Cross-linkers for improved high temperature performance of ROMP adhesives

Article abstract of DOI:10.1002/adsc.200600455

Ring-opening metathesis polymerization (ROMP) has been shown to be an effective methodology to make adhesive bonds between various substrates. Excellent adhesion has been demonstrated on a variety of substrates including post-vulcanized elastomers, polyol

Full text of DOI:10.1002/adsc.200600455

FULL PAPERS
DOI: 10.1002/adsc.200600455
Cross-Linkers for Improved HighTemperature Performance of
ROMP Adhesives
Christopher G. Keck,a Jonathan L. Kendall,b,c and Kenneth C. Castera,d,*
a
Lord Corporation, Materials Division, 110 Lord Drive, Cary, North Carolina 27511, USA
Lord Corporation, Chemical Products Division, 110 Lord Drive, Cary, North Carolina 27511, USA
Current address: Hanover Dental Company, P.O. Box 241301, Charlotte, North Carolina 28224, USA
Current address: Center for Biologically Inspired Materials and Material Systems, Duke University, Durham, North
b
c
d
Carolina 27708, USA
Fax : (+1)-919-549-4310; e-mail: ken.caster@duke.edu
Received: September 5, 2006
Abstract: Ring-opening metathesis polymerization using a 1808 peel test in a temperature range from
(ROMP) has been shown to be an effective method- À408C to 668C at various cross-linker concentra-
ology to make adhesive bonds between various sub- tions. Significant improvements in bond strength
strates. Excellent adhesion has been demonstrated were observed at elevated temperature with little
on a variety of substrates including post-vulcanized detrimental affect on low temperature strengths even
elastomers, polyolefins, and metals. Strong adhesion at high cross-linker concentration. Self-bonding of
results when one substrate surface is coated with the polyolefins assemblies was also investigated. Addi-
metathesis catalyst and the other substrate surface is tion of a peroxide radical initiator to the initial
coated with a ROMPable monomer. Concerns over ROMP formulation and cross-linking the bonded as-
using ROMP-based adhesion in dynamic applications sembly in a post-cure reaction was also found to lead
where the bond-line could experience higher-than- to improved high temperature bond strength. Syn-
ambient temperatures prompted this investigation. thetic details for the preparation of several cross-
In this report, a variety of bis-norbornadiene cross- linkers are reported.
linkers were used to bond tire carcass, tire tread,
post-vulcanized EPDM, and grit-blasted steel in a
variety of self- and cross-bonded configurations. The Keywords: adhesion; cross-linker; metathesis; perox-
resulting bonded assemblies were then evaluated ides; polymerization; ROMP
Introduction
portant materials like post-vulcanized rubber, EPDM-
based elastomers, polyethylene, and polypropylene.
The increasingly wide range of industrial products Finally, ROMP-cured adhesives contain olefins that
that could best be fabricated or repaired using struc- are suitable for secondary reactions within the adhe-
tural adhesives has created a continual need for new sive or between the adhesive and the substrate.
adhesive curing methods. These structural adhesives
ROMP has been found to be an effective polymeri-
must withstand exceptionally high stresses and ad- zation methodology to create adhesive bonds between
verse environmental conditions that are placed on the substrates. Strong primary adhesion has been demon-
adhesive joint. Adhesives that cure by ring-opening strated both with dispensed adhesives[1] and with sur-
metathesis polymerization (ROMP) possess several face-activated approaches. In the latter approach
particularly attractive traits compared to traditional (contact metathesis polymerization, CMP), a well-de-
commercial adhesives. Very fast cure speeds, resulting fined catalyst is applied directly to one and the mono-
in high product throughput, can be achieved with mer to the other substrate surface – adhesion occurs
well-defined catalysts such as Grubbsꢀ 1st and 2nd gen- when the two substrates come in direct contact. This
eration ruthenium catalysts. Another important ad- grafting-off process was investigated for cold bonding
vantage results from the hydrocarbon nature of most of pre- and post-vulcanized elastomers, polyolefin-
available ROMP monomers, which readily wet low filled elastomers, and metal-elastomer composites
surface-energy substrates. Wetting is thought to be a using Grubbsꢀ 1st generation ruthenium catalyst and
key factor in obtaining good adhesion to these diffi- dicyclopentadiene (DCPD), ethylidenenorbornene
cult-to-bond substrates, which include industrially-im- (ENB), and methylidenenorbornene (MNB) as mono-
Adv. Synth. Catal. 2007, 349, 165 – 174
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
165

Christopher G. Keck et al.
FULL PAPERS
mers.[2] Room temperature bond strengths were very and di-N-alkyldicarboxyimidonorbornenes of differ-
high, with rubber tear being the primary mode of sub- ent chain tether lengths;[20,21] cross-metathesis of am-
strate failure. Good bond strengths also resulted when phiphilic copolymers;[22] and the synthesis of novel
a similar, but slightly modified, process was used to cross-linking agents.[23,24]
adhere natural rubber elastomers to different multifil-
The patent literature has many examples where
ament fibers including nylon, polyester, and multifunctional ROMPable monomers have been
Kevlarꢂ.[3] This methodology also works well for gen- added to formulations to affect cross-linking and thus
erating coatings with excellent protective properties. improve properties. Examples include systems for
Others have used grafting-off, whereby polymeri- making storage-stable compositions that can be poly-
zation is initiated from a physisorbed initiator, to merized by action of actinic radiation or by short-
make polymer-coated pigment particles[4] and to func- term irradiation and subsequent thermal curing;[25]
tionalize poly(dimethylsiloxane).[5]
preparation of highly filled systems;[26] composites
Because adhesively-bonded industrial products, es- with sized reinforcement material;[27] fiber reinforced
pecially those using natural rubber and EPDM elasto- composites;[28] polymerizable compositions;[29] cross-
mers, are often subjected to high temperature, adhe- linking compositions;[30,31] dental compositions;[32] im-
sives for these purposes must have good high-temper- proving solvent resistance;[8,33] producing molded arti-
ature strength. We were concerned that adhesives cles;[34] making flame-retarding articles;[35] polymeriz-
based on the above-described technology might dis- ing low grade DCPD;[36] and improving heat distor-
play low bond strengths at high temperature above tion temperature.[37]
the glass transition temperature of the linear polymer.
