Specific nonlinear viscoelasticity behaviors of ... - Wiley Online Library

Specific Nonlinear Viscoelasticity Behaviors of Natural Rubber and Zinc Dimethacrylate Composites Due to Multi-Crosslinking Bond Interaction by Using Rubber Process Analyzer 2000

Yukun Chen,1 Chuanhui Xu2 1 School of Mechanical and Automotive Engineering, South China University of Technology, Guangzhou 510640, People’s Republic of China 2

College of Material Science and Engineering, South China University of Technology, Guangzhou 510640, People’s Republic of China

In this paper, the dynamic mechanical properties of the natural rubber (NR) which was filled with in situ zinc dimethacrylate (ZDMA) were investigated using a Rubber Process Analyzer (RPA2000). A weak rigid filler– filler network was formed in the uncured NR/ZDMA compounds. Pronounced Payne effect and stress-softening behavior were observed in the vulcanizates, which indicated that a strong and special filler structure was formed after vulcanization. However, an ionic network might be formed with a high ZDMA loading, which was not favorable to rebuild the poly-ZDMA network. Finally, two tentative regions of networks deformation during the dynamic strain sweep were discussed: filler–filler and ionic bonds dominating region and rubber network dominating region. POLYM. COMPOS., 32:1593–1600, 2011. ª 2011 Society of Plastics Engineers

INTRODUCTION Various studies on unsaturated carboxylates/elastomers composites have revealed that many elastomers can be reinforced by adding metal salt of dimethacrylate cured by peroxides [1–12]. Metal salt of dimethacrylate are effective crosslinking agents for elastomers and can be used over a wide loading range to tailor mechanical properties for different applications [13]. Mechanical properties of the vulcanizates are influenced significantly by the ionic crosslinks [7]. The metal salts of unsaturated carboxylic acids can be polymerized during vulcanization when peroxides are used as curing agents. Both homo-polymerization and graft-polymerization take place simultaneously. Correspondence to: Yukun Chen; e-mail: [email protected] DOI 10.1002/pc.21195 Published online in Wiley Online Library (wileyonlinelibrary.com). C 2011 Society of Plastics Engineers V

POLYMER COMPOSITES—-2011

The ionic crosslinks are formed by metal salts graft-polymerized onto the rubber chains, while the poly-(metal salt) chains repulse from the rubber matrix but attracted each other. Therefore, nano-scaled fine particles of poly(metal salt) chains are formed, which is key to form salt crosslinks [10]. Thus, not only conventional covalent crosslinks but also ionic crosslinks which exhibit special crosslinking structure and morphology coexist in the metal salts of unsaturated carboxylic acids-reinforced rubbers [10]. Although the nano-particles have been observed under transmission electron microscopy (TEM) [5, 6, 14], and many microstructure models [15] have been proposed, some aspects still need to be studied to improve the recognition of the essential microstructure. Dynamic viscoelasticity characterizes the internal molecular motion, which can help us to understand the internal structure of the composites. Few researches have been reported about the nonlinear viscoelastic behaviors of unsaturated carboxylates/elastomers. In this work, the nonlinear viscoelasticity of NR/ZDMA system was studied using a Rubber Process Analyzer (RPA2000). RPA2000 is a tool that can provide proper understanding of melt elasticity and process ability of polymer which is filled, or crosslinked, or compounded, or original polymers [16]. It was understood in terms of the dynamic functions like dynamic modulus of elasticity and viscosity over a wide range of strain amplitude and frequency. In this article, we also paid attention to the dynamic stresssoftening behavior of ZDMA-filled NR by RPA2000 in strain sweep mode. Stress softening is a typical phenomenon of many materials during cyclic tension tests. When a given filled elastomer is applied to a load–unload–reload cycle, the stress difference between successive loading cycles has the largest value in the first cycles and becomes negligible after several cycles, depending on the

