Thermooxidative Stability of Medical Natural Rubber Latex Products Prepared with a Sulfur Vulcanization System and a Peroxide Vulcanization System”, J. Appl. Polym. Sci., 98, 591-597, 2005. 72 W. Hopkins, H. Von, W. Carl, EP 1016691, “Curing of Butyl Rubber with a Phenolic Resin”, 2000. 73 J.N. Coker, US 3275607, “Curing Thiocarbonyl Fluoride-allyl Chloroformate Copolymers with Polyvalent Metal Oxides”, 1966. 74 L. Bateman, ed., “The Chemistry and Physics of Rubber-like Substances,” John Wiley & Sons, NY, 1963. 75 W. Hofmann, “Vulcanization and Vulcanizing Agents”, MacLaren, London, England 1967. 259 76 K.S. Loganathan, ed., “Rubber Vulcanization”, in “Rubber Engineering”, Indian Rubber Institute, McGraw, NY, 2000. 77 M. Porter, “The Chemistry of the Sulfur Vulcanizatioin of Natural Rubber”, in “The Chemistry of Sulfides,” A. V. Tobolsky, ed., Interscience Publishers, New York, 1968. 78 J.R. Wolfe, T.L. Pugh, A.S. Killian, “Chemistry of Sulfur Curing. III. Effects of Zinc Oxide on the Mechanism of the Reaction of Cyclohexene with Sulfur”, Rubber Chem. Technol., 41, 1329-1338, 1968. 79 P.A. Ciullo, N. Hewitt, “The Rubber Formulary”, Noyes Publications/William Andrew Publishing, LLC, 1999. 80 “Vulcanization”, in “Encyclopedia of Polymer Science and Engineering”, H. Mark, N. M. Bikales, C. Overberger, G. Menges, Eds., Wiley, New York, v. 17., 1989. 81 M.R. Krejsa, J.L. Koenig, “A Review of Sulfur Crosslinking Fundamentals for Accelerated and Unaccelarated Vulcanization”, Rubber Chem. Technol., 66, 376-410, 1993. 82 D.J. Elliot, “Developments with Natural Rubber”, in “Developments in Rubber Technology-1”, A. Whelan, K. S. Lee, ed., Applied Science Publishers, London, 1979. 83 A.Y. Coran, “Vulcanization. V. The Formation of Cross-links in the System Natural Rubber-sulfur-MBT-zinc Ion”, Rubber Chem. Technol., 37, 679-688, 1964. 84 A.Y. Coran, “Vulcanization. VI. A Model and Treatment for Scorch Delay Kinetics”, Rubber Chem. Technol., 37, 689-697, 1964. 85 A.M. Zaper, J.L. Koenig, “Solid-state Carbon-13 NMR Studies of Vulcanized Elastomers. II. Sulfur Vulcanization of Natural Rubber”, Rubber Chem. Technol., 60, 252-277, 1987. 86 A.M. Zaper, J. L. Koenig, “Solid-state Carbon-13 NMR Studies of Vulcanized Elastomers. III. Accelerated Sulfur Vulcanization of Natural Rubber”, Rubber Chem. Technol., 60, 278-297, 1987. 87 H. Krebs, “The Mechanism of Accelerator Action”, Rubber Chem. Technol., 30, 962-971, 1957. 260 88 R.H. Campbell, R.W. Wise, “Vulcanization. I. Fate of Curing System During the Sulfur Vulcanization of Natural Rubber Accelerated by Benzothiazole Derivatives”, Rubber Chem. Technol., 37, 635-649, 1964. 89 P.J. Flory, “Principles of Polymer Chemistry”, Cornell University Press, NY, 1953. 90 A.Y. Coran, “Vulcanization”, in “Science and Technology of Rubber”, J. E. Mark, B. Erman and F. R. Eirich, ed., Academic Press, NY, 1994. 91 N.J. Morrison, “The Reactions of Crosslink Precursors in Natural Rubber”, Rubber Chem. Technol., 57, 86-96, 1984. 92 B. Saville, A.A. Watson, “Structural Characterization of Sulfur-vulcanized Rubber Networks”, Rubber Chem. Technol., 40, 100-148, 1967. 93 A. Dogadkin, “Vulcanizate Structures and Their Changes during Vulcanization, Hot Stress Relaxation, and Fatigue”, Kauts. und Gummi, 12, 5-10, 1959. 94 A. Dogadkin, V. Shershnev, “Vulcanization of Rubber in the Presence of Organic Accelerators”, Rubber Chem. Technol., 35, 1-56, 1962. 95 B. Adhikari, D. De, S. Maiti, “Reclamation of Waste Rubber,” Prog. Polym. Sci., 25, 909-948, 2000. 