To solve this problem we reasoned that cross-linking
the in situ generated ROMP polymer should improve Results and Discussion
high temperature strength.
Two approaches to address this high temperature Cross-Linker Synthesis and Solubility
stability problem were considered. Initially, we con-
centrated on building cross-links into the polymer A variety of structures can be envisioned to be effec-
network as the polymer propagated. The key to this tive ROMPable cross-linking agents. For ease of
approach is to use ROMPable monomers that will availability and synthesis, we chose to examine struc-
cross-link into the system and not significantly change tural types that all contained only two norbornene
the reaction rate. Cross-linker solubility was expected units per molecule. They were either available com-
to be important, so we focused on hydrocarbon based mercially, prepared from literature synthetic proce-
bis-norbornadiene systems as they should be more dures, or newly synthesized bis-norbornadiene sys-
soluble in the hydrocarbon ROMP monomer and tems.
compatible with its resultant polymer. In the other ap-
Initially, self-bonding of post-vulcanized EPDM
proach, cross-linking of the network by a different using the grafting-off process was used to test the
chemistry was performed after the ROMP. Both ap- effect of cross-linker on bonding performance as we
proaches resulted in improvements in high tempera- had considerable experience with this system.[2] In a
ture bond strength and are described below. Finally, typical procedure, Grubbsꢀ 1st generation catalyst was
we examined how adding cross-linking agents affected applied to the steel surface and a mixture of cross-
self-bonding of polyolefins.
linker in monomer applied to the elastomer. After
Multiple approaches have been reported to cross- mating the surfaces, loading the assembly with a
link ROMP polymers including post-cross-linking of weight of approximately 100 grams, and allowing the
methacrylate[6,7] and epoxide[7] side-chains which are assembly to stand at room temperature for at least
attached to the ROMP polymer backbone; by addi- 30 min, the assembly was subjected to a 1808 peel test
tion of peroxide initiators to the ROMP polymer on an Instron test apparatus and failure analysis per-
matrix;[8,9] by molecular recognition and self-assembly formed.
either by H-bonding;[10–12] or the formation of transi-
Adhesion is determined from the load-at-maximum
tion metal pincer complexes;[12] or through metathe- load value, which is related to bond strength. While
sis. energy-to-break (i.e., the area under the force-dis-
Numerous examples of metathesis-based cross-link- placement curve) relates to bond toughness and dis-
ing have been reported. Examples include polymeriz- placement-at-maximum load correlates to bond elas-
ing the secondary olefin in DCPD;[13] while others in- ticity, both measurements also reflect the elasticity of
volve the preparation of monoliths from functional- the bonded assembly. After Instron testing, the speci-
ized norbornenes and multicyclic dienes;[14–17] synthe- mens were inspected to determine the mode of fail-
sis of molecularly imprinted polymers from DCPD ure. This subjective analysis provides information
and functionalized norbornenes;[18] ligand exchange about the relative adhesive strength between the vari-
with MMA by cross-metathesis;[19] the use of mono- ous interfaces. The most desirable failure mode is
166
www.asc.wiley-vch.de
ꢁ 2007Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 165 – 174

Cross-Linkers for Improved High Temperature Performance of ROMP Adhesives
FULL PAPERS
rubber tear, where a thick layer of rubber elastomer and subsequent trapping of the intermediate benzyne
remains on one surface. Rubber tear indicates that with furan or freshly cracked cyclopentadiene. While
the adhesive is stronger than the elastomeric material some tetrabromides were commercially available, p-
and that failure occurs within the elastomer and not ethyltoluene and 1,4-diethylbenzene had to be con-
at the elastomer-adhesive interface. Other failure verted to their tetrabromoaromatics by bromination
modes are 1) adhesive failure (ad) which occurs be- over Fe. The diethyl compound was prepared on a
tween the elastomer and adhesive interface, 2) thin- 10” larger scale with no difficulties except that addi-
layer cohesive failure (tla) in which a thin layer of ad- tion of the benzene to the Br2/Fe mixture was exo-
hesive remains on one surface, 3) cohesive failure thermic so the reaction temperature had to be care-
occurs towards the center of the adhesive layer show- fully controlled during this step. Work-up involved re-
ing the adhesive interface to be stronger than the ad- moval of the excess Br2 followed by recrystallization
hesive itself, and 4) stock break occurs when the of the crude products from EtOH. Their structures
bonded substrate fails. Typically in our systems, deep were corroborated by 1H and 13C NMR analysis.
rubber tear was observed and was indicative of high
bond strength and very strong primary adhesion.
Starting with the appropriate tetrabromobenzene,
metal-halogen exchange was done under the usual
conditions[38,39] to generate the benzyne intermediate
that was trapped in situ with freshly cracked cyclopen-
tadiene or furan to give bis-norbornadienes 3–7 as
white to cream-colored solids (Scheme 1). 1H and
13C NMR confirmed the products to be of high purity
Our first experiments involved using commercially
available anthraquinone (1) as the cross-linker at two
different concentrations (0.55 and 2.5 mol%) in ENB
and Grubbsꢀ catalyst as described above. High bond
strength was observed when the assembly was peeled
at room temperature, but after 30 min at 708C, no im-
provement in high temperature strength was observed
as all bonded specimens, which included control speci-
mens that lacked cross-linker, could be peeled apart
by hand. Attempts to dissolve relatively polar anthra-
quinone 1 in hydrocarbon ENB were problematic, so
this result was not entirely unexpected.
