Adhesion phenomena of BIMS.

Terpolymers of isobutylene, p-methylstyrene and p-bromomethylstyrene are a new generation of isobutylene-based elastomers. This polymer, denoted as BIMS throughout this article, has enhanced thermal stability and resistance to oxidative attack because of its saturated backbone. Three different studies involving the adhesion of BIMS and its compounds will be discussed. The first study was performed to understand the adhesion behavior of blends of BIMS polymer and tackifier(s) for pressure sensitive adhesive (PSA) applications. Hubbard et al., and Briddell and Hubbard have shown that useful adhesives in either solvent or cured tape form can be formulated using BIMS in combination with block copolymers (refs. 1 and 2). Peel strength values at 70[degrees]C after aging seven days at 70[degrees]C ranged from 0.58 to 1.02 kN/m for compositions which included BIMS, whereas comparable compositions containing conventional butyl and halobutyls had values from 0.07 to 0.33 kN/m. Our work focused on the behavior of BIMS as the sole polymer. The second and third studies were carried out to understand the adhesion of BIMS to a polar substrate, nylon 6, and to another elastomer, isoprene rubber (IR), respectively. The effect of benzyl bromide (BzBr) groups in BIMS on adhesion to nylon and IR will be discussed in this work.

A previously established adhesion model for PSA and hot melt adhesive (HMA) is used to understand all the above adhesion phenomena (refs. 3 and 4). This model states that:

P = [P.sub.0] BD (1) where [P.sub.0] is the intrinsic (or interfacial) adhesion term between the adhesive and the substrate surface. This term is high if interfacial diffusion or chemical bonding occurs. In the absence of chemical bonding, this term is maximized when the adhesive and the substrate have similar surface characteristics. The B term is the bonding term, which characterizes the bonding process and represents the bonding capability of the adhesive to the substrate. It depends on the bonding time and the bonding pressure of a particular adhesion test and also on the storage modulus of the adhesive. The D term is the debonding term, which is the rheological contribution in the debonding process. It represents the energy dissipation in the deformed adhesive when it is being debonded from the substrate surface. Under the conditions of constant [P.sub.0] and B, the higher the amount of the input energy dissipated by the bulk adhesive, the higher is the strength of the adhesive bond. Therefore, rheological and stress-strain measurements are employed to characterize the bonding, debonding and failure phenomena of the adhesion of the BIMS polymer system. For example, effects of transition temperature(s) and the plateau modulus of BIMS on the PSA behavior will be described. Also, model compound reactions coupled with FTIR and NMR analyses are used to elucidate surface interactions of the elastomer/substrate interface. The purpose of this work is to examine whether this model can be used to understand the adhesion of BIMS polymers.

Experimental

Polymers

An unbrominated BIMS (IMS) polymer and three different commercially available BIMS polymers used in this work are described in table 1. BIMS-1, BIMS-2 and BIMS-3 are Exxpro 93-5, Exxpro 90-10 and Exxpro 95-4 elastomers, respectively, of Exxon Chemical Company. Another BIMS polymer (denoted as BIMS-4) having a Mw [tilde] 300,000 was also used. This polymer contains [tilde] 3 mole % p-methylstyrene (PMS) of which 1 mole % of PMS is brominated. The IR used was Natsyn 2200 rubber, which has an ML-4 of 80 at 100C (ASTM D 1646).

[TABULAR DATA NOT REPRODUCIBLE IN ASCII]

Pressure sensitive adhesives (PSAs)

Two PSAs, B and C, formulated from BIMS-2 were studied and compared to the base polymer, A, and the PSA based on a styrene-isoprene-styrene triblock copolymer (SIS), D (table 2). This SIS (Shell Chemical's Kraton 1107 block copolymer) has a [M.sup.n] = 147,000 and a wt. % polystyrene = 15, and Tackifier-1 is Escorez 1310LC tackifier, the tackifying resin for SIS (ref 3). Tackifier-2 is Escorez 2520 tackifier, an aliphatic/aromatic liquid tackifier with a [M.sub.w] = 380, [M.sub.w]/[M.sub.n] = 1.5 and a [T.sub.g] = 16[degrees]C. Tackifier-3 is Escorez 5380 tackifier, a hydrogenated cyclopentadiene solid tackifier with a Mw = 360, [M.sub.w/[M.sub.n] = 2.4 and a [T.sub.g] = 36[degrees]C. All three tackifiers are products of Exxon Chemical Company. The PSA test used is 180[degrees] peel (PSTC-1). Samples used were adhesive formulations coated from toluene onto 38.1 [Mu]m mylar film. The substrate used was clean, smooth stainless-steel surfaces.

