Rubber Compounding - Chemistry and Applications Part 3 pptx

2
General-Purpose Elastomers
Howard Colvin
Riba-Fairfield, Decatur, Illinois, U.S.A.
I. INTRODUCTION
General-purpose elastomers played a critical role in the history of the last half
of the 20th century. In 1942 the Rubber Reserve program developed both the
basic technology and manufacturing capability to make emulsion styrene
butadiene rubber (SBR) just a few years after World War II had interrupted
natural rubber supplies. Historians have noted that the scientific contribution
to that effort is comparable to the nuclear research program at Los Alamos
that occurred at the same time (1). After the petroleum shortages of the 1970s,
fuel economy became a primary driving force in the automotive industry, and
the tire industry was challenged to develop new products that would improve
gas mileage. New elastomers based on solution SBR technology proved to be
part of the answer.
Today the tire industry is challenged to meet new environmental
standards while maintaining or improving the vehicle handling, ride, and
durability that has already been achieved. To meet this challenge, the rubber
technologist must have a thorough understanding of how general-purpose
elastomers (i.e., polybutadiene, styrene/butadiene, and styrene/butadiene/
isoprene) affect compound processability, tire rolling resistance, tire traction,
tire treadwear, and overall cost of tire components. Use of these elastomers
outside of the tire industry requires the same type of understanding of
fundamental polymer characteristics and how they affect the final applica-
tion. This review will describe the basic structure–property relationships
between general-purpose elastomers and end-use properties, with a focus on
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 51
Copyright © 2004 by Taylor & Francis
the tire industry. The processes used to make the general-purpose elastomers

will be described with an emphasis on how the polymerization variables
(mechanism, catalyst, process) affect the macrostructure and microstructure
of the polymer. It is polymer microstructure and macrostructure that
determine whether a polymer is suitable for a particular application, not
the type of process or catalyst used to produce the polymer.
Some important terms used in this chapter are defined in Table 1.
II. STRUCTURE–PROPERTY RELATIONSHIPS FOR
GENERAL-PURPOSE ELASTOMERS USED
IN TIRE APPLICATIONS
A. Laboratory Testing Methods
Prediction of tire properties based on laboratory properties has met with
various degrees of success, depending on which property was being predicted.
There is a good correlation between the rolling resistance of tires and the tread
compound tangent delta at 60jC and 40 Hz (2). There is a reasonable
Table 1 Definitions
Polymer microstructure Monomers incorporated into the polymer and the stereo-
chemistry of enchainment (i.e., cis, trans, vinyl).
Polymer macrostructure Polymer molecular weight and molecular weight distribu-
tion, molecular geometry (linear, branched, comb), and the order in which mono-
mers are incorporated (block, tapered block, or random).
Number-average molecular weight (M
n
) Summation of the number of polymer
chains (N) with a given molecular weight (m) times the molecular weight of each
chain divided by the total number of polymer chains: Sm
i
N
i
/SN
i

.
Weight-average molecular weight (M
w
) Summation of the number of polymer
chains (N) with a given molecular weight (m) times the square of the molecular
weight of each polymer chain divided by the total number of polymer chains times
the molecular weight of each chain: Sm
i
2
N
i
/Sm
i
N
i
.
Molecular weight distribution M
w
/M
n
.
Glass transition temperature (T
g
) Temperature at which local molecular motion in
a polymer chain virtually ceases. General-purpose elastomers behave like a glass
below this temperature.
Weight-average T
g
Average T
g

of a compound:
X

wt: polymer X
n
total polymer wt:

ðT
g
polymer X
n
Þ

4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 52
Copyright © 2004 by Taylor & Francis
correlation between tire traction and tangent delta of the tread compound at
0jC and 40 Hz (2). Tire wear is more difficult to predict, with one researcher
observing, ‘‘Despite more than 50 years of effort to devise laboratory abraders
that give a good prediction of the wear resistance in real-world situations, no
abrasion device currently exists that does an acceptable job’’ (3). Typically,
DIN abrasion or some type of blade abrader is used as a general indicator,
however. Rubber processability has been defined in a number of ways (4) but
is usually determined by what type of equipment will be used to process the
rubber. Mooney stress relaxation time to 80% decay (MSR t-80) is a rapid,
effective processability test that works well with both emulsion (5) and
solution SBR (6). Other more sophisticated instruments such as the rubber
processability analyzer (RPA) or capillary rheometer are now becoming more
popular.
B. Glass Transition Temperature

