Relationship between the rheology and morphology of dynamically vulcanized thermoplastic elastomers based on EPDM/PP.
Since dynamically vulcanized thermoplastic elastomers (TPVs) were first presented by Fisher [1, 2], they have been widely used in the rubber and plastic industries [3-5]. These materials behave, to a good approximation, like vulcanized rubber at room temperature, and undergo a thermoplastic melt process at temperatures above the melting point for matrix polymers.
While considerable attention has been paid to the morphological types and room-temperature performance of these materials [6-11], there has been relatively little interest in their rheological and viscoelastic properties at elevated temperatures [12-15]. Few studies have examined in detail the mechanism behind the morphological development of TPVs. In previous works [16-18] we studied the mechanism of morphological development, particularly the parameters that affect the formation of rubber particle agglomerates in TPVs based on EPDM/PP. We demonstrated that the dynamic vulcanization process can lead to cured rubber agglomeration formed in the PP matrix by a joint shell mechanism. The extent of this mechanism is governed by several parameters, including the blend composition, viscosity ratio, shear forces, and interfacial interaction between two phases.
Over the last three decades, the rheological behavior of immiscible polymer blends has been the subject of many studies [19-21]. However, despite some valuable advances, a successful theory describing the flow behavior while taking into account the contributions of the interfacial tension and flow-induced blend morphology is still lacking.
The main objective of the present work was to study the relationship between the rheological properties and morphology of these materials.
The basic characteristics of the polymers used in this study are listed in Table 1. Injection-grade isotactic polypropylene from Himont Co., and EPDM based on ethylidene norbornene (ENB) as a diene monomer (Keltan), supplied by DSM Elastomers Europe, were used as commercially available.
Preparation of Blends
The melt reactive blending process for preparing TPV samples composed of 20/80, 40/60, 60/40, EPDM/PP w:w, was carried out in a small-scale laboratory Haake internal mixer (Rhecord 90) at a rotor speed of 60 rpm and temperature of 180[degrees]C. PP was added to the blend, and after 2 min the EPDM was added. The mixing was continued for 4 min, and then crosslinking ingredients were fed into the melt mixture. The uncured blend samples (TPO) were prepared with the same procedure. The variations of mixing torque versus time were recorded for all blending processes. To follow the morphological development during the melt blending, we removed the samples from the hot running mixer at different mixing stages without interrupting the mixing process. The samples were then cooled at room temperature before they were examined by SEM.
Morphological studies were performed on the cryogenically fractured surface of the samples, which was etched by hot xylene for 30 sec. The treated samples were then sputter-coated with gold and viewed with a scanning electron microscope (model XL 30; Philips).
[FIGURE 1 OMITTED]
The rheological behavior and melt viscoelastic properties of the samples were studied using a rheometric mechanical spectrometer (Paar Physica USD200) with a parallel plate (diameter = 2.5 cm; gap = 1 mm). All measurements were performed at a temperature of 220[degrees]C, frequency range of 0.01-1000 [sec.sup.-1], and strain amplitude of 1%.
RESULTS AND DISCUSSION
To determine the strain amplitude range over which linear viscoelasticity prevails, we conducted amplitude sweep experiments at a frequency of 5 rad/sec in the overall amplitude range of 0.1-100. Figure 1 shows the storage modulus, G', against the strain amplitude obtained for the EPDM/PP (60/40, w:w) uncured blend (TPO) and dynamically vulcanized (TPV) samples. The linear viscoelastic region in the cured samples where G' remains constant with respect to the strain amplitude is much shorter than that in the uncured blend samples. Figure 2 shows the dynamic viscosity, [eta]', and storage modulus, G', versus angular frequency, [omega], at 220[degrees]C for the uncured blend samples containing 20%, 40%, or 60% of EPDM. By comparing these results, one can see that the sample containing 60% of EPDM shows an appreciable increase in viscosity at low shear rates. The elastic modulus of this sample at low shear rates is also much greater than those observed for the other two samples. These results can be attributed to a three-dimensional association of the dispersed EPDM phase as a result of the predominating coalescence process. This is demonstrated by scanning electron micrographs of two samples (Fig. 3) that were taken at two different shear rates and immediately frozen in liquid nitrogen.