96 http://www.epa.gov/garbage/tires/basic.htm 97 J. Paul, “Rubber Reclaiming,” Encyclopedia of Polym. Sci. & Eng., 14, 787-804, 1988. 98 A.A. Phadke, S.K. De, “Vulcanization of Cryo-ground Reclamied Rubber,” Kautsc. Gummi Kusntst., 37, 776-779, 1994. 99 H.F. Mark, N.M. Bikales, C. G. Overberger, G. Menges, ed., “Encyclopedia of Polymer Science and Engineering”, Vol. 14, 2nd ed., Wiley, New York, p. 787-804, 1985. 100 S. Lee, F.O. Azzam, B.S. Kocher, US 5516952, “Oxidative Decoupling of Scrap Rubber for Fuels or Feedstocks without Environmental Pollution”, 1996. 101 L. Adkins, US 5618852, “Recovery of Products from Recycling Used Tires Using Bath Mixture Containing Soybean Oil,” 1997. 261 102 W. Kaminsky, “Possibilities and Limits of Pyrolysis”, Makromolekulare Chemie, Macromolecular Symposia, 57, 145-60, 1992. 103 A.V. Chapman, M. Porter, “Sulfur Vulcanization Chemistry”, in “Natural Rubber Science & Technology”, A.D. Roberts, ed., Oxford University Press, New York, 1988. 104 B. Saville, A.A. Watson, “Structural Characterization of Sulfur-vulcanized Rubber Networks”, Rubber Chem. Technol., 40, 100-148, 1967. 105 D.S. Campbell, “Structural Characterization of Vulcanizates, Part X, Thiol-disulfide Interchange for Cleaving Disulfide Crosslinks in Natural Rubber Vulcanizates,” Rubber Chem. Technol., 43, 210-221, 1970. 106 M.L. Selker, A.R. Kemp, “S Linkage in Vulcanized Rubbers. I. Reaction of MeI with S Compounds”, Journal of Industrial and Engineering Chemistry, 36, 16-20, 1944. 107 M.L. Selker, A.R. Kemp, “S Linkage in Vulcanized Rubbers. II. The Reaction of MeI with Vulcanizates”, Journal of Industrial and Engineering Chemistry, 36, 20-28, 1944. 108 M.L. Selker, “Sulfur Linkage in Vulcanized Rubber. Reaction of Methyl Iodide with Sulfur Compounds”, Journal of Industrial and Engineering Chemistry, 40, 1467- 1470, 1948. 109 O. Holst, B. Stenberg, M. Christiansson, “Biotechnological Possibilities for Waste Tyre-Rubber Treatment”, Biodegradation, 9, 301–310, 1998. 110 A. Tsuchii, T. Suzuki, K. Takeda, “Microbial Degradation of Natural Rubber Vulcanizates”, Appl. Environ Microbiol., 50, 965-970, 1985. 111 A. Tsuchii, K. Takeda, “Rubber-Degrading Enzyme from a Bacterial Culture”, Appl. Environ Microbiol., 56, 269-274, 1990. 112 S. Sato, Y. Honda, M. Kuwahara, H. Kishimoto, N. Yagi, K. Muraoka, “Microbial Scission of Sulfide Linkages in Vulcanized Natural Rubber by a White Rot Basidiomycete, Ceriporiopsis Subvermispora”, Biomacromolecules, 5, 511-515, 2004. 113 K. Rose, A. Steinbuechel, “Biodegradation of Natural Rubber and Related Compounds: Recent Insights into a Hardly Understood Catabolic Capability of Microorganisms”, Applied and Environmental Microbiology, 71, 2803-2812, 2005. 262 114 M. Myhre, D.A. Mackillop, “Rubber Recycling”, Rubber Chem. Technol., 75, 429- 474, 2002. 115 A.H. Pelofsky, US 3725314, “Rubber Reclamation Using Ultrasonic Energy”, 1973. 116 M. Okuda, Y. Hatano, JP 62121741, “Devulcanization Method of Rubber Using Ultrasound”, 1987. 117 D. Mangaraj, N. Senapati, US 4548771, “Ultrasonic Vulcanization”, 1946. 118 A. Tukachinsky, D. Schworm, A. I. Isayev, “Devulcanization of Waste Tire Rubber by Powerful Ultrasound”, Rubber Chem. Technol., 69, 92-103, 1996. 119 A.I. Isayev, S.P. Yushanov, J. Chen, “Ultrasonic Devulcanization of Rubber Vulcanizates. Part I: Process Model”, J. Appl. Polymer Sci., 59, 803-813, 1996. 120 A. I. Isayev, S. P. Yushanov, J. Chen, “Ultrasonic Devulcanization of Rubber Vulcanizates. Part II: Simulation and Experiment,” J. Appl. Polymer Sci., 59, 815- 824, 1996. 