The most important elastomer variable in determining overall tire perform-
ance is the glass transition temperature, T
g
. Aggarwal et al. (2) showed that
the tangent delta at 60jC of filled rubber vulcanizates made from ‘‘conven-
tional rubbers’’ correlated with tire rolling resistance and then determined
that the tangent delta values were approximately a linear function of the
compound’s T
g
value. This was true whether the polymers were made by a
solution process or an emulsion process. They did not compare solution and
emulsion polymers at the same glass transition temperature.
Oberster et al. (7) showed that traction and wear properties were not
dependent on the way the polymer was manufactured but were functions of
the overall glass transition temperature of the compound, as shown in Figures
1 and 2. In actual tire tests, results are more complicated. The weight-average
T
g
of the tread compound is still a major variable, but it is not as dominant as
in laboratory tests. A comprehensive study of tire wear under a variety of
environmental and road conditions showed that tire wear improves linearly as
the ratio of BR to SBR is increased in BR–SBR tread compounds (lower
weight-average T
g
). The wear behavior was more complex in BR–NR blends
with low carbon black levels and was shown to be a function of ambient test
temperature (3).
Nordsiek (8) expanded the concept of using the glass transition tem-
perature to using the entire damping curve to predict tire performance. He
divided the damping curve into regions that influenced various tire properties

(Fig. 3). The damping curves for an emulsion SBR, a high-vinyl polybutadi-
ene, and a medium-vinyl SBR at the same T
g
were compared and shown to be
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 53
Copyright © 2004 by Taylor & Francis
different at temperatures of 20–100jC. This led to the proposal of an ‘‘ integral
rubber’’ that would have a compilation of damping curves from a number of
polymers and would incorporate damping behavior that would lead to the
‘‘ideal’’ elastomer for tread compounds. It was implied that this elastomer
consisted of segmented blocks of different elastomers with different glass
transition temperatures. An ‘‘integral rubber’’ was prepared and compared to
Figure 1 Effect of T
g
on traction of (x) solution polymers and (n) emulsion poly-
mers. (From Ref. 7.)
Figure 2 Effect of T
g
on wear of (x) solution polymers and (n) emulsion polymers.
(From Ref. 7.)
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 54
Copyright © 2004 by Taylor & Francis
natural rubber and SBR 1500 controls in a laboratory compounding study.
The ‘‘integral rubber’’ had a hot rebound within one point of the natural
rubber control and was three points higher than the SBR 1500 control.
Abrasion resistance was better than that of the natural rubber control but
slightly worse than that of the SBR 1500. The 0jC rebound was lower than
that of either control.

C. Molecular Weight and Molecular Weight Distribution
The molecular weight aspect of polymer macrostructure affects the rolling
resistance (via hysteresis) and processability of the tread compound. As the
molecular weight is increased, the total number of free chain ends in a rubber
sample is reduced, and energy loss of the cured compound is reduced. This
leads to improved rolling resistance, but at the expense of processability.
Caution should be used in extrapolating lab data on high molecular weight
rubbers to factory-mixed stocks, because filler dispersion is not as efficient
with large-scale equipment. Thus, low hysteresis in lab compounds may not
translate into low hysteresis in commercial tire compounds. There is an
optimum balance between molecular weight and processability that is defined
Figure 3 Damping curve of ESBR 1500 tread compound. (From Ref. 8.)
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 55
Copyright © 2004 by Taylor & Francis
by the type of mixing equipment used. Increasing the molecular weight
distribution at equivalent molecular weight by branching produces more free
chain ends and more hysteresis but at moderate levels can improve other
properties. Saito (9) showed that in silicon-branched solution SBR the effect
on hysteresis could be minimized and ultimate tensile strength could be
improved because of better carbon black dispersion. In emulsion polymers,
the branching is uncontrolled and the polymers have poorer hysteresis than
the corresponding solution polymer (10). From a practical standpoint, some
branching in tire polymers is necessary to prevent cold flow and ensure that
the elastomer bales will retain their dimensions on storage.
Polymer scientists have worked hard to take advantage of the relation-
ship between free chain ends and hysteresis. In one case, an attempt was made
to eliminate chain ends completely by preparing cyclic polymers. Hall (11)
polymerized butadiene with a cyclic initiator and claimed to have made a
mixture of linear and cyclic polybutadiene. Cyclic structure was inferred from

