Professional Documents
Culture Documents
A Dissertation
Presented to
In Partial Fulfillment
Doctor of Philosophy
Ximei Sun
May, 2007
THE DEVULCANIZATION OF UNFILLED AND CARBON BLACK FILLED
Ximei Sun
Dissertation
Approved: Accepted:
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.
fraction and crosslink density. Increasing the ultrasonic amplitude yielded a further
previous work on natural rubber (NR), the natural counterpart of IR which showed a
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
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
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
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
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
CHAPTER
I. INTRODUCTION..................................................................................................... 1
vi
2.7.6 Biotechnological Method .............................................................................. 30
3.1 Materials................................................................................................................ 50
3.2 Compounding........................................................................................................ 52
3.4 Revulcanization..................................................................................................... 55
vii
3.6.5 Mechanical Properties ................................................................................... 67
IV. CURE KINETICS STUDY OF THE UNFILLED AND THE CB FILLED IR..... 71
7.3 Vulcanization of the Virgin Filled IR without the Processing Oil...................... 155
ix
7.3.1 Curing.......................................................................................................... 155
7.4 Devulcanization of the Filled IR without the Processing Oil ............................. 163
7.7 Blending of the Devulcanized IR (dIR) with the Virgin IR................................ 191
x
8.4 Molecular Mobility of Treated Gum IR and their Vulcanizates ......................... 205
REFERENCES................................................................................................................ 254
xi
LIST OF TABLES
Table Page
2.4 Sulfur and accelerator of different vulcanization systems for NR82 ............. 20
4.1 Induction time and cure kinetic constants obtained from individual
isothermal fitting of the unfilled IR............................................................... 86
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
4.6 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
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
xiii
LIST OF FIGURES
Figure Page
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
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
xiv
4.4 Determination of the experimental induction time ....................................... 82
4.5 Finding the Γmax,0 − Γmin,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
4.13 Finding Γmax,0 − Γmin,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
5.5 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
6.3 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
6.7 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
6.10 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
7.2 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
7.7 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
7.12 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
7.15 Stress-strain curves of the virgin cured and the revulcanized 35 phr CB
filled IR obtained at three ultrasonic amplitudes ........................................ 173
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
7.21 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
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
7.27 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
8.3 Proton T2 for the gum IRs ultrasonically treated at different amplitudes ... 209
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
8.13 Transverse 1H relaxation decay at 70.5oC for the virgin IR fitted with the
two-component model containing the Weibull exponent ........................... 219
8.15 Proton T2 for the virgin, the cured and the devulcanized IR as a function
of extracted sol ............................................................................................ 221
8.18 Transverse 1H relaxation decay for the 35 phr CB filled gum IR at 70.5oC
fitted with the two-component model ......................................................... 224
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
xxii
CHAPTER I
1.INTRODUCTION
Natural rubber (NR) is widely used in the tire industry and is produced naturally
behavior has made it unique among all the other synthetic elastomers as far as mechanical
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
processing (mixing, extrusion, molding and calendering), purity and particularly that it
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
Rubber recovery is not an easy business since we are dealing with three-
chemical process. It would be reasonable to think that breaking this network inevitably
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
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
without the involvement of any chemicals and the rubbers treated by ultrasound are soft,
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
unfilled and carbon black filled synthetic IR vulcanizates; 3) understand the mechanism
3
CHAPTER II
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
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
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.
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
Figure 2.1 Isoprene monomer, Natural rubber (cis-1, 4-addition) and Gutta-
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
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.
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
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
environmental damage due to the unsaturated unit in its macromolecular chain. The
atmospheric oxygen (especially at high temperatures) and ozone easily attack the
6
2.2 Synthetic Isoprene Rubber (IR)
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
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
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
natural rubber. In this way almost 100 percent cis-1, 4 polyisoprene, “synthetic natural
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
tetrachloride, TiCl4)33. Soon afterwards, Firestone Tire & Rubber Corporation revealed a
synthesis of cis-1, 4-polyisoprene with a catalyst based on lithium metal and yielded a
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
catalyzed IR (Ti-IR) with a cis-1, 4 content of 98.5%4, finally allowing the benefits of
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.
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
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
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.
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.
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
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.
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
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
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
commercial natural rubber products has been detrimental for its use in certain medical
channel and acetylene blacks44. It is essentially elemental carbon in the form of fine
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
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).
Until 1968, carbon black nomenclature was informal and inconsistent. It was just
size, and electrical conductivity. In 1968, the ASTM committee on carbon black
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 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
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
2) The size, shape and distribution of aggregates. These parameters are generally
3) Surface activity. It is related to the reactivity of the chemical groups on the carbon
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
13
2.5 Carbon Black Filled Rubber
(stiffness) and the improvement of fracture properties such as tensile, tear and abrasion
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
World War II because its strength as a pure gum vulcanizate is so low that SBR needed
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
14
decreasing particle size and increasing structure and carbon black loading. Also, carbon
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
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
black64. 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
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
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
deformation after removal of the deforming force. In order to develop the full potential of
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
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
16
changes are proportional to the number of crosslinks and their length. Excessive
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
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
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
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.
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.
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
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
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
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
20
Table 2.5 Typical NR vulcanizate structures and properties at optimum cure time82
CV Semi-EV EV
Monosulfidic crosslinks, % 5 50 80
resistance
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
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
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
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
been studied since the 1950s, the exact mechanism of accelerated sulfur vulcanization
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
22
Nevertheless, there is some agreement on the general mechanism. In detail,
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
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
S S
NR1 NR1
S6
NR1 NR1
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-
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
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
terminated polysulfides.
C
C
Sx
S C
R
C
SA
monosulfidic pendant
side group
C C
C SB C C
S
S C
cyclic sulfides
S multifunctional
vicinal
C - C=C-C=C-C=C-C=C -
disulfidic conjugated unsaturation
C C
C
SA SB C=C C=C
Sx
C C
C
cis/trans isomerization
polysulfidic vicinal
The final stage of vulcanization is the over cure period of elastomers. At this
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.