Scheme 1. Formation of bis-norbornadiene cross-linkers by
trapping of in situ generated benzynes.
We expected that cross-linker solubility could be and to be a mixture of both syn- and anti-isomers.
tuned by making slight modifications of the bis-nor- Furan-derived compounds 3 and 4 were found to
bornene structure, and thus turned our attention to have low solubility in the ENB, MNB, and DCPD at
bis-norbornadienes 2, which are readily synthesized room temperature, but dissolved in all three at elevat-
by Diels–Alder cycloaddition of benzynes with cyclo- ed temperatures. Unfortunately, they came out of so-
pentadiene derivatives.[38] Structure 2 offers synthetic lution on cooling. Less polar derivatives (5, 6, 7) were
latitude in which polarity and symmetry can be ma- synthesized to allow control of hydrocarbon solubility
nipulated to adjust solubility while maintaining the (e.g., substitution on the aromatic ring to reduce sym-
high ROMP reactivity associated with the benzonor- metry, melting point, and increase hydrocarbon solu-
bornadiene system.
bility). They readily dissolved in hydrocarbon mono-
The cross-linkers were prepared by dilithiation of mers such as ENB at room temperature, but too
tetrabromobenzenes followed by elimination of LiBr showed differential solubility.
Adv. Synth. Catal. 2007, 349, 165 – 174
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
167

Christopher G. Keck et al.
FULL PAPERS
As more cross-linker was needed for these studies,
compound 6 was prepared following the procedure
worked out previously, but on a larger reaction scale.
However, this time, the white powder contained a
1
previously unobserved impurity. H and 13C NMR re-
vealed the main product to be bis-norbornene (6,
87%) and the impurity to be dibromobenzonorborna-
diene 8 (13%). Apparently, the reaction mixture was
not stirred long enough at low temperature for com-
plete lithiation of either the tetrabromobenzene or in-
termediate 8. The product was easily separated from
the intermediate by column chromatography.
In order to gain some insight into how other nor-
bornadiene-based systems might behave as cross-link-
ers, two other compounds were synthesized from 2,5-
norbornadiene (NBD). The first was 1,4-bis(2’,5’-nor-
bornadienyl)butane (9), which contains two NBD
units tethered with tetramethylene through one of
their olefins. This compound is made by treating
NBD with potassium tert-butoxide and n-butyllithium
followed by 1,4-dibromobutane.[40] Purification by
column chromatography (silica, hexanes) yielded a
colorless liquid that is miscible with the hydrocarbon
monomers used above.
The second cross-linking agent was made by the
UV irradiation of NBD in the presence of a
(Ph3)2Ni(CO)2 catalyst.[41] This reaction produced a
1
mixture 10, which was shown by H NMR spectrosco-
py to contain three dimers and one trimer in a grit-blasted steel was also evaluated. In general, good
22:5:1:4 mole ratio. Other products were present at performance was observed for nearly all conditions
<1%. Because we were unsuccessful in separating investigated.
the products by sublimation, as reported in the litera-
ture, we used the mixture for bonding experiments.
The mixture was soluble at least up to 10 w/w% in Tire Carcass and Tread Stock Bonding
ENB.
The initial studies on cross-bonding carcass-to-tread
looked at temperature using cross-linkers 5, 6, and 7
at slightly different concentrations in ENB (2.8 to 7.9
Variable Temperature Bond StrengthEvaluation
mol%). Figure 1 shows that at À408C, the control
As described earlier, we know that adhesive bonds (ENB without cross-linker) and the samples that con-
produced by grafting-off of surfaces using ROMP tained cross-linker 6 had virtually the same load –
chemistry performed very well at room temperature. they stretched on the Instron during testing whereas
However, many industrial applications require that an cross-linker 7 samples showed higher strength and
adhesive bond must be able to survive wide tempera- stock failure. Unfortunately, no low temperature data
ture swings and cycling. We wanted to ascertain how were obtained for cross-linker 5 as this may have pro-
well a rudimentary cross-linked ROMP adhesive vided insight whether the higher bond strength was
would perform at different temperatures in both self- simply a function of cross-linker concentration. Signif-
and cross-bonded configurations. Tire carcass and tire icantly, at these high cross-linker concentrations, the
tread stocks are industrially important substrates and adhesive bond would have likely been thought to be
were chosen for most of the cross-bonding work. brittle. However, the bonded assembly remained elas-
Three different cross-linkers (5, 6, 7) were examined tic. The higher bond strength observed at the lower
at different concentrations (6, 7) at temperatures temperature may be explained by increased stiffness
ranging from À408C to 668C. Cross-bonding of these of the elastomer. The lack of joint failure is a more
stocks to grit-blasted steel was examined to develop significant finding. As expected, all four specimens
understanding of how different elastomeric materials showed stock failure at room temperature. At 668C,
performed using grafting-off ROMP. Self- and cross- incorporation of all three cross-linkers greatly in-
bonding of post-vulcanized EPDM to itself and to creased bond strength. The control showed no rubber
168
www.asc.wiley-vch.de
ꢁ 2007Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 165 – 174

Cross-Linkers for Improved High Temperature Performance of ROMP Adhesives
FULL PAPERS
ure was observed at a lower concentration (3.4
mole%). However with cross-linker 6, bonded sam-
ples showed no rubber tear at 0.5 mole%, whereas at
2.8 mole%, stock failure was observed. While two dif-
ferent cross-linkers were used in this study, Figure 2
clearly shows trends in cross-linker performance: a)
at room temperature, bond strength decreases at very
high cross-linker concentration and b) at elevated
temperature, bond strength increases to a maximum
and then decreases again.