loading of filler and maximum extension. Stress softening is significantly evident for vulcanized rubber named as Mullins effect [17–21]. This special phenomenon is due to the rubber network and filler network. The authors have already demonstrated in previous work, in the case of uniaxial tension [22], in which NR/ZDMA had pronounced stress softening and showed the good reinforcement of ZDMA. The stress softening of NR filled with ZDMA was attributed to the exchange reaction of ionic bonds when stress was applied. The slippage of ionic bonds for vulcanizate was due to the breaking of ionic bonds under stress and forming new ionic crosslinks [22]. However, our previous work not well analyzed the stress-softening behavior in the case of periodic oscillating shear. Furthermore, the possible network deformations during the dynamic strain sweep were discussed in this paper. EXPERIMENTAL AND METHOD Raw Materials and Recipes NR (Malaysia 1#) was provided by Guangzhou Rubber Industry Research Institute (China). Methacrylic acid (MAA), purchased from Guangzhou Xin’gang Chemical Factory (China), was purified by distillation under nitrogen at reduced pressure. Zinc oxide (ZnO) was purchased from Tianji Yaohua Chem. Co. (China). Dicumyl peroxide (DCP), purchased from Sinopharm Chemical Reagent Co. (China), was purified by anhydrous alcohol recrystallization before use. Theoretically, 1 mol ZnO and 2 moles MAA reacted to form ZDMA and water completely. Equivalent ZnO and MAA (ZnO/MAA molar ratio 0.5) were used to react in the NR matrix during mixing. A high degree of conversion to ZDMA from the in situ reaction of ZnO and MAA had been verified by many papers [7, 12]. Here, the neutralization was assumed to be reacted completely to form ZDMA. The compounds containing 100 weight parts of NR, 1.5 phr (parts per 100 parts of rubber) DCP and 0–40 phr ZDMA.

Samples Preparation Rubber compounds were prepared in a two-roll mill. ZnO and MAA were first added into NR and mixed for several minutes. DCP was added last in sequence. The compound was stored at room temperature for 8 h. Adequate amount of compounds was placed in the cavity of RPA2000 (Alpha technologies Co., UK), which is basically a torsional dynamic rheometer. The cavity house is a biconical test chamber that is closed by the action of a pneumatic ram operated at a pressure. A slight excess of test material was needed to ensure the cavity house fully. Tests were thus made under pressurized conditions to make sure that porosity did not develop in the samples when the instrument was under operating condition. 1594 POLYMER COMPOSITES—-2011

Frequency Sweep The temperature and the strain amplitude were kept constant at 608C and 18, respectively. Frequency sweeps were performed from 1 to 1,000 cpm (cycles per minute) to measure the storage modulus (G0 ) and the loss modulus (G00 ).

Strain Sweep The temperature and the frequency were kept constant at 608C and 60 cpm, respectively. The range of the strain sweep was 0.1-–208deg for the uncured compounds. After four consecutive scans on the uncured samples, the temperature was raised to 1558C to cure the compound for 20 min. Then, the temperature was reduced to 608C again. The range of the strain sweep was 0.1-–108deg for the vulcanizates. After three consecutive scans on the cured samples, the temperature was raised to 1008C for a delay of 30 min, keeping the samples intact within the confines of the cavity of the RPA2000 and again reduced to 608C to carry out the forth scan. This was done as measurement for recovery of the stress softening of the cured samples. Thus, the forth scan was defined as recovery sweep.

RESULTS AND DISCUSSION Frequency Sweep To understand the nature of the viscoelastic behavior for ZDMA in uncured compounds as well as vulcanizates, some samples were subjected to frequency sweep experiments. Figure 1 shows the dependence of storage modulus (G0 ) and loss modulus (G00 ) of the ZDMA/NR compounds as a function of frequency. Both the G0 and the G00 increase with increasing frequency. In general, the effect at high frequency is equivalent to the effect at low temperature because of reduce of the chains responsibility. Thus, the increase of modulus was attributed to a decrease in time available for molecular relaxation. Storage modulus (Fig. 1a) or loss modulus (Fig. 1b) of the compounds exhibits a relative large slope, indicating a weak time-independent elastic (or viscous) response, which supported the weak filler transient network in the compounds. In the case of frequency sweep for ZDMA/NR vulcanizates, as shown in Fig. 2, similar increasing tendency of G0 and G00 depending on frequency can be observed. As expected, the slope of storage modulus and loss modulus seems shallow compared to the uncured compounds. This suggested that a strong network in the vulcanizates displayed an apparent time-independent elastic (or viscous) response. Moreover, loss modulus increased with frequency, which was due to the higher energy required for molecular viscous response. Compared to G0 (Fig. 2a) at a given ZDMA loading, the value of G00 (Fig. 2b) was lower than that of G0 , implying the elastic-dominant response in all the blends. DOI 10.1002/pc