a comparison of the viscous modulus of the cyclic polymer to that of a linear
control. All of the cyclic polymers had a lower viscous modulus than the
controls. No compounding data were reported, however.
A more popular method of reducing the effective number of free chain
ends is to functionalize the end of the polymer chain with a polar group.
Functional end groups can enhance the probability of cross-linking near the
chain end and interact directly with the filler, thus reducing end effects.
Ideally, difunctional low molecular weight polymers would be mixed with
filler and then chemically react with the filler during vulcanization to give a
network with no free chain ends. This ideal can be approached, depending on
how effectively the polymer chains are functionalized and the strength of the
interaction of the functional group with the filler. This will be discussed
further in the section on anionic polymerization and anionic polymers
(Section IV).
D. Sequence Distribution in Solution SBR
Day and Futamura (12) compared different 35% styrene solution SBRs at
equivalent molecular weights and found that hysteresis is a linear function of
the block styrene content. The effect of the polystyrene block length on
hysteresis is shown in Figure 4.
Sakakibara et al. (13) made block polymers of polybutadiene and SBR
with anionic polymerization and compared them to an SBR with the same
overall microstructure. They found that the block polymers had broader glass
transition temperatures that resulted in better wet skid resistance and lower
rolling resistance than the corresponding random SBRs. They also found that
blocky styrene in the SBR block was detrimental to overall performance.
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 56
Copyright © 2004 by Taylor & Francis
III. EMULSION POLYMERIZATION AND EMULSION
POLYMERS

The copolymerization of styrene and butadiene is accomplished by dispersing
the monomers in water in the presence of a surfactant, an initiator, and a
chain transfer agent. The process offers limited control over polymer micro-
structure, and the polymers are branched. Emulsion SBR, however, has
played and continues to play an important role in tire compounds.
A. Polymerization
The best way to consider the overall emulsion process is to examine the
original recipe used to produce GR-S rubber at the beginning of World War II
(14) (Table 2).
It is important that the polymerization be done in the absence of oxygen.
Oxygen is removed from the water by bubbling nitrogen through it prior to
the polymerization, and the polymerization is conducted under a nitrogen
atmosphere. When the ingredients are mixed, the monomers are partitioned
between the water, micelles, and monomer droplets. The water solubility of
styrene and butadiene is very low, so there is little of either in the water phase.
Micelles are aggregates of surfactant (fatty acid soap) with the polar carbox-
ylic group on the outside oriented toward the polar water and the nonpolar
hydrocarbon tail oriented toward the inside of the micelle. The nonpolar
styrene and butadiene are ‘‘soluble’’ inside the nonpolar environment of the
micelle. Still, only a small portion of the monomer is located in micelles. There
Figure 4 Effect of block styrene on hysteresis in SBR. (From Ref. 12.)
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 57
Copyright © 2004 by Taylor & Francis
are approximately 10
17
–10
18
micelles per milliliter of emulsion (15). Most of
the monomer is contained in monomer droplets, which are in lower concen-

tration (10
10
–10
11
monomer droplets per milliliter emulsion) and much larger
than the micelles (15). When the mixture is heated to 50jC, the potassium
persulfate decomposes into radicals in the aqueous phase. Because the surface
area of the micelles is much greater than that of monomer droplets, the
radicals are more likely to inoculate the micelles to begin the polymerization.
A representation of this is shown in Figure 5.
As the polymerization proceeds, monomer migrates from the monomer
droplets to the micelles until the monomer droplets are gone. Chain transfer
to the mercaptan controls polymer molecular weight. Conversion is stopped
at approximately 70% by addition of a radical trap such as the salt of a
dithiocarbamate or hydroquinone. The latex is stabilized, then coagulated to
give crumb rubber.
A major improvement in this process was the development of the redox
initiation system shortly after World War II (16) (Table 3). With this recipe,
the polymerization could be conducted at 5jC by changing the initiator
system from potassium persulfate to cumene hydroperoxide. The iron(II) salt
lowers the activation energy for the decomposition of the cumene hydroper-
oxide and is oxidized to iron(III) during the process. The dextrose is present to
reduce the iron(III) back to iron(II) so more peroxide can be decomposed.
The importance of the lower polymerization temperature is shown in
Figure 6. As the polymerization temperature is decreased, the ultimate tensile
strength of cured rubber increases dramatically (17). This is because there is
less low molecular weight material and less branching at the lower polymer-
ization temperature (18).
There is little control over butadiene polymer microstructure in the
emulsion process. It remains fairly constant at 12–18% cis, 72–65% trans, and