In order to further our understanding of the me-
chanical performance of ROMP-based adhesive sys-
tems and to add to our elastomer bonding findings re-
ported earlier,[2] elastomer-to-metal bond strength
data were also obtained for both carcass and tread
stocks; however, no cross-linkers were used in this
study. Unsanded and sanded versions of both carcass
and tread stocks were bonded to grit-blasted steel
coupons with Grubbsꢀ 1st generation ruthenium cata-
lyst and ENB (Table 1). The carcass-to-metal speci-
Figure 1. Comparison of bond strength for three different
cross-linkers for bonding tire carcass-to-tire tread.
Table 1. Elastomer-to-metal bonding for unsanded and
sanded carcass and tread stocks.
tear while the samples containing each cross-linker
showed stock failure.
Elastomer Sanded Load at Max. Energy to
Failure
mode[b]
Load (N)[a]
Break-Point
Cross-linker concentration was also examined
(Figure 2). In these studies, we tried to push the limits
of cross-linker solubility in ENB to see how wide
changes in concentration affected bond strength.
Each cross-linker sample exhibited stock failure at
room temperature, independent of concentration.
However, at 668C, failure was concentration-depen-
dent. Cross-linker 7 samples showed no rubber tear at
very high concentration (19.6 mole%), but stock fail-
(J)[a]
Carcass
Tread
No
No
120
280
5.2
40.5
adh
stock
break
adh
Carcass
Tread
Yes
Yes
80
214
3.2
24.6
adh
[a]
Average result for three specimens.
Notations: rt, rubber tear; tlc, thin-layer cohesive; adh,
adhesive.
[b]
mens showed primarily adhesive failure to the metal
side with only a small amount of rubber tear seen on
one sample. All tread-to-metal specimens showed ex-
cellent rubber tear. In fact, the tread broke before
peeling fully apart in all three samples. Bond
strengths were lower for the sanded surfaces for both
elastomers. The carcass-to-metal specimens showed
only a very small amount of rubber tear along the pe-
rimeter of the coupons. Otherwise, the polymer film
was left on the carcass surface after pulling apart. Sur-
prisingly, the tread-to-metal specimens produced simi-
lar results – rubber tear was somewhat greater along
the perimeter of these coupons but greatly reduced
when compared with the unsanded tread. The lower
bond strengths observed with the carcass are expected
as this elastomer is considerable softer than the tread.
Figure 2. Effect of concentration on bond strength at elevat-
ed temperature for bonding tire carcass-to-tire tread.
Adv. Synth. Catal. 2007, 349, 165 – 174
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
169

Christopher G. Keck et al.
FULL PAPERS
EPDM Bonding
used to kick off the peroxide initiator and further
cross-link the poly
The results were very different when cross-linker 6 mance.
dissolved in ENB was used to self-bond EPDM using
A
While DCPD is known to cross-link during meta-
Grubbsꢀ 1st generation catalyst. At room temperature, thesis through a secondary reaction of the cyclopen-
bond strength was essentially unaffected by cross- tenyl ring,[13] we wanted to determine whether post-
linker (196 N, 0.5 mol%) vs the control (178 N). Simi- cure cross-linking could be used to improve high tem-
larly at 668C, no improvement in bond strength was perature adhesion using monomers that give linear
observed for the cross-linked systems (107N, 0.5
polymers and that do not readily undergo secondary
mol%; 98 N, 2.8 mol%) vs. the control (98 N). How- reaction. Bonding experiments were done to deter-
ever, there was an increase in bond strength when mine adhesion performance for specimens where
cross-linker 6 (0.5 mol%) was used to cross-bond cross-linking was performed by post-reaction using
EPDM-to-grit-blasted steel. Compared to the control radical initiators that had been added to ENB prior
(285 N), a slight improvement in bond strength was to ROMP with Grubbsꢀ 1st generation ruthenium cat-
seen even at low levels of cross-linker (320 N, 0.5 alyst.
mol%) at room temperature. All samples displayed
Before attempting to bond any assemblies however,
good rubber tear; in fact, some of the cross-linker 6 effort was spent to identify the radical initiator and
samples showed stock failure. There was an even reaction conditions that would be used during the
more dramatic effect at 668C. Here, the control sam- post-bake cycle. In these tests, mixtures of ENB and
ples (85 N) showed little rubber tear while samples Wako azo radical initiator V-65 (T1/2 =518C) were
formulated with cross-linker 6 (133 N) showed good heated under N2 in a glass vial for different time peri-
rubber tear.
ods. Changes in viscosity were used as evidence of re-
action. Unfortunately, no polymerization was noted
under any conditions for any reaction mixture [at
608C for 5 h at 1 w/w% V-65 in ENB and 9.2 w/w%
V-65 in ENB; 1108C for 2 h at 5 w/w% V-65 in
Polyolefin Adhesion
The tethered cross-linker 1,4-bis(2’,5’-norbornadie- ENB]. In each case a light-yellow liquid remained
nyl)butane 9 was used neat to bond unsanded poly- with no apparent polymerization.
propylene coupons as lap shear samples. This com-
The initiator was tested for activity by heating a 1
pound took several times longer to cure than the w/w% solution of V-65 in Cyclol Methacrylateꢂ 11 at
ENB control. However, it gave the same mean break 1108C under N2. Indeed, the solution bubbled into a
values at room temperature: 302Æ40 N versus 294Æ hard mass within a few minutes after placing it into
7 6 N.
the heated atmosphere thus demonstrating that the
The NBD dimmer/trimer mixture 10 was used to initiator was indeed active. While disconcerting that
bond sanded polypropylene lap shears assemblies at a no reaction appeared to occur with ENB alone, it was
series of concentrations. Compared with the control decided that poly
A
specimen (1548Æ169 N), a 17% improvement in
ENB as this would represent more realistic bonding
room temperature bond strength was observed for conditions. A polymer solution was then prepared
samples that contained cross-linker (2 w/w%, 1806Æ from ENB and Grubbsꢀ catalyst in toluene (M:I=
214 N; 5 w/w%, 1802Æ102 N; 10 w/w%, 1810Æ142 3000:1). Once the polymer had been prepared, the V-
N). The two NBD-based cross-linkers have not yet 65 was added to an aliquot of solution and heated at
been tested at elevated temperatures.