Table 2 - tackification of BIMS

             Requires a low [T.sub.g] tackifier

                      A           B            C           D

BIMS-2               100         100          100          -
SIS                   -           -            -          100
Tackifier-1           -           -            -          125
Tackifier-2           -          100          100          -
Tackifier-3           -          15           50           -

Rheological and stress-strain measurements

Dynamic mechanical experiments were performed using Rheometric Scientific's RMS-800 spectrometer with the 25 mm parallel plate and the torsion rectangle fixtures. Isothermal measurements were performed over a wide range of temperatures. All the isochronal experiments were run at 10 rad/s. For tensile stress-strain measurements, samples were the-cut into micro-dumbbell specimens (ASTM D1708). Pull speeds of 5 and 10 cm/min. were used. The stress was calculated based on the undeformed cross-sectional area of the tensile specimen.

All the BIMS and IR samples were compression-molded at 150[degrees]C for 25 minutes. PSA blends of BIMS and tackifier(s) were prepared by a solution blending method described previously (ref 3). These pressed or cast samples have a thickness of about 2 mm.

Bonding of BIMS to nylon and nylon/BIMS blend

The nylon 6 used was Capron 8209F polymer ([T.sub.m] = 230-235[degrees]C) from Allied-Signal Inc. The polymer is polymerized from [Epsilon]-caprolactam (ref. 5). It has the structure of [H.sub.2]N[([CH.sub.2]).sub.5]CO-[[NH[(CH.sub.2]).sub.5]-CO].sub.n]-NH [(CH.sub.2]).sub.5]COOH.

A pure nylon and two nylon/BIMS blend substrates with thickness of 0.1 mm were prepared by a Randcastle microextruder for the adhesion study. The three polymeric materials used are described in table 3. All three polymeric materials were in pellet form, dried at 90'C under vacuum overnight, and sealed in plastic bags before the extrusion.

Table 3 - nylon and nylon/BIMS blend substrates

Substrate        Composition           Comment

Nylon            Capron (100 wt.%)     1-pass through
                                       extruder
Nylon/BIMS-5     Capron/BIMS           [tilde] 0.22 mol% BzBr in
                 (80/20 wt. ratio)     BIMS
Nylon/BIMS-6     Capron/BIMS           [tilde] 0.74 mol% BzBr in
                 (80/20 wt. ratio)     BIMS

For the bonding of BIMS to nylon, a cloth backing was first bonded to the nylon substrate to avoid the unwanted contribution from the extension of the nylon substrate to adhesion. The cloth backing was a woven cotton fabric with a thickness of 0.3 mm. BIMS and nylon surfaces were then contacted at 165[degrees]C for 25 minutes under a pressure of 40 psi = 0.28 MPa. Contact was prevented at one end by inserting a piece of mylar to provide "arms" for subsequent T-peel measurements. After being conditioned overnight at room temperature, the laminate, cut into strip of 1/2" = 1.27 cm wide, was T-peeled apart at a separation speed of 2"/min. = 847 [Mu]/s. Without the cloth backing on the nylon, the work used to extend the nylon substrate will be included in the peel strength, which does not represent the true adhesion between BIMS and the nylon substrate. In this study, BIMS polymers of different BzBr levels: IMS, BIMS-1, BIMS-2 and BIMS-3 (table 1), bonded to a pure nylon substrate and two BIMS-modified nylon substrates (table 3) were used.

Bonding of BIMS to IR

A thin sheet of BIMS or IR (thickness, t [tilde] 1.2 mm), was prepared with one surface adhering to the thin cotton cloth backing mentioned above and the other molded against a Teflon coated aluminum foil to yield a smooth surface. The backing cloth served to minimize the extension of the separated parts under the action of the peel force. This eliminated the unwanted contribution of bulk elastomer extension to adhesion. BIMS and IR surfaces were then contacted at 60[degrees]C for 25 minutes under a pressure of 0.28 MPa. Contact was prevented at one end by inserting a piece of mylar film to provide "arms" for subsequent T-peel measurements. After being conditioned overnight at room temperature, the laminate, cut into strip of 1.27 cm wide, was T-peeled apart at a separation speed of 847 [Mu]m/s. In this study, IMS and BIMS polymers of different BzBr levels (table 1) were bonded to IR.