Vulcanization is the reason that gum rubber is useful in the tire industry.
of waste rubbers dumped and stockpiled. Unlike the thermoplastic polymers which can
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
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
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.
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
Ambient grinding, unlike its name, does result in the generation of heat. In this
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
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
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.
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
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
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.
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
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.
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
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.
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
bacteria from the genus Nacardia and the process led to a substantial weight loss 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
(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
processes since it consumes little energy and does not require hazardous chemicals.
organisms, which are affected by the environment they are cultivated in. Chemicals
growth of microorganisms and thus inhibit the biodegradation of rubber materials113. Zinc
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
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
achieve devulcanization. The researchers claimed that the properties attained after the
revulcanization process were similar to that of the original vulcanizate. Mangaraj and
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,
(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,
harmful or unharmful. The process can be operated continuously, which makes this
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
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
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
simultaneously minor main chain scission in the vulcanized rubbers. This process
involves no chemicals. The ultrasonically treated rubber becomes soft and remoldable.
the virgin rubber. Due to these benefits, the ultrasonic method treatment of IR is the
There are very limited ways of recycling synthetic isoprene rubber reported so far.
Currently the available methods include three types: high temperature pyrolysis, chemical
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
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
will be costly.
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.
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
low amplitude waves are used for analytical purpose to measure the velocity and
from such measurements can be used in medical imaging, chemical analysis and the
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
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.
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
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
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
Pressure
- amplitude wavelength
5000oC
1000 atm
Bubble Bubble grows in Reaches Undergoes
forms Successive cycles unstable size Violent collapse
that the final temperature and pressure generated, Tf and Pf, are:
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
equations are a value of 4000 – 6000 K for Tf and 1000 – 2000 bar for Pf 144. It is the
energy generated on the collapse of these bubbles which is the underlying reason for the
and pressure generated during ultrasonic exposure, numerous other beneficial effects are
mechanical rupture145.
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
it occurs at the points of inherent weakness within the polymer backbone. In contrast, 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
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.
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.
mechanism. In this case, the cavitation occurs in a solid body. The occurrence of
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
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
Nevertheless, it has been shown that devulcanization of rubber with high power
energy among all the chemical bonds is not capable of rubber devulcanization119, 159.
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.
A simplified theoretical model is the elastic model119. This model was based on
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
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
ultrasonic devulcanization was based on the cavitation collapse model119 and was in
disagreement with experimental data157, 161. The disagreement was caused by the
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
Standard linear viscoelastic solid and Rouse viscoelastic solid models were used for
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
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
overstressed chains on the concentration of the cavities and high amplitude cavitation
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
also investigated166. 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
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
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
transfer with chain scission (ktr); and (e) 1st 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
NMR is the direct investigation of nuclear energy levels based upon the angular
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,
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
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
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
When observing nuclei of elastomers by solid-state NMR, there are two important
factors that regulate relaxation rates: molecular motion and magnetic interaction between
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)
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
∂c ∂ 2c
= − Dc 2 (2.3)
∂t ∂x
where c is the concentration of the diffusing species and Dc is the diffusion constant178. In
∂c
= − Dc ∇ 2 c (2.4)
∂t
∂ ∂ ∂
where ∇ = i + j +k
∂x ∂y ∂z
By replacing the concentration gradient with the probability P of finding a particle with
∂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 , G0 ) 2
ln = − γ 2 DG02τ 3 (2.6)
A(2τ ,0) 3
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
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
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 2 δ 2
X = δ G ∆ − − δGG 0 (cos θ ) (t1 + t 2 ) + δ (t1 + t 2 ) + 2
2 2 2
− 2τ 2 (2.8)
3 3
48
t1
49
CHAPTER III
3.EXPERIMENTAL
3.1 Materials
• No contamination or proteins
50
• Low water swell
The information for all the materials used including the compounding ingredients
acid, processing oil, retarder and carbon black as well as the solvents benzene and
Company
sulphenamide
Fort Worth, TX
LN low viscosity
Company
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
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
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
denoted as Tmax. A strain amplitude of 4.2% (0.3o) 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
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
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
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
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.
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.
Devulcanized IR 100
ZnO 2.5
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-
ULTRASONIC
FEEDER POWER
SUPPLY WATTMETER
CONVERTER
EXTRUDER
TEMPERATURE
AND P RESSURE
GAUGES DIE HORN BOOSTER
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
57
D1: Die entry diameter, 25.4 mm
d: Die gap
V: Effective volume
Figure 3.2 Schematic drawing of the effective volume of the die opening
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.
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
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
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.
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
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:
2πR 3
S* = G *θ 0 (3.1)
3Φ
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
S*
G* =
[1.47 x10 −4
]
m 3 / rad θ 0
(3.2)
The measured torque is divided by 1.47 x 10-4 m3/rad and by the strain θ (in radians) to
θ0
γo = (3.3)
Φ
For the APA 2000, Φ = 2 x 3.58o where the cone angle = 3.58o. Thus for a movement of
1o
γo = = 0.140 (3.4)
7.16 o
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
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:
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
plots186, 187. In our case, the latter was not viable for the devulcanized samples due to 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
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 − Vr ) + Vr + χVr2
ξ (g) = − (3.7)
V1 (Vr1/ 3 − Vr / 2 )
where ξ(g) 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
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
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
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
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
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
Vro mΦ
= 1− (3.9)
Vr (1 − Φ )
( )
m = 3C 1 − Vro1/ 3 + Vro − 1 (3.10)
65
volume of rubber gel
Vr = (3.11)
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
Crosslink density, mol/m
200
150
100
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
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
F
σ= (3.12)
A0
where F is the instantaneous force and A0 is the initial cross-sectional area of the sample.
l − l0
ε (%) = × 100 (3.13)
l0
where l is the observed distance between the benchmarks on the extended specimen and
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
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
ultrasonically treated and untreated gum isoprene rubber was carried out by GPC (Gel
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
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
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
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
treated gum IR and devulcanized IR were conducted by solid-state NMR at 70.5oC. The
comprehensively described elsewhere191, 192. The pulse sequence employed for the
relaxation experiments was the standard principal Hahn two-pulse sequence conducted
experiments the stimulated spin echo sequence 90o-τ1-90o-τ2-90o-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
70
CHAPTER IV
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
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
dα
= (k1 + k 2α m )(1 − α ) n
dt
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
dα
= kα m (1 − α ) n
dt
dα n 2 −(1+ n ) dα
= α t and = nk 1 / 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
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
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
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.