Over the last two decades a considerable number of studies have attempted to establish a model to describe the correlation between the microstructure and viscoelastic properties of emulsion systems [19-21]. It has been shown that the melt linear viscoelastic properties of immiscible polymer blends can be predicted from a model that was originally developed for emulsion by Palierne , which is defined as follows:
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
[G*.sub.mix] = [G*.sub.m] [[1 + 3[phi]H]/[1 - 2[phi]H]]; (1)
H = [4[alpha]/R(2[G*.sub.m] + 5[G*.sub.i]) + ([G*.sub.i] - [G*.sub.m])(16[G*.sub.m] + 19[G*.sub.i])]/[40[alpha]/R([G*.sub.m] + [G*.sub.i]) + (2[G*.sub.i] + 3[G*.sub.m])(16[G*.sub.m] + 19[G*.sub.i])]
where G*, [G*.sub.m], and [G*.sub.i] are the complex modulus of the blend, polymer matrix, and dispersed phase, respectively; [PHI] is the volume fraction of the disperse phase; R is the particle radius of the disperse phase; and [alpha] represents the interfacial tension ([[gamma].sub.12]) between two phases. This model takes into account the viscoelasticity of the phases, and polydispersity in the size and nature of the inclusions.
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
To use the above-described model, we calculated the interfacial tension of TPO samples ([[gamma].sub.12]) using a well known harmonic equation :
[[gamma].sub.12] = [[gamma].sub.1] + [[gamma].sub.2] - [[4[[gamma].sub.1.sup.d][[gamma].sub.2.sup.d]]/[[[gamma].sub.1.sup.d] + [[gamma].sub.2.sup.d]]] - [[4[[gamma].sub.1.sup.Pp][[gamma].sub.2.sup.P]]/[[[gamma].sub.1.sup.P] + [[gamma].sub.2.sup.P]]] (2)
where [[gamma].sub.1], [[gamma].sub.2], [[gamma].sup.P], and [[gamma].sup.d] denote the [gamma] values of PP and EPDM and their associated nonpolar (dispersion) and polar components, respectively. The calculated results are given in Table 2.
The average rubber droplet size (R) of the uncured blend samples containing 20%, 40%, and 60% of EPDM measured by the image analysis method are given in Table 3. It should be noted that all of these samples were taken at high shear rates (200 [s.sup.-1]) and were immediately frozen in liquid nitrogen. These results clearly demonstrate the significant role of coalescence on the dispersed droplet size as a result of increasing rubber concentration.
Figure 4 compares the experimental and predicted results of the complex modulus, G*, based on Eq. 1 as a function of frequency, [omega], for uncured blend samples of three different compositions. It can be seen that for the sample containing 20% of rubber, a good agreement is obtained for a whole range of shear rates, while for the sample containing a higher concentration of rubber (EPDM/PP, 60/40 w:w), the agreement exists only for high shear rates. The disagreement observed at low shear rates for the EPDM/PP (60/40 w:w) blend sample may be due to the predominating coalescence effect resulting in a co-continuous morphology.
The flow behavior and linear viscoelastic properties of dynamically vulcanized TPV samples with compositions of 20/80, 40/60, and 60/40 (EPDM/PP w:w) are shown in Fig. 5. It can be seen that these samples (particularly the sample containing 60% EPDM) show a pronounced viscosity upturn at low shear rates. These results also show that the samples containing 40% and 60% EPDM exhibit a strong storage modulus value that tends to become independent of frequency (plateau) at low frequencies. The rheological behavior and viscoelastic properties exhibited by these TPV samples are similar to those reported for block copolymers and concentrated particulate composites. Thus, the viscosity upturn and strong elasticity leading to yield stress at low shear rates shown in this sample can be attributed to the physical three-dimensional network structure formed between the cured rubber particles.
The results discussed above justify our previously proposed mechanism for the morphological development of TPVs  regarding with the agglomeration and dis-agglomeration of the cured rubber particles during the dynamic vulcanization stage. Scanning electron micrographs of the TPV samples are presented in Fig. 6. These results show that the extent of agglomerate formation is enhanced by the increased rubber aggregate size resulting from the increased rubber content in the samples.
Figure 7 shows the relaxation time distribution H([lambda]) for PP and three TPV samples varying in EPDM content. The multiple elastic response exhibited by the sample containing 60% EPDM, in comparison with the single but broad elastic response observed for the other samples, suggests that in the TPV samples (60/40 EPDM/PP, w:w), apart from the contribution of the flow-induced molecular orientation of the PP matrix, there may also exist some elastic response induced by agglomerates formed between the cured rubber particles.