908C for two hours under N2. Again, no polymeri-
zation appeared to occur as evidenced by lack of in-
soluble product.
Post-Cross-Linking Reactions
An explanation for this apparent lack of reactivity
likely involves chain-transfer from ENB or poly-
As described above, examples of cross-linking ROMP
A
polymers in a post-polymerization reaction have been mation of allylic and doubly allylic radical species
reported.[8] Improved solvent resistance was observed through radical abstraction reactions (Scheme 2). In
by thermally post-curing a poly(DCPD) resin which addition to chain-transfer limiting the cross-link den-
U
contained a peroxide [e.g., Luperoxꢂ 130; 2,5-dimeth- sity, the stabilized allylic radicals would be less reac-
yl-2,5-di-(tert-butylperoxy)hex-3-yne] that had been tive than normal alkyl radicals.
added to the initial monomer/catalyst mixture prior to
If the above hypothesis is operative, then other rad-
ROMP. Higher cross-link densities likely resulted ical initiators should also be unsuccessful. However as
from this treatment than would occur by metathesis described above,[8] the peroxide initiator Luperoxꢂ
only. In practice, once the ROMP polymer has 130 has been shown to be effective at cross-linking
formed, then a separate post-bake cycle would be bulk poly(DCPD). Thus, freshly grit-blasted steel cou-
G
170
www.asc.wiley-vch.de
ꢁ 2007Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 165 – 174

Cross-Linkers for Improved High Temperature Performance of ROMP Adhesives
FULL PAPERS
Conclusions
Ring-opening metathesis polymerization was used to
investigate cross-linking and its affect on adhesion at
different temperatures. Surface-initiated ROMP, a
grafting-off process where the well-defined catalyst
was applied to one substrate surface and the mono-
mer/cross-linker to the other substrate surface, was
used to assemble self- and cross-bonded specimens
constructed from tire carcass and tread rubber stocks,
post-vulcanized EPDM, grit-blasted steel, and poly-
propylene. Cross-linkers evaluated in this study were
based on bis-norbornadienes and were synthesized
using cycloaddition reactions and literature syntheses.
Cross-linker solubility in the ROMP monomer was
found to be a critical variable as too low a concentra-
tion did not result in improve performance. Except
for self-bonded EPDM, significant improvements in
Scheme 2. Abstraction of some of the allylic and doubly-al-
lylic hydrogens on poly(ENB) by a radical initiator.
AHCTREUNG
pons were bonded to post-vulcanized EPDM using bond strength were observed for bonded specimens at
Grubbsꢀ 1st generation catalyst and a mixture of ENB elevated temperatures with little detrimental affect on
that contained 2.9 w/w% (1.2 mol%) of Luperoxꢂ low temperature strengths even at high cross-linker
130. Thirty minutes after bonding, the specimens were concentration. No decrease in room temperature
post-cured in an oven at 1208C for 1.5 h (purged with bond strength was observed compared with control
N2). The specimens were evaluated on the Instron specimens, which lacked cross-linker. In a second ap-
using the 1808 peel test at room temperature and proach, post-vulcanized EPDM was bonded to grit-
668C (Figure 3). These data show increased bond blasted steel, however here, a peroxide radical initia-
tor was added to the monomer and the bonded as-
sembly cross-linked in a post-cure reaction. This re-
sulted in improved high temperature bond strength.
The findings reported herein show that improved
bond strength performance can be expected by using
either in-situ or post-cross-linking chemistries. These
approaches demonstrate the latitude available to the
formulator for controlling cross-linking chemistry and
for matching application/performance criteria for a
new product.
Experimental Section
General Experimental Details
1H and 13C NMR spectra were recorded on a Bruker
250 MHz NMR spectrometer at 250.13 MHz and 62.9 MHz,
respectively. All chemical shifts (d) are positive and refer-
enced downfield from tetramethylsilane (TMS); coupling
Figure 3. Comparison of bond strength for post-vulcanized
EPDM-to-Grit-Blasted Steel using metathesis and peroxide
constants (J) are recorded in Hz. Grubbsꢀ 1st generation cat-
alyst was purchased from Boulder Scientific and used as re-
ceived. Dicyclopentadiene, ethylidenenorbornene (CaH2),
and methylidenenorbornene were purchased from Aldrich
and distilled using standard practices from the reagent speci-
fied. CHCl3 and CH2Cl2 were obtained from Fisher or
Pharmco and used as received. Diethyl ether was distilled
from sodium benzophenone ketyl. Reduced carbonyl iron
powder (R2430) was purchased from ISP Technologies, Inc.
Wako azo radical initiator V-65 and Luperoxꢂ 130 were ob-
tained from Wako Chemicals USA, Inc. and Elf Atochem
North America Inc., respectively. Cyclol Methacrylateꢂ 11
was obtained from Monomer-Polymer & Dajac Labs, Inc.
cross-linkers.
strength for the post-baked specimens over both con-
trol and cross-linker 6 bonded specimens at both
room and elevated temperatures. Failure analysis,
however. showed small amounts of metal-adhesive
and rubber-adhesive modes to be present, which were
absent with the bis-norbornadiene cross-linker 6.