16–17% vinyl as the polymerization temperature is increased from 5jCto
Table 2 GR-S Recipe for Emulsion SBR
a
Component Parts by weight
Styrene 25
Butadiene 75
Water (deoxygenated) 180
Fatty acid soap 5
Dodecyl mercaptan 0.5
Potassium persulfate 0.3
a
Polymerization conducted at 50jC.
Source: Ref. 15.
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 58
Copyright © 2004 by Taylor & Francis
50jC. Butadiene microstructure does not vary significantly as the styrene
content is changed (19). The glass transition temperature of emulsion SBR is
controlled by the amount of styrene in the polymer.
B. Functional Emulsion Polymers
It is easy to incorporate a functional monomer into an emulsion polymer as
long as there is some water solubility. Emulsion butadiene or styrene
Figure 5 Species present during emulsion polymerization. (From Ref. 15. Re-
printed by permission.)
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 59
Copyright © 2004 by Taylor & Francis
Table 3 ‘‘Custom’’ Recipe for Emulsion SBR
Component Parts by weight
Styrene 28

Butadiene 72
Water 180
Potassium soap of rosin acid 4.7
Mixed tertiary mercaptans 0.24
Cumene hydroperoxide 0.1
Dextrose 1.0
Iron(II) sulfate heptahydrate 0.14
Potassium pyrophosphate 0.177
Potassium chloride 0.5
Potassium hydroxide 0.1
Source: Ref. 16.
Figure 6 Effect of polymerization temperature on mechanical properties of ESBR.
(From Ref. 18. Reproduced with permission.)
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 60
Copyright © 2004 by Taylor & Francis
butadiene rubbers containing acrylate, amine, cyano, and hydroxyl groups
have been made. Although some recent work has been done in exploring the
interaction of functional emulsion rubbers with fillers, more work could be
done. Emulsion SBR containing 3–5% acrylonitrile displays better abrasion
resistance than the corresponding unfunctionalized rubber in carbon black
compounds (20). Emulsion SBRs containing one to four parts of copolym-
erized amines were compounded into silica-containing stocks and showed
good processability, improved tensile strength, lower hysteresis, and better
abrasion resistance than a corresponding emulsion SBR control (21).
C. Oil-Extended Emulsion Polymers
A substantial percentage of the rubber used in tire compounds is oil-extended
emulsion SBR, which is prepared by adding an emulsion of oil to SBR latex
prior to coagulation. Oil extension allows higher molecular weight elastomers
to be used without processing problems, and incorporating the oil into the

latex is much easier than putting it in the compound at the mixer. The oils used
in compounding rubber are classified as paraffinic, naphthenic, and aromatic
depending on the aromatic content of the oil. The different types of oils affect
rubber compounds differently, and they cannot be directly substituted for
each other without compounding changes. The more paraffinic the oil is, the
lower its T
g
, which will lead to different compound properties than a higher T
g
naphthenic or aromatic oil. Direct comparison of SBR 1712 (37.5 phr
aromatic oil) with SBR 1778 (37.5 phr of naphthenic oil) in a sulfur-
vulcanized stock showed that the 1778 stock had a six point higher room
temperature rebound and a higher 300% modulus but poorer wet traction
(22). Schneider et al. suggested using a higher surface area black and adding
small amounts of a higher T
g
SBR to match the 1712 performance. Since the
late 1980s the aromatic oil used in SBR 1712 has come under fire for
containing polycyclic aromatics that may be a factor in causing cancer.
Compounders must be ready to make the necessary changes to eliminate
the high aromatic oil if necessary.
D. Emulsion–Filler Masterbatches
Carbon black and carbon black–oil masterbatches of emulsion SBR have
been used commercially for a long time. They are prepared by blending a
dispersion of carbon black and oil with latex followed by coagulation.
Masterbatching offers the advantages of improved black dispersion and
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 61
Copyright © 2004 by Taylor & Francis
shorter mix times. A major problem with masterbatching is that it limits