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
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
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
k1 c1
s
k3
k2 c2 p
Figure 4.1 Sulfur reaction scheme (s: sulfur; c1 and c2: stable and unstable sulfur-
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
74
s c1 c2
sC = , α1 = , α2 =
s0 s0 s0
Han et al199 proposed that the change of kinetic constants and isothermal
E
ki = ki 0 exp − i , i = 1, 2, 3 (4.1)
RT
E
ti = ti 0 exp ti (4.2)
RT
where ki0 and ti0 are the pre-exponent factors; Ei and Eti are the activation energies of the
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 ≥ t i ) are
expressed as:
dsC
= −(k1 + k2 ) sC (4.3)
dt
dα1
= k1sC (4.4)
dt
dα 2
= k2 sC − k3α 2 (4.5)
dt
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
75
The experimental value of α with the reversion type of reaction was determined
as199:
Γt (T ) − Γmin (T )
α (T , t ) = (4.6)
Γmax,0 − Γmin,0
where Γmin (T ) and Γt (T ) are the minimum and current torques at experimental
temperature T, respectively. Γmax,0 − Γmin,0 is the difference of the maximum and the
~
C is the constant. Applying the initial condition t = t i , s = 1 , we have C = (k1 + k2 )ti .
dα1
= k1 exp [ −(k1 + k2 )(t − ti ) ] (4.9)
dt
k1
α1 = − exp [ −(k1 + k2 )(t − ti )] + C (4.10)
k1 + k2
k1
Applying the initial condition: t = t i , α 1 = 0 ,, we have C = . As a result,
k1 + k2
α1 =
k1
k1 + k2
{1 − exp [ −(k1 + k2 )(t − ti )]} (4.11)
76
dα 2
= k2 exp [ −(k1 + k2 )(t − ti )] − k3α 2 (4.12)
dt
(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
= − k3α 2 (4.13)
dt
dα 2 dC
= −k3C (t ) exp [ − k3 (t − ti ) ] + exp [ − k3 (t − ti ) ] , and couple with (4.12) by plugging
dt dt
dα 2 dC
= −k3C (t ) exp [ −k3 (t − ti )] + exp [ −k3 (t − ti )] = k2 exp [ −(k1 + k2 )(t − ti ) ] − k3C (t ) exp [ −k3 (t − ti ) ]
dt dt
dC
= k2 exp [ (k3 − k1 − k2 )(t − ti )] (4.16)
dt
k2
The solution is C (t ) = exp [ (k3 − k1 − k2 )(t − ti ) ] + C ' , after combining it with
k3 − k1 − k2
k2
(4.15) we have α 2 = exp [ (k3 − k1 − k2 )(t − ti )] + C ' exp [ − k3 (t − ti ) ]
k3 − k1 − k2
77
k2
Applying the initial condition: t = ti , α 2 = 0 , it leads to C ' = ,
k1 + k2 − k3
So α 2 =
k2
k3 − k1 − k2
{exp [ −(k1 + k2 )(t − ti )] − exp [ −k3 (t − ti )]}
+
k2
k1 + k2 − k3
{exp [ −k3 (t − ti )] − exp [ −(k1 + k2 )(t − ti )]} for t ≥ ti
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.
follows:
tI
dt
∫ t (T ) = 1
0 i
(4.18)
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
α =0 for t ≤ t I
for t ≥ t I
t τ t t τ
α = ∫ k1 (τ ) exp − ∫ [k1 (θ ) + k2 (θ )]dθ dτ + exp − ∫ k3 (τ )dτ ∫ k2 (τ ) exp − ∫ [ k1 (θ ) + k2 (θ ) − k3 (θ )] dθ dτ
tI tI tI tI tI
78
In the modeling, the correspondent numerical form was applied where numerical double
α =0 for t ≤ t I
for t ≥ t I
t τ t t τ
α = ∑ k1 (τ ) exp −∑ [(k1 (θ ) + k2 (θ )]∆θ ∆τ + exp −∑ k3 (τ )∆τ ⋅ ∑ k2 (τ ) exp −∑ [ (k1 (θ ) + k2 (θ ) − k3 (θ ) ]∆θ ∆τ
tI tI tI tI tI
(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.
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
In this section the modeling of the isothermal cure kinetics was carried out for the
unfilled IR.
Figure 4.3 for the unfilled IR using the recipe in Table 3.2. The strain amplitude was
79
Run isothermal cure to obtain Γ(t)
Obtain k10, E1 , k20, E2 , k30, E3 by simultaneous fitting α(t) with Eqn (4.17)
Comparison of α(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, 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
temperatures is due to not only the reduced viscosity but also to more reversion at
elevated temperatures.
5 140oC
150oC
160oC
170oC 130oC
4
Torque, dNm
180oC
0
0.1 1 10 100
Time, min
Figure 4.3 Isothermal cure curves of the unfilled IR (strain amplitude: 4.2%,
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
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.
Torque
reversion
post cure
induction curing
period
Vulcanization Time
experimental evolution of the state of cure, α, from the available torque data Γ(t). To do
so according to Equation 4.6, one has to find the hypothetical value Γmax,0 − Γmin,0 . This
82
value is experimentally unattainable. However, the value of Γmax − Γmin at each
isothermal condition is available. These data are utilized to extrapolate to some low
temperatures in order to find the Γmax,0 − Γmin,0 value. Therefore, a proper description of
Γmax − Γmin as a function of temperature is desirable. Figure 4.5 shows the decrease of
Γmax − Γmin 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:
(Γmax,0 − Γmin,0 )
[Γmax (T ) − Γmin (T )] = (4.20)
(Γmax,0 − Γmin,0 )T 1− n
1+{ }
τ
4.2
4.0
3.8
(Γmax-Γmin), dNm
3.6
3.4
3.2
3.0 experiment
curve fitting
2.8
400 410 420 430 440 450 460
o
T, K
Figure 4.5 Finding the Γmax,0 − Γmin,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-
square regression technique and it leads to the Γmax,0 − Γmin,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
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.