Results and discussion

PSAs of BIMS

Master curves of storage modulus (G'), loss modulus (G") and loss tangent (tan [Delta] = G"/G') of Adhesive B (table 2) based on the time (or frequency)-temperature superposition principle are shown in figure I over an extended frequency scale (ref 6). The average G' value of the plateau zone in the mid-frequency region of figure I is known as the plateau modulus, [G.sub.N[degrees]] (ref 7). Values. of [G.sub.N[degrees]], and the glassy onset frequency, which is the frequency at which G' and G" curves meet in the high frequency region, are shown in table 4. Values of the plateau modulus and the glassy onset frequency of the pure BIMS-2 polymer can be obtained from a previous study (ref 7).

Table 4 - viscoelastic and adhesion data of PSAs

                              A              B                C

BIMS-2                       100            100              100
Tackifier-2                   -             100              100
Tackifier-3                   -              15              50
Plateau modulus, Pa       2.71x10(5)      8.14x10(4)      2.94x10(4)
Glassy onset                1,100           240              20
  frequency, rad/s
180% peel to steel,         0.2             0.5             0.8
  kN/m

As described elsewhere, the tackification of a low [T.sub.g] elastomer such as a SIS polymer requires a high level of a high [T.sub.g] tackifier (ref. 3). The incorporation of a high [T.sub.g] tackifier has two functions. The tackifier "loosens up" or reduces the number of entanglements in the entangled polymer chains, resulting in a lower plateau modulus. Also, it shifts the damping of the low [T.sub.g] elastomer. This is accomplished by reducing the high elastomer glassy onset frequency to approach the peel frequency of the PSA debonding test. In this way, the adhesive is able to dissipate input mechanical energy to resist debonding. BIMS, IR and SIS have very similar [T.sub.g]s. BIMS-2 has a [T.sub.g] of -65[degrees]C (ref. 7). Why a high level of a high [T.sub.g] tackifier is not needed to tackify BIMS (table 2) can be explained as follows. As shown by the loss tangent curves in figure 2, this isobutylene-based polymer has an extra damping transition ([T.sub.d]) occurring at -33[degrees]C, which corresponds to an additional natural energy dissipation or damping mechanism, above [T.sub.g] (ref 7). If a high level of a high [T.sub.g], tackifier is added to BIMS, the glassy onset frequency of the elastomer/tackifier blend will be reduced too much. In other words, one needs to peel and deform this adhesive at an extremely low rate to obtain a high level of energy dissipation.

As discussed above (equation 1), there are three basic contributing factors to adhesion: surface interaction (represented by [P.sub.0], the interfacial adhesion term), bonding capability (represented by B, the bonding term) and energy dissipation due to deformation of the bulk adhesive (represented by D, the debonding term). An attempt to establish PSA design pathways from certain bonding and debonding criteria is described below (ref 3). Plateau modulus is chosen as the bonding parameter because it is a measure of how readily the adhesive conforms and wets the substrate surface. Glassy onset frequency is chosen as the debonding parameter because this frequency represents approximately the time or rate scale one deforms the adhesive to achieve maximum energy dissipation. The behavior of the BIMS systems (BIMS-2 and Adhesive B) compared to that of conventional SIS adhesive systems (SIS and Adhesive D) is shown in figure 3. Different from SIS, BIMS fulfills the Dahlquist criterion (the adhesive plateau modulus [tilde] 0.33 MPa or lower in order to establish extensive molecular contact with microscopically rough substrate surfaces), and therefore BIMS has inherent tack (ref. 8). Consistent with the PSA bond formation and bond breaking requirements, addition of a tackifier to either BIMS or SIS decreases both the plateau modulus and the glassy onset frequency, accompanied by a higher peel strength (ref 3).