121 J. Yun, J.S. Oh, A.I. Isayev, “Ultrasonic Devulcanization Reactors for Recycling of GRT: Comparative Study,” Rubber Chem. Technol., 74, 317-330, 2001. 122 V.Y. Levin, S.H. Kim, A. I. Isayev, J. Massey, E. von Meerwall, “Ultrasound Devulcanization of Sulfur Vulcanized SBR: Crosslink Density and Molecular Mobility”, Rubber Chem. Technol., 69, 104-114, 1996. 123 M. Tapale, A.I. Isayev, “Continuous Ultrasonic Devulcanization of Unfilled NR Vulcanizates”, J. Appl. Polym. Sci., 70, 2007-2019, 1998. 124 J. Yun, A.I. Isayev, “Superior Mechanical Properties of Ultrasonically Recycled EPDM Rubber”, Rubber Chem. Technol., 76, 253-270, 2003. 125 S. Ghose, A.I. Isayev, “Recycling of Unfilled Polyurethane Rubber Using High Power Ultrasound”, Polym. Eng. Sci., 88, 980-989, 2003. 126 B. Diao, A.I. Isayev, V.Y. Levin, “Basic Study of Continuous Ultrasonic Devulcanization of Unfilled Silicone Rubber,” Rubber Chem. Technol., 72, 152-164, 1999. 127 A.I. Isayev, S. Ghose, “Ultrasonic Devulcanization of Used Tires and Waste Rubbers”, in “Rubber Recycling”, S. K. De, A. I. Isayev, K. Khait edited, Boca Raton, FL.: Taylor & Francis/CRC Press, 2005. 263 128 F. Cataldo, “Thermal Depolymerization and Pyrolysis of cis-1, 4-Polyisoprene: Preparation of Liquid Polyisoprene and Terpene Resin”, Journal of Analytical and Applied Pyrolysis, 44, 121-130, 1998. 129 F. Chen, J. Qian, “Studies on the Thermal Degradation of cis-1, 4-Polyisoprene”, Fuel, 81, 2071-2077, 2002. 130 M. Kojimaa, S. Kohjiya, Y. Ikeda, “Role of Supercritical Carbon Dioxide for Selective Impregnation of Decrosslinking Reagent into Isoprene Rubber Vulcanizate”, Polymer, 46, 2016–2019, 2005. 131 M. Kojimaa, K. Ogawa, H. Mizoshima, M. Tosaka, S. Kohjiya, Y. Ikeda, “Devulcanization of Sufur-Cured Isoprene Rubber in Supercritical Carbon Dioxide”, Rubber Chem. Technol., 76, 957-968, 2003. 132 H.B. Bode, A. Zeeck, K. Pluckhahn, D. Jendrossek, “Physiological and Chemical Investigations into Microbial Degradation of Synthetic Poly(cis-1,4-isoprene)”, Appl. Environmental Microbiology, 66, 3680–3685, 2000. 133 A. Linos, M.M. Berekaa, R. Reichelt, U. Keller, J. Schmitt, H. Flemming, R.M. Kroppenstedt, A. Steinbuchel, “Biodegradation of cis-1,4-Polyisoprene Rubbers by Distinct Actinomycetes: Microbial Strategies and Detailed Surface Analysis”, Appl. Environmental Microbiology, 66, 1639–1645, 2000. 134 T.J. Mason, “A General Introduction to Sonochemistry”, in “Sonochemistry: the Use of Ultrasound in Chemistry”, T.J. Mason, ed., Cambridge, [England]: Royal Society of Chemistry, 1990. 135 T.J. Mason, J.P. Lorimer, “Applied Sonochemistry: The Uses of Power Ultrasound in Chemistry and Processing”, Wiley-VCH, Weinheim, 2002. 136 P. Riesz, T. Kondo, “Free Radical Formation Induced by Ultrasound and Its Biological Implications”, Free Radical Biology & Medicine, 13, 247-270, 1992. 137 T.J. Mason, “Ultrasonics”, in “Sonochemistry: Theory, Applications and Uses of Ultrasound in Chemistry”, T.J. Mason, J.P. Lorimer, ed., Chichester: Ellis Horwood ; New York, NY: Wiley, 1988. 138 T.J. Mason, “Introduction”, in “Chemistry with Ultrasound”, T.J.Mason, (editor), Elsevier Applied Science, London & NY, 1990. 139 S.J. Doktycz, K.S. Suslick, “Interparticle Collisions Driven by Ultrasound”, Science, 247, 1067-1069, 1990. 264 140 T.J. Mason, “An Introduction to the Uses of Power Ultrasound in Chemistry”, in “Sonochemistry”, Oxford University Press, 1999. 141 T.J. Mason, ed., “Advances in Sonochemistry”, Volume 5; Oxford University Press: New York, 1999. 