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
The predictions of isothermal state of cure by the simultaneous modeling are shown as
84
5
k1
4
k2
4
k3
2 ti
-1
ln ti, ln min
3
ln ki, ln min
0
2
-2
1
-4
0
2.1x10-3 2.2x10-3 2.3x10-3 2.4x10-3 2.5x10-3
T-1, K-1
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
R&
R
&
R R&
& R&
R&
$9,9041.:70α
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
Γmax − Γmin 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
investigated.
Table 4.2 Cure kinetic constants obtained from simultaneous isothermal fitting 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
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)
Before testing the applicability of the kinetic model for the non-isothermal cure,
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
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
concept proposed by Isayev and Deng196. The predicted tI is used in the calculation of the
80
o
exp, 80-180 C
70 o
exp, 120-180 C
60 prediction
Induction time, min
40
30
20
10
0
0 2 4 6 8 10
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.
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
1.0
o
exp-1.93 C/min
o
exp-2.13 C/min
0.8 exp-2.52oC/min
exp-3.16oC/min
o
exp-5.81 C/min
State of cure α
0.6
Prediction
0.4
0.2
0.0
0 20 40 60
Time, min
Figure 4.9 Non-isothermal state of cure for the unfilled IR: experiments (starting
90
1.0
4
H[S &PLQ
4
H[S &PLQ
0.8 4
H[S &PLQ
H[S4&PLQ
State of cure α 0.6 SUHGLFWLRQ
0.4
0.2
0.0
0 10 20 30
Time, min
Figure 4.10 Non-isothermal state of cure for the unfilled IR: experiments (starting
91
1.0 200
exp α
0.8 prediction α 180
Temperature
160
Temperature, C
State of cure, α
0.6
o
140
0.4
120
0.2
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
Carbon black is one of the most widely used fillers in the rubber industry. The
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
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
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
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
93
reversion kinetic constants k30 between the unfilled IR (Table 4.2) and the filled IR which
12
120oC 110oC
o o
o 140 C 130 C
150 C
10 160oC
o
170 C
180oC
8
Torque, dNm
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%,
Figure 4.13 shows the Γmax − Γmin 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
dNm for the Γmax,0 − Γmin,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
94
10.0
9.5
(Γmax-Γmin), dNm
9.0
8.5
8.0
7.5 experiment
curve fitting
7.0
360 380 400 420 440 460
T, oK
Figure 4.13 Finding Γmax,0 − Γmin,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
95
4 4
2 2
0 0
-1
ln ti, ln min
ln ki, ln min
-2 -2
-4 -4
-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
Table 4.4 Induction time and cure kinetic constants obtained from individual
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
1.0 o 130oC 120 C
o
150 C140 C 110oC
o
160 C
170oC
0.8
180oC
State of cure α
0.6
0.4
0.2
0.0
Figure 4.15 Isothermal state of cure for the 35phr CB filled IR: experiments (symbols)
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
Table 4.5 Cure kinetic constants obtained from simultaneous isothermal fitting of
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.
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
Step temperature profile (oC) 5 min stay at each T 80 120 160 180
50
o
exp, 80-200 C
exp, 120-200oC
40 prediction
exp, step cure
Induction time, min
20
10
1 2 3 4
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
o
exp-1.80 C/min
o
exp-2.32 C/min
0.8 prediction
0.4
0.2
0.0
0 20 40 60
Time, min
Figure 4.17 Non-isothermal state of cure for the 35phr CB filled IR: experiments
100
1.0
0.8
0.4
0.2
4
0.0 H[S &PLQ
4
H[S &PLQ
0 10 20 30 40
Time, min
Figure 4.18 Non-isothermal state of cure for the 35phr CB filled IR: experiments
101
1.0
180
exp α
0.8 prediction α
Temperature 160
Temperature, C
State of cure, α
0.6
o
140
0.4 120
0.2 100
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
4.8 Conclusions
The simplified reversion type of cure kinetic model proposed by Han et al199 with
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
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
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
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
103
CHAPTER V
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-
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
has a long history. As early as 1933 Szalay211 described how ultrasonic waves
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.
chain near the center of the macromolecules151, 152. The mechanism of ultrasonic
cavitation can produce sufficiently high local pressure and temperature143, 144 to induce
of the presence of these radicals due to ultrasonic exposure was obtained by carrying out
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
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
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
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
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,
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
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
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
7 1600
1400
6
1000
4 800
3 600
400
2
200
1
0
0
0 2 4 6 8 10 12
Amplitude, µm
Figure 5.1 Die pressure and power consumption as a function of the ultrasonic
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
5
160oC cure
IR5µm IR
4 IR0µm
IR7.5µm
Torque, dNm
IR10µm
3
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,
crosslink density of the vulcanizates were evaluated and the results are shown in Figure
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
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
Crosslink density, kmol/m
90
0.20
Gel fraction, %
0.18
80
0.16
0.14
70
0.12
60 0.10
0 2 4 6 8 10 12
Amplitude, µm
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
dWf/dLog Mw
0.6
0.4
IR5µm
0.2
IR
0.0
IR10µm IR7.5µm IR0µm
3 4 5 6 7 8
Log Mw
107 7
(b)
6
Mw/Mn
Mw
5
Mn or Mw
Mw/Mn
106
Mn 4
105 2
IR 0 µm 5µm 7.5µm 10µm
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)
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
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
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
|η*|, Pa-s
106
Vulcanizates
105
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
114
102
IR
101 IR0µm
Gums IR5µm
IR7.5µm
100
IR10µm
tan δ
10-1
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
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
103
(a)
Vulcanizates
102
G', kPa
Gums
IR
101
IR0µm
IR5µm
IR7.5µm
IR10µm
100
10-2 10-1 100 101 102 103
ω, rad/s
102
(b)
101 Gum
G", kPa
100
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
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
| η 0* (T ) | T
| η * |= | η 0* (T ) |= A exp b
| η 0* (T ) | ω 1− n T
1+[ ]
τ
with A, Tb, n and τ being the fitting parameters, η 0* (T ) is the zero-frequency viscosity
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
activation energy E than the correspondent linear polymers225. 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
o
IR 60 C
104 o
IR10µm 60 C
o
IR 90 C
o
103 IR10µm 90 C
o
IR 120 C
o
IR10µm 120 C
102
10-2 10-1 100 101 102 103
ω, rad/s
107
106
105
|η*|, Pa-s
o
104 IR-5µm 60 C
o
IR-7.5µm 60 C
o
IR-5µm 90 C
103 o
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
118
Table 5.1 Rheological parameters of the modified Cross model for the virgin and
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
25
500mm/min
20 IR
0 µm
15
5 µm
σ, MPa
7.5 µm
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
σ , MPa
ε ,%
15 900
B
B
800
10
virgin IR 700
5 600
0 2 4 6 8 10 12
Amplitude, µm
1.0
(b)
0.9
E100
0.8
E100, E300, MPa
0.7
E300
0.6
0.5 virgin IR
0.4
0 2 4 6 8 10 12
Amplitude, µm
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
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
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
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
only detectable difference is that the Tg values of vulcanizates are 3-4oC 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
substantial Tg differences among the treated EPDM gums, namely 4 to 6oC. 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.