compound flexibility to compounds that contain the type of black that is in
the masterbatch. There can also be unexpected effects on the vulcanization
rate (23). Surprisingly, there is no commercial counterpart in an emulsion
SBR silica masterbatch, although there have been a number of patents on the
subject (24–27). In most of these patents, a dispersion of silica and some ma-
terial to reduce the filler–filler interaction is blended with the latex prior to
coagulation. The problems encountered with carbon black masterbatch are
also expected in silica masterbatches.
E. Commercial Emulsion Polymers and Process
The International Institute of Synthetic Rubber Producers (IISRP) classifies
commercial emulsion polymers as shown in Table 4. Specifics (soap type,
Mooney viscosity, coagulation, and supplier) for different grades of polymers
are provided in the detailed section of the IISRP Synthetic Rubber Manual
(28).
A schematic representation of a commercial continuous emulsion SBR
process is shown in Figures 7 and 8. Most of the ingredients are mixed and
cooled, then combined with a solution of initiator immediately before they
enter the first reactor. The number of reactors is chosen to control the
residence time to reach 60–65% conversion in 10–12 hr. The polymerization
is shortstopped, and the latex is pumped to a blowdown tank and flash tanks
to remove most of the residual butadiene. A dispersion of an antioxidant is
added to protect the polymer through the subsequent processing steps and
Table 4 Numbering System for Commercial Emulsion
Polymers
Series no. Description
1000 Hot nonpigmented rubbers
1500 Cold nonpigmented rubbers
1600 Cold black masterbatch with 14 or
less parts of oil per 100 parts SBR
1700 Cold oil masterbatch

1800 Cold oil black masterbatch with more
than 14 parts of oil per 100 parts SBR
1900 Emulsion resin rubber masterbatches
Source: Ref. 28.
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 62
Copyright © 2004 by Taylor & Francis
Figure 7 Emulsion polymer process—polymerization. (Courtesy of G. Rogerson, Goodyear Tire & Rubber Co.,
Akron, OH.)
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 63
Copyright © 2004 by Taylor & Francis
Figure 8 Emulsion polymer process—finishing. (Courtesy of G. Rogerson, Goodyear Tire & Rubber Co., Akron, OH.)
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 64
Copyright © 2004 by Taylor & Francis
storage prior to use. The latex is then steam stripped to remove the rest of the
butadiene and all of the styrene. Crumb rubber is produced by coagulation in
a solution of acidic sodium chloride. After washing, the crumb is dried and
baled (19).
IV. ANIONIC POLYMERIZATION AND ANIONIC POLYMERS
Anionic polymerization offers the rubber technologist the maximum versa-
tility in preparing new elastomers. The procedure involves reaction of a
lithium alkyl with a diene or combination of styrene and diene(s) in a
hydrocarbon solvent. The polymerization typically produces a polymer with
a narrow molecular weight distribution because each initiator molecule
produces one polymer chain, and initiation is fast relative to propagation.
Polymer microstructure is strongly influenced by a judicious choice of polar
modifier. The resulting polymer can be further treated with electrophiles to
prepare functional polymers. The polymerization process is straightforward,

although care must be given to purification of all reagents, and the polymer-
ization must be run in an inert atmosphere. A laboratory reactor setup for
preparative quantities of polymer has been described in the literature (29).
A. Initiation
Conventional organolithium species are highly associated in hydrocarbon
media, and the resulting aggregates are not very reactive in polymerization
(30). The aggregates are in equilibrium with less associated organolithium
species, which actually initiate most if not all of the polymerization (Fig. 9).
Conducting the polymerization in more polar solvents such as diethyl
ether or tetrahydrofuran (THF) increases the concentration of less associated
species and increases the reaction rate. Typically, however, small amounts of
polar compounds are added to the polymerization in nonpolar media to
achieve the same effect. These materials complex with the lithium to break up
the agglomerates. In ‘‘modified’’ polymerizations (polymerizations where a
small amount of a polar compound is added), most alkyllithium compounds
are suitable initiators, but for an unmodified polymerization secondary or
Figure 9 Aggregation of organolithium species.
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 65
Copyright © 2004 by Taylor & Francis
tertiary lithium compounds are required to rapidly initiate the polymeriza-
tion. This is because primary organolithium compounds such as n-butyl-
lithium are more associated than the secondary organolithium compounds
and thus are less reactive (31,32).
Functional organolithium reagents are used to make functional poly-
mers (33). This technique is generally better than functionalizing a living
polymer by reaction with an electrophile, because there are fewer side
reactions with initiation. The reactivity of the lithium portion of the initiator
requires that the functional group be protected in most cases, but the available
functionality is surprisingly diverse. The key issues with functional initiators