142 A.N. Gent, D.A. Tompkins, “Nucleation and Growth of Gas Bubbles in Elastomers”, J. Appl. Phy., 40, 2520-2525, 1969. 143 B.E. Noltingk, E.A. Neppiras, “Cavitation Produced by Ultrasonics: Theoretical Conditions for the Onset of Cavitation;” Proc. Phys. Soc. Ser. B, 64, 1032-8, 1951. 144 A. Henglein, “Sonochemistry: Historical Developments and Modern Aspects”, Ultrasonics, 25, 6-16, 1987. 145 D. Ensminger, “Ultrasonics: Fundamentals, Technology, Applications”, 2nd, rev. and expand ed.; M. Dekker: New York, 1988. 146 D.W. Ovenall, G. W. Hastings, P. E. M. Allen, “Degradation of Polymer Molecules in Solution under Influence of Ultrasonic Waves. I. Kinetic Analysis”, J. Polym. Sci., 33, 207-212, 1958. 147 D.W. Ovenall, G.W. Hastings, P. E. M. Allen, “Degradation of Polymer Molecules in Solution under Influence of Ultrasonic Waves. II. Rate Equation and the Limiting Degree of Polymerization”, J. Polym. Sci., 33, 213-225, 1958. 148 J.R. Thomas, “Sonic Degradation of High Polymers in Solution”, J. Phy. Chem., 63, 1725-1729, 1959. 149 K.S. Suslick, G. J. Price, “Applications of Ultrasound to Materials Chemistry”, Annu. Rev. Mater. Sci., 29, 295-326, 1999. 150 G.J. Price, P.F. Smith, “Ultrasonic Degradation of Polymer Solutions .3. The Effect of Changing Solvent and Solution Concentration”, Eur. Polym. J., 29, 419-424, 1993. 151 B.M.E. Van der Hoff, C.E. Gall, “A Method for Following Changes in Molecular Weight Distributions of Polymers on Degradation: Development and Comparison with Ultrasonic Degradation Experiments”, J. Macromol. Sci., Chem., A11, 9, 1739- 1758, 1977. 152 B.M.E. Van der Hoff, P.A.R. Glynn, “Rate of Degradation by Ultrasonication of Polystyrene in Solution”, J. Macromol. Sci., Chem., A8, 2, 429-449, 1974. 265 153 M. Tabata, T. Miyazawa, O. Kobayashi, J. Sohma, “Direct Evidence of Main-Chain Scissions Induced by Ultrasonic Irradiation of Benzene Solutions of Polymers”, Chem. Phys. Lett., 73, 178-180, 1980. 154 M. Tabata, J. Sohma, “Spin Trapping studies of Poly (methyl methacrylate) Degradation in Solution”, Eur. Polym. J., 16, 589-595, 1980. 155 K. Gaddam, H. M. Cheung, “Ultrasound-assisted Emulsion Polymerization of Methyl Methacrylate and Styrene”, J. Appl. Polym. Sci., 76, 101-104, 2000. 156 C.K. Hong, A.I. Isayev, “Ultrasonic Devulcanization of Unfilled SBR under Static and Continuous Conditions”, Rubber Chem. Technol., 75, 133-142, 2002. 157 S.P. Yushanov, A.I., Isayev and S.H. Kim, “Ultrasonic Devulcanization of SBR Rubber: Experimentation and Modeling Based on Cavitation and Percolation Theories”, Rubber Chem. Technol., 71, 168-190, 1998. 158 S.H. Kim, “Continuous Ultrasonic Devulcanization of Sulfur Cured SBR Vulcanizates”; PhD Dissertation, The University of Akron, OH, 1998. 159 A.I. Isayev, S.P. Yushanov, D. Schworm, A. Tukachinsky, “Modeling of Ultrasonic Devulcanization of Tire Rubbers and Comparison with experiments”, Plast. Rubber and Compos. Process. Appl., 25, 1-12, 1996. 160 A.N. Gent, A.C. Thomas, “Forms for the Stored Energy Function for Vulcanized Rubber”, J. Polym. Sci., 28, 625-628, 1958. 161 A.I. Isayev, S.P. Yushanov, S.H. Kim, L.V. Yu., “Ultrasonic Devulcanization of Waste Rubbers: Experimentation and Modeling”, Rheol. Acta, 35, 616-630, 1996. 162 V.V. Yashin, A.I., Isayev, “A Model for Rubber Degradation under Ultrasonic Treatment: Part I. Acoustic Cavitation in Viscoelastic Solid”, Rubber Chem. Technol., 72, 741-757, 1999. 