ultrasonic amplitude.
100
80 IR
IR-0µm
Weight, %
60 IR-5µm
IR-7.5µm
40 IR-10µm
Vulcanizates
20
Gums
0 ~ 5.0%
Figure 5.11 TGA curves for the virgin and ultrasonically treated IR gums and their
123
1.0
Gums IR
0.5 IR0µm
Heat Flow (W/g)
IR5µm
IR7.5µm
0.0 IR10µm
Vulcanizates IR
IR0µm
-0.5
IR5µm
IR7.5µm
IR10µm
-1.0
-90 -70 -50 -30 -10
Exo up Universal V3.9A TA Instruments
Temperature (oC)
Figure 5.12 DSC curves for the virgin and ultrasonically treated IR and their
Table 5.2 Tg of the virgin and the ultrasonically treated IR gums and their
vulcanizates
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
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
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
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
126
CHAPTER VI
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
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.
prompted increased usage of the more pure synthetic polyisoprene in some applications4.
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-
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.
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
the processes taking place during the ultrasonic devulcanization of the unfilled IR. In
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
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.
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
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
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
Torque, dNm 3
2
virgin
1 5µm
7.5µm
10µm
0
0 5 10 15 20 25
Time, min
7
(b) NR
6
5
Torque, Nm
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
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
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
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
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
the viscosity. In contrast, for IR and NR vulcanizates the power consumption shows
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,
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
Power, watts
800
600
Flow rate, g/s
400 0.47
0.63
1.07
200 2.10
2.55
0
4 5 6 7 8 9 10 11
Amplitude, µm
1800
(b) NR
1500 Flow rate, g/s
0.32
1200 0.63
Power, watts
1.26
2.52
900
600
300
0
4 5 6 7 8 9 10 11
Amplitude, µm
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
10
0
0 2 4 6 8 10 12
Amplitude, µm
16 1.26
2.52
12
0
0 2 4 6 8 10 12
Amplitude, µm
Figure 6.3 Die pressure for IR (a) and NR (b)181 rubbers devulcanized at various flow
134
chains do not have sufficient time to meet each other and to combine together
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
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
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
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
135
100
(a) IR
90
Gel fraction, %
80
100
(b) NR
95
90
Gel fraction, %
85
80
75
Flow rate, g/s
70 0.32
0.63
65 1.26
2.52
60
0 2 4 6 8 10 12
Amplitude, µm
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
0.47
0.10
0.05
(a) IR
0.00
0 2 4 6 8 10 12
Amplitude, µm
0.25
Flow rate, g/s
0.32
Crosslink density, kmol/m3
0.63
0.20
1.26
2.52
0.15
0.10
0.05
(b) NR
0.00
0 2 4 6 8 10 12
Amplitude, µm
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
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
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
undergo revulcanization than IR. The results also indicate that in the presence of
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
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
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
139
2.0
(a)
1.5 5µm
dWf/d(logMw)
IR
1.0 7.5µm
10µm
0.5
0.0
2 3 4 5 6 7 8
logMw
107 3.6
(b)
3.2
106
2.8
Mw/Mn
Mw
105 2.4
2.0
104
1.6
103 1.2
0 2 4 6 8 10 12
Amplitude, µm
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
a barrel temperature of 120oC. 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
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
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
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
distribution of IR than those of NR (Mn: 180,400; Mw: 1,116,000; Mw/Mn: 6.19)182. Also
NR has a higher amount of cis- 1, 4 isoprene structure (>99%) than IR (98%). The higher
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
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
5µm
108 7.5µm
10µm 0.6
|η*|, Pa-s 107
tan δ
106 0.4
105
0.2
104
0.0
103
102
10-2 10-1 100 101 102 103
ω, rad/s
108
(b)
107
106
|η*|, Pa-s
105
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)
G', kPa
102
virgin IR
IR cured
5µm
7.5µm
10µm
101
10-2 10-1 100 101 102 103
ω, rad/s
102
(b)
101
G", kPa
100
virgin IR
IR cured
10-1 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
σ, MPa
10
7.5 µm
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
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
discussed earlier. This phenomenon is consistent with the power consumption, gel
146
22 1000
(a)
20
18 800
16 ε
B
σ , MPa
14 600
ε ,%
B
12
B
10 σ 400
B
8
6 IR 200
NR
4
0 2 4 6 8 10 12
Amplitude, µm
1.0
(b)
0.9
E100
0.8
E100, E300, MPa
0.7
E300
0.6
0.5
IR
NR
0.4
0 2 4 6 8 10 12
Amplitude, µm
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
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
Rubbers Tg, oC
Devulcanized at 5 µm -59.4
Devulcanized at 10 µm -59.3
149
6.10 Conclusions
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,
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
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
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
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
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
7.1 General
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
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
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,
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
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
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
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,
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
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
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
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,
seen to be more severe in CB filled IR. The obtained induction time following the method
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
16
60 phr
14
12
10 35 phr
Torque, dNm
8
25 phr
6
15 phr
4
0 phr
2
0 5 10 15 20 25
Time, min
Figure 7.1 Cure kinetics of the filled IR at various CB loadings at 160oC (strain
156
10
ti=8.43*e-0.06*CB(phr) (160 C)
o
6
ti, min
0 20 40 60 80 100
CB, phr
157
107 o
160 C Virgin IR
IR in Banbury
106
105
|η*|, Pa-s
104
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
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
6
15phr virgin IR
5
unfilled IR in Banbury
4
Torque, dNm
3
unfilled virgin IR
2
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)
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
with the rubber matrix causes the formation of bound rubber layer at the surface of CB.