are storage stability and solubility in solvents suitable for polymerization.
Lithiated acetals (34) and lithiated trialkylsilyl ethers (35) are used to form
hydroxyl-terminated polymers after deprotection. Amine-terminated poly-
mers have proven to be more useful for the preparation of tire elastomers. The
synthetic routes diagrammed in Figure 10 can prepare these initiators.
The reaction of imine 1 with n-butyllithium produced initiator 2. SBR
was prepared with this initiator, but the number-average molecular weight
was much higher than predicted, which indicates that the alkyllithium
reaction with the imine produced less than 100% of 2 or that the initiator is
not completely efficient for initiation. The compounded SBR did exhibit
improved hysteresis compared to a butyllithium-initiated control (36,37). The
reaction of secondary amines with butyllithium seems like an easy way to
prepare n-lithium amides, but most of them are insoluble in nonpolar media.
Figure 10 Synthesis of lithium amide initiators. (From Refs. 36–38.)
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 66
Copyright © 2004 by Taylor & Francis
Cheng (38) prepared a series of simple secondary lithium amides, but in all
cases they were insoluble in hexane. The heterogeneous initiators were used to
polymerize dienes, but the polymerizations did not go to completion and the
resulting polymers most likely had a very broad molecular weight distribu-
tion. Lawson et al. (39a,39b) showed that preparation of lithium amides in the
presence of two equivalents of THF gave soluble initiators that could be used
to make a medium vinyl SBR at high conversion. The resulting polymer was
coupled with tin tetrachloride and showed a 40% reduction in hysteresis as
measured by tan y at 50jC compared to a butyllithium-initiated control
polymer. A partial list of the amide initiators studied and their solubilities is
given in Table 5.
Interestingly, although almost all of the amide initiators effectively
initiated polymerization, not all of the resulting polymers showed reduced

hysteresis on compounding.
N-Lithiohexamethyleneimine 3 and N-lithio-1,3,3-trimethyl-6-azabicy-
clo[3.2.1]octane 4 were studied further. They were both shown to be stable for
‘‘several days.’’ Initiator 4 produced polymers with a broader molecular
weight distribution than initiator 3 (40). One difficulty in working with these
initiators is that the amine group is lost during polymerization by the
mechanism shown in Figure 11. This reaction becomes more significant in
the presence of excess initiator and at temperatures above 80jC.
Initiators 5 and 6 (Fig. 12) can eliminate head group loss because the
additional carbon atom between the nitrogen and lithium prevents elimina-
tion (41).
The difficulty with the lack of solubility of simple lithium amides can be
overcome by in situ formation of the initiator. Immediately after charging a
reactor with solvent, monomer, randomizer (THF or potassium amylate),
and butyllithium, a secondary amine is added to the mixture. The amide is
made in situ, and high molecular weight polymers are formed that have lower
hysteresis than the corresponding polymers made with butyllithium. Approx-
imately 85–90% of the chains have amine head groups when this procedure is
used (42).
Tin-containing initiators are also important compounds used to prepare
high-performance tire rubbers. Addition of lithium metal to tributyltin
chloride in an ether solvent produces a solution of the desired initiator that
is filtered to remove lithium chloride (43) (Fig. 13). The initiator is stable at
room temperature and can be stored for approximately 8 weeks before a loss
in activity is observed. Polymer with a lower vinyl content and narrower
molecular weight distribution is obtained if the initiator is made in dimethyl
ether rather than THF. This is illustrated in Table 6 for the polymerization of
butadiene. Carbon black compounds based on these polymers have lower
hysteresis than corresponding unfunctionalized controls.
4871-9_Rodgers_Ch02_R2_052404

MD: RODGERS, JOB: 03286, PAGE: 67
Copyright © 2004 by Taylor & Francis
Table 5 Solubility and Effectiveness of Lithium Amide Initiators
Source: Ref. 39.
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 68
Copyright © 2004 by Taylor & Francis
B. Propagation
Propagation takes place at typical reaction temperatures (20–75jC) in inert
solvents such as hexane or benzene without chain transfer or termination. At
high temperature, however, the growing polymer chain can eliminate lithium
hydride, which stops the polymerization and broadens the molecular weight
distribution. The mechanism is shown in Figure 14.
Elimination of lithium hydride is a first-order process that yields a
polymer terminated with a diene. Addition of living polymer doubles the
molecular weight of the chain and provides an active site that can react with
additional butadiene to form a branched polymer (44).
The ratio of monomer to initiator has a major influence on the cis/trans
ratio in the homopolymerization of both butadiene and isoprene in unmod-
ified polymerizations, as shown in Table 7 (45,46). The higher the ratio of
Figure 12 Functional initiators to avoid head group loss.
Figure 11 Head group loss in functional polymers.
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 69
Copyright © 2004 by Taylor & Francis
monomer to initiator, the higher the cis/trans ratio produced with both
butadiene and isoprene.
Two kinetic factors affect the diene microstructure. The first involves
the relative rates of propagation versus isomerization of the initially formed
allyl anion. Monomer is inserted initially to form the allyl anion in the anti