this secondary filler–filler structure. The rise in G0 (at the initial low strain) as a function of ZDMA was attributed to the increase in filler volume fraction (/), known as the hydrodynamic effect. The loss modulus showed a similar behavior, as shown in Fig. 3b. G00 is related to the energy dissipated during the deformation that it is generally associated with the breakdown of the filler transient network as well as chains slippage during deformation. A high loading of ZDMA leading to form a high developed secondary filler–filler network, therefore, more energy dissipated during the rupture of the filler structure at low shear strain region. At large shear strain region, the weak filler structure had been destroyed completely. In this case, the micro-size particles of ZDMA played a dilution effect on rubber chains, resulting in a lower G00 at high ZDMA loading. Thus, the G00 was decreased as a function of ZDMA at large shear strain region. In the case of ZDMA/NR vulcanizates, Fig. 4a shows a pronounced Payne effect of the composites. The LVE region of vulcanizate with high loading of ZDMA corresponded to an elastic modulus independent to deformation, which could be observed at middle strain amplitudes

FIG. 1. Frequency sweep on ZDMA/NR compounds.

Strain Sweep Nonlinear viscoelastic behaviors of uncured ZDMA/ NR compounds as a function of strain are given in Fig. 3. It is evident that there was a linear viscoelastic region (LVE) where the storage modulus was independent of shear strain existing in NR gum and the compound with 10 phr ZDMA, as shown in Fig. 3a. With the loading of ZDMA increasing, the LVE region disappears and G0 at initial shear strain increases. In addition, the dependence of G0 on shear strain is more pronounced as ZDMA increases. Especially at a high loading of ZDMA, for example, the 30 and 40 phr, the storage modulus shows a significant drop with increasing the strain, which is a typical characteristic of structure breakdown. When ZnO and MAA were added into NR and mixed in the two-roll mill, the in situ ZDMA was formed via the neutralization of metal oxides and acids. A weak filler–filler transient network could be formed at this moment, which was consisted with the analysis of frequency sweep. This secondary filler–filler structure was so weak that it could be disrupted at the very low shear strain amplitudes. This indicated that the rubber chains did not get involved in DOI 10.1002/pc

FIG. 2. Frequency sweep on ZDMA/NR vulcanizates.


1595

of poly-ZDMA was big enough, the poly-(metal salt) chains would separate from the rubber matrix and aggregate into nano-scaled fine particle which had been verified by many papers [5, 6, 14]. Thus, it is reasonable to regard the pronounced Payne effect of vulcanizate was attributed to the filler–filler network formed by nano-scaled polyZDMA. However, the ionic crosslinks introduced by poly-ZDMA made the crosslink structure complex, adding obscurity to the essential mechanism of Payne effect of ZDMA/NR vulcanizates. Compared to uncured compounds, the vulcanizates yielded much higher G0 and G00 , due to the presence of chemically molecular crosslinks and formed nano-scaled poly-ZDMA. As shown in Fig. 4b, the G00 of vulcanizates did not seem to be regular. Increasing rapidly of the G00 at high strain amplitudes (near 108) indicated that more energy dissipated in deforming of the rubber crosslink network. For the sample of 20, 30, and 40 phr ZDMA, a small loss peak appeared at about 38strain amplitude, which was marked by a shadow region in Fig. 4b. Please note that the G0 of above samples exhibited their drops at this strain amplitude, showing in Fig. 4a. Considering the ionic bands, we regarded this small loss peak as the

FIG. 3. Strain sweep on ZDMA/NR compounds.

(even to 18). This experimental result demonstrated that a developed and strong filler–filler network was formed. Payne effect [23, 24] is regarded as a result of breakage and reforming of physical bonds between the filler aggregates. These bonds are assumed to build filler agglomerates of different size and, above a certain threshold, a filler–filler network within the rubber matrix. It is well known that the modulus is dependent on the crosslink network of the rubber, the hydrodynamic effect, the filler– rubber interactions, and the filler–filler network. Filler–filler networks are dependent on the deformation degree at smaller deformations, whereas the other parameters remain constant. Thus, when the deformation increased, the filler network eliminated, reducing the shear modulus. Therefore, a conclusion drawn from Fig. 4a was that the stronger the rebuilt ability of the filler–filler network, the higher the strain amplitudes used to break down the network. However, this was not applicable to the ZDMA/NR vulcanizate that would be given a discussion in detail later. As mentioned in Introduction, during the curing reaction, the polymerization of ZDMA led to increase molecular weight of poly-ZDMA. When the molecular weight 1596 POLYMER COMPOSITES—-2011

FIG. 4. Strain sweep on ZDMA/NR vulcanizates.