Such mixed failure modes are common.
Adv. Synth. Catal. 2007, 349, 165 – 174
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
171

Christopher G. Keck et al.
FULL PAPERS
Cross-linkers 3,[39] 4,[39] 9,[40] and 10[41] were prepared follow-
ing published procedures. Combustion analyses were per-
formed by Galbraith Laboratories. EPDM elastomer (BTR
96616) was obtained from British Tyre and Rubber under
the designation 96616 and was molded into strips
(149.4 mm”25.4 mm”3.2 mm) and cured at 1548C for
9 min. Tire carcass and tread stocks were obtained from B.F.
Goodrich. For bonding experiments, the elastomer strips
were washed with acetone and allowed to dry for several
minutes prior to application of catalyst or monomer. Grit-
blasted steel coupons were prepared by blasting 1010 fully
hardened cold rolled steel (60.3 mm”25.4 mm”1.6 mm)
with GS40 steel grit and were obtained from Northcoast
Tool, Erie, PA. They were washed with acetone and dried
prior to application of catalyst solution. Older unused metal
coupons were freshly grit-blasted with silica sand (grit size
30–100; obtained from McMaster Carr, Inc.) prior to bond-
ing. Spray application of catalyst was done using a Badger
hobby art sprayer and Badger PROPEL (1,1-difluoroethane
and butane) as propellant. Mechanical testing was per-
formed on an Instron Model 4204 Materials Tester equipped
with a 5 kN load cell.
(1”50 mL), and dried under vacuum to give 23.29 g of gray
powder. This was recrystallized from 700 mL of EtOH. The
solid was isolated by vacuum filtration, washed with cold
EtOH, and dried under vacuum to give the product as give
tan, fluffy needles; yield: 22.84 g (56%); mp 104–1058C;
1H NMR (CDCl3): d=1.15 (t, 3H), 2.78 (s, 3H), 3.22 (q,
2H); 13C NMR (CDCl3): d=12.0, 28.9, 35.3, 126.9, 127.8,
139.1, 143.9.
Bromination of 1,4-Diethylbenzene to give 1,4-
Diethyl-2,3,5,6-tetrabromobenzene
The procedure was equivalent as for the tetrabromotoluene
above, except that 7.8 mL (6.72 g, 0.050 mol) of 1,4-diethyl-
benzene were used. This gave 23.29 g of gray powder after
washing. This was recrystallized from 600 mL of EtOH. The
solid was isolated by vacuum filtration, washed with cold
EtOH, and dried under vacuum to give the product as tan,
shiny, needles; yield: 19.46 g (87%); mp 114.5–115.5 8C;
1H NMR (CDCl3): d=1.15 (t, 6H), 3.22 (q, 4H); 13C NMR
(CDCl3): d=12.0, 35.4, 127.3, 144.1.
Synthesis of Cross-Linker 6
Mechanical Testing of Bonded Assemblies
A flame-dried, 250-mL, round-bottom flask, which was
fitted with a thermometer, rubber septum, reflux condenser
with gas adapter, and stirring bar under an argon flow, was
charged with 5.00g (0.011 mol) of diethyltetrabromobenzene
and 130mL of diethyl ether. The clear, light-yellow solution
was cooled to À628C in a dry-ice bath at which point the
bromobenzene precipitated. To the cooled reaction mixture
was charged 9.1mL (7.30g, 0.110 mol, 9.9 equivs.) of freshly
distilled cyclopentadiene. Finally, 9.5mL (0.0237mol,
2.1 equivs.) of 2.5M n-BuLi solution in hexanes was charged
into the reaction flask drop-wise over a 40min period using
a gas-tight syringe. The reaction mixture was slowly brought
to room temperature and stirred overnight. The reaction
mixture was quenched with 1.5mL of MeOH and vacuum
filtered to remove undissolved salts. The filtrate was washed
with deionized, distilled water (3”25mL) and the organic
phase separated and dried with MgSO4 and vacuum filtered.
The clear-yellow filtrate was rotovaped at 398C under parti-
al vacuum to give a light-yellow, solid residue which was
dried under vacuum overnight. The residue was washed with
cold MeOH (3”5mL) and then dried under vacuum to give
cross-linker 6 as a fluffy, white powder; yield: 2.33g (80%);
mp 139–1498C; 1H and 13C NMR confirmed that the product
had been synthesized in high purity and that it was a mix-
ture of both syn and anti-isomers. 1H NMR (CDCl3): d=
1.19 (6H), 2.30 (4H), 2.78 (4H), 4.02 (4H), 6.88 (4H);
13C NMR (CDCl3): d=16.5, 23.1, 47.9, 69.6, 69.7, 129.8,
129.9, 143.3, 143.4, 145.9. Sublimation at 808C/0.4 Torr gave
shiny, white crystals; anal. calcd. for C20H22: C 91.55, H 8.45;
found: C 91.15, H 8.50.
Primary adhesion of bonded specimens was measured for
self- and cross-bonded specimens by pulling them on an Ins-
tron in a 1808 peel configuration according to ASTM-D 429
Method B. A typical test specimen was constructed from a
grit-blasted steel coupon of dimensions 60.3 mm”25.4 mm”
1.6 mm bonded to a strip of elastomer 149.4 mm”25.4 mm”
3.2 mm. Samples were pulled using a crosshead speed of
50.8 mm/minute. The load-at-maximum load, energy-to-
break, and displacement-at-maximum load were measured.
Load-at-maximum load correlates to bond strength, energy-
to-break (the area under the force-displacement curve) re-
lates to bond toughness, and displacement-at-maximum load
correlates to bond elasticity.