Adhesion of BIMS to nylon

Figure 6 shows peel strengths of IMS and BIMS polymers (table 1) bonded to the nylon 6 substrate. The IMS polymeric adhesive has a very low adhesion and the failure mode after bond peeling is interfacial. As the BzBr level in the BIMS polymeric adhesive increases beyond 0 mole %, the peel adhesion is increased significantly and the failure mode after bond peeling is cohesive. For instance, consider the bonding of IMS and BIMS-1 to nylon. The first polymer has no BzBr groups, whereas the second polymer has 14 BzBr groups per polymer chain (table 1). These 14 BzBr groups in the BIMS chain improve its adhesion to nylon from 0.13 to 3.1 kN/m, close to a 30-fold increase.

The viscoelastic master curves of G', G" and tan [Delta] of BIMS-2 have been shown previously (ref. 7). Actually IMS, BIMS-1, BIMS-2 and BIMS-3 all show similar viscoelastic behavior over a wide range of frequency (figures 5-6). According to equation 1, the debonding term, D, depends on the rheological behavior of the bulk adhesive. The debonding frequency, [Omega], in the T-peel test is calculated according to the following equation:

[Omega] = 2[Pi]v/2[t.sub.a] = [Eta]v/[t.sub.a] (2)

where v is the Instron cross-head separation speed, 847 [Mu]/s, and [t.sub.a] is the adhesive thickness, which is equal to 0.61-0.66 mm (ref. 4). Figure 6 indicates that all four polymers have similar values of G" in the frequency range of 4.0-4.4 rad/s, which is the range of T-peel debonding frequency,[Omega], in these peel strength measurements according to equation 2. Therefore, D terms for these polymers should be similar. Of course, B terms of these polymeric adhesives to nylon are also similar because of identical bonding conditions and similar G' values for these polymers (figure 5). Therefore, under conditions of constant B and D, the higher strength of adhesion of the higher BzBr polymers (BIMS-1, BIMS-2 and BIMS-3) to nylon compared to the IMS polymer is due to an increase in the interfacial adhesion term, [P.sub.0], in equation 1. This increase in [P.sub.0] could be due to some specific interactions between BzBr groups in BIMS and functional groups in nylon 6.

To further understand the interaction at the BIMS/nylon interface, we carry out reactions of some model compounds. First, as described before, the nylon 6 substrate has the structure of [H.sub.2]N[([CH.sub.2]).sub.5]CO-[[NH[ ([CH.sub.2]).sub.5]-CO].sub.n]-[NH([CH.sub.2]).sub.5]COOH. Groups which may react with BzBr groups in BIMS are the in-chain amide groups and the chain end amine group. To test the first possibility, we react benzyl bromide (98%; used as received from Aldrich) with [Epsilon]-caprolactam (99+%; used as received from Aldrich):

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where benzyl bromide and [Epsilon]-caprolactam are chosen to simulate BIMS and nylon without the amine chain end, respectively. The benzyl bromide/[Epsilon]-caprolactam mix is in an equimolar ratio. Reaction conditions are identical to bonding conditions of BIMS to nylon. According to the FTIR analysis, no reaction occurs because the reaction product still exhibits the -NH- peak at about 3,200 [cm.sup.-1] (refs. 9 and 10). To test the second possibility, we dissolve BIMS-2 in toluene, add some triethylene tetramine ([H.sub.2]N-[CH.sub.2][CH.sub.2]-NH-[CH.sub.2][CH.sub.2]-NH-[CH.sub.2] [CH.sub.2]-[NH.sub.2]) and remove the solvent at 70-80[degrees]C. The polymer becomes crosslinked. Therefore, it is believed that the improved adhesion between BIMS and nylon is due to the chemical bonding of BzBr groups in BIMS and amine groups at nylon 6 chain ends:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

For the IMS polymer bonded to nylon, the interfacial adhesion term, [P.sub.0], is low due to the absence of BzBr groups. The results are a low adhesion and a clean separation of the adhesive from the substrate surface.