DOI 10.1002/pc

FIG. 5. Stress softening of G0 of ZDMA/NR compounds.

energy dissipated involving ionic crosslinks slippage and rubber covalent crosslinks deformation. For the case of 10 phr, the ionic network was not formed. Therefore, the sample with 10 phr ZDMA and NR gum did not produce such a loss peak.

Dynamic Stress-Softening Behavior In a filled rubber, stress-softening effect is generally accepted as a result from the filler-related structures and therefore can yield some insight into the filler structures [25, 26]. The stress-softening effect of the G0 of uncured ZDMA/NR compounds is shown in Fig. 5. The NR (Fig. 5a) did not display a significant reduction of shear elastic modulus after the first strain scan, whereas the compound with 40 phr ZDMA (Fig. 5b) showed an apparent stress-softening behavior. The stress-softening effect in ZDMA/NR compounds was mostly from the contribution of the filler–filler rigid three-dimensional network structures. For the compound with a high loading of 40 phr ZDMA, the increasing strain magnitude in the first strain sweep obviously caused the filler network to break DOI 10.1002/pc

down. This again indicated a weak but developed filler– filler transient network was formed at high loading of ZDMA. However, an interesting behavior could be observed in the filled compound is that the G0 of third and forth sweep showed an increase compared with the second sweep. It is well known that the polarity of filler is a benefit to aggregate of filler. The strain sweeps could soften the rubber molecules to a certain degree which did good to accelerate the aggregating of ZDMA. As a result, the G0 showed an enhancement in the third and forth sweep. In the recycling strain sweep of ZDMA/NR vulcanizates, after three consecutive scans at 608C, the temperature was first raised to 1008C and kept at this temperature for 30 min. After that, the sample was kept intact within the confinement of the cavity of RPA2000 and was reduced to 608C to undergo the forth scan. The sample was kept at 1008C isothermally for 30 min, to accelerate the recovery of stress softening. The stress softening of G0 is shown in Fig. 6. Differing to the uncured compounds, the vulcanizates showed an apparent stresssoftening behavior. The higher loading of ZDMA was added, the more significant softening of the G0 could be observed. After the first scan, the LVE region of vulcanizates was shortened with increasing ZDMA, the recovery degree of G0 at the forth scan was also reduced with increasing ZDMA. The tendency of shortening LVE region was so apparent that the LVE region of 30 phr (Fig. 6d) and 40 phr ZDMA (Fig. 6e) almost disappeared even at small strain amplitudes (0–0.18). It was accepted that reinforcement was related to the combinational contribution of the filler–filler and the filler–rubber interactions. Breakdown of these interactions was considered to be responsible for strain softening. As discussed before, poly-ZDMA was in a nano-scale which might be contributed to a developed and strong filler network. Moreover, the chemical crosslink structure of composites contained not only the covalent crosslinks but also the ionic crosslinks introduced by the graft poly-ZDMA [10]. Thus, it was reasonable to deduce that a more developed ionic crosslink network might be formed at a high loading of ZDMA. In fact, the total crosslink density and the ionic crosslink density showed a significant increase with increasing ZDMA loading, whereas the covalent crosslink density decreased [4–12]. When the strain amplitudes increased to a certain degree, the covalent crosslink network and ionic crosslink network deformed, and some debondings like in the situation of carbon black [18, 22], but lots of ionic bonds slipped and broken, specifically at a high content of ZDMA. At the same time, the new ionic crosslinks might be formed rapidly [22]. Once the tear stress was relieved, most of the slippage of ionic bonds ceased back to the initial structure. However, in the case of conventional reinforcing filler such as carbon black [22, 26], the recovery of the G0 was remarkable that was attributed to the reformation of the filler structures. On the other hand, the filler–filler network and covalent netPOLYMER COMPOSITES—-2011

1597

FIG. 6. Stress softening of G0 of ZDMA/NR vulcanizates.