Bromination of p-Ethyltoluene to give 1-Ethyl-4-
methyl-2,3,5,6-tetrabromobenzene
A 250-mL, 3-neck, round-bottom flask was fitted with a
rubber septum, thermometer, reflux condenser with gas
adapter, and a stirring bar. The apparatus was connected to
a dry ice cold-finger and bubbler to trap Br2 vapor and was
charged with 1.40 g (0.025 mol, 0.5 equivs.) of iron powder.
The flask was cooled to 08C in a 2-propanol/dry-ice bath
and then charged with 49 mL (152 g, 0.95 mol, 19 equivs.) of
Br2. To the stirring mixture was added 7.0 mL (6.027 g, 0.050
mol) of p-ethyltoluene via syringe over a 2 h period. The re-
action temperature was not allowed to increase above 68C.
After addition, the reaction mixture was stirred at 08C for
1 h and then slowly brought to room temperature. The
rubber septum on the flask was replaced with a glass tube
connected to a N2 tank. Inert gas was then bubbled into the
flask to sweep out excess Br2. The flask was immersed in an
oil bath at 358C in order to facilitate Br2 evaporation. After
2 d of gas sweep at elevated temperature, a dark-red to
black residue remained. As this residue was stirred and
broken-up, Br2 gas was given off and eventually became a
powder, which was washed with deionized distilled H2O (4”
50 mL), NaHCO3 (2”30 mL), with deionized distilled H2O
Isolation of 6,7-Dibromo-5,8-diethylbenznorborna-
diene (8)
During the synthesis of compound 6 on the 50-g scale, a
new compound that was later identified as dibromobenzo-
norbornene 8 (13%) was isolated. Purification either by
flash column chromatography (100% hexanes) or by subli-
mation (758C/0.3 Torr) was effective. Recrystallization from
172
www.asc.wiley-vch.de
ꢁ 2007Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 165 – 174

Cross-Linkers for Improved High Temperature Performance of ROMP Adhesives
FULL PAPERS
EtOH to give analytically pure shiny, white crystals: mp 70–
718C; H NMR (CDCl3): d=1.2 (t, 6H), 2.2 (dd, 2H), 2.8
allowed to cure for 24 h. A set of five lap shear samples
were prepared for each monomer/cross-linker test.
1
(q, 4H), 4.0 (s, 2H), 7.8 (s, 2H); anal. calcd. for C15H16Br2:
C 50.59, H 4.53; found: C 50.89, H 4.77.
Acknowledgements
Synthesis of Cross-Linker 7
The procedure was equivalent as for cross-linker 6, except
that 5.00 g (0.0115 mol) of the tetrabromotoluene and
9.5 mL (7.62 g, 0.115 mol, 10.0 equivs.) of cyclopentadiene,
and 9.5 mL (0.0237mol, 2.1 equivs.) of 2.5 M n-BuLi in hex-
anes were used. Work-up as described above gave com-
pound 7 as a cream-colored powder; yield: 1.62 g (57%);
mp 139–1498C; 1H and 13C NMR confirmed that the product
had been synthesized in high purity and that the product
was a mixture of both syn and anti-isomers. Recrystallization
We thank the Lord Corporation Executive Officers and
Board of Directors for supporting this work. R.D. Walls and
M. E. Hontz are acknowledged for their suggestions and help
with mechanical testing during the course of this work. Pro-
fessor K. B. Wagener (University of Florida) provided en-
couragement and helpful feedback during the course of this
work.
1
gave one isomer: mp 84–858C; H NMR (CDCl3): d=1.18
(3H), 2.28 (7H), 2.75 (2H), 3.99 (4H), 6.85 (4H); 13C NMR
(CDCl3): d=14.8, 16.5, 23.1, 47.9, 48.0, 48.2, 69.8, 123.0,
123.1, 129.6, 129.7, 143.2, 143.4, 145.8, 146.6; anal. calcd. for
C19H20·0.2 H2O: C 90.57, H 8.28; found: C 90.69 H 8.04.
References
[1] a) J. L. Kendall, K. C. Caster, (Lord Corporation), US
Patent 6,800,170, 2004; b) J. L. Kendall, K. C. Caster,
(Lord Corporation), US Patent 6,872,792, 2005.
[2] K. C. Caster, E. F. Tokas, C. G. Keck, M. E. Hontz, J.
Mol. Catal. A: Chem. 2002, 190, 65.
General Procedure for Preparing Cross-Linker
Formulations and Bonding by CMP
[3] K. C. Caster, R. D. Walls, Adv. Synth. Catal. 2002, 344,
764.
Cross-linker and ENB were mixed at a specified concentra-
tion or series of concentrations and applied to the substrate
surface. Catalyst was applied directly to the other surface to
be bonded using the brush procedure[2] which resulted in an
average of 5.8Æ1.8 mg of catalyst per elastomer strip or
coupon. The substrates were mated, placed under a 100 g
weight and allowed to stand at room temperature for at
least 30 min to overnight. The mated assembly was then
heated in an oven at a specified temperature (i.e., 708C for
30 min) and then subjected to a 1808 hand peel while hot. If
the assembly could not be pulled apart, it was then pulled
on the Instron as described above. At least 3–5 assemblies
were bonded for each test.
[4] J. H. Chen, F. E. Schubert, (Copytele, Inc.), US Patent
6,194,488, 2001.
[5] S. Hu, W. J. Brittain, Macromolecules 2005, 38, 6592.
[6] B. R. Maughon, R. H. Grubbs, Macromolecules 1996,
29, 5765.
[7] B. R. Maughon, T. Morita, C. W. Bielawski, R. H.
Grubbs, Macromolecules 2000, 33, 1929.
[8] R. H. Grubbs, C. S. Woodson, (California Institute of
Technology), US Patent 5,728,785, 1998.