form. If propagation is rapid, the microstructure of the penultimate unit will
be cis. If, however, the allyllithium has sufficient time to isomerize to the
thermodynamically more stable syn form, then the penultimate unit will be
trans. Thus, at a high monomer/initiator ratio that favors rapid propagation,
the microstructure is primarily cis. As the monomer is depleted and the
monomer/initiator ratio decreases, more trans microstructure will be formed
(Fig. 15) (47,48). The second factor is the relative rate of addition of monomer
to the syn or anti isomer. Butadiene will add approximately twice as fast to the
anti form as to the syn form. With isoprene the factor is eight times as fast (49).
In addition to increasing the rate of polymerization, polar solvents or
polymerization modifiers also affect the vinyl content and sequence distribu-
tion in polybutadiene, as shown in Table 8 (50,51).
Large amounts of weak complexing agents such as diethyl ether or
triethylamine must be used to significantly affect the microstructure, but
strongly chelating modifiers such as tetramethylethylenediamine (TMEDA)
or 1,2-dipiperidinoethane increase the vinyl content dramatically at low
levels. The effect of polymerization temperature and its interaction with
modifier is also illustrated by the data. Vinyl content is increased as the
temperature is reduced for all polymerizations, but the effect is more
pronounced at low modifier/lithium ratios.
In the copolymerization of styrene and butadiene, the sequence distri-
bution is strongly affected by the addition of polar modifiers or salts. In
Figure 13 Synthesis of tributyltin lithium.
Table 6 Polymerization of Butadiene with Tributyltin Lithium
Solvent–initiator makeup THF Dimethyl ether
T
g
(onset) À85jC À93jC
Vinyl content 21% 11%
M

n
223,000 206,000
M
w
/M
n
1.25 1.11
Source: Ref. 43.
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 70
Copyright © 2004 by Taylor & Francis
hydrocarbon solvents without polar materials, most of the butadiene will
polymerize first, followed by the styrene. This process is used to prepare
‘‘tapered’’ block polymers where there is a butadiene block, a mixed butadi-
ene–styrene block, and a styrene block (52).
Addition of polar compounds will randomize the styrene and increase
the rate of polymerization. Choice of modifier is critical to get the proper
degree of randomization and control the vinyl content. Modifiers such as
potassium tert-butyl alkoxide (t-BuOK) are used to randomize the styrene
without significantly increasing the vinyl content. At a ratio of t-BuOK/n-
BuLi of 0.1, there is only a small increase in vinyl content (Fig. 16), but this is
sufficient to randomize styrene in an SBR (53).
For higher vinyl SBR, a more powerful randomizer such as TMEDA is
used that produces high vinyl polymers at relatively low modifier/lithium
ratios (54). Very high vinyl SBR and polybutadiene can be prepared with a
modifier consisting of a mixture of TMEDA and an alkali metal salt of an
alcohol (55).
Table 7 Effect of Monomer/Initiator Ratio on Microstructure
Polymerization conditions Microstructure
Monomer Solvent Initiator

Monomer/
initiator Cis Trans 1,2 3,4
Butadiene Hexane Li 5
Â
10
4
0.68 0.28 0.04 N/A
17 0.30 0.62 0.08 N/A
Isoprene Cyclohexane Li >5
Â
10
4
0.94 0.01 — 0.05
15 0.76 0.19 — 0.05
Source: Refs. 45, 46.
Figure 14 Mechanism for branching in lithium polymerization. (From Ref. 44.)
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 71
Copyright © 2004 by Taylor & Francis
Figure 15 Microstructure formation during lithium polymerization. (Adapted
from Refs. 47 and 48.)
Table 8 Effect of Polar Modifiers on Polybutadiene Microstructure During
Lithium Polymerization
% 1,2-Addition at
Modifier Modifier/Li 30jC50jC70jC
Triethylamine 270 37 33 25
Diethyl ether 12 22 16 14
96 36 26 23
Tetrahydrofuran 5 44 25 20
85 73 49 46