work of rubber possess some recoverability more or less. As a result, the recovery of G0 was weak at a high loading of ZDMA, and a high stress-softening behavior was shown in the vulcanizate with high loading of ZDMA too. After the first sweep, the disappearance of the LVE region of 30 and 40 phr ZDMA revealed that the rigid filler–filler network became fractured that initiated the sharp 1598 POLYMER COMPOSITES—-2011

decrease of modulus at low strain amplitude. This demonstrated that a more developed ionic crosslink network might handicap the reformation of filler–filler networks. Figure 7 exhibits the stress-softening behavior of elastic torque (S0 ) of the ZDMA/NR vulcanizates. Similar to the G0 , the stress-softening effect of S0 increases with increasing the loading of ZDMA. This was due to the DOI 10.1002/pc

reaction of ionic bonds leading to a relative stable new ionic network to ceases back to initial state. These dynamic stress-softening experimental results consisted with the results of a uniaxial tension experiment [22].

A Simple Discussion on the Possible Networks Deformation in the Dynamical Strain Sweep As discussed before, the ionic crosslinks resulted in an obscure mechanism of the Payne effect of the ZDMA/NR vulcanizates. Here, we chose the composite with 30 phr ZDMA as an instance to illustrate a tentative district of networks deformation during dynamic strain sweep. As shown in Fig. 8, the strain amplitude was necessary to be divided into two major regions according to G0 . It was well known that, using black carbon as a reinforcing filler, a rigid three-dimensional network was formed which contributes mainly to the modulus of composites. At a very small strain, the filler network could not be fractured. With the increase of strain amplitude, the rigid network became fractured, which initiated the sharp decrease of modulus. General speaking, the apparent decrease occurred at about 1% strain [26] which was equal to about 0.078in RPA2000. In the region from 0 to about 0.18in this particular experiment, the filler structure seemed weak since the drop of G0 was relative remarkable. The drop the G0 was due to the rupture of polyZDMA aggregates network. This could be explained that the developed ionic crosslinks maintain the instantaneous structure, which was not favorable to rebuild the polyZDMA network. To our surprise, another small loss peak appeared at about 0.18strain amplitude when the coordinate axe Y was amplified. This was a good support of energy dissipating caused by rupture of poly-ZDMA aggregates network, at least to some extent. We proposed that a developed and strong filler–filler network was formed because of a LVE region. In fact, this ‘‘LVE

FIG. 7. Stress softening of the elastic torque of ZDMA/NR vulcanizates.

increase of crosslink density with adding ZDMA. For the NR, the four S0 curves seemed essentially coincident with each other, indicating no rupture of network. Without a doubt, the ionic crosslinks contributed much to the softening behavior. However, the S0 of the forth sweep did not show any recovery compared with the third sweep for all the samples, which was a good support for the exchange DOI 10.1002/pc

FIG. 8. Illustration of the different mechanism dominating regions on strain.


1599

region’’ was mainly due to the slippage and exchange reaction of ionic bonds under the dynamical stress instead of rebuilt of poly-ZDMA aggregates network. The ionic crosslinks was considered to play an important role in the remarkable mechanical properties of ZDMA/NR composites [7, 10]. See the stress-softening behavior of S0 starting at the strain amplitudes exceed 0.18 (Fig. 8). This demonstrated that the crosslink network was got involved in the deformation of structure at such a strain amplitude. Furthermore, the G0 of a black carbon-filled rubber vulcanizate [26] generally shows a notable drop in the region from about 0.1–18 which is equal to about 1.4–14%. Generally speaking, the rupture of rubber covalent crosslink could not take place at this strain. Therefore, the ‘‘LVE region,’’ observed in the region from about 0.1–18, is related to the action of the ionic bonds. However, it is difficult to distinguish the filler network effect and ionic bonds effect, thus we summarized the two types network deformation together. Obviously, the rubber network deformed when the strain amplitudes exceeded about 18. G0 decreased to a low level which was mainly contributed by rubber matrix. G00 and S0 showing an abrupt increase at the high strain amplitudes was associated with the rubber network deformation. It was worth mentioned that the second small loss peak locating in this region was mainly attributed to the complex results in combination of ionic crosslinks and rubber covalent network. Final Comments As for the uncured compounds, the in situ ZDMA, formed via the neutralization of metal oxides and acids, could aggregate to form a weak rigid filler–filler network. The interfacial adhesion between ZDMA and NR was weak because of the great difference of polarity between ZDMA and NR. In the case of cured composites, the poly-ZDMA could be acted as nano-particles which involved the NR chains and ionic crosslinks, forming a strong network. Thus, a pronounced Payne effect and stress-softening behavior were observed. The elastic torque of the forth scan showing no recovery indicated the exchange reaction of ionic bonds leading to a relative stable new ionic network to cease back to initial state. After first sweep, the shortening of LVE region also demonstrated that the ionic crosslinks may handicap the reformation of filler–filler networks. Two tentative districts of networks deformation, namely filler–filler network and ionic bonds dominating region and rubber network dominating region, was necessary to be introduced to explain the deformations during strain sweep. However, the division of these two regions sounds somewhat farfetched, which was needed to be further studied and improved.