163 V.V. Yashin, A.I. Isayev, “A Model for Rubber Degradation under Ultrasonic Treatment: Part II. Rupture of Rubber Network and Comparison with Experiments”, Rubber Chem. Technol., 73, 325-339, 2000. 164 M. Gordon, “Good's Theory of Cascade Processes Applied to the Statistics of Polymer Distributions”, Proc. Roy. Soc., A268, 240-259, 1962. 165 G.R. Dobson, M. Gordon, “Theory of Branching Processes and Statistics of Rubber Elasticity”, J. Chem. Phy., 43, 705-713, 1965. 266 166 V.V. Yashin, A.I. Isayev, S.H. Kim, C.K. Hong, S.E. Shim, J. Yun, “Degradation of Rubber Network During the Ultrasonic Treatment”, SPE ANTEC Technical Papers, 49, 2485-2489, 2003. 167 L. Reich, S.S. Stivala, “Elements of Polymer Degradation”, McGraw-Hill, New York, 1971. 168 S.L. Madorsky, “Thermal Degradation of Organic Polymers”, Interscience Publishers, New York, 1964. 169 H.H.G. Jellinek, “Depolymerization”, in “Encyclopedia of Polymer Science and Technology: Plastics, Resins, Rubbers, Fibers”, H.F. Mark, N.G. Gaylord, N.M. Bikales eds., Interscience Publishers, New York, 4, 740, 1964. 170 V.V. Yashin, C.K. Hong, A.I. Isayev, “Thermomechanical Degradation of SBR during Ultrasonic Treatment under Static Condition”, Rubber Chem. Technol., 77, 50-77, 2004. 171 K. Schmidt-Rohr, H.W. Spiess, “Multidimensional Solid-State NMR and Polymers”, Academic Press: London; San Diego, 1994. 172 C.A. Fyfe, “Solid-state NMR for Chemists”, C. F. C. Press: Guelph, Ontario, 1983. 173 J.L. Massey, “NMR Investigation of Ultrasound Devulcanized Natural Rubber and Styrene-Butadiene Rubber”, PhD Dissertation, University of Akron, 2004. 174 E. von Meerwall, J.L. Massey, S. T. Johnston, S. H. Kim, V. Y. Levin, A. I. Isayev, “NMR Relaxation and Diffusion Study of Ultrasound Devulcanization of SBR Networks”, Polymer Preprints (American Chemical Society, Division of Polymer Chemistry), 38, 861-862, 1997. 175 V.J. McBrierty, K.J. Packer, “Nuclear Magnetic Resonance in Solid Polymers”, Cambridge University Press, Cambridge [England]; New York, NY, USA, 1993. 176 F.A. Bovey, P.A. Mirau, “NMR of Polymers”, San Diego: Academic Press, 1996. 177 J. Crank, “The Mathematics of Diffusion”, 2nd ed.; Oxford: Clarendon Press, 1979. 178 M. Doi, S.F. Edwards, “The Theory of Polymer Dynamics”, Oxford: Clarendon Press, New York: Oxford University Press, 1986. 179 E.L. Hahn, “Nuclear Induction Due to Free Larmor Precession”, Physical Review, 77, 297-298, 1950. 267 180 E.O. Stejskal, J.E. Tanner, “Spin Diffusion Measurements: Spin Echoes in the Presence of a Time-dependent Field Gradient”, J. Chem Phys., 42, 288–292, 1965. 181 M. Tapale, “Continuous Ultrasonic Devulcanization of Unfilled Natural Rubber Vulcanizates”, Master’s Thesis, University of Akron, 1998. 182 C.K. Hong, “The Application of Powerful Ultrasound to Devulcanization of Rubbers and Compatibilization of Polymer Blends”, PhD Dissertation, University of Akron, 2001. 183 G. Bielstein, “Some Aspects of High Temperature Vulcanization”, Rubber Chem. Technol., 34, 319-333, 1961. 184 U. Shankar, “Investigations of the Reversion of Vulcanized Rubber under Heat”, Rubber Chem. Technol., 25, 241-250, 1952. 185 H. Pawlowski, X. Xu, “The Measurement of Thermoset Cure Properties with a Novel Rheometer”, Technical Paper of Alpha Technologies (www.alpha- technologies.com). 186 M. Mooney, “A Theory of Large Elastic Deformation”, Journal of Applied Physiology, 11, 582–592, 1940. 187 R.S. Rivlin, “Large Elastic Deformations of Isotropic Materials”, I. II. III., Trans. R. Soc., A240: 459–490, 491–508, 509–525, 1948. 188 P.J. Flory, J. Jr., Rehner, “Statistical Mechanics of Cross-linked Polymer Networks”, J. Chem. Phys., 11, 512-520, 1943. 