The results of H([lambda])[lambda] vs. [lambda] for the sample containing 60% rubber are shown in Fig. 8. The pronounced peak corresponding to long relaxation times (low shear rates) can be considered as a characteristic elastic response of a three-dimensional network structure formed at low shear rates.
[FIGURE 6 OMITTED]
From the above-discussed results and the results of relative viscosity, [eta]', vs. the EPDM content shown in Fig. 9 for these samples, one can define a maximum packing volume [[PHI].sub.m], for the cured rubber particles above which the viscosity tends to infinity. After we considered these results, we attempted to find a rheological model that took into account the flow-induced morphological changes and cured rubber concentration. The most appropriate model for this purpose was that proposed by Jarzebski  for suspensions, in the following form:
[[eta].sub.sup] = [[[tau].sub.y]/[dot.[gamma]]] + [9/8]k[[([phi]/[[phi].sub.m])]/[1 - ([phi]/[[phi].sub.m])[.sup.1/3]]][.sup.n][bot.[gamma].sup.n-1] (3)
where [PHI] is the dispersed phase concentration; [[PHI].sub.m] is the maximum packing volume of the rubber particles, which can be assumed to be about 0.64 ; [[tau].sub.y] is the yield stress; and n, k are rheological parameters of the polymer matrix. Figure 10 compares the experimental results and the results predicted from the Jarzebski model. One can see that there is good agreement between the experimental and predicted results.
In the present work we attempted to use an appropriate linear viscoelastic model to gain more insight into the relationship between the microstructure and viscoelastic properties of TPV samples. In the Palierne model (Eq. 1), when H tends to 1/2, the model becomes approximately similar to the Einstein model introduced for the dilute suspensions with spherical particles , in the following form:
[G*.sub.susp] = [G*.sub.m][[1 + 3[phi]H]/[1 - 2[phi]H]] [approximately equal to] [G*.sub.m](1 + 2.5[phi]), H = 1/2. (4)
The experimental and predicted complex modulus, G*, values for three samples varying in rubber concentration are shown in Fig. 11. Although agreement is observed at high shear rates, at low shear rates the experimental results are higher compared to the predicted values. A comparison of these results also shows that the discrepancy between the experimental and predicted results is increased as a result of increasing the rubber content of the samples.
Since the Palierne model is based on dilute suspensions, this discrepancy is not unexpected for TPVs. Therefore, in the present work we modified the above model, considering the [[PHI].sub.m] associated with the agglomerate structure, in the following form:
[G*.sub.susp] = [a.sub.1]([phi]/[[phi].sub.m]) + [a.sub.2]([phi]/[[phi].sub.m])[.sup.2] + [G*.sub.m] [[1 + 3[phi]H]/[1 - 2[phi]H]]; H = 1/2
[a.sub.1] [approximately equal to] -6.1E4 (5)
[a.sub.2] [approximately equal to] 2.2E5.
The results predicted from the modified model were compared with the experimental values shown in Fig. 11. The excellent agreement between these results supports the validity of the modified model for predicting the melt viscoelastic properties of TPVs.
[FIGURE 7 OMITTED]
[FIGURE 8 OMITTED]
The melt flow and viscoelastic properties of uncured blend samples containing 20% and 40% EPDM were found to be predominantly controlled by the PP matrix. The appreciable viscosity increase and high elasticity observed at low shear rates for the sample containing 60% EPDM suggest a co-continuous morphology at low shear rates, and a matrix dispersal at high shear rates resulting from a droplet break-up mechanism.
It has been demonstrated that there is a close relationship between the rheological behavior and the viscoelastic properties and morphological features of TPVs, particularly regarding the parameters that affect the formation of the network agglomerate structure between the cured rubber particles.
An appreciable viscosity upturn and strong storage modulus, G', which tends to plateau at the low shear rates exhibited by the dynamically vulcanized EPDM/PP (60/40, w:w) sample are comparable to those reported for block copolymers and highly concentrated particulate composites. The melt flow and linear viscoelastic behavior of the TPV samples depend not only on the extent of the agglomerate formed between the cured rubber particles, but also on the agglomerate size and size distribution.
[FIGURE 9 OMITTED]
The rheological behavior of the TPV samples prepared in the present work was found to be in close agreement with that predicted by the Jarzebski model for suspensions when maximum packing volume, [[PHI].sub.m], of cured rubber particles is considered to be 0.64.
We have proposed a linear viscoelastic model similar to the Palierne model for dilute suspensions, taking into account the maximum packing volume, [[PHI].sub.m], for cured rubber particles, to predict the melt linear viscoelastic properties of TPV samples. A good agreement was found between the predicted and experimental results.