1600 POLYMER COMPOSITES—-2011

We expected that our study could help to explore the essence of the internal network structure of the unsaturated carboxylates/elastomers composites. REFERENCES 1. A. Nomura, J. Takano, A. Toyoda, and Y. Saito, Nippon Gomu Kyokaishi, 66, 830 (1993). 2. Y. Saito, K. Nishimura, M. Asada, and A. Toyoda, J. Jpn. Rubber Soc., 67, 867 (1994). 3. G.X. Gao, Z.C. Zhang, Y.S. Zheng, and Z.H. Jin, J. Appl. Polym. Sci., 113, 3901 (2009). 4. D. Yin, Y. Zhang, Y. Zhang, Z. Peng, Y. Fan, and K. Sun, J. Appl. Polym. Sci., 85, 2667 (2002). 5. D. Yin, Y. Zhang, Z. Peng, and Y. Zhang, Eur. Polym. J., 39, 99 (2003). 6. Y.L. Lu, L. Liu, C. Yang, M. Tian, and L.Q. Zhang, Eur. Polym. J., 41, 577 (2005). 7. Z. Peng, X. Liang, Y. Zhang, and Y. Zhang, J. Appl. Polym. Sci., 84, 1339 (2002). 8. X.H. Yuan, Z.L. Peng, Y. Zhang, and Y.X. Zhang, Polym. Polym. Comp., 7, 431 (1999). 9. A. Du, Z. Peng, Y. Zhang, and Y. Zhang, J. Appl. Polym. Sci., 89, 2192 (2002). 10. Y.J. Nie, G.S. Huang, L.L. Qu, P. Zhang, G.S. Weng, and J.R. Wu, J. Appl. Polym. Sci., 115, 99 (2010). 11. A.H. Du, Z.L. Peng, Y. Zhang, and Y.X. Zhang, Polym. Test., 21, 889 (2002). 12. Y. Lu, L. Liu, D. Shen, C. Yang, and L. Zhang, Polym. Int., 53, 802 (2004). 13. R.C. Klingender, M. Oyama, and Y. Saito, Rubber World, 202, 26 (1990). 14. A.H. Du, Z.L. Peng, Y. Zhang, and Y.X. Zhang, J. Polym. Sci.: Part B, 42, 1715 (2004). 15. Y.L. Lu, L. Liu, M. Tian, H.P. Geng, and L.Q. Zhang, Eur. Polym. J., 41, 589 (2005). 16. S. Datta, K. Naskar, Y.K. Bhardwaj, and S. Sabharwal, Polym. Bull., 66, 637 (2011). 17. F. Bueche, J. Appl. Polym. Sci., 10, 107 (1960). 18. D. Roozbeh and I. Mikhail, Int. J. Solids. Struct., 46, 2967 (2009). 19. Z. Guo and L.J. Sluys, Eur. J. Mech A/Solid, 25, 877 (2006). 20. F. Bueche, J. Appl. Polym. Sci., 10, 107 (1960). 21. F. Yoshihide, J. Appl. Polym. Sci., 95, 60 (2005). 22. Y.K. Chen and C.H. Xu, J. Appl. Polym. Sci., DOI 10.1002/app.34584. (2011) 23. A.R. Payne, J. Appl. Polym. Sci., 9, 1073 (1965). 24. A.R. Payne and R.E. Whitaker, Rubber Chem. Technol., 44, 440 (1971). 25. M.J. Wang, Rubber. Chem. Technol., 71, 520 (1998). 26. L. Jong, J. Polym. Environ., 13, 329 (2005).

DOI 10.1002/pc