189 A. Barton, “CRC Handbook of Polymer-Solvent Interaction Parameters and Solubility Parameters”, Chapter 32, CRC Press, Boston, 1990. 190 G.R. Hamed, “Energy Dissipation and the Fracture of Rubber Vulcanizates”; Rubber Chem. Technol., 64, 493-500, 1991. 191 E. Von Meerwall, P. Palunas, “The Effects of Polydispersity, Entanglements and Crosslinks on Pulsed-gradient NMR Diffusion Experiments in Polymer Melts”, J. Polym. Sci.: Part B: Polym. Phys., 25, 1439-1457, 1987. 192 E. von Meerwall, R.D. Ferguson, “Pulsed-field Gradient NMR Measurements of Diffusion of Oil in Rubber”, J. Appl. Polym. Sci., 23, 877-885, 1979. 268 193 P.E. Willard, “Determination of Cure of Diallyl phthalate Using Differential Scanning Calorimetry”, Polym. Eng. Sci., 12, 120-124, 1972. 194 M.R. Kamal, M.E. Ryan, “The Behavior of Thermalsetting Compounds in Injection Molding Cavities”, Polym. Eng. Sci., 20, 859-867, 1980. 195 G.O. Piloyan, I.D. Ryabchikov, O.S. Novikova, “Determination of activation energies of chemical reactions by differential thermal analysis”, Nature, 212, 1229, 1966. 196 A.I. Isayev, J.S. Deng, “Non-isothermal Vulcanization of Rubber Compounds”, Rubber Chem. Technol., 61, 340-361, 1988. 197 J.S. Deng, A.I. Isayev, “Injection Molding of Rubber Compounds: Experimentation and Simulation”, Rubber Chem. Technol., 64, 296-324, 1991. 198 M.R. Kamal, S. Sourour, “Kinetics and Thermal Characterization of Thermoset Cure”, Polym. Eng. Sci., 13, 59-64, 1973. 199 I. Han, C. Chung, J. Lee, “Optimal Curing of Rubber Compounds with Reversion Type Cure Behavior”, Rubber Chem. Technol., 73, 101-113, 2000. 200 G. Rimondi, W. J. Toth, J. Kounavis, “Predicative Model for Reversion-Type Cures”, Tire Sci. Technol., 24, 77-91, 1996. 201 A.I. Isayev, B. Sujan, “Non-isothermal Vulcanization of Devulcanized GRT with Reversion Type Behavior”, J. Elastomers and Plastics, 38, 291-318, 2006. 202 B. Sujan, “Injection and Compression Molding of Devulcanized GRT: Cure Kinetics and Physical Properties”, Master Thesis, University of Akron, 2005. 203 M. Arroyo, M.A. Lo´pez-Manchado, B. Herrero, “Organo-montmorillonite as Substitute of Carbon Black in Natural Rubber Compounds”, Polymer, 44, 2447-2453, 2003. 204 H. M. da Costa, L. L. Y. Visconte, R. C. R. Nunes, C. R. G. Furtado, “Rice Husk Ash Filled Natural Rubber. I. Overall Rate Constant Determination for the Vulcanization Process from Rheometric Data”, J. Appl. Polym. Sci., 87, 1194-1203, 2003. 205 A.S. Aprem, S. Thomas, K. Joseph, N.M. Barkoula, J.K. Kocsis, “Sulphur Vulcanisation of Styrene Butadiene Rubber using New Binary Accelerator Systems”, J. Elastomers and Plastics, 35, 29-55, 2003. 269 206 C. Lewis, R. Buanpa, S. Kiatkamjornwong, “Effect of Rubber Ratio, Carbon Black Level, and Accelerator Level on Natural Rubber/Bromobutyl Rubber Blend Properties”, J. Appl. Polym. Sci., 90, 3059-3068, 2003. 207 S. Park, M. Seo, C. Nah, “Influence of Surface Characteristics of Carbon Blacks on Cure and Mechanical Behaviors of Rubber Matrix Compoundings”, Journal of Colloid and Interface Science, 291, 229-235, 2005. 208 H.J. Stern, “Rubber: Natural and Synthetic”, Maclaren and Sons: London, 1967; p. 89. 209 A.K. Bhowmick, S.K. De, “Kinetics of Crosslinking and Network Changes in Natural Rubber Vulcanizates with a Dithiodimorpholine Based Accelerator System”, Rubber Chem Technol, 53, 1015-1022, 1980. 210 P.K. Pal, A.K. Bhowmick, S.K. De, “The Effects of Carbon Black-Vulcanization System Interactions on Natural Rubber Network Structures and Properties”, Rubber Chem Technol, 55, 23-40, 1982. 