[FIGURE 10 OMITTED]
[FIGURE 11 OMITTED]
TABLE 1. Characteristics of the materials used. Materials Characteristics PP MFI 6.0 g/10 min Density 0.91 g/[cm.sup.3] Tm 165[degrees]C EPDM Mooey viscosity ML(1 + 4) 125[degrees]C 77 Ethylene content 52% Termonomer content 4.5% Density 0.89 g/[cm.sup.3] Curing rate Fast TABLE 2. Interfacial tension characteristic parameters. Polymer [gamma] [[gamma].sup.p] [[gamma].sup.d] [[gamma].sub.12] mN/m mN/m mN/m mN/m PP 20.8 0.416 20.314 EPDM 22.8 0.2 22.6 EPDM/PP 0.18 TABLE 3. The average rubber particle size [R] of the uncured blend samples containing 20, 40, and 60% of epdm. Composition (EPDM/PP, w:w) 20/80 40/60 60/40 R ([micro]m) 1.1 1.8 2.4
1. W.K. Fisher, U.S. Patent 3,758,643 (1973).
2. W.K. Fisher, U.S. Patent 3,862,106 (1975).
3. A.Y. Coran and R.P. Patel (Monsanto Co.), U.S. Patent 4,104,210 (1978).
4. A.Y. Coran, B. Das, and R.P. Patel (Monsanto Co.), U.S. Patent 4,130,535 (1978).
5. A.Y. Coran and R.P. Patel, Rubber Chem. Technol., 53, 141 (1980).
6. B. Kuriakose and S.K. De, J. Appl. Polym. Sci., 32, 5509 (1986).
7. C.S. Ha, D.J. Ihm, and S.C. Kim, J. Appl. Polym. Sci., 32, 6281 (1986).
8. E.N. Kresge, Rubber Chem. Technol., 64, 469 (1991).
9. S. Abdov-Sabet, R.C. Puydak, and R.P. Patel, Rubber Chem. Technol., 64, 769 (1991).
10. E.N. Kresge, in Polymer Blends, Vol. 2, D.R. Paut, S. Newman, eds., Academic Publishers, New York, 2 (1978).
11. C. Qin, J. Yin, and B. Hung, Rubber Chem. Technol., 63, 77 (1990).
12. L.A. Goettler, J.R. Richwine, and N. Nakajima, J. Non-Newtonian Fluid Mech., 8, 103 (1988).
13. P.K. Han and J.L. White, Rubber Chem. Technol., 68, 728 (1995).
14. Z. Krulis and I. Fortelny, Eur. Polym. J., 33, 513 (1997).
15. M.A. Lopez-Manchado, J. Biagiotti, and J.M. Kenny, J. Appl. Polym. Sci., 81, 1 (2001)
16. F. Goharpey, A.A. Katbab, and H. Nazockdast, J. Appl. Polym. Sci., 81, 2531 (2001).
17. F. Goharpey, A.A. Katbab, and H. Nazockdast, Rubber Chem. Technol., 76, 239 (2003).
18. S. Bazgir, A. Katbab, and H. Nazockdast, J. Appl. Polym. Sci., 92, 2000 (2004).
19. J.F. Palierne, Rheol. Acta, 29, 204 (1990).
20. D. Graebling, R. Muller, and J.F. Palierne, Macromolecules, 26, 320 (1993).
21. D. Graebling and R. Muller, J. Rheol., 34, 193 (1990).
22. S. Wu, Polymer Interface and Adhesion, Marcel Dekker, New York (1982).
23. G.J. Jarzebski, Rheol. Acta, 20, 280 (1981).
24. L.E. Nielsen and R.E. Landel, Mechanical Properties of Polymers and Composites, 2nd ed., Marcel Dekker, New York (1994).
F. Goharpey, H. Nazockdast, A.A. Katbab
Polymer Engineering Department, Amirkabir University of Technology, Tehran, Iran
Correspondence to: H. Nazockdast; e-mail: Nazdast@aut.ac.ir
|Printer friendly Cite/link Email Feedback|
|Author:||Goharpey, F.; Nazockdast, H.; Katbab, A.A.|
|Publication:||Polymer Engineering and Science|
|Date:||Jan 1, 2005|
|Previous Article:||Spherulitic crystallization behavior of linear low-density polyethylene.|
|Next Article:||Comparison of four methods for monitoring the kinetics of curing of a phenolic resin.|