Adhesion of BIMS to nylon/BIMS blend

Peel strengths of IMS and BIMS polymers (table 1) to nylon/BIMS substrates compared to those of the same set of polymers to the pure nylon substrate are shown in figure 7. It is interesting to note that the IMS polymer exhibits high strengths of adhesion with the two BIMS-modified nylon substrates and failure modes after bond peeling are cohesive. Actually, adhesion to the two modified nylon substrates remains fairly constant with increasing amount of BzBr groups in BIMS. All bonds fail cohesively. With the addition of BIMS in the nylon substrate, the surface of the modified nylon substrate should be enriched with BIMS due to the lower surface energy of BIMS compared to the more polar nylon material. Therefore, when a BIMS polymeric adhesive is bonded to these modified nylon substrates, interdiffusion of BIMS polymers should occur under the bonding conditions of 165[degrees]C, 25 min. and 0.28 MPa, thus enhancing the interfacial adhesion term, [P.sub.0], in equation 1. A higher force is needed to separate these bonded materials and the separated surfaces exhibit cohesive failure.

It is also interesting to note that peel strengths of these IMS and BIMS polymers to nylon/BIMS-5 substrate appear slightly higher than those to nylon/BIMS-6 substrate (table 3). This could be explained as follows. BIMS-6 in nylon/BIMS-6 substrate has a higher concentration of BzBr groups, which react more readily with -NH2 groups in nylon chain ends. The result is a depletion of "free" NH2 groups in the nylon substrate to interact with BzBr groups in the BIMS polymeric adhesive and/or a lower amount of BIMS polymer migrating to the surface. The consequence of either event is a lowering in interfacial interactions, [P.sub.0].

To further understand the adhesion phenomena of BIMS to nylon/BIMS blends, we study the adhesion of BIMS to itself. Each BIMS polymeric layer is about 1 mm thick. Results are shown in table 5. In all cases, the interface between the two elastomers seems to disappear or "heal" as determined by visual inspection and the strength of the junction becomes identical to the cohesive strength of one of the elastomers. It is interesting to note that failure occurs in the mechanically weaker BIMS-1 for the bonding of BIMS-2 to BIMS-1. BIMS-1 is mechanically weaker than BIMS-2 based on stress-strain data shown in figure 8. At two different pull speeds, BIMS-2 has a higher tensile yield stress and tensile breaking energy (area under the stress-strain curve) than BIMS-1 (data for the low pull speed not shown in figure 8). Both elastomers show yielding and then stress-softening (negative slope) beyond the yield point (ref. 11).

Table 5 - adhesion of BIMS or IMS polymers

Elastomer A Elastomer B Peel strength, Failure mode

                                 kN/m
BIMS-2      BIMS-2      4.6 [plus or minus] 0.2      Cohesive(*)
BIMS-2      BIMS-1      3.5 [plus or minus] 0.1      Cohesive in B
IMS         IMS         3.9 [plus or minus] 0.2      Cohesive(*)

(*) Failure does not occur at the contacting interface. It occurs in one of the elastomer layers.

The above findings agree with the bonding results of various IMS and BIMS polymers to two BIMS-modified nylon substrates discussed earlier. It was speculated that BIMS in the modified nylon. substrates migrated to the substrate surface which would result in bonding of two BIMS polymers. Since the peel strengths are similar, in the 3.5+ kN/m region, and all the bonds fail cohesively, the speculation seems supported.

The independence of bond strength on functionality level of BIMS when bonding to BIMS-modified nylon suggests potential applications in composite structures. For example, some designs for automotive air conditioning hoses have complex multi-layer constructions such as 1) an inner elastomeric tube, 2) an inner adhesive layer, 3) a barrier film, 4) an outer adhesive layer and 5) an elastomeric cover compound. Use of the current findings could potentially lead to a simplified construction consisting of 1) a BIMS-based elastomeric tube, 2) a BIMS-based nylon barrier film and 3) a BIMS-based elastomeric cover. Of course the tube and cover could be independently formulated to fulfill different requirements without sacrificing adhesion. Other composite structures could also be considered including those based on BIMS-modified nylon fibers, for example, tires, air-springs and conveyor belts.