[9] C. Liu, S. B. Chun, P. T. Mather, L. Zheng, E. H. Haley,
E. B. Coughlin, Macromolecules 2002, 35, 9868.
[10] U. Drechsler, R. J. Thibault, V. M. Rotello, Macromole-
cules. 2002, 35, 9621.
[11] L. P. Stubbs, M. Weck, Chem. Eur. J. 2003, 9, 992.
[12] J. M. Pollino, K. P. Nair, L. P. Stubbs, J. Adams, M.
Weck Tetrahedron 2004, 60, 7205.
Cross-Linker Formulation Conditions – Preliminary
Bonding Results
[13] T. A. Davidson, K. B. Wagener, D. B. Priddy, Macromo-
Anthraquinone (1): No improvement in high temperature
strength, as determined by hand peeling the bonded speci-
mens after being placed in an oven at 708C for 30 min, was
observed at either a 40:1 ENB/catalyst molar ratio (2.5
mole% catalyst) or a 180:1 ENB/catalyst molar ratio (0.55
mole% catalyst).
lecules 1996, 29, 786.
[14] F. Sinner, M. R. Buchmeiser, Macromolecules 2000, 33,
5777.
[15] M. R. Buchmeiser, Macromol. Rapid Commun. 2001,
22, 1081.
[16] A. D. Martina, L. Garamszegi, J. G. Hilborn, J. Poly.
Sci.: Part A: Polym. Chem. 2003, 41, 2036.
[17] M. R. Buchmeiser, N. Atzl, G. K. Bonn, J. Am. Chem.
Soc. 1997, 119, 9166.
Polyolefin Bonding
Polypropylene lap shear samples were prepared from
101.6 mm”25.4 mm”3.2 mm coupons according to the fol-
lowing procedure. If the polypropylene was sanded, 100-grit
sandpaper was used to lightly roughen the bonding area of
the lap shear samples. A solution of 200 mg of Grubbsꢀ 1st
generation catalyst in 15 mL of dichloromethane was
sprayed onto the 645 mm2 bonding area of 10 polypropylene
coupons. After the solvent was dry, about 3.5 to 4.0 mg of
catalyst had been delivered to each coupon. About 150 mL
of monomer/cross-linker were placed on the conjugate
coupon, the catalyst-containing and the monomer/cross-
linker-containing coupons were mated, and the adhesive was
[18] A. Patel, S. Fouace, J. H. G. Steinke, Chem. Commun.
2003, 88.
[19] D.-J. Liaw, J.-S. Tsai, Pi-Ling Wu, Macromolecules.
2000, 33, 6925.
[20] E. Khosravi, Macromol. Symp. 2002, 183, 121.
[21] P. J. Hine, T. Leejarkpai, E. Khosravi, R. A. Duckett,
W. J. Feast, Polymer 2001, 42, 9413.
[22] K. Breitenkamp, T. Emrick, J. Am. Chem. Soc. 2003,
125, 12070.
[23] H. R. Allcock, W. R. Laredo, C. R. deDenus, J. P.
Taylor, Macromolecules. 1999, 32, 7719.
Adv. Synth. Catal. 2007, 349, 165 – 174
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
173

Christopher G. Keck et al.
FULL PAPERS
[24] C. G. Coleman, T. J. McCarthy, Polymer Preprints 1988,
28, 283.
[25] P. A. Van Der Schaaf, A. Hafner, A. Mꢃhlebach, (Ciba
Specialty Chemicals Corporation), US Patent 6,093,779,
2000.
[26] F. Setiabudi, A. Mꢃhlebach, Y. Naganuma, (Ciba Spe-
cialty Chemicals Corporation), US Patent 6,100,323,
2000.
[32] P. Bissinger, (Espe Dental AG), US Patent 6,075,068,
2000.
[33] R. J. Mionchak, P. C. Lane, (B. F. Goodrich), US Patent
4,701,510, 1987.
[34] S. Hara, U. Nakatani, (Hercules Inc.), US Patent
5,069,943, 1991.
[35] M. Warner, M. A. Giardello (A. O. Smith Corp.), US
Patent 6,071,459, 2000.
[36] C. S. Woodson, R. H. Grubbs (Advanced Polymer
Technologies, Inc.), US Patent 6,020,443, 2000.
[37] A. S. Matlack, (Hercules Corp.), US Patent 4,703,098,
1987.
[27] M. W. Warner, S. D. Drake, M. A. Giardello, (A. O.
Smith Corporation), US Patent 6,040,363, 2000.
[28] F. Setiabudi, T. Kainmuller, (Ciba Specialty Chemicals
Corp.), US Patent 5,840,238, 1998.
[29] A. Mꢃhlebach, P. A. Van Der Schaaf, A. Hafner, (Ciba
Specialty Chemicals Corp.), US Patent 6,162,883, 2000.
[30] K. W. Scott, N. Calderon, (Goodyear), US Patent
4,049,616, 1977.
[38] K. C. Caster, C. G. Keck, R. D. Walls, J. Org. Chem.
2001, 66, 2932.
[39] K. Shahlai, S. O. Acquaah, H. Hart, OrganicSyntheses
Coll. Vol. 10, 2004, 67 8.
[40] Adapted from: C. Yip, S. Handerson, R. Jordan, W.
Tam, Org. Lett. 1999, 1, 791.
[41] P. W. Jennings, G. E. Voecks, D. G. Pillsbury, J. Org.
Chem. 1975, 40, 260.
[31] A. Mꢃhlebach, A. Hafner, P. A. Van Der Schaaf, (Ciba
Specialty Chemicals Corporation), US Patent 5,973,085,
1999.
174
www.asc.wiley-vch.de
ꢁ 2007Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 165 – 174