211 A. Szalay, “Destruction of Highly Polymerized Molecules by Ultrasonic Waves”, Z. physik. Chem., A164, 234-240, 1933. 212 G. Schmid, O.Z. Rommel,“The Disintegration of Macromolecules by Ultrasonic Waves.”, Zeitschrift fuer Elektrochemie und Angewandte Physikalische Chemie, 45, 659-661, 1939. 213 C. Sarswati, P. Roy-Chowdhury, A.B. Biswas, “Ultrasonic Degradation of Macromolecules in Solution. Study of Degradation Kinetics of Butyl Rubber by Estimation of Free-radical Scavenger DPPH (a,a-diphenyl-b-picrylhydrazyl) and Solution Viscosity Measurements”, J Appl Polym Sci, 10, 1089-1098, 1966. 214 M.A. Bradley, S.W. Prescott, H.A.S. Schoonbrood, K. Landfester, F. Grieser, “Miniemulsion Copolymerization of Methyl Methacrylate and Butyl Acrylate by Ultrasonic Initiation”, Macromolecules, 38, 6346-6351, 2005. 215 A.I. Isayev, C.K. Hong, “Novel Ultrasonic Process for In-situ Copolymer Formation and Compatibilization of Immiscible Polymers”, Polym Eng & Sci, 43, 91-101, 2003. 216 C.K. Hong, A.I. Isayev, “Continuous Ultrasonic Devulcanization of Carbon Black- Filled NR Vulcanizates”, J Appl Polym Sci, 79, 2340-2348, 2001. 217 S.E. Shim, A.I. Isayev, “Ultrasonic Devulcanization of Precipitated Silica-filled Silicone Rubber”, Rubber Chem Technol, 74, 303-316, 2001. 270 218 V.Y. Levin, S.H. Kim, A.I. Isayev, “Vulcanization of Ultrasonically Devulcanized SBR Elastomers”, Rubber Chem Technol, 70, 120-128, 1997. 219 J. Yun, A.I. Isayev, “The Structure and Property Control of EPDM Rubber Using High Power Ultrasound”, Soc Plast Eng Annu Tech Conf, 48, 3212-3216, 2002. 220 J. S. Oh, S. Ghose, A.I. Isayev, “Effects of Ultrasonic Treatment on Unfilled Butadiene Rubber”, J Polym Sci Part B: Polym Phys, 41, 2959-2968, 2003. 221 W. Feng, A.I. Isayev, “High-power Ultrasonic Treatment of Butyl Rubber Gum: Structure and Properties”, J Polym Sci Part B: Polym Phys, 43, 334-344, 2005. 222 F. Dimier, B. Vergnes, M. Vincent, “Relationships between Mastication Conditions and Rheological Behavior of a Natural Rubber”, Rheol Acta, 43, 196-202, 2004. 223 N. Sombatsompop, C. Kumnuantip, “Rheology, Cure Characteristics, Physical and Mechanical Properties of Tire Tread Reclaimed Rubber/Natural Rubber Compounds”, J Appl Polym Sci., 87, 1723-1731, 2003. 224 M.M. Cross, “Relation between Viscoelasticity and Shear-thinning Behavior in Liquids”, Rheol Acta, 18, 609-614, 1979. 225 G.V. Vinogradov, A.Y. Malkin, “Rheology of Polymers: Viscoelasticity and Flow of Polymers”, Mir: Moscow, 1980. 226 A.I. Isayev, “Rubber Recycling”, in “Rubber Technologist’s Handbook”, S.K. De, J.R. White, ed., Shawbury, Shrewsbury, England: Rapra Technology Ltd., 2001. 227 W.C. Warner, “Methods of Devulcanization”, Rubber Chem. Technol., 67, 559-566, 1994. 228 J. S. Oh, A. I. Isayev, “Continuous Ultrasonic Devulcanization of Unfilled Butadiene Rubber”, J. Appl. Polym. Sci., 93, 1166-1174, 2004. 229 W. Feng, A. I. Isayev, “Continuous Ultrasonic Devulcanization of Unfilled Butyl Rubber”, J. Appl. Polym. Sci., 94, 1316-1325, 2004. 230 V.M. Makarov, V.F. Drozdovski, “Reprocessing of Tires and Rubber Wastes”; Ellis Horwood, New York, Chapter 2, 1991. 231 J. Yun, A. I. Isayev, S. H. Kim, M. Tapale, “Comparative Analysis of Ultrasonically Devulcanized Unfilled SBR, NR, and EPDM Rubbers”, J. Appl. Polym. Sci., 88, 434-441, 2003. 271 232 M. Mori, J.L. Koenig, “Solid-state 13C-NMR and Equilibrium-swelling Studies of Filled, TBBS-accelerated Sulfur Vulcanization of Natural Rubber”, J. Appl. Polym. Sci., 70, 1391-1399, 1998. 