Adhesion of BIMS to IR

Results are shown in figure 9. In all cases, the thicknesses of both IMS or BIMS and IR layers are about 1 mm. Peel strengths of the IMS and all three BIMS polymers to IR are quite similar. Peeled failure modes are mixed interfacial (I) and cohesive (C), but only a minor amount of IR transferred to the BIMS surface, i.e., the interfacial failure region is larger than the cohesive failure region (I > C). Again it is interesting to note that IR is mechanically weaker than BIMS-1 and BIMS-2 based on stress-strain data shown in figure 8: Both BIMSs have higher tensile yield stresses and tensile breaking energies than IR at two different pull speeds (data for the low pull speed not shown). Based on peel strength results in figure 9, it appears that BzBr groups in BIMS have no specific interfacial interactions with, for example, the unsaturation in IR. To confirm this, we carry out a reaction of the following model compounds:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where benzyl bromide and squalene (97%; used as received from Aldrich) are chosen to simulate BIMS and IR, respectively. The benzyl bromide/squalene mix is in a 6:1 molar ratio because each squalene molecule has six double bonds. Reaction conditions are identical to bonding conditions of BIMS to IR. No reaction occurs. The benzyl bromide/squalene mix in deuterated chloroform before and after the reaction shows identical proton NMR spectra.

Finally, viscoelastic master curves of IR are shown in figure 10 for the completeness of this study. Compared to BIMS polymers, IR has a much narrower loss tangent minimum due to the presence of only a single [T.sub.g] (no [T.sub.d]; figure 2), although the value of loss tangent minimum is similar: IR = 0.10 and BIMS = 0.14 (ref. 7). IR also has a higher plateau modulus than BIMS (IR = 0.310 MPa and BIMS = 0.271 MPa).

Conclusions

Results of three different studies involving the adhesion of BIMS and its compounds are summarized as follows:

* Compared to SIS, BIMS has inherent tack and requires only a low [T.sub.g] tackifier for further tackification due to its low plateau modulus and extra damping transition, [T.sub.d].

* Benzyl bromide (BzBr) groups in BIMS improve adhesion to nylon 6 due to the chemical bonding between BzBr groups in BIMS and [-NH.sub.2] groups at nylon 6 chain ends based on results of model compound reactions. In other words, peel strengths of BIMS polymers with various BzBr levels to nylon 6 vary due to differences in interfacial adhesion (interfacial chemical bonding between BzBr groups and [-NH.sub.2] groups), not due to differences in bulk rheological behavior.

* BzBr groups in BIMS have no effect on the adhesion to IR. This agrees with the result that there is no reaction between the model compounds: benzyl bromide and squalene (an isoprene oligomer), which are used to simulate BIMS and IR, respectively. In other words, peel strengths of different BIMSs to IR are similar, indicating no specific interactions between BzBr groups in BIMS and unsaturation in IR at the BIMS/IR interface.

* The adhesion model, P = [P.sub.0]BD (equation 1), developed for PSA and HMA, is applicable to the adhesion of the BIMS elastomeric system. Model compound reactions coupled with FTIR add proton NMR techniques can elucidate the interfacial adhesion term, [P.sub.0], whereas rheological and stress-strain measurements can characterize the bonding term, B, the debonding term, D, and the adhesive failure mode.

References

[1.] M.J. Hubbard, B.J. Briddell and D.K. Fisher, U.S. Patent 5,234,987, August 10, 1993.

[2.] B.J. Briddell and M.J. Hubbard, U.S. Patent 5,242,727, September 7, 1993.

[3.] M. F. Tse, J Adhesion Sci. Technol 3(7), 551 (1989); M.F. Tse and L Jacob, J. Adhesion 56, 79 (1996).

[4.] M.F. Tse, J. Adhesion 48, 149 (1995).

[5.] "Nylon plastics," M.I. Kohan Ed., Wiley-Interscience, New York, 1973, p. 18-20.

[6.] J.D. Ferry, "Viscoelastic properties of polymers," 3rd Ed., John Wiley & Sons, New York, 1980.

[7.] M.F. Tse, H.-C Wang and J.E. Rogers, "Rheological behavior of a new isobutylene-based elastomer blended with a filler, " Rubber World, June 1997, p. 39-45.

[8.] C.A. Dahlquist, in: Adhesion: Fundamentals and practice, MacLaren, London (1966).

[9.] T Mizazawa, T Shimanouchi and S. Mizushima, J. Chem. Phys. 24, 408 (1956); 29, 611 (1958).

[10.] T Mizazawa, J. Mol Spectrosc. 4, 155 (1960).

[11.] M.F. Tse, H.-C. Wang, J.L. White and A.A. Galuska, "Interactions of a new isobutylene-based elastomer with fillers, " paper no. 40, presented at a meeting of the Rubber Division, American Chemical Society, Montreal, Quebec, Canada, May 5-8, 1996.