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
1.00 400
3
Crosslink density, mol/m
0.90 300
Gel fraction
0.80 200
0.70 100
0.60 0
0 20 40 60
CB, phr
Figure 7.5 Gel fraction and crosslink density of virgin IR vulcanizates versus carbon
black loading
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
to very high tensile strength and elongation at break. This indicates that the addition of
160
500mm/min CB, phr
0
30
15
25
35
60
σ, MPa
20
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
σB, MPa 20
εB, %
600
15
10
400
σB
5
εB
0 200
0 10 20 30 40 50 60 70
CB, phr
8
E100 (b)
7
E300
6
E100, E300, MPa
0
0 10 20 30 40 50 60 70
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 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.
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
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
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
1400
CB, phr
0
1200 15
Power consumption, watts
25
1000 35
60
800
600
400
200
4 5 6 7 8 9 10 11
Amplitude, µm
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
Die pressure, MPa
25
25
35
60
20
15
10
5
0 2 4 6 8 10 12
Amplitude, µm
The gel fraction and the crosslink density of the devulcanized IR (open symbols)
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
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
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
1.00
0.95
0.90
Gel fraction
0.85
CB, phr
0.80 0
15
35
0.75
35 (revulc)
0.70
0 2 4 6 8 10 12
Amplitude, µm
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 15
3
Crosslink density, mol/m
35
200 35 (revulc)
150
100
50
0
0 2 4 6 8 10 12
Amplitude, µm
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
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
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
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
|η*|, Pa-s 107
0.6
tan δ
106
0.4
105
0.2
104
103 0.0
10-2 10-1 100 101 102 103
ω, rad/s
108
(b)
107
|η*|, Pa-s
106
105
virgin IR/35CB
IR/35CB cured
104 5µm
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
(b)
10-1
G", kPa
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
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
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
unfilled IR shown earlier in Chapter VI Section 6.3. Again, it is due to the presence of
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
171
14
35phr virgin
12 35phr 5 µm
35phr 7.5 µm
10 35phr 10 µm
Torque, dNm
8
0
0 5 10 15 20 25
Time, min
Figure 7.14 Revulcanization curves at 160oC for the devulcanized 35 phr CB filled IR
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
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
σ, MPa
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
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
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
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
CB/oil, phr
12
0/0
35/0
10 35/10
Torque, dNm
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
20
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
25
800
σB, MPa
εB, %
20
600
15
400
10
5 200
0 0
IR IR/35CB IR/35CB/10oil
Vulcanizates
4
E100 (b)
E300
3
E100, E300, MPa
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
Power consumption, watts
1200
1000
800
600
400
200
4 5 6 7 8 9 10 11
Amplitude, µm
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
15
10
0
0 2 4 6 8 10 12
Amplitude, µm
Figure 7.20 Die pressure as a function of ultrasonic amplitude for the 35 CB filled IR
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
Gel fraction
0.85
35/0 (devulc)
200 35/0 (revulc)
150
100
50
0
0 2 4 6 8 10 12
Amplitude, µm
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
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
|η*|, Pa-s 106
(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
tan δ
(b)
10-1
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
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
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
7.5.4 Revulcanization
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
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
G', Pa
100
10-1
10-1
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
184
10
8 0-0
10-0
0-5
Torque, dNm 6 0-7.5
10-5
10-7.5
10-10
4 0-10
2
oil(phr)-amplitude(µm)
0 5 10 15 20
Time, min
(containing 0 and 10 phr processing oil) at 160oC, a strain amplitude of 4.2% and a
(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
ingredients is difficult. In this case, the processing oil present in the devulcanized rubber
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
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
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
decreased with the increase of amplitude. The modulus behavior is consistent with that of
186
10 500
(a)
8 400
εB, %
4 oil, phr 200
0 (σB)
10 (σB)
2 100
0 (εB)
10 (εB)
0 0
0.0 2.5 5.0 7.5 10.0
Amplitude, µm
2.5
(b)
2.0
E100, MPa
1.5
1.0
strain E100 (c) as a function of ultrasonic amplitude for the revulcanizates of 35 phr CB
187
7.6 Effect of Retarder on Vulcanization and Revulcanization of Filled IR
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
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
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
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
10
0-0 1-0
8
Torque, dNm
1-5
1-7.5
6
0-5
1-10
0-10
4 0-7.5
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
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
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 (σB)
600
20 1 (σB)
0 (εB) 500
σB, MPa 15 1 (εB)
εB, %
400
10
300
5
200
0 100
0.0 2.5 5.0 7.5 10.0
Amplitude, µm
2.8
(b) SAFE, phr
2.6 0
1
2.4
2.2
E100, MPa
2.0
1.8
1.6
1.4
1.2
0.0 2.5 5.0 7.5 10.0
Amplitude, µm
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
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
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
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
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
2.54 mm. The cure kinetics and the mechanical properties are studied. Another objective
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
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
6
1
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
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
1
10
2
3
8 4
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
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
improves the processability and the mixing of the two different types of rubbers:
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
σB, MPa
10
5
dIR/IR
gIR/IR
0
0 25 50 75 100
Concentration of virgin IR in blends, wt%
500 3.0
(b)
2.5
400
2.0
E100, MPa
300
εB, %
1.5
200
dIR/IR (εB)
1.0
gIR/IR (εB)
100 dIR/IR (E100) 0.5
gIR/IR (E100)
0 0.0
0 25 50 75 100
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
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
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
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
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
The retarder introduced in the revulcanization recipe was effective to improve the
scorch safety and minimize the reversion in revulcanization. However, it was ineffective
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
198
CHAPTER VIII
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
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
Solid-state NMR technique has been applied to study the molecular dynamics and
butyl rubber244 and BR245. Proton NMR transverse magnetic relaxation reflects mainly
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
long T2 relaxation and the unextractable gel for short T2 relaxation122. In contrast, in the
study of BR proton T2 relaxation decay, it was shown that three-component model was
relaxation, the unentangled light sol and the dangling network chain ends for intermediate
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.