233 W.W. Barbin, M.B. Rodgers, “The Science of Rubber Compounding”, in “Science and Technology of Rubber”, edited by E.M. James, E. Burak and R.E. Frederick, 2nd ed., Academic Press, Inc., 1994. 234 http://www.akrochem.com/pdfs/plasticizer/X-plasticizer%20ln.pdf 235 http://www.akrochem.com/pdfs/retarders/x-retarder%20safe.pdf http://www.akrochem.com/pdfs/retarders/x-retarder%20rsf.pdf 236 S. Ghose, “Recycling of Polyurethane Rubber and Foam by the Application of High Power Ultrasound”, PhD Dissertation, University of Akron, 2004. 237 C.K. Hong, A.I. Isayev, “Blends of Ultrasonically Devulcanized and Virgin Carbon Black Filled NR”, J. Mat. Sci., 37, 385-388, 2002. 238 J. Yun, A.I. Isayev, “Blends of Ultrasonically Devulcanized and Virgin Carbon Black Filled Ethylene–propylene–diene Monomer Rubbers”, J. Appl. Polym. Sci., 92, 132-138, 2004. 239 D. Gibala, G.R. Hamed, “Cure and Mechanical Behavior of Rubber Compounds Containing Ground Vulcanizates. Part I. Cure Behavior”, Rubber Chem. Technol., 67, 636-648, 1994. 240 S.T. Johnston, J. Massey, E. von Meerwall, S.H. Kim, V. Yu. Levin, A.I. Isayev, “Ultrasound Devulcanization of SBR: Molecular Mobility of Gel and Sol”, Rubber Chem. Technol., 70, 183-193, 1997. 241 S.E. Shim, J.C. Parr, E. von Meerwall, A.I. Isayev “NMR Relaxation and Pulsed Gradient NMR Diffusion Measurements of Ultrasonically Devulcanized Poly(dimethylsiloxane)”, J. Phys. Chem. B, 106, 12072-12078, 2002. 242 S.E. Shim, A.I. Isayev, E. von Meerwall, “Molecular Mobility of Ultrasonically Devulcanized Silica-filled Poly (Dimethyl Siloxane)”, J. Polym. Sci., part B: Polym. Phys. Ed., 41, 454-465, 2003. 243 S. Ghose, A. I. Isayev, E. von Meerwall, “Effect of Ultrasound on Thermoset Polyurethane: NMR Relaxation and Diffusion Measurements”, Polymer, 45, 3709– 3720, 2004. 272 244 W. Feng, A. I. Isayev, E. von Meerwall, “Molecular Mobility in Ultrasonically Treated Butyl Gum and Devulcanized Butyl Rubber”, Polymer, 45, 8459–8467, 2004. 245 J.S. Oh, A.I. Isayev, T. Wagler, P.L. Rinaldi, E. von Meerwall, “Molecular Mobility and Structure of Ultrasonically Treated Unfilled Butadiene Rubber”, J. Polym. Sci., part B: Polym Phy., 42, 1875–1887, 2004. 246 S. Meiboom, D. Gill, “Modified Spin-echo Method for Measuring Nuclear Relaxation Times”, Rev, Sci. Instr., 29, 688-691, 1958. 247 E. von Meerwall, D. Thompson, “A Fortran Program for Reduction of NMR Relaxation Data”, Computer Physics Communications, 31, 385-392, 1984. 248 S.P. Yushanov, A.I. Isayev, V.Y. Levin, “Percolation Simulation of the Network Degradation during Ultrasonic Devulcanization”, J. Polym. Sci., part B: Polym. Phys. Ed., 34, 2409-2418, 1996. 249 P.G. De Gennes, “Critical Behavior for Vulcanization Processes”, J. Phys. (Paris), Lett., 38L, 355-358, 1977. 250 A.I. Isayev, S.H. Kim, W. Feng, “Continuous Ultrasonic Treatment of Uncured and Sulfur Cured SBR: Effect of Styrene Content”, Rubber Chem & Technol., 78, 606- 619, 2005. 251 T. Kelen, “Polymer Degradation”, Van Nostrand Reinhold Co., New York, Chapter 4, 1983. 252 A. I. Isayev, C. K. Hong, K. J. Kim, “Continuous Mixing and Compounding of Polymer/Filler and Polymer/Polymer Mixtures with the Aid of Ultrasound”, Rubber Chem. Technol., 76, 923-947, 2003. 253 K.S. Loganathan, ed., “Atomic Structure and Bonding”, in “Rubber Engineering”, Indian Rubber Institute, McGraw, NY, 2000. 254 E.M. Dannenberg, “Bound Rubber and Carbon Black Reinforcement”, Rubber Chem Technol, 59, 512-524, 1986. 273