Tetramethylethylenediamine 1.14 76 61 46
1,2-Dipiperidinoethane 1 99 68 31
10 99 95 84
Source: Refs. 50, 51.
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 72
Copyright © 2004 by Taylor & Francis
C. Termination
Termination is easily accomplished by reaction of the living polymer with an
electrophile. In early anionic polymerization studies, the electrophile was a
proton donor and termination resulted in a hydrocarbon polymer. Reaction
with other electrophiles such as carbon dioxide (carboxylic acid), sultones
(sulfonates), ethylene oxide (alcohol), or imines (amines) produce functional
polymers, but unless conditions are carefully controlled the functional poly-
mer is contaminated with other materials (56). Virtually every electrophile
known has been tested as a terminating agent for lithium polymerizations. In
one patent alone, the following were claimed for terminating a living trans-
polybutadiene polymerization—isocyanates, isothiocyanates, isocyanuric ac-
id derivatives, urea compounds, amide compounds, imides, N-alkyl-substi-
tuted oxazolydinones, pyridyl-substituted ketones, lactams, diesters,
xanthogens, dithio acids, phosphoryl chlorides, silanes, alkoxysilanes, and
carbonates (57), Amine- and tin-containing electrophiles provide the greatest
interaction with carbon black. Epoxy compounds and alkoxysilanes are most
beneficial for silica-filled compounds. The early work focused on termination
with amine-containing functional groups such as EAB [4,4V-bis-(diethylami-
no)benzophenone] (58–60). Black compounds made with these polymers
showed higher rebound, lower heat buildup, higher compound Mooney
Viscosity, and more bound rubber than the corresponding control rubber.
Another study by Kawanaka et al. (61) suggested that the mechanism of the
Figure 16 Effect of potassium butoxide/lithium ratio on polybutadiene micro-

structure. (n) Percent trans; (E) percent vinyl. (From Ref. 53.)
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 73
Copyright © 2004 by Taylor & Francis
rubber–filler interaction was through an iminium salt formed from the
reaction product of the amide and living polymer chain end (Fig. 17). The
authors inferred this because rubber functionalization with amides that could
not easily form iminium salts did not interact well with carbon black.
Termination with tin-containing compounds provides more flexibility
than with amine compounds. R
x
SnCl
y
(where x + y = 4) can be chosen to
give different levels of branching and thus assist in macrostructure control.
Phillips pioneered the coupling of solution polymers with tin halides to make
radial polymers in the 1960s but the Japanese Synthetic Rubber Company
(JSR) was the first to use the nature of the carbon–tin bond for tire com-
pounds. Tsutsumi et al. (62) outlined the synthesis of tin-coupled solution SBR,
the mechanism of how it improves hysteresis, structure–property relation-
ships to maximize the effect of tin, and pitfalls to avoid in compounding (62).
They first demonstrated that coupling solution SBR with tin tetrachloride
provided a superior polymer compared to other coupling agents (Table 9).
The SBR polymerization was terminated with tin tetrachloride such
that 50% of the chain ends were coupled. The only major difference in
performance among the coupling agents was the low hysteresis exhibited by
the tin-coupled polymer. Tsutsumi et al. compared a series of tin-coupled
polymers with a polymer containing trialkyltin groups along the backbone.
Only tin located at the end of the polymer chain (or branch point) was
effective in reducing hysteresis (Fig. 18).

Figure 17 Termination of lithium polymerization with a cyclic amide. (From Ref.
61.)
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 74
Copyright © 2004 by Taylor & Francis
In another study the same group showed that putting the tin group on a
butadienyl chain end was more effective in reducing compound hysteresis
than putting it on a styryl chain end. Finally, they postulated that the
mechanism of interaction with carbon black is by formation of a bond
between the polymer chain and the quinone groups on the carbon black.
This was based on a model study of the reaction of tributyltin-capped low
Figure 18 Effect of tin content position on dynamic properties of tin-coupled SBR.
(x) Polymer modified on backbone. (n) Polymer modified at chain end. (From Ref.
62.)
Table 9 Coupling of Solution SBR
a,b
Coupling agent
ML-1+4
(100jC)
Compounded
ML-1+4
(100jC)
Tensile
strength
(MPa)
Elongation
at break
Tan y
at 50jC
Tan y

at 0jC
None 54 93 22.3 400 0.121 0.235
Divinylbenzene 51 70 22.5 400 0.125 0.241
Diethyladipate 47 74 21.6 410 0.126 0.237
Silicon 57 89 23.5 400 0.126 0.240
Tetrachloride
Tin tetrachloride 57 76 25.0 400 0.096 0.239
a
Formulation (phr): Polymer 100, HAF black 50, zinc oxide 3, stearic acid 2, antioxidant 1.8,
accelerator 1.8, sulfur 1.5.
b
SBR: 24% bound styrene, 40% vinyl.
Source: Ref. 62.
4871-9_Rodgers_Ch02_R2_052404
MD: RODGERS, JOB: 03286, PAGE: 75
Copyright © 2004 by Taylor & Francis