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
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.
All the NMR measurements (T2 relaxation decay and molecular diffusion) were
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
Oil — — 100 —
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
sensitive detection of the echo signal recorded 3 kHz off-resonance was followed by
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
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
entangled network, would not have interfere with the direct measurement of sol or
expected low sol diffusion coefficient confirmed that the T2 values measured with the
The 1H 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
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
A(2τ ) 2τ E 2τ 2τ
= f S exp(− ) + f L exp(− ) + (1 − f S − f L ) exp(− ) (8.1)
A(0) T2 S T2 L T2 M
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<E<2) accounting
When the diffusion rates display an arbitrary distribution, they might be modeled
as several components with fractional amplitude ai and diffusion coefficients Di. The
A(2τ , X )
= ∑ exp ai (−γ 2 Di X ) (8.2)
A(2τ , 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
A(2τ , X )
= f fast exp(−γ 2 D fast X ) + (1 − f fast ) exp(−γ 2 Dslow X ) (8.3)
A(2τ , 0)
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
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:
1 1
'
where f long (2τ ) = f long + (1 − f short − flong ) exp 2τ − (8.5)
T2long T2 med
The molecular mobility of the treated gum IR and their vulcanizates was
molecular weight and glass transition temperature of the virgin and treated gum IR was
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
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
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
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
decay together with a three-component modeling is given in Figure 8.2. While the
206
Figure 8.1 Spectral decomposition of the 1H T2 relaxation decay at 70.5oC for the
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-
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
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
(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
T2, ms
10
T2M
T2S
1
0 2 4 6 8 10 12
Amplitude, µm
Figure 8.3 Proton T2 for the gum IRs ultrasonically treated at different amplitudes
209
100
T2L
T2, ms
10
T2M
T2S
1
0 2 4 6 8 10 12
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
210
100
80 f2S
Proton T2 fractions, %
60
40
20 f2M
f2L
0
0 2 4 6 8 10 12
Amplitude, µm
Figure 8.5 Proton T2 fractions for the gum IR ultrasonically treated at different
amplitudes
100
f2S
80
Proton T2 fractions, %
60
40
20
f2M
0 f2L
0 2 4 6 8 10 12
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
spacing τ chosen resulted in essentially the full intensity of the long 1H T2 component
being observed with a minor contribution from the intermediate component, and avoiding
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
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
Log (D, cm /s)
-7.0
2
Dfast Dslow
treated gum
-7.5 vulcanizate
-8.0
-8.5
0 2 4 6 8 10 12
Amplitude, µm
Figure 8.8 Diffusion coefficient for the ultrasonically treated IR gum and their
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
oligomer molecules into the chemical network, with the rest of the light molecules left for
diffusion.
Fast-diffusing echo fraction (at t=2τ) ffast, %
80
60
40
20
treated gum
vulcanizate
0
0 2 4 6 8 10 12
Amplitude, µm
Figure 8.9 Fast-diffusing echo fraction for the ultrasonically treated gum IRs and
214
10
0
0 2 4 6 8 10 12
Amplitude, µm
Figure 8.10 Fast-diffusing sample fraction for the ultrasonically treated gum IRs and
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
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
in BR245. But it is different from SBR122 and silicone rubber241 where the glass transition
Table 8.3 Sol fraction and glass transition temperature of the virgin gum, vulcanized,
5 2.49 -59.4
10 10.91 -59.3
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
Therefore, the two-component model without including the Weibull exponent is not
sufficient to describe the 1H 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
217
Figure 8.12 Spectral decomposition of the 1H 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
for the cured and devulcanized rubber is shown in Figure 8.14. It is evident that an
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
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
35
flow rate, g/s
30 0.63
1.07
25 2.10
Sol fraction, %
20
15
10
0
0 2 4 6 8 10 12
Ultrasonic amplitude, µm
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
H T2 (ms)
T2L
10
1
T2S
1
0 20 40 60 80 100
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
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
221
100 fS
80
H T2 fractions, % 60
fsol
40
20
1
0 fL
0 20 40 60 80 100
Extracted sol, %
Figure 8.16 Proton T2 fractions for virgin, cured and devulcanized IR as a function of
extracted sol
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
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
for the proton transverse relaxation decay of the black filled IR.
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
224
Figure 8.19 Spectral decomposition of the 1H T2 relaxation decay for the 35 phr CB
225
30
oil, phr
25 0
10
20
Sol fraction, %
15
10
0
0 2 4 6 8 10 12
Ultrasonic amplitude, µm
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
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
100
filled without oil
filled with oil
oil only
10
H T2S, ms
1
0.1
0 20 100
Extracted sol, %
Figure 8.21 Proton T2S for the virgin uncured, cured and devulcanized 35 phr CB filled
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
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
10
1
1
0 20 100
Extracted sol, %
Figure 8.22 Proton T2L for the virgin uncured, cured and devulcanized 35 phr CB filled
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
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
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.
H T2 long component fractions fL, %
60
40
20
0 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
1
H 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
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
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
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
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
Analysis from three different rubber compounds (ultrasonically treated gums plus
their vulcanizates, the unfilled and the CB filled IR) indicated that NMR relaxation
(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
234
CHAPTER IX
9.1 General
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
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
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
method was already successfully applied to many other rubbers127, 229, 250.
Firstly it was assumed that the primary polymer chains obey the Flory molecular
weight distribution
∞
wn = (1 − p ) 2 np n −1 and ∑w
n =1
n =1 (9.1)
where wn is the weight fraction of chains having n monomeric units, and p is the Flory
1 1+ p
N n ( p) = and N w ( p) = (9.2)
1− p 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
236
forming crosslinks. Based on the classical gel theory, if α exceeds a certain critical
value:
1
α ≥ α crit = (9.3)
Nw −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
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:
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 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 )
237
ρ ξ ( elast ) ( p, α )
ξ (g)
( p, α ) = (9.7)
M 0 N n ( p ) g ( p, α )
where ρ and M 0 are the rubber density and the molecular weight of a monomeric unit of
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
equations is applied below for the calculations of the normalized crosslink density in gel
ξ devulc
(g)
ξ ( g ) ( p, α )
ξ * ( p, α ) = =
ξ 0( g ) ξ ( g ) ( p0 , α 0 )
g devulc 1 − S ( p, α )
g * ( p, α ) = = (9.8)
g0 1 − S ( p0 ,α 0 )
where p0 and p are the initial and current values of the Flory distribution parameter and
α 0 and α 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
respectively. The above two equations can be combined together and re-written in the
following form
k p / kα
p (t ) α (t )
= (9.9)
p0 α
0
M0
The p0 value was determined as p0 = 1 − , where M n is the number-average
Mn
molecular weight of the rubber. The α 0 can be found using the experimental data on
crosslink density ξ 0( g ) and gel fraction g 0 in the virgin rubbers using Equation (9.7) with
(g) ρα 0
ξ 0 ( p 0 , α 0 ) = M g ( p , α ) (1 − υ 0 ) (1 + 2 υ 0 )
3
0 0 0 0
−2
1
υ 0 = 1 + − 1α 0 (1 − υ 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
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
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
kp/kα=4.1x10-3 (IR)
0.4 kp/kα=4.2x10-3 (NR)
kα=0
kp=0 (IR)
0.2 kp=0 (NR)
IR
NR
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Normalized crosslink density
Figure 9.1 Experimental (symbols) and fitted (lines) values of normalized gel fraction
number average molecular weight Mn of NR used was 180,400252, 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
values. For IR and NR, the kp/kα value of 4.1×10-3 and 4.2×10-3 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-
sulfidic) and Sx (multi-sulfidic) with the bond energies of 285, 268 and 251 kJ/mol253,
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.
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
vulcanizates of the correspondent CB level. The detailed values are listed in Table 9.2.
IR NR182
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
Normalized gel fraction
kα=0
0.6
kp=0 (IR)
kp=0 (NR)
0.4 -3
exp predic CB/oil (phr) kp/kα (x10 )
15/0 5.3
0.2 IR 35/0 17.1
35/10 4.0
15/0 6.7
NR
35/0 8.1
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Normalized crosslink density
Figure 9.2 Experimental (symbols) and fitted (lines) values of normalized gel fraction
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 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:
6.19) used in earlier study182 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
9.5 Conclusions
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
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
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
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
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
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
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
the ultrasonic amplitudes, the gel fraction and crosslink density were lower than those of
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
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
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
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
The retarder was combined in the revulcanization recipe and it was effective to
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
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
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
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
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
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
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
253
REFERENCES
1 L.K. Opeke, “Tropical Tree Crops”, John Wiley and Sons, New York, 1982.
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.
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.
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.
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.
29 P.W. Allen, “Natural Rubber and the Synthetics”, John Wiley & Sons, New York,
1972.
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
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.
45 http://www.continentalcarbon.com/downloads/whatis.pdf
47 ASTM D1765, Standard Classification System for Carbon Blacks Used in Rubber
Products, 1968.
257
50 B.B. Boonstra, “Role of Particulate Fillers in Elastomer Reinforcement: a Review”,
Polymer, 20, 691-704, 1979.
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.
55 A.K. Bagchi, B.G. Sharma, “Reinforcement and Physical Properties of Filled Rubber
System”, Indian Journal of Technology, 19, 368-372, 1981.
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.
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.
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.
68 A.M. Bueche, “The Curing of Silicone Rubber with Benzoyl Peroxide”, J. Polym.
Sci., 15, 105-120, 1955.
259
76 K.S. Loganathan, ed., “Rubber Vulcanization”, in “Rubber Engineering”, Indian
Rubber Institute, McGraw, NY, 2000.
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.
84 A.Y. Coran, “Vulcanization. VI. A Model and Treatment for Scorch Delay Kinetics”,
Rubber Chem. Technol., 37, 689-697, 1964.
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.
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.
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.
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.
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.
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.
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.
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.
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.
148 J.R. Thomas, “Sonic Degradation of High Polymers in Solution”, J. Phy. Chem., 63,
1725-1729, 1959.
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.
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.
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.
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.
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.
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.
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.
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.
190 G.R. Hamed, “Energy Dissipation and the Fracture of Rubber Vulcanizates”; Rubber
Chem. Technol., 64, 493-500, 1991.
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.
197 J.S. Deng, A.I. Isayev, “Injection Molding of Rubber Compounds: Experimentation
and Simulation”, Rubber Chem. Technol., 64, 296-324, 1991.
199 I. Han, C. Chung, J. Lee, “Optimal Curing of Rubber Compounds with Reversion
Type Cure Behavior”, Rubber Chem. Technol., 73, 101-113, 2000.
202 B. Sujan, “Injection and Compression Molding of Devulcanized GRT: Cure Kinetics
and Physical Properties”, Master Thesis, University of Akron, 2005.
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.
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.
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.
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.
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.
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.
230 V.M. Makarov, V.F. Drozdovski, “Reprocessing of Tires and Rubber Wastes”; Ellis
Horwood, New York, Chapter 2, 1991.
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.
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.
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.
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